Lecture Notes in Physics

Gianguido Dall’Agata
Marco Zagermann

Supergravity
From First Principles to Modern
Applications
Lecture Notes in Physics
Founding Editors
Wolf Beiglböck, Heidelberg, Germany
Jürgen Ehlers, Potsdam, Germany
Klaus Hepp, Zürich, Switzerland
Hans-Arwed Weidenmüller, Heidelberg, Germany

Volume 991

Series Editors
Roberta Citro, Salerno, Italy
Peter Hänggi, Augsburg, Germany
Morten Hjorth-Jensen, Oslo, Norway
Maciej Lewenstein, Barcelona, Spain
Angel Rubio, Hamburg, Germany
Wolfgang Schleich, Ulm, Germany
Stefan Theisen, Potsdam, Germany
James D. Wells, Ann Arbor, MI, USA
Gary P. Zank, Huntsville, AL, 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
non-specialist 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
responsible editor at Springer:

Dr Lisa Scalone
Springer Nature
Physics
Tiergartenstrasse 17
69121 Heidelberg, Germany
[email protected]

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

Gianguido Dall’Agata • Marco Zagermann

Supergravity
From First Principles to Modern
Applications
Gianguido Dall’Agata Marco Zagermann
Physics and Astronomy Hamburg, Germany
University of Padua
Padua, Italy

ISSN 0075-8450 ISSN 1616-6361 (electronic)

Lecture Notes in Physics
ISBN 978-3-662-63978-8 ISBN 978-3-662-63980-1 (eBook)
https://doi.org/10.1007/978-3-662-63980-1
© Springer-Verlag GmbH Germany, part of Springer Nature 2021
This work is subject to copyright. All 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, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer-Verlag GmbH, DE part of
Springer Nature.
The registered company address is: Heidelberger Platz 3, 14197 Berlin, Germany
Preface

The idea of supersymmetries between bosons and fermions continues to be a major

driving force behind many interesting developments in contemporary high-energy
and mathematical physics. This statement applies in particular to supergravity,
which may be described either as the result of supersymmetrizing Einstein’s general
theory of relativity or, equivalently, as the result of turning supersymmetry into a
local gauge symmetry.
Given the success of gauge symmetries in the Standard Model, it may seem rather
natural to treat also supersymmetry as a local gauge symmetry. On the other hand,
gravitational interactions between particles are notoriously irrelevant at accessible
collider energies. One might therefore be concerned that the difference between
global supersymmetry and supergravity could only affect Planckian energy scales
and hence be without measurable implications for present-day phenomenology. This
naive expectation, however, is in general not true, because many supergravity effects
also involve the scale of supersymmetry breaking, which may be much lower than
the Planck scale. In fact, supergravity may be directly relevant for issues such as
supersymmetry breaking, dark matter, dark energy, Big Bang Nucleosynthesis and
inflation, i.e. for topics of strong present interest.
From the viewpoint of unified frameworks of all interactions such as string
theory, the inclusion of gravity is of course not just an option, but a necessity. In
fact, the low-energy limits of all supersymmetric string theories are supergravity
theories. Due to their inherently non-linear structure and the protective properties of
supersymmetry, the supergravity approximations to string theory are often the only
available tools for probing the non-perturbative structure of string theory.
In spite of all these applications, many researchers who use aspects of supergrav-
ity in their daily work never really studied the theoretical origins of these aspects or
the derivations of equations they have encountered many times. This is largely due
to supergravity’s reputation of being an extremely complicated subject with only
very few pedagogically suitable references.
Part of this perception is not completely wrong: compared to the literature on
global supersymmetry or string theory, there are much fewer introductory texts
on supergravity that treat the subject in a sufficiently deep but comprehensible
manner. Admittedly, some supergravity Lagrangians may at first sight also look
quite intimidating, and browsing through this book will certainly reveal a few

v
vi Preface

examples. But the reader should be assured that this intimidation goes away rather
quickly once one has learned where to focus on in these equations.
Another part of the frustration many students experience when they try to learn
supergravity is that many approaches use advanced techniques such as superfields
on curved superspace or superconformal tensor calculus, which are powerful
computational tools, but require a dedicated study of these techniques before the
really interesting stuff can be extracted. It is also fair to say that the usefulness
of some of these techniques is often greater for global supersymmetry than for
supergravity, or may, in fact, not even be practically available, as is the case for
many extended or higher-dimensional supergravity theories.
Especially for particle phenomenologists, an additional obstacle is that many
texts on supersymmetry work with the two-component spinor notation instead of
the more familiar four-component spinor formalism.
Lowering these obstacles towards a deeper understanding of supergravity and
explaining also those things that are usually not explained or even overlooked were
precisely the central motivations for writing this book.
In order to reach this goal, we completely avoid the introduction of a technical
machinery that involves unphysical fields, auxiliary symmetries or other artificial
objects. Instead, we work directly with the physical on-shell field content and use
straightforward and explicit computations as well as simple geometrical reasoning
to arrive at the results. Two major themes of this book are also the differences
between supergravity and global supersymmetry and the physical and mathematical
consequences of these differences.
Throughout this book, we use four-component spinors instead of two-component
spinors. This formalism should make the text more accessible to phenomenologi-
cally oriented readers and has the advantage that it can be easily extended to other
spacetime dimensions.
This book is suited for beginning graduate or advanced undergraduate students in
high-energy physics or mathematical physics, as well as for researchers working in
these or related areas. It assumes familiarity with basic notions of general relativity,
differential geometry and global supersymmetry.
The book is divided into three parts:
Part I discusses foundational material such as our spinor conventions in Chap. 1,
the transition from global to local supersymmetry in Chap. 2, as well as spinors
in curved spacetime in Chap. 3. It culminates in the detailed discussion of pure
supergravity with and without a cosmological constant in Chap. 4.
Part II is devoted to the couplings of matter fields in global supersymmetry
(Chap. 5) and supergravity (Chap. 6) and discusses the phenomenological conse-
quences for particle physics and cosmology in Chap. 7.
In Part III, three more formal topics are introduced: extended supergravity
in Chap. 8, gauged supergravity in Chap. 9 and supergravity in higher spacetime
dimensions in Chap. 10.
Preface vii

Most chapters contain some exercises to illustrate special cases, discuss some
extensions or prove statements from the main text. Moreover, an incomplete
collection of references is given at the end of each chapter, where we refer to some
original papers, review articles or textbooks we found useful for preparing this text.

Padova, Italy Gianguido Dall’Agata

Hamburg, Germany Marco Zagermann
29 June 2021
Acknowledgements

Over the years, we have benefitted much from the interaction with many colleagues
on subjects related to this book, and it is a great pleasure to thank them all for the
numerous fruitful discussions and collaborations. We feel guilty in not being able to
list them all, but we do have to give special thanks to those colleagues who played a
pivotal role in shaping our understanding of supergravity: Anna Ceresole, Riccardo
D’Auria, Sergio Ferrara, Murat Günaydin, Antoine Van Proeyen and Fabio Zwirner.
Without their support and their willingness to share with us so many of their deep
insights into this subject, this book would have never been written.
This book grew out of introductory lectures on supergravity and closely related
topics presented by G. Dall’Agata at SISSA (a.y. 2006/07–2009/10), Scuola
Normale Superiore di Pisa (2011, 2017–2021), Padova Uni. (2008/09,2011/12–
2012/13, 2016/17–2017/18), EPFL Lausanne (2009), GGI, Florence (LACES 2009)
and at SEPnet (Queen Mary U. London, Univ. of Southampton and Royal Hol-
loway U. London, 2010) and by M. Zagermann at LMU Munich (2006/07),
Sogang University, Seoul (2007), MPI for Gravitational Physics, Potsdam (“String
Steilkurs” 2005 and 2008), MPI for Physics, Munich (2009), Leibniz Universität
Hannover (2009/10), DESY, Hamburg (2012) and GGI, Florence (LACES 2019).
We thank C. Angelantonj, M. Bertolini, D. Cassani, O. Lechtenfeld, J. Louis, D.
Lüst, J.-H. Park, R. Russo, A. Sagnotti, C. Scrucca, F. Steffen and S. We thank
Theisen for inviting us to these lectures and for encouraging us to give such
courses. It is a pleasure to thank all the participants of these courses for their
countless insightful comments and questions that greatly helped us to improve the
presentation, especially Johannes Brödel, Daniel Junghans, Stefano Lanza, Enrico
Pajer, Dario Partipilo, Erik Plauschinn, Timm Wrase and Matteo Zatti. Moreover,
we thank Vincent Bettaque, Niccolò Cribiori, Gianluca Inverso and Karl Kortum for
their careful reading of parts of the manuscript and their valuable feedback.
We are indebted to Christian Caron and Lisa Scalone from Springer for their
helpful support over the years and their remarkable patience.
Finally, we thank Luis Álvarez Gaumé, who first suggested to turn our notes into
a book. We hope that the published version embodies the hands-on and physics first
approach that he liked in our original presentation.

ix
Contents

Part I Foundations and Pure Supergravity

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 3
1.1 The Many Facets of Supergravity .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 4
1.2 Plan of the Lectures.. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 7
1.3 A Quick Guide Through Our Spinor Conventions . . . . . . . . . . . . . . . . . 8
1.3.1 Two-Component Spinors.. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 9
1.3.2 Four-Component Spinors . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 10
1.3.3 Susy Algebra in Four Dimensions .. . .. . . . . . . . . . . . . . . . . . . . 18
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 18
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 19
2 From Global to Local Supersymmetry . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 21
2.1 Promoting Supersymmetry to a Local Symmetry .. . . . . . . . . . . . . . . . . 21
2.2 The Gravitino . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 25
2.2.1 The Gravitino Action .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 25
2.2.2 The Gravitino Multiplet .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 29
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 30
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 30
3 Gravity and Spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 31
3.1 The Standard Metric Formulation . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 31
3.2 The Vielbein Basis and Cartan’s Formalism . . .. . . . . . . . . . . . . . . . . . . . 33
3.3 Spinors in Curved Spacetime . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 41
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 42
4 Pure N = 1 Supergravity in Four Dimensions . . . . .. . . . . . . . . . . . . . . . . . . . 43
4.1 Pure Supergravity: The Action and SUSY Rules . . . . . . . . . . . . . . . . . . 43
4.1.1 Second-Order Formalism . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 49
4.1.2 First-Order Formalism . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 50
4.1.3 1.5-Order Formalism .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 50
4.2 Adding a Cosmological Constant .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 51
4.2.1 Construction of the Action . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 55
4.2.2 Mass in AdS . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 60

xi
xii Contents

4.A Appendix: Gauging the Poincaré Algebra .. . . . .. . . . . . . . . . . . . . . . . . . . 65

4.A.1 Gauging the Super Poincaré Algebra .. . . . . . . . . . . . . . . . . . . . 69
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 69
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 70

Part II Matter Couplings and Phenomenology

5 Matter Couplings in Global Supersymmetry . . . . . . .. . . . . . . . . . . . . . . . . . . . 73

5.1 Our Approach .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 73
5.2 Chiral Multiplets in Global Supersymmetry . . .. . . . . . . . . . . . . . . . . . . . 74
5.2.1 The Renormalizable Wess–Zumino Model . . . . . . . . . . . . . . 74
5.2.2 Non-linear Sigma Models I: The Holonomy Group . . . . . 76
5.2.3 Non-linear Sigma Models II: Fermions
and Supersymmetry . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 78
5.2.4 4D Supersymmetry and Kähler Manifolds .. . . . . . . . . . . . . . 80
5.3 Globally Supersymmetric Gauge Theories .. . . .. . . . . . . . . . . . . . . . . . . . 85
5.3.1 Super Maxwell Theory .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 85
5.3.2 Super Yang-Mills Theory . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 85
5.3.3 Coupling Super Maxwell/Yang–Mills Theories to
Chiral Multiplets . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 86
5.3.4 Non-minimal Kinetic Terms for Vector Multiplets:
The Gauge Kinetic Function .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . 86
5.3.5 Non-linear σ -Models III: Global and Local
Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 87
5.3.6 Killing Prepotentials, D-Terms, and the General
Globally Supersymmetric Lagrangian .. . . . . . . . . . . . . . . . . . . 90
5.4 Supersymmetry Breaking . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 98
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 100
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 101
6 Matter Couplings in Supergravity .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 103
6.1 New Supergravity Couplings.. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 103
6.1.1 Coupling Chiral Multiplets to Supergravity .. . . . . . . . . . . . . 104
6.1.2 The Kähler Covariant Derivative . . . . .. . . . . . . . . . . . . . . . . . . . 105
6.1.3 Additional Bare Superpotential Terms . . . . . . . . . . . . . . . . . . . 109
6.1.4 Inclusion of Vector Multiplets . . . . . . . .. . . . . . . . . . . . . . . . . . . . 113
6.1.5 More on D-Terms . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 115
6.1.6 The Gradient Flow Relations . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 117
6.1.7 Final Remarks . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 119
6.2 Kähler–Hodge Manifolds . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 122
6.2.1 An Example: Quantization of Newton’s Constant .. . . . . . 124
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 126
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 127
Contents xiii

7 Phenomenological Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 129

7.1 Spontaneous Supersymmetry Breaking.. . . . . . . .. . . . . . . . . . . . . . . . . . . . 129
7.1.1 Vacua .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 129
7.1.2 General Features of Spontaneous Supersymmetry
Breaking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 131
7.1.3 Mass Scales Related to Supersymmetry Breaking .. . . . . . 134
7.1.4 Rigid Limits . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 136
7.2 Gravitino, Goldstino, and Super-Higgs Mechanism . . . . . . . . . . . . . . . 137
7.2.1 The Goldstino in Global Supersymmetry . . . . . . . . . . . . . . . . 138
7.2.2 The Goldstino and the Gravitino in Supergravity .. . . . . . . 140
7.2.3 Gravitino Couplings .. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 144
7.2.4 Generalizations . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 147
7.3 Mass Sum Rules and Mediation Mechanisms ... . . . . . . . . . . . . . . . . . . . 148
7.3.1 Mass Sum Rules, Hidden Sectors, and Mediation
Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 148
7.3.2 Gravity-Mediated Supersymmetry Breaking and
the Polonyi Model .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 150
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation . . . . . . . . . . . . . . . 155
7.4.1 Moduli Stabilization and Moduli Masses . . . . . . . . . . . . . . . . 155
7.4.2 No Scale Models . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 161
7.4.3 Dark Energy and de Sitter Vacua . . . . .. . . . . . . . . . . . . . . . . . . . 165
7.4.4 Inflation and the Supergravity η-Problem .. . . . . . . . . . . . . . . 167
7.A Appendix: Proof of Eq. (7.114) . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 169
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 171
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 172

Part III Extended, Gauged and Higher-Dimensional Supergravity

8 Extended Supergravities .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 177

8.1 Electric–Magnetic Duality . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 177
8.2 N = 2 Supergravity . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 184
8.2.1 N = 2 Vector Multiplets and Special Kähler
Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 185
8.2.2 N = 2 Hypermultiplets and Hyper-Kähler vs.
Quaternionic Kähler Geometry . . . . . . .. . . . . . . . . . . . . . . . . . . . 196
8.3 Extended Supergravity with N ≥ 3 .. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 201
8.A Appendix: Details on and Origin of Local Special Kähler
Geometry .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 205
8.A.1 The Symplectic Section and Its Kähler
Transformation . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 205
8.A.2 The Kähler Potential . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 208
8.A.3 The Existence of a Prepotential .. . . . . .. . . . . . . . . . . . . . . . . . . . 209
8.A.4 The Gauge Kinetic Matrix . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 211
xiv Contents

8.B Appendix: Quaternionic-Kähler vs. Hyper-Kähler

Manifolds of Hypermultiplets .. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 214
8.B.1 Sp(nH ) × SU (2)-Adapted Vielbein .. . . . . . . . . . . . . . . . . . . . 215
8.B.2 Holonomy and Curvature . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 217
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 222
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 222
9 Gauged Supergravity .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 223
9.1 Supergravities and Scalar Potentials .. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 224
9.2 Duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 227
9.2.1 From Electric–Magnetic Duality to U -Duality . . . . . . . . . . 228
9.3 Gauging and Symplectic Frames. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 229
9.4 Coset Manifolds and Gauging . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 231
9.4.1 Vielbein, Metric, and Isometries of G/H . . . . . . . . . . . . . . . . . 235
9.4.2 The Special-Kähler Manifold SU(1,1)/U(1) and
Inequivalent Symplectic Frames .. . . . .. . . . . . . . . . . . . . . . . . . . 238
9.5 Gauging and the Embedding Tensor .. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 241
9.5.1 Constraints on the Embedding Tensor .. . . . . . . . . . . . . . . . . . . 243
9.5.2 Couplings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 244
9.5.3 An Example: The Maximal Theory . .. . . . . . . . . . . . . . . . . . . . 249
9.6 Classifying Gaugings .. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 253
9.6.1 The Quotient Space S . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 254
9.6.2 The S Space of SO(8) Maximal Gauged
Supergravity . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 258
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 262
10 Supergravity in Arbitrary Dimensions . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 263
10.1 Higher-Dimensional Theories .. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 263
10.2 Example: D = 11 Supergravity .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 265
10.3 Dimensional Reduction and Ten-Dimensional Supergravities . . . . 267
10.4 Dimensional Reduction and the Origin of Gauged
Supergravities .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 272
10.5 Example: D = 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 275
10.5.1 N = 2 in 5D . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 276
10.5.2 N = 4 in 5D . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 279
10.5.3 N = 6 in 5D . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 279
10.5.4 N = 8 in 5D . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 280
10.5.5 Gaugings and Tensor Fields . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 280
10.A Appendix: Clifford Algebras and Spinors in Arbitrary D . . . . . . . . . 284
10.A.1 Irreducible Representations of Cliff(1, D − 1) . . . . . . . . . . 284
10.A.2 Irreducible Spinor Representations
of SO0 (1, D − 1) . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 285
10.A.3 Majorana and Weyl Condition .. . . . . . .. . . . . . . . . . . . . . . . . . . . 289
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 290
References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 290

Index . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 293
 Part I
Foundations and Pure Supergravity
Introduction
1

The quest for a fundamental theory of all elementary particles and their interactions
has been one of the most fascinating scientific endeavors during the past century.
One of the main guiding principles in the construction of the ever more refined
theories of high energy physics has been the systematic use of symmetry principles.
They form the basic language in terms of which the Poincaré invariant quantum field
theory based on the gauge group SU(3) × SU(2) × U(1) we now call the Standard
Model (SM) is formulated.
Among the many interesting ideas for physics beyond the Standard Model,
the most fruitful one has arguably been the introduction of the concept of super-
symmetry, i.e., a symmetry between bosonic and fermionic degrees of freedom.
Indeed, supersymmetric extensions of the Standard Model have been put forward
as possible solutions to the hierarchy problem, though the simplest models are
challenged by current Large Hadron Collider results; they improve the unification of
the three Standard Model gauge couplings at the Grand Unified Theory (GUT) scale,
MGUT ∼ = 2 × 1016GeV; and they might provide interesting dark matter candidates.
The one missing major player in these constructions, however, is the gravitational
interaction with its elegant geometric description in terms of Einstein’s general
theory of relativity. Combining the principles of supersymmetry with gravity defines
what is called supergravity, the topic of these lecture notes. As we now briefly
explain, the idea of supergravity has many other intriguing applications in various
areas of particle physics, cosmology, string theory, and mathematics that go far
beyond the simple desire to marry supersymmetry with gravity and can serve as
separate motivations to study supergravity theories.

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 3

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_1
4 1 Introduction

1.1 The Many Facets of Supergravity

Supergravity as a Gauge Theory of Supersymmetry

Supersymmetric field theories in rigid Minkowski spacetime feature supersymmetry
as a global (or “rigid”) symmetry. In view of the success of local gauge invariances
in the Standard Model, it is natural to try to promote supersymmetry likewise to a
local gauge symmetry. This leads directly to supergravity, and in fact this was the
way supergravity was first constructed. Supergravity theories can hence equivalently
be defined as the gauge theories of supersymmetry.

Just as for an ordinary Yang–Mills symmetry, the gauging of supersymmetry

requires the introduction of a suitable gauge field that transforms into the spacetime
derivative of the infinitesimal symmetry parameter:

δgauge (gauge field) = ∂μ (gauge parameter) + . . . (1.1)

In supersymmetry, the symmetry parameter is a spinorial quantity, α , with α being

a spinor index. The supersymmetry gauge field is therefore not a vector field as
in ordinary Yang–Mills theories but a vector-spinor field, ψμα . On-shell (and for
unbroken supersymmetry), this field propagates two helicity ±3/2 states, whereas
off-shell it also contains states with helicity ±1/2 (just as an ordinary gauge field,
Aμ , contains helicities ±1 on-shell, but off-shell also two states with helicity 0).
We will see later that supersymmetry further implies that the graviton, gμν , and
the supersymmetry gauge field, ψμα , sit in the same supersymmetry multiplet, the
supergravity multiplet, so that local realizations of supersymmetry necessarily need
to include gravity, which explains the name supergravity. Being the superpartner of
the graviton, ψμα is called the gravitino.

Extended Supergravity and Unification

In extended supergravity theories with N ≥ 2 supersymmetries, the supermultiplet
of the graviton contains N gravitini but also fields with spin ≤ 1. Extended
supersymmetry can thus interpolate between the graviton and ordinary gauge fields,
thereby leading to some sort of unification of the interactions mediated by these
fields. For N > 2, the supergravity multiplet also contains spin 1/2 particles, and
one may wonder whether sufficiently extended supergravity could even provide a
unified theory of all interactions and matter particles. The attempts in this direction
culminated in the construction of the maximally extended N = 8 supergravity
with the maximally possible compact gauge group SO(8) by de Wit and Nicolai
in 1982 [1]. Unfortunately, this and other extended supergravity theories generally
suffer from the problem that the corresponding gauge interactions are non-chiral, in
contrast to what we see in the electroweak sector of the Standard Model. Moreover,
the SO(8) gauge group studied in [1] is too small to accommodate the Standard
Model gauge group SU(3) × SU(2) × U(1) as a subgroup, and the theory also
has a very large negative cosmological constant. Although unification via extended
1.1 The Many Facets of Supergravity 5

supergravity theories has not met with phenomenological success, the study of many
extended supergravity theories in the early 1980s has proven to be a very valuable
resource for many modern theoretical developments, notably in string theory; see
below.

Better Behavior of Ultraviolet Divergencies

Another interesting motivation for studying supergravity is the better behavior of
ultraviolet divergencies of its corresponding quantum theory in comparison with
ordinary general relativity (GR). In the quantization of GR, one can remove the
potential on-shell one-loop divergencies by field redefinitions. However, there is
an ultraviolet (UV) divergence at two loops which cannot be removed [2, 3], and
if matter fields are added, divergencies can appear already at one loop. However,
if the matter content is consistent with supersymmetry, the situation improves
again. It is even still under debate at the time of this writing whether maximally
supersymmetric gravity in four dimensions is divergent or not. Recent calculations
show that divergencies in N = 8 supergravity do not show up before five loops [4]
and current consensus is that the first counterterm appears at seven loops, though it
is still possible that its coefficient is vanishing [5]. The study of the UV properties
of N = 8 supergravity and its connections with amplitudes in N = 4 super Yang–
Mills theory is presently a very active research area, and new techniques developed
for their computation are now used in a much wider context.

Phenomenology
Whereas the lack of chiral gauge interactions precludes any direct use of extended
supersymmetry and supergravity for phenomenological applications, N = 1
supergravity theories are phenomenologically very interesting and could resolve
some issues in globally supersymmetric extensions of the Standard Model. For
instance:

• Supergravity can remove the large tree-level cosmological constant of sponta-

neously broken rigid supersymmetry.
• Supergravity suggests new mechanisms of supersymmetry breaking and their
transmission to the Standard Model sector.
• When supersymmetry is broken in supergravity, the goldstino is “eaten” by the
gravitino, which provides one way of explaining why the goldstino has not been
seen.

Supergravity can also have important implications for the dark matter sector or early
universe cosmology, as we will also explain later.

Supergravity as an Effective Theory of String Theory

While supergravity theories are in general not renormalizable, they do arise as the
low-energy effective actions of (super)string theory, the best viable candidate for a
consistent unified quantum theory of gravity and all gauge interactions. From this
point of view, supergravity forms the interface between string theory and most of
6 1 Introduction

its potentially observable low-energy phenomena, with its non-renormalizability

no longer being an issue. Moreover, even as the infrared limit of string theory,
supergravity with all its non-linear interactions still captures some of the non-
perturbative properties of string theory, which would hardly be accessible via the
conventional world sheet conformal field theory approach. For instance, string
theory is not only a theory of strings but also contains other extended objects
such as D-branes or other solitonic p-branes. Some of these objects first arose as
solutions of supergravity models, and the supergravity perspective has often given
interesting insights into their physics. Similarly, various duality symmetries have
been first, or better, understood by looking at the effective supergravity theories and
their solutions. Often when different compactifications lead to the same low-energy
spectra and vacua, one may uncover some new underlying symmetry of the higher
dimensional theories.

The Gauge/Gravity Correspondence

One of the biggest revolutions in our understanding of string theory came from
the observation that certain string models on curved spaces can be dual to non-
gravitating (often conformal) gauge field theories: this is the gauge/gravity cor-
respondence or anti-de Sitter/conformal field theory (AdS/CFT) correspondence
[6, 7]. The weak coupling limit of string models, i.e., supergravity, is then an
important tool to compute quantities that are related to the strong coupling regime
in the dual field theory. A special role in this correspondence is played by
gauged supergravity,1 because global symmetries in the CFT become local in the
corresponding supergravity model.

Geometry
Supergravity solutions that preserve some of the supersymmetry of the action are
particularly interesting also from a mathematical point of view. For one thing
this is due to the circumstance that supersymmetric solutions often satisfy first-
order differential equations, which in general are much easier to solve than the
usual second-order field equations of non-supersymmetric solutions. These first-
order differential equations define Killing spinors, which prove to be a powerful
concept in different areas of mathematics. The supersymmetric compactification
backgrounds of string or M-theory, for example, define compactification manifolds
with interesting mathematical structures such as restricted holonomy groups or spe-
cial types of gauge bundles, which are consequences of the corresponding Killing
spinor equations. But also the spaces of fluctuations about such compactification
backgrounds carry non-trivial geometric structures that result in various interesting
scalar field geometries or “moduli spaces” in the dimensionally reduced field
theories, such as, e.g., Kähler, special Kähler, or quaternionic Kähler geometries, or
various types of coset spaces (see later chapters). Various cascades of dimensional

1Gauged supergravity refers to supergravity theories that contain also non-trivial conventional
gauge interactions, as we will see in Chap. 9.
1.2 Plan of the Lectures 7

reductions then reveal unexpected mathematical relations between different classes

of restricted geometries that would have been difficult to obtain or even suspect
otherwise.

Fake Supergravity
As a final motivation, we would like to mention the concept of fake supergravity
[8, 9]. Fake supergravity describes classes of solutions (e.g., domain walls or
black holes) of non-supersymmetric gravity theories for which the second-order
field equations can be rewritten in a first-order form that resembles the Killing
spinor equations of genuine supergravity theories. This is especially useful for
the discussion of the stability of these gravity solutions, as it allows the use of
the Nester–Witten argument [10] in a large class of theories without the usual
consistency requirements and limitations imposed by actual supersymmetry. In this
same framework, also cosmological solutions may be under better control [11].

1.2 Plan of the Lectures

Clearly, the present lecture notes cannot cover all the above topics and applications
in full detail. Instead, their purpose is to give a survey of the basic ingredients needed
to construct a supergravity action and to discuss its physical implications (mainly
for particle physics), without introducing too much technical formalism or spending
too much time on the many possible applications. Another focus will be on the
differences between globally supersymmetric theories and supergravity, so that the
reader may better understand what really needs supergravity and what can already
be implemented in a globally supersymmetric theory without gravity.
More explicitly, the outline of the lecture notes is as follows: In the first part,
which consists of Chaps. 1–4, we work our way toward the construction of the
simplest possible supergravity theory in four dimensions, namely, pure 4D, N = 1
supergravity. To this end, we first introduce our spinor conventions in the remainder
of Chap. 1 and then explain, in Chap. 2, how the requirement of local supersymmetry
naturally leads to the inclusion of the graviton supermultiplet. In Chap. 2, we also
motivate the form of the supergravity action and the supersymmetry transformation
laws. In order to write these down properly, we review spinors in curved spacetime
and the vierbein formulation of general relativity in Chap. 3. Chapter 4 then is
devoted to the detailed discussion of pure 4D, N = 1 supergravity with and
without a cosmological constant and gives the complete proof of its invariance under
local supersymmetry. In these first chapters, we work out all the details and often
show pedantically all the steps necessary for this construction. The serious reader is
recommended to go through this material in full detail, as it greatly contributes to
developing a good intuition that will also be helpful for the remaining chapters.
In the second part of the book, we describe how the minimal supergravity
sector discussed in Part I can be coupled to matter multiplets and what the
implications of these matter couplings for phenomenology are. To this end, we
first discuss matter couplings in global supersymmetry in Chap. 5 and then explain,
8 1 Introduction

in Chap. 6, the differences that arise when supergravity and local supersymmetry
are introduced. Various consequences of supergravity for the phenomenology of
particle physics and cosmology are explored in Chap. 7. In these chapters, we
emphasize in particular the differences between rigid and local supersymmetry, both
at a formal/mathematical and at a physical level.
In the third part of the book, the presentation will be more formal, and we will
discuss a number of more advanced topics. In Chap. 8, we introduce the concept of
electric-magnetic duality and show how its interplay with the R-symmetry group
essentially fixes the geometrical structures encountered in models with extended
supersymmetry, detailing especially the case of N = 2 theories. We then give, in
Chap. 9, a brief but modern introduction to gauged supergravity models, which are
playing a prominent role in many interesting recent developments in string theory.
Special emphasis will be given here to the case of N = 8 gauged supergravity.
Finally, we conclude with some remarks on higher-dimensional theories and the
relations between these models and four-dimensional gauged supergravities in
Chap. 10.
We used a number of references throughout the lectures, mainly to point to
some additional sources of information on the discussed topics. There are obviously
already many very good reviews on various aspects of supergravity theories, which
we used as inspiration to prepare these lectures; some of them are [12–29]. Also,
very good complementary recent references on supergravity are [30, 31].

1.3 A Quick Guide Through Our Spinor Conventions

Before we embark on our journey through the world of supergravity theories, we

briefly summarize the 4D spinor conventions used in this book.2 For a generic
treatment of spinors in any spacetime dimension, see appendix 10.A. Very good
references are also [18, 32].
Assuming an orthonormal set of basis vectors, ea , of Minkowski space
(a, b, . . . = 0, 1, 2, 3), the Minkowski metric with our choice of signature is

ηab = diag(−1, +1, +1, +1). (1.2)

The generators of Lorentz transformations are denoted by Mab = −Mba when they
are taken as anti-Hermitian generators or as Mab = −iMab when they are taken as
Hermitian generators and satisfy the Lorentz algebra so(1, 3),

[Mab , Mcd ] = −2 ηc[a Mb]d + 2 ηd[a Mb]c (1.3)

2 Throughout the book, it is understood that c = 1 and h̄ = 1.

1.3 A Quick Guide Through Our Spinor Conventions 9

In these notes, symmetrization () and antisymmetrization [] are always taken

with weight one, i.e., (ab) = 1/2(ab + ba), [ab] = 1/2(ab − ba), etc.

A spinor representation is a representation of the above Lorentz algebra that

does not integrate to an ordinary (i.e., “single-valued”) representation of the
corresponding Lorentz group. Instead, it gives rise only to a “double-valued”
representation of the Lorentz group in the sense that spatial rotations by 2π give
minus the identity.
Mathematically, this is possible because the Lorentz group3 is not simply
connected but contains closed loops that cannot be continuously contracted to a
point. The universal covering group of the Lorentz group is a group that is locally
isomorphic to the Lorentz group but with a different global structure such that all
closed curves are fully contractible. Spinor representations are then equivalently
described as single-valued representations of this universal covering group that
project to double-valued representations of the Lorentz group itself. The universal
covering group of the 4D Lorentz group happens to be isomorphic to SL(2, C), the
group of unimodular complex (2 × 2) matrices.
In four dimensions, there are two commonly used notations for spinor represen-
tations: the two-component spinor notation and the four-component spinor notation.
The two-component spinor notation is based on the abovementioned accidental
isomorphism between the universal cover of the 4D Lorentz group and SL(2, C),
which does not have a direct analogue in a generic spacetime dimension. The
four-component spinor notation, by contrast, readily generalizes to any spacetime
dimension and is based on the notion of Clifford algebras and gamma matrices.
In this book, we will use the four-component spinor formalism, because it is
more common in the supergravity literature. The two-component spinor notation,
on the other hand, is frequently used in texts on global supersymmetry, so we briefly
include it here as well to facilitate the translation of one formalism into the other.

1.3.1 Two-Component Spinors

The group SL(2, C) has two equivalence classes of irreducible two-dimensional

complex representations, corresponding to the defining representation of SL(2, C)
and its complex conjugate representation. These two representations are the minimal
spinor representations of the 4D Lorentz group and are often denoted by (1/2, 0) and
(0, 1/2) or by dotted and undotted two-component spinors, λA and χȦ∗ . The Lorentz

3 The full isometry group O(1, 3) of 4D Minkowski spacetime decomposes into four disconnected
components. In this book, we mean by Lorentz group only the component, SO(1, 3)0 , of Lorentz
transformations that are continuously connected to the identity element. This subgroup is often
called the “proper orthochronous Lorentz group”.
10 1 Introduction

group generators, Mab , act on these via representation matrices that can be chosen
as
1
(1/2, 0) : ρ(Mab ) = − (σa σ b − σb σ a ) (1.4)
4
 
1
(0, 1/2) : ρ (Mab ) = e − (σ a σb − σ b σa ) e−1 ,
∗
(1.5)
4

where σi = −σ i (i = 1, 2, 3) are the Pauli matrices, σ0 = σ 0 = 12 , and

 
0 1
e≡ . (1.6)
−1 0

That (1.5) is indeed the complex conjugate of (1.4) follows from the identity
e−1 σ μ e = σ μ∗ = σ μT .
The matrix e is an invariant of SL(2, C) in the sense that

AT eA = e ∀A ∈ SL(2, C), (1.7)

which implies that quantities such as λT eω are Lorentz invariant products of two
spinors λA and ωA . This is often written as λA ωA or λA ωA using suitable raising
and lowering conventions for spinor indices with the matrix e.
All finite-dimensional irreducible representations of SL(2, C) can be obtained
from symmetrized tensor products of these elementary building blocks, often
denoted by (n/2, m/2), where n and m count the (1/2, 0) and (0, 1/2) factors,
respectively. For (m + n) =even/odd, these describe single-/double-valued repre-
sentations of the Lorentz group, corresponding to bosons/fermions.

1.3.2 Four-Component Spinors

The above two-component spinor formalism has many nice features and, as already
mentioned, is frequently used in the literature on global supersymmetry. In these
lectures, however, we will use four-component spinors instead, as these are more
standard in the four-dimensional supergravity and particle phenomenology literature
and generalize easily to other spacetime dimensions.
More precisely, we will use four-component Majorana spinors to describe all
fermions in 4D. Majorana spinors satisfy a reality condition (see below) and natu-
rally describe fermions that are gauge invariant or transform in real representations
of the gauge group. This is true in particular for the gauge-invariant gravitino,4
the superpartner of the graviton, but also for the gaugini, the superpartners of the

4As we will discuss later, an exception to this gauge invariance of the gravitino arises in the
presence of Fayet–Iliopoulos terms in N = 1 supergravity. These terms require a gauging of
1.3 A Quick Guide Through Our Spinor Conventions 11

gauge bosons, which (before spontaneous symmetry breaking) always transform in

the adjoint and hence a real representation of the gauge group. Fermions that sit in
N = 1 chiral multiplets, on the other hand, may have chiral gauge interactions and
hence may sit in complex gauge group representations. While these fermions can
be nicely described by two-component spinors, they may equally well be described
in terms of the chiral components of different sets of four-component Majorana
spinors, as we will do in this book.

1.3.2.1 Dirac Spinors

A general four-component Dirac spinor, ψα (α = 1, . . . , 4), has four complex
components and corresponds to a direct sum of two two-component spinors of the
form (1/2, 0) ⊕ (0, 1/2), i.e., it combines one undotted two-component spinor, λA ,
and one dotted spinor, χȦ∗ . A very convenient way to do so is to define
 
e · χ∗
ψ= , (1.8)
λ

so that the Lorentz generators are represented by the manifestly reducible matrices
 
eρ ∗ (Mab )e−1 0
Σab = . (1.9)
0 ρ(Mab )

The (4 × 4) matrices Σab can conveniently be expressed in terms of Dirac gamma

matrices,
 
0 iσ a
γa = , (1.10)
iσa 0

so that
1
Σab = [γa , γb ]. (1.11)
4
The gamma matrices defined above satisfy the Clifford algebra

{γa , γb } = 2 ηab 14 . (1.12)

As one easily verifies, these Clifford algebra relations already imply the Lorentz
algebra commutation relations for the matrices Σab if one uses (1.11) as their
definition. From this point of view, the particular expression (1.9) is just a special
case that corresponds to the particular representation (1.10) of the Clifford algebra

the R-symmetry group, under which also the gravitino is charged, but these theories are often
anomalous.
12 1 Introduction

(1.12). This representation (1.10) is called the Weyl representation, and its main
virtue is that the Lorentz generators assume the block diagonal form (1.9), making
their reducibility manifest.5 However, there are infinitely many other irreducible
representations of (1.12) that one could also use; apart from the Weyl representation,
the most common representations are the Dirac representation and the Majorana
representation.6 Fortunately, all these representations are equivalent to the Weyl
representation, and we rarely will make use of particular representations. For
convenience, however, we will always assume “friendly” representations whose
defining properties are

γ0† = −γ0 (1.13)

γi† = +γi (1.14)

γaT = ±γa , (1.15)

where the last equation means that each gamma matrix is either symmetric or
antisymmetric. The Weyl, Dirac, and Majorana representations are obviously all
friendly representations.
An important object in the following will be the completely antisymmetrized
products of several gamma matrices,

γa1 ...ap ≡ γ[a1 γa2 . . . γap ] , (1.16)

where, as usual, the antisymmetrization involves a prefactor 1/p!, so that, e.g.,

γab = 1/2[γa , γb ], etc. Note that the Clifford relation (1.12) implies

γa1 γa2 . . . γap if all ai are different
γa1 a2 ...ap = (1.17)
0 otherwise

The 16 matrices γM = {14 , γa , γab , γabc , γabcd } are linearly independent and form
a basis of the complex (4 × 4) matrices.
As we have seen, an irreducible representation of the Clifford algebra (1.12)
induces a spinor representation (1.11) of the Lorentz group. This is true in
any spacetime dimension, which makes the Clifford algebra approach to spinor
representations so useful for higher-dimensional supergravity theories. As was
stressed in footnote 5, however, the Lorentz algebra representations obtained in

5 Note that the Weyl representation is irreducible as a representation of the Clifford algebra (1.12).

It is only the induced representation Σab of the Lorentz algebra

 that is reducible.

6 The Dirac representation corresponds to γ = iσ ⊗ 1 ≡ i 12 0
, γi = σ2 ⊗ σi , and the
0 3 2
0 −i 12
Majorana representation is given by γ0 = iσ2 ⊗ σ3 , γ1 = −σ1 ⊗ 12 , γ2 = σ2 ⊗ σ2 , γ3 = σ3 ⊗ 12 ,
but they are not really needed for this book.
1.3 A Quick Guide Through Our Spinor Conventions 13

this way are in general not irreducible, even if the underlying Clifford algebra
representation is irreducible.
For supersymmetry, it is convenient to work with minimal representations of
the Lorentz group, as these typically correspond also to the minimal amount of
supersymmetry one can have in the respective spacetime dimension. There are
essentially two ways to reduce the number of degrees of freedom of a Clifford alge-
bra spinor so as to obtain irreducible representations (irreps) of the corresponding
Lorentz algebra. One possibility is to impose a chirality condition, which leads to
Weyl spinors, and the other one is to impose a reality condition, which leads to
Majorana spinors.7 This is not always possible in every spacetime dimension: The
Weyl condition can only be imposed in even dimensions, whereas the possibility
to impose a Majorana condition shows a somewhat more complicated dependence
on the spacetime dimension, as we will see in Chap. 10. Moreover, the Weyl and
Majorana condition can often not be imposed simultaneously. For the moment, we
restrict ourselves to four spacetime dimensions, where one can impose a Weyl or a
Majorana condition, but not both of them at the same time.

1.3.2.2 The Weyl Condition

The Weyl condition projects out the part of a spinor that has a particular handedness.
It is imposed with the γ5 matrix:

γ5 ≡ γ 5 ≡ −iγ 0 γ 1 γ 2 γ 3 = +iγ0 γ1 γ2 γ3 . (1.18)

The Clifford algebra (1.12) implies

(γ5 )2 = 14 (1.19)
{γ5 , γa } = 0 ⇒ [γ5 , Σab ] = 0 (1.20)

so that the chirality projectors

1  1 
PL ≡ 1 + γ5 , PR ≡ 1 − γ5 , (1.21)
2 2
can be used to define left- and right-handed spinors.

ψL ≡ PL ψ, ψR ≡ PR ψ. (Weyl condition) (1.22)

7 Note that imposing a Weyl or Majorana condition in 4D does not necessitate the use of the Weyl or

the Majorana representation of the gamma matrices. The Weyl condition just takes on a particularly
simple form in the Weyl representation, and the Majorana condition leads to a particularly simple
result in the Majorana representation. We usually do not make use of these simplified forms and
write down the conditions in a covariant way.
14 1 Introduction

Because of (1.20), this projection is consistent with Lorentz covariance, and left-
and right-handed spinors form separate representations of the Lorentz group. In the
Weyl representation, γ5 = σ3 ⊗12 , i.e., the left- and right-handed spinors, are simply
the upper or lower two components of ψ as in (1.8). Note that γ 5 ψL = ψL , while
γ 5 ψR = −ψR .
In a friendly representation, γ0 γa† γ0 = γa ⇒ Σab †
γ0 = −γ0 Σab , and the Dirac
conjugate of a general four-component spinor is defined by

ψ ≡ iψ † γ 0 = −iψ † γ0 (1.23)

so that bilinears such as ψχ are Lorentz invariant.

In a friendly representation, we also have that γ5† = γ5 , and hence PL,R
†
= PL,R .
Moreover PL γ0 = γ0 PR and PR γ0 = γ0 PL , and this implies

ψR = ψPL , and ψL = ψPR , (1.24)

where

ψR ≡ (PR ψ) = −i(PR ψ)† γ0 . (1.25)

Because of this we will often write

ψ L ≡ ψPL = ψR , ψ R ≡ ψPR = ψL . (1.26)

The matrix γ5 also enters a number of very useful duality relations between the
antisymmetrized products of gamma matrices, as the reader is asked to verify in
Exercise 1.1.

1.3.2.3 The Majorana Condition

The Majorana condition is a reality condition that can be written as

ψ ∗ = Bψ, (Majorana condition) (1.27)

with some matrix B. This condition is self-consistent (i.e., ψ ∗∗ = ψ) and Lorentz

covariant if B satisfies

B ∗ B = 14 , γa∗ = Bγa B −1 ⇒ Σab

∗
B = BΣab . (1.28)

In the Weyl basis (1.10), one can choose

 
0 e
B= , (1.29)
−e 0
1.3 A Quick Guide Through Our Spinor Conventions 15

so that a Majorana spinor would be of the form (1.8), but with λ = χ. From this
we see that a Majorana spinor describes the same number of independent degrees of
freedom as a Weyl spinor.
Another equivalent, but for many purposes more convenient, way to write the
Majorana condition is via the charge conjugation matrix, C, which satisfies

C T = −C, γaT = −Cγa C −1 . (1.30)

In a friendly representation, one can moreover choose C such that it also satisfies

C −1 = −C = C † . (1.31)

In terms of C, the charge conjugate spinor is defined as

ψ c = Cψ = iCγ 0T ψ ∗
T
(1.32)

and a Majorana spinor is defined as

ψ c = ψ. (alternative form of Majorana condition) (1.33)

This is equivalent to (1.27) if we identify

B = (i Cγ0T )−1 , (1.34)

so that, in terms of B, charge conjugation reads

ψ c = B −1 ψ ∗ . (1.35)

The advantage of C is that for a Majorana spinor the Dirac conjugate can be
written as

ψ = ψ T C. (1.36)

Notice that using (1.30) one finds the symmetry properties

C T = −C, (Cγ abc )T = −(Cγ abc ), (Cγ abcd )T = −(Cγ abcd ),

(Cγ a )T = (Cγ a ), (Cγ ab )T = (Cγ ab ).

(1.37)

For anti-commuting Majorana spinors, this then implies

⎧
⎨ +ψ 2 Mψ1 for M = 14 , γabc , γabcd
ψ 1 Mψ2 = (1.38)
⎩
−ψ 2 Mψ1 for M = γa , γab
16 1 Introduction

Unless stated otherwise, we will, in the following, always use anti-commuting

Majorana spinors but often also take in addition the chiral projections ψL and ψR
of these Majorana spinors, which therefore are not independent.
More specifically, we have, in our conventions,

(ψL )c = ψR , (ψR )c = ψL . (1.39)

To show this, we use

B −1 γ5∗ = −γ5 B −1 ⇔ B −1 PL∗ = PR B −1 (1.40)

so that

(ψL )c = (PL ψ)c = B −1 PL∗ ψ ∗ = PR B −1 ψ ∗ = PR ψ c = PR ψ = ψR . (1.41)

Note that (1.39) implies that ψR is no longer a Majorana spinor, because that
would require (ψR )c being equal to ψR . Thus, in 4D, a four-component spinor
cannot be simultaneously chiral and Majorana. Nevertheless, it makes sense to talk
about the projection ψR or ψL of a given Majorana spinor ψ.
From the definition of γ5 and (1.30), one also gets

Cγ5 = γ5T C ⇔ CPL = PLT C (1.42)

so that for a Majorana spinor ψ,

ψ L = ψPL = ψ T CPL = ψ T PLT C = (ψL )T C (1.43)

even though ψL is not Majorana. From this we can obtain more symmetry properties
for the chiral projections that are very similar to those for the Majorana spinors
themselves,

χ L ψL = χLT CψL = −ψLT C T χL = ψ L χL ,

χ L γ a ψR = −ψ R γ a χL , χ L γ ab ψL = −ψ L γ ab χL (1.44)
χ L γ abc ψR = ψ R γ abc χL .

We also note that under charge conjugation

(γa )c = γa , (γ5 )c = −γ5 , (1.45)

in the sense that (γ a ψ)c = γ a ψ c , etc.

In all the subsequent formulae, the Hermitian conjugate, +h.c., of a field operator
is denoted with a superscript ∗ , whereas the superscript † is reserved for matrix
expressions when in addition to Hermitian conjugation also a transposition of the
matrix is involved. On ordinary complex numbers and classical fields, the Hermitian
1.3 A Quick Guide Through Our Spinor Conventions 17

conjugation acts as complex conjugation, where, however, the order of anti-

commuting spinor fields is exchanged to mimic the effect of Hermitian conjugation
on the corresponding quantum fields. This results in a minus sign when the original
spinor order is restored. Fortunately, the effect of this Hermitian conjugation can
simply be obtained by writing down the charge conjugate expression with all the
rules obtained so far, including (1.45), but without exchanging the order of the
spinors.
As an example, we show (ψ L γ a χR )∗ = (ψ L γ a χR )c = ψ R γ a χL :

(ψ L γ a χR )∗ = (ψLT Cγ a χR )∗ = −ψL† C ∗ γ a∗ χR∗ = −ψL† C ∗ γ a∗ B(χR )c

= −ψL† C ∗ Bγ a χL . (1.46)

Inserting (1.34), C ∗ = C, and (γ 0T )−1 = −C(γ 0 )−1 C −1 = −C −1 (γ 0 )−1 C, this

becomes

(ψ L γ a χR )∗ = −iψL† (γ 0 )−1 γ a χL = ψL γ a χL = ψ R γ a χL = (ψ L )c (γ a )c (χR )c

= (ψ L γ a χR )c , (1.47)

where in the second equation we used (γ 0 )−1 = −γ 0 , which follows from (γ 0 )2 =

−14 .
In the following, we will often need to rewrite three or four Fermi terms to
complete our analysis of the supersymmetry properties of an action, and hence Fierz
identities will be extremely useful. We list here the main ones for two spinors:

1 1
ψR χ R = − χ R ψR PR + χ R γab ψR γ ab PR , (1.48)
2 8
1
ψR χ L = − χ L γ a ψR γa PL , (1.49)
2
where for the sake of clarity we explicitly left the projectors on the right hand side.
In the rest of these lectures, we will also often make use of spinor one-forms
ψ = dx μ ψμ . Exchanging such spinor one-forms then leads to an additional minus
sign from the anti-commutativity of the wedge product, and hence we have

1
ψR ∧ ψ R = − ψ R ∧ γab ψR γ ab PR , (1.50)
8
1
ψR ∧ ψ L = ψ ∧ γ a ψR γa PL , (1.51)
2 L
18 1 Introduction

where now ψ L ∧ ψL = 0 because of (1.44) and the wedge product. A crucial

consequence is the cyclic identity:

γ a ψL ∧ ψ L ∧ γa ψR = 0. (1.52)

1.3.3 Susy Algebra in Four Dimensions

Using the conventions described so far, the N = 1 supersymmetry algebra in four

dimensions has the following form:

{Q, Q̄} = −2iγ a Pa ,

[Pa , Q] = 0,
i
[Mab , Q] = γab Q,
2
[R, Q] = i γ5 Q, (1.53)

[Pa , Pb ] = 0,
[Pa , Mbc ] = −2i ηa[b Pc] ,
[Mab , Mcd ] = 2i ηc[a Mb]d − 2i ηd[a Mb]c .

Here, Q is the supersymmetry generator described by a Majorana spinor; Mab

and Pa denote, respectively, the usual generators of Lorentz transformations and
spacetime translations; and R is the U(1) internal R-symmetry generator. In (1.53),
we used Hermitian generators Pa and Mab . Just as in (1.3), we will sometimes also
use their anti-Hermitian counterparts,

Pa = i Pa , Mab = i Mab (1.54)

when this is more convenient. Note that, for simplicity, we have not included internal
bosonic symmetry generators other than the R-symmetry, as they commute with the
above generators.

Exercises

1.1. Using the totally antisymmetric epsilon tensor with

0123 = 1, (1.55)
References 19

check the duality relations

1
γ abc = i  abcd γd γ5 , i γa γ5 = abcd γ bcd ,
3!
1
γ abcd = −i  abcd γ5 , i γ5 = abcd γ abcd , (1.56)
4!
i abcd
γ ab =  γcd γ5 .
2
1
1.2. Verify that Mab = γab satisfies [Mab , Mcd ] = −2 ηc[a Mb]d + 2 ηd[a Mb]c .
2
1.3. Using just the Clifford algebra (1.12) in an arbitrary representation and the
12
definition (1.11) of Σab , compute the rotation matrix R(θ ) = eθΣ for a rotation in
the (1, 2)-plane by a finite angle θ and read off from your result that R(2π) = −1,
i.e., that Σab is indeed a spinor representation.

References
1. B. de Wit, H. Nicolai, N = 8 Supergravity. Nucl. Phys. B208, 323 (1982)
2. M.H. Goroff, A. Sagnotti, The ultraviolet behavior of Einstein gravity. Nucl. Phys. B266, 709
(1986)
3. A.E.M. van de Ven, Two loop quantum gravity. Nucl. Phys. B378, 309–366 (1992)
4. Z. Bern, L.J. Dixon, R. Roiban, Is N = 8 supergravity ultraviolet finite?. Phys. Lett. B644, 265–
271 (2007) [hep-th/0611086]
5. J. Carrasco, Generic Multiloop Methods for Gauge and Gravity Scattering Amplitudes, a
Guided Tour with Pedagogic Aspiration (Strings, Munich, 2012). http://wwwth.mpp.mpg.de/
members/strings/strings2012/strings_files/program/Talks/Tuesday/Carrasco.pdf
6. J.M. Maldacena, The large N limit of superconformal field theories and supergravity. Int. J.
Theor. Phys. 38, 1113–1133 (1999) [arXiv:hep-th/9711200 [hep-th]]
7. E. Witten, Anti-de Sitter space and holography. Adv. Theor. Math. Phys. 2, 253–291 (1998)
[arXiv:hep-th/9802150 [hep-th]]
8. D.Z. Freedman, C. Nunez, M. Schnabl, K. Skenderis, Fake supergravity and domain wall
stability. Phys. Rev. D69, 104027 (2004) [hep-th/0312055]
9. A. Celi, A. Ceresole, G. Dall’Agata, A. Van Proeyen, M. Zagermann, On the fakeness of fake
supergravity. Phys. Rev. D71, 045009 (2005) [hep-th/0410126]
10. K. Skenderis, P.K. Townsend, Gravitational stability and renormalization group flow. Phys.
Lett. B 468, 46 (1999) [hep-th/9909070]
11. K. Skenderis, P. Townsend, Pseudo-supersymmetry and the domain-wall/cosmology corre-
spondence. J. Phys. A A40, 6733–6742 (2007) [hep-th/0610253]
12. P. Van Nieuwenhuizen, Supergravity. Phys. Rept. 68, 189–398 (1981)
13. H.P. Nilles, Supersymmetry, supergravity and particle physics. Phys. Rept. 110, 1 (1984)
14. B. de Wit, D.Z. Freedman, Supergravity: the basics and beyond. Bonn Superym. ASI 0135
(1984). MIT-CTP-1238
15. L. Castellani, R. D’Auria, P. Fre, Supergravity and Superstrings: A Geometric Perspective, vol.
1. Mathematical Foundations (World Scientific, Singapore, 1991), pp. 1–603
16. L. Castellani, R. D’Auria, P. Fre, Supergravity and Superstrings: A Geometric Perspective, vol.
2. Supergravity (World Scientific, Singapore, 1991), pp. 607–1371
20 1 Introduction

17. J. Wess, J. Bagger, Supersymmetry and Supergravity (Princeton University Press, Princeton,
1992)
18. A. Van Proeyen, Tools for supersymmetry. hep-th/9910030
19. J.-P. Derendinger, Introduction to Supergravity. Lectures at the Summer School “Gif 2000” in
Paris
20. B. de Wit, Supergravity. hep-th/0212245
21. A. Van Proeyen, Structure of supergravity theories. hep-th/0301005. To be published in the
series Publications of the Royal Spanish Mathematical Society
22. P. van Nieuwenhuizen, Supergravity as a Yang-Mills theory. hep-th/0408137
23. F. Zwirner, Supersymmetry Breaking in Four and More Dimensions. Lectures at the 2005 Parma
School of Theoretical Physics
24. H. Samtleben, Lectures on gauged supergravity and flux compactifications. Class. Quant. Grav.
25, 214002 (2008) [arXiv:0808.4076]
25. R. Kallosh, L. Kofman, A.D. Linde, A. Van Proeyen, Superconformal symmetry, supergravity
and cosmology. Class. Quant. Grav. 17, 4269–4338 (2000) [Erratum: Class. Quant. Grav. 21,
5017 (2004)] [arXiv:hep-th/0006179 [hep-th]]
26. L. Andrianopoli, M. Bertolini, A. Ceresole, R. D’Auria, S. Ferrara, P. Fre, T. Magri, N = 2
supergravity and N = 2 super Yang-Mills theory on general scalar manifolds: symplectic
covariance, gaugings and the momentum map. J. Geom. Phys. 23, 111–189 (1997) [arXiv:hep-
th/9605032 [hep-th]]
27. B. de Wit, H. Samtleben, M. Trigiante, On Lagrangians and gaugings of maximal supergravi-
ties. Nucl. Phys. B 655, 93–126 (2003) [arXiv:hep-th/0212239 [hep-th]]
28. B. de Wit, H. Samtleben, M. Trigiante, The maximal D = 4 supergravities. JHEP 06, 049 (2007)
[arXiv:0705.2101 [hep-th]]
29. S. Weinberg, The Quantum Theory of Fields, vol. 3. Supersymmetry (Cambridge University
Press, Cambridge, 2000), 419 p.
30. D.Z. Freedman, A. Van Proeyen, Supergravity (Cambridge University Press, Cambridge, 2012)
31. E. Lauria, A. Van Proeyen, N = 2 Supergravity in D = 4, 5, 6 Dimensions. Lect. Notes Phys.
966 (2020) [arXiv:2004.11433 [hep-th]]
32. P.C. West, Supergravity, brane dynamics and string duality [arXiv:hep-th/9811101 [hep-th]]
From Global to Local Supersymmetry
2

In this chapter, we revisit the simplest globally supersymmetric field theory in four
dimensions, the free massless Wess–Zumino model for one chiral multiplet, and
discuss how this theory has to be changed when supersymmetry is turned into a
local symmetry. As we will see, making supersymmetry local by a simple iterative
procedure (the “Noether method”) directly exhibits the need for the gravitino field
and its superpartner, the graviton, and suggests the supersymmetry transformation
laws of these fields. We end this chapter by a discussion of the basic properties of
the gravitino in Sect. 2.2.

2.1 Promoting Supersymmetry to a Local Symmetry

Consider the free massless Wess–Zumino model for one chiral multiplet (φ, χ),
where φ(x) is a complex scalar and χ(x) a Majorana spinor field with Lagrangian

L = −∂μ φ∂ μ φ ∗ − χ R ∂/χL + χ L ∂/χR . (2.1)

We recall that the mass dimensions of the fields are D[φ] = 1 and D[χ] = 3/2.
Before we continue, we should make a short remark on our index conventions. In
this book, we generally use Greek indices μ, ν, . . . = 0, 1, 2, 3 to denote the local
coordinates, x μ , of 4D spacetime manifolds. For consistency, we also do this for
4D Minkowski spacetime when, as in the above Lagrangians, it is considered as a
special example of a differentiable manifold. On the other hand, when we consider
Minkowski space as a vector space (e.g., a tangent space at a point of a differentiable
manifold) with orthonormal basis vectors, ea , we use the Latin indices a, b, . . . =
0, 1, 2, 3, as we did in Sect. 1.3. For Minkowski spacetime, this distinction is of
course not really necessary if one works with coordinates x μ that correspond to
a global inertial frame, as we also do here, because then the coordinate-induced
tangent vectors, ∂μ , are orthonormal and could be identified with the orthonormal

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 21

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_2
22 2 From Global to Local Supersymmetry

basis vectors ea . The gamma matrices γ μ in Minkowski spacetime are then also
simply the same constant matrices γ a described in Sect. 1.3. We will see later,
however, that on a general curved spacetime manifold, this distinction between local
coordinate indices μ, ν and the indices a, b of an orthonormal basis of the tangent
spaces is in fact crucial for a proper definition of spinor fields on curved spacetimes.
The Lagrangian (2.1) is supersymmetric under the variations1

δ  φ =  L χL ⇐⇒ δ  φ ∗ =  R χR (2.2)
1 1 ∗
δ  χL = ∂/φR ⇐⇒ δ  χR = ∂/φ L , (2.3)
2 2
with D[] = −1/2. In these conventions, it follows that

1 1
δ χ L = −  R ∂/φ ⇐⇒ δ χ R = −  L ∂/φ ∗ , (2.4)
2 2

as one can easily check by using (1.43) as well as RT (γ μ )T C = −RT (Cγ μ ), or by
simply taking the charge conjugate of both sides. The Lagrangian (2.1) is invariant
under the supersymmetry transformations (2.2)–(2.3) up to a total derivative.
To check this explicitly, we first use (1.44) to write the fermionic term of the
Lagrangian (2.1) as Lfer = −χ R ∂/χL + ∂μ (χ R )γ μ χL . For the supersymmetry
variation of the Lagrangian, it is sufficient to trace the terms involving L , which
come only from the variation of φ and χ R :

δL = −∂μ (δφ)∂ μ φ ∗ − δχ R ∂/χL + ∂μ (δχ R )γ μ χL + h.c. (2.5)

Integrating by parts now the first and the second term gives
 
δL = δφφ ∗ + 2∂μ (δχ R )γ μ χL + ∂μ − δφ∂ μ φ ∗ − δχ R γ μ χL +h.c.
  
≡K μ

(2.2),(2.4)
=  L χL φ ∗ − ∂μ ( L ∂/φ ∗ )γ μ χL + ∂μ K μ
+ h.c.

1 Note that, since we are not using auxiliary fields, the supersymmetry algebra closes only on-shell:

[δ2 , δ1 ]φ = 1 μ
2 ( 1 γ 2 )∂μ φ
 
δ2 , δ1 χL = 1 μ
2 ( 1 γ 2 )∂μ χL + [. . .]∂/χL ,

where [. . .] denotes a non-vanishing expression of the fields and supersymmetry parameters. The
last term then1 vanishes due to the field equation ∂/χL = 0, and one obtains the usual susy algebra
δ2 , δ1 = 2 ( 1 γ μ 2 )∂μ on all fields.
2.1 Promoting Supersymmetry to a Local Symmetry 23

=  L χL φ ∗ − ∂μ ( L )∂/φ ∗ γ μ χL −  L ∂μ ∂ν φ ∗ γ ν γ μ χL + ∂μ K μ
+ h.c.
  
φ ∗ 14

= −∂μ ( L )∂/φ ∗ γ μ χL + ∂μ K μ
+ h.c. (2.6)

As promised, the result is that under global supersymmetry, where the supersymme-
try parameter is constant, ∂μ  = 0, the Lagrangian transforms into a total derivative:

δ L = ∂μ (K μ
+K μ∗
) ≡ ∂μ K μ . (2.7)

When dealing with local supersymmetry, however, the parameter  becomes a

local function of the coordinates,  = (x), and the Lagrangian is no longer invariant
up to a total derivative. The new non-invariant part of the Lagrangian reads
μ μ
δ L = (∂μ )j μ = (∂μ  L )jL + (∂μ  R )jR , (2.8)

where

jL ≡ −∂/φ ∗ γ μ χL ,
μ μ
jR ≡ −∂/φγ μ χR (2.9)
μ μ
give the super-Noether current j μ = jL + jR . In fact, it can be easily checked
that this supercurrent is a conserved current, namely, that ∂μ j μ = 0, upon using the
equations of motion for the fields φ and χ. It should also be noted that the dimension
μ
of these currents is D[jL,R ] = 7/2.
We can now apply Noether’s method and associate to the supercurrent (2.9) a
gauge field that compensates the non-invariance of the Lagrangian (2.8). This gauge
field, ψμα , has to have a spinorial index (i.e., the index α = 1, 2, 3, 4, which we
will suppress again in the following), such that

δ ψμ L,R = MP ∂μ L,R , δ ψ μ L,R = MP ∂μ  L,R , (2.10)

where MP is a mass parameter that is needed to relate the mass dimension 3/2 of
the fermionic field ψμ and the dimension of the supersymmetry parameter D[] =
−1/2. As suggested by the notation, MP will later be identified with the (reduced)
Planck mass.
The Noether procedure now tells us that we need to add a new piece to the
Lagrangian:

 1 μ μ
LWZ =− ψ μL jL + ψ μR jR . (2.11)
MP

Again MP is needed to get a Lagrangian density whose total mass dimension is 4,

and this dimensionful coupling in the action can be viewed as a first sign that we
eventually need gravity in local supersymmetry.
24 2 From Global to Local Supersymmetry

Using (2.10) in the variation of (2.11), we now precisely compensate the variation
of the original Wess–Zumino multiplet, but now there is a new piece to compensate
μ
in the variation of (2.11) from δ jR,L , which is in general non-vanishing. To see
μ
this, it suffices to consider the variation of the term ψ μL jL that is quadratic in the
μ
scalar fields. This term comes from the variation of χL inside jL :

1
ψ μL ∂/φ ∗ γ μ δ χL = ψ γ ν γ μ γ ρ R ∂ν φ ∗ ∂ρ φ = ψ μL γν R T μν + . . . (2.12)
2 μL
where, using some gamma matrix algebra,

1
T μν = ∂ (μ φ∂ ν) φ ∗ − ημν (∂σ φ∂ σ φ ∗ ), (2.13)
2
and the dots stand for terms involving γ νμρ . One can show that variations bilinear
in χ likewise give the energy momentum tensor for the field χ. So,

 1
δLWZ ∼ γμ ψν T μν + . . . (2.14)
MP

In order to cancel this term, we now introduce a new current which is a symmetric
tensor gμν with transformation rule

1
δgμν ∼ γ(μ ψν) (2.15)
MP

and add a new piece to the Lagrangian with a coupling between the tensor field gμν
and energy momentum tensor:

LWZ ∼ −gμν T μν . (2.16)

As only the spacetime metric can couple to the energy momentum tensor, local
supersymmetry requires the coupling of the Wess–Zumino multiplet to gravity
described by a dynamical spacetime metric, gμν , and ψμ must be its superpartner,
the gravitino, as follows from the transformation law (2.15). As in ordinary gauge
theories, one also adds kinetic terms for these new “gauge” fields, and we thus
expect a final result of the form

L = Lkin (φ) + Lkin (χ) + Lint (φ, χ, gμν , ψμ )

     
LWZ 
LWZ + LWZ  +... (2.17)
+ Lkin (gμν ) + Lkin (ψμ )

where the dots indicate possible further interaction terms.

2.2 The Gravitino 25

We used the chiral multiplet to guess the supersymmetry transformation rules of

the supergravity multiplet. These rules, however, should hold also in the absence of
the chiral multiplet, and we thus arrive at a motivated guess for the Lagrangian and
transformation laws of pure N = 1 supergravity:

Lpure sugra = Lkin(gμν ) + L (ψμ ) +Lint (gμν , ψμ ),

    kin  (2.18)
MP √
2

2 −gR − 12 ψ μ γ μνρ ∂ν ψρ |cov

using

1
δgμν  γ(μ ψν) |cov , (2.19)
MP

and

δψμ  MP ∂μ |cov , (2.20)

where cov stands for a proper spacetime covariantization, and Lint denotes
possible interaction terms that are not contained in the covariantizations of the
kinetic terms (e.g., four Fermion terms). In Chaps. 3 and 4, we will discuss this
spacetime covariantization, which requires an appropriate description of spinors in
curved spacetimes. As we will see in Chap. 4, the additional interaction terms not
related to spacetime covariantization can, in fact, elegantly be absorbed into the
(covariantized) kinetic terms by working with covariant derivatives with non-trivial
torsion. Before we come to this, however, let us briefly pause and take a quick look
at some properties of the gravitino.

2.2 The Gravitino

As we have seen from the previous discussion, a new fundamental ingredient in the
construction of locally supersymmetric actions is the gravitino, which acts as the
gauge field associated to the supersymmetry gauge transformation. In this section
we want to discuss two important points regarding the gravitino field: its action and
the role of the gravitino multiplet.

2.2.1 The Gravitino Action

The gravitino field should propagate spin 3/2 degrees of freedom, because it is
the superpartner of the metric field, which is known to propagate spin 2 degrees
26 2 From Global to Local Supersymmetry

of freedom.2 The correct Lorentz representation for describing such degrees of

freedom is (1, 1/2) ⊕ (1/2, 1), where we labeled the representations of the Lorentz
group as described in Sect. 1.3, which means that one should use ψ(AB)Ȧ and ψ(ȦḂ)A
fields. These fields generically contain both spin 3/2 and spin 1/2 degrees of freedom
with respect to the spatial rotation group appropriate for the description of a massive
particle:

1 ⊗ 1/2 = 3/2 ⊕ 1/2. (2.21)

Unfortunately, as shown by Fierz and Pauli [1, 2], it is impossible to write down a
local, Lorentz invariant action with such fields only. One needs at least an auxiliary
spin 1/2 field. As discussed by Rarita and Schwinger [3], one can build a consistent
action by employing a gravitino field that is a vector-spinor, ψμ , sitting in a reducible
representation of the Lorentz group. It does indeed contain both the gravitino
and an auxiliary spinor degree of freedom as follows from the Clebsch–Gordan
decomposition of its index structure

(1/2, 1/2) ⊗ [(1/2, 0) ⊕ (0, 1/2)] = (1, 1/2) ⊕ (1/2, 1) ⊕ (1/2, 0) ⊕ (0, 1/2) . (2.22)

The action for the free field is of the form

1 1
L3/2 = − ψ μ γ μνρ ∂ν ψρ + m3/2ψ μ γ μν ψν , (2.23)
2 2
where m3/2 is a (real) mass parameter that we will identify below with the physical
mass of the gravitino in Minkowski spacetime.
We now show that the equation of motion following from (2.23) propagates
2 degrees of freedom in the massless case and 4 degrees of freedom in the
massive case. Let us first focus on the massless case, where m3/2 = 0. In this
case, the Lagrangian L3/2 is invariant, up to a total derivative, under the gauge
transformations

δψμ = ∂μ Λ, (2.24)

where Λ(x) is an arbitrary Majorana spinor-valued function. For m3/2 = 0, the

equation of motion following from L3/2 is

γ μνρ Ψνρ = 0, (2.25)

2For the sake of readability, we do not distinguish carefully here between spin and helicity, i.e.,
“spin s” should be understood as “helicity ±s” in the massless case.
2.2 The Gravitino 27

where we introduced the gravitino field strength, Ψμν = 2∂[μ ψν] , which is invariant
under (2.24). This equation is also equivalent to

γμ γσ γ μνρ Ψνρ = 4γ μ Ψμσ = 0. (2.26)

In order to see that this is the correct equation that propagates only spin 3/2
degrees of freedom, we can look at it in momentum space. Assuming that we
have plane-wave solutions with momentum k μ , the Fourier transform of Ψμν can
be decomposed as

j
Ψμν (k) = ai (k) k[μ ν]
i
+ bi (k) k̃[μ ν]
i
+ c(k) k[μ k̃ν] + dij (k) [μ
i
ν] , (2.27)

where we introduced a complete set of longitudinal and transverse vectors


k μ = (k 0 , k), 
k̃ μ = (k 0 , −k),  μi , (2.28)

 = 0, and
with i = 1, 2, k 0 = |k|

kμ k̃ μ < 0, kμ  μi = k̃μ  μi = 0, μi  μj = δ ij . (2.29)

It is straightforward to see that bi = dij = 0 because of the Bianchi identity

∂[μ Ψνρ] = 0, which in momentum space is

j
i
bi (k) k[μ k̃ν ρ] + dij (k) k[μ νi ρ] = 0. (2.30)

Hence off-shell the independent degrees of freedom are contained in the spinors
ai and c, which are Majorana, and therefore sum up to 12 independent real
components in four dimensions. In momentum space, the equation of motion (2.26)
now becomes

μi kν γ ν ai + k̃μ kν γ ν c − kμ (νi γ ν ai + k̃ν γ ν c) = 0. (2.31)

Contracting with  μj , k̃ μ , and k μ , this implies kμ γ μ ai = 0, kμ γ μ c = 0, and

μi γ μ ai + k̃μ γ μ c = 0, which further brings us to

c = 0. (2.32)

Thus, all the on-shell degrees of freedom are contained in the two spinors ai (k),
subject to the constraints

kμ γ μ ai = 0, μi γ μ ai = 0. (2.33)
28 2 From Global to Local Supersymmetry

Overall we therefore have only two degrees of freedom, because the first equation is
a projector halving the number of degrees of freedom contained in ai and the second
shows that they are not independent.
In the massive case, the Lagrangian (2.23) leads to the equations of motion

γ μνρ ∂ν ψρ = m3/2 γ μν ψν , (2.34)

which are also not in the standard form proposed by Dirac for a fermion field. This,
however, was to be expected, because the gravitino field is a sum of representation
of the Lorentz group of different dimension.
We can obtain the Dirac equation of motion for a spin 3/2 field by looking at two
different contractions of (2.34). First, take its divergence:

0 = ∂μ γ μνρ ∂ν ψρ = m3/2 ∂μ γ μν ψν , (2.35)

where the first term vanishes because of symmetry properties. Then contract (2.34)
with γμ :

0 = 2γ νρ ∂ν ψρ = 3 m3/2 γ ν ψν , (2.36)

which is vanishing because of (2.35). The result is that we removed the auxiliary
spin 1/2 component ψ./ Finally, by using some gamma matrix algebra and by
applying the constraints (2.35) and (2.36) to the original equations of motion, we
get

γ μνρ ∂ν ψρ = γ μ γ νρ ∂ν ψρ − γ ρ ∂ μ ψρ + ∂/ψ μ = ∂/ψ μ , (2.37)

for the left-hand side of (2.34), and

γ μν ψν = γ μ γ ν ψν − ψ μ = −ψ μ (2.38)

for the right-hand side of (2.34). We thus proved that the massive Rarita–Schwinger
action leads to equations of motion equivalent to

∂/ψμ + m3/2 ψμ = 0, ∂μ ψ μ = 0, γ μ ψμ = 0, (2.39)

where the second equation follows from the divergence of (2.38), using (2.36). The
final result is the standard Dirac form for the equations of motion of the irreducible
component of the gravitino field. We can thus interpret m3/2 as the physical mass of
the gravitino in Minkowski space.3 A counting of the degrees of freedom, similarly
to the massless case, would reveal four independent states [1, 2].

3We will see in Sect. 4.2.2 that in the presence of a (negative) cosmological constant, the concept
of mass will be slightly different from the one we are used to in the flat case.
2.2 The Gravitino 29

2.2.2 The Gravitino Multiplet

In the previous section, we mentioned that the gravitino is associated to the graviton
by a supersymmetry transformation and that therefore these two fields will sit in
the same supermultiplet. However, from the representations of the supersymmetry
algebra, one can see that, for minimal supersymmetry, there is also a supermultiplet
that pairs the gravitino with a vector field, namely,

the gravitino multiplet: {ψμ , Aμ }. (2.40)

One may expect this multiplet to play a role in a supergravity theory, too. However,
this is not the case.
In fact, it has been shown that trying to add interactions to the free gravitino
multiplet or to use a massless gravitino multiplet as a matter multiplet in a
supergravity theory leads to undesired pathologies. For instance, minimal couplings
to an external U(1) vector field are already problematic. In fact, once one introduces
covariant derivatives in the gravitino kinetic term ∂μ → Dμ = ∂μ − eAμ , the
divergence of the massless gravitino equation of motion imposes γ μνρ Fμν ψρ = 0
and therefore improperly sets to zero some of the degrees of freedom. On the
other hand, the minimal coupling of a massive gravitino to external electromagnetic
fields results in equations of motion which exhibit faster-than-light propagation of
signals [4]. The only known way outs, so far, are given either by Vasiliev’s higher
spin theories [5], where one introduces an infinite number of non-normalizable
couplings in a fixed spacetime with a non-vanishing cosmological constant, or by
the promotion of the gravitino’s gauge invariance to an additional supersymmetry
transformation relating it to the metric field.
As we will see later on, in Chap. 8, this important fact leads also to one of the
main striking differences between global and local supersymmetry: In the case of
globally supersymmetric theories, the models with N > 1 supersymmetry can
always be rewritten in terms of the N = 1 language. More precisely, even highly
(globally) supersymmetric models can always be considered as special models in
the class of minimally supersymmetric theories in the sense that their couplings
are fully compatible with all N = 1 requirements but just sit at a special point
in parameter space that allows the existence of additional global supersymmetries.
In other words, several global supersymmetries do not interfere with each other’s
individual consistency conditions and can peacefully coexist.
In supergravity this is not the case. The graviton multiplet with more than
minimal supersymmetry has to be rewritten in terms of an N = 1 graviton
multiplet coupled to N − 1 massless gravitino multiplets. However, as we just saw,
this is not consistent unless one introduces new scales and couplings which lead
to higher spin models or when the gauge invariance of the additional gravitini is
realized as an additional supersymmetry. A local supersymmetry, however, always
introduces various non-renormalizable and non-linear couplings of the scalar fields,
as we will discuss in detail in later chapters. Due to their non-linear nature,
30 2 From Global to Local Supersymmetry

these additional scalar field couplings cannot simply be “superimposed” for the
different local supersymmetries, but they distort each other. The resulting scalar
field couplings therefore satisfy distorted consistency conditions compared to the
situation when only one local supersymmetry is present, and, generically, the
consistency conditions for precisely one local supersymmetry are no longer satisfied
in extended supergravity. Thus, extended supergravity models can in general not be
described in terms of N = 1 language. We will see examples of this when we
discuss the scalar manifolds of, e.g., hypermultiplets in N = 2 supergravity or of
the scalars in N = 8 supergravity.
We close this chapter by noting that despite the abovementioned difficulties for
gravitino multiplet couplings, it is nevertheless possible to construct a (globally)
supersymmetric action for this multiplet, with the gravitino and vector field
satisfying free equations of motion and both being massless fields:

1 1
L = − ψ μ γ μνρ ∂ν ψρ − F μν Fμν . (2.41)
2 4
The gravitino has the expected gauge invariance δψμ = ∂μ Λ, like the gauge boson
δAμ = ∂μ Σ, but this is unrelated to the supersymmmetry transformation rule,
which, on the other hand, relates the gravitino to the vector field strength [6]:

1 1
δψμ = γμνρ  F νρ − γ ρ  Fμρ , δAμ = ψμ . (2.42)
4 2

Exercises

2.1. Verify the invariance of (2.41) under the supersymmetry transformations

(2.42).

References
1. M. Fierz, On the relativistic theory of force-free particles with any spin. Helvetica Physica Acta
12(I), 3–37 (1939)
2. M. Fierz, W. Pauli, On relativistic wave equations for particles of arbitrary spin in an
electromagnetic field. Proc. Roy. Soc. Lond. A 173, 211 (1939)
3. W. Rarita, J. Schwinger, On a theory of particles with half integral spin. Phys. Rev. 60, 61 (1941)
4. G. Velo, D. Zwanziger, Propagation and quantization of Rarita-Schwinger waves in an external
electromagnetic potential. Phys. Rev. 186, 1337–1341 (1969)
5. M.A. Vasiliev, Higher spin gauge theories in four-dimensions, three-dimensions, and two-
dimensions. Int. J. Mod. Phys. D 5, 763–797 (1996) [arXiv:hep-th/9611024 [hep-th]]. Higher
spin gauge theories: star product and AdS space [arXiv:hep-th/9910096 [hep-th]]
6. B. de Wit, J.W. van Holten, Multiplets of linearized SO(2) supergravity. Nucl. Phys. B155, 530
(1979)
Gravity and Spinors
3

Supergravity is a supersymmetric theory of gravity based on the principles of

general relativity. In this subsection we set up our notation and review some
elementary material on general relativity with particular emphasis on those aspects
that are relevant for the formulation of supergravity theories.

3.1 The Standard Metric Formulation

General relativity is concerned with the dynamics of Lorentzian metric tensors, g =

gμν (x)dx μ ⊗ dx ν , on differentiable manifolds, M . In these lectures, we always use
the metric signature (−, +, . . . , +), and local coordinate bases of tangent vectors
are denoted as ∂μ ≡ ∂x∂ μ , i.e., gμν (x) = g(∂μ , ∂ν ), (See Fig. 3.1).
A connection, ∇, on a differentiable manifold M is a prescription for a covariant
derivative of the tensor fields of M . It is fixed by its action on the coordinate
basis vectors ∂μ (or the dual covectors dx μ ), i.e., by the (generically x-dependent)
ρ
connection coefficients Γμν ,

∇μ (∂ν ) ≡ ∇∂μ (∂ν ) = Γμν

ρ
∂ρ ⇐⇒ ∇μ (dx ρ ) = −Γμν
ρ
dx ν . (3.1)

Once the action on the basis vectors is defined, we can compute the action of ∇ on
any vector (and covector):

∇V W = V μ ∇∂μ W ν ∂ν = V μ ∂μ W ν + Γμρ
ν
W ρ ∂ν . (3.2)

With an obvious abuse of notation, we therefore define

∇μ W ν ≡ ∂μ W ν + Γμρ
ν
Wρ , (3.3)

and analogously for more general tensor fields.

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 31

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_3
32 3 Gravity and Spinors

Fig. 3.1 The tangent space Tp M to a point p ∈ M with tangent vectors in a coordinate basis

On a manifold with a metric, there is a natural connection, the Levi–Civita

connection, which is uniquely specified by the following requirements

(i) ∇μ gνρ = 0 (metric compatibility)

ρ ρ
(ii) ∇μ ∂ν − ∇ν ∂μ = 0 ⇔ Γμν = Γνμ (vanishing torsion)

The metric compatibility (i) means that the inner product of two vectors is
unchanged under parallel transport, whereas the vanishing torsion1 condition (ii)
can, e.g., be interpreted as requiring that two covariant derivatives commute when
ρ
they act on a scalar function or that Γμν vanishes at the origin of a Riemannian
normal coordinate system. For later reference we note that, using the one-form
ρ
Γ ρ ν ≡ Γμν dx μ , the vanishing torsion condition (ii) can alternatively be written
in terms of the dual basis of one-forms, dx ρ , as

• !
(ii)’ ∇dx ρ ≡ (d + Γ • ∧) dx
ρ = Γ ρ ν ∧ dx ν = 0.

Either way, the two requirements of metric compatibility and vanishing torsion
completely fix the Levi–Civita connection in terms of the metric via the standard
Christoffel symbols,

1 ρσ
ρ
Γμν = Γμν
ρ
(g) = g ∂μ gνσ + ∂ν gμσ − ∂σ gμν . (3.4)
2

1The torsion tensor, T , maps two vector fields, V , W , to another vector field defined by T (V , W ) =
ρ ρ
∇V W − ∇W V − [V , W ], implying Tμν ρ = Γμν − Γνμ .
3.2 The Vielbein Basis and Cartan’s Formalism 33

From the connection, one obtains the Riemann tensor

μ μ
Rρσ μ ν ≡ 2 ∂[ρ Γσ ]ν + 2 Γτ [ρ Γστ]ν , (3.5)

the Ricci tensor

Rμν ≡ Rρμ ρ ν , (3.6)

and the Ricci scalar

R ≡ Rρμ ρ ν g μν = Rμν g μν . (3.7)

The Einstein–Hilbert action then is


1 √
SEH = d 4x −g R, (3.8)
16πGN

where g ≡ det(gμν ), and GN denotes Newton’s constant. It is customary to

introduce the reduced Planck mass, MP , related to Newton’s constant by2

1 M old
MP = √ = √P = 2.44 · 1018 GeV, (3.9)
8πGN 8π

so that the Einstein–Hilbert action can also be written as


MP2 √
SEH = d 4x −g R. (3.10)
2

3.2 The Vielbein Basis and Cartan’s Formalism

In D-dimensional Minkowski spacetime, a vector field, V , and a spinor field, ψ,

transform, respectively, in the fundamental and in the spinor representation of the
Lorentz group,

V μ → Λμ ν V ν (3.11)

ψα → ρ(Λ)αβ ψβ , (3.12)

2The difference between the “old” Planck mass value MPold = 1.22 · 1019 GeV and the reduced
Planck mass (3.9) used in supergravity computations should be kept in mind when quantitative
predictions are required.
34 3 Gravity and Spinors

Fig. 3.2 Orthonormal non-coordinate basis for Tp M

where Λμ ν ∈ SO(1, D − 1), and ρ(Λ)αβ is a spinor representation matrix (i.e.,

a double-valued representation matrix of SO(1, D − 1); see Sect. 1.3) with α, β
denoting suitable spinor indices.
On a curved manifold with arbitrary local coordinate transformations x → x  (x),
the transformation (3.11) is naturally generalized by replacing Λμ ν by the Jacobi
matrix (∂x μ /∂x ν ), which in general is an (x-dependent) element of GL(D, R). A
spinor representation ρ(Λ)αβ , by contrast, does not have an analogous extension to
the full group GL(D, R).
In order to define spinor fields in curved spacetimes, one therefore makes use
of the equivalence principle and switches to a local inertial frame at every point.
More precisely, one chooses as tangent space basis a set of local vector fields, ea
(a = 0, 1, . . . , (D − 1)), that are mutually orthonormal at each point (see Fig. 3.2),

g(ea , eb ) = ηab . (3.13)

Obviously, this condition does not fix the vectors ea uniquely; rather they can
be rotated by arbitrary local (i.e., x-dependent) Lorentz transformations, ea →
eb Λb a (x), without violating the orthonormality (3.13). It should also be noted that
these orthonormal vector fields can in general only be defined on local patches
unless M is parallelizable and that they can in general not be written as coordinate
vectors ∂x∂ a for some local coordinates x a unless M happens to be flat. The non-
coordinate basis ea is often called an orthonormal frame, vielbein, or D-bein (in
D = 4: vierbein or tetrad). In the following we will follow common practice and
use the word vielbein also for the matrix eμa (x) that mediates between the local, ea ,
and the coordinate, ∂μ , bases,

ea = eaμ (x)∂μ , ∂μ = eμa (x)ea (3.14)

3.2 The Vielbein Basis and Cartan’s Formalism 35

μ
with eμa eaν = δμν , ea eμb = δab . This vielbein can be used to convert the “curved” (or
“world”) indices μ, ν, . . . of any tensor field to “flat” (or “local Lorentz”) indices
a, b, . . ., e.g., V a = eμa V μ , etc. In particular, they interpolate between the curved
metric gμν (x) and the flat metric ηab :

gμν (x) = eμa (x)eνb (x)ηab . (3.15)

This statement is clearly invariant under the abovementioned local Lorentz trans-
formations of the vielbein, eμa  = eμb Λa b (x), as ΛT ηΛ = η, and we see
that the vielbein contain the same amount of information as the metric tensor.
Equation (3.15) also implies
√
−g = e ≡ det eμa . (3.16)

A general relativistic theory written in terms of a vielbein thus has two local
invariances: general coordinate transformations acting with Jacobian matrices on all
curved indices μ, ν, . . ., and local Lorentz transformations acting with Λb a (x) on
all flat indices a, b, . . .. One can therefore think of the vielbein eμa also as elements
parameterizing the coset space

GL(D, R)
.
SO(1, D − 1)

In terms of the orthonormal basis ea , a connection is now specified by its action

on these basis vectors,

∇μ ea = ωμ b a (x) eb (3.17)

where we introduced new connection coefficients ωμ b a (x). A connection ωμ b a is

ρ
equivalent to a given connection Γμν in a curved basis if ∇μ ea = eaν ∇μ ∂ν , which is
equivalent to the condition

∇μ eνa ≡ ∂μ eνa + ωμ a b eνb − Γμν

ρ a
eρ = 0, (3.18)

where, as suggested by the notation, the left-hand side can be viewed as a total
covariant derivative acting on curved and flat indices. Equation (3.18) is sometimes
referred to as the “vielbein postulate,” but here it is really just the statement that the
ρ
two connections ωμ b a and Γμν are actually equivalent connections expressed in two
different basis systems.
Just as for the curved indices, one can now demand that ωμ b a preserve the metric
tensor and be torsion-free so as to arrive at the flat index equivalent of the Levi–
Civita connection. The corresponding connection coefficients ωμ b a = ωμ b a (e)
are then uniquely determined by the vielbein eμa and define the (torsion-free) spin
connection. We derive the expression for ωμ b a further below. In the following, it
36 3 Gravity and Spinors

will be useful to define also a derivative operator, Dμ , that is covariant only with
respect to local Lorentz transformations, but not necessarily with respect to general
coordinate transformations, i.e., all local Lorentz indices will be contracted with
spin connections, but there are no Christoffel symbols that contract any world index.
As an example, consider this Lorentz covariant derivative acting on a vielbein eνa :

Dμ eνa ≡ ∂μ eνa + ωμ a b eνb . (3.19)

ρ ρ
In view of (3.18), this expression does not vanish but is equal to Γμν eρa , where Γμν
is the curved connection equivalent to ωμ b a . Recalling that the antisymmetrization
ρ
Γ[μν] = 12 Tμν ρ is just the torsion tensor of this connection, we can thus write

1 1
a
D[μ eν] = Tμν a = Tμν ρ eρa . (3.20)
2 2
Note that, even though the Lorentz covariant derivative, Dμ , is not covariant with
respect to general coordinate transformations, the above antisymmetrized derivative
still forms a proper tensor field. In supergravity, all equations can similarly be
expressed in terms of antisymmetrized Lorentz covariant derivatives only. This
allows one to use compact differential form notation, which, as we will see in
Chap. 4, can substantially reduce the index clutter in some computations. To this
end, we introduce the co-frame of one-forms, ea , dual to the vectors eb :

ea ≡ eμa dx μ . (3.21)

These one-forms are at the basis of Cartan’s formulation of general relativity, as we

briefly review in the remainder of this subsection.

This box summarizes our conventions regarding differential forms.

The wedge product between basis one-forms is defined as

dx i1 ∧ . . . ∧ dx ip = sign(σ ) dx σ (i1 ) ⊗ . . . ⊗ dx σ (ip ) ,
σ ∈Sp

where Sp is the permutation group of p elements. The components of a p-

form Fp ∈ Λp (M ) are normalized as

1 μ1
Fp = dx ∧ . . . ∧ dx μp Fμ1 ...μp .
p!

(continued)
3.2 The Vielbein Basis and Cartan’s Formalism 37

The exterior differential acts from the left as

1 ρ
dFp = dx ∧ dx μ1 ∧ . . . ∧ dx μp ∂ρ Fμ1 ...μp .
p!

For instance, for a connection one-form A, the components of the curvature

F = dA are defined as Fμν = 2 ∂[μ Aν] . The inner product also starts from
the left,

ıV (dx μ1 ⊗ . . . ⊗ dx μp Tμ1 ...μp ) = dx μ2 ⊗ . . . ⊗ dx μp V μ1 Tμ1 μ2 ...μp ,

so that
1
ıV Fp = dx μ2 ∧ . . . ∧ dx μp V μ1 Fμ1 μ2 ...μp .
(p − 1)!

Another useful tool is the Hodge star dual, relating elements in Λp (M )

with those in Λ(D−p) (M ), where D is the dimension of the manifold M . Its
definition on a basis of Λp (M ) is defined by:

1
dx μ1 ∧ . . . ∧ dx μp = dx μp+1 ∧ . . . ∧ dx μD μp+1 ...μD μ1 ...μp ,
(D − p)!

where indices are raised and lowered with the metric gμν , and μ1 ...μD =
eμ1 a1 . . . eμD aD a1 ...aD with 012...(D−1) = 1, so that

1
Fp = dx μp+1 ∧ . . . ∧ dx μD μp+1 ...μD μ1 ...μp Fμ1 ...μp .
p!(D − p)!

This also implies that

2 Fp = −(−1)p(D−p)Fp .

In terms of the ea , the expression (3.20) for the torsion tensor reads

1 μ
Dea ≡ dea + ωa b ∧ eb = T a = dx ∧ dx ν Tμν a , (3.22)
2

whereas the curvature tensor of the connection ωμ b a is given by the two-form

R a b = dωa b + ωa c ∧ ωc b . (3.23)
38 3 Gravity and Spinors

The conditions for the connection ωμ b a to be equivalent to the Levi–Civita

connection (i.e., for ωμ b a to be the torsion-free spin connection) can now be written
as

• Dηab = 0 (metric compatibility)

• T a = 0 (vanishing torsion)

The first condition implies that the spin connection is antisymmetric in its indices
when both are raised or lowered with η: Dηab ≡ dηab + ωa c ηcb + ωb c ηac , and
dηab = 0, so that ω(ab) = 0. The second equation (in combination with (3.22)) has
an obvious similarity with the alternative way (ii)’ of writing the vanishing torsion
condition for the curved indices,

Dea = 0. (3.24)

Just as for the Levi–Civita connection in curved indices, the vanishing of the
torsion tensor can be used to deduce an expression for the spin connection in terms
of the vielbein, ωμab = ωμab (e). It is instructive to do this calculation once, as it will
also be useful later, and the same trick can be used for T a = 0. To this end, consider
 
t dc,a ≡ eν[d eμc] ∂μ eνa + ωμ a b eνb .

This tensor is zero because of the assumed vanishing torsion. Hence we can consider
the sum

t dc,a − t ca,d − t ad,c = 0. (3.25)

From this sum there is only one term in the spin connection that survives, eρa ωρcd ,
and, multiplying by eμa , we obtain

d]
ωμcd (e) = 2eν[c ∂[μ eν] − eaμ eν[c eσ d] ∂ν eσa . (torsion-free spin connection)
(3.26)

It can be checked that the expression (3.4) of the Levi–Civita connection implies the
spin connection (3.26) and vice versa.
In Cartan’s formalism the Einstein–Hilbert action now becomes

MP2
SEH = R ab ∧ ec ∧ ed abcd , (3.27)
4

where 0123 = 1. Recovering the standard form (3.10) is rather straightforward if

we use the identification of the four-dimensional measure

dx μ ∧ dx ν ∧ dx ρ ∧ dx σ = −d 4 x e  μνρσ , (3.28)
3.2 The Vielbein Basis and Cartan’s Formalism 39

where μνρσ = eμa eνb eρc eσd abcd , and express the forms through their components:
    
M2 1 μ
SEH = P dx ∧ dx Rμν
ν ab
∧ dx ρ eρc ∧ dx σ eσd abcd
4 2
  
MP2
= dx μ ∧ dx ν ∧ dx ρ ∧ dx σ Rμν τ υ eτa eυb eρc eσd abcd
8
(3.29)

MP2
=− d 4 x e  μνρσ τ υρσ Rμν τ υ
8

M2
= P d 4 x e R.
2

In order to derive the equations of motion coming from the Einstein–Hilbert

action, it is useful not to assume (3.26) from the outset, but rather to consider the
vielbein and spin connection as independent fields, which then leads to the two field
equations

δS
= 0, (3.30)
δeμa

δS
= 0. (3.31)
δωμab

Given the Einstein tensor,

1
Gμν ≡ Rμν − gμν R, (3.32)
2
one finds
 
δS a
d 4x δe = −MP2 d 4 x e eaσ Gσ μ δeμa (3.33)
δeμa μ

and then the vacuum Einstein equation for the torsion-free spin (i.e., the Levi–
Civita) connection,

δS
= 0 ⇔ Gμν = 0, (3.34)
δeμa

δS
= 0 ⇔ T a = 0, (3.35)
δωμab

as the reader is urged to verify in the exercises at the end of this chapter.
40 3 Gravity and Spinors

Apart from encoding the Einstein equation, these equations are especially
important also for the discussion of the invariance of the action under a symmetry.
A general symmetry acting on the vielbein and the spin connection transforms the
action as
  
δS a δS
δS = d x 4
δe + δω ab
, (3.36)
δeμa μ δωμab μ

where the functional derivative with respect to the vielbein is meant to include only
the explicit appearance of eμa .
Depending on how we interpret the variation with respect to the spin connection,
we can now distinguish the following formalisms:

• Second-order formalism: Here the spin connection is meant to be the explicit

functional of the vielbein given in (3.26), and, therefore, the variation of the
action has to be understood as
  
δS δS δωνcd
δS = d x 4
+ cd δeμa .
δeμa δων δeμa

The name “second order” comes from the fact that the equations of motion for
the vielbein (or the metric) in this formalism involve second-order derivatives.
• First-order formalism: Here the spin connection and the vielbein are considered
as independent fields, just as we did when we derived the Einstein equation
above. This is also called the Palatini formalism, and in this formalism we have
a priori Dea = 0. As we have seen before, it is the variation with respect to
the spin connection that leads to the torsion constraint, while the variation with
respect to the vielbein gives the Einstein equation. The action invariance under a
symmetry is established by requiring a suitable variation for the spin connection
δωab . We will see an example of this later in Chap. 4.
• 1.5-order formalism: In this mixed formalism, one considers the spin connection
and the vielbein again as independent but uses the fact that the variation of the
action with respect to the spin connection is the spin connection equation of
motion (or torsion constraint (3.35)) and hence vanishes on-shell. The invariance
of the action then needs to be checked only by considering the variation with
respect to the vielbein, and the spin connection only plays the role of an auxiliary
field whose variation does not have to be considered in this formalism. Though
this is not very useful in an ordinary gravity theory, it will simplify a lot the
calculations when we check the supersymmetry invariance of the supergravity
action.
3.3 Spinors in Curved Spacetime 41

Of course, these three formalisms must reduce to the same set of transformation
rules on-shell. This happens as long as the spin connection ωab appears in the
action at most quadratically, as in (3.27). This is not always the case for actions
including higher order terms in the Riemann tensor like R ab ∧ Rab or R 4 terms,
etc., which however we do not consider in these lecture notes. Also, while classically
equivalent, these formalisms may not be equivalent at the quantum level.
All these results can be extended easily to cases with non-vanishing (but fixed)
torsion. We just need to solve (3.22) for a given T a . This gives a spin connection
which is the sum of (3.26) and a contorsion tensor, κμab , defined as
 
κμab = eμc T ab c − Tc a b − T b c a , (3.37)

where indices are made flat by the use of vielbeins and raised and lowered with
the flat metric. This solution will follow again from the variation of the action with
respect to the spin connection, when it is considered as an independent field. When
the action has more terms than just the Einstein–Hilbert term, and especially when
there are new couplings of the spin connection to other fields, the variation of the
action with respect to the spin connection will lead to a spin connection with non-
vanishing torsion, as we will see Chap. 4.
We close this subsection by mentioning that the treatment of the vielbein and
the spin connection as independent quantities can have a conceptual meaning that
goes beyond the mere simplification of some mathematical computations. As we
explain in Appendix 4.A, an interesting perspective on gravity can be gained by
considering it as a gauge theory of the Poincaré group. This analogy will work only
to a certain extent, but it will be very useful in understanding many specific new
features that have to be introduced when one wants to promote supersymmetry to a
local symmetry of nature. In fact supergravity is the gauge theory of supersymmetry,
and therefore there must be a way to describe it as a theory where the gauge group
is the Poincaré supergroup (or some other supergroup).

3.3 Spinors in Curved Spacetime

As we already mentioned, spinors transform in double-valued representations of

the Lorentz group and cannot be extended to analogous representations of the
general linear group. To describe spinors in curved spacetime, we therefore use a
vierbein basis for the tangent space, so that local Lorentz transformations are still
meaningful, and a spinor field now transforms as

ψα → ρ(Λ(x))αβ ψβ . (3.38)

Here, as indicated, the Lorentz transformation Λ(x) in general is allowed to be

spacetime dependent. Meaningful derivatives acting on spinors thus have to be
covariant with respect to such local Lorentz transformations. As the spin connection
42 3 Gravity and Spinors

is the connection for local Lorentz transformations, and the latter are generated by
1
2 γab in the spinor representation, the correct Lorentz covariant derivative is

1
Dμ ψ = ∂μ ψ + ωμ ab γab ψ, (3.39)
4
where we suppressed the spinor indices, and the additional factor of 1/2 comes
from the sum over a double index. This derivative satisfies Dμ [ρ(Λ(x))ψ] =
ρ(Λ(x))Dμ ψ.
We finally mention that gamma matrices with a curved index μ are obtained from
the constant γa via contraction with a vierbein:

γμ ≡ eμa γa . (3.40)

The γμ are then in general no longer constant, and they transform non-trivially under
a variation of the vierbein.

Exercises

3.1. Prove the relations (3.4) and (3.26).

3.2. Prove (3.34)–(3.35).

3.3. Check that

{γμνρ , γ ab } = −12 γ[μeνa eρ]

b
.

3.4. Using the definition of the Riemann curvature from the spin connection (3.23)
and the definition of the covariant derivative on a spinor (3.39), prove that

1
[Dμ , Dν ]ψ = Rμν ab γab ψ.
4
Pure N = 1 Supergravity in Four Dimensions
4

In this chapter, we discuss pure supergravity in four dimensions with minimal

supersymmetry and show explicitly the construction of its action, also proving the
invariance under supersymmetry transformations. We then extend this construction
to the case of a non-trivial cosmological constant and finally discuss the concept of
mass in spacetimes with negative curvature.

4.1 Pure Supergravity: The Action and SUSY Rules

In this section, we will work out in detail the supersymmetry transformation of

the pure supergravity Lagrangian, so that the origin and inevitability of the various
terms in the action should become as clear as possible. In order to simplify several
computations, we often use the language of differential forms. For instance, the
starting supergravity action is the sum of the Einstein–Hilbert and Rarita–Schwinger
actions,
   
M 2
1
S = d 4 x (LEH + LRS ) = d 4 x e P
R − ψ μ γ μνρ Dν ψρ , (4.1)
2 2

which, in the language of differential forms, becomes

 
M2 i
S= P R ab
∧ e ∧ e abcd
c d
+ ea ∧ ψ ∧ γ5 γa Dψ, (4.2)
4 2

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 43

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_4
44 4 Pure N = 1 Supergravity in Four Dimensions

as follows from using the duality relations on the gamma matrices derived in the
Exercises to Chap. 1:
 
i i
ea ∧ ψ ∧ γ5 γa Dψ = dx μ ∧ dx ν ∧ dx ρ ∧ dx σ eμa ψ ν γ5 γa Dρ ψσ
2 2

i
=− d 4 x e  μνρσ ψ ν γ5 γμ Dρ ψσ (4.3)
2

1
=− d 4 x e ψ μ γ μνρ Dν ψρ .
2

Since we are coupling the spin 3/2 field ψμ to gravity, the covariant derivative
in the kinetic term should a priori be the full covariant derivative, ∇, and not just
the Lorentz covariant derivative, D, we have used in the above expressions. The
full covariant derivative ∇ contains both the Levi–Civita connection, Γ , coupling
to the vector index of the gravitino, and the spin connection, ω, coupling to the
spinor index. However, even if we had used ∇, it would appear in the action only in
antisymmetrized form,1

1 ab
∇[ν ψρ] = ∂[ν ψρ] + ω[ν γab ψρ] − Γ[νρ]
σ
ψσ , (4.4)
4
and the last term is identically zero so that ∇[ν ψρ] = D[ν ψρ] , and we can indeed
use the Lorentz covariant derivative D in the kinetic term of the gravitino. In fact,
the Levi–Civita connection in terms of Christoffel symbols will never really appear
in the following.
We now discuss the invariance under supersymmetry of (4.1). We start by making
one simple assumption that is motivated by our discussion in Chap. 2, namely, that
the gravitino transformation rule is proportional to the (covariant) derivative of the
supersymmetry parameter, so that the gravitino can be viewed as the gauge field
of local supersymmetry and the theory is covariant with respect to local Lorentz
transformations:
 
1
δ ψμ = MP Dμ  ≡ MP ∂μ + ωμab γab . (4.5)
4

1 It is important here that Γ really denotes the torsion-free Levi–Civita connection. As we will see
later, it is useful to include a torsion piece bilinear in the gravitini in the spin connection (but not in
the connection Γ , which should stay torsion-free). The connections defined by Γ and ω are then
no longer equivalent connections.
4.1 Pure Supergravity: The Action and SUSY Rules 45

It is worth pointing out that (4.5) is the transformation rule that can also be
inferred from the action of the supersymmetry algebra on the fields as explained
in Appendix 4.A. The conjugate field satisfies
 
1
δ ψ μ = MP ∂μ  −  γab ωμab ≡ MP Dμ . (4.6)
4

Having specified only the gravitino supersymmetry transformation so far, the

next thing we would like to obtain is the transformation rule of the vierbein. We
could simply make an educated guess in line with our considerations leading to
Eq. (2.15), but let us try to actually derive the vierbein transformation law from what
we already have. From the variation of SEH (3.33), we see that the only contribution
with δea comes multiplied by the curvature R ab . We therefore try to single out from
δSRS all possible terms that give the same type of contributions proportional to the
curvature of the spin connection. Supersymmetry invariance will then determine
δea , and we will then check the invariance of the full action.
The variation of the gravitini in the Rarita–Schwinger Lagrangian gives
e e
δLRS = − ψ μ γ μνρ Dν δψρ − δψμ γ μνρ Dν ψρ + . . .
2 2
e e (4.7)
= − ψμ γ μνρ
Dν δψρ − Dν ψρ γ μνρ δψμ + . . . ,
2 2

where we used the identity χγ μνρ λ = λγ μνρ χ for anti-commuting Majorana

spinors, and the dots refer to the variations of the vierbein, δeμa , and the spin
connection, ωμab , which we do not consider for now because they give terms that
are not of the form we need. Inserting (4.5) in (4.7), we obtain
e e
δLRS = −MP ψ γ μνρ Dν Dρ  − MP Dν ψρ γ μνρ Dμ  + . . . (4.8)
2 μ 2
Integrating the last term by parts, we can replace it by
e  e
− ∂μ MP Dν ψρ γ μνρ  + MP Dμ Dν ψρ γ μνρ  (4.9)
2 2

plus terms involving derivatives of the vielbein, Dμ eνa , which we also neglect in this
first step, because they will not give contributions proportional to the curvature R ab .
The equivalence of (4.9) to the last term in (4.8) can easily be checked either by
recalling that the γ -matrices are covariantly constant in the sense that

1
Dμ γ a = ∂μ γ a + ωμ a b γ b + ωμbc [γbc , γ a ] = 0, (4.10)
4
46 4 Pure N = 1 Supergravity in Four Dimensions

or that Dμ (scalar) = ∂μ (scalar). As shown in Exercise 3.4, from the definition of

the curvature, we find

1
[Dμ , Dν ] = Rμν ab γab , (4.11)
4

and therefore, using γab ψρ = −ψρ γab ,

1
D[μ Dν ψρ] = − R[μν ab ψ ρ] γab . (4.12)
8
Hence the variation of the Rarita–Schwinger term becomes
e e
δLRS = − MP ψ μ γ μνρ γab  Rνρ ab − MP ψ ρ γab γ μνρ Rμν ab + . . .
16 16
e  
=− MP ψ μ γ μνρ , γab Rνρ ab + . . .
16
e ρ e μ
= MP ψ μ γ μ  Rνρ ab eaν eb + MP ψ μ γ ν  Rνρ ab eaρ eb + . . . (4.13)
4 2
 
e 1
= − MP ψ μ γ ν  Rν μ − δνμ R + . . .
2 2
e
= − MP ψ μ γ ν Gν μ + . . . ,
2
where the dots contain all terms that do not multiply the curvature of the spin
connection, and we have used the results of Exercise 3.3 in the third line. As
expected, this can be compensated by (cf. Eq. (3.33))

δLEH a
δeμ = −MP2 e eaν Gν μ δeμa , (4.14)
δeμa

which is also proportional to the same combination of the curvature, provided we

define the variation of the vielbein as
1
δeμa = γ a ψμ (4.15)
2MP

(recall that γ a ψμ = −ψ μ γ a ). Note that, proceeding in this way, we did not
simply guess the variation of the vielbein from our considerations in Chap. 2 but
instead really derived it. On the other hand, we see immediately that (4.15) is indeed
consistent with (2.15).
4.1 Pure Supergravity: The Action and SUSY Rules 47

In order to complete the proof of the invariance of the action (4.1), we still need
to discuss the following variations:

δLEH
(i) δ ωμab ;
δωμab

δLRS
(ii) δ ωμab ;
δωμab

δLRS
(iii) δ eμa ;
δeμa

δLRS
(iv) Terms involving Dea from the partial integration in δ ψμ .
δψμ

We also need to understand and specify δ ωab . As we will see, the variation of the
spin connection will depend on the formalism (first, second, or 1.5 order) used to
prove the invariance of the action.
To do this calculation, we go back to the form expression (4.2). The variation of
the action is then

MP2 M2
δS = Dδωab ∧ ec ∧ ed abcd + P R ab ∧ δec ∧ ed abcd
4   2  
B1 A1

i a i
+ δe ∧ ψ ∧ γ5 γa Dψ + ea ∧ δψ ∧ γ5 γa Dψ
(4.16)
2   2  
B2 A2

i i
− ea ∧ ψ ∧ γ5 γa γcd ψ ∧ δωcd + ea ∧ ψ ∧ γ5 γa Dδψ .
 8   2  
B3 A3

We know from previous computations that the term A1, coming from δLEH /δea ,
and the terms involving D 2 , D 2 ψ coming from δLRS /δψ cancel. In detail, the
D 2 -term is A3, where one uses the explicit expression for δψ, and the D 2 ψ-term
can be extracted from A2 using the same steps that also led to (4.15). To do so,
we switch the two-form Dψ and the one-form δψ, using (1.38) and the result of
Exercise 1.1, so that

i a i i
A2 ≡ e ∧ δψ ∧ γ5 γa Dψ = ea ∧ Dψ ∧ γ5 γa δψ = MP ea ∧ Dψ ∧ γ5 γa D
2 2 2
(4.17)
48 4 Pure N = 1 Supergravity in Four Dimensions

and, integrating again by parts,

 
i a i
A2 = −MP d e ∧ Dψ γ5 γa  + MP Dea ∧ Dψ γ5 γa 
2 2  
A2 (4.18)
i
− MP ea ∧ DDψ γ5 γa  .
 2  
A2

The term A2 then cancels A1 and A3 as before, and we are left with δS = B1 +
B2 + B3 + A2 plus boundary terms.
To proceed further, we integrate by parts the term B1 and get
 
M2 MP2 ab
B1 = P δωab ∧ Dec ∧ ed abcd + d δω ∧ e ∧ e abcd .
c d
2 4

In order to write B3 in a very similar form, we write, using the results of the
Exercise 1.1 and Eq. (1.38),

ψ ∧ γ5 γa γcd ψ = ψ ∧ γ5 (γacd + ηac γd − ηad γc )ψ = −i ψ ∧ γ e ψacde (4.19)

so that, after some relabeling and reordering,

1
B3 = − δωab ∧ ψ ∧ γ c ψ ∧ ed abcd . (4.20)
8
Discarding boundary terms and inserting also (4.15) in B2, we then have
 
MP2 ab 1
δS = δω ∧ De − c
ψ ∧ γ ψ ∧ ed abcd
c
2 4MP2 (4.21)
i i
+ γ a ψ ∧ (ψ ∧ γ5 γa Dψ) + MP Dea ∧ Dψ γ5 γa .
4MP 2

This expression
 can be simplified
 by rewriting the second line so that the torsion
piece Dea − 4M1 2 ψ ∧ γ a ψ can also be factored out. While the last term of
P
(4.21) obviously contains a derivative of the vielbein, the other term needs a
reshuffling of the gravitini in order to produce the right bilinear without derivatives.
We can achieve this by using the Fierz identity

1 1  
ψ ∧ψ= ψ ∧ γ a ψ γa − ψ ∧ γ ab ψ γab (4.22)
4 8
4.1 Pure Supergravity: The Action and SUSY Rules 49

and the fact that γ c γ ab γc = 0 and γ c γ a γc = −2γ a :

i i
γ a ψ ∧ (ψ ∧ γ5 γa Dψ) = − (ψ ∧ γa ψ) ∧ (Dψγ5 γ a ). (4.23)
4MP 8MP

Altogether, δS can then be written as

 
MP 1  
δS = De −
a
ψ ∧ γ aψ ∧ iDψγ5 γa  + MP δωbc ∧ ed abcd
2 4MP2

   
MP 1 1
= Dea − ψ ∧ γ aψ ∧ − Dψ γ bcd  + MP δωbc ∧ ed abcd , (4.24)
2 4MP2 6

where we have used γ5 γa = (i/6)abcd γ bcd . At this point the variation of the spin
connection assumes a primary role, and we can try to set (4.24) to zero in various
different ways.

4.1.1 Second-Order Formalism

In this case, one imposes the conventional constraint,

1
Dea = ψ ∧ γ a ψ, (4.25)
4MP2

which determines the spin connection, ωμab = ω̂μab (e, ψ), as the solution to this
equation. The spin connection is thus treated from the very beginning as a dependent
field, whose supersymmetry variation follows from the supersymmetry variations of
eμa and ψμ via the chain rule and the explicit functional dependence of ω̂μab (e, ψ).
By simple inspection of (4.24), however, we see that (4.25) already implies that
δ S = 0, and we don’t really need to know δ ωμab .
Let us nevertheless use
1
Ta = ψ ∧ γ a ψ, (4.26)
4MP2

to solve Dea = T a for the spin connection, which yields

1  [a [a

ω̂μab (e, ψ) = ωμab (e) − 2
ψ γμ ψ b] − ψ μ γ [a ψ b] − ψ γ b] ψμ . (4.27)
4MP
50 4 Pure N = 1 Supergravity in Four Dimensions

This is used in the original approach of Ferrara, Freedman, and van Nieuwenhuizen
in [1]. It is interesting to point out that the supersymmetry variation of this
connection, inherited from the variations of eμa and ψμ , does not contain derivative
terms, ∂, of the supersymmetry parameter:

1  
δ ω̂μab = γ ρ D σ ψ τ 2eρ[a eτb] gμσ − eτ[a eσb] gμρ .
MP

This is the reason why ω̂μab is often called supercovariant.

4.1.2 First-Order Formalism

This is the approach followed by Deser and Zumino in their original paper [2].
Asking for the invariance of the action, δS = 0, via the vanishing of the term in
square brackets in (4.24) fixes the variation of the spin connection to

1 c be 1 b ce
δωμbc = Bμbc − e B + eμ Be , (4.28)
2 μ e 2
with
i
Bμbc = γμ γ5 Dρ ψσ  ρσ bc . (4.29)
2MP

This can be extracted using the same trick that was previously used to derive the
form of the spin connection in terms of the vielbein from the torsion constraint. It
should be noted that (4.28) is not the same as the variation derived using second-
order approach in the previous subsection. However, they become equivalent upon
using the gravitino equations of motion (see Exercise 4.1).

4.1.3 1.5-Order Formalism

In the 1.5-order formalism, one uses the fact that (4.25) can be obtained as a
field equation from varying the action with respect to ωμab (as is obvious from the
terms proportional to δωbc in (4.24)). Thus, when we determine the supersymmetry
variation of the action and require ωμab to be determined by

1 δS
Dea = ψ ∧ γ aψ ⇔ = 0,
4MP2 δωμab

we can immediately drop all terms proportional to δωμab , as these are proportional
to δωδSab , which vanishes on-shell. Obviously, for the simple action we consider here,
μ
the only advantage over the second-order formalism is that we would not have to
4.2 Adding a Cosmological Constant 51

keep track of the δωμab terms in (4.24). Just as in the second-order formalism, the
vanishing of the supersymmetry variation (4.24) is thus obtained by using (4.25),
with the difference that (4.25) is now not imposed by hand but arises as a field
equation for the independent field ωμab . It is in this sense that the 1.5-order formalism
combines elements from the first-order formalism (the a priori independence of
the field ωμab ) and from the second-order formalism (the use of (4.25) for the
cancellation of (4.24)).
It should be stressed that the 1.5-order trick of using on-shell field equations in
the supersymmetry variation can only be used for auxiliary fields such as ωμab .

4.2 Adding a Cosmological Constant

So far we considered the construction of a supergravity action around a Minkowski

background, whose non-linear completion led to Einstein gravity coupled to a
gravitino field without a cosmological constant. In ordinary Einstein gravity,
however, we can always add a cosmological constant Λ to obtain the action
  
√ 1
S = MP2 d 4 x −g R−Λ (4.30)
2

and find a vacuum with

Λ > 0 ↔ de Sitter (dS),

Rμν = Λgμν (4.31)
Λ < 0 ↔ Anti-de Sitter (AdS).

It is natural to ask whether these solutions and the corresponding actions

can be supersymmetrized in a natural way. In this section, we will focus on
pure supergravity theories (without matter multiplets). If we want to construct a
supergravity action generalizing (4.30), we should be able to find a supergroup that
contains the symmetry group of AdS and/or dS spacetime. We will now see that
minimal supersymmetry constrains the closure of the algebra in a way that only one
of the two options is consistent.2

2 There are consistent de Sitter superalgebras with extended supersymmetries, but they do not allow
for positive weight representations, and hence their realizations have the wrong sign in front of the
kinetic terms of some of their fields [3, 4].
52 4 Pure N = 1 Supergravity in Four Dimensions

(A)dS Spacetimes. These spaces are maximally symmetric spaces that can be
expressed as cosets

SO(1, d) SO(2, d − 1)
dSd = , AdSd = . (4.32)
SO(1, d − 1) SO(1, d − 1)

Their geometrical properties can be derived by looking at these spaces

as hypersurfaces embedded in a spacetime with d + 1 dimensions. The
embedding equations can be written as


d−1
∓ (Xd )2 − (X0 )2 + (Xi )2 = ∓2 , (4.33)
i=1

in a flat (d + 1)-dimensional space with signature


d−1
ds 2 = ∓dXd2 − dX02 + dXi2 , (4.34)
i=1

where the upper sign describes the AdS space and the lower sign dS and
where  is a constant of dimension length that parameterizes the radius
of curvature of the (A)dS hypersurface. The d-dimensional metrics can be
obtained by choosing a proper parameterization of the d + 1 embedding
coordinates. A typical (and useful) choice for AdS is given by global
coordinates ρ, τ, ξi (note that time is compact τ ∈ [0, 2π])

X0 =  cosh ρ cos τ, Xd =  cosh ρ sin τ,

 (4.35)
Xi =  sinh ρ ξi , ( i ξi = 1),
2

which give a metric covering once the entire hypersurface (4.33)

ds 2 = 2 (− cosh2 ρ dτ 2 + dρ 2 + sinh2 ρdΩ 2 ), (4.36)

where dΩ 2 is the line element on the (d − 2)-sphere parameterized by

the constrained ξi coordinates. AdS spacetime as given by the hypersurface
(4.33) has closed time-like curves. This can be avoided if one passes over to

(continued)
4.2 Adding a Cosmological Constant 53

the universal covering surface. Another useful choice is given by choosing

coordinates t, r, xi ,

X0 = 1 + r 2 (2 + x 2 − t 2 ) /(2r), Xd =  r t,
Xi =  r xi , (i = 1, . . . , d − 2), (4.37)
Xd−1 = 1 − r 2 (2 − x 2 + t 2 ) /(2r),

which give a metric that only covers half of the hypersurface (4.33) but is
useful in the context of the gauge/gravity correspondence,

dr 2  
ds 2 = 2 2
+ 2 r 2 −dt 2 + d x 2 . (4.38)
r
From these metrics we can also derive the Ricci tensor and hence the relation
between the cosmological constant in (4.31) and the radius of curvature of the
(A)dS hypersurface,

3
Λ=− . (4.39)
2
Similar derivations apply to dS.

Before discussing the corresponding superalgebras, we note that the symmetry

groups of both AdS and dS in d dimensions, SO(1,d) and SO(2,d − 1), respectively,
can be embedded in SO(2, d). For d = 4, the (A)dS algebra is described by ten
anti-Hermitian generators MAB satisfying the commutator relations

[MAB , MCD ] = −2 ηC[A MB]D + 2 ηD[A MB]C , (4.40)

with ηAB = diag{− + + + −} for AdS and ηAB = diag{− + + + +} for dS

space. The explicit (A)dS algebra follows by identifying M5a =  Pa , where we
split A = {a, 5}, with a, b, . . . = 0, 1, 2, 3, and  is the radius of curvature of
(A)dS:

[Mab , Mcd ] = −2 ηc[a Mb]d + 2 ηd[a Mb]c ,

[Pa , Mbc ] = 2 ηa[b Pc] , (4.41)
1
[Pa , Pb ] = ± 2 Mab ,

where the last commutator is equivalent to [M5a , M5b ] = −η55 Mab and η55 = −1
for AdS space and η55 = +1 for dS; hence the upper sign is for AdS and the lower
54 4 Pure N = 1 Supergravity in Four Dimensions

one for dS spacetime. Clearly, when  → ∞, the (A)dS curvature goes to zero, and
one gets back the Poincaré algebra.
To construct the full superalgebra, one needs to specify also the commutators
with the supercharges. In particular, [Pa , Q] cannot be zero anymore, as it used to
be in the super-Poincaré case, because we would no longer close the super Jacobi
identities, as

[[Pa , Pb ], Q] + [[Pa , Q], Pb ] − [[Pb , Q], Pa ] = 0. (4.42)

        
∼Mab 0 0

We therefore need to impose new commutator relations such that the momenta
do not commute with the supercharges. To respect Lorentz covariance and the
graded algebra structure, the result of the commutator should be proportional to
the supercharges and come with some gamma matrices. In principle there are two
possibilities that respect the Majorana condition on the supercharges

[Pa , Q] ∼ γa Q, or [Pa , Q] ∼ γa γ5 Q. (4.43)

In the first case, the coefficient multiplying the right-hand side should be real, while
in the second it should be imaginary. If we use a chiral notation, the sign and the
ambiguities can be reabsorbed in a single dimensionful complex coefficient g̃. We
stress this fact, because there are sometimes wrong statements in the literature about
this. Once we introduce the chiral notation, the new commutators are

g̃ g̃ ∗
[Pa , QR ] = − γa QL , [Pa , QL ] = − γa QR . (4.44)
2 2
Once we introduce these new non-trivial commutators, the super Jacobi identity
can be satisfied, though only for the AdS case. This is readily seen by explicitly
computing the results of the various commutators:

!
0 = [Pa , [Pb , QL ]] + [QL , [Pa , Pb ]] − [Pb , [Pa , QL ]]
|g|2 1
= (γb γa − γa γb ) QL ± 2 [QL , Mab ] (4.45)
4 
|g|2 1
=− γab QL ± 2 γab QL ,
2 2
where we used the commutation relations in (1.53) but with the anti-Hermitian
generators Pa = i Pa and Mab = i Mab . It is now clear that only for the plus
sign we can get a solution:

1
|g|2 = . (4.46)
2
4.2 Adding a Cosmological Constant 55

Hence, only for AdS we can write a consistent supersymmetric completion with a
single supercharge. We finally note that the closure of the super Jacobi identities
requires that another commutator gets modified, namely,

{QL , QL } = g̃ ∗ γ ab Mab . (4.47)

The constraint imposing that the superalgebra can be defined only for the AdS
supergroup and not for dS implies a very important fact: a positive cosmological
constant will always break supersymmetry, while a negative cosmological constant
may be compatible with supersymmetry.
Although matter couplings or extended supersymmetries may allow for de
Sitter vacua in a supersymmetric theory, the vacuum itself will always break
supersymmetry. We will come back to supersymmetry breaking and to the vacuum
selection in Chap. 7.
Once supersymmetry is broken, one could describe this phase of the theory by
using non-linear realizations, as it is customary for any other symmetry whose
linear action is broken. This has been the subject of intense scrutiny (see, for
instance, [5–7]), and one can indeed write actions for theories with dS vacua where
supersymmetry is non-linearly realized. Since an effective discussion of this topic
requires some additional technical introduction to superfields in supergravity, we
will not deal with it here but refer the reader to the literature on the subject, such
as [8].

4.2.1 Construction of the Action

Now that we established that the anti-de Sitter group can be consistently extended
to a supergroup, we would like to realize it in terms of a supersymmetric action that
includes a negative cosmological constant. We will proceed in a fashion similar to
what has been done in the flat case, starting from the supersymmetry transformation
of the gravitino and then trying to close the action of the supersymmetry transforma-
tion on the free Lagrangian for the gravity multiplet, possibly introducing interaction
terms. We therefore need to fix first the supersymmetry transformation rule of the
gravitino. Since the algebra has been modified with respect to the case without
cosmological constant, we expect that also the supersymmetry transformations get
modified accordingly.
As detailed in Appendix 4.A, we can generically deduce the supersymmetry
transformation properties of the various fields from the structure constants of the
superalgebra. In fact, gravity in flat space can be thought of as a gauge theory of
the Poincaré group (and by extension supergravity of the super-Poincaré group),
but with some constraints needed to relate the vielbein and the spin connection
degrees of freedom. From this point of view, the vielbein and the spin connection
can be thought of as gauge fields for the translation generators and for the Lorentz
rotations, while gravitini are the gauge fields for the supersymmetry generators. The
56 4 Pure N = 1 Supergravity in Four Dimensions

transformation rules of the same fields under the various generators of the “gauge
algebra” are then fixed by its structure constants in the usual way.
By using this trick, we can now deduce the supersymmetry transformation of
the gravitino by looking at the structure constants of the AdS superalgebra coming
from commutators which have a supersymmetry generator on the right-hand side.
This inspection shows that a new term in the supersymmetry transformation of
the gravitino should appear because of the non-zero commutator (4.44) between
the translation generators and supersymmetry generators. If one interprets the
spin connection term in the Lorentz covariant derivative in the original gravitino
transformation (4.5) as due to the non-vanishing commutator of Mab with Q, the
new non-vanishing commutator (4.44) between Pa and Q should then analogously
lead to an additional contribution to the gravitino transformation so as to make the
transformation covariant with respect to the full AdS isometry group. In δψμL , this
additional contribution should be of the form given in the first equation of (4.44)
contracted with the gauge field of the translation generator Pa , i.e., with the vierbein
eμa . We therefore should have

g 2
δψμL = MP Dμ L − M γμ R , (4.48)
2 P
where now we have a dimensionless constant g ∈ C, because of the introduction
of the dimensionful MP and MP2 factors. Clearly, this has to go together with the
conjugate relation:

g∗ 2
δψμR = MP Dμ R − M γμ L . (4.49)
2 P
Once again we stress that g here can be any complex number, because there are
sometimes wrong statements in the literature about this.
To construct the action, we start from the action (4.1) with the vierbein
transformation rule
1
δeμa =  L γ a ψμR + h.c.,
2MP

and (4.48) for the gravitino

g 2
δψμL = MP Dμ L − M γμ R .
2 P
The reason we start from the action without the cosmological constant, rather than
adding explicitly the cosmological constant among the bosonic terms right from the
beginning, is that it will automatically be enforced by supersymmetry in an iterative
procedure at higher order in g, as will become clear momentarily.
The gravitino relation differs from the one in (4.5) by a shift term proportional
to the constant g. Clearly this shift breaks the supersymmetry of the original action
4.2 Adding a Cosmological Constant 57

(4.1), and we need to restore it by adding additional terms to it. In the following,
we will establish again the invariance under supersymmetry of a modified action.
To our knowledge this was first done in [9].
The first supersymmetry breaking effect of the shift term (proportional to g) is
that of generating new terms in the variation of the Rarita–Schwinger part of the
Lagrangian. To compute these terms, we use the supersymmetry variation of the
conjugate gravitino, which, in form notation, reads

g∗ 2
δψ R = MP D R + M  L γa e a , (4.50)
2 P
as one may easily verify. Denoting by δg the variations due to the O(g) shift
term in the gravitino transformation law, the uncancelled variation of LRS under
supersymmetry is then

i a i
δg LRS = e ∧ δg ψ R ∧ γ5 γa DψL + ea ∧ ψ L ∧ γ5 γa Dδg ψR + h.c.
2 2
i ∗ 2 a
= g MP e ∧ eb ∧  L γb γ5 γa DψL
4
i  
− g ∗ MP2 ea ∧ ψ L ∧ γ5 γa D eb γb L + h.c.
4
i ∗ 2 a i
= g MP e ∧ eb ∧  L γ5 γab DψL − g ∗ MP2 ea ∧ Deb ∧ ψ L γ5 γa γb L
4 4
i ∗ 2 a
− g MP e ∧ eb ∧ ψ L ∧ γ5 γab DL + h.c. (4.51)
4
Integrating by parts the first term in the last equality, we get
 
i ∗ 2 a 1
δg LRS = g MP e ∧ De ∧  L γ5 γab ψL − ψ L γ5 γa γb L
b
2 2
i
− g ∗ MP2 ea ∧ eb ∧ ψ L ∧ γ5 γab DL + h.c. (4.52)
2
The first term, proportional to Dea , plays a similar role as in the case without
a cosmological constant and will be discussed later after Eq. (4.59). Since the
remaining terms are proportional to the derivative of the supersymmetry parameter,
we can try and use the supersymmetry transformation rule of the gravitini, δψL =
MP DL + O(g), to cancel them. For this reason we add a mass-like term to the
Lagrangian:

i ∗ i
LMψ = g MP ea ∧ eb ∧ ψ L ∧ γ5 γab ψL + g MP ea ∧ eb ∧ ψ R ∧ γ5 γab ψR ,
4 4
(4.53)
58 4 Pure N = 1 Supergravity in Four Dimensions

so that the variation of ψ in (4.53) compensates for (4.52) at order g. We point

out here the reduced Planck mass factors, so that the whole coefficient has mass
dimension 1, as well as the overall factor 1/4, which is due to the double variation
required to match (4.52).
While the introduction of (4.53) allows the cancellation of the Dε terms in (4.52),
it also gives rise to two further new variations we have to take care of. One variation
comes at order g from the variation of the vierbein in (4.53). We will discuss its
cancellation together with the cancellation of the Dea terms in (4.52) further below.
The other new variation of (4.53) is of order g 2 and arises when the order g shift
term in the gravitino transformation is used in the variation of the gravitini in (4.53).
Indeed, the order g 2 variation of the gravitino mass term produces (suppressing the
wedges)

i
δg LMψ = − |g|2 MP3 ea eb ψ L γ5 γab ec γc R −  R ec γc γ5 γab ψL + h.c. (4.54)
8
Putting together the gamma matrices and using the duality relation γ5 γabc =
−i abcd γ d , we obtain

|g|2 a b c  
δg LMψ = MP3 e e e abcd ψ L γ d R + ψ R γ d L
4
|g|2 a b c  R γ d ψL +  L γ d ψR
= −MP4 e e e abcd
2 2MP (4.55)
|g| 2
= −MP4 ea eb ec abcd δed
2
M4  
= −|g|2 P δ ea eb ec ed abcd .
8

We further recall that ea eb ec ed abcd = +4! d 4x e and then realize that we need
to add a single term of order |g|2 to the Lagrangian to cancel (4.55):
 
3 d 4 x e MP4 |g|2 = −MP2 d 4 x e Λ. (4.56)

This is a cosmological constant term. Notice that there is no choice of the sign of
this cosmological constant

3
Λ = −3 MP2 |g|2 = − < 0. (4.57)
2
This agrees with the discussion following from the supersymmetry algebra.
It is also extremely important to note that the variation of (4.56) does not generate
terms of order g 3 and that therefore supersymmetry closes at order g 2 .
4.2 Adding a Cosmological Constant 59

The only variation left is the vierbein variation in the gravitino mass term together
with the already mentioned first term in (4.52). Using

4(ψ R ∧ γab ψR ) ∧ (εγ b ψ) = 3(ψ R γab εR ) ∧ (ψ ∧ γ b ψ) + (ψ R εR ) ∧ (ψ ∧ γa ψ),

(4.58)
which one can derive from the Fierz identities (1.48)–(1.52), one obtains for all
remaining uncancelled variations
 
MP 1
δL = De −
a
ψ ∧γ ψa
2 4MP2
  
1
∧ i g ∗ MP eb ∧  L γba ψL − ψ L γb γa L
2
 
1
−i g MP eb ∧  R γba ψR − ψ R γb γa R
2
 
1
+abcd − Dψ γ  + MP δω ∧ e
bcd bc d
, (4.59)
6

where the last term proportional to εabcd is the same as in the case without
cosmological constant.
This completes the proof of the invariance of the action in any of the formalisms
described above. In the second-order formalism, this variation vanishes because of
the torsion constraint. In the first-order formalism, we deduce from this variation the
expression for δωab that makes it vanish. Finally, in the 1.5 formalism, the equations
of motion for the spin connection do not change, and hence once again the full
Lagrangian is invariant under supersymmetry.
The final Lagrangian is therefore the following

MP2 e e
L = eR − ψ μR γ μνρ Dν ψρL − ψ μL γ μνρ Dν ψρR
2 2 2
g g∗
−e MP ψ μR γ μν ψνR − e MP ψ γ μν ψνL (4.60)
2 2 μL

+3 e MP4 |g|2 ,

where we should remember that in the second-order formalism ωab = ωab (e, ψ),
and the final supersymmetry transformations are

1
δeμa =  L γ a ψμR + h.c., (4.61)
2MP
g 2
δψμL = MP Dμ L − M γμ R . (4.62)
2 P
60 4 Pure N = 1 Supergravity in Four Dimensions

Now that we completed the construction of a supersymmetric action for super-

gravity with a (negative) cosmological constant, we can make some comments.
First of all, we have seen from the construction that we have performed only
minimal modifications. After shifting the supersymmetry transformation of the
gravitino field, we introduced the smallest set of terms needed to cancel O(g)
and O(g 2 ) terms in the supersymmetry variations. As already pointed out above,
supersymmetry closes at order |g|2 . There is no need to introduce any term of order
g 3 or more.
All the modifications can be summarized in three main pieces:

• A shift in the fermionic supersymmetry rules at O(g)

• A mass-like term of O(g) for the fermions
• A potential term at order O(g 2 )

Although these modifications have been forced by the presence of the cosmological
constant in a pure gravity theory, we will see that the pattern outlined above
is actually underlying all gauged supergravity theories for models with extended
supersymmetries. In fact, as we will see, in extended supergravities the appearance
of non-Abelian gauge groups is tied to the presence of a non-trivial scalar potential,
which may act as an effective cosmological constant. The result is that the gauging
procedure introduces the same three main modifications listed above, where the
mass-like term for the fermions in the general case with scalar fields becomes
a Yukawa-like coupling and the cosmological constant term becomes a scalar
potential (see Sect. 9.1).
It can also be seen that this scalar potential (in this case a pure cosmological
constant) can be expressed as the square of the shifts of the supersymmetry
variations of the fermionic fields:
3 μ
V  L γ a R = −3MP4 |g|2  L γ a R = − δg ψ μR γ a δg ψL . (4.63)
2
Note the minus sign in front of the squared gravitino shifts. This identity is called
the supersymmetric Ward identity [10].

4.2.2 Mass in AdS

A further crucial point in establishing susy invariance is that the gravitino has to
have a mass term proportional to the square root of the cosmological constant

g g∗
− MP ψ μR γ μν ψνR − MP ψ μL γ μν ψνL . (4.64)
2 2
4.2 Adding a Cosmological Constant 61

So we have a massless graviton gμν , a “massive” gravitino ψμ , but we still preserve

supersymmetry! This is a consequence of a general point that should always be
remembered in supergravity: Supersymmetry multiplets are degenerate in mass only
in Minkowski space. On curved spaces we can have different “masses” for different
components of the same multiplet, where the quotes here refer to the fact that the
whole concept of mass becomes somewhat ambiguous in curved spacetimes when
the masses are comparable to the curvature scale.
Again, the superalgebra construction comes to our help in explaining this fact.
Without a cosmological constant, supergravity is invariant under the super-Poincaré
group. In the corresponding superalgebra, the supersymmetry generator and the
momentum commute, [Q, P ] = 0. This means that also P 2 commutes with the
supersymmetry generator,

[P 2 , Q] = 0,

and hence P 2 , defining the mass, is a good Casimir operator, and it will take the
same value on all the fields of the same supersymmetry multiplet.
This is not true for the AdS superalgebra. In this case P 2 does not commute with
the supersymmetry generator, and therefore it is not the same for states in the same
multiplet. Actually, P 2 is not even an SO(2,3) invariant, and therefore we cannot
even classify states by that quantity.
There is, however, another good invariant for SO(2,3), which is defined as

1
C = − MAB M AB , (4.65)
2
where the SO(2,3) generators are taken anti-Hermitian
†
MAB = −MAB (4.66)

and are represented by MAB = {Mab , M5a =  Pa }. Since C is a good invariant,

we can classify states by its value. It would then be important to understand if we
can associate to the values of C = − 12 Mab M ab + 22 Pa P a something we can
call mass and spin of the field. SO(2,3) has SO(2) × SO(3) as the maximal compact
subgroup. The states can thus be labeled by E and s, where E is the SO(2)  U(1)
energy eigenvalue and s is the spin. Recall that global AdS has closed time-like
curves.3 Hence the generator of translations along the time-like circles, namely,
M50 , can be interpreted as a dimensionless energy operator

H = i P 0 = −i M50 (4.67)

3 In practice, we always consider its universal covering space, CAdS, as mentioned earlier.
62 4 Pure N = 1 Supergravity in Four Dimensions

and we call its eigenvalues E. The generators of three-dimensional spatial rotations

are the generators Mij , for i, j = 1, 2, 3, which commute with H . The remaining
generators can be combined into three pairs of mutually conjugate operators

L±
i = −i M0i ± M5i , (4.68)

so that the commutators become

[H, L± ±
i ] = ±Li , (4.69)

[L+ −
i , Lj ] = −2H δij − 2Mij , (4.70)

[L± ±
i , Lj ] = 0. (4.71)

Clearly, the generators L± i are playing the role of raising (lowering) operators for
E. When applied to an eigenstate of H with eigenvalue E = E0 , they give states
with eigenvalues E = E0 ± 1. Assuming that the spectrum of H is bounded
from below, the lowest eigenvalue E0 is realized by a state |E0 , s0  satisfying
L−i |E0 , s0  = 0, for any i. From this ground state, one can then construct the other
states by application of products of creation operators L+ i .
In this basis, the Casimir operator can be written as (sum over i implied)

1 1
C = H 2 + J 2 − L+ L− − L− L+ , (4.72)
2 i i 2 i i

where J 2 = − 12 Mij M ij . Using the commutator relations above and Mii = 0, this
simplifies to

C = H (H − 3) + J 2 − L+ −
i Li . (4.73)

On the lowest energy state, we obtain that

1
C |E0 , s0  = − MAB M AB |E0 , s0  = [E0 (E0 − 3) + s0 (s0 + 1)] |E0 , s0 .
2
(4.74)

Actually, any state belonging to the corresponding unitary irreducible representation

must satisfy

E, s|C |E, s = [E0 (E0 − 3) + s0 (s0 + 1)] E, s|E, s, (4.75)

precisely because C is a Casimir operator of the AdS algebra. This allows us to

obtain a unitarity constraint on the allowed quantum numbers. In order to get it, we
use (4.75) for an excited state with energy E = E0 +1 and spin s = s0 −1 (assuming
4.2 Adding a Cosmological Constant 63

the ground state has s0 ≥ 1) and compare the result with the direct computation of
C on the same state:

(4.75) 2
E0 + 1, s0 − 1|C |E0 + 1, s0 − 1 = N [E0 (E0 − 3) + s0 (s0 + 1)]
(4.73)
= E0 + 1, s0 − 1|(E0 + 1)(E0 − 2) + (s0 − 1)s0 |E0 + 1, s0 − 1
−|L−
i |E0 + 1, s0 − 1| ,
2

(4.76)

where N 2 = E0 + 1, s0 − 1|E0 + 1, s0 − 1 is the norm squared of the state. This

implies that

|L−
i |E0 + 1, s0 − 1| = 2(E0 − s0 − 1) N
2 2
≥0 (4.77)

and shows that in order to have unitary representations, one has to require

E0 ≥ s0 + 1. (4.78)

Actually, the AdS algebra has a special unitary representation, which is the
singleton,4 for s0 = 0 and E0 = 1/2 (together with its partner s0 = 1/2, E0 = 1
state in the supersymmetric case), while all the other representations have to fulfill
the above unitarity bound [12].
If we come to the definition of mass, we should stress that there is no
unambiguous definition. In fact, there are two main different conventions in the
literature.
Most of the older supergravity literature identifies massless representations with
those living at the boundary of the unitarity bound threshold, i.e., with E0 = s0 + 1.
For the ground states, this means

C |m = 0, E0 = s0 + 1, s0 
1
= − MAB M AB |m = 0, E0 = s0 + 1, s0 
2
= [(s0 + 1)(s0 − 2) + s0 (s0 + 1)]|m = 0, E0 = s0 + 1, s0 
= 2(s02 − 1)|m = 0, E0 = s0 + 1, s0 . (4.79)

4 The singleton representation is a very special representation of the AdS algebra, with no Poincaré
counterpart. Its four-dimensional field representations are pure gauge degrees of freedom and do
not propagate in the bulk of spacetime. However, any massless field in AdS can be constructed by
taking the product of two of them [11].
64 4 Pure N = 1 Supergravity in Four Dimensions

Setting therefore m2 = 0 for these representations, one can define the mass in AdS
as
1
m2 2 = − MAB M AB − 2(s02 − 1) = E0 (E0 − 3) − (s0 + 1)(s0 − 2). (4.80)
2
Another natural convention that is widely used more recently, especially after
the birth of the AdS/CFT correspondence, is to simply identify the mass with the
result of the direct computation of the Laplace–Beltrami operator. In this case, the
relations between the masses and the (E0 , s0 ) quantum numbers are [13, 14]

m20 2 = E0 (E0 − 3), (4.81)

 
 3 
m1/2  = E0 − . (4.82)
2

m21 2 = (E0 − 2)(E0 − 1), (4.83)

 
 3 
m3/2  = E0 − , (4.84)
2

m22 2 = E0 (E0 − 3). (4.85)

In the case of scalar fields, (4.81) gives an interesting outcome if we invert the
relation and write the energy E0 as a function of the mass

3 9
E0 = ± + m2 2 . (4.86)
2 4

It is easy to see that in order to satisfy the unitarity bound E0 ≥ 1, we can have both
the plus and minus signs, as long as one allows negative squared masses, m2 < 0,
satisfying

9 1 3
m2 ≥ − 2
= − |Λ|. (4.87)
4 4
This is the Breitenlohner–Freedman bound [15], and it is easy to check that it
corresponds to a minimum of the Casimir operator for s = 0. This bound is
very important in establishing the stability of vacua with a negative cosmological
constant. In generic field theories, scalar fields with negative mass signal instabilities
of the vacuum against small fluctuations of the same fields. In a theory with
many scalar fields, this is related to the fact that the scalar potential is at a local
maximum or a saddle point rather than at a local minimum. However, for a gravity
theory, a maximum or a saddle point of the potential with a negative value of the
cosmological constant can be stable against small fluctuations, provided the negative
4.A Appendix: Gauging the Poincaré Algebra 65

eigenvalues of the mass matrix satisfy the Breitenlohner–Freedman bound, so that

the fluctuations do not violate unitarity. The outcome of this discussion is once more
that, whatever convention we use for the definition of mass, the true parameters that
unambiguously label the irreducible representations are (E0 , s0 ), while m is just a
bookkeeping parameter related to the wave equation or to the AdS algebra.
Everything we discussed so far is valid in any matter coupled gravity theory
around an anti-de Sitter vacuum. Supersymmetry further imposes additional inter-
esting constraints. For instance, supersymmetric vacua with a negative cosmological
constant always satisfy the Breitenlohner–Freedman bound, and therefore they
always provide stable vacua, also for potentials unbounded from below (which is
often the case in supergravity). Moreover, fields come organized in supermultiplets.
In four dimensions (with minimal supersymmetry), there are three possible multi-
plets [12]:

• The singleton multiplet: this representation contains only one bosonic and one
fermionic AdS representation with ground states

|E0 = 1/2, s0 = 0, |E0 = 1, s0 = 1/2. (4.88)

• Regular multiplets: here the ground states are

|E0 = Δ, s = s0 , |Δ + 1/2, s0 + 1/2, |Δ + 1/2, s0 − 1/2, |Δ + 1, s0 ,

(4.89)

and they form massless multiplets if E0 = s0 + 1; otherwise they form massive

multiplets. Notice also that the states with spin s0 −1/2 are obviously not present
in the case of chiral multiplet representations.

We therefore see that, for instance, according to (4.81)–(4.85), a massive vector

multiplet whose lowest energy state has E0 = 5/2 and s0 = 1/2 contains a massless
scalar field, two fermions of masses m = 1 and m = 2, and a vector field of mass
squared m2 2 = 2.

4.A Appendix: Gauging the Poincaré Algebra

When introducing general relativity as well as supergravity, we discussed the possi-

bility of considering the vierbein and the spin connection as independent quantities.
Do we have any conceptual reason behind this, in addition to the simplification of
some computations? We will now see that an interesting perspective on gravity,
which can help when dealing with supergravity, is that of considering gravity
itself as a sort of a gauge theory where the gauge group is the Poincaré group.
This analogy will work only to a certain extent, but it will be very useful for
understanding many specific new features that have to be introduced when one
wants to promote supersymmetry to a local symmetry of nature. In fact, supergravity
66 4 Pure N = 1 Supergravity in Four Dimensions

is the gauge theory of supersymmetry, and therefore there must be a way to describe
it as a theory where the gauge group is the Poincaré supergroup (or some other
supergroup). For the sake of simplicity in this appendix, we set MP = 1.
Consider an ordinary gauge transformation δ =  A TA , where TA are the gauge
generators satisfying

[TA , TB ] = fAB C TC .

If this is a global symmetry of an action, it can be made local by introducing vector

fields AA μ for each symmetry so that the algebra
 
[δ(1A ), δ(2B )] = δ 2B 1A fAB C , (4.90)

where fAB C are the structure constants, has a faithful realization on them,

μ = ∂μ  +  Aμ fBC ,
δ AA A C B A
(4.91)

and we can introduce covariant derivatives

Dμ = ∂μ − AA
μ TA (4.92)

acting non-trivially on fields which transform in non-trivial representations of the

gauge group. The curvature, defined as

[Dμ , Dν ] = −Fμν
A A
T ⇔ A
Fμν ≡ 2∂[μ AA
ν] + Aμ Aν fBC ,
B C A
(4.93)

transforms covariantly:
A
δ Fμν =  C Fμν
B
fBC A . (4.94)

Let us now imagine that we want to make local the symmetries of the Poincaré
group. The usual procedure is to introduce gauge fields in correspondence with
the generators of the algebra. For the Poincaré algebra, this means introducing two
fields, one eμa with a vector gauge index and one ωμab with two, so as to match the
gauge generators

1 ab
μ TA = eμ Pa +
AA a
ω Mab . (4.95)
2 μ

The gauge curvatures of these vectors are precisely T a and R ab as defined in (3.22)
and (3.23), where the spin connection and the vielbein are so far independent fields.
4.A Appendix: Gauging the Poincaré Algebra 67

This construction is perfectly legitimate. However, it clearly leads to an ordinary

gauge theory and not to a gravity theory as we would like. From the T a and R ab
curvatures, we could construct kinetic terms giving the propagation of independent
degrees of freedom and discuss the resulting gauge theory, where the Poincaré group
is realized on the vector fields as

δP ωab = 0, (4.96)

δP ea = D a , (4.97)

δM ωab = dΛab − ωa c Λcb − Λac ωc b , (4.98)

δM ea = Λa c ec . (4.99)

If, on the other hand, we want to get only the metric degrees of freedom, we have
to impose a constraint between ωab and ea . This is the conventional constraint or
torsion constraint

T a = 0. (4.100)

This constraint however is not invariant under (4.96)–(4.97):

δP T a = δP Dea = DδP ea = DD a = −R a b  b = 0. (4.101)

This means that if we impose the conventional constraint, translation symmetry is

broken. Moreover, it is also clear that now the spin connection ωab cannot be treated
as independent of the vielbein anymore and hence the transformation (4.96) will not
be valid anymore. Since ωab = ωab (e), the spin connection is also not invariant
under translations

δωab
δP ω ab
= d 4x δP ec = 0.
δec

The final outcome of this discussion is that, when the conventional constraint
is imposed, the Poincaré gauge algebra is deformed and translational symmetry
is replaced by a new invariance under diffeomorphisms. This can be seen by
considering the commutator of two translation generators on the vielbein:

[δP 2 , δP 1 ]ea = −δP [2 (D1]

a
) = −δP [2 ωa c 1]
c
, (4.102)

and now δP ωab = 0. The resulting algebra then has a non-vanishing commutator

[P , P ] = 0,
68 4 Pure N = 1 Supergravity in Four Dimensions

as is appropriate for general coordinate transformations, which do not commute.

Actually one can check that using the constraint (4.100), the translation generators
on the vielbein take the form of general coordinate transformations.

The Lie derivative of a p-form Ap along the flow of a vector field V is defined
as
1 ∗
LV Ap = lim σ Ap (σt (x)) − Ap (x) ,
t →0 t t

where σt∗ is the pullback of the differential form along the flow generated by
the vector field V . When applied to a scalar valued p-form, this reduces to

LV Ap = (ıV d + dıV )Ap .

The action constructed from the curvatures and the vielbein then is invariant
with respect to local Lorentz group transformations and diffeomorphisms. The
infinitesimal change of a function under a diffeomorphism is given by the Lie
derivative L , and therefore the action is going to be invariant if
 
L S = d(ı L ) + ı dL = 0,

but the first term is a total derivative that can be discarded while the second is zero
because dL has one degree more than the top form. Finally, in the construction
of an action, we will not make use of a kinetic term of the form R ab ∧ Rab
because of the conventional constraint which makes it quartic in the derivatives.
The appropriate quadratic term is the Einstein–Hilbert action above.
The method we have outlined in this section can be easily extended to generic
supergravity theories by extending the Poincaré algebra to the super Poincaré
algebra, by including fermionic generators and possibly other bosonic generators
for the internal symmetries. The power of this approach lies in the ease of guessing
the transformation laws under the various symmetries, including supersymmetry.
This means that this approach can be used as a guide to derive and construct the
Lagrangian and/or the equations of motion of systems respecting any symmetry
group we would like to realize. Once again, we stress that one has to be careful with
its application because of the constraints that will be needed to obtain a consistent
gravity theory (invariant under diffeomorphisms). Imposing these constraints will
break the transformation rules that do not preserve them.
Exercises 69

4.A.1 Gauging the Super Poincaré Algebra

We end this appendix with a few remarks on the gauging of the super Poincaré
group. We could gauge this algebra by adding new vector fields ψμA for the
fermionic generators QA . From the algebra we then have

1 ab A
μ TA = eμ Pa +
AA ω Mab + ψ μA QA + ψ μ QA ,
a
(4.103)
2 μ
and we can read the supersymmetry transformations by applying

μ = ∂μ  +  Aμ fBC .
δ A A A C B A

For instance, for N = 1 supergravity, we would get that the spin connection is
invariant,

δ ωab = 0, (4.104)

because the Lorentz generator never appears on the right-hand side of any com-
mutator involving the supersymmetry generator. However, just like for the bosonic
case, we should impose a torsional constraint in order for the vielbein and spin
connection not to be independent. Doing so, we fix the form of the spin connection
as ωab = ωab (e, ψ) and check the new realization of the algebra on the fields.
One last interesting remark involves the definition of the gauge curvatures for
the super-Poincaré algebra. From the structure constants of the supersymmetry
algebra (1.53), one can deduce a new definition for the curvatures, including the
one of the translation generators, namely, the torsion T a . Since the translation
generators Pa appear on the right-hand side of the commutator of two supercharges,
the corresponding curvature definition is now

1 A a
T a = Dea − ψ γ ψA (4.105)
4
and involves a fermion bilinear. This means that imposing the conventional con-
straint T a = 0 results in a spin connection depending on the gravitino fields. Hence,
supergravity is often referred to as a theory with non-trivial torsion for the spin
connection, because T a = 0 implies Dea = 0.

Exercises

4.1. Prove that the supersymmetry variation of the spin connection in the first-order
formalism and in the second-order formalism is equivalent upon using the gravitino
equation of motion.
70 4 N = 1 Supergravity in Four Dimensions

4.2. In the first-order formalism, compute the new piece in the variation of the spin
connection due to the cosmological constant.

4.3. Find the embedding coordinates that give origin to

ds 2 = −dt 2 + e2t / d x 2 ,

for de Sitter and to

 
ds 2 = −dt 2 + 2 sin2 (t/) dψ 2 + sinh2 ψ dΩ 2 ,

for anti-de Sitter spacetime. Discuss the “cosmological” meaning of these metrics.

4.4. Check that the gauge curvatures T a and R ab coming from (4.95) and (4.93)
match (3.22) and (3.23).

References
1. D.Z. Freedman, P. van Nieuwenhuizen, S. Ferrara, Progress toward a theory of supergravity.
Phys. Rev. D13, 3214–3218 (1976)
2. S. Deser, B. Zumino, Consistent supergravity. Phys. Lett. B62, 335 (1976)
3. S. Ferrara, Algebraic properties of extended supergravity in de Sitter space. Phys. Lett. B69,
481 (1977)
4. K. Pilch, P. van Nieuwenhuizen, M.F. Sohnius, De Sitter superalgebras and supergravity.
Commun. Math. Phys. 98, 105 (1985)
5. E.A. Bergshoeff, D.Z. Freedman, R. Kallosh, A. Van Proeyen, Pure de Sitter supergrav-
ity. Phys. Rev. D 92(8), 085040 (2015) [Erratum: Phys. Rev. D 93(6), 069901 (2016)]
[arXiv:1507.08264 [hep-th]]
6. G. Dall’Agata, E. Dudas, F. Farakos, On the origin of constrained superfields. JHEP 05, 041
(2016) [arXiv:1603.03416 [hep-th]]
7. N. Cribiori, G. Dall’Agata, F. Farakos, From linear to non-linear SUSY and back again. JHEP
08, 117 (2017) [arXiv:1704.07387 [hep-th]]
8. J. Wess, J. Bagger, Supersymmetry and Supergravity (Princeton University Press, Princeton,
1992)
9. P.K. Townsend, Cosmological constant in supergravity. Phys. Rev. D15, 2802–2804 (1977)
10. S. Cecotti, L. Girardello, M. Porrati, Constraints on partial superhiggs. Nucl. Phys. B268, 295–
316 (1986)
11. M. Flato, C. Fronsdal, Lett. Math. Phys. 2, 421 (1978); Phys. Lett. 97B, 236 (1980); J. Geom.
Phys. 6, 294 (1988)
12. H. Nicolai, Representations of supersymmetry in Anti-de Sitter space, in *Trieste 1984,
Proceedings, Supersymmetry and Supergravity ’84*, 368–399 and CERN Geneva – TH. 3882
(84, REC.JUL.) 32 p.
13. S. Ferrara, C. Fronsdal, A. Zaffaroni, On N = 8 supergravity on AdS(5) and N = 4 superconfor-
mal Yang-Mills theory. Nucl. Phys. B 532, 153–162 (1998) [arXiv:hep-th/9802203 [hep-th]]
14. O. Aharony, S.S. Gubser, J.M. Maldacena, H. Ooguri, Y. Oz, Large N field theories, string
theory and gravity. Phys. Rept. 323, 183–386 (2000) [arXiv:hep-th/9905111 [hep-th]]
15. P. Breitenlohner, D.Z. Freedman, Positive energy in anti-De Sitter backgrounds and gauged
extended supergravity. Phys. Lett. B 115, 197–201 (1982)
 Part II
Matter Couplings and Phenomenology
Matter Couplings in Global Supersymmetry
5

In the previous chapter, we discussed N = 1 pure supergravity, which only

involves the N = 1 supergravity multiplet. In order to include also ordinary
matter fields and Yang–Mills-type gauge interactions, one has to couple the pure
supergravity sector to N = 1 chiral and vector multiplets in a way consistent
with local supersymmetry. Some aspects of the coupling of free chiral multiplets to
supergravity were already sketched in Chap. 2. It is the purpose of the following two
chapters to extend this preliminary discussion to a more systematic and complete
analysis of the general matter couplings in 4D, N = 1 supergravity. As a
preparation, we first discuss, in this chapter, general matter couplings in global
supersymmetry.

5.1 Our Approach

Our way to introduce matter couplings in global and local supersymmetry is

motivated by two main goals:

• Derive general matter couplings in supergravity in an as direct and physical way

as possible so that it can also be readily generalized to extended supergravity and
higher dimensions.
• Exhibit the differences between matter couplings in global and local supersym-
metry as clearly as possible.

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 73

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_5
74 5 Matter Couplings in Global Supersymmetry

In order to reach the first goal, we will consistently work with on-shell com-
ponent fields and four-component spinors, using a combination of the Noether
method and geometrical reasoning, i.e., methods that are easy to generalize also to
extended supergravity and other spacetime dimensions with a minimal theoretical
apparatus. In order to reach the second goal, we will introduce these methods and
our conventions already at the level of global supersymmetry, i.e., in a context most
readers will have some familiarity with. This also includes a discussion of non-
renormalizable interactions already at the level of global supersymmetry. Most of
these non-renormalizable interactions can be nicely packaged in terms of elegant
geometrical structures that greatly facilitate the transition to local supersymmetry.
Our approach is to iteratively construct the most general matter couplings by first
discussing the most important building blocks in isolation and then show how they
can be patched together to obtain the general theory. We will do this first for global
supersymmetry and then, later, in a completely analogous way also for supergravity.

5.2 Chiral Multiplets in Global Supersymmetry

In this section, we discuss the possible couplings of chiral multiplets in global

supersymmetry.

5.2.1 The Renormalizable Wess–Zumino Model

We already encountered the Wess–Zumino model in Sect. 2.1. This model can
be generalized in various ways. The most general renormalizable Wess–Zumino
model describes nC chiral multiplets (φ m , χ m ) (m = 1, . . . , nC ) whose mass and
interaction terms (Yukawa interactions as well as cubic and quartic scalar potentials)
are all encoded in a single holomorphic function, W (φ m ), the superpotential.
Renormalizability restricts W (φ m ) to be a polynomial of at most cubic order. In
the following, we use barred indices m, m, . . . to denote the complex conjugates of
the scalar fields, φ m , and the right-handed projections of the Majorana fermions,
χ m , and write:

φ n ≡(φ n )∗
(5.1)
∂ ∂
∂m ≡ , ∂m ≡
∂φ m ∂φ m

χRm ≡ PR χ m , χLm ≡ PL χ m . (5.2)

5.2 Chiral Multiplets in Global Supersymmetry 75

The Lagrangian then is1,2

 
L = − δmn (∂μ φ m )(∂ μ φ n ) + χ m
L
/
∂ χ n
R + χ n/ m
R ∂ χ L

∗ m n (5.3)
− (∂m ∂n W )χ m
L χL − (∂m ∂n W )χ R χR
n

− VF ,

with supersymmetry transformations given by

δφ m =  L χLm , (5.4)

δφ m =  R χRm , (5.5)

1 m 1
δχLm = ∂/φ R − δ mn (∂n W ∗ )L , (5.6)
2 2
1 m 1
δχRm = ∂/φ L − δ mn (∂n W )R . (5.7)
2 2
The scalar potential in (5.3) is

VF = δ mn (∂m W )(∂n W ∗ ), (5.8)

which is manifestly non-negative. This potential is called F-term potential, as it

arises, in an off-shell formulation, from the auxiliary fields of the chiral multiplets,
which are often denoted as F m and which are on-shell proportional to δ mn (∂n W ∗ ).
Note that these quantities also appear as an additional contribution (also called
“fermionic shifts”) in the supersymmetry variation of the chiral fermions in (5.6)
and (5.7). The fact that the scalar potential is a sum of squares of these fermionic
shifts is a generic feature of supersymmetric theories that we already saw in (4.63)
and that we are going to encounter again in the following.
If one does not insist on renormalizability, e.g., in the context of low-energy
effective field theory descriptions, one could admit the following generalizations:

• An arbitrary holomorphic superpotential W (φ m ) (i.e., one that is not necessarily

a cubic polynomial)

 
1 In terms of superfields, Φ n , this corresponds to d 4 θ δmn Φ m (Φ n )† + d 2 θ W (Φ) + h.c. after
integrating out auxiliary fields.
2 In the following we will often denote the Kronecker symbol as δ
mn̄ to make manifest covariance
of the equations, even though one should interpret it as δmn .
76 5 Matter Couplings in Global Supersymmetry

• Non-minimal kinetic terms,

δmn (∂μ φ m )(∂ μ φ n ) → gmn (φ, φ ∗ ) (∂μ φ m )(∂ μ φ n ) (5.9)

• Higher derivative terms (i.e., terms with more than two spacetime derivatives)

For standard applications in particle physics with small external momenta and
small field gradients, higher derivative terms are usually irrelevant and will not
be discussed further in this book. Since we already discussed the rôle of the
superpotential (all equations above are equally valid for an arbitrary holomorphic
W (φ m )), we now turn to the consequences of non-minimal kinetic terms. We first
recall a few basic geometrical features of supersymmetric non-linear sigma models.
The reader familiar with this may jump ahead to Sect. 5.3.

5.2.2 Non-linear Sigma Models I: The Holonomy Group

Let us first consider ns real scalar fields ϕ i (x) (i = 1, . . . , ns ). The most general
Lagrangian for this field content with at most two spacetime derivatives contains a
scalar potential and a non-minimal kinetic term:

1
L = − gij (ϕ)(∂μ ϕ i )(∂ μ ϕ j ) − V (ϕ), (5.10)
2

where the coefficients gij may depend on the scalars ϕ k , as indicated. Under general
field redefinitions ϕ i → ϕ i (ϕ j ), the kinetic term transforms into the analogous
expression − 12 gij (ϕ)(∂μ ϕ i )(∂ μ ϕ j ), with

∂ϕ k ∂ϕ l
gij (ϕ) = i
gkl (ϕ(ϕ)) j (5.11)
∂ϕ ∂ϕ

so that gij (ϕ) can be interpreted as a metric on the scalar manifold (or “target
space”), Mscalar , which is the space parameterized by all possible values of the scalar
fields (ϕ 1 , . . . , ϕ ns ). Unitarity of this field theory requires gij (ϕ) to be positive
definite (i.e., Riemannian). A very simple example for a scalar manifold is given
by the Euclidean two-dimensional plane, Mscalar ∼ = R2 , which can be equivalently
parameterized by, for instance, two real Cartesian field variables (ϕ 1 , ϕ 2 ) with
metric gij = δij , or by fields R(x) and θ (x) corresponding to polar coordinates
with gRR = 1, gθθ = R 2 (x), gRθ = 0, or by a complex field z = √1 (ϕ 1 + iϕ 2 )
2
with gzz = 1, gzz = gzz = 0. Obviously, the field dependence or independence
of the metric gij (ϕ) strongly depends on the chosen field variables. For a generic,
possibly curved, scalar manifold, it is usually not possible to choose global Cartesian
field variables with gij (ϕ) = δij . The coordinate independent flatness criterion for
Mscalar is the vanishing of its Riemann tensor, Rij k l , which is defined in terms of
5.2 Chiral Multiplets in Global Supersymmetry 77

Christoffel symbols, Γijk , in the usual way,

1 kl
Γijk = g (∂i gj l + ∂j gil − ∂l gij ), (5.12)
2

Rij k l = ∂i Γjkl + Γim

k
Γjml − (i ↔ j ), (5.13)

where ∂i ≡ ∂/∂ϕ i .

5.2.2.1 The Holonomy Group of the Scalar Manifold

The Christoffel symbols, Γijk , define a covariant derivative, Di , on Mscalar , e.g.,

j
Di V j ≡ ∂i V j + Γik V k (5.14)

for a tangent vector, V j , on M  a curve γ : [0, 1] → Mscalar on the scalar

scalar . For
dγ i
manifold with tangent vectors dt (t) , one can then define the notion of parallel
transport in the usual way, i.e., by saying that V i is parallel transported along γ , if
 
dγ i
Di V j = 0 ∀t ∈ [0, 1]. (5.15)
dt

If the curve is closed, γ (0) = γ (1), the vectors V j (γ (0)) and V j (γ (1)) live in the
same tangent space and can hence be compared via the linear transformation that
connects them,

V j (γ (1)) = M j i (γ )V i (γ (0)). (5.16)

The matrix M j i (γ ) depends in general on the chosen path γ and is called the
holonomy along γ . Since the metric is covariantly constant, Di gj k = 0, the parallel
transport by the Γ connection is length preserving and hence M j i (γ ) ∈ SO(ns ).
The holonomy group, Hol(Mscalar ), of the scalar manifold Mscalar is then defined
to be the group formed by the matrices M j i (γ ) of all possible closed curves3 γ .
Obviously, Hol(Mscalar ) ⊆ SO(ns ), and a generic Riemannian manifold usually
also sweeps out the full orthogonal group. Manifolds with additional structure,
however, often have restricted holonomy groups, Hol(Mscalar ) ⊂ SO(ns ). A
simple example is given by a flat manifold, which has a trivial holonomy group
that just consists of the identity transformation. As we will see in a moment,
the scalar manifolds that appear in supersymmetric field theories also have a

3 More precisely, one defines the holonomy group, Hol(p), at a point p ∈ Mscalar to be the group
generated by all M j i (γ ) of curves that begin and end at p. This pointwise holonomy group is the
same for all points on Mscalar if (as we will always assume) Mscalar is connected, so that one can
elevate it to the holonomy group, Hol(Mscalar ), of the entire manifold.
78 5 Matter Couplings in Global Supersymmetry

restricted holonomy group. In fact, the peculiar mathematical structures of many

scalar manifolds in supersymmetry can be fully characterized by their restricted
holonomy groups. The holonomy groups themselves can often be derived by very
simple arguments that follow essentially from the supermultiplet structure. Thus, the
holonomy group of a scalar manifold is a very important and convenient concept
from which many features of supersymmetric non-linear sigma models can be
derived quite easily.

5.2.3 Non-linear Sigma Models II: Fermions and Supersymmetry

Let us consider an (oversimplified) toy model consisting of ns real scalars ϕ i (x)

and ns Majorana “fermions”, ψ i (x), (i, j, . . . = 1, . . . , ns ). We have put the word
“fermions” in quotes here because we are temporarily neglecting any complication
that may arise from chirality properties. Also, we are neglecting the fact that the
above field content contains twice as many fermions as allowed by an honest chiral
supermultiplet, because the scalars are real here. These are our temporary over-
simplifications, which we will correct soon. Assume supersymmetry transformation
rules of the schematic form

δϕ i = ψ i , (5.17)
δψ = ∂/ϕ .
i i
(5.18)

These equations motivate the interpretation of the fermions ψαi as tangent vectors
on the scalar manifold (while they are at the same time spinors in spacetime, as
indicated by the (usually suppressed) spinor index α).4
Now consider the Lagrangian

1 1 i
L = − gij (ϕ)(∂μ ϕ i )(∂ μ ϕ j ) − gij (ϕ)ψ ∂/ψ j . (5.19)
2 2
Varying the bosonic term, one obtains

δLbos = gij (δϕ i )ϕ j + δϕ k (∂μ ϕ i )(∂ μ ϕ j )gkl Γijl , (5.20)

     
A B

where the term B comes from varying the field-dependent metric, δgij = gij,k δϕ k ,
as well as from partial integrations that act on gij .

4More properly speaking, they are sections in the tangent bundle of Mscalar and sections in the
spacetime spinor bundle.
5.2 Chiral Multiplets in Global Supersymmetry 79

i j
Using δψ ∂/ψ j = −∂μ ψ γ μ δψ i and a partial integration, one obtains for the
variation of the fermionic term:

i 1 j
δLfer = −gij ψ ϕ j − gij,k (∂μ ϕ k )(∂ν ϕ i )ψ (γ μν + ημν ) +O(ψ 3 )-terms,
   2  
C
D
(5.21)

where the term D is due to partial integration acting on the field-dependent metric
gij , and the O(ψ 3 )-terms come from the variation of gij .
We see that, just as in the case with minimal kinetic terms, the two terms A and
C cancel. However, the field dependence of gij introduces the uncancelled terms B,
D, and the O(ψ 3 )-terms. In order to cancel B and D, one has to add a new term to
the Lagrangian:

1 i j
Lnew = − gij ψ (∂/ϕ l )Γlm ψ m . (5.22)
2
One can indeed verify that varying the fermions in Lnew precisely cancels B and
D, whereas the variation of the scalar field-dependent terms gives rise to additional
O(ψ 3 )-terms. The cancellation of the O(ψ 3 )-terms would require adding quartic
terms in ψ i to the Lagrangian and terms of the form ψψ to the transformation of
ψ i , which, however, we do not want to discuss here. Instead, we would like to focus
on the geometric meaning of the new term (5.22). This term can be interpreted as a
covariantization of the derivative of the fermions in the original Lagrangian (5.19),

1 i 1 i
− gij ψ ∂/ψ j → − gij ψ Dψ
/ j, (5.23)
2 2
where
j
Dμ ψ j ≡ ∂μ ψ j + (∂μ ϕ l )Γlm ψ m . (5.24)

More precisely, this derivative is a covariant derivative with respect to arbitrary

coordinate transformations on Mscalar, provided the fermions transform as tangent
vectors on Mscalar ,

∂ ϕi j ∂ ϕi
ϕ i → ϕ i (ϕ j ), ψi → ψi = ψ ⇒ Dμ ψ i → Dμ ψ i = Dμ ψ j .
∂ϕ j ∂ϕ j
(5.25)

The kinetic term (5.23) then transforms into − 12 gij ψ i D

/ ψ j with gij as in (5.11),
i
supporting our earlier interpretation of the ψ as components of a tangent vector on
Mscalar .
80 5 Matter Couplings in Global Supersymmetry

It should be stressed that the above covariantization with respect to coordinate

transformations on Mscalar has nothing to do with a covariantization with respect
to arbitrary spacetime coordinate transformations (we still work in flat Minkowski
spacetime) or Yang–Mills-type gauge transformations.

5.2.4 4D Supersymmetry and Kähler Manifolds

While the above toy model was sufficient for the purpose of understanding the
source of the covariantization of the fermionic derivatives with respect to general
coordinate transformations on Mscalar , it was still an oversimplified toy model.
When dealing with four-dimensional N = 1 supersymmetry, the scalar fields
naturally combine into complex scalars, and the number and chirality of the fermions
are important.
Indeed, the N = 1 supersymmetry transformation laws for free chiral multiplets
are (cf. Eqs. (5.4)–(5.7))

δφ m =  L χLm , δφ m =  R χRm ,
(5.26)
1 m 1 m
δχLm = ∂/φ R , δχRm = ∂/φ L ,
2 2

which show that the scalars φ m are superpartners of the left-handed components,
χLm , whereas the complex conjugate scalars, φ m , transform into the right-handed
components, χRm . As chirality is a spacetime property, the internal geometry of
Mscalar cannot interfere with it and should respect the above natural splitting of the
fields. This has profound consequences for the geometry of Mscalar . For one thing it
implies that Mscalar must be a complex manifold, i.e., it looks locally like CnC and
can be covered by mutually biholomorphic coordinate systems. Furthermore, the
holonomy group Hol(Mscalar ) should respect the natural splitting expressed in the
supersymmetry transformation laws (5.26), i.e., it should not mix φ m with φ m . Thus,
if we combine δφ m and δφ m into a 2nC -dimensional column vector, an element, M,
of the holonomy group of Mscalar has to be block diagonal:
     
δφ m δφ m Am n δφ n
→M = , (5.27)
δφ m δφ m A∗ m n δφ n

or,
 
A 0
M= , (5.28)
0 A∗

where, a priori, A ∈ GL(nC , C).

5.2 Chiral Multiplets in Global Supersymmetry 81

Switching to the 2nC -dimensional real basis (Re(δφ n ), Im(δφ m )), GL(nC , C) is
naturally embedded into GL(2nC , R) as the subgroup given by the (2nC × 2nC )-
matrices
 
 Re(A) −Im(A)
M = . (5.29)
Im(A) Re(A)

Recalling that holonomies must also be orthogonal to preserve the metric, we

have the additional requirement M  ∈ SO(2nC ) ⊂ GL(2nC , R) so that, putting
everything together,

Hol(Mscalar ) ⊂ GL(nC , C) ∩ SO(2nC ) ⊂ U (nC ), (5.30)

where the last inclusion follows because SO(2nC ) is compact and the maximal
compact subgroup of GL(nC , C) is U (nC ).5
Complex manifolds of complex dimension nC whose holonomy group is con-
tained in U (nC ) are called Kähler manifolds. We thus learn that the scalar manifold
of chiral multiplets in global N = 1 supersymmetry must be a Kähler manifold.
Kähler manifolds can also be defined in a different way by means of differential
conditions on certain globally defined invariant tensors, and this definition is very
useful to make explicit some of the properties of such manifolds. We therefore
review the steps necessary to arrive at this alternative, though equivalent, definition.
The first step is to consider a complex manifold. A complex manifold of complex
dimension nC is a differentiable manifold of real dimension 2nC that can be covered
by local complex coordinate systems, φ m (m = 1, . . . , nC ), such that the transition
functions on overlapping coordinate patches are biholomorphic. On a complex
manifold, one has a natural and well-defined tensor field, J , of type (1,1) given
by

∂ ∂
J = i dφ m ⊗ − i dφ m̄ ⊗ , (5.31)
∂φ m ∂φ m̄

where dφ m and ∂/∂φ m and their complex conjugates are the coordinate bases of the
complexified cotangent and tangent spaces. Being a (1,1) tensor field, J maps at
each point of the manifold tangent vectors to tangent vectors via J · V := ıV J
and acts on ∂m ≡ ∂/∂φ m and ∂m ≡ ∂/∂φ m as

J · ∂m = i∂m , J · ∂m = −i∂m . (5.32)

5 The intersection of GL(nC , C) and SO(nC ) inside GL(2nC , R) depends on how SO(2nC ) is
embedded inside GL(2nC , R) relative to the embedding (5.29) of GL(nC , C). The intersection is
maximal if the metric to be preserved by SO(2nC ) in the basis (Re(δφ n ), Im(δφ m )) is the unit
matrix 12nC . In that case, the orthogonality condition reads M T M  = 12nC , which is equivalent
to A† A = 1nC , which means that in that case the intersection of GL(nC , C) and SO(2nC ) sweeps
out the full U (nC ).
82 5 Matter Couplings in Global Supersymmetry

J thus squares at each point to minus the identity operator on the corresponding
tangent space,

J 2 = −id (5.33)

and is called an almost complex structure on the complex manifold.

The name “almost” here refers to the fact that the converse of the above statement
is in general not true: A differentiable manifold of real dimension 2nC with real
coordinates ϕ i (i = 1, . . . , 2nC ) and an everywhere well-defined (1,1) tensor field

∂
J = dϕ i ⊗ Ji j , (5.34)
∂ϕ j

satisfying

j
Ji k Jk j = −δi , (5.35)

does not necessarily admit compatible complex coordinate systems with biholo-
morphic transition functions and is consequently called only an almost complex
manifold. To give rise to a complex manifold, J has to satisfy an additional
differential identity, which we don’t need here, however, because we always assume
a complex manifold from the start.
If a complex manifold carries a metric, g, that is compatible with J in the sense
that, in terms of real coordinates ϕ i ,

Ji k Jj l gkl = gij (5.36)

or, equivalently,

Ji k gkj = −Jj k gik , (5.37)

one calls the complex manifold a Hermitian manifold. In terms of the complex
coordinates, φ m , φ m̄ , the components of a Hermitian metric satisfy

gmn = 0, (5.38)
gmn = 0, (5.39)
∗ ∗
gmn = gnm = (gmn ) = (gnm ) , (5.40)

as one easily verifies by inserting the complex basis vectors ∂m , ∂m into the
coordinate independent version, g(J · , J · ) = g(·, ·), of (5.36). Thus, the line
element can be written as
 
ds 2 = dϕ i ⊗ dϕ j gij = dφ m ⊗ dφ n̄ + dφ n̄ ⊗ dφ m gmn̄ . (5.41)
5.2 Chiral Multiplets in Global Supersymmetry 83

From these ingredients, we can in turn define the fundamental two-form or Kähler
form on a Hermitian manifold,

1 i 1
J = dϕ ∧ dϕ j Jij ≡ dϕ i ∧ dϕ j Ji k gkj , (5.42)
2 2
which can also be expressed in complex coordinates as

J = i dφ m ∧ dφ n̄ gmn̄ (φ p , φ q ). (5.43)

A Kähler manifold is then defined as a complex manifold with Hermitian metric

and closed fundamental form, namely,

dJ = 0, (5.44)

which means that J is also a symplectic form on Mscalar .6 We will later make use of
this symplectic structure on the scalar manifold by exploiting some analogies with
the Hamiltonian formulation of classical mechanics on phase space.
Condition (5.44) implies that locally there must exist a real function, K(φ n , φ m ),
called the Kähler potential, such that

gmn = ∂m ∂n K, (5.45)

J = i ∂m ∂n K dφ m ∧ dφ n . (5.46)

Note that the metric and the Kähler form do not change for any transformation
that maps the Kähler potential to a new one by the addition of the real part of a
holomorphic function:

K(φ n , φ n ) → K(φ n , φ n ) + h(φ n ) + h∗ (φ n ). (5.47)

These are the Kähler transformations, which are also used to match different local
expressions of the Kähler potential.
The restricted form of the metric further implies that the non-vanishing compo-
nents of the Levi–Civita connection are
l
Γmn = g lk ∂m gnk , l
Γmn = g lk ∂m gnk , (5.48)

where g lk denotes the inverse of the matrix glk . This further implies simple
expressions for the Riemann tensor,

Rnm l k = ∂n Γmk
l
, etc. (5.49)

6Kähler manifolds can in fact equivalently be defined as manifolds that are complex and real
symplectic at the same time.
84 5 Matter Couplings in Global Supersymmetry

In particular, the restricted form of the Riemann tensor (i.e., the fact that when
indices are lowered one never has pairs of indices of the same type) shows that the
Lie algebra of the holonomy group, generated by the curvature tensor, is indeed
contained in u(nC ).
From the above considerations, it follows that the geometry of the scalar manifold
of N = 1 chiral multiplets can be parameterized by a suitable Kähler potential K.
The most general globally N = 1 supersymmetric Wess–Zumino model of nC
chiral multiplets can therefore be expressed entirely in terms of the two functions
K(φ m , φ m ) and W (φ m ):
 
L DχR + χ R DχL
L = −gmn (∂μ φ m )(∂ μ φ n ) + χ m / n n / m

−(Dm ∂n W )χ m ∗ m n
L χL − (Dm ∂n W )χ R χR
n (5.50)

−g mn (∂m W )(∂n W ∗ ) + O(χ 4 ),

where

Dμ χRn ≡ ∂μ χRn + (∂μ φ m )Γml

n l
χR , (5.51)

Dμ χLn ≡ ∂μ χLn + (∂μ φ m )Γml

n l
χL , (5.52)

Dm ∂n W ≡ ∂m ∂n W − Γmn
l
∂l W, (5.53)

Dm ∂n W ∗ ≡ (Dm ∂n W )∗ . (5.54)

These new derivatives are covariant with respect to arbitrary holomorphic coordi-
nate transformations,

∂ φm n
φ m → φ m (φ n ), χLm → χLm = χ (5.55)
∂φ n L

(and analogously for the conjugate fields) on the scalar manifold, so that the entire
Lagrangian is form invariant under such field redefinitions.
The above Lagrangian is further invariant under the following supersymmetry
transformations

δφ m =  L χLm , (5.56)

δφ m =  R χRm , (5.57)

1 m 1 mn
δχLm = ∂/φ R − g (∂n W ∗ )L + O(χχ), (5.58)
2 2
1 1 mn
δχRm = ∂/φ m L − g (∂n W )R + O(χχ). (5.59)
2 2
5.3 Globally Supersymmetric Gauge Theories 85

Note that in global supersymmetry, K and W are completely independent functions

and that the metric gmn , and hence the Lagrangian L , are invariant under the
Kähler transformations introduced in (5.47). As we will see in the next chapter,
the independence of Kähler and superpotential will be lost when one couples matter
to supergravity.

5.3 Globally Supersymmetric Gauge Theories

In this section, we recall the form of globally supersymmetric theories that involve
vector multiplets (either alone or in connection with chiral multiplets).

5.3.1 Super Maxwell Theory

An on-shell N = 1 vector multiplet contains one vector field, Aμ (x), and one
Majorana fermion, λ(x), (“gaugino”). Pure super Maxwell theory describes the
coupling of nV such vector multiplets (AIμ , λI ) (I, J, . . . = 1, . . . , nV ) with
I ≡ ∂ AI − ∂ AI . The Lagrangian is
Abelian field strengths Fμν μ ν ν μ

1 1 I
L = − δI J Fμν
I
F μνJ − δI J λ ∂/λJ . (5.60)
4 2
The corresponding action is invariant under the supersymmetry transformations

1
δAIμ = − γμ λI , (5.61)
2
1
δλI = γ μν Fμν
I
, (5.62)
4
as will be verified in the exercises.

5.3.2 Super Yang-Mills Theory

The generalization to pure N = 1 super Yang-Mills theory is obtained by making

the replacements
I
Fμν → Fμν
I
≡ 2∂[μ AIν] + fJ K I AJμ AK
ν , (5.63)

∂μ λI → !
∂μ λI ≡ ∂μ λI + AJμ fJ K I λK , (5.64)
86 5 Matter Couplings in Global Supersymmetry

where fJ K I are the structure constants of the gauge algebra. Here and in the
following, we will use a hat on a derivative to indicate that it is a gauge covariant
derivative in the conventional sense. The symbol D will instead be reserved for
derivatives that are covariant with respect to scalar field redefinitions as explained
in the previous subsection. The above covariant derivative of the gaugini reflects
the fact that the gaugini have to transform in the adjoint representation of the
gauge group, just as their superpartners, the vector fields. We mention here that we
have implicitly set all gauge couplings equal to 1. For general values of the gauge
couplings, gI , the δI J in the kinetic terms has to be replaced by δI J gI−2 (no sum),
where gI can be different for each gauge group factor.

5.3.3 Coupling Super Maxwell/Yang–Mills Theories to Chiral

Multiplets

We now discuss the general couplings of nV vector multiplets AIμ , λI (I, J, . . . =

1, . . . , nV ) and nC chiral multiplets (m, n, . . . = 1, . . . nC ). The coupling of these
two sectors can occur in two ways:

1. Non-minimal kinetic terms for the vector fields and gaugini:

 1 1 I J  1 1 I J
δI J − Fμν
I
F μνJ − λ !
∂/λ → NI J (φ m , φ m ) − Fμν
I
F μνJ − λ !
∂/λ ,
4 2 4 2
(5.65)

where NI J (φ m , φ m ) denotes a scalar field-dependent kinetic matrix

2. Gauge couplings for the scalars and fermions in the chiral multiplets:

∂μ φ m → !
∂μ φ m = ∂μ φ m + AIμ (. . .),
(5.66)
!μ χ m = Dμ χ m + AI (. . .),
Dμ χ m → D μ

where !∂μ and D!μ denote suitable gauge covariant derivatives that contain
minimal couplings to the vector fields AIμ as specified by some yet to be
determined couplings (. . .)

We will now discuss these two types of couplings in turn.

5.3.4 Non-minimal Kinetic Terms for Vector Multiplets: The Gauge

Kinetic Function

A modification of the kinetic terms of the vector multiplets with non-trivial

functions NI J (φ m , φ m ) can be made consistent with supersymmetry only when
the matrix elements NI J (φ m , φ m ) are the real parts (or imaginary, depending on
5.3 Globally Supersymmetric Gauge Theories 87

conventions) of holomorphic functions, fI J (φ m ),

 
NI J φ m , φ m = Re fI J (φ m ) . (5.67)

The set of fI J (φ m ) is usually referred to as the gauge kinetic function (or gauge
kinetic matrix) .
Moreover, supersymmetry also requires a term of the form

1 1 
− i!
I
(Im fI J )  μνρσ Fμν
I
Fρσ
J
∂μ (λ γ5 γ μ λJ ) . (5.68)
4 2
The first term in (5.68) can be viewed as a generalized field-dependent θ -angle
term. Note that, for constant fI J , the entire expression (5.68) is a total spacetime
derivative that would not change the classical field equations.
A non-trivial gauge kinetic function fI J (φ m ) also requires additional fermionic
terms, e.g., terms of the form λλ, χλFμν , but it leaves the scalar potential
unchanged: V = VF , with VF = g mn (∂m W )(∂n W ), as before. Instead of giving
the full Lagrangian here, we will first also include genuine gauge interactions of the
form (5.66) and then display the most general globally supersymmetric Lagrangian
with vector and chiral multiplets and at most two spacetime derivatives in total.
In order to do this, however, we first have to understand the rôle of symmetries in
non-linear σ -models, in particular in those that are based on Kähler manifolds.

5.3.5 Non-linear σ -Models III: Global and Local Symmetries

Let us again consider a toy model consisting of ns real scalar fields ϕ i (i, j, · · · =
1, . . . , ns ) with a Lagrangian of the form

1
L = − gij (ϕ)(∂μ ϕ i )(∂ μ ϕ j ). (5.69)
2

An internal symmetry of this Lagrangian is a map I : Mscalar → Mscalar , ϕ i →

ϕ i (ϕ j ), that preserves the metric on Mscalar :

∂ϕ i ∂ϕ j
glk ≡ gij = glk , (5.70)
∂ ϕl ∂ ϕk

or, in other words, an isometry of Mscalar . The simplest example is the case
Mscalar = Rns with gij = δij . The isometries in this case consist of orthogonal
field rotations,

ϕi → M i j ϕj , M i j ∈ SO(ns ) (5.71)
88 5 Matter Couplings in Global Supersymmetry

and rigid translations (“shift symmetries”),

ϕ i (x) → ϕ i (x) + a i , a i ∈ Rns , (5.72)

which together generate the isometry group Iso (Mscalar ) ∼ = SO(ns )  Rns , with 
denoting a semidirect product.
At the infinitesimal level, isometries are generated by the corresponding Killing
vectors, ξIi (ϕ) (I = 1, . . . , dim (Iso (Mscalar )), on Mscalar ,

ϕ i → ϕ i + α I ξIi (ϕ). (5.73)

Here, α I are infinitesimal real parameters, and the ξIi satisfy the Killing equation

LξI g = 0 ⇔ gij,k ξIk + gkj ξI,i

k
+ gik ξI,j
k
= 0, (5.74)

⇔ Di ξIj + Dj ξI i = 0, (5.75)

where a comma denotes a partial derivative with respect to a scalar field, and LξI
is the Lie derivative along ξI . The Killing equation is the infinitesimal version of
(5.70) and hence ensures the invariance of the Lagrangian (5.69) under (5.73) if
α I = const.
We recall that Killing vectors are the generators of a Lie algebra

[ξI , ξJ ] = fI J K ξK , (5.76)

where [ξI , ξJ ]i = ξIi ∂i ξJ − ξJi ∂i ξI . In the above example with Mscalar ∼

j j
= Rns ,
the isometry group is the Euclidean group in ns dimensions, where the orthogonal
rotations are generated by

δϕ i = α I TI i j ϕ j , TI i j ∈ so(ns ), (5.77)
  
=ξ i (ϕ)
I

whereas infinitesimal shift symmetries are generated by constant Killing vectors

!
δϕ i = α I d!
i
, i
d! = const. (5.78)

I I

=ξ!
i
I

Here, we have split the index I = I , ! I , corresponding to the rotations labeled by

I = 1, . . . , dim(so(ns )) and the translations with !
I = 1, . . . , ns .
The above symmetries of the non-linear sigma model are global internal
symmetries in the sense that the infinitesimal symmetry parameters α I have to be
constant (i.e., spacetime independent).
5.3 Globally Supersymmetric Gauge Theories 89

If we allow for spacetime-dependent parameters, α I = α I (x), the Lagrangian

(5.69) is no longer invariant and transforms into terms proportional to ∂μ α I . In order
to make (5.69) invariant also under such local (or, “gauge”) symmetries, we have
to turn the partial derivatives of the scalar fields into gauge covariant derivatives ∂ˆμ :

∂μ ϕ i → !
∂μ ϕ i ≡ ∂μ ϕ i − AIμ ξIi (ϕ). (5.79)

This gauge covariant derivative transforms under the gauge transformations

δgaugeϕ i = α I (x)ξIi (ϕ), (5.80)

δgaugeAIμ = ∂μ α I (x) + fJ K I AJμ α K (x), (5.81)

as
   
δgauge !∂μ ϕ i = α I (x) ∂j ξIi !∂μ ϕ j , (5.82)

which then again leaves the Lagrangian (5.69) invariant due to the Killing equation
(5.74).
In the above example with Mscalar ∼
= Rns , the gauge covariant derivatives become

! ! i
∂μ ϕ i = ∂μ ϕ i − AIμ TI i j ϕ j − AIμ d!
I
. (5.83)

We emphasize again that the gauge covariantization (5.79) should not be

confused with the covariantization with respect to general field reparameterizations
ϕ i → ϕ i (ϕ j ) that we encountered earlier, e.g., for the fermionic derivative

Dμ ψ i ≡ ∂μ ψ i + (∂μ ϕ j )Γjik ψ k (5.84)

in a supersymmetric non-linear sigma model. This derivative is covariant with

respect to such general field reparameterizations (under which the fermions trans-
form as in Eq. (5.25)), but not yet with respect to the gauge transformations (5.80),
(5.81). Moreover, the connection in (5.79) corresponds to an elementary vector field,
while the one in (5.84) is a composite connection built from scalar fields and their
derivatives. In order to see that (5.84) is not gauge covariant, one recalls that the
(gauge covariantized) supersymmetry transformation law of the fermions ψ i would
be7

δsusy ψ i = !
∂/ϕ i . (5.85)

7We remind the reader that this is still a toy model where we are neglecting any chirality properties
of the fermions. The complete model will be discussed below.
90 5 Matter Couplings in Global Supersymmetry

Consistency with (5.82) then demands that the fermions also transform with the
derivative of the Killing vectors under gauge transformations,

δgaugeψ i = α I (x)(∂j ξIi )ψ j . (5.86)

Because of this, the proper gauge covariantization of (5.84) is

!μ ψ i ≡ ∂μ ψ i + (!
Dμ ψ i → D ∂μ ϕ j )Γjik ψ k − AIμ (∂j ξIi )ψ j
(5.87)
= Dμ ψ i − AIμ (Dj ξIi )ψ j ,

as will be verified in the exercises.

5.3.6 Killing Prepotentials, D-Terms, and the General Globally

Supersymmetric Lagrangian

In the above toy model, we have neglected that the scalar manifold in a proper
supersymmetric non-linear sigma model in 4D has an additional structure, namely,
it must be a Kähler manifold of complex dimension nC , where nC is the number of
chiral multiplets. This means, in particular, that the metric on the scalar manifold
derives locally from a Kähler potential, gmn = ∂m ∂n K, where K itself is defined
only modulo Kähler transformations K → K + h(φ) + h∗ (φ ∗ ).
The isometries on a Kähler manifold should therefore have the following
properties [1]:

1. Iso (Mscalar ) should respect the complex structure and not mix φ m and φ m :

δgaugeφ m = α I ξIm (φ n ), (5.88)

δgaugeφ m = α I ξIm (φ n ), (5.89)

where ξIm (φ n ) is a holomorphic Killing vector and ξIm ≡ (ξIm )∗ .

2. Iso (Mscalar ) should leave the Kähler potential invariant up to Kähler transforma-
tions:

δI K ≡ ξIm ∂m K + ξIm ∂m K = rI (φ m ) + rI∗ (φ m ), (5.90)

where the rI (φ m ) are some holomorphic functions corresponding to Kähler

transformations (5.47) with h(φ m ) = rI (φ m ) that measure the non-invariance
of the Kähler potential under gauge transformations. The condition (5.90) is
equivalent to the fact that the Killing vectors preserve the metric gmn .
5.3 Globally Supersymmetric Gauge Theories 91

Since the holomorphic Killing vectors preserve the metric on Mscalar as well as
its complex structure, they also preserve the Kähler form J :

LξI J = ıξI dJ + dıξI J = 0, (5.91)

where we expressed the Lie derivative, LξI , on a differential form in terms of the
inner product, ıξI , and the exterior derivative.
For a Kähler manifold, one has dJ = 0 so that locally there must exist functions
PI with

ıξI J = dPI . (5.92)

The functions PI are called Killing prepotentials. They are real, because the full
Killing vectors ξI = ξIm ∂m + ξIm ∂m and the Kähler form J are real. Viewing J as
a symplectic form and Mscalar as the phase space of a fictitious mechanical system,
the Killing prepotentials simply correspond to the moment maps8 whose gradients
generate the flows along the integral curves of the ξI [2]. In terms of its holomorphic
and anti-holomorphic components, (5.92) becomes

iξIm gmn = ∂n PI ,
(5.93)
−iξIn gmn = ∂m PI .

Recall now that gmn = ∂m ∂n K, and ∂n ξIm = 0. This implies that iξIm gmn =
∂n iξIm ∂m K so that

PI = iξIm ∂m K − i rI (φ)
= −iξIn ∂n K + i rI∗ (φ ∗ ) (5.94)
i  m 
= ξI ∂m K − ξIn ∂n K − rI (φ) − rI∗ (φ ∗ )
2

where the equalities are due to PI = PI∗ , and rI (φ) denote some as yet arbitrary
holomorphic functions. This expression is compatible with the request (5.90) that
the Kähler potential be invariant up to Kähler transformations as follows from

δξ K = ξIm ∂m K + ξIm ∂m K = −iPI + rI + iPI + rI∗ = rI + rI∗ . (5.95)

The integration functions rI are then completely determined up to imaginary

constants, ΔrI = −i ηI , called Fayet–Iliopoulos terms. If Eq. (5.95) were the
only constraints on the Killing prepotentials, the Fayet–Iliopoulos constants would

8In classical mechanics, if the moment maps Poisson commute with the Hamiltonian, they are
conserved along the time evolution generated by the Hamiltonian.
92 5 Matter Couplings in Global Supersymmetry

be completely undetermined for each gauge group generator. There is, however,
another identity the PI have to satisfy, the equivariance condition,

{PI , PJ } = fI J K PK , (5.96)

where on the left-hand side are the Poisson brackets of the prepotentials, defined as
the inner product of the corresponding Killing vectors with the fundamental two-
form,

{PI , PJ } ≡ −J (ξ I , ξ J ) = −ıξ J ıξ I J = ıξ I ıξ J J, (5.97)

which is also equivalent to

{PI , PJ } = −iξIm gmn ξJn + iξJm gmn ξIn = ξIm ∂m PJ + ξIn ∂n PJ

= δI PJ = −δJ PI . (5.98)

The equivariance condition is proven in the next subsection. The prepotentials enter
explicitly in the supersymmetric Lagrangian via the D-term9

D I ≡ (Ref )−1 I J PJ (“D-term”), (5.99)

which, among other places, appear in the supersymmetry transformations of the

gaugini as a shift term and in the scalar potential, where they contribute the so-
called D-term potential,

1 1
VD = (RefI J )D I D J = (Ref )−1 PI PJ . (5.100)
2 2

5.3.6.1 Proof of the Equivariance Condition

As we saw above, the equivariance condition (5.96) is rather important in fixing the
form of the prepotentials and hence of the D-terms. Its origin can be understood as a
consequence of gauge invariance, and for the sake of completeness, we sketch here
the proof, which may be omitted in a first reading.
Taking the exterior derivative of the right-hand side of (5.97) and recalling that
on scalar valued differential forms (from now on ıI ≡ ıξI ) the Lie derivative acts as
LI = ıI d + dıI , one obtains

d{PI , PJ } = dıI ıJ J = (LI ıJ − ıI dıJ ) J, (5.101)

where the last term vanishes identically because ıI J = dPI .

9Similar to the F-terms, the above D-terms are on-shell expressions for a certain auxiliary field,
often called D(x), of an off-shell vector supermultiplet.
5.3 Globally Supersymmetric Gauge Theories 93

Moreover, we know that isometries must preserve the Kähler form, i.e., LI J =
0, and therefore we can rewrite the previous expression as

d{PI , PJ } = [LI , ıJ ]J. (5.102)

This can be further manipulated using the algebraic identity [LI , ıJ ] = ı[I,J ] and
the fact that Killing vectors generate a Lie algebra, [ξI , ξJ ] = fI J K ξK . Using these
two identities, we see that

d{PI , PJ } = ı[I,J ] J = fI J K ıK J = fI J K dPK , (5.103)

and hence obtain the equivariance condition up to additive constants,

{PI , PJ } = fI J K PK + cI J , (5.104)

where the constants cI J = −cJ I should also satisfy

f[I J L cK]L = 0, (5.105)

because Poisson brackets fulfill the Jacobi identity.

The last step is to show that such constants should either disappear or be
reabsorbed in the definition of the prepotentials, in case this is possible (see
below). To understand this, we recall that the invariance of the gauge kinetic term,
RI J Fμν
I F μν J , where R
I J ≡ Re (fI J ), requires the kinetic matrix to transform
according to

!
δK RI J = LK RI J = 2fK(I L RJ )L (5.106)

in order to compensate the transformations of the field strengths, which also trans-
form with the structure constants. The inverse of the gauge kinetic matrix, R I J ≡
(RI J )−1 , enters in the definition of the D-term potential, VD ∼ PI R I J PJ ,
so that the transformations δI R J K = −2fI(K LR
J )L and δ P
I J = LI PJ =
{PI , PJ } = fI J PK + cI J imply that gauge invariance requires
K

!
cI J R J K PK = 0. (5.107)

This cannot be satisfied for arbitrary scalar field values when cI J = 0 and therefore
imposes that either the cI J have to be absorbed in the definition of the prepotentials
(for semi-simple gauge groups, see below), or they lead to inconsistent gaugings.
Let us first consider the case of semi-simple gauge groups. For semi-simple
gauge groups, the condition (5.105) implies that the constants cI J must also satisfy

cI J = fI J K cK , (5.108)
94 5 Matter Couplings in Global Supersymmetry

for some constants cK , which is related to the fact that semi-simple Lie algebras
have a trivial second cohomology group. If one now shifts PK by cK (which is
allowed because a priori PK is only defined up to additive constants by (5.92)), one
can eliminate the coefficients cI J from the Poisson bracket (5.104) and reproduce
the required equivariance condition (5.96) without cI J . Note that this is equivalent
to adding a very specific Fayet–Iliopoulos term to PK even though the gauge group
is non-Abelian. The important point here is that the Fayet–Iliopoulos constant is
fixed by the equivariance condition.
Let us now consider Abelian gaugings. For Abelian gaugings, there are two
possible cases. The first case corresponds to situations where no cI J appears in
the Poisson bracket (5.104) right from the beginning. In that case, the gauging is
consistent for arbitrary Fayet–Iliopoulos constants, because no cI J is generated
by such shifts in PK due to the vanishing structure constants. The second case
corresponds to situations where a non-trivial cI J does appear in (5.104). Due to the
vanishing structure constants for Abelian gaugings, it is not possible to absorb this
cI J by adding Fayet–Iliopoulos constants to the PK . In this case, the gauging is
thus inconsistent, because the scalar potential is not gauge invariant, no matter how
one chooses the Fayet–Iliopoulos constant.
We will now give two examples that illustrate the above points. The first example
covers the case where a non-Abelian gauging requires a specific Fayet–Iliopoulos
constant to eliminate a non-trivial cI J . The second example covers the case where
an Abelian gauging is inconsistent due to a non-vanishing cI J .

5.3.6.2 Two Examples of Gaugings with Fayet–Iliopoulos Constants

As an example of the existence of non-trivial integration constants cI J for non-
Abelian gauge groups and their absorption by the prepotential, we follow [3], where
the gauge group G = SU(2) results from the gauging of the isometries of the scalar
manifold Mscalar = CP1 with Kähler potential
 
1
K = −3ν log − (1 + φφ ∗ )− 3 ,
1
(5.109)
3

where ν is an arbitrary positive real coefficient for the time being. In details, the
isometry group SU(2) acts non-linearly on the complex scalar field φ

i 2 1 2
δ1 φ = (φ − 1), δ2 φ = (φ + 1), δ3 φ = −iφ, (5.110)
2 2
so that [ξI , ξJ ] = I J K ξK , and we can obtain the Killing prepotentials from (5.94).
Let us start by checking the transformation of the Kähler potential under such
isometries, recalling that it can vary by a Kähler transformation δI K = rI + rI∗ . We
see that
ν ν
δ1 K = i (φ − φ ∗ ), δ2 K = (φ + φ ∗ ), δ3 K = 0, (5.111)
2 2
5.3 Globally Supersymmetric Gauge Theories 95

which fixes
ν ν
r1 = i φ + i η1 , r2 = φ + iη2 , r3 = i η3 , (5.112)
2 2
where ηI are real constants. The prepotentials then follow by applying (5.94):
 "
ν φ + φ∗ ν φ − φ∗ φφ ∗
PI = + η1 , −i + η2 , ν + η3 . (5.113)
2 1 + φφ ∗ 2 1 + φφ ∗ (1 + φφ ∗ )

In these expressions we still have three undetermined real constants ηI . The

equivariance condition forces these constants to vanish with the exception of η3 ,
which has to be η3 = − ν2 , so that

{P1 , P2 } = δ1 P2 = P3 . (5.114)

We therefore see that in this example, the existence of a non-trivial constant shift
of the prepotential (usually called Fayet–Iliopoulos term) is necessary in order to
satisfy the equivariance condition, even though the gauge group is non-Abelian.
We now provide a different example [4], where the non-trivial constants cI J can
not be absorbed in the prepotentials. This example starts from the complex plane as
scalar manifold, with Kähler potential K = φφ ∗ . As gauge group, we choose the
Abelian group given by the product of two shift symmetries:

δ1 φ = λ, δ2 φ = iκ. (5.115)

The Kähler potential is not invariant under such isometries, and hence one has
non-trivial integration functions rI (φ), which moreover allow for arbitrary constant
Fayet–Iliopoulos terms:

δ1 K = λ(φ + φ ∗ ) ⇒ r1 =λφ + iη1 ,

(5.116)
δ2 K =−iκ(φ − φ ∗ ) ⇒ r2 = − iκφ + iη2 .

Once again, we can construct the respective Killing prepotentials using (5.94)

P1 = i δ1 φ ∂φ K − ir1 = −iλ(φ − φ ∗ ) + η1 , (5.117)

P2 = i δ2 φ ∂φ K − ir2 = −κ(φ + φ ∗ ) + η2 , (5.118)

and compute the equivariance condition

{P1 , P2 } = δ1 P2 = −2λκ = c12 = 0. (5.119)

96 5 Matter Couplings in Global Supersymmetry

As one can see, the constant coefficient c12 has to be there if one wants to gauge both
isometries, and, since we have vanishing structure constants, it cannot be absorbed
into the definition of the prepotentials PI , i.e., the cI J are independent of any
Fayet–Iliopoulos shifts ηI . How do we deal with such a case then? As we discussed
above, one can see that requiring full gauge invariance of the action and in particular
of the gauge kinetic term and the potential forbids such extensions, and therefore one
has to understand the previous example as an inconsistent gauge group choice for a
supersymmetric theory.

5.3.6.3 Lagrangian and Susy Rules

In order to summarize the results of this section, we now present the most general
N = 1 globally supersymmetric Lagrangian of nC chiral and nV vector multiplets
with at most two spacetime derivatives10
 
Lglobal = −gmn (!∂μ φ m )(!
∂ μφn) + χ m  ! Rn + χ nR D
L Dχ  !χLm
 1 1 I 
+(RefI J ) − Fμν
I
F μν J − λ !
∂λJ
4 2
1  
− 2i!
I
+ (ImfI J ) FμνI
FρσJ μνρσ
 ∂μ (λ γ5 γ μ λJ ) (5.120)
8
# 1 i I
+ − fI J,m Fμν I
χmL γ λL + D fI J,m χ L λ
μν J m J
4 2
1
+ (∂m W )g mn fI∗J,n λR λJR
I
4
$
I m
−(Dm ∂n W )χ m χ
L L
n
− 2ξ I
n
gmn λ χ L + h.c.

−V (φ m , φ n ) + L4F ,

where L4F denotes four Fermion terms, and

Dm ∂n W ≡ ∂m ∂n W − Γmn
l
∂l W, (5.121)
1
V (φ m , φ n ) = VF + VD = g mn (∂m W )(∂n W ∗ ) + (RefI J )D I D J ≥ 0. (5.122)
2

10 Incertain theories with non-gauge invariant gauge kinetic functions, generalized Chern–Simons
terms may also occur, which are terms of the form AI ∧ AJ ∧ dAK and AI ∧ AJ ∧ AK ∧ AL [5].
5.3 Globally Supersymmetric Gauge Theories 97

The supersymmetry transformations are (neglecting three Fermion terms):

δφ m =  L χLm ,
1 ! m 1
δχLm =  ∂φ R − g mn (∂n W ∗ )L ,
2 2
1 (5.123)
δAIμ = − γμ λI ,
2
1 μν I i
δλ = γ Fμν  + γ5 D I .
I
4 2

5.3.6.4 A Familiar Special Case: Canonical Kähler Potential and Minimal

Gauge Kinetic Function
In the special case

K = δmn φ m φ n and fI J = δI J , (5.124)

the Lagrangian (5.120) simplifies to

 
Lglobal = −δmn (!
∂μ φ m )(!
∂ μφn) + χ m ! n n ! m
L  ∂χR + χ R  ∂χL

 1 1 I 
+δI J − Fμν I
F μν J − λ ! ∂λJ (5.125)
4 2
# $
I m
− (∂m ∂n W )χ m
L χL + 2ξI δmn λ χL + h.c. − (VF + VD ),
n n

where

!
∂μ χLm ≡ ∂μ χLm − AIμ (∂n ξIm )χLn (5.126)

is the gauge covariant derivative of the chiral fermions (remembering that the
Christoffel symbols on Mscalar now vanish).
Assuming that the gauge action on the scalar fields is linear, i.e.,

δI φ m = −i TI m n φ n ,

δI φ m = +i TI m n φ n , (5.127)

where TI m n is Hermitian (i.e., TIm l δmn = δlk TIk n ), and TI m n ≡ (TI m n )∗ , one reads
off the Killing vectors

ξIm = −i TI m n φ n , ξIm = +i TI m n φ n . (5.128)

Due to the Hermiticity of the TI m n , the Kähler potential is gauge invariant.

According to (5.90), the holomorphic functions rI can then at most be an imaginary
98 5 Matter Couplings in Global Supersymmetry

constant, −iηI (the Fayet–Iliopoulos terms) provided the corresponding gauge

group factor is Abelian. The Killing prepotentials are then

PI = i[ξIm ∂m K − rI ] = [φ n TI nm φ m − ηI ], (5.129)

where TI nm ≡ TI l m δln . The gauge covariant derivative (5.126) then reduces to the
familiar expression

!
∂μ χLm = ∂μ χLm + i AIμ TI m n χLn , (5.130)

and the D-term potential becomes

1  n 2
VD = φ TI nm φ m − ηI , (5.131)
2
I

with ηI non-zero at most for Abelian factors.

5.4 Supersymmetry Breaking

Before concluding this chapter, we collect here a few remarks on supersymmetry

breaking in globally supersymmetric field theories.
This is a deep and extremely well-studied subject, and entire books could be
written only on this topic. Here we discuss only some very basic facts that will be
strikingly different in supergravity.
Global supersymmetry can be broken explicitly or spontaneously. In the first
case, one loses the restrictions, but also the control on the theory, which comes from
supersymmetry. Some control is retained, however, if the supersymmetry breaking
terms are “soft,” by which one means that they do not generate quadratic divergences
in quantum loop corrections and hence preserve some of the good UV behavior of
theories with unbroken supersymmetry.
An attractive way to generate such soft supersymmetry breaking terms is by
considering a supersymmetric Lagrangian that exhibits spontaneous supersymmetry
breaking at a certain energy scale, MS . The effective field theory below that energy
scale will then contain in general soft supersymmetry breaking terms.
Spontaneous supersymmetry breaking means that one considers the theory in a
vacuum state |0 that is not left invariant by all the supersymmetry transformations,
Q|0 = 0. The supersymmetry transformation of a field operator Φ(x) is given
by i [Q, Φ(x)] ≡ δsusyΦ(x), so that there must be at least one field Φ(x)
with δsusyΦ = 0. If we assume a maximally symmetric spacetime, only scalar
fields could possibly have a (constant) non-vanishing vacuum expectation value
(vev), whereas the vev of all fermions and tensor fields must vanish. As bosonic
fields transform into fermionic fields, we thus automatically have δsusy Φbos  =
5.4 Supersymmetry Breaking 99

Fermions = 0, and all supersymmetry breaking can only come from a non-
vanishing δsusy Φfer .
Inspection of the supersymmetry transformations of the chiral fermions and the
gaugini in (5.123) then shows that spontaneous supersymmetry breaking in global
supersymmetry could occur via a non-vanishing F-term vev F  ∼ = ∂W  = 0
and/or a non-vanishing D-term vev D = 0. Hence a useful order parameter of
supersymmetry breaking is the scalar potential V = |F |2 + D 2 ≥ 0:

V (φ m , φ m ) > 0 ⇔ susy breaking,

(5.132)
V (φ m , φ m ) = 0 ⇔ susy preserved.

It should be noted that the F - and D-term contributions are completely independent
in global supersymmetry. Supersymmetry is broken whenever at least one of the two
is non-vanishing, and a positive cosmological constant is always generated. √
The scale of supersymmetry breaking at tree level is defined by MS2 ≡ V ,
which becomes MS2 ∼ F for pure F-term SUSY breaking or MS2 ∼ D for pure
D-term SUSY breaking.
In the case of spontaneous supersymmetry breaking, we should also note that
several aspects of the resulting effective theory are dictated by the supertrace of the
squared masses

StrM 2n = (−1)2J (2J + 1)MJ2n , (5.133)
J

where MJ is the mass of a field of spin J , and we are summing over all
fields and spins, with opposite signs for bosons and fermions. For instance, for a
renormalizable model, the supertrace of the squared masses computed at tree level
in the non-supersymmetric vacuum by the action is proportional to the charge matrix
of the chiral fields, Q, and the expectation value of the D-terms [6]:

StrM 2 = 2 trQD. (5.134)

This imposes phenomenological constraints on the supersymmetry breaking

patterns and the resulting spectrum of particles. Another example is the one-loop
effective potential, which can be computed for an explicit cutoff Λ:

1 1 1 M2
V1 = V0 + StrM 0 4
Λ + StrM 2 2
Λ + StrM 4
log +...
64π 2 32π 2 64π 2 Λ2
(5.135)

where the dots stand for contributions with negative powers of Λ. For spontaneous
supersymmetry breaking, the quartic divergence is absent because of the equal
number of bosonic and fermionic degrees of freedom. The same is generally true for
the second most divergent term, which is the quadratically divergent contribution,
100 5 Matter Couplings in Global Supersymmetry

proportional to Str M 2 . In the Standard Model, Str M 2 depends on the Higgs field
and induces a quadratically divergent contribution to the Higgs squared mass, the
well-known source of the gauge hierarchy problem. In unbroken supersymmetry,
Str M 2 is identically vanishing, and, following (5.134), its vanishing persists if
we have broken supersymmetry without anomalous U(1) factors. However, without
anomalous U(1) factors, Str M 2 = 0, and therefore, at least some of the sleptons
and squarks should be lighter than some of the SM fermions (a more detailed
discussion can be found in [7]). Hence direct supersymmetry breaking within the
Standard Model sector is problematic.
The usual way out is to consider explicit supersymmetry breaking terms coming
from the relation of the observable SM sector with some hidden supersymmetry
breaking sector. The sparticle spectrum then depends on the messenger of super-
symmetry breaking and on whether the corrections arise from direct couplings
or radiative corrections. As we will discuss in Chap. 7, the non-renormalizable
interactions that arise from the coupling to supergravity provide a compelling
example for the first possibility.

Exercises

5.1. Prove that the Lagrangian (5.60) transforms under supersymmetry into

δL = J ρ ∂ρ  (= 0 in global supersymmetry)

with the supercurrent

1 I J ρ μν
J ρ ≡ − δI J Fμν λ γ γ .
4
5.2. Compute non-vanishing components of the Levi–Civita connection and of the
Riemann tensor in terms of the metric for a Kähler manifold.

5.3. Show that the derivative D !μ in (5.87) is simultaneously covariant with respect
to field reparameterizations (5.25) and with respect to gauge transformations (5.80),
(5.81), and (5.86).

5.4. Consider a model with Kähler potential

K = |φ|2 + |A|2 + |B|2

and superpotential

W = φAB.
References 101

The three scalar fields are charged under a U(1) symmetry, which is gauged and
generates a Fayet–Iliopoulos term η. Their charges are Qφ = 0, QA = 1, and
QB = −1. The kinetic term for the vector field is canonical.

(a) Compute the scalar potential.

(b) Find the supersymmetric and non-supersymmetric vacua.
(c) Compute the boson and fermion masses at the vacua.
(d) Find the range where the non-supersymmetric vacuum can be a minimum (at
least in the field directions A and B).
(e) Discuss the type of supersymmetry breaking.

References
1. J. Bagger, E. Witten, The gauge invariant supersymmetric nonlinear sigma model. Phys. Lett. B
118, 103–106 (1982)
2. H. Goldstein, Classical Mechanics, 2nd edn. (Addison-Wesley, Boston, 1980). ISBN 978-0-
201-02918-5
3. R. Kallosh, L. Kofman, A.D. Linde, A. Van Proeyen, Superconformal symmetry, supergravity
and cosmology. Class. Quant. Grav. 17, 4269–4338 (2000) [arXiv:hep-th/0006179 [hep-th]]
4. C.M. Hull, A. Karlhede, U. Lindstrom, M. Rocek, Nonlinear σ models and their gauging in and
out of superspace. Nucl. Phys. B 266, 1–44 (1986)
5. J. De Rydt, J. Rosseel, T.T. Schmidt, A. Van Proeyen, M. Zagermann, Symplectic structure of
N = 1 supergravity with anomalies and Chern-Simons terms. Class. Quant. Grav. 24, 5201–5220
(2007). 0705.4216
6. S. Ferrara, L. Girardello, F. Palumbo, A general mass formula in broken supersymmetry. Phys.
Rev. D20, 403 (1979)
7. S. Weinberg, The Quantum Theory of Fields, vol. 3. Supersymmetry (Cambridge University
Press, Cambridge, 2000), 419 p.
Matter Couplings in Supergravity
6

In this chapter, we study the coupling of 4D, N = 1 chiral and vector multiplets to
supergravity, using the geometric language developed in the previous chapter. Our
main emphasis will be on the difference between the matter couplings in global and
local supersymmetry.

6.1 New Supergravity Couplings

As we have seen in the previous section, the general N = 1 globally supersymmet-

ric theory of chiral and vector multiplets is completely specified by the following
data:

• The numbers, nC and nV , of, respectively, chiral and vector multiplets

• The Kähler potential K(φ m , φ m ) that determines the geometry of Mscalar
• The holomorphic superpotential W (φ m ) that encodes the self-interactions of the
chiral multiplets
• The holomorphic gauge kinetic function fI J (φ m ) related to the kinetic terms of
the vector multiplets
• The action of the gauge group on Mscalar , as specified by the holomorphic Killing
vectors ξIm (φ n ) and the corresponding Killing prepotentials PI (φ m , φ m )
• The real Fayet–Iliopoulos terms, ηI , which might be non-zero for Abelian gauge
group factors

When we couple such a theory to supergravity, making it locally supersymmetric,

there will be additional couplings of the matter multiplets to the supergravity multi-
plet but also new and modified couplings among the fields of the matter multiplets
themselves [1–3]. All these additional or modified couplings are still completely
specified by the data that already specified a theory in global supersymmetry
and that were mentioned above. As we will now explain, they will appear in the
Lagrangian with inverse powers of MP .

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 103

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_6
104 6 Matter Couplings in Supergravity

6.1.1 Coupling Chiral Multiplets to Supergravity

Let us start with the modifications that are necessary in order to make a globally
supersymmetric Wess–Zumino model also invariant under local supersymmetry.
The globally supersymmetric Lagrangian is (cf. (5.50))
 
LW Z = −gmn (∂μ φ m )(∂ μ φ n ) + χ m
L DχR + χ R DχL
/ n n / m

∗ m n
L χL − (Dm ∂n W )χ R χR
−(Dm ∂n W )χ m n
(6.1)

−g mn (∂m W )(∂n W ∗ ) + O(χ 4 ).

If we now allow for a spacetime-dependent supersymmetry parameter  = (x), the

derivative in the kinetic terms of the fermions χ m will produce new terms when it
acts on (x) coming from the supersymmetry transformations of the chiral fermions

1 m 1 mn
δχLm = ∂/φ R − g (∂n W ∗ )L + O(χχ), (6.2)
2 2
1 1 mn
δχRm = ∂/φ m L − g (∂n W )R + O(χχ). (6.3)
2 2
The result is an uncancelled variation of the form
μ μ
δLWZ = J R ∂μ R + J L ∂μ L , (6.4)

where the supercurrents are

μ ∗ μ μ
J L = −gmn χ m
L γ ∂ φ + χ R γ ∂n W ,
μ/ n n μ
JR = (JL )c . (6.5)

As we have already shown in Chap. 2 for the special case of a free Wess–Zumino
model, the cancellation of these terms is achieved by adding the Noether couplings
to the gravitino

1  μ μ

LNoether = − J R ψμR + J L ψμL . (6.6)
MP

Using δψμ = MP ∂μ , one then finds that everything cancels modulo terms that
come from the variation of the supercurrents themselves:

1  μ

δ(LW Z + LNoether) = − (δJ R )ψμR + h.c. (6.7)
MP
6.1 New Supergravity Couplings 105

These terms are of the form

δ(LW Z + LNoether) = −δg μν Tμν + Z1 + Z2 . (6.8)

Just as discussed in Chap. 2, Tμν is the energy momentum tensor of LW Z , and the
new field gμν is identified with the spacetime metric, signaling the necessity for
a coupling to gravity. The minimal coupling to a dynamical metric is achieved by
covariantizing everything with respect to general spacetime coordinate and local
Lorentz transformations and by adding the pure supergravity Lagrangian. The
metric variation of this covariantized Lagrangian then precisely cancels the first
term in (6.8), and the theory would be supersymmetric if there weren’t also the two
additional terms Z1 and Z2 in Eq. (6.8) that we have neglected so far. As we will
now show, these two terms are actually quite important, as they lead to additional
MP−2 -suppressed interactions between the fields of the chiral multiplets themselves
that have some far-reaching consequences.
In order to make this more precise, let us first state what Z1 and Z2 are:
e
Z1 = − gmn ψ μ γ μνρ γ5  (∂ν φ m )(∂ρ φ n ), (6.9)
2MP
e  
Z2 = ψ μL γ μν L (∂ν W ∗ ) + ψ μR γ μν R (∂ν W ) . (6.10)
MP
μ
The first term Z1 comes from the variations of the form δχLm ∼ 12 ∂/φ m R in JL and
its conjugate, which give rise to terms with three antisymmetrized gamma matrices
as well as terms with one gamma matrix. The latter are part of the energy momentum
tensor terms in (6.8) (because δgμν involves only one gamma matrix), whereas the
terms with three antisymmetrized gamma matrices are precisely given by Z1 . The
first term of Z2 is due to the variations δχLm ∼ − 12 g mn (∂n W ∗ )L in the first term
in (6.5) and due to the variation δχRm ∼ 12 ∂/φ m L in the second term in (6.5). The
μ
second term in Z2 arises from the analogous variations of JR .
We will now see that the cancellation of Z1 and Z2 requires the introduction of
new terms with important consequences.

6.1.2 The Kähler Covariant Derivative

In order to cancel Z1 , we first rewrite it by using the relation between the metric of
the scalar manifold and the Kähler potential:
 
Z1 ∼ψ μ γ μνρ γ5  ∂ν φ m ∂ρ φ n ∂m ∂n̄ K
(6.11)
1  
=ψ μ γ μνρ
γ5  ∂ρ φ n̄ ∂ν ∂n̄ K − ∂ρ φ m ∂ν ∂m K .
2
106 6 Matter Couplings in Supergravity

Using the last expression, we then integrate by parts the spacetime derivative that
acts on the Kähler potential. This produces in particular terms where the derivative
acts on  and terms where it acts on ψ μ . The former term is

e 1 
ψ μ γ μνρ γ5 (Dν ) ∂ρ φ n̄ ∂n̄ K − ∂ρ φ m ∂m K . (6.12)
2MP 2

We now repeat our old trick and simply add the negative of this term (times a factor
1/2) to the Lagrangian, but with Dν  replaced by ψν ,
 
e i
LKähler cov = − ψ μ γ μνρ Qν (φ)γ5 ψρ , (6.13)
2 2MP2

where Qν is a composite vector field,

i  
Qν (φ) ≡ (∂n K)∂ν φ n − (∂m K)∂ν φ m . (6.14)
2
Varying the two gravitini in this expression would then precisely cancel (6.12).
The cancellation of the remaining term in Z1 , where the derivative acts on ψ μ ,
will be discussed later (see footnote 1 in this chapter).
The new interaction term LKähler cov , however, now poses another problem:
as one easily verifies, it is not invariant under Kähler transformations K → K +
h + h∗ . LKähler cov would thus seem to single out a particular Kähler potential,
even though a specific Kähler potential is not an intrinsic geometrical object on a
Kähler manifold. In general, the Kähler potential is in fact only locally defined and
requires Kähler transformations on the overlaps of local coordinate patches. So if the
Lagrangian was not Kähler invariant, the physics would in general also be different
for different coordinate patches of the scalar manifold.
To understand the resolution of this problem, we observe that the term
LKähler cov can be absorbed into the Rarita–Schwinger action by modifying
the covariant derivative with a new term,
e
LRS + LKähler cov = − ψ μ γ μνρ Dν (ω, Q)ψρ , (6.15)
2
where
i
D[ν (ω, Q)ψρ] ≡ D[ν (ω)ψρ] + Q[ν γ5 ψρ] , (6.16)
2MP2

with Dν being the Lorentz covariant derivative. To understand the significance of

this modification, one notes that Qμ transforms under Kähler transformations like a
U(1) connection:

Qμ → Qμ + ∂μ Im(h). (6.17)
6.1 New Supergravity Couplings 107

More precisely, Qμ is a composite U(1) connection, i.e., it is not an elementary

vector field, but rather a function of the scalar fields and their derivatives.
We now see that we can render the Lagrangian invariant if we require that Kähler
transformations, K → K + h + h∗ , be accompanied by chiral rotations of the
gravitino:
% &
i
ψμ → exp − Im(h(φ))γ5 ψμ . (6.18)
2MP2

Indeed, the derivative (6.16) then transforms covariantly,

% &
i
D[μ ψρ] → exp − Im(h(φ))γ5 D[μ ψρ] , (6.19)
2MP2

and the combination (6.15) is Kähler (and obviously also locally Lorentz) invariant.
These geometric arguments thus suggest that, in supergravity, Kähler transforma-
tions on the scalar manifold also act on the gravitino as a chiral U(1) symmetry,
with Qμ being the corresponding (composite) U(1) connection. If this is to make
sense, this non-trivial action of Kähler transformations on the gravitini should also
be compatible with supersymmetry. As we will now show, this requirement will lead
to further interesting differences with respect to global supersymmetry and provides
further consistency checks.
First we note that if the gravitino transforms under Kähler transformations,
the consistency with the supersymmetry transformation law δψμ ∼ MP Dμ  also
requires that  transforms under Kähler transformations,
% &
i
 → exp − Im(h(φ))γ5  (6.20)
2MP2

and that its derivative (as it appears in δψμ ) should also be covariantized,1

i
Dμ (ω, Q) ≡ Dμ (ω) + Qμ γ5 . (6.21)
2MP2

This in turn implies, because of δχLm = 12 ∂/φ m R + . . ., that also the chiral fermions
transform under Kähler transformations,
% &
i
χ → exp +
m
Im(h(φ))γ5 χ m , (6.22)
2MP2

1 This is indeed confirmed by computing the gravitino variations of LRS with the new Kähler
covariant transformation law, δψμ ∼ MP Dμ , which leads to a new term that precisely cancels
the remaining uncancelled part of Z1 (i.e., the part of Z1 with a derivative acting on ψ μ ).
108 6 Matter Couplings in Supergravity

and that their derivatives have to be Lorentz, Mscalar reparameterization, and Kähler
covariant, e.g.,2

i
Dμ χLm ≡ Dμ χLm + (∂μ φ n )Γnlm χLl − Qμ χLm . (6.23)
2MP2

Note that there is a different sign in (6.22) (and hence also in (6.23)) compared
to the corresponding terms of the gravitino or the supersymmetry transformation
parameter (cf. (6.18) and (6.20) as well as (6.16) and (6.21)). This sign difference
arises because one has to move the γ5 matrix in (6.20) through one gamma matrix
in the supersymmetry transformation δχLm = 12 ∂/φ m R + . . ..
Although we will discuss gauge multiplets later, we already mention here that
δλI ∼ 14 γ μν Fμν
I  + . . . implies that also the gaugini transform non-trivially under

Kähler transformations (with the same sign as ψμ and )

% &
i
δλ → exp −
I
Im(h(φ))γ5 λI (6.24)
2MP2

and that likewise all their derivatives have to be properly covariantized with respect
to Kähler transformations (again with the same sign as for ψμ and ).
To conclude, all fermion fields and not just the gravitino are charged with respect
to a composite chiral U(1) symmetry that is related to Kähler transformations
and that is not present in the global case. It should be emphasized that in the
limit of global supersymmetry, MP → ∞, these chiral rotations become trivial,
as is signaled by the inverse powers of MP . This is consistent with the rigid
supersymmetry results of the previous chapter, where this chiral composite U(1)
is not encountered.
Interestingly, the above non-trivial transformations of the fermions under Kähler
transformations also imply that the superpotential and its derivatives have to
transform as we will show in Sect. 6.1.3. The result is that
% &
1
W → exp − 2 h(φ m ) W (φ m ) (6.25)
MP

and its derivatives have to be Kähler covariantized as follows

% &
K K
2 2 (∂n K)
∂n W → e 2MP
Dn W ≡ e 2MP
∂n + W. (6.26)
MP2

2For the sake of simplicity, we do not introduce a new symbol for the Kähler covariantized
derivative and still call it Dμ .
6.1 New Supergravity Couplings 109

To summarize: The cancellation of Z1 by adding LKähler cov gives rise to the

interpretation that the fermions and the superpotential should transform non-trivially
under Kähler transformations. In order to ensure this, all derivatives of the fermions
and the superpotential have to be Kähler covariantized, and the superpotential terms
have to be dressed with an exponential of the Kähler potential. One can show that all
these modifications are indeed also necessary for the cancellation of various other
variations we have not discussed here in detail. In general, we define the Kähler
covariant derivatives in field space as
 
p
Dm Φ = ∂m + 2 ∂m K Φ,
MP
  (6.27)
p
Dm Φ = ∂ m − ∂ m K Φ,
MP2

where p is the Kähler “charge” of the field Φ.

6.1.3 Additional Bare Superpotential Terms

In global supersymmetry, all superpotential terms always appear with at least one
derivative with respect to the scalar fields (see, e.g., the Lagrangian (5.120)). As
we saw in the previous subsection, the coupling to supergravity (in particular the
cancellation of the term Z1 ) requires a Kähler covariantization of these derivatives of
W , which then introduces “bare” W -terms inside these Kähler covariant derivatives,
i.e., W -terms that are not differentiated with respect to any scalar field. In this
subsection, we show that there are additional “bare” superpotential terms in the
Lagrangian and the supersymmetry transformation laws. Their necessity follows
from the cancellation of the term Z2 to which we now turn.
In order to cancel the Z2 -term
e  
Z2 ≡ ψ μL γ μν L (∂ν W ∗ ) + ψ μR γ μν R (∂ν W ) (6.28)
MP

we proceed as we did for Z1 and first perform an integration by parts. This will
then give again terms with a derivative acting on the supersymmetry parameter 
and terms where the derivative acts on the gravitini ψ μ . To cancel the former, we
then again add to the Lagrangian a term where the derivatives of  are replaced by
gravitini, or, more precisely,
e  ∗ 
2
W ψ μL γ μν ψνL + W ψ μR γ μν ψνR . (6.29)
2MP
110 6 Matter Couplings in Supergravity

This term is an obvious mass-like term for the gravitino, and therefore, following
the rules we have learned in the case of pure supergravity in the presence of a
cosmological constant, we have to further modify the variation of the gravitino field
by adding a new term of the form

1
δnew ψμL ∼ W γμ R . (6.30)
2MP

This new variation applied to the Rarita–Schwinger action also gives the term
required to cancel the second piece coming from the partial integration of Z2 ,
namely, the term with the derivative acting on the gravitino.
Before proceeding further, let us come back to the Kähler covariantization of the
superpotential terms and prove (6.25). Subjecting (6.30) to Kähler transformations
tells us that the left-hand side transforms as
% &
1
exp − (h(φ) − h∗ (φ)) , (6.31)
4MP2

while the epsilon parameter on the right-hand side transforms with the opposite sign
due to the opposite chirality:
% &
1 ∗
exp + (h(φ) − h (φ)) . (6.32)
4MP2

At this point it is obvious that in order for the Kähler transformation to be compatible
with supersymmetry, we need to transform also the superpotential, as we already
mentioned earlier. The superpotential, on the other hand, is a holomorphic function
by construction and hence can transform only with a holomorphic factor,
% &
α
W → exp − 2 (h(φ)) W, (6.33)
MP

where α is a real constant. In order to get the same rotation on the left- and on
the right-hand side of (6.30), we still need something that transforms under Kähler
transformations with the exponential of h + h∗ , like the exponential of the Kähler
2
potential itself, eβK/MP . The right coefficients follow then by equating the two sides:

1 1 α
− 2
(h(φ) − h∗ (φ)) = + 2
(h(φ) − h∗ (φ)) − 2 (h(φ))
4MP 4MP MP
β
+ (h(φ) + h∗ (φ)). (6.34)
MP2
6.1 New Supergravity Couplings 111

This fixes α = 1 and β = 1/2 and tells us that we have to replace the superpotential
with the combination
2
eK/(2MP ) W (6.35)

and that indeed W transforms under Kähler transformations as in (6.25).

Coming back to the check of supersymmetry invariance, we now see that the
new transformation law for the gravitino (6.30) applied to the new bilinear term
(6.29) gives a new variation of the form |W |2 ψγ . Not too surprisingly, this can
then finally be cancelled by adding a new contribution ∼ −e|W |2 to the scalar
potential and varying the vierbein determinant e. This is the generalization of the
procedure derived in Sect. 4.2 for the case of a constant superpotential, i.e., for pure
supergravity with a cosmological constant.
Although it may be hard to believe, it turns out that, after proper Kähler covari-
antizations, the above modifications are sufficient to ensure also the cancellations of
all the other variations we have not considered explicitly here.
The end result is the Lagrangian

MP2 1
e−1 L = R(e, ω(e)) − ψ μ γ μνρ Dν (ω(e), Q)ψρ
2 2
 
L DχR + χ R DχL
−gmn (∂μ φ m )(∂ μ φ n ) + χ m / n n / m

# 2
$
− eK/2MP (Dm Dn W )χ m χ
L L
n
+ h.c. (6.36)

1 # 2
$
+ gmn ψ μL γ ν γ μ χLm (∂ν φ n ) + ψ μR γ μ χLm eK/2MP Dm W + h.c.
MP

1 # K/2M 2 $
+ e P Wψ
μR γ μν
ψνR + h.c − V (φ m , φ n ),
2MP2

with the scalar potential given by the sum of two contributions

% &
K/MP2 ∗ 3|W |2
V =e g mn
(Dm W )(Dn W ) − , (6.37)
MP2

where the first term is the Kähler covariantization of the F-terms from global super-
symmetry and the second is a genuine contribution from gravitational couplings, in
the sense that it is a variation of the vierbein determinant that leads to a cancellation
of the |W |2 terms mentioned after (6.35). In the Lagrangian (6.36), the first line is
the Kähler covariantization of the pure supergravity action. The second and third line
112 6 Matter Couplings in Supergravity

correspond to the Kähler and spacetime covariant Wess–Zumino action (without the
potential). Note that now
  %  &  
∂m K ∂n K ∂p K
Dm Dn W = ∂m + ∂n + 2 W − Γmn p
∂p + W.
MP2 MP MP2
(6.38)

The fourth line is the Kähler and spacetime covariant

 Noether coupling
 of the
μ μ
supercurrents to the gravitino, LNoether = − MP J R ψμR + J L ψμL (with the
1

fermions moved into a different order). The fifth line, finally, contains the W -
dependent extra terms as well as the (Kähler covariantized) scalar potential of the
Wess–Zumino model.
The supersymmetry transformation rules, up to three fermion terms, are

1
δeμa = γ a ψμ ,
2MP
1 K/2M 2
δψμL = MP Dμ (ω(e), Qν )L + e P Wγ  ,
μ R
2MP (6.39)
δφ m =  L χLm ,
1 m 1
∂/φ R − g mn eK/2MP (Dn W ∗ )L .
2
δχLm =
2 2
Obviously, in the MP → ∞ limit, these equations reduce to the globally supersym-
metric theory discussed earlier in Eqs. (5.50) and (5.56)–(5.59). One also notices
that truncating out the chiral multiplets and keeping a constant superpotential
K
2
e 2MP W = −g MP3 gives back the pure supergravity Lagrangian with cosmological
constant, Eq. (4.60).
Note further that in supergravity the Kähler potential and the superpotential
are no longer independent, as one can shift terms back and forth via Kähler
transformations. In fact, as long as W is not equal to zero, one can even make the
superpotential equal to MP3 by performing a Kähler transformation with h(φ) =
MP2 log(W/MP3 ) (cf. Eq. (6.25)). More generally, instead of using the two functions
K and W , one can express the entire Lagrangian in terms of the function

|W |2
G = K + MP2 log , (6.40)
MP6

which is manifestly Kähler invariant. For instance, the part of the scalar potential
coming from the superpotential becomes
 
V = eG /MP MP2 g mn̄ Gm Gn̄ − 3MP4 .
2
(6.41)
6.1 New Supergravity Couplings 113

Note, however, that by doing so, one cannot recover the W = 0 case, which has to
be discussed separately. Hence the usefulness of leaving explicit both K and W in
our approach.

6.1.4 Inclusion of Vector Multiplets

The inclusion of vector multiplets requires the following changes:

1. All terms that were already present in global supersymmetry (cf. Eq. (5.120)) are
also present in supergravity, but they all have to be made spacetime and Kähler
covariant.
2. A new Noether coupling of the vector multiplet supercurrent
 
μ J 1 i
J V M ≡ eλ − (RefI J )Fνρ γ γ − PJ γ μ γ5
I μ νρ
(6.42)
4 2

to the gravitino has to be introduced:

 1 μ
LNoether =− J ψμ (6.43)
MP V M

in order to cancel terms of the form

μ
δL = J V M ∂μ  (6.44)

that arise due to the derivative in − 12 e (RefI J )λ !

I
∂/λJ when it acts on the  in δλ.
3. The composite Kähler connection Qμ receives an additional contribution pro-
portional to AIμ PI for each of the gauged isometries:

i  
Qμ = Qμ (φ m , φ n , AIμ ) = (∂n K)∂μ φ n − (∂m K)∂μ φ m +AIμ PI (6.45)
2
i  
= (∂n K)!
∂μ φ n − (∂m K)!
∂μ φ m
2
+AIμ Im(rI ), (6.46)

where the last equality follows from the form (5.94) of the prepotentials (we will
see more on this in Sect. 6.1.5). This additional term is needed, e.g., in order to
cancel a variation proportional to Fμν
J P γ μνρ γ ψ that occurs in the variation
J 5 ρ
μ
− MP δ(JV M )ψμ and is not of the form −δg μν Tμν .
1

It should be noted that this additional contribution to Qμ has another

important consequence. Namely, if one shifts the Killing prepotential PI of an
Abelian factor by a Fayet–Iliopoulos constant, PI → PI + ηI , one introduces
114 6 Matter Couplings in Supergravity

new chiral gauge interactions for all fermions, including, e.g., the gravitino,

i
Dμ ψν → Dμ ψν + AIμ ηI γ5 ψν , (6.47)
2MP2

which can easily lead to quantum anomalies [4]. Thus, the introduction of Fayet–
Iliopoulos constants in N = 1 supergravity requires some care. We will actually
see later on that in supergravity the Fayet–Iliopoulos terms are related to the
non-invariance of the superpotential under gauge transformations.

Ignoring four fermion terms, the end result of all these modifications is the
following general matter-coupled Lagrangian3

MP2 1
e−1 L = R(e, ω(e)) − ψ μ γ μνρ Dν (ω(e), Q)ψρ
2 2
 
−gmn (! ∂μ φ m )(!
∂ μφn) + χ m Dχ
 ! n + χ n Dχ ! m
L R R L
 
1 I 1 I ! J
+(RefI J ) − Fμν F μν J − λ D  λ
4 2
1  
+ (ImfI J ) Fμν I
Fρσ
J μνρσ
 !μ (eλI γ5 γ μ λJ )
− 2i D
8

1 i I
+ − fI J,m Fμν I
χmL γ λL + D fI J,m χ L λ
μν J m J
4 2
1
+ eK/2MP (Dm W )g mn fI∗J,n λR λJR
2 I
4
"
2 I m
−eK/2MP (Dm Dn W )χ m χ
L L
n
− 2ξ n
g
I mn λ χ L + h.c.

1
+ (RefI J )ψ μ γ νρ γ μ λJ Fνρ I
4MP
 "
1
+ gmn ψ μL γ ν γ μ χLm (!
∂ν φ n ) + h.c.
MP
   "
1 μ i I m K/2MP2
+ ψ μR γ λ PI + χL e Dm W + h.c.
MP 2 L
1 # K/2M 2 $
+ e P Wψ
μR γ μν
ψ νR + h.c − V (φ m , φ n ), (6.48)
2MP2

3As mentioned earlier, when the gauge kinetic function is not gauge invariant, generalized Chern–
Simons terms of the form AI ∧ AJ ∧ dAK and AI ∧ AJ ∧ AK ∧ AL may be possible. Their form,
however, is the same as in global supersymmetry [5].
6.1 New Supergravity Couplings 115

with the scalar potential

% &
K/MP2 3|W |2 1
V =e g mn
(Dm W )(Dn W ) − + (RefI J )D I D J . (6.49)
MP2 2

The supersymmetry transformation rules, up to three fermion terms, are

1
δeμa = γ a ψμ ,
2MP
1 K/2M 2
δψμL = MP Dμ (ω(e), Q)L + e P Wγ  ,
μ R
2MP
δφ m =  L χLm ,
1 ! m 1
 ∂φ R − g mn eK/(2MP ) (Dn W ∗ )L ,
2
δχLm =
2 2
1
δAIμ = − γμ λI ,
2
1 i
δλI = γ μν FμνI
 + γ5 D I . (6.50)
4 2
It is again easy to see that, in the global limit, MP → ∞, the above equations reduce
to the globally supersymmetric theory discussed in Eqs. (5.120) and (5.123).
For completeness, we display the full (i.e., local Lorentz, scalar coordinate,
Kähler, and gauge covariant) derivative of λI and χ m :

!μ χ m = Dμ χ m + (! i
D L L ∂μ φ n )Γnlm χLl − AIμ (∂n ξIm )χLn − Qμ χLm ,
2MP2

!μ λI = Dμ λI + AJ f I λK + i
D μ JK Qμ γ5 λI , (6.51)
2MP2

where, as usual, Dμ denotes the Lorentz covariant derivative. The full covariant
derivatives of ψμ and  are just as for λI , except for the gauge covariantization term
AJμ fJIK λK , which is absent for these fermions (hence, we can omit the hat on their
derivatives).
More details on the action and the four Fermi terms can be found in [5, 6].

6.1.5 More on D-Terms

Although the D-terms and the D-term potential take the same form as in global
supersymmetry, local supersymmetry does have some interesting implications also
for the D-terms. To understand this, we recall that the general matter-coupled
supergravity Lagrangian is invariant under Kähler transformations that act at the
116 6 Matter Couplings in Supergravity

same time on the Kähler potential, the superpotential, and the fermions. As in
global supersymmetry, a gauge transformation therefore does not necessarily have
to leave the Kähler potential invariant but may in general transform it with a Kähler
transformation,

δgauge K ≡ ξIm ∂m K + ξIm ∂m K = rI + rI∗ . (6.52)

However, in supergravity theories this requires a non-trivial action also on the

superpotential (if W = 0)
rI
δgaugeW ≡ ξIm ∂m W = − W, (6.53)
MP2

so that the combination G in (6.40) remains invariant:

δgauge G = ξIm ∂m G + ξIm̄ ∂m̄ G = 0 (6.54)

For all the points in field space where W = 0, we can then also rewrite rI as

∂m W
rI = −MP2 ξIm , (6.55)
W
so that the Killing prepotentials can be also expressed in terms of the gauge-invariant
quantity G :
% &
M 2 ∂m W MP2 Dm W
PI = iξIm ∂m K − irI = iξIm ∂m K + P = iξIm = iξIm ∂m G .
W W
(6.56)

The D-terms are thus

D I = i (Re f )−1I J ξJm ∂m G , (6.57)

and the total scalar potential with F-terms and D-terms can be written in a very
compact and suggestive form
 
V = eG /MP hmn Gm Gn − 3 MP2 ,
2
(6.58)

where

e−G /MP
2

hmn ≡ g mn + (Re f )−1I J ξIm ξJn , (6.59)

2
so that the new metric contains the Kähler metric giving the F-term potential (6.41)
and the additional term coming from the D-term potential.
6.1 New Supergravity Couplings 117

We can now comment on some of the differences between global and local
supersymmetry, which may lead to relevant physical differences.
First of all, we see that in supergravity, D-terms and F-terms are not independent
of one another; rather the D-terms are (for W = 0) a particular combination of the
F-terms (6.57).
The next important difference concerns the Fayet–Iliopoulos constants. Just as in
global supersymmetry, the gauge transformation (6.52) of the Kähler potential fixes
rI only up to an additive imaginary constant iηI and hence PI up to an additive
real constant ηI . The equivariance condition (5.96), together with gauge invariance
of the other terms in the action, again restricts the possible values for these constants
except for U(1) factors. The difference to global supersymmetry now is that the
superpotential W also transforms under gauge transformations as in (6.53) so that a
shift of rI by an additive constant iηI implies that W transforms with an additional
phase factor under the corresponding U(1) transformation. In other words, changing
ηI changes the U(1) charge of W . Note that such U(1) transformations due to FI
constants may even occur when the Kähler potential is invariant under this U(1)
factor, because, according to (6.52), this only implies rI (φ) = i ηI .
Another important effect of a FI constant is that it leads to a chiral U(1)
transformation of the fermions, as follows from their non-trivial transformations
under Kähler transformations described in Sect. 6.1.2. As explained around (6.47),
this may then easily lead to anomalous gauge couplings and requires some care.

6.1.6 The Gradient Flow Relations

In supergravity theories, supersymmetry imposes some interesting relations between

various quantities. Particularly interesting are the relations between the potential and
the supersymmetry transformations of the fermions as well as the relations between
the supersymmetry transformations of the fermions themselves, known as gradient
flow relations [7].
When the metric and the scalar fields are the only non-trivial fields, the
supersymmetry rules can be written as

δψμL = MP Dμ L + γμ S R , (6.60)

1 ! m
δχLm =  ∂φ R + N m
L , (6.61)
2
1 μν I
δλIL = γ Fμν L + N I L . (6.62)
4
Here, S is the fermionic shift in the gravitino transformation,
K
1 2
S≡ e 2MP W, (6.63)
2MP
118 6 Matter Couplings in Supergravity

whereas the shift of the chiralini is the Kähler covariantized F-term,

K
1
≡ − g mn e 2MP Dn W ∗ , ≡ (N m )∗ ,
2
N m
N m
(6.64)
2
and the shift of the gaugini is given by the D-terms,

i I
NI ≡ D . (6.65)
2
These fermionic shifts also enter various bilinear terms in the fermions in the
Lagrangian as well as the scalar potential, which is simply given by the squares
of the shifts in (6.60)–(6.62), with appropriate signs and prefactors:

V = −12SS ∗ + 4gmn N m
N n
+ 2Re(fI J )N I N ∗J . (6.66)

Writing the potential in this way makes a number of interesting properties of

the scalar potential obvious, as we will explain in our discussion on spontaneous
supersymmetry breaking in Sect. 7.1.2. Here, we would like to list instead a number
of useful differential relations between the shifts. These are often called gradient
flow relations because for supersymmetric configurations, the behavior of the scalar
fields is fixed in terms of gradients of the gravitino shift and the other fermion shifts
and fermion masses are further determined by additional gradients in field space.
Let us start by noting that the gravitino shift S is covariantly holomorphic4

Dm S = 0. (6.67)

The shift of matter fields is on the other hand equal to the Kähler covariant derivative
of the shift of the gravitino with respect to the scalar fields,

2
Dm S = − gmn N n . (6.68)
MP

Furthermore, full covariant derivatives of the shift of the chiralini give the mass
matrices for the gravitini

S
Dm N n
= −δm
n
, (6.69)
MP

or for the chiralini

1
Dm N n gln = Mml , (6.70)
2

4 Recall the definition (6.27) for the covariant derivatives, which here are taken in fields space. In
the current context p(S) = −p(S ∗ ) = 12 .
6.1 New Supergravity Couplings 119

where
2
Mmn = eK/2MP (Dm Dn W ). (6.71)

Finally, the full covariant derivative of the shift of the gaugini is proportional to the
gauged isometries:
  1
Dm N I ∗ RefI J = gmn ξJn . (6.72)
2
The origin of these relations is obviously related to the request of invariance of
the action under supersymmetry variations. Indeed all these relations are needed to
compensate for the supersymmetry variation of the scalar fields in the potential and
of the mass terms. For instance, schematically, we can see that

∂V
∼ SNm + Mmn N n
+ gmn̄ ξIn̄ N I , (6.73)
∂φ m

and therefore the gradient flow relations mentioned above are necessary to
close supersymmetry, because such a relation is needed for the cancellation of
δLfermion mass against δV .

6.1.7 Final Remarks

The discussion on the matter couplings we outlined in this chapter clearly depends
heavily on what kind of matter we allow in these couplings. More general
Lagrangians could be obtained by introducing additional matter multiplets, such as
tensor multiplets. The tensor fields in such tensor multiplets, however, can in general
be dualized to either massless scalar or massive vector fields so that the theory will
be eventually of the standard form we described above. On the other hand, these
dualities are often non-perturbative or may require complicated field redefinitions.
It may therefore be interesting to study these models directly with tensor fields, also
because such tensor fields naturally arise from string theory compactifications We
are not going to do so in the following, though we would like to show here explicitly
the duality relations.
The best way to exhibit the duality relations in four dimensions between tensor
and scalar fields (when massless) or between tensor and vector fields (when massive)
is by means of Lagrange multipliers. Consider the action
 
1
S= V1 ∧ V1 − B2 ∧ dV1 , (6.74)
2
120 6 Matter Couplings in Supergravity

where V1 is a generic one-form and B2 is the two-form built from an antisymmetric

tensor field Bμν . From the variation of the action, we get two equations of motion

δS
= 0, ⇔ dV1 = 0, (6.75)
δB2

δS
= 0, ⇔ V1 = dB2 ≡ H3 . (6.76)
δV1

Recall that in four dimensions 2 Fp = −(−1)p(4−p)Fp . If one solves the first

equation, then V1 = dφ locally, and one can use the second equation to replace
the two-form B in the action to go back to the action for a free scalar field
 
1 1
S= dφ ∧ dφ = − d 4 x e ∂μ φ∂ μ φ. (6.77)
2 2

On the other hand, the solution to the second equation (6.76) gives a replacement
rule for the one-form, V1 , in terms of B2 and leads to the standard kinetic term
for B2 :

1
S= H3 ∧ H3 . (6.78)
2

Thus, we see that in four spacetime dimensions a massless scalar field is equivalent
to a massless antisymmetric tensor field. This implies that in a supersymmetric
theory one can replace one of the scalar fields in a massless chiral multiplet by
a two-form field, producing an off-shell linear representation of supersymmetry
called tensor multiplet, which contains a scalar field, a fermion, and a tensor field.
In theories with extended or with on-shell representations of supersymmetry, the
number of scalar fields that can be dualized simultaneously may vary, and one
can then obtain tensor multiplets, double-tensor multiplets, and also vector-tensor
multiplets. Although we do not have a direct interest in these multiplets here, it
is important to mention that tensor fields do play an important role in extended
supergravities when the gauging procedure is performed, as we explain in Chap. 9.
For the massive case, we can also produce a duality relation between a two-form
tensor and a vector field. Once again, we can introduce the duality relation by means
of Lagrange multipliers:
  
1
S=− H3 ∧ H3 − 2H3 ∧ dB2 + m2 B2 ∧ B2 . (6.79)
2
6.1 New Supergravity Couplings 121

The equations of motion for H3 and B2 are respectively

δS
= 0, ⇔ H3 − dB2 = 0, (6.80)
δH3

δS
= 0, ⇔ d(H3 ) + m2  B2 = 0. (6.81)
δB2

The first one can be solved in the usual way giving back the Lagrangian for a massive
tensor

1
S=− [dB2 ∧ dB2 − m2 B2 ∧ B2 ]. (6.82)
2

The second one is instead solved by

1
B2 =  d  H3 , (6.83)
m2
which, inserted back in the action, gives the action for a massive one form if we
identify A1 = H3 :
  
1
S= dA1 ∧ dA1 − A1 ∧ A1 . (6.84)
m2

Generically, we do not introduce massive fields directly in our Lagrangian, but

rather vector fields become massive by means of gauge symmetry breaking, where
some of the scalar fields that are charged under a gauge symmetry are eaten by the
vector fields. Whenever this happens, we cannot dualize anymore the (massless)
scalar fields involved in the Higgs mechanism as described before. However, we
can see that, just like massless scalars are dual to massless tensor fields, the massive
vectors fields resulting from the spontaneous symmetry breaking are dual to massive
tensor fields. Once we are in the broken phase, the massive vector can be described
by a Stückelberg coupling everywhere in field space, and therefore we can always
find a field redefinition such that any of the scalar fields involved in the gauging is
coupled electrically to a vector field as

Aμ ≡ ∂μ φ + e Ãμ , (6.85)

which may be seen as the definition of the massive vector field Aμ , so that (6.84)
becomes
  2    
e
S= d à 1 ∧ d à 1 − dφ + e à 1 ∧  dφ + e à 1 . (6.86)
m2
122 6 Matter Couplings in Supergravity

The dual procedure by which the tensor field gets a mass is similar, but with vector
fields coupled to tensors

Bμν ≡ 2 ∂[μ Aν] + m B̃μν , (6.87)

so that (6.82) becomes


1     
S=− d B̃2 ∧ d B̃2 − dA1 + m B̃2 ∧  dA1 + m B̃2 . (6.88)
2

Other interesting couplings arise by considering higher derivative corrections to

the Lagrangian, which may become important when one discusses supergravity as
an effective theory.
For what concerns us, we stop here, because we have introduced all the main
ingredients needed for extracting the most relevant phenomenological differences
between supergravity and global supersymmetry.

6.2 Kähler–Hodge Manifolds

Gravity interactions have an impact also on the mathematical properties of the scalar
manifold, which remains a Kähler manifold, but now of a restricted type. Since this
is a more technical aspect, first-time readers may want to skip this section.
As seen in Sect. 6.1.2, local supersymmetry and invariance under Kähler trans-
formations, K → h + h∗ , require that the fermions and the superpotential transform
non-trivially under Kähler transformations:
% &
Im h
ψμ , , λ → exp −i
I
γ5 ψμ , , λI , (6.89)
2MP2
% &
Im h
χm → exp +i γ5 χ m , (6.90)
2MP2
% &
h
W → exp − 2 W. (6.91)
MP

In general, the Kähler potential is not a globally defined function on the Kähler
manifold but may be subject to Kähler transformations when one switches between
two overlapping coordinate patches UA and UB :

KA = KB + hAB + h∗AB , (6.92)

6.2 Kähler–Hodge Manifolds 123

where KA and KB are the local Kähler potentials on UA and UB , respectively, and
hAB is a holomorphic function on the overlap UA ∩ UB .
Because of (6.91), this means that W likewise needs to be patched together on
the manifold by means of Kähler transformations
% &
hAB
WA = exp − 2 WB (6.93)
MP

and hence cannot in general be viewed as a simple function on Mscalar , but rather
as a section of a holomorphic line bundle, L , over Mscalar .5 In order for such a line
bundle to be well defined, the local transition functions must fit together in a globally
consistent way. The important point for us is that the local transition functions are
not arbitrary holomorphic functions, but instead they are the exponentials of the
functions hAB that also describe how the local Kähler potentials KA , KB are patched
together. This relates the global consistency condition of the line bundle gluings to a
global restriction on the possible Kähler geometries that could occur in supergravity.
The same is true for fermions, from which we get the strongest restrictions.
Unlike the superpotential, the fermions transform with the exponential of i Imh
which means that they can be viewed as sections in a principal U(1) bundle that is
associated with the line bundle L . The connection on this U(1)-bundle is essentially
(6.14) but now viewed as a connection on the scalar manifold,

i  
Q≡ (∂n K)dφ n − (∂m K)dφ m , (6.94)
2
whose curvature is related to the Kähler form

dQ = J. (6.95)

Similar to the Dirac magnetic monopole (see also the next subsection for an explicit
example), the flux of the field strength of this connection through any topologically
non-trivial two sphere in Mscalar must then be quantized, and this in turn implies
that the Kähler form J has to obey a similar non-trivial quantization condition. In
the supergravity literature, one therefore sometimes uses the special term “Kähler–
Hodge manifolds” to denote Kähler manifolds that admit a line bundle, L , whose
principal bundle has a curvature equal to the Kähler form. To summarize: In 4D,
N = 1 supergravity, the scalar manifold of the chiral multiplets cannot just be
an arbitrary Kähler manifold but must be a Kähler–Hodge manifold, i.e., a Kähler

5A line bundle is simply a vector bundle where the fiber is a one-dimensional vector space, R or
C. A holomorphic line bundle is a line bundle with fiber C over a complex base manifold (here
our Kähler manifold Mscalar ) where the transition function between two local trivializations can be
chosen to be holomorphic. According to (6.93), this is the case for the superpotential W , which is
hence a section of a holomorphic line bundle, L , over Mscalar .
124 6 Matter Couplings in Supergravity

manifold that admits a holomorphic line bundle whose curvature is related to the
Kähler form. In global supersymmetry, by contrast, the superpotential and the
fermions do not transform under Kähler transformations. In the terminology of
this subsection, this corresponds to a trivial line bundle with flat curvature that is
unrelated to the Kähler form, which hence remains unconstrained. Thus, any Kähler
manifold of the right dimension can serve as a target space for the chiral multiplet
scalars in global supersymmetry.
Just like the N = 1 case outlined here, also for N > 1 supersymmetry, the
transition from global to local supersymmetry imposes different constraints on the
scalar manifold. Once again this leads to the fact that N > 1 supergravity theories
cannot be described in terms of N = 1 models. For instance, N = 2 supergravity
allows quaternionic scalar manifolds (their definition will be given later), which are
not Kähler. Hence they cannot be used as the scalar σ -model of an N = 1 theory.
On the other hand, the rigid N = 2 counterpart is given by hyper-Kähler manifolds,
which are also Kähler. Hence in rigid N = 2 supersymmetry, one can always view
the theory as a special classes of N = 1 models.

6.2.1 An Example: Quantization of Newton’s Constant

Before discussing other properties of the supergravity action and its differences with
respect to global supersymmetry, we now present an explicit example of how global
conditions constrain the scalar manifold in supergravity, illustrating some of the
abstract discussion of the previous subsection. One curious non-trivial effect of these
global conditions is that Newton’s constant can become quantized in these models
[8] in a sense to be described below.
As an example of a non-trivial manifold admitting two cycles, consider

Mscalar = CP1  S 2 .

This manifold can be locally parameterized by one complex scalar field φ (which
we now take of canonical dimension zero after a suitable rescaling with MP ). More
precisely, we need to cover the manifold with two coordinate patches. We start with
the stereographic projection of the sphere onto the complex plane from one of the
poles as one coordinate system (Fig. 6.1). A good Kähler potential for this manifold
is then K/MP2 = N log(1 + φφ ∗ ), resulting in the metric

N
gφφ = . (6.96)
(1 + φφ ∗ )2

This gives an effective theory whose bosonic sector reads

  
1 μ ∗
MP2 4
d xe R − gφφ ∂μ φ∂ φ . (6.97)
2
6.2 Kähler–Hodge Manifolds 125

C
S2

Fig. 6.1 Stereographic projection coordinates

The coordinate system and the Kähler potential discussed above are good for the
entire S 2 except for the north pole. We therefore need to cover CP1 by an alternative
chart, with coordinate z = φ1 and Kähler potential Kz /MP2 = N log(1 + zz∗ ). As
expected, in the overlapping patch, the two Kähler potentials are related by a Kähler
transformation:

1 + φφ ∗ 1 + φφ ∗
Kφ − Kz = MP2 N log = M 2
N log (6.98)
1 + zz∗ P
1 + φφ1 ∗

= MP2 N log φ + MP2 N log φ ∗ (6.99)

= h(φ) + h∗ (φ ∗ ). (6.100)

The problem of this transition function h = MP2 N log φ is that it is multivalued

on C. However, in relating the two descriptions, what we really need is that the
transition function for the fermions,
% &  
1 ∗ N φ
exp h−h = exp log ∗ ,
4MP2 4 φ

be single valued. This is the case if and only if N is an even integer. Computing
explicitly this expression at the equator
   
N φ i
exp log ∗ = exp Nα ,
4 φ 2

where α is the angle parametrizing the equator, we see that N has to be in 2Z for
making the fibers fit together when we complete a full revolution (α = 2π compared
to α = 0). Hence, while in global supersymmetry N ∈ R, supergravity forces a
quantization of its value and N ∈ 2Z.
A curious fact is that now the quantization of N provides as a consequence the
quantization of Newton’s constant. In fact, when discussing scalar σ -models (φ is
126 6 Matter Couplings in Supergravity

like a pion field), most of the physics is summarized in a dimensionful coupling

constant, f 2 , in front of the kinetic term of the scalars (which we generically
reabsorbed in the fields). In cases where the σ -model is a coset manifold G/H,
f is related to the vacuum expectation value of the field which breaks the global
symmetry group G to a subgroup H in the effective theory. In any case, from the
kinetic term of the σ -model under consideration, we have that

N
f 2 = NMP2 = , (6.101)
8πGN

or, turning this around,

N
GN = . (6.102)
8πf 2

We therefore see that Newton’s constant is quantized in units of f .

Exercises

6.1. Verify that (6.25) and (6.26) are required by the consistency of the last of (6.39)
and the holomorphicity of W .

6.2. Consider the same model considered in the globally supersymmetric case with
three chiral multiplets and one vector multiplet (cf. Problem 5.4). This model has
Kähler potential

K = |φ|2 + |A|2 + |B|2

and superpotential

W = φAB.

The three scalar fields are charged under a U(1) symmetry, which is gauged and
generates a Fayet–Iliopoulos term. Their charges are qφ , qA , and qB , not specified
for the moment. The kinetic term for the vector field is canonical. Considering the
coupling to gravity:

(a) Compute the scalar potential.

(b) Find the supersymmetric and non-supersymmetric vacua.
(c) Compute the boson masses at the vacua.
(d) Find out for which value of the charges we can get a potential of the form V =
g2 2
2 η , where g is the gauge coupling constant.
(e) Discuss the type of supersymmetry breaking.
References 127

6.3. Find the Killing vectors of the isometries of the scalar manifold generated by
 
i
K = − log − (s − s̄)((t − t¯)2 − (u − ū)2 ) .
2

Compute its prepotentials and discuss the invariance of the Kähler potential under
the action of these Killing vectors.

References
1. E. Cremmer, S. Ferrara, L. Girardello, A. Van Proeyen, Coupling supersymmetric Yang-Mills
theories to supergravity. Phys. Lett. B 116, 231–237 (1982)
2. E. Cremmer, S. Ferrara, L. Girardello, A. Van Proeyen, Yang-Mills theories with local
supersymmetry: Lagrangian, transformation laws and SuperHiggs effect. Nucl. Phys. B 212,
413 (1983)
3. J.A. Bagger, Coupling the gauge invariant supersymmetric nonlinear sigma model to supergrav-
ity. Nucl. Phys. B 211, 302 (1983)
4. H. Elvang, D.Z. Freedman, B. Körs, Anomaly cancellation in supergravity with Fayet–Iliopoulos
couplings. JHEP 11, 068 (2006). [arXiv:hep-th/0606012 [hep-th]]
5. J. De Rydt, J. Rosseel, T.T. Schmidt, A. Van Proeyen, M. Zagermann, Symplectic structure of
N=1 supergravity with anomalies and Chern-Simons terms. Class. Quant. Grav. 24, 5201–5220
(2007). [0705.4216]
6. P. Binetruy, G. Dvali, R. Kallosh, A. Van Proeyen, Fayet–Iliopoulos terms in supergravity and
cosmology. Class. Quant. Grav. 21, 3137–3170 (2004). [hep-th/0402046]
7. R. D’Auria, S. Ferrara, On fermion masses, gradient flows and potential in supersymmetric
theories. JHEP 05, 034 (2001). [arXiv:hep-th/0103153 [hep-th]]
8. E. Witten, J. Bagger, Quantization of Newton’s constant in certain supergravity theories. Phys.
Lett. B115, 202 (1982)
Phenomenological Aspects
7

In previous chapters, we gave a complete picture of the Lagrangian and the

supersymmetry rules for N = 1 supergravity coupled to vector and chiral
multiplets. In the present chapter, we now take a closer look at these theories
from a phenomenological point of view. Our main focus will again be on those
phenomenological aspects that differ in important ways between locally and
globally supersymmetric theories. Most of these phenomenological differences can
be traced back to:

• The different form of the scalar potential in global and local supersymmetry
• The presence of the gravitino
• The presence of gravity and other non-renormalizable interactions in supergrav-
ity

Our discussion includes the important topic of supersymmetry breaking, various

aspects associated with the gravitino, as well as some cosmological consequences.

7.1 Spontaneous Supersymmetry Breaking

7.1.1 Vacua

Having a supersymmetric theory that involves gravity, it would now be very

tempting to investigate the general interplay between supersymmetry, spacetime
geometry, and the various matter fields. For many phenomenological questions,
however, it is sufficient to restrict oneself to solutions of 4D, N = 1 matter-coupled
supergravity in which the spacetime metric is maximally symmetric, i.e., either
Minkowski, de Sitter, or anti-de Sitter spacetime. We will refer to such solutions
as “vacua” of the corresponding supergravity theory. The only non-trivial field
configurations that could be consistent with the 4D homogeneity and isotropy of

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 129

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_7
130 7 Phenomenological Aspects

these spacetimes are the metric and constant scalar fields (or possibly also constant
Lorentz invariant condensates of fermionic fields, such as gaugino condensates,
which, however, we will not discuss here).1
Given a theory with nS real scalar fields, ϕ i (x) (i, j, . . . = 1, . . . , nS ), the scalar
field equation with constant scalars ϕ i (x) = ϕci and all other non-gravitational fields
set to zero reduces to

∂V (ϕ) 
∂i V |ϕ i =ϕci ≡ = 0, (7.1)
∂ϕ i ϕ i =ϕci

i.e., the scalar fields have to be at a critical point of the scalar potential. In the
following, we will, in a semi-classical sense, often refer to quantities that are
evaluated at the critical point ϕ i = ϕci as the vacuum expectation value (vev) of
that quantity and denote it with . . ., so that, e.g., (7.1) becomes ∂i V  = 0.
The value V  ≡ V (ϕc ) of the potential at the critical point then is the vacuum
energy density, ρvac , which enters the Einstein equation as a cosmological constant
Λ = ρvac MP−2 = V MP−2 , so that V  = 0, V  > 0 and V  < 0 correspond to,
respectively, Minkowski, de Sitter, and anti-de Sitter spacetime.
A vacuum of the above type is stable against small fluctuations, ϕci → ϕci + δϕ i ,
of the scalar field if the Hessian, ∂i ∂j V , at the critical point is positive definite.
If some eigenvalues of the Hessian are zero, the potential may have flat directions
at the critical point, which correspond, for V  ≥ 0, to marginal stability. Field
directions along which the Hessian has a negative eigenvalue are called tachyonic
field directions or tachyons and imply a perturbative instability of the vacuum along
that direction if V  ≥ 0. For V  < 0, on the other hand, the solution may be
stable even for tachyonic field directions, as long as their mass eigenvalue satisfies
the Breitenlohner–Freedman bound we discussed in (4.87).
In supergravity, we deal with complex scalar fields, φ m (x) (m, n, . . . =
1, . . . , nC ), and their complex conjugates, φ m (x). In terms of these, the critical
point condition is correspondingly ∂m V  = 0, and the stability is then dictated by

1 Our definition of a “vacuum” here should not be confused with the notion of a “vacuum solution”

in general relativity, which is more general and refers to solutions to Einstein’s equation without
contributions to the energy-momentum tensor Tμν from any kind of matter excitation. Note that a
constant scalar field φ(x) = φ0 = const. does not count as an excitation here and hence should
be allowed in such a vacuum solution, because its contribution to Tμν arises only from its constant
potential V (φ(x)) = V (φ0 ), which is indistinguishable from a contribution due to a cosmological
constant, Λ, in the Einstein equation. In other words, our maximally symmetric “vacua” are
special cases of the “vacuum solutions” of general relativity, which, however, also encompass less
symmetric matter-free solutions such as the Schwarzschild metric or a gravitational wave in empty
space. Viewing the inhomogeneities of the metric in these solutions as gravitational excitations,
one could also characterize our “vacua” as solutions that are free of matter and gravitational
excitations and hence form the natural semi-classical analogues of the Poincaré-invariant vacua
in conventional quantum field theories on Minkowski spacetime.
7.1 Spontaneous Supersymmetry Breaking 131

the eigenvalues of the matrix

 
∂m ∂n V  ∂m ∂n V 
. (7.2)
∂m ∂n V  ∂m ∂n V 

The stability and many other phenomenologically important properties of a given

vacuum crucially depend on whether supersymmetry is preserved or broken in this
vacuum. We will therefore now introduce some basic terminology of spontaneous
supersymmetry breaking in supergravity.

7.1.2 General Features of Spontaneous Supersymmetry Breaking

In this subsection, we analyze the conditions for spontaneous supersymmetry

breaking in matter-coupled 4D, N = 1 supergravity and explore some of their
most immediate phenomenological consequences. As explained in the previous
subsection, we will do this only for maximally symmetric spacetimes, but it
should be mentioned that supersymmetry breaking can also be discussed for other
physically interesting backgrounds such as black hole spacetimes, cosmic strings,
domain wall solutions, or instantons with many interesting theoretical applications.
For simplicity, we will also not consider non-trivial fermion condensates, so that the
only fields that can have a non-trivial vev are the scalar fields (with constant vevs)
and the metric.2
In matter-coupled 4D, N = 1 supergravity, the scalar potential is the sum of a
(Kähler covariantized) F-term contribution, VF , a D-term contribution, VD , and a
new supergravity contribution, VG :
 
|W |2
K
2 1
V = VF − VG + VD = e MP
|DW |2 − 3 2 + RefI J D I D J , (7.3)
MP 2

where, for convenience, in this section we introduced the splitting

K
2
VF = e MP |DW |2 , (7.4)

1
VD = RefI J D I D J , (7.5)
2
K
2 |W |2
VG = 3e MP , (7.6)
MP2

2 Effects such as gaugino condensation, where strongly coupled gauge dynamics induce a non-
trivial vev for gaugino bilinears, λλ = 0, can, in an effective field theory below the condensation
scale, often be described by an additional scalar field and contribute to spontaneous supersymmetry
breaking in the effective scalar field sector. This effective scalar field dynamics would be contained
in our analysis.
132 7 Phenomenological Aspects

as this is useful in the following.3 When the metric and the (constant) scalar fields
are the only non-trivial fields, the supersymmetry rules can be written as

δψμL = MP Dμ L + γμ S R , (7.7)

δχLm = N m L , (7.8)

δλIL = N I L . (7.9)

Here, S is the fermionic shift in the gravitino transformation, which enters the
Lagrangian also as a mass term for the gravitino (cf. also the next subsection),
K
1 2
S≡ e 2MP W, (7.10)
2MP

whereas the shift of the chiralini is the covariantized F-term,

K
1
≡ − g mn e 2MP Dn W ∗ , ≡ (N m )∗ ,
2
N m
N m
(7.11)
2
and the shift of the gaugini is given by the D-terms,

i I
NI ≡ D . (7.12)
2
As we already explained in Sect. 6.1.6, supersymmetry gives an extremely useful
relation between the “shifts” in the supersymmetry transformations of the fermions
and the scalar potential. More concretely, the latter is simply given by the square of
the shifts in (7.7)–(7.9), with appropriate signs and prefactors:

V = −12 SS ∗ + 4 gmn N m
N n
+ 2 Re(fI J ) N I N ∗J . (7.13)

We see that the fermionic shifts from the matter fields always give positive
semi-definite contributions, while the shifts from the gravitini give a negative semi-
definite contribution. This relation between fermionic shifts in the supersymmetry
transformations and the scalar potential is an important consequence of a Ward
identity, similar to (4.63), which also generalizes to extended supersymmetry and
other dimensions.

3Note that what we call VF − VG here is usually called the “F-term potential” and would usually
be denoted by VF .
7.1 Spontaneous Supersymmetry Breaking 133

Having expressed the scalar potential in terms of the fermionic shifts, we can
easily read off some important properties of the scalar potential and its vacua:

• The scalar potential can be positive, negative, or zero. This is the first striking
difference with respect to the globally supersymmetric case. The reason is the
negative contribution coming from the shift of the gravitino supersymmetry
transformation rule.
• Supersymmetry is preserved if and only if the supersymmetry variations of the
fermionic fields are vanishing in the background under consideration: δψμ  =
δχ m  = δλI  = 0. This requires both (we recall that for W  = 0, the second
condition follows from the first).

N m
=0 ⇔ Dm W  = 0 (7.14)

and

N I  = 0 ⇔ D I  = 0. (7.15)

On the other hand, there is no condition on the gravitino shift S, because a non-
trivial S may be cancelled by a non-trivial Dμ . Taking the covariant derivative
of MP Dμ L = −γμ SR , however, one finds that this is possible only in anti-de
Sitter space (see the discussion in Sect. 4.2 and the next item).
• For supersymmetric vacua the potential is negative (semi-)definite:
' (
∗
K
2 |W |2
V  = −12SS  = −3 e MP
≤ 0. (7.16)
MP2

This condition is also different from the globally supersymmetric case, where
the vev of V had to vanish for unbroken supersymmetry. A supersymmetric
Minkowski spacetime in supergravity is recovered only when the superpotential
is vanishing on the vacuum: W  = 0. In this case, unbroken supersymmetry is
equivalent to ∂m W  = W  = D I  = 0. By contrast, when the superpotential
is not vanishing on the vacuum, W  = 0, the scalar potential has a negative
value, and therefore we obtain a supersymmetric configuration with negative
cosmological constant: the vacuum is an anti-de Sitter space.
• Conversely, having a negative or vanishing potential in the vacuum does not
imply unbroken supersymmetry. This is also different from the globally super-
symmetric case, where only a positive cosmological constant could break
supersymmetry. For instance, we could have Dm W  = 0 and W  = 0 at the
same time in a way that they compensate in the potential so that V  = 0. A
nice example for this are the no-scale models, which we introduce in Sect. 7.4.2.
Also, it should be clear that supersymmetry breaking does not imply W  = 0,
because we could have Dm W  = 0, but W  = 0, which would imply a positive
cosmological constant.
134 7 Phenomenological Aspects

• In supergravity, D-terms and F-terms are related through (6.57). Hence, when
both are present in the potential, there is no pure D-term breaking unless the
superpotential vanishes on the vacuum. In this case, the relation (6.57) may
yield a finite result even for vanishing Dm W . In any case, we cannot uplift
supersymmetric AdS vacua (where necessarily W  = 0) to non-supersymmetric
ones only by introducing a pure D-term breaking.

7.1.3 Mass Scales Related to Supersymmetry Breaking

The spontaneous breakdown of supersymmetry in a phenomenologically realistic

model defines a number of important mass scales. In this subsection, we define
these mass scales and exhibit the model-independent relation between some of them.
Other relations are more model-dependent and will be discussed in the subsequent
sections of this chapter, where we will focus on the sectors that define these mass
scales in more detail.

7.1.3.1 The Supersymmetry Breaking Scale Msusy

As we have just seen, the relevant quantity that signals spontaneous supersymmetry
breaking in supergravity are the vevs of the F-term shifts of the chiralini and the
D-term shifts of the gaugini. Just as in global supersymmetry, we therefore define
the supersymmetry breaking scale, Msusy , by

4
Msusy ≡ VF  + VD  (7.17)

as the natural order parameter of spontaneous supersymmetry breaking. In a

generic supersymmetry breaking scenario, we expect Msusy to be larger than the
electroweak scale, Mew , because no sparticle has been observed yet.

7.1.3.2 The Vacuum Energy Scale Mvac

In spontaneously broken global supersymmetry, the supersymmetry breaking scale
Msusy coincides with the mass scale, Mvac , of the tree-level vacuum energy
density, ρvac = Mvac 4 , which, in the presence of gravity, would be related to the

cosmological constant Λ as Mvac4 = MP2 Λ. In supergravity, however, Msusy and

Mvac decouple, because in supergravity the latter also depends on the contribution,
VG , from the gravitino shift,
4
Mvac = VF  + VD  − VG  = Msusy
4
− VG . (7.18)

This means that in supergravity a phenomenologically realistic value, Msusy >

102 GeV, can coexist with the tiny measured value of the vacuum energy density,
4
Mvac ∼ (10−3 eV)4 , already at tree level. It is obvious from these considerations
that it does not make much sense to compute the vacuum energy density in
supersymmetry without taking into account gravity effects. An interesting class of
7.1 Spontaneous Supersymmetry Breaking 135

models that have a tree-level Minkowski vacuum for arbitrary large Msusy are the
no-scale models, which we discuss in Sect. 7.4.2.

7.1.3.3 The Gravitino Mass M3/2

Another important mass scale is the mass of the gravitino after spontaneous
supersymmetry breaking. It has a simple relation to Msusy if one neglects the small
observed cosmological constant and approximates Mvac ≈ 0. The (real) gravitino
mass is then (omitting the brackets for vevs here)
*
) + 4
K
|W | + Msusy − Mvac
4 2
Msusy
M3/2 = e
2
2MP
=
VG
= , ≈ √ (7.19)
MP2 3MP2 3MP2 3MP

We see that, for Msusy > Mew = 246 GeV , the gravitino mass could a priori be
anywhere between the Planck and the milli-electronvolt scale. In order to constrain
it further, one would have to feed in the desired mass scale for the superpartners of
the Standard Model particles and to specify a particular supersymmetry breaking
mechanism. We will discuss the gravitino sector in Sect. 7.2.

7.1.3.4 The Soft Masses Msoft

The abovementioned masses of the superpartners of the Standard Model particles
are among the soft supersymmetry breaking terms that one can add to a supersym-
metric extension of the Standard Model without introducing quadratic divergences
to the Higgs mass. These “soft masses” must be above the energy scales that have
already been probed by accelerators, i.e., at least near the TeV scale, with the precise
values depending on the particle type and the assumed model.
In order to address naturalness issues of the Standard Model, the soft masses
would have to be not too far above the electroweak scale either, which would lead
to a typical scale, Msoft , of the soft masses somewhere near the TeV scale.
It should be emphasized, however, that Msoft is not a precise value for a given
model but that the actual soft masses may of course cover a certain range. In models
that address the hierarchy problem, this range cannot be too big, but if one abandons
supersymmetry as a solution to the hierarchy problem, this mass range can be
substantially bigger and in fact define several hierarchically different mass scales.
This is in particular the case in models with split supersymmetry [1, 2], where the
fermionic superpartners are assumed to be near the TeV scale, whereas the scalar
masses could be substantially bigger without spoiling coupling unification or the
existence of possible dark matter candidates among the fermionic superpartners.
In any case, the soft masses are related to the gravitino mass and the super-
symmetry breaking scale in different ways depending on the particular scheme of
supersymmetry breaking. This will be discussed in Sect. 7.3.

7.1.3.5 Moduli Masses Mmod

Moduli are scalar fields that couple to ordinary matter such as the Standard Model
particles only with very weak, typically MP -suppressed, interactions. Moduli are
136 7 Phenomenological Aspects

a typical ingredient of string compactifications but may also be introduced for

purely phenomenological reasons, e.g., as a sector that realizes spontaneous
supersymmetry breaking. As we will discuss, a natural scale for the moduli masses
is the gravitino mass, but one can easily construct models where at least some
moduli are much lighter or much heavier than the gravitino. We will discuss moduli
masses in Sect. 7.4.1.

7.1.3.6 The Inflationary Energy Scale Minf

This energy scale is not directly related to our present vacuum but refers to the
potential energy density, ρinf = Minf 4 , of the inflaton during early Universe

inflation. This potential energy must be high enough to enable the standard hot Big
Bang scenario after reheating that leads to the present Universe. It therefore is also
associated with a strong breaking of supersymmetry during the period of inflation
so that we include it in this list. Some basic aspects of inflation in supergravity will
be discussed in Sect. 7.4.4.

7.1.4 Rigid Limits

In our discussions above, we have seen that the question of vacuum energy can
only be studied in a meaningful way when one also takes into account the effects
from supergravity. This means that if one is interested in the limit of global
supersymmetry, such a limit should only be taken after a specific vacuum has been
chosen in the supergravity theory.
Once a vacuum has been identified in supergravity, it is clear that a rigid limit
should involve sending MP → ∞. This limit kills all the supergravity related
interactions in the Lagrangian and also modifies the supersymmetry transformation
rules, as we discussed earlier. However, the supersymmetry breaking scale and the
gravitino mass may also depend on MP so that some care is required when this limit
is performed and certain quantities are to be kept fixed. There are in fact two distinct
procedures that could be taken in order to recover a proper rigid limit in the presence
of supersymmetry breaking:

• Keep M3/2 fixed.

This means that the supersymmetry breaking scale is also taken to be high
2
Msusy ∼ M3/2 MP → ∞. (7.20)

The resulting global theory features explicitly broken supersymmetry with tree-
level soft terms of O(M3/2) and a decoupled goldstino (eaten up by the
gravitino).4

4 The couplings of the goldstino are proportional to inverse powers of Msusy and hence vanish in
this limit.
7.2 Gravitino, Goldstino, and Super-Higgs Mechanism 137

• Keep Msusy fixed. This implies

2
Msusy
M3/2 ∼ → 0. (7.21)
MP

Hence we obtain a scenario with spontaneously broken supersymmetry (with a

goldstino interacting with matter) and decoupled gravitino.

One thing that has to be stressed in this procedure is that the rigid limit may affect
the geometry of the scalar manifold in a non-trivial way. Indeed, the rigid and local
geometries of the scalar σ -models are generically different, although for minimal
supersymmetry the local case can be seen as a subcase of the rigid one and only
concerns the global structure of the manifold (see Sect. 6.2). As we will discuss in
Chap. 8, the difference between the geometries are more pronounced in extended
supersymmetry and also concern the local structure. This phenomenon has not been
thoroughly studied in the literature, and the phenomenological literature does not
seem to be aware of the problem, mainly because usually only topologically trivial
models are discussed. In general [3], only a subset of the scalars of the supergravity
theory appears in the rigid limit for N > 1, though for N = 1 we can still keep
all of them, as we would expect.

7.2 Gravitino, Goldstino, and Super-Higgs Mechanism

When supersymmetry is spontaneously broken in a Minkowski or AdS vacuum,

the gravitino field acquires a Lagrangian mass term proportional to the (necessarily
non-vanishing) expectation value of the superpotential, though in an AdS vacuum,
as explained in Sect. 4.2.2, the Lagrangian mass term may not coincide with the
actual mass definition. In a de Sitter vacuum, supersymmetry is always broken, and
the vev of the superpotential may or may not be zero.
In any case, a massive gravitino should acquire new spin-1/2 polarizations with
respect to the massless case, just as a gauge boson acquires an additional spin-0
degree of freedom when a gauge symmetry is spontaneously broken through the
Higgs mechanism. Since we are now dealing with supersymmetry, this process
has been named super-Higgs mechanism [4]. The goldstino, the massless fermion
related to the spontaneous breaking of supersymmetry, is “eaten” by the gravitino,
giving the latter the missing degrees of freedom to become a massive particle.
The details of this process depend on the type of vacuum one considers. We
will discuss here in detail only the case of pure F-term breaking in a Minkowski
vacuum. A non-trivial cosmological constant or the presence of D-terms would
change some of the details of the computation, but the general scheme of the super-
Higgs mechanism should become clear already in our restricted setup and can be
worked out along the same lines [5].
138 7 Phenomenological Aspects

We recall that in ordinary gauge theories the Goldstone bosons correspond to

those directions in field space along which the broken symmetry generators act
non-trivially in a given vacuum. Alternatively, they can be identified as the scalar
field fluctuations that have a non-trivial coupling to the gauge fields of the broken
symmetries in the vacuum. They are zero eigenvectors of the scalar mass matrix and
can be absorbed by a redefinition of the vector fields of the broken symmetries. All
these properties of the Goldstone boson have a close analogue for the goldstino.

7.2.1 The Goldstino in Global Supersymmetry

In order to exhibit the difference between global and local supersymmetry most
clearly, let us briefly recapitulate the properties of the goldstino in a globally
supersymmetric theory with nC chiral multiplets and pure F -term supersymmetry
breaking.
The fermionic part of the Lagrangian is
 
LF = −gmn χ m
L DχR + χ R DχL
/ n n / m

(7.22)
∗
− (Dm ∂n W )χ m n
L χL − (Dm ∂n W )χ m n
R χR

and the supersymmetry transformations of the chiralini are

1 m 1 1 m 1
δχLm = ∂/φ εR − F m εL , δχRm = ∂/φ εL − F m εR , (7.23)
2 2 2 2

where F m ≡ g mn ∂n W ∗ and F m ≡ (F m )∗ . Considering the fermions on a

Minkowski background with constant scalar fields, we can replace Dμ → ∂μ , and
the supersymmetry transformations become

1 1
δχLm = − F m εL , δχRm = − F m εR . (7.24)
2 2
Here and in the following, all scalar field-dependent quantities such as F m , gmn ,
etc. are meant to denote their constant vevs (i.e., we drop the brackets for vevs
for simplicity), which are then also inert under supersymmetry transformations.
Obviously the supersymmetry transformation is non-trivial only along the direction
defined by F m and its complex conjugate in field space, and hence we define the
goldstino field, ζ , as the projection of the chiralini along that direction,

Fm χ m
ζL := √ L n (7.25)
Fn F

and similarly for ζR , where the denominator ensures canonical normalization. The
chiralini χLm can then be decomposed into ζL and its orthogonal complement, χLm⊥ ,
7.2 Gravitino, Goldstino, and Super-Higgs Mechanism 139

with respect to the vev of the Kähler metric,

F m ζL F m ζL
χLm = √ + χ m
L − √ , (7.26)
Fn F n F Fn
  n 
:=χLm⊥

so that the kinetic term in (7.23) becomes

 
− [ζ L ∂/ζR + ζ R ∂/ζL ] − gmn χ m⊥
L
/
∂ χ n⊥
R + χ n⊥ / m⊥
R ∂ χ L , (7.27)

and

Fm χLm⊥ = 0. (7.28)

The supersymmetry transformations of ζL and χLm⊥ are

1-
δζL = − Fm F m εL , (7.29)
2
δχLm⊥ = 0, (7.30)

so that indeed the goldstino ζ captures the full supersymmetry transformation.

To show that the goldstino is a massless fermion, we need the critical point
condition, ∂l V = 0, which can be written as

0 = ∂l V = Dl V = Dl (∂m Wg mn ∂n W ∗ ) = (Dl ∂m W )g mn ∂n W ∗ = (Dl ∂m W )F m ,

(7.31)
where Dl g mn = 0 and Dl ∂n W ∗ ≡ ∂l ∂n W ∗ − Γln ∂p W ∗ = 0 have been used. As the
p
p
matrix (Dl ∂m W ) ≡ (∂l ∂m W − Γlm ∂p W ) is manifestly symmetric in l and m, we
thus obtain

0 = (Dl ∂m W )F m = (Dm ∂l W )F m . (7.32)

Using this equation, it is then easy to see that

− (Dm ∂n W )χ m
L χL = −(Dm ∂n W )χ L χL
n m⊥ n⊥
(7.33)

i.e., the goldstino drops out of the fermionic mass term and hence is indeed massless.
140 7 Phenomenological Aspects

7.2.2 The Goldstino and the Gravitino in Supergravity

As we will now discuss, the situation in local supersymmetry has many similarities
to the above but also some important differences that have to do with the presence
of the gravitino as the gauge field of local supersymmetry.
Before we start, we collect some useful relations for the Kähler invariant
function, G ≡ K + MP2 log(|W |2 /Mp6 ), and its derivatives that will help us keeping
the equations in a compact form. We define Gn ≡ ∂n G , G m ≡ g mn̄ ∂n̄ G , etc. so that

∂m W Dm W
Gm = ∂m K + MP2 = MP2 , (7.34)
W W
 
Dn W ∂m Dn W ∂m W Dn W
∂m Gn = MP2 ∂m = − MP2 , (7.35)
W W W2
 
Dm Dn W Dm W Dn W
Dm Gn = ∂m Gn − Γmn
k
Gk = − MP2 , (7.36)
W W2

where we also used the fact that we have full covariantization of the derivatives with
respect to Kähler transformations and field redefinitions, like in (6.38). We recall
the form of the scalar potential in terms of G , as in (6.41)
 
K
|W |2  
∗
= eG /MP MP2 g mn Gm Gn̄ − 3MP4 ,
2 2
V =e MP
g mn
Dm W Dn W − 3 2
MP
(7.37)

so that the Minkowski condition, V = 0, implies

Gm G m = 3MP2 . (7.38)

Furthermore, the condition that we are at a critical point of the potential gives (by
recalling that ∂m Gn̄ = ∂m ∂n̄ K = gmn̄ and ∂l g mn̄ = −Γlkm g k n̄ , as in (5.48)):

G
2  
0 = ∂n V |V =0 = e MP MP2 G m Dn Gm + Gn , (7.39)

which implies

(Dn Gm )G m = −Gn . (7.40)

7.2 Gravitino, Goldstino, and Super-Higgs Mechanism 141

The fermionic part of the Lagrangian in a Minkowski background with constant

scalar fields and no vector multiplets is

1  
LF = − ψ μ γ μνρ ∂ν ψρ − gmn χ m
L
/
∂ χ n
R + χ n/ m
R ∂ χ L
2
% &
2S S
+ −Mmn χL χL + 2 ψ μR γ χL Gm +
m n μ m
ψ γ ψνR + h.c. ,
μν
MP MP μR
(7.41)

where
2 2S  2 
Mmn := eK/2MP (Dm Dn W ) = 3
MP Dm Gn + Gm Gn , (7.42)
MP

S denotes the gravitino shift (7.10), and we have dropped a mixing term between the
gravitino and the chiralini involving derivatives of the (here still constant) scalars.
The supersymmetry transformations of the fermions in such a background are

δψμL = MP ∂μ εL + Sγμ εR (7.43)

1 S∗ m
δχLm = − eK/2MP g mn Dn W ∗ εL = −
2
G εL , (7.44)
2 MP

We are now ready to discuss the goldstino and the super-Higgs effect. In
the case of pure F-term breaking at V = 0, the goldstino is, just as for global
supersymmetry, given by the linear combination of the chiralini along the direction
of supersymmetry breaking, i.e., along G m ,

Gm χ m Gm χ m
ζL := √ L n = √ L , (7.45)
Gn G 3MP

where we used (7.38) in the second equality.

Just as for global supersymmetry, we then decompose χLm into ζL and its
orthogonal complement, χLm⊥ ,
 
Gm Gm
χLm = √ ζ L + χ m
L − √ ζ L , (7.46)
Gn G n Gn G n
  
=:χLm⊥
142 7 Phenomenological Aspects

where the normalizations have again been chosen such that the kinetic term
factorizes and is canonically normalized,

−gmn [χ m
L ∂ χR + χ R ∂ χL ] = −[ζ L ∂ ζR + ζ R ∂ ζL ]
/ n n/ m / /
 
− gmn χ m⊥L ∂ χR + χ R ∂ χL
/ n⊥ n⊥ / m⊥
, (7.47)

and, moreover,

Gm χLm⊥ = 0. (7.48)

The supersymmetry transformation (7.44) then becomes

√
δζL = − 3S ∗ εL (7.49)
δχLm⊥ =0 (7.50)

so that the goldstino indeed captures the entire supersymmetry transformation.

In the case of global supersymmetry, the goldstino dropped out of the mass
term for the chiralini and was thus recognized as a massless fermion. In local
supersymmetry, the situation is a bit more complicated due to the presence of the
gravitino. In order to see this, we first note that Eqs. (7.38) and (7.40) now imply

4S
Mmn G n = Gm , (7.51)
MP

which does not vanish, unlike the analogous expression (7.32) in global supersym-
metry. This in turn implies that
 
4S
− Mmn χ m n
L χL =− ζ ζL + Mmn χ L χL ,
m⊥ n⊥
(7.52)
MP L

i.e., although the goldstino decouples from its orthogonal complement in the naive
mass matrix of the chiralini, it does get a non-vanishing mass contribution from it.
This is different from the analogous expression (7.33), where the goldstino drops
out. In contrast to global supersymmetry, however, this is not yet the end of the
story, as in supergravity there is also a mixing term between the gravitino and the
chiralini in (7.41), which upon using (7.46) becomes
√
2 3S
ψ μR γ μ ζL + h.c. (7.53)
MP
7.2 Gravitino, Goldstino, and Super-Higgs Mechanism 143

Putting everything together, the fermionic action (7.41) now becomes

1  
LF = − ψ̄μ γ μνρ ∂ν ψρ − [ζ L ∂/ζR + ζ R ∂/ζL ] − gmn χ m⊥
L
/
∂ χ n⊥
R + χ n⊥ / m⊥
R ∂ χ L
2
% √ &
4S 2 3S S
+ − ζ ζL + ψ μR γ ζL +
μ
ψ γ ψνR + h.c.
μν
(7.54)
MP L MP MP μR
 
− Mmn χL m⊥ χLn⊥ + h.c.

In order to understand the mass spectrum, we have to get rid of the mixing
term between the goldstino and the gravitino. This could be done by expressing
√
everything in terms of a redefined gravitino field of the form ψμL + γμ ζR / 3, but
this would introduce a mixing term in the kinetic terms. To also remove that one,
one instead has to use the redefined gravitino

1 1
ψμL ≡ ψμL + √ ∂μ ζL + √ γμ ζR , (7.55)
3S ∗ 3

in terms of which (7.54) becomes

1  
LF = − ψ μ γ μνρ ∂ν ψρ − gmn χ m⊥
L ∂ χR + χ R ∂ χL
/ n⊥ n⊥ / m⊥
2
   (7.56)
S 
+ ψ μR γ μν ψνR + h.c. − Mmn χL m⊥ χLn⊥ + h.c. .
MP

We notice that the goldstino has not only decoupled from the gravitino but that it
actually has disappeared completely from the action. Moreover, when we compare
(7.55) with the supersymmetry transformation (7.43), we realize that the transition
from ψμ to ψμ precisely
√ ∗ takes the form of a local supersymmetry transformation
with εL = ζL /( 3S ), which in turn would precisely transform ζL to zero. We
thus realize that in local supersymmetry, the goldstino becomes a pure gauge degree
of freedom that can be entirely absorbed as the longitudinal mode of the massive
gravitino. The gauge where ζ = 0 is called the unitary gauge.
We can finally read off the gravitino mass from the above equation:

2|S|
M3/2 = . (7.57)
MP

This is equivalent to the expression given in (7.19).

Sometimes it is easier to compute the fermion masses directly from the original
χLm basis, but taking into account the field redefinition (7.55) for the gravitino. In
this case one diagonalizes a mass matrix that contains nC fields, among which there
144 7 Phenomenological Aspects

is the goldstino. This matrix is

 
. 2 S 2S 1
M mn ≡ M mn − G G
m n = M 2
D
P m nG + G G
m n , (7.58)
3 MP3 MP3 3

which indeed satisfies

.
M mn G = 0.
n
(7.59)

7.2.3 Gravitino Couplings

So far, we considered constant scalar fields and a fixed Minkowski background, but
the super-Higgs mechanism also works when all fields are dynamical. In order to
understand a subtlety of the gravitino couplings to matter fields in theories with
spontaneously broken supersymmetry, it is useful to briefly also sketch this general
case.
To this end, we start from the general supergravity Lagrangian with nC chiral
multiplets,

MP2 1  
e−1 L = R − ψ μ γ μνρ Dν ψρ − gmn (∂μ φ m )(∂ μ φ n ) + χ m
L DχR + χ R DχL
/ n n / m
2 2
 2
 1  μ μ

− eK/2MP (Dm Dn W )χ m χ
L L
n
+ h.c. − J R ψμR + J L ψμL (7.60)
MP

1  
+ Sψ μR γ μν ψνR + S ∗ ψ μL γ μν ψνL − V + e−1 L4F ,
MP

with the supercurrent

μ 2S ∗
J L = −gmn χ m
L γ ∂φ +
μ/ n
Gn χ nR γ μ (7.61)
MP

and the supersymmetry transformation laws

1
δeμa = εγ a ψμ , (7.62)
2MP

δψμL = MP Dμ εL + Sγμ εR , (7.63)

δφ m = εL χLm , (7.64)

1 m S∗ m
δχLm = ∂/φ εR − G εL . (7.65)
2 MP
7.2 Gravitino, Goldstino, and Super-Higgs Mechanism 145

As the scalar fields are now dynamical, we have to distinguish carefully again the
dynamical fields φ m (x) and their vevs φ m  and define the shifted fields

Δφ m (x) := φ m (x) − φ m  (7.66)

and similarly for ΔG m , ΔS, etc. The goldstino, ζ , and its orthogonal complements
χ m⊥ are then given by

Gm  G m  m
ζL : = √ χ m
= √ χL , (7.67)
Gm G m  L 3MP
G m 
χLm⊥ := χLm − √ ζL . (7.68)
Gm G m 

For the goldstino we then have the following supersymmetry variation

 
Gm  1 m S∗ m
δζL = √ ∂ φ εR −
/ G εL (7.69)
Gm G m  2 MP
 
Gm G m S ∗  Gm  1 m Δ(S ∗ G m )
=− √ ε L + √ /
∂ φ ε R − ε L (7.70)
MP Gm G m  G G m  2 MP
 m  
=:δ  ζL
√
= − 3S ∗ εL + δ  ζL . (7.71)

In the case of
√ constant scalar fields, one has δ  ζL = 0, and one can simply choose
∗
εL = ζL /( 3S ) to gauge away the goldstino. In the general case with non-
vanishing δ  ζ m , the gaugino can still be gauged away, but the explicit form of the ε
that achieves this is different. However, if, as we will assume from now on, the scalar
field fluctuations and gradients are much smaller than the supersymmetry breaking
scale, Msusy = [Gm G m |S|2 /MP2 ]1/4 , i.e., if
   ∗ m 
 Δ(S ∗ G m )   
  %  S G   ∼ M 2 , (7.72)
 M   M  susy
P P
 ∗ m 
   
∂μ φ m  %  S G   ∼ M 2 , (7.73)
 M  susy
P

then δ  ζL is only a small correction to the leading term,

√
δζL = − 3S ∗ εL + δ  ζL . (7.74)
   
O(Msusy
2 ) %Msusy
2
146 7 Phenomenological Aspects

The ε necessary to gauge away ζ can then be expanded in powers of the small
2
quantities ∂μ φ m /Msusy and Δ(G m S ∗ )/(MP Msusy
2 ) with the leading term given by
√ ∗
the value ζL /( 3S ) for constant scalar fields, i.e.,

ζL
εL = √ [1 + x], (7.75)
3S ∗ 

where x are small corrections of order ∂μ φ m /Msusy2 and Δ(G m S ∗ )/(MP Msusy2 )
that could be determined explicitly in an iterative procedure but won’t be needed
here. Using this ε, we can then again eliminate the goldstino from the original
supergravity action by expressing it in terms of the transformed fields eμa = eμa +δeμa ,
ψμ = ψμ + δψμ , etc. As the original action is supersymmetric, the action with the
transformed fields takes the same form as the original action, but now with ζ ≡ 0
everywhere. In particular, we now have the Noether coupling

1  μ μ 
− J L |ζL =0 ψμL + J R |ζR =0 ψμR , (7.76)
MP

where, as indicated, we have to set ζ = 0 in the supercurrents. This coupling

describes the main coupling of the physical gravitino to the remaining matter
fields φ m and χLm⊥ . These couplings have the schematic form M1P ∂ φ m χLm⊥ ψμ and
 ∗ 
1 m⊥ S Gm
MP Rχ Δ MP ψμ . For large enough momenta, the first term usually dominates
and gives a contact interaction of strength E/MP , which is highly suppressed at
accessible energies E % Msusy % MP .
For energies much higher than the gravitino mass, i.e., in the regime M3/2 %
E % Msusy , however, an enhanced gravitino coupling is often used based on
the equivalence theorem, which states that the matter couplings of the spin-1/2
polarization states of the gravitino are effectively given by the corresponding
goldstino couplings.
To understand what is meant here, we start with the transformed gravitino field
2
that has eaten the goldstino (ignoring small corrections of order ∂μ φ m /Msusy or
m 2
Δ(SG )/(MP Msusy )),

MP 1 2 1
ψμL ∼
= ψμL + √ Dμ ζL + √ γμ ζR = ψμL + √ Dμ ζL + √ γμ ζR ,
3S ∗  3 3M3/2 3
(7.77)

which, for energies E & M3/2 , is dominated by the derivative term. The dominant
coupling of the gravitino to the supercurrent then takes the schematic form

1 1
∂ φ m χLm⊥ Dμ ζL ∼ 2 ∂ φ m χLm⊥ Dμ ζL , (7.78)
MP M3/2 Msusy
7.2 Gravitino, Goldstino, and Super-Higgs Mechanism 147

which, for E & M3/2 , scales as

E2 E2 E
2
∼ & , (7.79)
Msusy M M
P 3/2 MP

i.e., it is enhanced by a factor E/M3/2 & 1 relative to the naive coupling. This
would be particularly relevant for scenarios with small gravitino mass such as, e.g.,
scenarios with gauge-mediated supersymmetry breaking (see the next section).
It should be noted, however, that there is a subtlety with the above coupling when
one takes the rigid limit MP → ∞, M3/2 → 0 with Msusy fixed. In this limit,
the above coupling between the derivative of the goldstino and the supercurrent
survives, and one would conclude that in global supersymmetry, there is a goldstino
coupling of the form

1 μ
J L |ζL =0 ∂μ ζL (7.80)
M2 susy

1
which, in particular, contains couplings of the schematic form (∂ φ m χLm⊥ )∂μ ζL .
MP2
In the original Lagrangian (5.50), however, there is no analogous coupling with two
spacetime derivatives and two fermions.5 The resolution of this is that the above
coupling still refers to the transformed matter fields φ m and χLm⊥ , which are related
to the original fields, φ m and χLm⊥ , by a local supersymmetry transformation with
supersymmetry parameter (7.75). While this is a symmetry in supergravity, it is not a
symmetry in global supersymmetry so that the form of the action does change when
one switches from the original fields φ m , χLm⊥ to the transformed fields φ m , χLm⊥ .
In fact, we know precisely how the action changes from our discussion in Chap. 2,
where we found that a local supersymmetry transformation in a globally supersym-
μ
metric theory leads to an uncancelled variation of the form J L ∂μ εL , which for an
εL as in (7.75) precisely reproduces (7.80). Thus, the above goldstino coupling to
the supercurrent only occurs in global supersymmetry upon a field redefinition to
fields that are adapted to the vacuum of the corresponding supergravity Lagrangian
but that are not particularly natural in global supersymmetry.

7.2.4 Generalizations

When both D- and F-terms are present, the combination ζ giving the goldstino
can be found in an analogous way as above by examining the supersymmetry
transformation laws in the vacuum. A simpler way is to read it off directly from

5In fact, it is easy to construct globally supersymmetric theories in which the goldstino completely
decouples from the χ m⊥ .
148 7 Phenomenological Aspects

the mixing terms with the gravitino in the Lagrangian (6.48):

− K
2M 2
Dm W m e P MP i I
ζL = MP χL + λ PI . (7.81)
W W 2 L
Again, one finds that after an analogous redefinition of the gravitino, this field
disappears from the action. The gravitino now also has Noether couplings to the
gauge fields and the gaugini, and similar remarks apply to these couplings as in the
case of chiral multiplets.
As mentioned earlier, we will not discuss in detail this case and the ones with a
non-trivial cosmological constant, which however can be found in [5].

7.3 Mass Sum Rules and Mediation Mechanisms

In this section, we discuss how the soft supersymmetry breaking terms in supersym-
metric extensions of the Standard Model might be generated and how the scale of the
soft masses, Msoft , would be related to the fundamental supersymmetry breaking
scale, Msusy , and the gravitino mass, M3/2 , in some popular scenarios.

7.3.1 Mass Sum Rules, Hidden Sectors, and Mediation Mechanisms

In our discussion of spontaneous supersymmetry breaking in the global case in

Sect. 5.4, we mentioned a phenomenological problem with the corresponding mass
sum rule StrM 2 = 0 that is valid for renormalizable couplings at tree level. If
we restrict to canonical Kähler potentials, constant gauge kinetic functions, and no
gauging (i.e., no D-terms), the corresponding tree-level expression in supergravity
is [6]

StrM 2 = (−1)2J (2J + 1)MJ2 = 2(nc − 1)M3/2
2
, (7.82)
J

where nc is the number of chiral multiplets, and we considered pure F-term breaking
on a Minkowski vacuum. This modified mass sum rule indicates that soft masses
generically get a supergravity contribution of order M3/2 that is absent in the
globally supersymmetric case. On the other hand, this contribution may alleviate
the mass splitting problem encountered in global supersymmetry provided M3/2
turns out large enough. In the following, we discuss some of the most popular
supersymmetry breaking scenarios with special attention to the relation between
the soft masses, Msoft , and the gravitino mass, M3/2 , resp. the supersymmetry
breaking scale Msusy in those scenarios. As explained in Sect. 5.4, we will assume
that supersymmetry is broken in a hidden sector without renormalizable couplings
at tree level to the Standard Model sector and mostly restrict ourselves to the case
of soft masses near the TeV scale.
7.3 Mass Sum Rules and Mediation Mechanisms 149

Gravity Mediation
The non-renormalizable tree-level interactions present in supergravity Lagrangians
generically transmit effects of supersymmetry breaking from the hidden sector to
the visible fields, even if these have no renormalizable couplings to one another.
When this is the dominant source for the soft masses in the visible sector, one calls
this gravity-mediated supersymmetry breaking [7].6 Due to the special role played
by supergravity in this scenario, we will discuss it in more detail in Sect. 7.3.2. In
the simplest version, the soft masses are of order Msoft ∼ Msusy
2 /MP and hence
of the order the gravitino mass, as may also be estimated from the mass sum rule
(7.82). For soft masses around the TeV scale, the gravitino mass would then also
be in this regime, whereas the fundamental supersymmetry breaking scale Msusy
would be around 1011 GeV, depending on the details.

Anomaly Mediation
For special forms of the Kähler and superpotential, the abovementioned gravity
mediated tree-level contribution to the soft masses may be suppressed. This
happens, for example, in the no-scale models to be discussed below or in so-
called sequestered models. A set of loop corrections that can also be related
to certain quantum anomalies can then give the dominant contribution to soft
g2
masses and generate gaugino and sfermion masses of order 16π 2 M3/2 , where g
represents a Standard Model gauge coupling. This mechanism is generally referred
to as anomaly-mediated supersymmetry breaking [8]. Due to the loop suppression
factors, the gravitino mass in pure anomaly mediation scenarios is one or two orders
of magnitude larger than the soft masses in the Standard Model sector. For TeV
scale soft masses, this would imply M3/2 ∼ O(102 TeV), corresponding to a SUSY
breaking scale around 1012 or 1013 GeV.

Gauge Mediation
Another important scenario in which the tree-level terms from gravity mediation
(as well as from anomaly mediation) are subdominant (for a different reason than
in anomaly mediation, see below) is gauge-mediated supersymmetry breaking [9].
In gauge mediation, the dominant soft terms in the Standard Model sector are
generated by loop diagrams that involve messenger fields, a set of new fields charged
under the Standard Model gauge group and with renormalizable couplings to the
hidden sector where the fundamental supersymmetry breaking takes place. The soft
g2
masses are typically of order 16π 2 Mmess , where g again represents a Standard
Model gauge coupling, and Mmess denotes the mass scale of the messenger
particles.7 If Msusy and Mmess turn out to be of the same order, TeV scale soft

6 This name is a bit misleading, as it is not exactly gravity that induces the soft terms but other
Planck-suppressed contact interactions that are (in part) related to gravity by supersymmetry; see
Sect. 7.3.2.
7 The gaugino masses arise at one-loop, whereas the squares of the sfermion masses are due to a

two-loop correction.
150 7 Phenomenological Aspects

masses would be generated for Mmess ∼ Msusy ∼ O(102 TeV), which would
correspond to a very small gravitino mass in the eV regime. Making Msusy larger
while keeping the soft masses fixed, one can reach larger gravitino masses in the
keV or even MeV regime at the expense of introducing larger and larger hierarchies
between Msusy and Mmess . Thus, while in scenarios dominated by gravity or
anomaly mediation the gravitino mass is comparable to the soft masses or slightly
larger, it is generically much smaller than the soft masses for gauge mediation. This
is self-consistent in the sense that it is only because of this small gravitino mass that
one can neglect the effects of gravity and anomaly mediation (which are generically
present) relative to the gauge mediation effects.

The above describes the most commonly discussed scenarios of supersymmetry

breaking in their simplest form. If supersymmetry is indeed realized in nature, it
might of course be broken in a way that cannot be embedded unambiguously in one
of the above frameworks. It could also be that supersymmetry breaking works in a
way that makes use of an effective field theory that seems non-generic from a low-
energy point of view but may look more natural from a top-down approach such as
string theory. Likewise the simple picture of having all soft masses in the TeV region
may be wrong as, e.g., in the various incarnations of split supersymmetry, or simply
because supersymmetry could turn out to be irrelevant for low energy physics.
However, one may hope that if signs of supersymmetry are to be found in the
future and a particular susy breaking scheme emerges as favored, we can learn
something about the physics at even higher energy scales, possibly way outside
the reach of direct accelerator experiments.

7.3.2 Gravity-Mediated Supersymmetry Breaking and the Polonyi

Model

7.3.2.1 The Polonyi Model

The simplest model for generating a hidden sector breaking supersymmetry at
vanishing cosmological constant was suggested by Polonyi in an unpublished
preprint [10]. This model has a canonical Kähler potential for a single scalar field

Kh = φφ ∗ (7.83)

and a superpotential with a constant contribution and an arbitrary mass scale fixed
by m and α ∈ R,

Wh = mMP (φ + αMP ) . (7.84)

7.3 Mass Sum Rules and Mediation Mechanisms 151

In terms of the dimensionless field, Z := MφP = v+iw, the resulting scalar potential
is
∗
 
V = m2 MP2 eZZ |1 + Z ∗ (Z + α)|2 − 3|Z + α|2 . (7.85)

The critical point condition ∂Z V |V =0 = 0 admits the solution w = 0, to which we

restrict ourselves in the following. The critical point condition ∂Z V |V =0 = 0 then
becomes

(2v + α)(1 + v(v + α)) − 3(v + α) = 0, (7.86)

whereas the Minkowski condition, V = 0, implies

√
1 + v(v + α) = ± 3(v + α). (7.87)
√
Inserting (7.87) into (7.86) implies (v+α) =√± 3−v, which
√ can be re-inserted
√ into
to give the four solutions α = −2 ∓ 3, v = 1 ± 3, and α = 2 ± 3, v =
(7.87) √
−1 ± 3, out of which, however, only two are local minima,
√ √
α = −2 + 3, φ = MP (1 −
3),
√ √ (7.88)
α = 2 − 3, φ = MP (−1 + 3),

The scale of supersymmetry breaking in this model then turns out to be

√ √
2
Msusy = 3mMP e2− 3 , (7.89)

and the mass of the gravitino is

2 Wh √
M3/2 = eKh /(2MP ) = e 2− 3
m. (7.90)
MP2

Gravity mediation would then lead to soft masses in the TeV range if m is chosen
to be in the TeV range. This simple analysis shows two unpleasant features of
this model: the parameter α has to be extremely fine-tuned, and also m needs to
be chosen appropriately to obtain interesting phenomenology. We can, however,
consider this setup as a simple toy model for a hidden sector that can serve as the
basis to illustrate the effects of gravity mediation.

7.3.2.2 Illustration of Gravity Mediation: Scalar Soft Terms

The idea of gravity mediation is that the effects of supersymmetry breaking in
the hidden sector are transmitted to the visible sector by the MP -suppressed
extra interactions that distinguish supergravity from globally supersymmetric field
theories. There are different ways to realize such a setup depending on the details of
152 7 Phenomenological Aspects

the hidden sector dynamics and how exactly the hidden and visible sectors appear in
the Kähler and superpotential. We discuss here only one of the simplest realizations
in which the hidden sector is modeled by a single chiral multiplet with complex
scalar φ that develops a Planck scale vev and breaks supersymmetry spontaneously.
For the sake of concreteness, we realize this with the above Polonyi model,

Kh = φφ ∗ , Wh = mMP (φ + αMP ), (7.91)

but the general conclusions do not depend on this choice. More crucial is the
coupling of this model to the observable sector (the MSSM, for instance), which
is done through a separable Kähler potential and superpotential

K = |φ|2 + |y i |2 , (7.92)
i

W = Wh (φ) + Wv (y i ). (7.93)

Here, y i denotes the scalars from the visible sector. Note that in global supersym-
metry, such a separable K and W would imply a complete decoupling of the hidden
sector and the visible sector. Hence, no matter how badly supersymmetry is broken
in the hidden sector, the visible sector would not feel this in global supersymmetry
in this model. Let us now investigate how this changes in supergravity.
We first write down the F-term potential,
⎡ 2
 |y i |2 |φ|2  
∗
i M2 + M2
V =e P P ⎣Dφ Wh + φ Wv 
 MP2 

 2 ⎤
  ∂Wv   2
y i∗ y i∗   W + W 
 i + 2 + 2 Wh  − 3 
h v ⎦
+  .
 ∂y M P M P
 M P
i
(7.94)

∂Wv
We assume that the vevs of Wv and ∂y i
are negligible compared to the vevs of,
∂Wh
respectively, Wh and so that the dynamics in the hidden sector is essentially
∂φ
unaffected by the presence of the visible sector and yields essentially the same
vevs in that sector as without the visible sector.8 We thus still have to a very good

8 This is not a strong assumption for the MSSM, because if SU (3)c and U (1)em as well as R-parity
are to be unbroken, only the neutral Higgs fields can get a vev in the visible sector, and this vev is
around the electroweak scale. The only term in Wv that can then get a vev is the quadratic term in
the Higgs fields, the μ-term, which will then also be near the electroweak scale. The vev of Wh in
the Polonyi model, by contrast, is of order MP2 M3/2 , which is many orders of magnitude higher.
Similar remarks apply to the derivatives of the superpotentials.
7.3 Mass Sum Rules and Mediation Mechanisms 153

approximation
' |φ|2
( ' |φ|2 (
2 2
√ √ 2 |Wh |
e 2MP
|Dφ Wh | = Msusy = 3MP M3/2 = 3 e 2MP
. (7.95)
MP

The mass of φ is of order m, i.e., of the same order as the gravitino mass, M3/2 ,
which, as we will see in a moment, also sets the scale of the gravity-mediated soft
terms in this model. We can thus not integrate out φ while keeping the other fields
based on simple energy considerations. However, we can split φ = φ + Δφ and
use the fact that the vev of φ is of order MP , so that MP -suppressed terms involving
φ may have significant contributions to the dynamics of the visible sector due to
the contribution from φ, whereas the contributions from the fluctuation Δφ will
just give rise to MP -suppressed interactions with the visible sector fields. We thus
ignore these interactions and replace φ, and consequently Dφ Wh and Wh , by their
vevs using the above relations to obtain
⎡ 2
|y i |2 √

φ ∗  ! 
V = ei ⎣ 3MP M3/2 + Wv 
MP2  MP2 

 2 ⎤
 2
  ∂ W
!v i∗
y ! 
  ! 
Wv  ⎦
+  i + 2W 
v + y M3/2  − 3 MP M3/2 +
i∗
 ∂y M  M  ,
i P P

(7.96)

where
' |φ|2
(
!v := e
W 2M 2
P Wv (7.97)

is an appropriately rescaled superpotential for the visible sector. Taking now the
rigid limit MP → ∞ with M3/2 and φ/MP fixed, we obtain
 
  ∂W
!v 2  i 2 2
V =  
 ∂y i  + |y | M3/2
i i
% √  &
3φ  ∂W
!v i
+ M3/2 −3 W!v + y + h.c. . (7.98)
MP ∂y i
i

The first term is the usual F-term potential for the fields y i in global supersymmetry.
The remaining terms are soft terms, with the second term being a mass term for
the sfermions and the terms in brackets corresponding to bilinear and trilinear soft
terms if W!v contains only quadratic and cubic terms as in the MSSM. We note that
154 7 Phenomenological Aspects

the universal sfermion masses, Msfermion = M3/2 , are a consequence of the simple
Kähler potential (7.92). Non-universal sfermion masses may, e.g., be obtained from
more general Kähler potentials such as

K = |φ|2 + hi (φ, φ ∗ )|y i |2 . (7.99)
i

7.3.2.3 Gaugino Masses

Gaugino masses can also be obtained if the gauge kinetic function, fI J , has a non-
trivial dependence on the hidden sector fields, because the general supergravity
action contains the term
1 K/2M 2
P Dm Wg mn (∂n̄ f ∗ )λ λJ + h.c.
I
e IJ R R (7.100)
4

If ∂n fI∗J ) is of order 1/MP , this term would yield a gaugino mass, M1/2 , of order
M3/2 , so that all soft masses would be comparable to the gravitino mass.
Note, however, that the above coupling (7.100) is already present in the most
general globally supersymmetric theory (5.120). Thus the gaugino masses are in this
sense not really gravity mediated, and the mass scale of ∂n fI∗J need not necessarily
be related to MP but could a priori be related to any mass scale μ < MP so that
∂n fI∗J  ∼ 1/μ. A simple example would, e.g., be f (φ) = 1 + μ1 (φ − φ), which
satisfies f (φ) = 1, but ∂φ f  = 1/μ, which would be much bigger than 1/MP
if μ is much smaller than MP . In at least two important cases, however, a different
mass scale μ % MP in f does not enter the physical gaugino masses:
φ φ
(i) f (φ) = = μ +μ (φ
1
− φ);
μ 
φ
(ii) f (φ) = 1 + g 2 log μ .

In case (i), ∂φ f  = 1/μ naively seems to result in a gaugino mass of the form
M3/2 MP /μ & M3/2 , but this is not correct, because f  = φ/μ & 1 so that
the kinetic term for the gaugini and the gauge fields would not be canonically
normalized. Expressing
√ the mass term in terms of the canonically normalized
gaugini λ̃ = MP /μλ then results again in a mass term of order M3/2 if φ is
of order MP .
In case (ii), one again splits off the vev of f ,
   
φ φ
f = 1 + g 2 log +g 2 log (7.101)
μ φ
  
=:X
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 155

√
and switches to canonically normalized fields λ̃ = Xλ so that the new gauge
kinetic function becomes9
 
˜ g2 φ
f =1+ log . (7.102)
X φ

If g is a gauge coupling, the canonical normalization of the gauge fields also rescales
g 2 to g̃ 2 = g 2 /X so that

∂ f˜ g2 1 1
= = g̃ 2 (7.103)
∂φ Xφ φ

which results in gaugino masses of order g̃ 2 M3/2 if φ is of order MP . Thus, up to

a possible slight suppression due to the factor g̃ 2 , one again obtains gaugino masses
of the same order as for the sfermions.
It should be noted, however, that there is a difference between the gaugino masses
and the sfermion masses also in the above two cases. Namely, in the case of the
sfermion masses, the mass due to gravity mediation does not rely on having φ
of order MP . In the derivation of the sfermion masses, we only used that the vevs
of Wh and Dφ Wh are much larger than the vevs of Wv and ∂y i Wv , which would
be the case for vevs of φ that are large compared to the electroweak scale, but not
necessarily of order MP . The gaugino masses in the above two examples, on the
other hand, are of order M3/2 only for φ ∼ MP .

7.4 Moduli Stabilization, de Sitter Vacua, and Inflation

In this section, we derive some model-independent constraints on masses of scalar,

in particular modulus-like, fields and discuss some issues associated with the
cosmologically interesting case of positive potential energy.

7.4.1 Moduli Stabilization and Moduli Masses

An interesting possible ingredient of supersymmetric extensions of the Standard

Model is moduli or modulus-like fields. These are scalar fields that are uncharged
with respect to the gauge interactions and couple to ordinary matter only very
weakly, typically with interactions suppressed by a very high mass scale such as
MP . Moduli can be useful for realizing spontaneous supersymmetry breaking, and
they are a natural ingredient of phenomenologically realistic string compactifica-

9The canonical normalization factor X is not parametrically large due to the logarithm, so unlike in
case (i), the canonical normalization does not change the result by at most an order of magnitude.
156 7 Phenomenological Aspects

tions, where they correspond to deformations of the compactification background

that cost little energy.10
A phenomenologically very important parameter is the mass of a modulus field,
because if the moduli are too light, they can give rise to various phenomenological
problems. Depending on the exact mass range, these may include, e.g., conflicts with
the successful predictions of Big Bang Nucleosynthesis (BBN), problems with the
measured energy density of the Universe or tensions with fifth force experiments.
The simplest way to avoid all these problems is to ensure that the moduli have
masses above about 30 TeV.11 The implementation of mechanisms that generate
such masses is generally referred to as moduli stabilization and forms an important
part of modern string phenomenology.
In many models, a natural mass scale of the moduli is the gravitino mass,
M3/2 , but moduli may also have masses considerably larger or smaller than M3/2 ,
depending on the details of the model. Moreover, some moduli may develop a
negative mass squared, which means that they are tachyonic directions of the scalar
potential that indicate an instability of that particular vacuum.

7.4.1.1 The sgoldstini

In this subsection,12 we restrict ourselves mainly to a few rather general statements
one can make based on the general form of the scalar potential. A crucial role in this
discussion is played by the sgoldstini, the two scalar superpartners of the goldstino.
Unlike the goldstino itself, the sgoldstini are not eaten by any other field and hence
are part of the physical scalar spectrum.
To define the sgoldstini, it is convenient to shift all scalar fields, φ m , by their vevs
such that we have φ m  = 0. In a given vacuum, the sgoldstini are then the two real
components of the complex scalar field

Gm 
Σ := √ φm. (7.104)
Gn G n 

10 In the original sense, moduli denote massless scalar fields, usually related to constant energy

deformations of the internal space in string compactifications or to exactly marginal deformations

of conformal quantum field theories. As described above, in phenomenologically realistic string
compactifications, such scalar fields should have a certain mass, i.e., the corresponding deforma-
tions of the internal space should cost some energy (e.g., due to the presence of p-form fluxes in the
internal space or as a consequence of non-perturbative quantum corrections). Nevertheless, these
massive scalar fields are still called moduli.
11 This bound is an approximate estimate based on the following assumptions [11]: (1) Moduli

couple with MP -suppressed interactions to other matter with a resulting decay width Γ ∼
O(MMod3 /MP2 ); (2) the Hubble scale after inflation is larger than MMod , and when it drops
to H ≈ MMod , the modulus starts oscillating around its potential minimum; (3) the moduli so
produced decay before Big Bang Nucleosynthesis (BBN).
12 This subsection is based on [12].
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 157

With respect to the scalar field metric in the vacuum, gmn , its orthogonal
complement is

G m 
φ m⊥ := φ m − - Σ, (7.105)
Gl G l 

which satisfies

Gm φ m⊥ = 0 (7.106)

so that indeed

− gmn (∂μ φ m )(∂ μ φ n ) = −gmn (∂μ φ m⊥ )(∂ μ φ n⊥ ) − (∂μ Σ)(∂ μ Σ ∗ ), (7.107)

and Σ and φ m⊥ are the superpartners of, respectively, the goldstino, ζL , and its
orthogonal complement, χLm⊥ , in the given vacuum,

δΣ = εL ζL , δφ m⊥ = εL χLm⊥ . (7.108)

Several interesting conclusions can be drawn from the behavior of the scalar
potential in the complex plane given by Σ. In order to probe the scalar potential
along arbitrary real straight lines through the origin of this plane, we introduce the
rotated field

Z := eiθ Σ, (7.109)

where eiθ is a constant phase factor, and split it into canonically normalized real and
imaginary parts, X and Y ,

1
Z = √ (X + iY ). (7.110)
2

Studying V along X for all possible θ is then equivalent to studying V along all
possible straight real lines through the origin of the Σ-plane.
If all other fields are frozen, the canonically normalized mass of X would be

1
m2X = ∂X ∂X V = (VZZ + 2VZZ ∗ + VZ ∗ Z ∗ )
2
1 −2iθ
= (e VΣΣ + 2VΣΣ ∗ + e2iθ VΣ ∗ Σ ∗ )
2
= VΣΣ ∗ + Re[e−2iθ VΣΣ ], (7.111)

where VZZ ≡ ∂Z ∂Z V , etc., and all quantities are meant to be evaluated at the critical
point.
158 7 Phenomenological Aspects

Setting now VΣΣ = Aeiα with A := |VΣΣ | and α being a possible phase, we
can write

m2X = VΣΣ ∗ + A cos(α − 2θ ). (7.112)

Thus, for the particular field direction given by θ = α/2 − π/4, the angular term
can be eliminated, and we obtain

1
m2X = VΣΣ ∗ = G l Vlk G k , (7.113)
Gn G n

where again all quantities are meant to be constant vevs.

In Appendix 7.A at the end of this chapter, we show that, at a critical point of the
potential, (7.113) can be written as [12]
⎡ ⎤
Rklpm G k G l G p G m
m2X = MP2 eG /MP ⎣2 − ⎦.
2
(7.114)
Gn G n

A number of interesting conclusions regarding moduli mass scales and the stability
of the vacuum can be drawn from this equation.

7.4.1.2 A Constraint on the Lightest Modulus Mass

In this subsection, we discuss how, under certain assumptions, Eq. (7.114) provides
a bound on the mass of the lightest modulus-like field [12]. To this end, we use
= MP2 eG /MP and neglect the small cosmological constant in the present
2 2
M3/2
Universe, setting Gn G n = 3MP2 . Equation (7.114) then becomes

MX2 = ∂X ∂X V = M3/2
2
[2 − r], (7.115)

where
1
r := R G k G l G nG m. (7.116)
3MP2 klnm

Now, MX2 is not necessarily an eigenvalue of the mass matrix, because the Hessian
of V in general contains non-trivial off-diagonal terms. However, MX2 = ∂X ∂X V
is a diagonal element of the mass matrix, and as the smallest eigenvalue of a
diagonalizable real matrix cannot be bigger than any of its diagonal elements, we
can conclude that the smallest scalar mass eigenvalue, Mmin 2 , is bounded from
2
above by MX ,

2
Mmin ≤ MX2 = M3/2
2
[2 − r]. (7.117)
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 159

Thus, if the curvature of the scalar manifold along the sgoldstino directions is at
most of order MP−2 , such that |r| = O(1) or smaller, Eq. (7.117) implies that the
lightest modulus-like field cannot be parametrically heavier than the gravitino.
Thus, if one wants the moduli masses to be above 30 TeV, the gravitino mass
could then likewise not be parametrically smaller than this mass scale. This in
turn would be a challenge for scenarios with small gravitino masses such as gauge
mediation, and it might be part of an explanation for why the superpartners of the
Standard Model particles might be somewhat heavier than what simple naturalness
arguments suggest (see, e.g., [13]).

7.4.1.3 Possible Caveats

While it would be a fascinating possibility, there are also a number of caveats
to the above argument. More precisely, there are two types of possible caveats.
The first type concerns all the assumptions that underlie the phenomenological
30 TeV estimate for the lightest modulus mass (see footnote 11). If any of these
assumptions is too strong, the moduli masses might be smaller without conflicting
observations, which would then also relax the above-described constraints on the
gravitino mass from Eq. (7.117). The second type of caveats concerns the conclusion
that the lightest modulus cannot be parametrically heavier than the gravitino due to
Eq. (7.117). We focus here only on this latter type of caveats, which include the
following:

• The smallest eigenvalue of the scalar mass matrix need not necessarily be
a modulus-like field but could in principle be another type of scalar, e.g., a
charged scalar field. In that case, the bound (7.117) would not constrain the
lightest modulus-like field but this other type of scalar field. However, if there
is no significant mixing between the moduli and these other scalar fields and if
supersymmetry is predominantly broken in the moduli sector, one can, to a very
good approximation, repeat the above analysis in the moduli sector alone and
would then again conclude that the lightest modulus is indeed constrained by an
equation of the form (7.117).
• We did not consider D-terms in our analysis. If they play a significant role in
supersymmetry breaking and/or moduli stabilization, the above analysis would
have to be re-done with D-terms.
• If the scalars, φ m , are not canonically normalized with respect to the vev of the
scalar field metric in the kinetic term, i.e., if gmn  = δmn , the eigenvalues of
the Hessian of V do not give the physical masses of the canonically normalized
fields. The transition to canonically normalized scalars would instead rescale
the physical mass eigenvalues. In the case at hand, however, this would not
change (7.117) and the conclusions drawn from it. The reason is that the real
sgoldstino fields X and Y are already canonically normalized and orthogonal
to the other scalars φ m⊥ . Thus a redefinition of the φ m⊥ to make them also
canonically normalized would not change the diagonal element ∂X ∂X V of the
mass matrix but at most the terms involving at least one derivative with respect
to φ m⊥ . Thus, after canonical normalization, the smallest eigenvalue of the new
160 7 Phenomenological Aspects

mass matrix would still be bounded by the unchanged diagonal element (7.117)
with r being expressed in terms of the new canonically normalized fields. But as
r is a scalar under scalar field redefinitions, it does not change its value under
these redefinitions, and hence also the canonically normalized fields satisfy the
bound (7.117).
• If r is negative and |r| is extremely large, i.e., if −r & 1, the lightest modulus
may be much heavier than the gravitino, and the above bound would allow for
much smaller gravitino masses. A special case of this would be the limit r →
∞, which would typically correspond to a particular realization of non-linear
supersymmetry, which is thus also not directly covered by the argument above.

7.4.1.4 A Counterexample with Large Curvature

Let us briefly illustrate the last item in the above list with a simple toy model [14]
given by a single chiral multiplet with scalar φ and

φ2φ∗2
K = φφ ∗ − , W = μ2 (φ + cMP ), (7.118)
Λ2
where μ and Λ are mass parameters with Λ % MP , and c is a numerical constant.
For suitably fine-tuned c, the resulting F-term potential has a non-supersymmetric
Minkowski minimum with a very small vev of φ. This may be seen by a perturbative
expansion in the small parameter Λ/MP % 1, which, after some algebra, leads to
  
1 Λ2
c = √ 1+O , (7.119)
3 MP2
   
∗ 1 Λ2 Λ2
φ = φ  = MP √ 1+O % MP . (7.120)
2 3 MP2 MP2

The curvature parameter r at this vacuum is

  
M2 Λ2
r = −12 P2 1+O % −1 (7.121)
Λ MP2

so that the bound (7.117) allows for moduli masses parametrically larger than M3/2 .
And indeed, one finds that, to lowest order in the small parameter Λ2 /MP2 , the two
2 = μ4 /3M 2 ,
eigenvalues of the scalar mass matrix are, using M3/2 P
     
2
4μ4 Λ2 2 MP Λ2
M12 = M22 = 2 1+O = 12M3/2 1+O & M3/2
2
.
Λ MP2 2 Λ MP2
(7.122)
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 161

In the limit Λ → 0, the curvature parameter r and with it the scalar masses diverge,
and the low-energy theory becomes effectively a specific version of a local non-
linearly realized supersymmetry with the metric and a massive gravitino as the only
light fields, which one may also describe with the help of nilpotent superfields [14,
15].
Interestingly, even though some of the above caveats might in principle apply,
many models derived from string theory do indeed have a lightest modulus that is
not parametrically heavier than the gravitino [12].

7.4.2 No Scale Models

In the counterexample above, we saw that the sgoldstini can in principle be

parametrically heavier than the gravitino when the curvature in the sgoldstino plane
takes on extreme (negative) values. There is, however, also the opposite possibility
that an exact or approximate cancellation of the two terms in (7.117) could make the
sgoldstini and hence the lightest modulus much lighter than the gravitino, or even
massless. An extreme example for this are no-scale models, where scalars that do
not appear in W are exactly massless at tree level, while the gravitino mass can take
on arbitrarily high values. At the same time, the tree-level cosmological constant
remains exactly zero, no matter how high the scale of supersymmetry breaking
is [16]. Subleading corrections in general lift this vacuum degeneracy and could
generate a hierarchically small supersymmetry breaking scale [17]. Unfortunately,
the smallness of the cosmological constant relative to the supersymmetry breaking
scale in general does not survive these corrections. Nevertheless, no-scale models
play an important role as a first step in model building due to their natural emergence
in string theory compactifications.

7.4.2.1 The Simplest Example

The simplest version of a no-scale model is based on one chiral multiplet with
dimensionless complex scalar, T , and

K = −3MP2 log(T + T ∗ ), W = αMP3 (7.123)

with α ∈ R. In that case, one has ∂T W = 0 and

3MP2 3MP2 ∗
∂T K = − , gT T ∗ = ⇒ K T K T gT T ∗ = 3MP2
T +T∗ (T + T ∗ )2
(7.124)

so that the F-term contribution from T to the scalar potential satisfies

∗ |W |2
g T T DT W DT ∗ W ∗ = 3 , (7.125)
MP2
162 7 Phenomenological Aspects

which exactly cancels the negative term coming from the gravitino contribution.
As a result, the potential is completely flat, V ≡ 0, so that we have an infinite
set of degenerate Minkowski vacua parameterized by the vev of an unfixed massless
modulus T . The supersymmetry breaking scale and the gravitino mass, however, are
not zero but depend on the vev of T ,
√
3αMP2 αMP
MSUSY =
2
, M3/2 = . (7.126)
(T + T ∗ )3/2  (T + T ∗ )3/2 

7.4.2.2 Generalizations
The crucial property of the above example that leads to the cancellation of the
negative term in the potential is (apart from ∂T W = 0) the geometric property

KT K T = 3MP2 . (7.127)

A straightforward generalization of this for n complex dimensionless scalars t α

(α, β, . . . = 1, . . . , n) is obtained when K and W satisfy

K α Kα = 3MP2 , ∂α W = 0, (7.128)

where, as usual, Kα ≡ ∂α K, etc. This would then again lead to V ≡ 0.

For a concrete realization of such Kähler potentials, we split the n scalars t α
into n1 complex scalars, T A (A, B, . . . = 1, . . . n1 ), and n2 complex scalars, ϕ a
(a, b, . . . = 1, . . . , n2 ), and take the Kähler potential to be of the following form

K = −MP2 log[F (J A )], J A := T A + T A − N A (ϕ a , ϕ a ), (7.129)

where F is a homogeneous function of degree three of the composite variables J A ,

∂F
JA = 3F, (7.130)
∂J A

and N is an arbitrary real function of the fields ϕ a .

To verify the no-scale property,

K α Kα = K A KA + K a Ka = 3MP2 , (7.131)

we first note that

∂K ∂K ∂K MP2 ∂F
KA ≡ = = = − (7.132)
∂T A ∂T A ∂J A F ∂J A

so that, with (7.130),

J A KA = −3MP2 . (7.133)
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 163

With this, we compute

∂K
0 = ∂α (J A KA ) = (∂α J A ) A
+J A KαA = ∂α K + gαA J A , (7.134)
∂T
  
∂K
∂J A

or, raising the index α with g αα ,

K α = −δA
α A
J , (7.135)

so that (7.133) indeed implies Kα K α = 3MP2 .

Frequently encountered examples are
 3  
• K = −MP2 log T + T ∗ − cδab ϕ a ϕ b = −3MP2 log T + T ∗ − cδab ϕ a ϕ b .
For the special case c = 1/3, this is the symmetric space SU (1, 1+n2 )/S(U (1)×
U (1 + n2 )).  
• K = −MP2 log dABC (T A + T A )(T B + T B )(T C + T C ) , where dABC is a real
symmetric tensor.

Although the above form of the Kähler potential looks quite peculiar, it should
be stressed that no-scale models are in fact not uncommon in string theory
compactifications and other types of dimensional reductions. Upon reducing a
theory from D to d < D dimensions, one finds, in particular, that the volume
modulus, and more generally the Kähler moduli, behaves like the T A in the leading
order Kähler potential.

7.4.2.3 Adding Scalars Without No-Scale Property

If one adds scalars, zi (i, j, . . . = 1, . . . , m), to a no-scale model with scalars, t α ,
as above such that

! i , zi ),
K = K̃(t α , t α ) + K(z W = W (zi ) (7.136)

with K̃α K̃ α = 3MP2 , the F-term of the t α still cancels the negative gravitino
contribution in the potential, which then becomes positive semi-definite,
 
K(t, z, t ∗ , z∗ ) ¯
V = exp g i j Di W Dj W ∗ ≥ 0, (7.137)
MP2

where i, j run only over the zi fields, and the only dependence on t α is in the
exponential factor. This implies that the Minkowski vacua with Di W  = 0 are
the global minima of the potential at which the zi are in general stabilized. This
degenerate valley of Minkowski minima is therefore parameterized by the vevs of
164 7 Phenomenological Aspects

the unstabilized t α , because the t α -dependence of the potential drops out in the
valley, where V =0. Just as in the simple example (7.123), for fixed zi , the vevs of
the t α determine the supersymmetry breaking scale and the gravitino mass, which
can take on arbitrary values even though the cosmological constant vanishes at tree
level.
In general, however, the flatness of the potential along the no-scale moduli t α
is lifted by quantum corrections, and the t α cannot be chosen at will but need to
be evaluated at a suitable minimum of the corrected potential (if such a minimum
exists). As they are generated only at subleading order, the masses of the t α may be
hierarchically smaller than the masses of the zi . If that is the case, one can integrate
out the heavier zi and work with an effective theory of the t α alone. A well-known
example for this is the KKLT scenario in type IIB string theory [18], where the zi are
complex structure moduli that are stabilized at tree level by fluxes of antisymmetric
tensor fields in the compact space, and the t α correspond to Kähler moduli with
smaller masses generated by non-perturbative quantum corrections to W .

7.4.2.4 A D-Term Analogue

While the above examples of no-scale models are all based on pure F-term
potentials, there is also a construction involving D-terms that shares some of the
features of the pure F-term models [19]. To this end, we couple N = 1 supergravity
to one chiral multiplet with the following Kähler potential and superpotential

K = −2 MP2 log(φ + φ ∗ ), W = a MP3 , (7.138)

where a is a real constant and φ is dimensionless. This is obviously not of the

no-scale type discussed above and by itself would not lead to a flat potential. To
achieve this, however, we can now also couple one vector multiplet (Aμ , λ) to the
above theory with a gauge kinetic function f (φ) given by

1
f (φ) = (7.139)
g2

where g is a real constant (i.e., the φ dependence is trivial here). Obviously, the
Kähler metric on the scalar manifold spanned by the complex field φ only depends
on the real part of φ so that the field transformation

φ → φ + α ξφ = φ + i α (7.140)

is an isometry on the scalar manifold with symmetry parameter α and Killing vector
ξ φ = i. Gauging this isometry then leads to a D-term potential given by

1
VD = (Ref )−1 PP, (7.141)
2
7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 165

where P ≡ iξ φ ∂φ K is the Killing prepotential of ξ φ . The total potential in the

special case 2g 2 = a 2 = 0 is then again flat, V ≡ 0, and the supersymmetry
breaking scale can take on any value depending on the undetermined vev of φ.
Also in this case, one can imagine adding additional scalars and vector fields
without spoiling the no-scale property of the resulting potential, provided certain
requirements are satisfied. The Kähler potential and superpotentials can be extended
to

K = −2 MP2 log(φ + φ ∗ ) + ΔK(φ, φ ∗ , zi , zi ), W = a MP3 + ΔW (zi ),

(7.142)

and the gauge kinetic functions generalized to

(0) i (1) i
fab = fab (z ) + fab (z )φ, (7.143)

provided ΔK = ΔW  = 0 at the minimum and the supersymmetric conditions

Di W = 0 = Da are satisfied for the zi directions and for the new gauge fields Aaμ .

7.4.3 Dark Energy and de Sitter Vacua

In Sect. 7.4.1, we were primarily concerned with the absolute value of moduli
masses in a Minkowski vacuum and how these masses compared to the mass of the
gravitino based on the bound (7.117). What we did not yet discuss is the possibility
that the sign of some squared moduli masses might actually be negative at a given
critical point. In fact, (7.114) indicates that for scalar manifolds with sufficiently
strong positive curvature in the sgoldstino plane, the curvature term in (7.114) may
overcompensate the positive first term and imply that the lightest modulus must
have a negative mass squared. In other words, the critical point would not be a
local minimum of the potential but would have a tachyonic direction in field space
indicating a perturbative instability of that vacuum.
The danger of the sgoldstino bound becoming negative is generally smaller for
AdS than for Minkowski or even de Sitter vacua. To understand this, we go back to
the original equation (7.114), which is valid for any value and sign of the vacuum
energy, and write it as13

eG /MP  
2

m2min ≤ m2X = ∂X ∂X V = MP2 2G l G l

− R G G G G
k l p m
. (7.144)
Gn G n klpm

13 This subsection is based on [20].

166 7 Phenomenological Aspects

Obviously, a tachyon is implied if the term in brackets becomes negative, i.e., if

the quartic curvature term overcompensates-the first term quadratic in G m . For sub-
Planckian curvatures and small |G m | ≡ Gl G l % MP2 , however, the first term
should always dominate and make MX2 positive so that a negative Mmin 2 is not
necessarily implied.14 Remembering V = MP2 eG /MP [Gm G m − 3MP2 ], however,
2

the case |G m | % MP2 is seen to correspond to negative vacuum energy, i.e., an AdS
vacuum.15 √
For a Minkowski vacuum, by contrast, we have instead |G m | = 3MP so that
a (positive) curvature Rklpm of order 1/MP2 would make the quartic term in (7.144)
comparable to the quadratic term so that a negative MX2 becomes possible. For a
√
de Sitter vacuum, one even has |G m | > 3MP so that a positive curvature can
imply a negative MX2 more easily the larger |G m | and hence Mvac is. In fact, for
some models one can even rule out a dS minimum no matter how small the vacuum
energy is, as we now describe.
To this end, we rewrite the quantity in brackets in (7.144) as

λ := 2Gm G m − Rklpm G k G l G p G m
 
2 2
= − (Gn G − 3)Gl G +
n l
(Gl G ) − Rklpm G G G G
l 2 k l p m
(7.145)
3 3
2! !
=− V (V + 3) + σ,
3
! ≡ (Gn G n − 3) and
where V
 
1
σ := (g gmp + glp gmk ) − Rlkmp G l G k G m G p . (7.146)
3 lk

! > 0, a tachyon would necessarily be present if

Thus, in a dS vacuum, with V

σ ≤ 0. (7.147)

Note that the sign of σ only depends on the orientation of the G m as well as on
purely geometric quantities determined by K, but not on the absolute value |G m |
and hence also not on the magnitude of the (positive) cosmological constant.
We conclude this subsection with a few examples, where, for simplicity, we set
MP = 1 and work with dimensionless fields.

14 This does of course not guarantee that there is really no tachyon, because (7.144) is just an upper
bound.
15 As we have discussed in Chap. 4, stable AdS vacua would even be consistent with tachyonic

scalars as long as they satisfy the Breitenlohner–Freedman bound.

7.4 Moduli Stabilization, de Sitter Vacua, and Inflation 167

• K = φφ ∗ ⇒ Rφφ ∗ φφ ∗ = 0 ⇒ σ = 23 (Gl G l )2 > 0. Hence, the bound

(7.144) does not necessarily imply the existence of a de Sitter tachyon for this
case. In fact, the Polonyi model with appropriately chosen constant term in the
superpotential would give a simple example with a dS minimum.
• K = −3 log(T + T ∗ ) ⇒ σ = 0. Hence, unless K receives corrections, this
no-scale model does not allow for a dS minimum with a pure F-term potential,
no matter how W is chosen (see also [21]).
• K = −3 log(T + T ∗ − 1/3 i |φ i |2 ) ⇒ σ = 0. Hence, in this model, there
are no dS minima possible if one just uses an F-term potential and K does not
receive corrections.

In this section, we focused on the stability of a given de Sitter extremum of an

F-term potential and found that there are some dangers of developing a tachyonic
instability in the sgoldstino plane. A somewhat related problem occurs when one
starts from no-scale potentials with Minkowski vacua and tries to deform them
into a de Sitter vacuum with a small cosmological constant by introducing small
corrections to the no-scale potential. As shown in [22], many of such corrections
may lead to a tachyon at the de Sitter extremum, which is not in the sgoldstino plane
but along a field direction which only aligns with one of the sgoldstino directions in
the Minkowski limit.
Of course tachyons can occur in other field directions as well, and as there is no
analogue of a Breitenlohner–Freedman bound for de Sitter spacetime, any tachyon
would lead to an instability, making this a general issue for de Sitter vacua.
Moreover, many F-term potentials that descend from tree-level dimensional
reductions of higher-dimensional supergravity theories do not even have de Sitter
critical points due to surprisingly simple no-go theorems [23]. These no-go theorems
may be evaded, e.g., by introducing objects of negative energy density such as
orientifold planes in type II string theories and/or by taking into account quantum
or stringy corrections or by allowing for more exotic branes and fluxes.

7.4.4 Inflation and the Supergravity η-Problem

Another interesting cosmological problem that can be addressed in supergravity is

inflation [24]. Inflation denotes a postulated period of accelerated cosmic expansion
in the very early Universe that could solve various cosmological naturalness
problems such as the flatness and homogeneity problem and provides a mechanism
for the generation of density fluctuations. The simplest realization of inflation is via
a scalar field, φ, (“inflaton”) that slowly rolls down a relatively flat potential, V (φ),
with its momentary potential energy driving the accelerated cosmic expansion. The
168 7 Phenomenological Aspects

flatness of the potential is usually expressed in terms of the slow-roll parameters 

and η, which have to satisfy
 2
1 V
 ≡ MP2 % 1, (7.148)
2 V
  
V 
η≡ MP2   % 1. (7.149)
V 

If one embeds this model in a supersymmetric theory, one might hope that
supersymmetry could protect the inflaton direction from getting a large slope or
curvature. However, in supergravity, this is surprisingly difficult to achieve, even at
tree level. In order to see this, we assume that the inflaton potential during slow-roll
inflation is dominated by the F-term potential,
 
∼ 2 |W |2
Vinfl = VF = eK/MP |DW | − 3 2
2
(7.150)
MP

Assuming now a canonical Kähler potential, K = Kcan = φφ ∗ , one finds that the
derivatives of the exponential of K produce a term of the form

1
∂φ ∂φ ∗ Vinfl = Vinfl + . . . (7.151)
MP2

where the ellipsis denotes terms with derivatives acting also on the other terms. This
then implies

η = 1 + ..., (7.152)

i.e., a generic contribution of order one to the eta parameter.16 This would have to be
cancelled by the terms denoted by the ellipsis, which usually requires very special
superpotentials and/or fine-tunings. This is the supergravity η-problem of F-term
inflation17 [25]. Possible solutions are the following:

• W is of a special form such that η % 1 is achieved. An example is given in the

exercises.
• The Kähler potential might have a shift symmetry along, say, the imaginary part
∗ )2
of φ [26]. More precisely, define ϕ ≡ Im(φ) and let K = (φ+φ 2 , which
would still give rise to a canonical kinetic term for the scalars. In this case, the

16 Note that we have been a bit sloppy here with the fact that φ in supergravity is complex and that

in single field inflation the inflaton is real. One thus has to go over to the diagonalized mass matrix
of the real and imaginary part of φ first.
17 The inflaton mass term can alternatively be viewed as due to gravity-mediated spontaneous

supersymmetry breaking during slow-roll inflation.

7.4 Appendix: Proof of Eq.(7.114) 169

exponential of K would not depend on ϕ, and, taking ϕ as the inflaton, no η-

problem would arise at this level. Note that this new Kähler potential is related
to the canonical one by a Kähler transformation. In global supersymmetry, we
can therefore always bring the Kähler potential to this shift symmetric form, but
in supergravity, the Kähler transformation would also act on W , which would
then pick up an exponential of the Kähler transformation, thereby maintaining
the η-problem (it would just be shifted from K to W ).
• Simply accept an O(10−2 ) fine-tuning (eta is constrained only to order 10−2 ).
• Use D-term inflation, i.e., an inflaton potential dominated by D-terms [27, 28].
The problem with standard versions of D-term inflation, however, is that they
usually produce cosmic strings after inflation. Moreover, in models in which
moduli fields have to be stabilized (e.g., in the generic string theory low-energy
effective actions), the stabilization usually works with F-term potentials, which
can re-introduce the η-problem also in D-term inflation models [29].
• Use a potential that is not a supergravity potential or one with non-linearly
realized supersymmetry [30, 31].

7.A Appendix: Proof of Eq. (7.114)

In this Appendix, we show that, at a critical point of an F-term potential in N = 1

supergravity, one has
% &
G /MP2 Rklnm G k G l G n G m
G G Vkl =
k l
MP2 e Gp G p
2− , (7.153)
Gq G q

which then implies Eq. (7.114) upon division by Gp G p .

As a preparation, we first define

! := Gm G m − 3MP2
V (7.154)

so that

!
V = MP2 eG /MP V
2
(7.155)

and rewrite the critical point condition in a number of useful ways. Using (7.155),
we first have
% &
G !+V !l = 0
∂l V = MP2 eG /MP
2 l
V (7.156)
MP2
!
Gl V
!l = −
⇔V , (7.157)
MP2

where Gl ≡ ∂l G , etc. and all quantities are meant to be vevs.

170 7 Phenomenological Aspects

To compute V !l directly from the definition (7.154), we use Dl g mn = 0 and

Dl Gn = gln to obtain

!l = ∂l V
V ! = Dl V
! = (Dl Gm )G m + Gm g mn (Dl Gn ) = G m (Dl Gm ) + Gl (7.158)

Hence, (7.157) becomes

 
!
V
G Dl Gm = −Gl
m
+1 . (7.159)
MP2

We are now ready to prove (7.153) and first compute

(7.156)  G /M 2 ! + MP2 V
!l )

Vkl = ∂k (∂l V ) = ∂k e P (Gl V (7.160)
⎡ ⎤
⎢ ⎥
Gk ⎢
G /MP2 ⎢
⎥
= +e ! ! ! ⎥
∂ V ⎢(∂k Gl ) V + Gl V +MP2 V lk ⎥ (7.161)
MP 
l
⎣   k
2
⎦
=0 gkl (7.157) !/M 2
= −Gl Gk V P
% &
glk Gl Gk
=V − ! .
+ MP2 eG /MP V
2
(7.162)
lk
MP2 MP4

! , one can replace ∂ by D , because the Christoffel symbols with mixed indices
In V lk k k
vanish, so that

! =D V
V ! (7.158)
= (Dk Gn )(Dl Gm )g mn + G m (Dk Dl Gm ) + gkl (7.163)
lk k l

= (Dn Gk )(Dm Gl )g mn + G m (Dk Dl Gm ) + gkl (7.164)

p
where Dl Gm ≡ ∂l Gm − Γlm Gp = Dm Gl and the complex conjugate thereof was
p p
used in the last step. Using now ∂l gkm = Γlm gkp and ∂k Γlm = Rkl p m , one finds
Dk Dl Gm = −Rklpm G p so that

! = (Dn G )(Dm Gl )g mn − R
klpm G G + gkl .
m p
V lk k (7.165)
Exercises 171

Inserting this into (7.162) and contracting with G k G l then gives

% &
Gl G l (Gl G l )2
G l G k Vlk = V 2
−
MP MP4
⎡ ⎤
⎢ ⎥
⎢ ⎥
⎢ l⎥
+MP2 eG /MP
2
⎢( G Dn Gk ) g
k mn
(G Dm Gl ) −Rklpm G G G G + Gl G ⎥ .
l k l p m
⎢       ⎥
⎣   
!
 ⎦
! −Gm V
+1
−Gn V
2 +1 MP2
MP

(7.166)

! + 3M 2 ) in the remaining term, one

Factoring out Gl G l and replacing Gn G n by (V P
!-dependent terms cancel and (7.153) is obtained.
finds, after some algebra, that all V

Exercises

7.1. Compute the masses of the fermions at the non-supersymmetric vacuum of the
model with Kähler potential

K = |φ|2 + |A|2 + |B|2

and superpotential

W = φAB.

Discuss the super-Higgs mechanism at this vacuum.

7.2. Consider an N = 1 supersymmetric theory of three chiral multiplets with

complex scalars S, φ+ , φ− and the following Kähler and superpotentials:

K = |S|2 + |φ+ |2 + |φ− |2 (7.167)

 
W = S κφ+ φ− − μ2 . (7.168)

Here, μ is a mass parameter and κ denotes a dimensionless coupling constant.

(a) Show that in global supersymmetry the resulting classical

√ F-term potential has
positive mass terms for φ+ and φ− if |S| > Sc = μ/ κ.
(b) For |S| > Sc , one can thus assume φ±  = 0. Show, again in global
supersymmetry, that the classical F-term potential is flat along S for vanishing
172 7 Phenomenological Aspects

vevs of φ+ and φ− and that supersymmetry is spontaneously broken in this

valley.
(c) The abovementioned classical flatness along S is lifted by two effects. One is
due to the spontaneously broken supersymmetry, which induces a logarithmic
loop correction. This correction is welcome as it may induce a gentle slope of
the potential along S, which may then serve as an inflaton. If S slowly rolls down
its potential and reaches the critical point |S| = Sc , one of the scalars φ+ and φ−
becomes tachyonic and condenses, and the scalar potential drops to zero. This
would then be a supersymmetric realization of a hybrid inflation model.
This naive picture, however, could be ruined by the second type of corrections,
namely, the classical supergravity corrections to the F-term potential, i.e., all
the MP -suppressed terms in the supergravity F-term potential that are absent in
global supersymmetry. As discussed above, these generically give rise to a too
large inflaton mass/eta parameter. Compute the classical supergravity F-term
potential along the line φ+ = φ− = 0 and show that, for this particular model,
the supergravity η-problem is actually absent, i.e., that the quadratic terms in S
induced by K and W precisely cancel.

References
1. N. Arkani-Hamed, S. Dimopoulos, Supersymmetric unification without low energy supersym-
metry and signatures for fine-tuning at the LHC. JHEP 06, 073 (2005). [arXiv:hep-th/0405159
[hep-th]]
2. G.F. Giudice, A. Romanino, Split supersymmetry. Nucl. Phys. B 699, 65–89 (2004).
[arXiv:hep-ph/0406088 [hep-ph]]
3. M. Billo, F. Denef, P. Fre, I. Pesando, W. Troost, A. Van Proeyen, D. Zanon, The rigid limit
in special Kahler geometry: From K3 fibrations to special Riemann surfaces: a detailed case
study‘. Class. Quant. Grav. 15, 2083–2152 (1998). [arXiv:hep-th/9803228 [hep-th]]
4. E. Cremmer, B. Julia, J. Scherk, P. van Nieuwenhuizen, S. Ferrara, L. Girardello, Super-higgs
effect in supergravity with general scalar interactions. Phys. Lett. B 79, 231–234 (1978)
5. S. Ferrara, A. Van Proeyen, Mass formulae for broken supersymmetry in curved space-time.
Fortsch. Phys. 64(11–12), 896–902 (2016). [arXiv:1609.08480 [hep-th]]
6. M.T. Grisaru, M. Rocek, A. Karlhede, The superhiggs effect in superspace. Phys. Lett. B 120,
110–118 (1983)
7. H.P. Nilles, Phys. Lett. B 115, 193 (1982); A.H. Chamseddine, R. Arnowitt, P. Nath, Phys.
Rev. Lett. 49, 970 (1982); R. Barbieri, S. Ferrara, C.A. Savoy, Phys. Lett. B 119, 343 (1982);
L.E. Ibáñez, Phys. Lett. B 118, 73 (1982); L.J. Hall, J.D. Lykken, S. Weinberg, Phys. Rev. D
27, 2359 (1983); N. Ohta, Prog. Theor. Phys. 70, 542 (1983)
8. L. Randall, R. Sundrum, Out of this world supersymmetry breaking. Nucl. Phys. B 557, 79–118
(1999); G.F. Giudice, M.A. Luty, H. Murayama, R. Rattazzi, Gaugino mass without singlets.
JHEP 12, 027 (1998); A. Pomarol, R. Rattazzi, Sparticle masses from the superconformal
anomaly. JHEP 05, 013 (1999)
9. M. Dine, W. Fischler, M. Srednicki, Nucl. Phys. B 189, 575 (1981); M. Dine, W. Fischler, Phys.
Lett. B 110, 227 (1982); M. Dine, A.E. Nelson, Phys. Rev. D 48, 1277 (1993) [arXiv:hep-
ph/9303230]; M. Dine, A.E. Nelson, Y. Shirman, Phys. Rev. D 51, 1362 (1995) [arXiv:hep-
ph/9408384]
10. J. Polonyi, Generalization of the Massive Scalar Multiplet Coupling to the Supergravity,
Hungary Central Inst. Res. KFKI-77-93 preprint (1977, rec. July 1978), 5 pages, unpublished
References 173

11. G.D. Coughlan, W. Fischler, E.W. Kolb, S. Raby, G.G. Ross, Phys. Lett. B 131, 59–64 (1983);
J.R. Ellis, D.V. Nanopoulos, M. Quiros, Phys. Lett. B 174, 176–182 (1986); B. de Carlos,
J.A. Casas, F. Quevedo, E. Roulet, Phys. Lett. B 318, 447–456 (1993); T. Banks, D.B. Kaplan,
A.E. Nelson, Phys. Rev. D 49, 779–787 (1994)
12. B.S. Acharya, G. Kane, E. Kuflik, Bounds on scalar masses in theories of moduli stabilization.
Int. J. Mod. Phys. A 29, 1450073 (2014)
13. M.R. Douglas, The string landscape and low energy supersymmetry (2012). https://doi.org/10.
1142/9789814412551_0012. [arXiv:1204.6626 [hep-th]]
14. R. Casalbuoni, S. De Curtis, D. Dominici, F. Feruglio, R. Gatto, Nonlinear realization of
supersymmetry algebra from supersymmetric constraint. Phys. Lett. B 220, 569–575 (1989)
15. G. Dall’Agata, E. Dudas, F. Farakos, On the origin of constrained superfields. JHEP 05, 041
(2016) [arXiv:1603.03416 [hep-th]]
16. E. Cremmer, S. Ferrara, C. Kounnas, D.V. Nanopoulos, Phys. Lett. B 133, 61 (1983); J.R. Ellis,
C. Kounnas, D.V. Nanopoulos, Nucl. Phys. B 247, 373–395 (1984); S. Ferrara, C. Kounnas,
F. Zwirner, Nucl. Phys. B 429, 589 (1994). Erratum: [Nucl. Phys. B 433 (1995) 255] [hep-
th/9405188]
17. J. Polchinski, L. Susskind, Phys. Rev. D 26, 3661 (1982); H.-P. Nilles, M. Srednicki, D. Wyler,
Phys. Lett. B 124, 337 (1983); J.R. Ellis, A.B. Lahanas, D.V. Nanopoulos, K. Tamvakis, Phys.
Lett. B 134, 429 (1984); J.R. Ellis, C. Kounnas, D.V. Nanopoulos, Nucl. Phys. B 241, 406
(1984) and B 247, 373 (1984); C. Kounnas, F. Zwirner, I. Pavel, Phys. Lett. B 335, 403 (1994)
[hep-ph/9406256]
18. S. Kachru, R. Kallosh, A.D. Linde, S.P. Trivedi, De Sitter vacua in string theory. Phys. Rev. D
68, 046005 (2003). [arXiv:hep-th/0301240 [hep-th]]
19. G. Dall’Agata, F. Zwirner, New class of N = 1 no-scale supergravity models. Phys. Rev. Lett.
111(25), 251601 (2013). [arXiv:1308.5685 [hep-th]]
20. L. Covi, M. Gomez-Reino, C. Gross, J. Louis, G.A. Palma, C.A. Scrucca, de Sitter vacua in
no-scale supergravities and Calabi-Yau string models. JHEP 06, 057 (2008). [arXiv:0804.1073
[hep-th]]
21. R. Brustein, S.P. de Alwis, Moduli potentials in string compactifications with fluxes: Mapping
the discretuum. Phys. Rev. D 69, 126006 (2004). [arXiv:hep-th/0402088 [hep-th]]
22. D. Junghans, M. Zagermann, A universal tachyon in nearly no-scale de Sitter compactifica-
tions. JHEP 07, 078 (2018) [arXiv:1612.06847 [hep-th]]
23. G.W. Gibbons, Aspects of supergravity theories, in Supersymmetry, Supergravity and Related
Topics, ed. by F. del Aguila, J.A. de Azcárraga and L.E. Ibáñez; B. de Wit, D.J. Smit,
N.D. Hari Dass, Residual supersymmetry of compactified D=10 supergravity. Nucl. Phys.
B 283, 165 (1987); J.M. Maldacena, C. Nunez, Supergravity description of field theories on
curved manifolds and a no go theorem. Int. J. Mod. Phys. A 16, 822–855 (2001). [arXiv:hep-
th/0007018 [hep-th]]
24. A.H. Guth, The inflationary universe: a possible solution to the horizon and flatness problems.
Phys. Rev. D 23, 347–356 (1981); A.D. Linde, A new inflationary universe scenario: a possible
solution of the horizon, flatness, homogeneity, isotropy and primordial monopole problems.
Phys. Lett. B 108, 389–393 (1982); A. Albrecht, P.J. Steinhardt, Cosmology for grand unified
theories with radiatively induced symmetry breaking. Phys. Rev. Lett. 48, 1220–1223 (1982)
25. E.J. Copeland, A.R. Liddle, D.H. Lyth, E.D. Stewart, D. Wands, False vacuum inflation with
Einstein gravity. Phys. Rev. D 49, 6410–6433 (1994). [arXiv:astro-ph/9401011 [astro-ph]]
26. M. Kawasaki, M. Yamaguchi, T. Yanagida, Natural chaotic inflation in supergravity. Phys. Rev.
Lett. 85, 3572–3575 (2000). [arXiv:hep-ph/0004243 [hep-ph]]
27. P. Binetruy, G.R. Dvali, D term inflation. Phys. Lett. B 388, 241–246 (1996). [arXiv:hep-
ph/9606342 [hep-ph]]
28. P. Binetruy, G. Dvali, R. Kallosh, A. Van Proeyen, Fayet–Iliopoulos terms in supergravity and
cosmology. Class. Quant. Grav. 21, 3137–3170 (2004). [arXiv:hep-th/0402046 [hep-th]]
29. L. McAllister, An Inflaton mass problem in string inflation from threshold corrections to
volume stabilization. JCAP 02, 010 (2006). [arXiv:hep-th/0502001 [hep-th]]
174 7 Phenomenological Aspects

30. S. Ferrara, R. Kallosh, A. Linde, Cosmology with nilpotent superfields. JHEP 10, 143 (2014).
https://doi.org/10.1007/JHEP10(2014)143. [arXiv:1408.4096 [hep-th]]
31. G. Dall’Agata, F. Zwirner, On sgoldstino-less supergravity models of inflation. JHEP 12, 172
(2014). [arXiv:1411.2605 [hep-th]]
 Part III
Extended, Gauged and Higher-Dimensional
Supergravity
Extended Supergravities
8

Whereas in N = 1 supersymmetry scalar fields are only mapped to certain spin-

1/2 fermions, the larger multiplets of extended supersymmetry may link scalars
also to fields of spin greater than 1/2, in particular vector fields. As we will see
in the following, this heavily constrains the structure and couplings of extended
supergravities. The main reason for this is that in a theory without charged fields,
the field equations of Abelian vector fields in four spacetime dimensions exhibit
the phenomenon of electric–magnetic duality. In extended supersymmetry, this
electric–magnetic duality structure of the vector field sector is then linked to the
scalar field geometry by supersymmetry and constrains the scalar manifolds of most
supermultiplets. Moreover, this same duality structure is going to be at the basis of
the consistent construction of deformations that make some of the fields charged
under the gauge group. We therefore first discuss in Sect. 8.1 the main features of
electric–magnetic duality in preparation for our discussion of extended supergravity.
In the subsequent sections of this chapter, we then explain how electric–magnetic
duality and the structure of the R-symmetry group constrain the geometries of the
corresponding scalar manifolds. In Sect. 8.2, this will lead us to special Kähler
and quaternionic manifolds for the scalar fields in, respectively, the vector and
hypermultiplets in N = 2 supergravity. The scalar manifolds for supergravity
theories with N ≥ 3 supergravity are then discussed in Sect. 8.3. The appendix to
this chapter contains computational details on the scalar field geometries in N = 2
supergravity.

8.1 Electric–Magnetic Duality

It is well-known that Maxwell’s equations in the vacuum are invariant under the
exchange of the electric and magnetic vector fields: E → B and B → −E.  In the
Lorentz covariant notation, this duality relation is expressed as the exchange of the

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 177

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_8
178 8 Extended Supergravities

gauge field strength, F = 1/2 dx μ ∧ dx ν Fμν , and its Poincaré dual,

1 μ
F = F = dx ∧ dx ν Fμν , (8.1)
2
where
1
Fμν ≡ μνρσ F ρσ . (8.2)
2
In fact, Maxwell’s equations in vacuum can be written as
5
dF = 0,
(8.3)
d  F = 0,

where the first equation is usually interpreted as a Bianchi identity for the curvature
F = dA of the gauge potential, A, and the second is its equation of motion.
However, by the introduction of the dual gauge field strength, F , these equations
can be written in the more symmetric form
5
dF = 0,
(8.4)
d F = 0,

which is invariant under F → F = F and F → F = ()2 F = −F . Actually, one

could mix the two curvatures with general linear transformations without changing
the content of (8.4). This duality transformation is obviously violated by minimal
electric couplings and/or non-Abelian gauge field strengths as these contain direct
couplings of the gauge potentials that cannot be written in terms of the Abelian
field strengths alone. However, we can introduce a new generalized form of duality
transformation for Abelian vector fields (and their field strengths) that is valid also
in the presence of non-minimal matter couplings.
The first thorough analysis of the consequences of the existence of electric–
magnetic duality for generic field theories containing vector fields was performed
in [1]. This work also constitutes the basis of the discussion for their generalization,
namely, U-duality symmetries, which we will give in the next chapter. Although
we borrow a lot from that paper, in the following, we provide an original presen-
tation, clarifying some aspects that are especially relevant for the construction of
supergravity theories.
8.1 Electric–Magnetic Duality 179

Consider a generic 2-derivative Lagrangian containing nV Abelian vectors, AI

(I = 1, . . . , nV ), through their field strengths, F I = dAI , and arbitrary couplings
to other fields ϕ i (denoting together bosonic and fermionic fields),

1 1 1 μν I
e−1 L (F I , ϕ i , ∂μ ϕ i ) = II J Fμν
I
F J μν + RI J Fμν
I
F J μν + OI Fμν
4 4 2
+ e−1 Lrest . (8.5)

Here, II J and RI J are symmetric matrices that may depend on the scalar fields,
μν
with II J being negative definite to ensure unitarity, OI is a generic tensor function
of the other fields containing at most a single derivative, and Lrest contains all the
terms that do not depend on the vector field strengths. By definition, the vector field
curvatures are F I = dAI and hence

dF I = 0 (8.6)

describe the Bianchi identities of the nV vector fields, AI . The equations of motion
for such fields then follow as usual from setting

∂L
∇μ = 0. (8.7)
∂F μνI
These equations can also take the form of Bianchi identities if we introduce dual
variables

∂L
GI μν = 2 (8.8)
∂F I μν
ρσ
and GI μν = 1
2 μνρσ GI , so that (8.7) can be written as

dGI = 0. (8.9)

The system of Bianchi identities and equations of motion is a priori invariant

under constant GL(2nV , R) transformations,
 

FI
F = SF, F≡ (8.10)
GJ

with S ∈ GL(2nV , R). However, in order to preserve the definition of the GI tensors
in terms of F I and via (8.8) for the transformed Lagrangian, we have to further
constrain the matrix S.
180 8 Extended Supergravities

In order to do so, notice that (8.9) implies the (local) existence of nV dual 1-
forms AI such that GI = dAI . Obviously, the Lagrangian will depend only on one
of the two, but we can trade one for the other by a Legendre-like transformation,
introducing a total derivative term to the action of the form
 
1 1
S = F I ∧ dAI = F I ∧ GI . (8.11)
2 2

Varying the action S + S  with respect to AI , one obtains the usual Bianchi identity
dF I = 0, while varying with respect to F I , one obtains the definition of GI =
dAI and hence one could rewrite the Lagrangian in terms of the dual potentials by
plugging the solution to these equations back into the action. Let us then perform a
Legendre-like transformation from the F to the G variables and introduce the dual
Lagrangian
 
−1 −1 1 I μν
e LD = e L − Fμν GI , (8.12)
2 F =F (G)

where, after adding (8.11) to the original action, we replace each instance of the
original field strengths, F I , with their expression in terms of the dual ones, GI . We
call this transformation Legendre-like because the dual Lagrangian is defined by the
difference between the Lagrangian itself and the product of the variables on which
the original Lagrangian depends (the F I field strengths, which we treat as black
boxes) and the dual variables (the GI forms we introduced above), like an actual
Legendre transformation. This analogy can be actually extended to the point that
the original variables are recovered by varying the dual Lagrangian with respect to
the dual variables:
∂LD
I
Fμν = −2 μν . (8.13)
∂ GI

In fact, we can expand the dual Lagrangian in terms of the dual field strengths as
follows:
1 IJ μν 1 μν 1
e−1 LD = 
A GI μν GJ + B I J GI μν GJ − O I μν GJ μν +Lrest , (8.14)
4 4 2

where, again, A I J and B I J are symmetric matrices, O I μν is a tensor that does

 is the part of the Lagrangian that does not
not contain the GI μν fields, and Lrest
depend at all on the GI μν fields. The fact that (8.14) has the form presented above
can be argued by the linearity of the relation (8.8) between the original vector field
strengths and their duals.
8.1 Electric–Magnetic Duality 181

Now that we have an explicit form for both L (F I ) and LD (GI ), we can recover
the explicit relations following from applying (8.8) and (8.13). The results are
collected in the following expressions:

GI μν = II J Fμν
J
+ RI J Fμν
J
+ OI μν ,

GI μν = −II J Fμν
J
+ RI J Fμν
J
− OI μν ,
(8.15)
I
Fμν = −A I J GJ μν − B I J GJ μν + Oμν
I
,

I
Fμν = A I J GJ μν − B I J GJ μν + Oμν
I
,

where we explicitly wrote the relations involving the Fμν curvatures and their Hodge
duals Fμν = 12 μνρσ Fρσ , for completeness. Consistency of these conditions then
gives the following conditions:

A = −(I + RI −1 R)−1 , (8.16)

B = A RI −1 = I −1 RA , (8.17)

O I = A I J OJ − B I J OJ , (8.18)

 1
Lrest = Lrest + O I OI , (8.19)
4

which come from inserting the expressions (8.15) for F I and F I into the ones for
GI and GI . The outcome is that we can express LD fully in terms of the quantities
appearing in the original L . Moreover, using these relations, we can then explicitly
check that (8.12) is indeed identically satisfied as it should.
At this point, we can impose the consistency constraints on the duality transfor-
mations (8.10) by applying them to one of the explicit relations (8.15) and imposing
its consistency. Let us see how. At the infinitesimal level, the duality transformation
(8.10) can be written in terms of four real matrices

δF I = AI J F J + B I J GJ ,
(8.20)
δGI = CI J F J + DI J GJ .

If we now apply this to the first line in (8.15), we see that

δ GI μν = −δII J F μν
J
+ δRI J F μν
J
− II J δF μν
J
+ RI J δ F μν
J
+ δOI μν , (8.21)
182 8 Extended Supergravities

which gives

J
CI J Fμν + DI J GI μν = − δII J F μν
J
+ δRI J F μν
J

 
− II J AJ K F μν
K
+ B J K GK μν (8.22)
 
+ RI J AJ K F μν
K
+ B J K GK μν + δOI μν .

This should be an identity once we express all GI in terms of F I or the opposite,

using again (8.15). Once we do so, we find the following transformation rules for
the matter couplings

δII J = DI K IKJ − II K AK J − II K B KL RLJ − RI K B KL ILJ ,

δRI J = CI J + DI K RKJ + II K B KL ILJ − RI K AK J − RI K B KL RLJ ,

δOI = DI J OJ − II J B J L OL − RI J B J L OL .
(8.23)

Moreover, consistency of δI T = δI and δR T = δR implies that

C = CT , B = BT , A = −D T , (8.24)

which are the conditions on the infinitesimal duality transformation

 to be in the
AB
algebra sp(2nv , R). In fact (8.24) is the statement that S = exp satisfies
CD
 
0 1nV
S ΩS = Ω,
T
Ω= , (8.25)
−1nV 0

or, at the infinitesimal level, (8.24). In the following, we will call a 2nV -component
object, Y , that transforms under electric–magnetic duality transformations in the
same way as the field strength vector Fμν ,

Y '→ S · Y, (8.26)

a symplectic vector and define the symplectically invariant inner product between
two symplectic vectors, Y and Z, as

Y, Z := Y T ΩZ. (8.27)

Before moving on, let us notice two important facts. The first one is that the
invariance of the system dF = 0 does not imply invariance of the Lagrangian.
8.1 Electric–Magnetic Duality 183

Indeed, applying an infinitesimal duality transformation, the Lagrangian transforms

as
1 1
δL = F C F + GB G + δLrest . (8.28)
4 4
Even if δLrest = 0, as we will prove later, this implies that the Lagrangian is
invariant, up to a total derivative (the F C F̃ term), only if B = 0. This case
corresponds to perturbative transformations, which in the quantum theory have to
be restricted to Sp(2nV , Z). The second one is that the full electric–magnetic duality
group acts with fractional transformations on the scalar matrices and O couplings of
the original action. This becomes clear if we introduce the complex kinetic matrix

NI J = RI J + i II J (8.29)

and the self-dual combination1

1
O+ = (O − i O) , (8.30)
2

which indeed satisfies O + = iO + . If we act with an infinitesimal duality

transformation (8.23) on these objects, we find

δN = C + DN − N A − N BN , (8.31)

and

δO + = O + (D T − BN ) = O + (−A − BN ), (8.32)

or (always at first order in A, B, C, and D)

N  = C + (1 + D)N (1 − A − BN ), (8.33)

O + = O + (1 − A − BN ). (8.34)

It is straightforward to see that these are the expansion of

N  = (Ĉ + D̂N )(Â + B̂N )−1 , (8.35)

1 We define the self-dual and anti-self-dual tensor field combinations

 
± 1 i
Tμν ≡ Tμν ∓ μνρσ T ρσ ,
2 2

which satisfy 1 ρσ ± ±.
= ± i Tμν
2 μνρσ T
184 8 Extended Supergravities

and

O + = O + (Â + B̂N )−1 , (8.36)

which are the fractional transformation of the kinetic couplings and the remaining
matter couplings once we identify
 
 B̂
S= , (8.37)
Ĉ D̂
 
AB
and we recover the infinitesimal expansion S = 1+ . This now also explains
CD
why B̂ = 0 corresponds to non-perturbative duality transformations, because from
(8.35) we see that B̂ = 0 involves inversions of the components of N , which may
change weak couplings to strong couplings or vice versa.
Finally, we close this section by noting that the introduction of self-dual and anti-
self-dual field tensors allows the rewriting of the kinetic Lagrangian in the form

1 1
e−1 Lkin = I
(ImNI J )Fμν F μνJ − (ReNI J ) μνρσ Fμν
I J
Fρσ (8.38)
4 8
1  I + μνJ +

= Im NI J Fμν F , (8.39)
2

so that the duality action F = SF becomes

+ +
Fμν = (Â + B̂N )Fμν , G+ +
μν = (Ĉ + D̂N )Fμν , (8.40)

which is compatible with the transformation (8.35) for the kinetic matrix N . Notice
also that

G+ J+
I μν = NI J Fμν . (8.41)

8.2 N = 2 Supergravity

We have seen in Chaps. 5 and 6 that the scalar manifold, Mscalar , of N = 1

chiral multiplets must be a Kähler manifold in 4D global and local N = 1
supersymmetry. The coupling to supergravity introduces a global restriction, which
leads to the subclass of Kähler-Hodge manifolds, as discussed in Sect. 6.2, but the
Kähler manifolds in N = 1 supersymmetry are otherwise arbitrary.
In N = 2 supersymmetry, matter can reside in N = 2 vector or hypermulti-
plets, which, at the linearized level, can be thought of as being composed of one
N = 1 vector and one chiral multiplet, or two chiral multiplets, respectively. The
8.2 N = 2 Supergravity 185

presence of global or local N = 2 supersymmetry, however, has quite distinct

consequences for the possible scalar geometries of these two types of N = 2
multiplets and leads to a factorization of the scalar manifold,
N =2
Mscalar = Mvec × Mhyper , (8.42)

because the superpartners of the scalars in vector and hypermultiplets involve

different spins and hence are not related by any symmetry, unless one has N ≥ 2
supersymmetry.
As we will now discuss, the scalar fields of the N = 2 vector multiplets parame-
terize a Special Kähler manifold, which comes in two different versions, depending
on whether one has global or local N = 2 supersymmetry. The hypermultiplet
scalars, on the other hand, are constrained to form a hyper-Kähler manifold in global
supersymmetry or a quaternionic-Kähler manifold in supergravity. Apart from
quaternionic-Kähler manifolds, all the above geometries are particular subclasses of
Kähler manifolds. For the globally supersymmetric theories, this is clear, because
any field theory with N -extended global supersymmetry is only a special case of
the field theories with N  < N global supersymmetry, whereas an analogous
statement is in general not true in supergravity due to the missing gravitini in
theories with lower N .
The above N = 2 scalar field geometries feature prominently in various
applications of string and field theory and define interesting mathematical structures
in their own right. We therefore devote an extra section to their structure and explain
how this structure is imposed by supersymmetry.
For the rest of this chapter, we set

MP = 1, (8.43)

unless stated otherwise.

8.2.1 N = 2 Vector Multiplets and Special Kähler Geometry

As N = 2 supersymmetry interpolates between the complex scalar and the vector

field of an N = 2 vector multiplet, the symplectic duality covariance of the vector
field sector discussed in Sect. 8.1 should leave an imprint on the scalar manifold
Mvec . This imprint is expected to be different for global and local supersymmetry,
because the latter involves an additional vector field from the N = 2 supergravity
multiplet. In the following, we will explain how precisely the resulting special
Kähler geometry arises for global and local supersymmetry and discuss their
differences.

8.2.1.1 Rigid (“Affine”) Special Kähler Geometry

The special Kähler geometry that arises in N = 2 global supersymmetry is called
“rigid” or “affine” special Kähler geometry (for a review see [2, 31]). The field
186 8 Extended Supergravities

theories with global N = 2 supersymmetry form a subclass of the field theories

with global N = 1 supersymmetry, defined by the invariance under the larger
N = 2 Poincaré superalgebra. The essential features of rigid special Kähler
geometry can be easily derived by imposing this larger invariance on the general
N = 1 field theories of Chap. 5. In fact, it is sufficient to just impose invariance
under the discrete R-symmetry [3]

Q(1) → Q(2) , Q(2) → −Q(1) , (8.44)

which exchanges the two prospective supersymmetry generators Q(1) and Q(2) . This
R-symmetry is an element of the full R-symmetry group U(2)R .
In terms of N = 1 language, an N = 2 vector multiplet is decomposed of an
N = 1 vector multiplet (Aμ , λ) and an N = 1 chiral multiplet (χ, φ). In a theory
with only nV N = 2 vector multiplets, one therefore has to restrict oneself to an
equal number, nV , of N = 1 vector and chiral multiplets, which we therefore label
by a common index I = 1, . . . , nV . To make the equations look more natural, we
furthermore slightly adjust the names and normalizations of the fields and write the
field content as
 I 
I λ
(Aμ , λ , λ , X ) = Aμ , χ , , φ ,
I I (1) I (2) I I I
(8.45)
2

where the right-hand side contains the original N = 1 fields, and we have
combined the χ I and λI /2 into the N = 2 gaugini λI (i) (i, j, . . . = 1, 2) as
indicated. Also, in the following, to adhere to usual N = 2 notation, we will denote
I
the complex conjugation on scalars with a bar X = (XI )∗ .
As the gaugini λI (i) are obtained by acting with Q(i) on the vector fields AIμ , the
discrete R-symmetry (8.44) now exchanges the gaugini λI (1) and λI (2) ,

λI (1) → λI (2) , λI (2) → −λI (1) (8.46)

and leaves the other fields invariant.

This means, in particular, that the kinetic terms of the fermions (cf. the N = 1
Lagrangian (5.120)),
 
I (1) J (1) J (1) I (1)
A := −gI J λL Dλ / R + λR Dλ / L (8.47)

I (2) I (2)
B := −2(RefI J )λ ∂/λJ (2) + i∂μ (ImfI J )λ γ5 γ μ λJ (2) , (8.48)
8.2 N = 2 Supergravity 187

where we integrated by parts in the last term, have to transform into each
other under (8.46). Using Dμ χLJ = ∂μ χLJ + (∂μ XI )ΓIJK χLK and ∂μ (Im fI J ) =
K
(∂μ XK )∂K Im fI J + (∂μ X )∂K Im fI J , this then leads to the constraints

gI J = gJ I = 2 Re fI J (8.49)

2i∂K (Im fI J ) = ΓKI J = ΓKJ I . (8.50)

The holomorphicity of fI J and the Kähler relation ΓKI J = ∂K gI J imply (8.50)

from (8.49), which is thus the only non-trivial condition. Due to the holomorphicity
of fI J and the Kähler relation gI J = ∂I ∂J K, (8.49) is (up to Kähler transforma-
tions) solved by
 
K = i XI FI − XI FI , (8.51)

where FI (X) are holomorphic functions of the scalars XI . In this expression, the
factor i is introduced in order to have a relative minus sign between the two terms,
which renders the Kähler potential manifestly Sp(2nV , R) invariant if we identify
XI and FJ as components of a symplectic vector,
 I
X
V := . (8.52)
FJ

Indeed, after we introduce the symplectic invariant matrix Ω, the Kähler potential
can be written as

K = iV , V  = iV T ΩV , (8.53)

which is manifestly Sp(2nV , R) invariant under

V  = SV , (8.54)

for any S ∈Sp(2nV , R).

From (8.51) and (8.49), one further obtains

gI J = ∂I ∂J K = −i(∂I FJ − ∂J FI ) = (fI J + fI J ), (8.55)

so that fI J = −i∂I FJ . The symmetry of fI J then implies ∂[I FJ ] = 0, i.e.,

FI = ∂I F, (8.56)

where F (X) is a holomorphic function of the scalars, and hence

fI J = −i ∂I ∂J F. (8.57)
188 8 Extended Supergravities

The holomorphic function F (X) is called the holomorphic prepotential. It deter-

mines the Kähler potential K and the gauge kinetic matrix fI J by its first and second
derivatives, respectively. As in Sect. 8.1, the gauge kinetic matrix is often expressed
in terms of its complex conjugate

NI J := −i fI J = ∂I ∂J F ≡ FI J , (8.58)

in terms of which (8.55) becomes

1
ImNI J = − gI J . (8.59)
2
Equations (8.52), (8.53), (8.56), and (8.58) capture the essence of rigid special
Kähler geometry, which describes the scalar field geometry of N = 2 vector mul-
tiplets in field theories with global N = 2 supersymmetry.2 From a mathematical
point of view, however, the above equations are still unsatisfactory as a basis for
an intrinsically geometric and coordinate independent definition of rigid special
Kähler geometry. The core of this problem lies in Eq. (8.54), which states that the
vector V transforms under a symplectic duality transformation S ∈ Sp(2nV , R)
in the same way as the vector field strengths Fμν I and their duals G
μνI . This
identification of V with a symplectic vector, however, then raises an important
technical question: How is the general holomorphic reparameterization invariance
of the scalars, XI '→ XI (X), on a Kähler manifold compatible with the symplectic
covariance of the theory, when the XI are restricted to belong to a symplectic vector
that only admits linear reparametrizations?
To understand this, we consider the supersymmetry transformations of the
fermions with respect to the original N = 1 supersymmetry, which we take to
be parameterized by  (1) (cf. Eqs. (5.123)),

I (1) 1
δ (1) λL = (∂/XI )R(1) (8.60)
2
I (2) 1 I (1) 1 I − (1)
δ (1) λL = γ μν Fμν L = γ μν Fμν L , (8.61)
8 8

where we have used L(1) = PL  (1) and γ μν = (i/2) μνρσ γρσ γ5 to convert Fμν
I to
I − . The transformation (8.60) suggests, just as in N = 1 supersymmetry, that
Fμν
I (1)
λL transforms as a holomorphic tangent vector under scalar reparameterizations
I (1) J (1)
XI '→ XI (X), i.e., λL '→ (∂ XI /∂XJ )λL . The R-symmetry (8.46) then
I (2)
implies that also λL has to transform in this way, so that (8.61) then ultimately
would imply that also the field strengths transform with the Jacobian, Fμν I − '→
J J −
(∂ X /∂X )Fμν . If the transformation X '→ X (X) is nonlinear, however, the
I I I

2 In the literature these equations often appear with different normalizations.

8.2 N = 2 Supergravity 189

Jacobian is XI -dependent, so that the field equations and Bianchi identities (8.6)
and (8.7) would no longer hold for the transformed field strengths and instead get
modified by spacetime derivatives of scalar fields.
The most natural resolution of this tension between symplectic duality and
scalar reparameterization invariance is achieved by separating these two types of
transformations. To this end, one considers the symplectic vector V (z) as an abstract
holomorphic 2nV -dimensional vector field with components labeled by I, J, . . . =
1, . . . , nV on the scalar manifold Mvec , which itself is locally parameterized by nV
complex coordinates, zm , (m, n, . . . = 1, . . . , nV ). The electric–magnetic duality
transformations then only act on the components of V with a symplectic matrix
S ∈ Sp(2nV , R) as in (8.54), but not on the coordinates zm of the scalar manifold.
In other words, V (z) is a section in a holomorphic vector bundle with fiber C2nV
and structure group Sp(2nV , R). As the symplectic matrix S does not depend on the
coordinates zm , this bundle is topologically trivial.
The physical scalar fields are then the zm , which in general are not equal to
the components XI (z). The special coordinate systems in which the coordinates
zm are equal to the components XI (z) (i.e., z1 = X1 (z), etc.) are called special
coordinates. Under a general holomorphic scalar field reparameterization zm '→
zm (z), the components of V do not get rotated, but behave as ordinary functions on
the scalar manifold. In other words, the symplectic vector bundle is not identified
with or related to the tangent bundle of the scalar manifold.
Just as in N = 1 supersymmetry, the left- and right-handed components of the
fermions transform under scalar reparameterizations zm '→ zm (z) as, respectively,
holomorphic and antiholomorphic tangent vectors on the scalar manifold and are
consequently also denoted with indices m, n, . . .:
   m

m(j ) ∂ zm n(j ) m(j ) ∂z n(j )
λL '→ λL , λR '→ λR . (8.62)
∂zn ∂z n

Just as the scalar fields, zm , the fermions are then considered invariant with respect
to symplectic duality transformations.
A covariant form of the supersymmetry transformations that respects all the
above symmetries and reduces to (8.60) and (8.61) for special coordinates is then
given by

m(1) 1 m (1)
δ (1) λL = (∂/z )R (8.63)
2
m(2) 1 J − μν (1)
δ (1) λL = − g mn ∂n XI (ImNI J )Fμν γ L , (8.64)
4

where g mn (z, z) denotes the inverse of the metric on the scalar manifold. For (8.63)
this is easy to see, as both sides are symplectically invariant and transform as
tangent vectors with respect to scalar reparameterizations with an obvious reduction
to (8.60) upon choosing special coordinates. As for (8.64), both sides are again
190 8 Extended Supergravities

manifestly transforming as holomorphic tangent vectors with respect to scalar field

reparameterizations. The left-hand side is also manifestly symplectically invariant.
It thus remains to show that the right-hand side is also symplectically invariant and
reduces to the right-hand side of (8.61) upon switching to special coordinates.
To show the symplectic invariance, we use GμνI − = NI J Fμν J − and compute

J− 1 J−
∂n XI (ImNI J )Fμν = ∂n XI (NI J − NI J )Fμν (8.65)
2i
1 J−

= ∂n XI NI J Fμν − ∂n XI GμνI − . (8.66)
2i

This is the symplectic inner product, 2i1 F−

μν , ∂n V , of the two symplectic vectors
   I− 
∂n X I Fμν
∂n V = and F−
μν = provided that
∂n FJ GμνJ −

∂n XI NI J = ∂n FJ . (8.67)

This relation, however, simply follows from NI J = ∂I ∂J F , FJ = ∂J F and

the chain rule, showing that the right-hand side of (8.64) really is symplectically
invariant.
The equivalence to (8.61) in special coordinates, finally, follows from ∂n XI = δnI
and (8.59), which are valid if the zm are identified with the XI .
We finally note that (∂n XI ) is the Jacobian of the change from general coordi-
nates zn to special coordinates and hence must be invertible. This implies that (8.67)
together with the relation FI = ∂I F fixes the gauge kinetic matrix to

NI J = (∂n XI )−1 · (∂n FJ ), (8.68)

or equivalently

NJ I = (∂n FJ ) · (∂n XI )−1 , (8.69)

because NI J = NJ I . In this notation, we can now prove that NI J transforms under

symplectic duality transformations in the way it should (cf. Eq. (8.35)). In fact, from
the action of (8.37) on the symplectic sections V , we inherit
   −1
NJI = C !J K ∂n FK · A
!J K ∂n XK + D !I L ∂n FL
!I L ∂n XL + B , (8.70)
8.2 N = 2 Supergravity 191

which can also be written as

  −1 −1  
NJI = C !J L ∂m FL · ∂m XK
!J K + D ∂n XK ·
(8.71)
 −1   −1 −1
· ∂n X L !I !
A L+B IM
∂m FM · ∂m X L ,

which finally gives (8.35), once (8.69) is used.

Notice that (8.69) is invariant under a constant phase redefinition of the symplec-
tic sections V , as well as under a constant shift V → V + c, with c ∈ C2nV . In fact
it is easy to see that the entire Lagrangian and the supersymmetry laws are invariant
under such symmetries, provided that the gaugini λm(i) and the supersymmetry
parameter  (i) transform as well with a suitable chiral phase factor. Because of these
symmetries, when we define a special Kähler manifold, we have to admit that on
different coordinate patches the symplectic sections may differ by constant phases
as well as constant complex shifts. This leads to the following general definition of
rigid special Kähler geometry [2, 4]:

Definition. A rigid special Kähler manifold is a n-dimensional Kähler

manifold of restricted type:

• it is equipped with a tensor bundle H given by the product of a flat U(1)

bundle and a vector bundle with an inhomogeneous symplectic structure
group, so that, on each patch, UA , of a good cover, a section of H is
described by a symplectic vector
 I
X
V := , I = 1, . . . , n, (8.72)
FJ

such that the transition functions between two different local trivializations
of H on UA and on UB have the form

VA = ei φAB SAB VB + cAB , (8.73)

where SAB is a constant matrix in Sp(2n, R), φAB ∈ R, cAB ∈ C2n ;

• the Kähler potential is given by

K = iV , V  = iV T ΩV , (8.74)

where ·, · is a Hermitian metric on H ;

(continued)
192 8 Extended Supergravities

• the sections satisfy

∂m V , ∂n V  = 0. (8.75)

This last equation guarantees the existence of a prepotential and the symmetry
of the kinetic vector matrix NI J .
The transition functions are subject to the usual consistency conditions on
triple overlaps:

φAB φBC φCA = 1, SAB SBC SCA = 1. (8.76)

The covariant transformation laws (8.63)–(8.64) and the relation (8.67) will have
important analogues in N = 2 supergravity to which we turn next.

8.2.1.2 Local (“Projective”) Special Kähler Geometry

Local or “projective” special Kähler geometry arises when N = 2 vector multiplets
are coupled to N = 2 supergravity. As we have just discussed, an N = 2 vector
multiplet contains one vector field, Aμ (x), two gaugini, λ(i) (x) (i = 1, 2), and one
complex scalar field, z(x). The N = 2 supergravity multiplet, on the other hand,
(i)
consists of the vierbein, eμa (x); two gravitini, ψμ (x) (i, j, . . . = 1, 2); and another

vector field, Aμ (x), often called the “graviphoton.” The index i corresponds to
the fundamental representation of the R-symmetry subgroup SU(2)R of the total
R-symmetry group U(2)R .
Thus, coupling nV N = 2 vector multiplets to N = 2 supergravity gives rise
to a theory with the field content

{eμa , ψμ(i) , AIμ , λm(i) , zm }, (8.77)

where I, J, . . . = 0, 1, . . . , nV , and m, n, . . . = 1, . . . , nV . Note that the vector

fields, AIμ , now include the graviphoton so that there are (nV + 1) vector fields and
only nV physical complex scalar fields, zm . In global supersymmetry, by contrast,
one has as many vector fields as physical complex scalars so that one can in principle
use the same index type to label them, as we first did at the beginning of Sect. 8.2.1.1.
As we saw there, however, even for global supersymmetry the clearest notation that
makes all symmetries manifest is to distinguish the index I for the vector fields and
the index m for the complex scalars and the gaugini. In N = 2 supergravity, the
mismatch in the number of vector and scalar fields makes this distinction even more
natural, so that we will use it here right from the beginning.
We are thus immediately led to the interpretation that the (nV + 1) Abelian
I + , and their magnetic duals, G J+
field strengths, Fμν μνI + = NI J Fμν , combine into
+
a (2nV + 2)-component symplectic vector, Fμν , that transforms in the fundamental
8.2 N = 2 Supergravity 193

representation of the electric–magnetic duality group Sp(2nV + 2, R),

 I+   I+ 
Fμν Fμν
F+μν = →
' S· = S·F+ μν , S ∈ Sp(2nV +2, R). (8.78)
GμνJ + GμνJ +

The matrix NI J in the above expression for GμνI + is the scalar field-dependent
gauge kinetic function of the theory, which thus transforms as in (8.35) under
symplectic duality transformations. Its precise dependence on the scalar fields zm
is of a different form than in global supersymmetry, however, and will be derived
further below.
The scalars zm and the gaugini λm(i) , by contrast, are symplectically invariant,
but transform under holomorphic scalar field reparameterizations as
 m
∂z
zm '→ zm (z), λm(i) '→ λn(i) , (8.79)
∂zn

with the vector fields AIμ being invariant.

As the coupling to supergravity does not involve new gaugini or new scalars, the
basic supersymmetry relation between the gaugini and the scalars should again link
left-handed fermion components with the zm and right-handed fermion components
with the complex conjugates zm . Just as discussed in Sect. 5.2.4, the holonomy
group of the scalar manifold of the zm should then again be contained in U (nV ),
i.e., the scalar manifold should again be a (Hodge-)Kähler manifold. In parallel
with the situation in global supersymmetry, however, we expect that the interplay
with the symplectic duality of the vector fields and the extended supersymmetry
will lead to restrictions on the possible Kähler geometries, and hence to another
version of “special Kähler geometry.” As the number of vector fields and the rank of
the symplectic duality group are now increased, it is plausible that the type of special
Kähler geometry will be different from the rigid special Kähler geometry discussed
in the previous subsection. This is indeed true, and the essential differences with the
rigid case turn out to be [2, 5–8]:

rigid N = 2 supersymmetry, there is a holomorphic symplectic section

1. Just as in 
XI (z)
V (z) = (I, J, . . . = 0, 1, . . . , nV ) that depends holomorphically on
FJ (z)
the scalar fields zm and transforms as

V (z) '→ S · V (z), S ∈ Sp(2nV + 2, R) (8.80)

under electric–magnetic duality transformations. An obvious difference to the

rigid case here is that V has two more components to match the number  of
I+
Fμν
+
components of the symplectic field strength vector Fμν = . The
GμνJ +
existence of this symplectic section ensures that the holomorphic reparameter-
ization invariance of the scalars and the symplectic covariance of the vector field
194 8 Extended Supergravities

sector can coexist and be consistent with the supersymmetry transformation law
of the gaugini, similar to the situation in rigid supersymmetry. This is explained
in more detail in Appendix 8.A.1, where we will also explain the origin of the
holomorphicity of V .
2. Under a general Kähler transformation, K(z, z) → K(z, z) + h(z) + h(z), the
holomorphic symplectic section V transforms non-trivially as

V (z) '→ e−h(z) V (z). (8.81)

In rigid special Kähler geometry, by contrast, at most an additive constant vector

to V (z) could generate a very specific type of Kähler transformation. This non-
trivial Kähler transformation of the symplectic section is needed to reconcile the
various Kähler transformation properties of the fermions with one another, as is
explained in detail in Appendix 8.A.1.
3. The Kähler potential of the scalar manifold parameterized by the zm is given by
 
K(z, z) = − log i XI (z)FI (X(z)) − FI (X(z))XI (z) = − log iV , V .
(8.82)

Similar to rigid supersymmetry, this is manifestly symplectically invariant, but

it differs from the rigid expression (8.51) by the logarithm. Just as in N = 1
supergravity, the transformation of the fermions under Kähler transformations
leads to a global restriction and requires the Kähler manifolds to be Hodge-
Kähler manifolds. The logarithmic form of the Kähler potential is due to the
Kähler transformation property (8.81), as is also explained in Appendix 8.A.2.
4. In contrast to rigid special Kähler geometry, the lower components FJ are not
necessarily the derivatives of a holomorphic prepotential F (X) with respect to
XJ . However, for nV > 1, there is always a symplectic matrix S ∈Sp(2n V +
XI
2, R) such that V = S·V does have this property, i.e., such that V = with
FJ
FJ = ∂ F (X)/∂ XJ .3 More precisely, for nV > 1, one has to impose an additional
constraint on the symplectic section V (z) in order to ensure the symmetry and
uniqueness of the gauge kinetic matrix NI J . This condition is

Dm V , Dn V  = 0, (8.83)

3 For symplectic frames with FJ = ∂F /∂X J , one often says that this is a section “without
prepotential” or “the prepotential does not exist.” A more precise statement would be that (for
nV > 1) there is always a symplectic frame where such a prepotential does exist, but that this is
just not the symplectic frame under consideration. We will nevertheless also often use the above
less precise terminology. In fact, in the case of gauge interactions, the electric–magnetic duality is
at least partially broken by the presence of “naked” vector fields AIμ without spacetime derivatives,
so that the symplectic frames with the standard prepotential form might not be accessible in the
usual way.
8.2 N = 2 Supergravity 195

where Dm V = (∂m +(∂m K))V is the Kähler covariant derivative of V . With this
condition, one can prove the existence of a symplectic frame with prepotential
for nV > 1. For the case nV = 1, however, the condition (8.83) is empty as it is
antisymmetric in m and n. And in fact, for nV = 1, counterexamples without a
prepotential in any symplectic frame have been constructed [4]. More details on
this issue are given in Appendix 8.A.3.
In symplectic frames where the identification FJ = ∂F /∂XJ is possible, there
is also an analogue of the special coordinates we encountered in rigid special
Kähler geometry, where one could (locally) always choose the zm such that they
can be identified with the XI (z). In local special Kähler geometry, this cannot
work exactly this way because there is now one more XI than zm . However,
due to the complex rescalings (8.81) induced by Kähler transformations, one
can interpret the XI as homogeneous coordinates of a projective space and then
identify the zm with the corresponding inhomogeneous coordinates, e.g., zm =
Xm /X0 . This is the reason for the alternative name “projective special Kähler
geometry.”
5. Another difference to the rigid case is that such a holomorphic prepotential F (X)
must be homogeneous of degree two, i.e., under a rescaling XI → κXI , one
has F (κX) = κ 2 F (X), whereas in the rigid case, F (X) can be an arbitrary
holomorphic function (as long as the resulting metric gmn is positive definite).
This can, e.g., be understood from the consistency of the non-trivial Kähler
transformation (8.81) with FI = ∂I F , but it is also necessary for ensuring the
right properties of the gauge kinetic matrix, NI J . Details on this can be found in
the Appendices 8.A.3 and 8.A.4.
6. In a symplectic frame with a holomorphic prepotential F (X), the gauge kinetic
matrix NI J (z, z) is given by

ImFI K (X)ImFJ L (X)XK XL

NI J = FI J (X) + 2i , (8.84)
ImFKL (X)XK XL

where FI J ≡ ∂I ∂J F . This expression differs from the corresponding expression

(8.58) by the second, non-antiholomorphic piece. The reason for this extra term
is that, in connection with the degree 2 of the prepotential, it makes the super-
symmetry transformation law of the gravitini symplectically invariant without
destroying the symplectic invariance of the supersymmetry transformation law
of the gaugini. This is further explained in Appendix 8.A.4.
196 8 Extended Supergravities

Altogether, we can therefore define the local special Kähler geometry as

Definition. A local special Kähler manifold is a n-dimensional Hodge–

Kähler manifold of restricted type:

• it is equipped with a tensor bundle, H , given by the product of a flat

holomorphic vector bundle with a symplectic structure group and of a
holomorphic line bundle, so that, on each patch, UA , of a good cover, a
section of H is described by a projective symplectic vector
 I
X
V := , I = 0, . . . , n, (8.85)
FJ

such that the transition functions between two different local trivializations
of H on UA and on UB have the form

VA = e−hAB SAB VB , (8.86)

where SAB is a constant matrix in Sp(2n + 2, R), hAB is a holomorphic

function;
• the Kähler potential is given by
 
K = − log iV , V  = iV T ΩV , (8.87)

where ·, · is a Hermitian and symplectic metric on H ;

• the sections satisfy

Dm V , Dn V  = 0. (8.88)

For n > 1, this last equation guarantees the existence of a prepotential in

some symplectic frame and the symmetry of the kinetic vector matrix NI J .
The transition functions are subject to the usual consistency conditions on
triple overlaps:

SAB SBC SCA = 1 = ehAB +hBC +hCA . (8.89)

8.2.2 N = 2 Hypermultiplets and Hyper-Kähler vs. Quaternionic

Kähler Geometry

In this subsection, we discuss the scalar field geometry of N = 2 hypermultiplets in

rigid supersymmetry and supergravity. Hypermultiplets are N = 2 field multiplets
with spins/helicities not exceeding 1/2. They can thus be viewed as N = 2
generalizations of N = 1 chiral multiplets. As we will now describe, one
8.2 N = 2 Supergravity 197

1
2
Q

0 0

Q
1
2

Fig. 8.1 The particle states corresponding to a CPT-completed chiral field multiplet in N = 1
supersymmetry

hypermultiplet has the same field content as two chiral multiplets, but the proper
embedding of N = 2 supersymmetry and the relation to particle multiplets are a
bit non-trivial. As we will see, this non-trivial structure straightforwardly leads to
the peculiar scalar field geometries of hypermultiplets:hyper-Kähler manifolds for
rigid supersymmetry and quaternionic Kähler manifolds for supergravity [9].
To begin with, we recall that, at the level of massless particle states, a massless
N = 1 chiral multiplet, (χ(x), φ(x)), corresponds to the direct sum of two
unitary irreducible representations of the N = 1 Poincaré superalgebra, as shown
schematically in Fig. 8.1, where we label the states by their helicities, h, and
indicate which states are superpartners with respect to the N = 1 supersymmetry
generator Q. The above two particle multiplets are CPT conjugates of one another,
and the presence of both is required to make the field theory CPT-invariant. The
transformation laws
1 1 ∗
δχL = ∂/φR , δχR = ∂/φ L (8.90)
2 2
suggest the associations

(χL , φ) ↔ (| + 1/2, |0) (8.91)

∗
(χR , φ ) ↔ (| − 1/2 , |0 ).
# #
(8.92)

Let us now turn to analogous multiplets with |h| ≤ 1/2 in N = 2 supersym-

metry. At the level of massless particle multiplets, there is only one type of N = 2
supermultiplet with helicities between +1/2 and −1/2. It consists of two scalar
particle states, one state with helicity +1/2 and one state with helicity −1/2. They
are connected by the two supersymmetry generators Q(1) and Q(2) as in Fig. 8.2.
The two spinless states |0, |0 form a doublet of the SU (2)R subgroup of the
R-symmetry U (2)R , which is possible because they are two linearly independent
vectors in a complex vector space. By analogy with the N = 1 chiral multiplet
198 8 Extended Supergravities

1
2
Q(1) Q(2)

0 SU(2)R 0

Q(2) Q(1)
1
2

Fig. 8.2 The irreducible multiplet of particle states with |h| ≤ 1/2 in N = 2 supersymmetry.
Q(1) and Q(2) denote the two independent supersymmetry generators, which are linked by the
SU (2)R subgroup of the R-symmetry group U (2)R . The two spinless states consequently form an
SU (2)R doublet

1 1
2 2
Q(1) Q(2) Q(1) Q(2)

0 SU(2)R 0 0 SU(2)R 0

Q(2) Q(1) Q(2) Q(1)

1 1
2 2

Fig. 8.3 An N = 2 hypermultiplet in four dimensions corresponds to the direct sum of two
irreducible N = 2 particle multiplets, as shown. The red states can be viewed as one CPT-
complete N = 1 chiral multiplet of the type given in Fig. 8.1 and the blue states as another
such multiplet

and Fig. 8.1, one might now be tempted to associate | + 1/2 and | − 1/2 with,
respectively, the left- and right-handed component of a Majorana fermion field, χ,
and |0 and |0 with, respectively, the superpartners φ and φ ∗ , so that N = 2
supersymmetry would ultimately be represented on a CPT-complete N = 1 chiral
multiplet. This is false, however, because the field space parameterized by the
scalars φ and φ ∗ is C ∼ = R2 , which does not permit a non-trivial representation
of SU (2). Put differently, φ and φ ∗ describe a particle and its anti-particle, which
cannot sit in one and the same SU (2)R -doublet, contrary to what the above particle
multiplet requires. We thus must not identify |0 as the anti-particle of |0 , but
instead need to add another particle multiplet of the same type as in Fig. 8.3.
8.2 N = 2 Supergravity 199

The anti-particles of | + 1/2 and |0 can then be identified with | − 1/2 and
|0
so that these four states can be viewed as one CPT-complete N = 1 chiral
multiplet, which is then accompanied by another, independent CPT-complete chiral
multiplet consisting of the states | + 1/2 , |0 , |0 , | − 1/2. Putting everything
together, a minimal N = 2 field multiplet with |h| ≤ 1/2 thus must contain two
independent Majorana fields χ 1 , χ 2 and two independent complex scalars φ 1 , φ 2 ,
which we associate with the above states as follows:

(χL1 , φ 1 ) ↔ (| + 1/2, |0) (8.93)

(χR1 , φ 1∗ ) ↔ (| − 1/2 , |0 ) (8.94)
(χL2 , φ 2 ) ↔ (| + 1/2 , |0 ) (8.95)
(χR2 , φ 2∗ ) ↔ (| − 1/2, |0 ). (8.96)

From (8.90), we then read off that (χ 1 , φ 1 ) and (χ 2 , φ 2 ) behave as two N = 1

chiral multiplets with respect to Q(1) , whereas (χ 1 , φ 2∗ ) and (χ 2 , φ 1∗ ) behave as
two N = 1 chiral multiplets with respect to Q(2) . More precisely,
1 1 (1) 1 2∗ (2)
δχL1 = ∂/φ R + ∂/φ R (8.97)
2 2
1 1 1∗ (2)
δχL2 = ∂/φ 2 R(1) − ∂/φ R , (8.98)
2 2
where the minus sign in front of the φ 1∗ term is to ensure the desired invariance
under the discrete R-symmetry (cf. (8.44)),
 (1)   (2)   1   2∗ 
  φ φ
−→ , −→ . (8.99)
 (2) − (1) φ 2∗ −φ 1

Analogously, one has

(1) (2)
δφ 1 =  L χL1 −  R χR2 (8.100)
(1) (2)
δφ 2 =  L χL2 +  R χR1 . (8.101)

Suppose now one has nH hypermultiplets with analogous (linearized) transforma-

tion laws, i.e.,

1 m (1) 1 ml (2)
δχLm = ∂/φ R + δ Eln ∂/φ n R (8.102)
2 2
1 1 ml
δχRm = ∂/φ m L(1) + δ Eln ∂/φ n L(2) (8.103)
2 2
δφ m =  (1) m mn (2) r
L χL − δ Enr  R χR (8.104)

δφ m =  (1) m mn (2) r
R χR − δ Enr  L χL , (8.105)
200 8 Extended Supergravities

where m, n, . . . = 1, . . . , 2nH , and

⎛ ⎞
e 0 ··· 0
⎜ .. ⎟  
⎜0 e .⎟ 0 1
⎜
(Emn ) := ⎜ . ⎟ = (Emn ), e= . (8.106)
. ⎟ −1 0
⎝ .. .. 0 ⎠
0 ··· 0 e

In Sect. 5.2.4, we showed that

the above
 m transformations
 for  (2) = 0 maintain their
χLm /
∂φ
structure if one acts on and with unitary (4nH × 4nH )-matrices of
χRm ∂/φ m
the form
 
A 0
U= , A ∈ U (2nH ). (8.107)
0 A∗

Preserving also the transformations with  (2) = 0 then imposes the additional
!
restriction E · A∗ = A · E, which because of A† = A−1 and the reality of E
then implies AT · E · A = E, showing that A also has to be symplectic,

A ∈ U (2nH ) ∩ Sp(2nH , C) ≡ Sp(nH )(≡ U sp(2nH )). (8.108)

Together with the SU (2)R R-symmetry subgroup, which acts on the nH doublets
 
φ 2l−1
(8.109)
(φ 2l )∗

for all l = 1, . . . , nH , we therefore have the invariance group Sp(nH ) × SU (2) of

the linearized supersymmetry transformations. This structure should be preserved
by the holonomy group of the scalar manifold.4
The supersymmetry parameters  (i) (i = 1, 2) transform as an SU (2) doublet. In
N = 2 supergravity, the supersymmetry parameters are no longer constant, and the
SU (2) part of the curvature has to be non-trivial. This is explained in more detail
in Appendix 8.B.2. Manifolds of real dimension 4nH with holonomy contained in
Sp(nH ) × SU (2) and non-trivial SU (2) holonomy are called quaternionic Kähler
manifolds. They are in general not Kähler so that they could not be coupled to just

4A priori, one only has that the tangent space group of Mhyper should allow a restriction to
Sp(nH )× SU(2) or a subgroup thereof. In other words, Mhyper has a Sp(nH )×SU(2) structure.
This implies that Mhyper admits a connection with holonomy group contained in Sp(nH )× SU(2).
A priori, this needs not be the torsion-free Levi–Civita connection, but supersymmetry invariance
of the action requires (see Appendix 8.B.2) that this must be the case. Hence it is indeed the
holonomy of the Levi–Civita connection that must be contained in Sp(nH ) × SU (2), and we can
speak of Riemannian manifolds with holonomy in Sp(nH ) × SU (2).
8.3 Extended Supergravity with N ≥ 3 201

N = 1 supergravity at the full nonlinear level. In other words, constructing an

N = 2 supergravity theory with hypermultiplets by coupling an N = 1 gravitino
multiplet to N = 1 supergravity with chiral multiplets requires (a) an even number
of chiral multiplets (so that one has 4n real scalars) and (b) a deformation of
the nonlinear couplings of the scalars among themselves, i.e., of the scalar field
geometry such that the geometry is quaternionic Kähler and in general no longer
Kähler.
In rigid N = 2 supersymmetry, by contrast, the  (i) are constants, and the SU (2)
part of the curvature must be trivial. For rigid supersymmetry, the holonomy group
of the scalar manifold is therefore contained in Sp(nH ). Such manifolds are called
hyper-Kähler. They form a subclass of Kähler manifolds that can be characterized
by the presence of three independent complex structures that satisfy the algebra of
the quaternions. The fact that they are still Kähler manifolds is consistent with the
fact that a rigid N = 2 theory is always also a particular rigid N = 1 theory. Or,
turning this around, one can infer from the fact that rigid N = 2 supersymmetry
must be a special case of rigid N = 1 supersymmetry that the SU (2) part of
the holonomy group must be trivial for hypermultiplets in rigid supersymmetry,
because otherwise the scalar manifold would be quaternionic Kähler instead of
hyper-Kähler and hence in general no longer Kähler, in contradiction with the
N = 1 requirements.

8.3 Extended Supergravity with N ≥ 3

We have seen that for N = 2 supersymmetry, the vector multiplet scalar geometry
is constrained by an interplay of symplectic duality invariance and the local
composite U (1) symmetry of the gaugini and gravitini, which is closely related
to Kähler transformations on the scalar manifold. This U (1) may be associated with
the U (1)-part of the R-symmetry group U (2) of N = 2 supersymmetry.
For the hypermultiplets, on the other hand, it is the non-trivial transformation
of the scalars under the SU (2) part of the R-symmetry group that leads to the
reduced holonomy group Hol(Mhyper) ⊂ Sp(nH )×SU (2) of a quaternionic Kähler
manifold, with the SU (2) curvature being non-trivial and of Planckian size (see
(8.207)).
For higher N supersymmetry, the multiplets get bigger, and the scalar fields
transform non-trivially under larger R-symmetry groups. By a reasoning completely
analogous to our discussion of hypermultiplets, this then leads to drastic restrictions
on the allowed holonomy groups of the scalar manifolds. In fact, for N ≥ 3, these
restrictions are so strong that they fix the entire scalar manifold once the number of
multiplets is specified, and these manifolds are certain symmetric spaces.
Let us look at this in more detail.

• N = 3: For N = 3 supersymmetry, the CPT-completed supergravity multiplet

contains the graviton, three gravitini, three vector fields, and one spin-1/2
fermion. Apart from this multiplet, there is only one type of matter multiplet,
202 8 Extended Supergravities

namely, the N = 3 vector multiplet. Its CPT-completed version has the same
field content as an N = 4 vector multiplet, namely, one vector field, four spin-
1/2 fields, and six real scalars. In rigid supersymmetry, N = 3 supersymmetry
therefore also implies N = 4 supersymmetry, but for local supersymmetry,
this is no longer true, because the supergravity multiplets for N = 3 and
N = 4 are different. The R-symmetry group of N = 3 supersymmetry is
U (3). With respect to the SU (3) subgroup thereof, the field content of a vector
multiplet splits into (cf. Fig. 8.4) a singlet vector field, Aμ , a triplet of gaugini,
λ(i) , (i, j, . . . = 1, 2, 3), one SU (3) singlet gaugino, λ(4) , and three complex
scalars, φ ij = −φ j i , which naturally transform in the antisymmetric tensor
representation of SU (3) (which is equivalent to the fundamental representation
via the epsilon tensor). The linearized supersymmetry transformation laws are

1
Q(1) Q(3)
Q(2)
1 (1) 1 (2) 1 (3) 1
2 2 2 2

Q(2) Q(1)
Q(3) Q(1) Q(2)
Q(3)
(31) (12) (23) (31)C (12)C (23)C
0 0 0 0 0 0
Q(3)
Q(3) Q(1) Q(2)
Q(2) Q(1)
1 (4) 1 (1) 1 (2) 1 (3)
2 2 2 2

Q(2)
Q(1) Q(3)
1

Fig. 8.4 A CPT-completed N = 3 vector multiplet consists of the direct sum of two irreducible
particle representations. The red states are SU (3)R singlets, whereas the blue states form obvious
SU (3)R triplets. The states of the right multiplet are the CPT conjugates of the states of the left
multiplet. The vector field Aμ is associated with | ± 1, the gaugini triplet λ(i) L corresponds to
| + 12 (i) , the gaugino singlet λ(4)
L to | + 2 , and the scalars φ to |0
1 ij (ij ) . The states can carry

non-trivial U (1)R charges, which follow from the number and chirality of the supercharges that
connect the states. This implies different U (1)R charge for the SU (3) singlet and the SU (3) triplet
of fermions, as described in (8.112). For N = 4 supersymmetry, there is a fourth supercharge
Q(4) connecting, e.g., 1 with | + 12 (4) , which then implies that the four states with helicity + 12
form an SU (4)R quartet and hence cannot carry different U (1)R -charges anymore. The U (1)R
must then be represented trivially on the vector multiplets, similar to the U (1)R for the N = 2
hypermultiplets
8.3 Extended Supergravity with N ≥ 3 203

(cf. Fig. 8.4 and the analogy with the hypermultiplet discussion)

1 ij j 1 − (i)
δλ(i)
L = ∂/φ R + γ μν Fμν L (8.110)
2 8
1
δλ(4)
L = εij k ∂/φ ij ∗ R(k) . (8.111)
4
The maximal internal symmetry group of these transformation laws is SU (3) ×
U (1) with the following U (1) charge assignments:

1 3 1
Aμ : 0, λ(i)
L : q, φ ij : +1q, λ(4)
L : − q, L(i) : q,
2 2 2
(8.112)

where q is some arbitrary real number, and charge conjugation changes the
sign of the charge. This can obviously be identified with the U (1) R-symmetry
subgroup (cf Fig. 8.4), so that the U (3) R-symmetry is actually the full invariance
group of the transformation laws of one vector multiplet. For n such vector
multiplets, there is thus just room for an additional SU (n) that rotates the
vector multiplets as a whole into each other. This is different from the N = 2
hypermultiplets, where the smaller R-symmetry group SU (2) was realized,
which leaves an additional symmetry factor Sp(1) ∼ = SU (2) even for one
hypermultiplet and Sp(nH ) instead of just SU (nH ) for nH hypermultiplets.
Correspondingly, the holonomy group of the scalar manifold of N = 3
vector multiplets must have holonomy group contained in U (3) × SU (n), with
the U (3) holonomy group being non-trivial. According to a classification of
special holonomy manifolds by Berger [10], a non-compact real 6n-dimensional
manifold with this property must be equivalent to

SU (3, n)
M N =3 = . (8.113)
S(U (3) × U (n))

• N = 4: For N = 4 supersymmetry, the supergravity multiplet now also

contains two real scalar fields in addition to the graviton, four gravitini, six
vector fields, and four spin-1/2 fermions. The N = 4 vector multiplet has
the same field content as the N = 3 vector multiplet just discussed. Now,
however, the vector multiplet is CPT self-conjugate, i.e., it does not have to be
completed by a CPT conjugate multiplet. This means that the six real scalars,
which transform in the six-dimensional, real representation of the R-symmetry
subgroup SU (4) ∼ = SO(6) must be their own anti-particles and hence cannot
carry a non-trivial charge with respect to the U (1) part of the U (4) R-symmetry
(4)
group. Likewise, different charge assignments for λL and the other gaugini as in
(8.112) are no longer possible as that would break SU (4). This can also be seen
from the linearized supersymmetry transformations, which are obtained from the
204 8 Extended Supergravities

N = 3 transformations by obvious SU (4) covariantization,

1 ij (j ) 1 μν − (i)
δλ(i)
L = ∂/φ R + γ Fμν L (8.114)
2 8

with i, j, . . . , = 1, 2, 3, 4, and the reality condition (φ ij )∗ = 12 εij kl φ kl , which

obviously does not allow any U (1) charge assignments for the scalars anymore.
This is similar to the N = 2 hypermultiplet scalars being neutral with respect
to the U (1) subgroup of the R-symmetry group U (2). The holonomy group
of the real 6n-dimensional scalar manifold of n vector multiplets in N = 4
supergravity must therefore be contained in SO(6) × SO(n), with the SO(6)
factor being non-trivial and the SO(n) arising here as the maximal subgroup of
SO(6n) that commutes with the action of SO(6) on the scalars. SO(n) simply
rotates the different multiplets into each other. Berger’s list then fixes the scalar
manifolds of the vector multiplets to be of the form SO(6, n)/(SO(6)×SO(n)).
Together with the scalars from the supergravity multiplet, which parameterize
SU (1, 1)/U (1), we therefore have the scalar manifold

SU (1, 1) SO(6, n)
M N =4 = × . (8.115)
U (1) SO(6) × SO(n)

• N = 5: For N = 5 supersymmetry, there are no longer matter multiplets with

helicities between 1 and −1, and the only supermultiplet one could use is the
supergravity multiplet. In its CPT-completed version, this multiplet contains one
graviton, five gravitini, ten vector fields, eleven spin 1/2 fields, and ten real scalar
fields. The holonomy group must be U (5), which then leaves the space

SU (5, 1)
M N =5 = . (8.116)
U (5)

• N = 6: Here, the CPT-completed supergravity multiplet contains the graviton,

6 gravitini, 16 vector fields, 26 spin 1/2 fermions, and 30 real scalar fields. The
holonomy group of the scalar manifold must be the R-symmetry group U (6),
which leaves as scalar manifold

SO ∗ (12)
M N =6 = . (8.117)
U (6)

• N = 8: As the CPT-completed N = 7 supergravity multiplet is equivalent

to the CPT-self-conjugate N = 8 supergravity multiplet, there is no N = 7
supergravity theory. The N = 8 supergravity multiplet contains the graviton,
8 gravitini, 28 vector fields, 56 spin 1/2 fermions, and 70 real scalars. As this
multiplet is CPT-self-conjugate, the scalars must be neutral with respect to the
U (1) subgroup of the U (8) R-symmetry group and only transform non-trivially
with respect to the SU (8) subgroup, which must therefore also be the holonomy
8.A Appendix: Details on and Origin of Local Special Kähler Geometry 205

group of the scalar manifold. This scalar manifold is given by

E7(7)
M N =8 = . (8.118)
SU (8)

8.A Appendix: Details on and Origin of Local Special Kähler

Geometry

In this appendix, we would like to understand the origin of the differences between
rigid and local special Kähler geometry that are listed in Sect. 8.2.1.2. Our goal
herein is not so much mathematical completeness or a full derivation of all terms
and prefactors in the Lagrangian, but rather to emphasize the geometrical roots of
these differences and to exhibit them as directly as possible and without introducing
additional formalism [2, 4–8].

8.A.1 The Symplectic Section and Its Kähler Transformation

We begin with the necessity of the existence of a scalar field-dependent symplectic

section, v. Just as for rigid special Kähler geometry, this necessity can be deduced
from the requirement that the transformation laws of the gaugini should be
symplectically invariant and covariant with respect to holomorphic scalar reparame-
terizations. More precisely, if we again consider the supersymmetry transformation
with respect to the original N = 1 supersymmetry parameter  (1), the analogy with
the rigid transformations (8.63), (8.64) suggests that in supergravity one has

m(1) 1 m (1)
δ (1) λL = (∂/z )R (8.119)
2
m(2) 1 J − μν (1)
δ (1) λL = − g mn fnI (ImNI J )Fμν γ L . (8.120)
4
Here, NI J (z, z) is the scalar field-dependent gauge kinetic matrix of the vector field
Lagrangian (8.39), and fnI (z, z) is a scalar field-dependent object that transforms as
a holomorphic covector, fnI '→ (∂zm /∂ zn )fmI , under scalar field reparameteriza-
tions and forms the upper part of a symplectic vector
 
fnI (z, z)
Un (z, z) = . (8.121)
hJ n (z, z)

To be most general as possible, we allow for a possible non-holomorphic zm -

dependence, but, as in rigid supersymmetry, we do expect that also Un will be related
to a suitable derivative of another symplectic (and possibly non-holomorphic)
section V (z, z) with respect to zn . In order to ensure the symplectic invariance of
206 8 Extended Supergravities

(8.120), however, we only need to require

fnI NI J = hJ n , (8.122)

because then

J− 1  J−
 1
fnI ImNI J Fμν = hJ n Fμν − fnI GμνI − = F− , Un , (8.123)
2i 2i μν
which is manifestly symplectically invariant.
As mentioned above, we expect that Un be related to a suitable derivative of a,
possibly non-holomorphic, symplectic section, which we denote as
 
LI (z, z)
V (z, z) = , (8.124)
MJ (z, z)

and which will be closely related to the advertised holomorphic section V (z)
mentioned in item 1.
In order to understand this better, it is necessary to address also item 2 in
Sect. 8.2.1.2, namely, the claim that in supergravity the scalar field-dependent
symplectic sections have to transform non-trivially under Kähler transformations
(see (8.81) for V (z)). The root of this behavior is the particular non-trivial Kähler
transformation property of the fermions we already discovered in N = 1 super-
gravity. Repeating the arguments of Sect. 6.1.2 for each of the two supersymmetries,
(1) (2)
one now finds that the two gravitini ψμ , ψμ have to transform under Kähler
transformations as (cf. (6.18))
 
i
ψμ(j ) '→ exp − (Im h)γ5 ψμ(j ) , (j = 1, 2) (8.125)
2

which immediately implies

 
i
 (j )
'→ exp − (Im h)γ5  (j ) , (j = 1, 2) (8.126)
2

(j )
via δψμ = Dμ  (j ) + . . ., where  (i) are the two Majorana spinor parameters.
(j )
This behavior is consistent with the SU (2)R R-symmetry rotating the two ψμ
and the two  (j ) into each other and requires the introduction of Kähler covariant
derivatives (cf. (6.16), (6.21)),

(j ) (j ) i (j )
D[ν ψρ] ≡ D[ν (ω)ψρ] + Q[ν γ5 ψρ] (8.127)
2
i
Dν  (j ) ≡ Dν (ω) (j ) + Qν γ5  (j ) (8.128)
2

with Qν (z, z) ≡ i/2[(∂n K)∂ν zn − (∂m K)∂ν zm ].

8.A Appendix: Details on and Origin of Local Special Kähler Geometry 207

While this is completely analogous to N = 1 supergravity, an important

difference occurs for the gaugini. If we again identify  (1) with the original
N = 1 supersymmetry parameter and (λm(1) , λm(2) ) = (χ m , λm /2), we expect
a supersymmetry transformation rule with respect to  (1) of the form (cf. (8.119),
(8.120))

m(1) 1 m (1)
δλL = ∂/z R (8.129)
2
m(2) 1 J − μν (1)
δλL = − g mn fnI ImNI J Fμν γ L . (8.130)
4
From the first equation and (8.126), we deduce
 
m(1) i m(1)
λL '→ exp + (Im h) λL , (8.131)
2

m(2)
which must then also hold for λL due to the assumed SU (2)R R-symmetry.
This, however, seems to be in conflict with the second equation (8.130), which
naively would imply a transformation with a minus sign in the exponent due to the
appearance of L(1) instead of R(1). In N = 1 supergravity, we encountered a similar
problem for the supersymmetry transformations of the fermions in the presence of a
non-trivial superpotential (see the discussion below Eq. (6.30)). This problem could
be solved by assuming that the superpotential transforms non-trivially under Kähler
transformations. To solve the analogous problem in (8.130), one likewise requires
that the field-dependent term in front of γ μν transform non-trivially under Kähler
transformations. In this term, the inverse scalar field metric g mn is by construction
Kähler invariant, and a Kähler transformation of NI J and/or Fμν J − would lead

to Kähler non-invariances in the gauge field sectors. This leaves fnI as the only
reasonable candidate for a non-trivial Kähler transformation, which then has to be
of the form

fnI '→ exp [−i(Imh)] fnI ⇐⇒ Un '→ exp [−i(Imh)] Un (8.132)

in order to yield the correct Kähler transformation of the right-hand side of (8.130).
Equation (8.132) now also implies that Un cannot be just the partial derivative
of a symplectic section, V , as that would also produce derivatives of h in (8.132).
Instead, Un can at most be a Kähler covariant derivative,
 
1
Un = Dn V ≡ ∂n + (∂n K) V , (8.133)
2

where
 
LI (z, z)
V = V (z, z) = (8.134)
MJ (z, z)
208 8 Extended Supergravities

is a symplectic section that transforms as

V '→ exp [−i(Imh)] V (8.135)

under Kähler transformations. This transformation property of V depends on h(z)

and h(z), which implies that V cannot depend holomorphically on the coordinates
zm of the scalar manifold, as indicated in (8.134).
We can however define a modified symplectic section, v, that does transform
holomorphically under Kähler transformations,
 
XI (z) K(z,z)
V (z) = := e− 2 V (z, z) ⇒ V (z) '→ exp [−h(z)] V (z)
FJ (z)
(8.136)

and hence may sensibly be restricted to have a holomorphic coordinate dependence,

as indicated. We thus have
 K   1  K  K
  K
Un = Dn e 2 V ≡ ∂n + (∂n K) e 2 v = e 2 ∂n + (∂n K) v ≡ e 2 Dn V
2
(8.137)

so that, in particular,
K K
 
fnI = e 2 Dn XI ≡ e 2 ∂n + (∂n K) XI (8.138)

and similarly for hJ n . We finally note that the holomorphicity of v, ∂n v = 0, is

equivalent to
 
1
Dn V ≡ ∂n − (∂n K) V = 0. (8.139)
2

8.A.2 The Kähler Potential

After having discussed the necessity of a holomorphic symplectic section, V (z), we

now come to the explicit form of the Kähler potential (8.82) mentioned in item 3
of Sect. 8.2.1.2. The analogy with Eq. (8.51) in rigid supersymmetry suggests that
the Kähler potential should be related to the real symplectic invariant iV , V  =
i(XI FI − XI FI ). As opposed to the rigid case, however, we now have the behavior
(8.136) under Kähler transformations, which implies

iV , V  −→ e−(h+h) iV , V . (8.140)

8.A Appendix: Details on and Origin of Local Special Kähler Geometry 209

Hence, iV , V  does not transform the right way to serve as the Kähler potential, but
the logarithm of it would do the job. We are thus led to the logarithmic expression
(8.82).
Let us summarize what we have shown so far. The scalar manifolds, Mvec ,
of vector multiplets in N = 2 supergravity are Kähler manifolds that allow
for a symplectic section, V (z, z) = eK(z,z)/2 V (z), that satisfies the following
properties:

• ∂m v = 0 ⇐⇒ Dm V = 0
• K = − log iV , V  ⇐⇒ V , V  = i
' e−h(z) V (z) under Kähler transformations,
• V (z) →

where it is understood that V (z) must be such that the resulting metric gmn =
∂m ∂n K is positive definite for the zm -domain of interest.
Geometrically, the above Kähler transformation property of V (z) means that
V (z) is a section not only in an Sp(2(nV + 1), R) bundle but, similar to the
superpotential in N = 1 supergravity, also in a holomorphic line bundle. This
line bundle obeys a topological restriction such that the (special) Kähler manifold
must actually be a (special) Hodge-Kähler manifold, as we discussed for N = 1
supergravity in Sect. 6.2.

8.A.3 The Existence of a Prepotential

The properties listed at the end of the previous subsection are still not sufficient
to allow for a consistent embedding into N = 2 supergravity and have to be
supplemented by the additional requirement

Dm V , Dn V  = 0, (8.141)

where Dn V = (∂n + (∂n K))V . Due to the antisymmetry of the symplectic product,
this condition is antisymmetric in m and n, and hence non-trivial only for nV > 1,
where (8.141) is necessary for the symmetry of the gauge kinetic matrix NI J and
ensures its uniqueness, as we will discuss in Appendix 8.A.4. Moreover, for nV > 1,
(8.141) implies that there is always a symplectic transformation that brings V (z)
into the familiar prepotential form
 
XI ∂F (X)
V = , with FJ = (8.142)
FJ ∂XJ

for a suitable holomorphic function F (X) that exists at least locally and, unlike
in rigid supersymmetry, is also restricted to be homogeneous of degree two. A
simple way to understand the homogeneity requirement is provided by the behavior
XI '→ e−h XI and FJ '→ e−h FJ , which is consistent only if the FJ = ∂F /∂XJ are
210 8 Extended Supergravities

homogeneous functions of the XI of degree one, implying homogeneity of degree

two for F (X).
The generic necessity of performing a suitable symplectic rotation in super-
gravity before the prepotential form (8.142) is achieved is an important difference
to rigid special Kähler geometry, where the section is of a prepotential form in
any symplectic duality frame. This difference is ultimately related to the fact
that in supergravity the (nV + 1) components XI cannot be swapped for the
nV independent coordinates zm . This is best illustrated by a simple example [7].
Choosing, e.g., nV = 1 with F (X) = −iX0 X1 and special coordinates X0 =
1, X1 = z, the section V (z) in the form (8.142) is
⎛ ⎞
1
⎜ z ⎟
V (z) = ⎜ ⎟
⎝ −iz ⎠ (8.143)
−i

which leads to the metric gzz = (z + z)−2 corresponding to the manifold Mvec =
SU (1, 1)/U (1). Performing now a symplectic transformation of the form
⎛ ⎞ ⎛ ⎞
10 0 0 1
⎜0 0 0 −1 ⎟ ⎜ ⎟
' SV = ⎜
V → ⎟V = ⎜ i ⎟, (8.144)
⎝0 0 1 0 ⎠ ⎝ −iz ⎠
01 0 0 z

we see that the transformed section can no longer be of the form (8.142), as the lower
two components cannot be functions of the upper two components. The prepotential
form thus in general only holds in some duality frames but not in all of them. If one
does not impose (8.141), on the other hand, one could even write down sections
V (z) that can never be rotated into a prepotential form (8.142), no matter what
symplectic transformation one uses. The proof that (8.141) ensures the existence of
a prepotential basis is a bit technical [2, 4] and will therefore not be repeated here.
Instead we content ourselves with showing the opposite direction, namely, that in a
basis (8.142) with second-degree prepotential, condition (8.141) follows. As (8.141)
is automatically true for nV = 1, we assume nV > 1 unless stated otherwise. We
start by showing that (8.141) is implied by5

V , ∂m V  = 0. (8.145)

5 For nV > 1, (8.145) also follows from (8.141). To see this, one acts with Dl on (8.141), where it
just acts as ∂l due to the Kähler weights of Dm V . This results in gml V , ∂n V +gnl ∂m V , V  = 0,
which implies (8.145) upon contraction with g lp and p = m = n.
8.A Appendix: Details on and Origin of Local Special Kähler Geometry 211

To see this, we first note that V , V  = 0 due to antisymmetry of the symplectic

product, so that (8.145) can be equivalently written as

V , Dm V  ≡ V , (∂m + (∂m K))V  = 0. (8.146)

Acting with Dn and antisymmetrizing in n and m and using (8.145) and V , V  = 0

then imply (8.141). Using now the prepotential form (8.142), one finds
 
∂XJ
V , ∂m V  = XI (∂m FI ) − FI (∂m XI ) = XI ∂J ∂I F − ∂m F
∂zm
∂XJ I ∂XJ
= X ∂I (∂J F ) − ∂m F = ∂J F − ∂m F = 0, (8.147)
∂zm ∂zm

where we used that ∂J F = XI ∂I ∂J F is homogeneous of degree one. This then

implies (8.141).
Before we come to the discussion of the gauge kinetic matrix and its relation to
(8.141), we comment on the case nV = 1. In that case, (8.141) is empty and hence
does not imply the existence of a prepotential. If one uses instead the (for nV = 1
non-equivalent) condition (8.145), on the other hand, the existence of a prepotential
basis can be proven in a similar way as for nV > 1 [2]. However, as the condition
(8.145) is no longer implied by the required symmetry of the gauge kinetic matrix
NI J (see below), there seems to be no physical reason that enforces (8.145). And
indeed, in [4], models without an action and one vector multiplet were constructed
where (8.145) does not hold.

8.A.4 The Gauge Kinetic Matrix

We finally come to the discussion of the gauge kinetic matrix and its relation to
condition (8.141). More precisely, we would like to show that (8.141) is necessary
for a sensible gauge kinetic matrix, NI J , for nV > 1, and that, in a prepotential
basis, this matrix can be written as announced in (8.84). We already have derived
the constraint

fmI NI J = hJ m , (8.148)

which follows from the requirement that the gaugino transformation law (8.120) be
symplectically invariant and covariant with respect to scalar reparameterizations. As
we will now show, the gravitino transformation law imposes the additional condition

LI NI J = MJ . (8.149)

Note that there is no complex conjugation involved in this equation, as opposed

to (8.148). In order to motivate this equation, we recall the rigid supersymmetry
212 8 Extended Supergravities

transformations of the N = 1 gravitino multiplet discussed in Sect. 2.2.2,

−
δψμL = −γ ρ R Fμρ (8.150)

δAμ = ψμ , (8.151)

where we used γμνρ = iμνρσ γ σ γ5 to write the gravitino transformation in terms

of the anti-self-dual field strength.
These transformations leave invariant the Lagrangian

1 1
L = − ψ μ γ μνρ ∂ν ψρ − F μν Fμν . (8.152)
2 4
This theory should be embeddable in pure N = 2 supergravity by making
the supersymmetry parameter x-dependent and by coupling it to the N = 1
supergravity multiplet in the usual way. This will introduce a derivative of  as
an additional term into the gravitino transformation law and require spacetime
covariantization, but as there are no scalar fields involved, there is not much room for
a modification of the above existing terms from rigid N = 1 supersymmetry. And
indeed, one can write the above transformation laws and the Lagrangian formally in
terms of local special Kähler geometry for the special case nV = 0. To this end, we
use the prepotential F (X0 ) = − 2i (X0 )2 , which leads to
   
X0 X0
V = = . (8.153)
F0 −iX0

This implies via (8.82) and (8.84)

K = − log |X0 |2 , N00 = −i (8.154)

so that, after choosing X0 = 1, one has V = (L0 , M0 )T = (1, −i)T and hence

δψμL = L0 ImN00 Fμρ

0− ρ
γ R . (8.155)

This expression is symplectically invariant, because

L0 N00 = −i = M0 (8.156)

0− = (1/2i)(M F 0− − L0 G
implies L0 ImN00 Fμρ 0 μρ μρ0− ) = (1/2i)Fμν , V .
For N = 2 supergravity coupled to nV vector multiplets, the gravitino
transformation law should then contain the analogous term

(1) J− ρ (2)
δψμL ∝ LI ImNI J Fμρ γ R , (8.157)
8.A Appendix: Details on and Origin of Local Special Kähler Geometry 213

because this ensures that both sides transform with e− 2 (Imh) and are symplectic
i

invariant provided the analogue of equation (8.156) holds, which is just (8.149).
The gauge kinetic matrix NI J thus has to obey the two constraints (8.148)
and (8.149). In fact, these conditions determine NI J uniquely. To see this, we
observe from the gaugino transformation (8.130) and the gravitino transformation
m(1)
(8.157) with respect to  (2) that the nV independent gaugini λL gmn transform into
J − , whereas the gravitino ψ
fnI ImNI J Fμν
(1)
μ transforms into the linear combination
L ImNI J Fμν− . For any fixed scalar field value zm , these must therefore be (nV + 1)
I I

linearly independent combination of vector field strengths. This implies that the
following ((nV + 1) × (nV + 1))-matrix, A, must be invertible:

A := (fmI , LI ), (8.158)

where I is the row index and m labels the first nV columns, whereas LI is the last
column. Defining similarly

B := (hI m , MI ), (8.159)

Eqs. (8.148) and (8.149) can be written as

AT · N = B T , (8.160)

so that, with the invertibility of A, one has a unique expression for the gauge kinetic
matrix

N = AT −1 · B T . (8.161)

This has to be symmetric, N − N T = 0, which is then equivalent to

 
Um , Un  Un , V 
B T A − AT B = 0 ⇐⇒ = 0. (8.162)
V , Un  V , V 

This implies

Um , Un  = 0, V , Un  = 0. (8.163)

The second of these equations simply follows from taking a covariant derivative
of V , V  = i, whereas the first one is equivalent to (8.141), which implies the
existence of a basis with a prepotential for nV > 1, as discussed earlier.
214 8 Extended Supergravities

It remains to verify that in a basis with a prepotential NI J takes the form (8.84).
To see this, we use

Dm V , V  = 0 (8.164)

(∂m XI )FI K = ∂m FK (8.165)

XI FI K = FK (8.166)

fnI FI K = hKm . (8.167)

The first of these equations is valid in any symplectic frame and equivalent to
Dm V , V  = 0, which follows from V , V  = −i and Dm V = 0. The second
uses the prepotential relations FI K = ∂J ∂K F and FK = ∂K F , which also imply
the third equation due to the homogeneity of F of degree two. The fourth equation
finally follows from the second and third upon using fnI = eK/2 (∂m + (∂m K))XK
and hKm = eK/2(∂m + (∂m K))FK .
We know that NI J is uniquely specified by (8.148) and (8.149), so it suffices to
show that (8.84) satisfies these two equations in a prepotential basis. We begin with
(8.148) and first show that

2ifmI (ImFI K )XK = fnI FI K XK − fnI FI K XK (8.168)

(8.166),(8.167) I (8.164)
= fm FI − hKm XK = Dm V , V  = 0
(8.169)

The second term on the right-hand side of (8.84) thus does not contribute to (8.148),
which then follows from fmI NI K = fmI FI K = hKm .
Equation (8.149), finally, follows from (LI , MK ) = eK/2 (XI , FK ) and (8.166),
with the non-antiholomorphic piece in NI J now playing a crucial role. This term is
thus the unique extension of FI J that ensures (8.149) (i.e., ultimately the symplectic
invariance of the gravitino transformation law) without destroying (8.148) (i.e., the
symplectic invariance of the gaugino transformation law).

8.B Appendix: Quaternionic-Kähler vs. Hyper-Kähler

Manifolds of Hypermultiplets

In this appendix we show in more detail how the invariance of the action implies a
non-trivial SU (2) curvature for hypermultiplets in N = 2 supergravity and hence
leads to the difference between hyper-Kähler geometry and quaternionic-Kähler
geometry [9].
8.B Appendix: Quaternionic-Kähler vs. Hyper-Kähler Manifolds. . . 215

8.B.1 Sp(nH ) × SU (2)-Adapted Vielbein

To begin with, we consider the real 4nH -dimensional scalar manifold, Mhyper, of
nH N = 2 hypermultiplets and parameterize it locally by 4nH real scalar fields q X
(X, Y, . . . = 1, . . . , 4nH ). Denoting an ordinary vielbein on Mhyper by fΓX (q) with
flat SO(4nH )-indices Γ, Δ, . . . = 1, . . . 4nH , the metric components, hXY (q), can
be expressed as

hXY (q) = fΓX (q)δΓ Δ fΔ

Y (q). (8.170)

We now fix a particular point p ∈ Mhyper and switch to a new basis of the vielbein
at that point p. To this end, we split the vielbein components into two sets,
 
f1A
(fΛ
X) = X , (A, B, . . . = 1, . . . , 2nH ). (8.171)
f2A
X

We then make a basis change with the non-orthogonal matrix (cf. (8.106) for the
definition of E)
   
1 i12nH −iE −1 1 −i12nH 12nH
M := √ ⇔M = √ , (8.172)
2 12nH E 2 −iE −E

which satisfies
 
0 E
MT M = e ⊗ E = , (M −1 )∗ = (e ⊗ E)M −1 . (8.173)
−E 0

More precisely, we define a new vielbein at p,

 
fX1A
fX = , (8.174)
fX2A

by

fX1 = √1 (−if1 + f2 )
X X
−1 2
fX = M fX ↔ (8.175)
fX2 = √ (−iEf − Ef2 )
1 1
X X
2

In terms of this new vielbein, we then have, with Eq. (8.173),

jB
hXY = fTX fY = fXT M T MfY = fXT (e ⊗ E)fY = fXiA eij EAB fy , (8.176)
216 8 Extended Supergravities

as well as the reality condition

 ∗
fX1A
fX∗ = = M −1∗ fX = (e ⊗ E)M −1 fX = (e ⊗ E)fX (8.177)
fX2A

i.e.,

(fXiA )∗ = fX ej i EBA =: fXiA .

jB
(8.178)

Equation (8.176) tells us that the new vielbein fXiA is adapted to an Sp(nH )×SU (2)
split of the tangent space at the point p. In Sect. 8.2.2, we used the linearized
supersymmetry transformations (8.102), (8.103) to argue that the tangent space
group of the scalar manifold, Mhyper, of nH N = 2 hypermultiplets should be
contained in Sp(nH ) × SU (2) ∈ SO(4nH ). In other words, the above Sp(nH ) ×
SU (2) covariant form of the vielbein at p can actually be extended smoothly across
all of Mhyper. We can thus consistently use a vielbein of the adapted form fXiA for
all points.
The reality condition (8.178), on the other hand, indicates that it is natural to
denote complex conjugation by lowering the SU (2) indices, i, j, . . ., with eij and
the Sp(nH ) indices, A, B, . . ., by EAB according to the convention

V i = eij Vj , V i = V j ej i ; V A = E AB VB , VA = V B EBA
(8.179)

j
where eij eik = δk and E AB EAC = δCB .
Before we come to the supersymmetry transformation laws, we note two
additional contraction identities,

fXiA fYj A + fYiA fXj A = hXY δji (8.180)

hXY fXiA fYj B = δji δBA (8.181)

j∗ j∗
The first of these is equivalent to fXiT fY +fYiT fX = hXY δji , which may be verified
by using (8.175), the reality of fX , and f1T
X fY + fX fY = fX fY = hXY . The second
1 2T 2 T
†
identity is equivalent to h fX fY = 14nH = 12 ⊗12nH , which follows from (8.175),
XY

the reality of fX , and the relations hXY fX fTY = 14nH and M −1 M −1† = 14nH .
Remark: In the literature, one often also finds the statement that the vielbein fXiA
satisfies the additional relation
1
fXiA fY iB + fYiA fXiB = hXY δBA . (False!) (8.182)
nH
8.B Appendix: Quaternionic-Kähler vs. Hyper-Kähler Manifolds. . . 217

This equation, however, is false, except for the special case nH = 1. This can be
X f Y lD , which, using (8.181),
verified in many ways, e.g., contracting (8.182) with fkC
results in
1 D A
δCA δBD + E AD EBC = δ δ . (8.183)
nH C B

Choosing A = C = D = B such that E AD = 0 (e.g., C = 1, D = B = 4 for

E = 1nH ⊗ e), one finds a contradiction. Alternatively, one can contract (8.182)
with fZkB and use (8.180), (8.182), and again (8.180) to arrive at
     
1 1 1
1− hY Z fX + 1 −
kA
hXZ fY + 1 −
kA
hXY fZkA = 0 (8.184)
nH nH nH

At the origin of Riemannian normal coordinates with hXY = δXY , one can then take
X = Y = Z to infer that, for nH > 1, fZkA = 0 for all Z = X = Y . Contracting
this with fZ  kA and setting Z = Z  then would give the contradiction hZZ = δZZ =
0. Yet another way to show that (8.182) cannot hold for nH > 1 is to go to the
origin of Riemannian normal coordinates with hXY = δXY and fΓX = δX Γ and to

use directly (8.175) for, e.g., X = Y = 1. This would lead (using E = 1nH ⊗ e)
to δ1A δ1B − δ AC EC1 E1D δ DB = δ1A δ1B + δ2A δ2B = (1/nH )δBA , which is not a valid
identity for nH > 1.

8.B.2 Holonomy and Curvature

Up to now, we have shown that the 4nH -dimensional scalar manifold, Mhyper,
of nH hypermultiplets in N = 2 supersymmetry admits a vielbein, fXiA (q),
that is adapted to the Sp(nH ) × SU (2) structure of MHyper and that satisfies
(8.176), (8.178), (8.180), and (8.181). We also saw that the reality condition
(8.178) motivates the convention that complex conjugation raises or lowers the
SU (2) index i and the Sp(nH )-index A. According to our discussion around
(8.102) and (8.103), the left-handed chiral fermions of hypermultiplets and the left-
handed supersymmetry parameter transform in the fundamental representations of,
respectively, Sp(nH ) and SU (2), so that we likewise use the analogous convention
that charge conjugation lowers the corresponding indices, i.e., we set

ζ A := χLA , ζA := χRA (8.185)

(i) (i)
 i := L , i := R , (8.186)

where, in contrast to the vielbein convention (8.178), the lower indices should not
be thought of as arising from lowering the indices with eij or EAB , as that cannot
change the chirality.
218 8 Extended Supergravities

The proper generalization of the linearized transformation laws (8.102)–(8.105)

can then be written in the form
i iA
δζ A = f (q)∂/q X i (8.187)
2 X
i
δζA = − fXiA (q)∂/q X  i (8.188)
2
δq X = −ifiA
X
(q) i ζ A + if XiA (q) i ζA . (8.189)

In fact, by setting
  √ 
qX X
√ 2δX
m Re φ
m
(q ) =
X
 =  (8.190)
qX 2δm Im φ m
1 m X m X
⇐⇒ φ m = √ (δX q + iδX q ), (8.191)
2

where X, Y , . . . = 1, . . . , 2nH , and X , Y  , . . . = 1, . . . , 2nH , and by going to the

origin of Riemann normal coordinates with the choice fΓX = δX Γ , i.e., f1A = δ A ,
X X
fX = δX and fX = fX = 0, the relation (8.175) brings (8.187)–(8.189) to the
2A A 2A 1A

form (8.102)–(8.105).
Using (8.189), the commutator of two supersymmetry transformations of q X is,
to lowest order in fermion fields,

1 X jA
[δη , δ ]q X = (f f + f Xj A fY iA )( i γ μ ηj )∂μ q Y + c.c., (8.192)
2 iA Y

which is equal to the desired result 12 ( i γ μ ηi )∂μ q X + c.c. precisely when (8.180)
holds. For the fermions ζ A , on the other hand, one uses the Fierz identities

1
i ηj = − ηj γ ν i γν PL (8.193)
2
1 1
i ηj = − ηj i PR + ηj γνρ i γ νρ PR (8.194)
2 4

to obtain for the terms involving ∂μ ζ A ,

1
[δη , δ ]ζ A = − fXiA fjXB (ηj γ ν i )γ μ γν ∂μ ζ B
4
 
1 1
+ fXiA f Xj B ηj i γ μ ∂μ ζB − ηj γνρ i γ μ γ νρ ∂μ ζB − ( ↔ η)
4 2
8.B Appendix: Quaternionic-Kähler vs. Hyper-Kähler Manifolds. . . 219

μ
Using the linearized field equation ∂/ζ A = 0 and γ μ γν = −γν γ μ + 2δν , this
becomes
1
[δη , δ ]ζ A = − fXiA fjXB (ηj γ μ i −  j γ μ ηi )∂μ ζ B
2
1
− fXiA f Xj B ηj γνρ i −  j γνρ ηi γ μ γ νρ ∂μ ζB .
8
The spinor bilinear in the second line is symmetric under exchange of i and j due to
(1.44) and hence vanishes upon contraction with the vielbein terms due to (8.181),
which also yields the desired result 12  i γ μ ηi ∂μ ζ A +c.c. for the first line. This shows
that (8.180), (8.181) essentially ensure the closure of the supersymmetry algebra and
that an equation of the (incorrect) form (8.182) is indeed not required.
We now come to the final part and discuss the invariance of the action. This will
tell us that the connection compatible with the Sp(nH )×SU (2) structure of MHyper
is actually the torsion-free Levi–Civita connection and that the SU (2) curvature
must be non-trivial in supergravity.
Starting point is the kinetic term of the hypermultiplet fields,

1
e−1 Lhyper,kin = e−1 (Lscalar + Lfermion) = − hXY (q)∂μ q X ∂ μ q Y − 2ζ Dζ
A
/ A,
2
(8.195)

where Dμ is the spacetime and Sp(nH ) covariant derivative with respect to an

Sp(nH ) connection compatible with the Sp(nH ) × SU (2) structure on Mhyper.
In the following, we will in general denote by Dμ the covariant derivative with
respect to spacetime coordinate transformations, local Lorentz transformations,
local composite Sp(nH ) × SU (2) transformations, and general field reparameteri-
zations q X → q X (q).
We first consider the supersymmetry variation of the scalar kinetic term with
respect to the scalar fields,
  1 1
δ e−1 Lscalar = − hXY,Z δq Z ∂μ q X ∂ μ q Y − hXY ∂μ (δq X )∂ μ q Y
2 2
1
− hXY ∂μ q X ∂ μ (δq Y )
2
1
= −hXY,Z + hZY,X + hXZ,Y δq Z ∂μ q X ∂ μ q Y
2
  
W h
ΓXY WZ

+hXY δq X ∇μ ∂ μ q Y
= hXY δq X Dμ ∂ μ q Y , (8.196)
220 8 Extended Supergravities

where we used partial integration between the first and second line and Dμ ∂ μ q Y =
Y ∂ q X ∂ μq Z = ∂ q Y + Γ μ ∂ ν q Y + Γ Y ∂ q X ∂ μq Z .
∇μ ∂ μ q Y + ΓXZ μ μ μν XZ μ
Now consider the variation of the fermionic term due to the variation of the
fermions,

δ[e−1 Lfermion ] = −2ζ D(δζ

A
/ A ) − 2δζ A Dζ / A = −2ζ A D(δζ
/ A ) − 2ζ A D(δζ
/ A
)
 
A i
= −2ζ D / − fXiA ∂/q X  i + c.c
2
   
A A
= iζ γ μ (Dμ fXiA )(∂/q X ) i + c.c + iζ fXiA (D/ ∂/q X ) i + c.c.
 
A
+ iζ γ μ fXiA (∂/q X )Dμ  i + c.c. , (8.197)

where we used partial integration and (1.44) in the first line. Using γ μ γ ν = γ μν +
ρ
g μν and that D[μ ∂ν] q X = 0 due to the symmetry of the Christoffel symbols Γμν
Z
and ΓXY , the second term becomes

ifXiA ζ  i Dμ ∂ μ q X + c.c. = −hXZ δq Z Dμ ∂ μ q X = −δ[e−1 Lscalar ],

A
(8.198)

and hence
 
δ[e−1 Lhyper,kin ] = iζ γ μ (Dμ fXiA )(∂/q X ) i + c.c + [Ji Dμ  i + c.c.]
A μ

(8.199)

with the supercurrent

μ A
Ji := ifXiA ζ γ μ ∂/q X . (8.200)

The vanishing of the first term in (8.199) requires

jA
DX fYiA ≡ ∂X fYiA − ΓXY
Z
fZiA + fY ωXj i + fYiB ωXB A = 0, (8.201)

where DX contains the Sp(nH ) and SU (2) connections, ωXA B (q) and ωXi j (q),
and the Christoffel connection, ΓXYZ (q), on M
hyper , as indicated. Just as for the
spin connection (cf. (3.18) in Sect. 3.2), this equation is simply the statement that
the Sp(nH ) × SU (2)-compatible connection is equivalent to the torsion-free Levi–
Civita connection, and we have Sp(nH ) × SU (2) holonomy with respect to the
Levi–Civita connection, as announced earlier.
To eliminate the second term in (8.199), we add the usual Noether term to the
Lagrangian,
μ
LNoether = −Ji ψμi + c.c. (8.202)
8.B Appendix: Quaternionic-Kähler vs. Hyper-Kähler Manifolds. . . 221

for which the gravitino variation δψμi = Dμ  i + . . . just leads to the negative of the
last term in (8.199). Just as in our discussion for N = 1 supergravity, the variation
μ
of the matter fields in Ji itself leads to terms involving the energy momentum
tensor and is cancelled by variations of the metric but also to a term involving an
antisymmetrized product of three gamma matrices that cannot be so absorbed. It is
instead cancelled by the non-trivial composite connection terms (here the composite
SU (2) connection) in the gravitino variation of the kinetic term of the gravitini. To
see how this works in detail, we use again γ μ γ ν = γ μν + g μν and (1.44) to write

e−1 LNoether = −ifXiA (∂ν q X )ψ μ (−γ μν + g μν )ζ A + c.c.

i
(8.203)

The variation of ζ A introduces one additional gamma matrix, so that a term

involving γ μνρ can only arise from the γ μν -term in (8.203). We thus have
 
−1 i i jA ρ
δζ [e LNoether]| γ μνρ = ifXiA (∂ν q X
)ψ μ γ μν f γ ∂ρ q j |γ μνρ + c.c.
Y
2 Y
1 jA i
= − fXiA fY (∂ν q X )(∂ρ q Y )ψ μ γ μνρ j + c.c. (8.204)
2

This is cancelled by the gravitino variation δψμi = Dμ  i + . . . in the kinetic term

i
Lgravitino = −ψ μ γ μνρ Dν ψρi ,

i
δLgravitino = −ψ μ γ μνρ Dν Dρ i + c.c. (8.205)
1 i
= − RXY i j (∂ν q X )(∂ρ q Y )ψ μ γ μνρ j , (8.206)
2

where RXY i j is the composite SU (2) curvature. Comparing with (8.204), we

conclude that

RXY i j = MP−2 f[XiA fY ] ,

jA
(8.207)

where we have reinstalled the Planck mass, which originates from δψμi ∝ MP
and LNoether ∝ MP−1 . Using the explicit relation (8.175) to the standard SO(4nH )
vielbein fΓX , one can verify that the right-hand side of the above equation (8.207)
is not identically zero and hence that the SU (2) curvature in N = 2 supergravity
must be non-trivial. Obviously, the above reasoning becomes empty in the case of
rigid supersymmetry, and the SU (2) curvature becomes flat, as also follows from
(8.207) by formally sending MP → ∞.
222 8 Extended Supergravities

Exercises

8.1. Construct the Kähler potential, the metric, and the Killing vectors and compute
the action of the isometries on the symplectic sections for the STU model, i.e., for
the prepotential

X1 X2 X3
F (X) = .
X0

References
1. M.K. Gaillard, B. Zumino, Duality rotations for interacting fields. Nucl. Phys. B193, 221
(1981). Dedicated to Andrei D. Sakharov on occasion of his 60th birthday
2. B. Craps, F. Roose, W. Troost, A. Van Proeyen, What is special Kahler geometry? Nucl. Phys.
B 503, 565–613 (1997). [arXiv:hep-th/9703082 [hep-th]]
3. S. Weinberg, The Quantum Theory of Fields. Vol. 3: Supersymmetry (Cambridge University
Press, Cambridge, 2000), 419 p.
4. P. Claus, K. Van Hoof, A. Van Proeyen, A symplectic covariant formulation of special Kahler
geometry in superconformal calculus. Class. Quant. Grav. 16, 2625–2649 (1999). [arXiv:hep-
th/9904066 [hep-th]]
5. B. de Wit, P.G. Lauwers, A. Van Proeyen, Lagrangians of N=2 Supergravity - Matter Systems.
Nucl. Phys. B 255, 569–608 (1985)
6. A. Strominger, Special geometry. Commun. Math. Phys. 133, 163 (1990)
7. A. Ceresole, R. D’Auria, S. Ferrara, A. Van Proeyen, Duality transformations in super-
symmetric Yang-Mills theories coupled to supergravity. Nucl. Phys. B 444, 92–124 (1995).
[arXiv:hep-th/9502072 [hep-th]]
8. A. Ceresole, R. D’Auria, S. Ferrara, The symplectic structure of N=2 supergravity and its
central extension. Nucl. Phys. Proc. Suppl. 46 , 67–74 (1996). [hep-th/9509160]
9. J. Bagger, E. Witten, Matter couplings in N=2 Supergravity. Nucl. Phys. B 222, 1–10 (1983)
10. M. Berger, Sur les groupes d’holonomie homogènes des variétés a connexion affines et des
variétés riemanniennes. Bull. Soc. Math. France 83, 279–330 (1953)
Gauged Supergravity
9

In the previous chapter, we discussed the consequences of extended supersymmetry

for the geometries of the scalar manifolds in supergravity. Some parts of these
geometrical structures are caused by the larger R-symmetry groups, whereas others
can be traced back to the electric–magnetic duality of the vector field sector. In the
present chapter, we go one step further and turn on gauge interactions in the form of
minimal couplings between the vector fields and the other matter fields in extended
supergravity. Due to the extended supersymmetry, the possible gauge interactions
are more restricted, and they are more tightly connected to the interactions of the
other fields, in particular they completely determine the scalar potential for N ≥ 2.
For this reason, one gives extended supergravity theories with gauge interactions a
special name and calls them gauged supergravity theories. Another important aspect
of the gauge interactions is that their presence breaks the electric–magnetic duality
of the ungauged theories. As this duality underlies many geometrical structures, it
is useful to discuss gauged supergravity also in a way that formally maintains the
symplectic covariance. This formalism is called the embedding tensor formalism
and will be used in this book. To arrive there, we first describe, in Sect. 9.1, some
general features of gauged supergravity and how their connection with the scalar
potential arises. In Sects. 9.2 and 9.3, we discuss some important subgroups of the
electric–magnetic duality group as well as the relevance of the symplectic duality
frames for gauge interactions. In Sect. 9.4, we introduce some terminology on
coset manifolds, whereas Sect. 9.5 puts all these things together and presents the
embedding tensor formalism for gauged supergravity, with the N = 8 theory as an
explicit illustration. In Sect. 9.6, we finally discuss how inequivalent gaugings can
be classified and show, as an application, how this formalism can be used to uncover
some gaugings in the N = 8 theory that were unknown until a few years ago.

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 223

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_9
224 9 Gauged Supergravity

9.1 Supergravities and Scalar Potentials

Generic matter multiplets in supergravity theories contain scalar fields, whose

expectation values determine the masses and the couplings of the other fields and
eventually the physics we would like to describe by these models. Such expectation
values are arbitrary as long as there is no scalar potential. On the other hand, the
existence of a scalar potential constrains their values and hence provides crucial
information on the resulting physics.
In N = 1 supergravity, the scalar potential receives contributions from both
the superpotential, W , and the D-terms. There is a large amount of freedom
involved in the choice of a superpotential. For one thing, any non-vanishing
superpotential can be transformed to any other non-vanishing superpotential by a
suitable Kähler transformation such that the Kähler invariant combination of Kähler
and superpotential,

|W |2
G = K + MP2 log , (9.1)
MP6

stays the same. But even if one fixes a particular Kähler potential, one usually has
a large amount of freedom in the choice of W , as long as one just respects any
gauge symmetries that might be present. It should also be emphasized that in N =
1 supergravity a superpotential and the resulting F-term potential can always be
chosen at will if there is no gauge symmetry whatsoever.
The other contribution to the N = 1 scalar potential, the D-term potential, by
contrast, is directly related to a gauging, i.e., to the process of turning some of the
global internal symmetries into local symmetries. These global internal symmetries
are essentially coincident (with few exceptions) with the isometries of the scalar
manifold, described by its Killing vectors, ξIi ,

δϕ i = α I ξIi (ϕ), where ∇(i ξI j ) = 0, (9.2)

with infinitesimal symmetry parameters, α I . An important difference to the scalar

potential that arises from a superpotential is that the D-term potential is fixed as soon
as the action of the gauge symmetries on the fields is fixed.1
In this chapter, we will mainly deal with this gauging procedure. The reason
is that extended supergravities have no scalar potential without gauging! In other
words, in extended supergravity there is no analogue of an F-term potential due
to a superpotential that can exist independently of any gauging. This means that if
we want to stabilize the scalar fields that inevitably appear in almost all extended
supergravity theories and that we know cannot be massless because of the limits
from fifth force experiments, we need to gauge the theory. This is also the case

1As described in Chap. 6, in N = 1 supergravity this is also true for any Fayet–Iliopoulos
constants, as they entail gaugings of the U (1) R-symmetry group that act on the fermions.
9.1 Supergravities and Scalar Potentials 225

if we want to break supersymmetry in the vacuum, because if there is no scalar

potential, the only maximally symmetric vacuum allowed is a fully supersymmetric
Minkowski spacetime.
One may use this as an argument to focus exclusively on N = 1 theories, which
anyway can lead more easily to realistic phenomenology due to their exclusive com-
patibility with chiral gauge interactions. In many string theory compactifications,
however, the closed string sector can be described by effective field theories with
extended supersymmetry, which is reduced to N = 1 only due to their coupling
to open strings living on branes used to describe supersymmetric extensions of the
Standard Model as well as orientifold projections.
Moreover, in many branches of string theory, it is only due to extended
supersymmetry that certain computations can be carried out with sufficient technical
control. In the context of the gauge/gravity correspondence, this is in particular
true for the gauge theories and their renormalization group flows at the boundary,
whose holographic description then also necessitates extended supersymmetry in
the gravity dual in the bulk. These extended supergravity theories must have a non-
trivial scalar potential to allow for AdS vacua or domain wall solutions, and hence
they are also gauged supergravity theories with extended supersymmetry.
In this chapter, we will give a brief introduction to gauged supergravity,
explaining how the gauging procedure works and what kind of generic features and
constraints they involve.
Before entering into the details of such constructions, let us give a brief descrip-
tion of what the necessary steps are and what we should expect. As discussed in
Sect. 5.3.5, isometries of the scalar manifold of a supergravity theory are generically
global symmetries of the Lagrangian (or at least of the equations of motion). In fact,
transformations of the form (9.2) obviously leave the kinetic couplings of the scalar
fields,

gij (ϕ) ∂μ ϕ i ∂ μ ϕ j , (9.3)

invariant (for the sake of simplicity, we do not discuss now the other matter
couplings).
As explained in Sect. 5.3.5, gauging such isometries means allowing for a non-
constant symmetry parameter, α I (x), which then implies that the variation of (9.3)
gives uncancelled terms like

∂μ α I (x)∂ μ ϕ j ξI j . (9.4)

As detailed in Sect. 5.3.5, these terms can be cancelled by introducing minimal

couplings to the vector fields of the theory,

∂μ ϕ i −→ !
∂μ ϕ i ≡ ∂μ ϕ i − AIμ ξIi , (9.5)
226 9 Gauged Supergravity

and by requiring that the Maxwell-type transformation, δAIμ = ∂μ ΛI , of the

ungauged theory is changed to

δAIμ = ∂μ α I + AJμ α K fJ K I . (9.6)

This procedure is in general not possible for all scalar field isometries. Instead, the
gaugeable isometries usually only form subgroups, G ⊂ Iso(Mscalar ), that have to
obey certain restrictions. The most obvious restrictions for G are that its dimension
be at most equal to the number, nV , of vector fields, dim G ≤ nV , and that the
vector fields transform in the adjoint representation of the subgroup G in the sense
of a global symmetry prior to the gauging.
In order to also make the kinetic terms of the vector fields gauge invariant, one
has to replace the Abelian field strengths, FμνI , by their non-Abelian counterparts,

Fμν I = 2∂ AI + f
[μ ν]
I J K
J K Aμ Aν .
Although we are not going to do this in later parts of this chapter, let us
temporarily follow a common practice and rescale the vector fields by a coupling
constant, A → g A, so that our modification (9.5) of the derivative of the scalar
fields is seen to be a modification of order g, and hence the scalar kinetic term
corresponds to a modification of order g 2 . Likewise, the field strengths of the vector
fields have been modified at order g relative to the Abelian part, so that also their
kinetic terms are modified at order g 2 relative to the ungauged case.
These modifications obviously break supersymmetry, because the scalars and
vector fields have fermionic superpartners whose kinetic terms do not yet contain
such gauge covariantizations. Covariantizing also these kinetic terms of the fermions
as well as the supersymmetry transformation rules, many supersymmetry variations
will in fact cancel in a way that is very similar to the ungauged theory. A
small number of uncancelled variations, however, remains and requires further
modifications of the action and the supersymmetry transformation laws at order g or
g 2 . More precisely, in order to compensate the new terms, we must further modify
the supersymmetry transformation rules of the fermions,

δSUSY (Fermions) = δSUSY,old (Fermions) + δSUSY,new (Fermions), (9.7)

where

δSUSY,new (Fermions) = O(g) (9.8)

are “fermionic shifts” involving scalar field-dependent terms and one power of g.
This, in turn, also enforces the introduction of Yukawa-like terms for the same
fermions,

LYuk = O(g). (9.9)

9.2 Duality 227

The variation of LYuk due to δSUSY,new will then lead to an uncancelled variation of
order g 2 , which can be cancelled by adding a scalar potential of order g 2 ,

V (ϕ) = O(g 2 ). (9.10)

For consistent gaugings, no O(g 3 ) terms are needed. These steps match precisely the
ones we described in Chap. 4, when we added a cosmological constant to minimal
pure supergravity, except that there was no gauge covariantization involved.
Although the procedure just described looks straightforward, it can be technically
very challenging, and its application leads to the discovery of many interesting
general properties of supergravity theories. For this reason, in what follows, we are
going to consider a simplified instance, namely, the case where the scalar σ -model
is described by a coset manifold,

G
Mscalar = . (9.11)
H
Actually, for theories with N ≥ 3, this is always true and also for N = 2 theories,
one is often interested in the analysis of scalar manifolds that have this form. This
case, as we will see, simplifies considerably many technical aspects, leaving the
substance of the gauging procedure unchanged.
Before we can enter this discussion, however, we first have to make more precise
what kind of global symmetries we could actually encounter when we talk about the
process of gauging.

9.2 Duality

As we are going to see in the following, the gauging procedure is intimately

related to the existence and action of duality symmetries. Actually, duality relations
are at the basis of our understanding of the vast majority of the most interesting
physical effects in supergravity and string theory. In the supergravity context, the
subject was opened by the seminal paper by Gaillard and Zumino [1], which
discusses the duality group of four-dimensional theories containing vector fields. We
already introduced the electric–magnetic duality in Chap. 8. This, however, does not
coincide with the U-duality group, which is the one playing the most prominent role
in the gauging procedure. For this reason, we now present the details that are needed
to reduce the electric–magnetic duality group to the U-duality group and how this
group is at the basis of all possible deformations of extended supergravities via the
gauging procedure.
228 9 Gauged Supergravity

9.2.1 From Electric–Magnetic Duality to U -Duality

In Chap. 8, we saw that the presence of nV vector fields in a four-dimensional

Lagrangian comes with a generalized electric–magnetic duality leaving invariant
the set of Bianchi identities and equations of motion of the vector fields under
Sp(2nV , R) rotations. On the other hand, under an infinitesimal Sp(2nV , R) transfor-
mation, the Lagrangian does in general not remain invariant but changes according
to (8.28),

1 1 μν
δL = I
CI J Fμν F J μν + B I J GI μν GJ + δLrest. (9.12)
4 4
If we want the electric–magnetic duality to leave all the equations of motion
unaffected, we should understand if and how it acts on the other fields present in the
Lagrangian. This, as we will see, will imply further restrictions so that the resulting
set of transformations is going to be encoded in the U-duality group GU ,

GU ⊂ Sp (2nV , R). (9.13)

In order to simplify the following discussion, we will focus on scalar fields, but the
proof can be extended to fermion fields as well.
If we call ϕ i the set of scalar fields on which the couplings and Lrest depend, we
define their equation of motion operator, Ei , as
 
∂ ∂
Ei [L ] ≡ i
− ∂μ L = 0. (9.14)
∂ϕ ∂∂μ ϕ i

If the duality transformations act non-trivially on the scalar fields, they are going to
transform as

δϕ i = k i (ϕ), (9.15)

where k i (ϕ) are some as yet unspecified functions. Once again, we would like that
the full set of Bianchi identities and equations of motion for all fields in L are
invariant under the action of the duality transformations. This is easily achieved if
the equations of motion of the scalar fields transform covariantly with respect to the
action of the duality symmetry2

∂k j
δEi [L ] = − Ej [L ]. (9.16)
∂ϕ i

from (9.15), recalling that for a contravariant vector V =

2 This follows i
 straightforwardly
∂ϕ i j ∂k i
∂ϕ j
V  δj +  ∂ϕ j + O( ) V and that a covariant vector transform with the inverse
i 2 j

Jacobian.
9.3 Gauging and Symplectic Frames 229

This requirement produces an interesting result, once we apply the equation of

motion operator to the variation (8.28). After some algebraic simplifications, we
obtain that

Ei [δLrest ] = 0, (9.17)

and the only way to consistently satisfy this identity is that

δLrest = 0. (9.18)

We therefore conclude that the subgroup of electric–magnetic duality transforma-

tions that leave the full set of equations of motion and Bianchi identities invariant is
the set of transformations that also leave invariant Lrest. If, as usual, we interpret
the scalar fields as coordinates on a manifold, k i (ϕ) are going to be identified
with the Killing vectors, ξ i (ϕ), associated with the isometries of this manifold, and
we therefore see that the U-duality group is going to coincide with the symmetry
group under which the reparameterization of the scalar fields leaves the Lagrangian
invariant.3
As we said, this proof can be extended to more general Lagrangians containing
also fermion fields, and therefore we will always have a reduction of the electric–
magnetic group Sp (2nV , R) to a U-duality subgroup that is the subgroup of
transformations leaving invariant Lrest . In general, the global symmetry group
leaving invariant the set of Bianchi identities and equations of motions is Gglobal =
GU ×Ginert, where Ginert is the global symmetry of the inert matter fields under
duality transformations, i.e., those fields that do not have direct couplings to the
vector fields, as, e.g., the hypermultiplets in N = 2 supergravity.
Finally, one can also see that when gravity is present, Einstein’s equation
contains a stress tensor that is invariant under the U-duality transformations, as a
consequence of its conservation law following from the equations of motion of the
other fields [1]. This is a very important fact that implies that one can map different
solutions of Einstein’s equation into each other by carefully acting on the original
set of fields defining the solution with the U-duality group.

9.3 Gauging and Symplectic Frames

By gauging we usually mean the procedure of making local a global symmetry, GL ,

of a given Lagrangian, L . However, when dealing with the duality group, GU , we
see that this is a symmetry group of the Bianchi identities and equations of motion

3We should note that in some instances there may be additional global symmetries that act trivially
on the scalar manifolds and that there are also examples where there is still a restricted non-trivial
duality group in models without scalar fields, like pure N = 2 supergravity. In most of these
cases, the additional factors come from the R-symmetry acting non-trivially on the fermion fields
of the theory.
230 9 Gauged Supergravity

and not necessarily of the Lagrangian. Moreover, we should also notice that for a
given theory with a fixed set of multiplets, we may have different Lagrangians with
different GL . Actually, we are often allowed to have entirely different groups for
different Lagrangian realizations. This is clear if we recall that the group of electric–
magnetic duality transformations, Sp(2nV , R), leaves invariant the equations of
motion and Bianchi identities of the vector fields, but only a subgroup, GU , can
be reabsorbed in a field redefinition of the other fields. In particular, only electric
vectors appear in the Lagrangian, and therefore GL transformations may only be
represented by block triangular matrices acting on the symplectically covariant field
strengths, F, introduced in (8.10). A generic Sp(2nV , R) transformation is going
to modify the Lagrangian, instead. Hence for the same theory, we can consider
different Lagrangians having different realizations of GL . It is therefore important
to understand which realizations of GL are available before the gauging procedure
is carried out.
The set of Lagrangians that cannot be mapped to each other by local field
redefinitions is identified with the double quotient space

GL(nV , R) \ Sp(2nV , R) / GU . (9.19)

This follows from the fact that we can always perform local field redefinitions of the
nV vector fields of the theory, which corresponds to the GL(nV , R) quotient, as well
as use the U-duality group GU corresponding to the isometry group of the scalar
manifold Mscalar to redefine the other fields. Each different Lagrangian corresponds
then to a distinct symplectic frame and is invariant under a particular electric
subgroup of the GU duality group acting locally on the physical fields. Hence we
can use this choice of frame to obtain different gauge groups by introducing minimal
electric couplings. The choice of the symplectic frame is therefore important to find
a purely electric realization of Ggauge. On the other hand, one should notice once
more that the resulting equations of motion and Bianchi identities are equivalent for
any Lagrangian defined by (9.19) before the gauging. In fact, even if the Lagrangians
differ in different symplectic frames, the full set of Bianchi identities and equations
of motion can still be mapped into each other by the symplectic field redefinition.
So, classically we have different Lagrangian descriptions of the same physics. As
we will see explicitly later, this is broken by the gauging, and hence this choice of a
symplectic frame is relevant for the gauging procedure.
Alternatively, as we will see shortly, keeping the symplectic frame fixed, we
can obtain Ggauge ⊂ GL (while obviously Ggauge ⊂ GU ) by introducing magnetic
charges and couplings for the non-perturbative symmetries that would not leave
invariant L . This requires writing the Lagrangian with the gauge interactions in a
duality invariant way, and this can be done by introducing the embedding tensor
formalism. As we will see, this formalism is especially suited for a general analysis
of the gaugings of a given supergravity theory. The reason is that once we fix a given
symplectic frame the embedding tensor describes precisely how Ggauge is embedded
in the duality group GU and eventually in the symplectic group Sp(2nV , R). This
allows for a general formulation of the gauging procedure that is also transparent
9.4 Coset Manifolds and Gauging 231

to the differences between theories that have the same Ggauge, but inequivalent
symplectic embeddings.

9.4 Coset Manifolds and Gauging

Before entering the discussion of the embedding tensor formalism, we recall a

few basic facts about coset manifolds, which play a prominent role as σ -model
geometries of supergravity theories. We illustrate this terminology with a simple
example that we are going to use throughout the chapter. As mentioned above, for
the sake of simplicity, we will limit our discussion and examples of the embedding
tensor applications to theories where the scalar manifold is homogeneous, so that we
can use the coset structure to obtain linear realizations of the symmetries involved
in the gauging procedure.
A homogeneous space is a manifold with a metric whose isometry group, G, has
a transitive action on the space, meaning that any point on the space can be reached
from any other by the group action. The subgroup, H, of G that leaves a chosen
point, x, of the manifold fixed, is called the isotropy subgroup of x. Because of the
transitive action of G, any other point, x  = gx (obtained by the action of g ∈ G,
g ∈/ H), is invariant under the subgroup gHg −1 of G, which is isomorphic to H.
This implies that any homogeneous space can be described as the coset space G/H,
defined as the set of equivalence classes of elements of G with respect to the (right)
action of H elements:

g ∼ g , if g = g h for g, g  ∈ G, h ∈ H. (9.20)

The dimension of such a coset space is then d = dim[G] − dim[H].

Given this identification, we can now describe homogeneous spaces directly by
their coset structure. The Lie algebra of the group G, g, can be split as

g = h ⊕ k, (9.21)

where h is the Lie algebra of H and k contains the remaining generators. We can
then classify coset manifolds according to the structure of their algebra

[ti , tj ] = fij k tk , (9.22)

[ti , ta ] = fia k tk + fia b tb , (9.23)

[ta , tb ] = fab k tk + fab c tc , (9.24)

where tA ∈ g, ti ∈ h, and ta ∈ k. A very useful concept is that of the Cartan–Killing

metric

ηAB = fAD C fBC D , (9.25)

232 9 Gauged Supergravity

which is non-degenerate for semi-simple groups. This implies that, for semi-simple
groups, we can always map the generators to a basis so that the Cartan–Killing
metric on g is diagonal. With respect to that basis, any coset manifold G/H is
reductive, i.e., the decomposition above satisfies

fia j = 0 ⇔ reductive, (9.26)

so that [h, k] ⊆ k. Actually, one can prove the existence of a reductive decomposition
also for the sum of a semi-simple and an Abelian Lie algebra [2]. In the following,
we will therefore focus on reductive coset spaces. Another useful definition appears
when the commutator of generators in k closes on h. This is called a symmetric coset
manifold:

fab i = 0 ⇔ symmetric, (9.27)

or [k, k] ⊆ h. This typically happens when G is simple and H is maximal, i.e., when
there is no other proper subgroup of G that contains H strictly. Moreover, this
is always the case when we have a non-compact reductive G/H and maximally
compact H. In fact, a reductive non-compact coset space can be obtained from its
compact counterpart by multiplying the generators in k by the imaginary unit i. This,
however, forces [k, k] ⊆ h, because if, in the compact case, the commutator between
elements of k also involved terms in k, the closure of the algebra could in general no
longer work with real structure constants when one multiplies such generators by i.
Another useful concept, when dealing with non-compact coset manifolds, is
given by the solvable or Iwasawa decomposition [2]. This decomposition assures
that for any Euclidean non-compact maximal homogeneous manifold G/H there is
a solvable subalgebra of g, solv ⊂ g, acting transitively on G/H, such that

g = h + solv, dim solv = dim G/H, (9.28)

where a solvable algebra is an algebra whose derivative algebras vanish at some

finite order, i.e., repeated commutators of the generators vanish at a finite order.
Such a solvable algebra can be constructed as follows. First one considers the
maximal Abelian subspace of k, Ck , which can be proven to be given by all the
non-compact elements of the Cartan subalgebra of g, named C . Hence, Ck = C ∩ k.
Then one fixes a set of positive roots Δ+ ⊂ g with respect to Ck , for instance,
by taking a hyperplane passing through the origin of the algebra root diagram that
does not contain any generator in addition to those appearing at the origin and then
considering all generators on one side of such hyperplane (Fig. 9.1). A solvable
algebra is then given by

solv = Ck + Δ+ . (9.29)
9.4 Coset Manifolds and Gauging 233

Fig. 9.1 Examples of the Iwasawa decomposition for the isometries of the SU(2,1)/[SU(2) ×
U(1)] manifold (root diagram on the left) and G2(2) /[SO(4)] manifold (root diagram on the right).
For su(2, 1) only one generator of the Cartan subalgebra is non-compact, the one whose weights
are plotted following the red arrow. The set of positive roots (in red) plus the non-compact Cartan
give the four generators of the corresponding manifold. For the g2(2) algebra, we have that C = Ck ;
hence the set of positive roots plus the two Cartan generators give a total of eight generators

Let us now give an example, by considering the coset manifold

M = SU(1, 1)/U(1). (9.30)

The generators of the corresponding algebras can be taken as

 "
1 i 1 i
tA = σ1 , σ2 , σ3 ∈ su(1, 1), th = t2 = σ2 ∈ u(1), (9.31)
2 2 2 2

satisfying

[t1 , t2 ] = −t3 , [t3 , t1 ] = t2 , [t2 , t3 ] = −t1 , (9.32)

so that the Cartan–Killing metric is

η = diag{1, −1, 1} . (9.33)

The root diagram can be described by placing the non-compact Cartan generator
t3 ∈ Ck at the origin and taking the positive and negative roots as t± = 12 (t1 ± t2 ),
which then have weight one,

[t3 , t± ] = ± t± . (9.34)
234 9 Gauged Supergravity

t t3 t

Fig. 9.2 Root diagram and decomposition for su(1, 1)

A generic coset element can be described by the exponential of the generators in

the quotient subspace, k = span{t1 , t3 },
⎛ φ φ
⎞
e 2 cosh ψ
e− 2 sinh ψ
L(φ, ψ) = exp (ψ t1 ) exp (φ t3 ) = ⎝ ⎠,
2 2
φ
(9.35)
ψ − φ2 ψ
e sinh
2
2 e cosh 2

or, using the Iwasawa decomposition, by the exponential of the generators of the
solvable algebra, solv = span{t3 , t+ },
⎛ φ φ
⎞
e 2 e− 2 C
L(φ, C) = exp (C t+ ) exp (φ t3 ) = ⎝ φ
⎠. (9.36)
0 e− 2

In this case, as expected, the result is an exponential for the coordinates correspond-
ing to the generators in the Cartan subalgebra and a polynomial in the coordinates
associated with the positive roots generators (Fig. 9.2).
As L(x) is just a coset representative, the action with a constant group element, g,
from the left in general gives the coset representative at the transformed coordinate,
x  , only modulo a local H transformation from the right,

gL(x) = L(x  )h(x). (9.37)

For instance, if we choose the finite O(1,1) transformation

 
1 31
g = exp (log 2 t1 ) = √ , (9.38)
2 2 13

then

gL(φ, C) = L(φ̃, C̃)h(φ, C), (9.39)

where
 
1 C + 3 −eφ
h(φ, C) = - (9.40)
(C + 3)2 + e2φ eφ C +3
9.4 Coset Manifolds and Gauging 235

is an element of U(1) and the new representative coordinates are specified by

% √ &
2 2 eφ/2
φ̃ = 2 log - , (9.41)
(C + 3)2 + e2φ

8(3 + C)
C̃ = 3 − . (9.42)
(C + 3)2 + e2φ

9.4.1 Vielbein, Metric, and Isometries of G/H

From the coset manifold representative, L, we can construct the left-invariant one-
form

Ω(x) = L−1 (x)dL(x). (9.43)

This is called the Maurer–Cartan form, and it is Lie algebra valued and may be
expanded in terms of the g generators

Ω(x) = ea (x) ta + ωi (x) ti , (9.44)

where ea is a covariant vielbein on G/H, and ω is called the H-connection. Recalling

(9.37), under left multiplication by a constant element of G, the Maurer–Cartan form
is not invariant, but transforms as

Ω(x  ) = hΩh−1 + hdh−1 . (9.45)

This is a very interesting property, because, once projected onto the subalgebras k
and h, this means that the vielbein rotates under the action of the H subgroup, and
ωi transforms as a connection:

ea (x  ) = eb (x)Db a (h−1 ), (9.46)

ωi (x  ) = ωj Dj i (h−1 ) + (hdh−1 )i , (9.47)

where DA B (g) is the adjoint representation of G acting on the algebra generators.

Once we construct a (left-invariant) metric from the vielbein obtained in (9.44),
we can ask ourselves how to derive the isometries, which should naturally respect
the free action of the group G on the manifold. Let us consider once again (9.37) at
the infinitesimal level. Expanding the constant g transformation as g = 1 + α A tA
and the local transformation h(x) = 1 + α A wA i
(x)ti and interpreting the result as
the action of the corresponding Killing vector

L(x  ) = L(x) + α A ξA β (x)∂β L(x), (9.48)

236 9 Gauged Supergravity

we obtain

tA L(x) = ξA (x)[L(x)] − L(x)ti wA

i
(x), (9.49)

where ξA (x)[L(x)] ≡ ξA β (x)∂β L(x), and wA

i (x) is often called the H-compensator.

It can be readily verified that

[ξA , ξB ] = −fAB C ξC , (9.50)

and multiplying (9.49) by L−1 from the left and projecting on the generators, one
gets an explicit expression for the Killing vectors and for the H-compensators

ξA α (x) = DA a (L(x))eaα (x), (9.51)

i
wA (x) = ωα i (x)ξAα (x) − DA i (L(x)), (9.52)

In order to make all this concrete, let us explicitly compute the metric and
isometries of the SU(1,1)/U(1) manifold. We start from the representative (9.36).
The Maurer–Cartan form is then
 dφ −φ 
2 e dC
Ω= = dφ t3 + e−φ dC t+ = e1 t1 + e2 t3 + ωU(1) t2 . (9.53)
0 − dφ 2

This means that we can identify the vielbein with

ea = {dφ, e−φ dC} (9.54)

and the U(1) connection with

ωU(1) = e−φ dC, (9.55)

whose curvature is

RU (1) = dωU(1) = e−φ dC ∧ dφ. (9.56)

This implies that the metric is proportional to ea eb δab with a coefficient of

proportionality fixing the curvature
 
ds 2 = 2 dφ 2 + e−2φ dC 2 . (9.57)

The space is indeed an Einstein space with negative Riemannian curvature:

1
Rαβ = − gαβ . (9.58)
2
9.4 Coset Manifolds and Gauging 237

For what concerns the isometries, we can then use (9.49) to obtain
 
1 0 e−φ/2 1
t1 L = = (1 + e2φ − C 2 )∂C L − C∂φ L − Lt2 e−φ , (9.59)
2 eφ/2 e−φ/2 C 2

 
1 0 −e−φ/2 1
t2 L = = (1 − e2φ + C 2 )∂C L + C∂φ L + Lt2 e−φ , (9.60)
2 eφ/2 e−φ/2 C 2

 
1 eφ/2 e−φ/2 C
t3 L = = C∂C L + ∂φ L. (9.61)
2 0 −e−φ/2

From these we then recognize the Killing vectors

1
ξ1 = (1 + e2φ − C 2 )∂C − C∂φ , (9.62)
2
1
ξ2 = (1 − e2φ + C 2 )∂C + C∂φ , (9.63)
2
ξ3 = C∂C + ∂φ . (9.64)

Notice that we can associate prepotentials, PI , to these isometries, through the U(1)
curvature,

dPI = ıI RU(1), (9.65)

so that
1 φ
P1 = (e − e−φ + e−φ C 2 ), (9.66)
2
1 φ
P2 = (e + e−φ + e−φ C 2 ), (9.67)
2

P3 = C e−φ . (9.68)

In fact, one can prove that SU(1,1)/U(1) is a Kähler manifold and that RU(1) can be
identified with the Kähler form J .
238 9 Gauged Supergravity

9.4.2 The Special-Kähler Manifold SU(1,1)/U(1) and Inequivalent

Symplectic Frames

The homogeneous manifold SU(1,1)/U(1) is also an instance of a special Kähler

manifold. One way to see this is by introducing the prepotential

F (X) = −i X0 X1 , (9.69)

where XI are the projective coordinates defined in Chap. 8. Choosing the indepen-
dent coordinate z according to X0 = 1, X1 = z, we have F0 = −iz and F1 = −i,
so that under Sp(4,R), the combination {XI , FI } forms a vector. This leads to the
Kähler potential

I
K = − log[i(X FI − XI F I )] = − log[2(z + z)], (9.70)

and metric
dz ⊗ dz
ds 2 = gzz dz ⊗ dz = . (9.71)
(z + z)2

Notice that we recover the metric of the previous section by setting

z = eφ + i C, (9.72)

with  = 1/2. Notice also that the real part of the complex coordinate z has to be
positive definite, because of the Kähler potential (9.70). Special geometry fixes also
the gauge kinetic matrix to be diagonal

i
N00 = −i z, N11 = − . (9.73)
z

Finally, the isometries we discussed in the previous section can be described by

holomorphic Killing vectors:

i  
ξ1 = 1 + z2 ∂z , (9.74)
2
i  
ξ2 = −1 + z2 ∂z , (9.75)
2
ξ3 = z ∂z . (9.76)

We can now see explicitly the existence of different symplectic frames for the
N = 2 supergravity theory constructed by supergravity coupled to a single vector
multiplet with prepotential (9.69). Let us focus on the part of the Lagrangian
9.4 Coset Manifolds and Gauging 239

quadratic in the vector field strengths, i.e., the first two terms in (8.5). In the frame
defined by the prepotential above, we have

eφ 0 0 μν eφ C 0 0 μν C
e−1 LF 2 = − Fμν F − F 1 F 1 μν + Fμν F − F 1 F 1 μν .
4 4|z|2 μν 4 4|z|2 μν
(9.77)

Now we can construct the full symplectic vector of field strengths on which one
acts with symplectic and duality transformations to find the inequivalent symplectic
frames. Focusing again on LF 2 , we have
0
G0μν = C Fμν + eφ F 0ρσ , (9.78)

C 0 eφ 0ρσ
G1μν = − F + F . (9.79)
|z|2 μν |z|2

Altogether, F is subject to Sp(4,R) transformations, whose generators can be split

into those of GU = SU(1, 1),

1
su(1, 1) = {σ1 ⊗ σ3 , −i σ2 ⊗ 1, −σ3 ⊗ σ3 } , (9.80)
2
the vector field redefinitions GL(2,R),

1
gl(2, R) = {−σ3 ⊗ σ3 , σ3 ⊗ 1, σ3 ⊗ σ1 , 1 ⊗ i σ2 , } , (9.81)
2
where the first generator is in common between gl(2,R) and su(1, 1), and finally the
remaining generators of the quotient

GL(2, R)\Sp(4, R)/SU(1, 1), (9.82)

namely,

1
{σ1 ⊗ 1, i σ2 ⊗ σ3 , σ1 ⊗ σ1 , iσ2 ⊗ σ1 } . (9.83)
2
The identification (9.80) can be explicitly confirmed by computing its action on the
holomorphic sections, V = (XI , FI ), and the resulting action on the independent
coordinate z:
 
X1 δI X 1 X1
δI z = δI = − δI X0 = ξI (z), (9.84)
X0 X0 (X0 )2

where δI XJ can be read from the direct action δI V = tI V .

240 9 Gauged Supergravity

Let us now see the effect of an equivalent frame, obtained by a duality

transformation, and that of an inequivalent one, obtained by the action of an element
of GL(2, R)\Sp(4, R)/SU(1, 1). The simplest example corresponds to the shift
symmetry C → C + α, which is generated by t+ = 12 (σ1 ⊗ σ3 − i σ2 ⊗ 1). The
application of the symplectic matrix

Sd = exp (α t+ ) (9.85)

on F produces the Lagrangian

eφ 0 0 μν eφ
e−1 LF 2 = − Fμν F − F 1 F 1 μν
4 4(e2φ + (C + α)2 ) μν
(9.86)
C + α 0 0 μν C +α
+ Fμν F − F 1 F 1 μν ,
4 4(e2φ + (C + α)2 ) μν

which straightforwardly goes back to the previous Lagrangian by the obvious redef-
inition of C  = C +α, which is also an isometry of the scalar manifold. On the other
hand, if we now use the transformation generated by tsp = 12 (σ1 ⊗ σ1 − i σ2 ⊗ σ3 ),
which belongs to GL(2, R)\Sp(4, R)/SU(1, 1), we obtain

eφ 0 0 μν eφ
e−1 LF 2 = − Fμν F − F 1 F 1 μν
4 4(e2φ + (C − α)2 ) μν
(9.87)
C + α 0 0 μν C +α
+ Fμν F − F 1 F 1 μν ,
4 4(e + (C − α)2 ) μν
2φ

which now cannot be reabsorbed by a field redefinition. Notice that while introduc-
ing C  = C − α produces just a total derivative difference, we explicitly see that the
two Lagrangians are not identical, a fact that appears even more explicitly for other
isometries. A particularly interesting case is that of the duality transformation (the
missing entries are zero)
⎛ ⎞
1
⎜ 1⎟  
S=⎜
0 ⎟ = exp π (i σ2 ⊗ 1 + i σ2 ⊗ σ3 ) , (9.88)
⎝ 1 ⎠ 4
−1 0

which, as explained in Chap. 8, moves us to a frame where the prepotential does not
exist. In this frame, the Lagrangian becomes

eφ 0 0 μν eφ 1 1 μν C 0 0 μν C 1 1 μν
e−1 LF 2 = − F F − Fμν F + Fμν F + Fμν F , (9.89)
4 μν 4 4 4
9.5 Gauging and the Embedding Tensor 241

and the transformations related to t+ and tsp are now lower-triangular

⎛ ⎞ ⎛ ⎞
0 0
⎜ 0 ⎟ ⎜ 0 ⎟
S −1 t+ S = ⎜
⎝1 0
⎟,
⎠ S −1 tsp S = ⎜
⎝1
⎟,
⎠ (9.90)
0
1 0 −1 0

hence showing again the fact that the first can be reabsorbed by a field redefinition
0 F 0 μν and
and the second cannot. In fact, in the first case, the term in front of Fμν
1
Fμν F 1 μν are corrected in the same way, sending C → C + α, while in the second
case, we have that the term in front of Fμν0 F 0 μν is shifted as C → C + α, while the
1
term in front of Fμν F 1 μν gets shifted as C → C − α.

9.5 Gauging and the Embedding Tensor

Let us now assume that we have fixed our starting Lagrangian by choosing a specific
symplectic frame and that we want to perform the gauging of a generic Ggauge ⊂
GU ⊂ Sp(2nV , R). As we mentioned in the previous sections, the best way to do
so is by means of the embedding tensor formalism [3, 4, 12], and we will make the
simplifying assumptions that the scalar manifold is homogeneous and that there is
no inert matter under the duality group. We therefore review in the following this
formalism by explaining the general idea and the constraints and discuss some of
the Lagrangian terms that are crucially related to the non-vanishing of some of the
components of the embedding tensor.
The Bianchi identities and equations of motion of a generic supergravity theory
are invariant under global GU transformations
5
δL = α β tβ L + L ti wβ i
(9.91)
μ = −α (tβ )N Aμ
δAM β M N

and local U(1)2nV gauge transformations

μ = ∂μ Λ (x).
δAM M
(9.92)

Only the subset that does not introduce magnetic vectors in the Lagrangian will have
a Lagrangian description, and an even smaller subset will be a global symmetry of
the Lagrangian, according to the discussion of Sect. 9.3.
To proceed with the gauging, we need to know which generators are going to
constitute the gauge algebra. This is done by selecting some of the generators tα ∈
gU ⊂ sp(2nV , R) and associating to these generators a linear combination of the
242 9 Gauged Supergravity

vector fields (electric and magnetic), which is going to become the gauge vector. To
do so we have to specify the embedding tensor

ΘM α , (9.93)

a 2nV × dim gU matrix, which will give us the gauge algebra generators as
combinations of the duality ones:

XM ≡ ΘM α tα . (9.94)

Once we know this, the formalism is so constrained that all the couplings and
interactions of the corresponding gauged supergravity Lagrangian are fixed. Notice
that, by construction, dim(Ggauge) = rank(ΘM α ), and consistency forces

dim(Ggauge ) = rank(ΘM α ) ≤ nV . (9.95)

We could think of this tensor as the charge matrix. In fact, we can construct
covariant derivatives in terms of the full set of Lagrangian and dual vectors AM =
{AIμ , AI μ }:

!
∂μ ≡ ∂μ − AM
μ ΘM tα = ∂μ − Aμ ΘI tα − AI μ Θ
α I α Iα
tα . (9.96)

These covariant derivatives appear in the local transformations of the vector fields
under a Ggauge transformation
5
δL = α M (x) XM L + L Xi wM i
(9.97)
M αP ≡ !
μ = ∂μ α + Aμ XNP
δAM M N ∂μ α M ,

and XNP M ≡ ΘN α (tα )P M . This covariantization procedure obviously cannot be

consistent for an arbitrary choice of Θ. For instance, we would like the X generators
to close into a Lie algebra (or at least be part of a free differential algebra [5], as
we will see later). For this reason we have to impose constraints on the embedding
tensor, which we will now analyze.
Before moving to the analysis of such constraints though, let us pause for some
comments.
Before the gauging procedure, one can rotate the electric F I and magnetic GI
vector field strengths among themselves, without changing the equations of motion
nor the Lagrangian, provided one uses a transformation that is part of the U-duality
group (see again Sect. 9.3). This means that we can choose which are the vectors we
consider as fundamental degrees of freedom in the Lagrangian. Still, one can have
many different Lagrangians for a given supergravity theory (once the number of
supersymmetries and the matter content has been fixed), because one could change
couplings by a symplectic transformation that is not in the U-duality group. Once the
Lagrangian is fixed, the embedding tensor ΘM α tells us which among the electric
9.5 Gauging and the Embedding Tensor 243

and magnetic vectors enter in the gauging procedure. This obviously breaks the
invariance of the equations of motion and Bianchi identities under the U-duality
group. Also, at most nV vectors can enter with their potential in the Lagrangian,
while the others should never appear naked.
The same gauge group may have different embeddings in the duality group for
different choices of the embedding tensor, and the embedding of the duality group
in the symplectic group may be chosen differently to generate different theories.
Since Θ is telling us how we make local the isometries of the U-duality group, the
resulting theory does not depend only on the gauge group but also on its embedding.

9.5.1 Constraints on the Embedding Tensor

Once the gauge group has been chosen and the embedding tensor fixed, the gauging
can be carried out in the standard way, by introducing the covariant derivatives and
possibly some topological terms, in order for the resulting Lagrangian to be invariant
under local Ggauge transformations. This procedure, however, generically breaks
supersymmetry, and therefore one has to restore it by further deforming the original
Lagrangian in order to restore supersymmetry and preserve Ggauge invariance. This
procedure can be completed successfully if we use embedding tensors that satisfy
certain constraints. We now discuss them according to their origin.

• Gauge invariance. The embedding tensor must be a singlet of the gauge group,
and hence it should be invariant under the action of the XM generators:

δM ΘN α = ΘM β δβ ΘN α =
(9.98)
= ΘM β (tβ )N P ΘP α + ΘM β fβγ α ΘN γ = 0,

where the first term is the action on the N index and the second term comes from
the identification of the adjoint generators with the structure constants (tβ )γ α =
−fβγ α . The contraction of this constraint with the duality algebra generators tα
gives

ΘM β (tβ )N P ΘP α tα + ΘM α [tα , tβ ]ΘN β = 0, (9.99)

which implies the closure of the gauge algebra

[XM , XN ] = −XMN P XP , (9.100)

so that X[MN] P play a role similar to that of structure constants. Note, however,
that this constraint also contains X(MN) P XP = 0, which is generically a non-
trivial condition for magnetic gaugings.
244 9 Gauged Supergravity

• Locality. We obviously want that at most nV mutually local vectors appear in the
gauging process. This is expressed by

ΘM α ΘN β Ω MN = 0 ⇔ Θ I [α ΘI β] = 0. (9.101)

This equation means that electric and magnetic charges should be mutually local,
which in turn implies that there always is a choice of symplectic frame such that
the gauging can be made purely electric.
• Supersymmetry. The embedding tensor ΘM α is in a definite representation of
the duality group, which is the product of the adjoint (α) and the fundamental
(M, also in the fundamental of the symplectic group). Supersymmetry restricts
the allowed representations with a linear projection,

X(MNP ) = X(MN Q ΩP )Q = 0, (9.102)

which generically removes the highest weight, together with other represen-
tations. This projection appears as one requires the cancellation of the O(g)
terms originating from the supersymmetry variations of the minimal couplings
by means of new terms depending on the fermion shifts we discussed in previous
chapters. For instance, for N = 8 supergravity the embedding tensor sits in
the 56 ⊗ 133 = 56 ⊕ 912 ⊕ 6480, while supersymmetry closes if Θ is in the
representation 912. (This condition can be relaxed, so as to leave also the 56
representation if we keep the trombone symmetry, a scaling symmetry of the
theory that can be gauged if one gives up a Lagrangian description.) We should
stress that (9.102) may get corrected for N = 1 theories whenever anomalies
are present [6].

In theories where all scalar fields sit in the same multiplets as the vectors, the
first two conditions become equivalent once the linear representation constraint
is imposed; otherwise the locality and closure conditions should be imposed
independently. In fact, the two conditions are independent whenever there are scalar
isometries with no duality action, which means that the symplectic representation
of the isometry algebra is not faithful, i.e.,

(tα )M N = 0, for some tα ∈ gI so(M ), (9.103)

This is the case for N ≤ 2 theories, where we have matter multiplets that contain
scalar fields, but no vectors, like chiral or hypermultiplets.

9.5.2 Couplings

Clearly, there are many different ways to solve the constraints on the embedding
tensor, and we will discuss in the next section how, at least in principle, a full
9.5 Gauging and the Embedding Tensor 245

classification of all the possibilities that can be obtained. Let us, for now, provide
some details on some important classes of gaugings and their physical properties.
Electric gaugings are gaugings where only the electric vectors are used:

ΘIα = 0 ⇒ XI = 0. (9.104)

The other constraints further restrict the form of the remaining generators to
 
XI J K 0
(XI )M N
= , (9.105)
XI J K −XI J K

with

XI J K = X[I K] J , and XI [J K] = 0, X(I KJ ) = 0, (9.106)

where the last expression may get modified in N = 1 theories with quantum
anomalies [6]. It is interesting to note that already in this case the structure constants
are not always restricted only to the adjoint, but there is also a term mixing the
electric and magnetic vectors. Actually, the XI,J K term generates Chern–Simons-
like couplings of the form
 
1 μνρσ 3
LCS =−  XI,J K Aμ Aν ∂ρ Aσ + XLP Aμ Aν ,
I K J J L P
(9.107)
3 8

which are needed to cancel the gauge variation of the kinetic term.
On the other hand, the generic case may also have magnetic gaugings, and in this
case

XMN P = X[MN] P + Z P MN , (9.108)

with Z P MN = Z P NM , which, from (9.100), must satisfy

Z P MN XP = 0. (9.109)

Note that now Jacobi identities do not close in the usual form because of the Z
terms:

X[MN] P X[QP ] R + X[QM] P X[NP ] R + X[NQ] P X[MP ] R = −Z R P [Q XMN] P .

(9.110)
246 9 Gauged Supergravity

The important consequence of this fact is that ordinary gaugings, where one has
an ordinary Lie algebra generating the gauge group, require Z = 0. On the other
hand, whenever Z = 0, gauge invariance does not close trivially and often imposes
the introduction of tensor fields so that the gauge transformations close in a free
differential algebra.
The introduction of tensor fields is needed for two reasons. The first is that
without tensor fields one would not have gauge invariance for a generic gauging.
The second is that electric and magnetic vectors are related by non-local field
redefinitions, and in order to have a local Lagrangian, the magnetic ones should
be related to the gauge degrees of freedom of tensor fields. In fact, the full set of
2nV gauge connections must transform as in (9.97), but this implies that the set of
2nV curvatures

FM
μν = 2∂[μ Aν] + X[NP ] Aμ Aν
M M N P
(9.111)

transform as

δFM = −α N XNP M FP + Z M NP (2α N FP − AN δAP ). (9.112)

so that the first term gives a covariant transformation, while the second is non-
covariant and must be absorbed by new terms in the Lagrangian. The second
problem is manifest when one looks at the Bianchi identities for such curvatures:
 
! M = dFM + XNP M AN ∧ FP = ZPMQ AP ∧ dAQ + 1 XRS Q AR ∧ AS = 0.
dF
3
(9.113)

This is related to the failure of the Jacobi identity and in particular implies that
!
dF I = 0. This is not shocking if we think that the true gauge fields are defined
as the combinations of the AM connections singled out by the contraction with the
embedding tensor ΘM α , and such combinations are well defined because

! M XM = 0.
dF (9.114)

The solution to restore gauge covariance (and invariance) is to introduce tensor

fields, coupled to the vectors by means of the non-trivial Z coefficients
M
Hμν = FM
μν + Z NP Bμν .
M NP
(9.115)

Note that these new field strengths no longer satisfy the standard Bianchi identities,
but rather

! 1
M
∂[μ Hνρ] = Z M NP Hμνρ
NP
, (9.116)
3
9.5 Gauging and the Embedding Tensor 247

NP have to be interpreted as the field strengths of the two-forms. These

where Hμνρ
newly defined curvatures indeed transform covariantly under the gauge group by
means of the free differential algebra relations

! M
μ = ∂μ α − Z NP Σμ ,
δAM M NP
(9.117)

MN
δBμν = 2! MN
∂[μ Σν] − 2 α (M Hμν
N)
+ 2 A(M N)
[μ δAν] , (9.118)

where Σ is the gauge transformation of B, and the magnetic AM (selected by the

Z coefficients) are pure gauge under this transformation, so that H M transforms
covariantly
M
δHμν = −XP N M α P Hμν
N
. (9.119)

It is interesting to note that the coupling (9.115) creates an analogue of the Higgs
mechanism, where the vector fields are used to give mass to a set of tensor fields
(originally dual to scalar fields when massless), as the new kinetic Lagrangian is

1 1
e−1 Lkin = II J Hμν
I
H J μν + RI J Hμν
I
H J μν . (9.120)
4 4

We should also point out, though, that the H M do not form the correct symplectic
vector we would expect from our discussion on electric–magnetic duality. In fact
the correct symplectic vector transforming linearly under electric–magnetic duality
transformations is
 I
H
G =
M
, (9.121)
GI

where now

∂L
GI μν = 2 , (9.122)
∂H I μν
and HI and GI do not necessarily coincide. This also means that we would
rather have GM and not H M to be the well-defined curvatures under gauge
transformations. One therefore introduces further modifications to achieve this, but
the beauty of the whole construction is that H M and GM as well as their gauge
transformations do coincide on-shell. This happens if the tensor equations of motion
give

(H M − GM )ΘM α = 0, (9.123)
248 9 Gauged Supergravity

which correctly produces the correct identification of the dual vectors. This happens
if we introduce a topological term in the action of the form
 
1
Ltop   μνρσ Θ I α ∂μ AI ν Bα ρσ + Θ I α ΘI β Bα μν Bβ ρσ + . . . , (9.124)
8

where we used the identification Bα = −tα NP B NP , so that the tensor equations

of motion imply the covariant version of the duality equation between the massive
vector and massive tensor degrees of freedom and hence allows us to remove one in
favor of the other. Notice also that whenever we switch off the gauging, i.e., Θ = 0,
such relations disappear and the tensor fields decouple.
In fact, we should not expect that the tensor fields BμνMN we just introduced add

new degrees of freedom to the theory, but rather they should just be dual to some of
the scalar fields of the original ungauged formulation. This is indeed the case once
one consistently builds all the couplings according to this formalism. Moreover, the
presence of tensor fields fits nicely with the derivation of some of these models from
flux compactifications of string theory, where form fields are naturally present and
where the reduction process may generically give rise to massive charged tensor
fields.
The duality between tensors and scalars becomes manifest, once we build the
full duality covariant Lagrangian and analyze the equations of motion of the vector
fields:

1 μνρσ MN ∂L
 Dν GM
ρσ = Ω , (9.125)
2 ∂AM
μ

where the right-hand side arises from the minimal couplings of the vectors to the
matter fields, which clearly vanish in the ungauged theory. Originating from the
minimal couplings, this “current” term must be proportional to the embedding tensor
ΘM α , and thus we can write

1 μνρσ
 Dν GM
ρσ = Ω
MN
ΘN α Jαμ . (9.126)
2
The locality constraint tells us then that

1 μνρσ
 Dν GM
ρσ ΘM = 0,
α
(9.127)
2
which are the Bianchi identities for the gauge fields. This is consistent with our
previous discussion where we showed that the Bianchi identities for the gauge
curvature have no magnetic source term, so that the gauge connections are well
defined. For what concerns the remaining equations, the relation (9.126) becomes
a dualization equation, relating the field strength of the two-form fields Bμν MN

to the scalars, whose covariant derivative appears in the right-hand side current.
9.5 Gauging and the Embedding Tensor 249

Substituting its solution in the action, we are eliminating the tensor fields by their
equations of motion and effectively performing a rotation to the electric frame.
While this discussion shows that somehow one can always find a symplectic
rotation so that the gauging becomes electric, generic string compactifications will
naturally give rise to models with tensor fields, and therefore it definitely pays off
to be general and write the Lagrangian and couplings in a symplectically invariant
form with the help of tensor fields.

9.5.3 An Example: The Maximal Theory

As a working example of the procedure outlined in the previous sections, we now

reconstruct the main points leading to the gauging of maximal supergravity in four
dimensions [7], making explicit the steps mentioned in Sect. 9.1.
Maximal supergravity contains a single gravity multiplet, whose fields are the
graviton gμν ; 8 gravitini, ψμi (i = 1, . . . , 8); 28 vector fields, AIμ (conventionally
I = 0, . . . , 27); 56 spin 1/2 dilatini, χij k = χ[ij k] ; and 70 real scalar fields, ϕ u
(u = 1, . . . , 70).
The scalar fields describe a nonlinear σ -model given by a homogeneous manifold

E7(7)
Mscalar = . (9.128)
SU(8)

Hence, we see that the U-duality group of the theory is GU =E7(7), which has
133 generators tα . The vector fields and their duals transform in the 56-dimensional
fundamental representation of E7(7), which is a symplectic representation, defining
an embedding of E7(7) in Sp(56, R). The coset representative is customarily
described by complex 56-dimensional vectors, LM ij = −LM j i , and their complex
conjugates, LM ij , which together build a matrix
 
LM N = LM ij , LM kl . (9.129)

This matrix transforms under rigid E7(7) transformations from the left and under
local SU(8) transformations from the right. We also note the following properties of
LM N , which follow from their definition,

LM ij LN ij − LN ij LM ij = i ΩMN , (9.130)

ij
Ω MN LM ij LN kl = i δkl , (9.131)

N = 0.
Ω MN LM ij Lkl (9.132)

The gauging procedure is going to promote a subgroup of GU to Ggauge. In order

to do so, we introduce the embedding tensor that specifies which generators of E7(7)
250 9 Gauged Supergravity

are going to be part of Ggauge and which linear combinations of the vector fields are
going to be the associated gauge connections. This means introducing the covariant
derivative (where we explicitly introduce the coupling g to keep track of the gauging
terms)

!μ LM ij = ∂μ LM ij − Q
D !μ kl ij LM kl − g APμ ΘP α (tα )M N LN ij , (9.133)

where Q! represents the composite H-connection for the coset manifold, derived in
the usual way from the Maurer–Cartan form (9.43), now including a gauging term,
 
!μ kl ij = 2 i δ [i LI l]m ∂μ LI j ]m − LIl]m ∂μ Lj ]m − i g AM
Q μΩ
NP
LN ij XMP Q Lkl
Q.
[k I
3
(9.134)

The scalar kinetic terms therefore now are modified with O(g) and O(g 2 ) terms,
following from the covariantization of the vielbein on the scalar manifold,

P !μ LN kl .
!μ ij kl = i Ω MN LM ij D (9.135)

Having introduced covariant derivatives that make local some of the isometries
of E7(7), we also have to modify the gauge field strengths, according to the rules
mentioned in the previous sections. As we explained above, consistency requires
that the embedding tensor satisfies a number of linear and quadratic constraints.
How is this reflected in the construction of the Lagrangian, though? Having
introduced O(g) terms in the covariant derivatives and then in the Lagrangian,
supersymmetry is broken and should be restored by the modification of the
supersymmetry transformations and by the addition of further O(g) and O(g 2 )
terms in the Lagrangian, as mentioned in Sect. 9.1. Since all these modifications
follow from the gauging procedure, they should in some way depend on the gauge
structure constants and eventually on the embedding tensor. In fact they are actually
encoded in the tensorial structure provided by the scalar-dressed structure constants,
called T -tensor. The T -tensor is defined by

TMN P [Θ, ϕ] = L−1

M
N −1 N
LN LP P XMN P . (9.136)

This is clearly a constrained tensor, whose constraints are induced by the constraints
on the embedding tensor. For the maximal theory, we know from the previous
discussion that such constraints force Θ to live in the representation 912 of E7(7).
Once we decompose this in terms of SU(8), we get that 912 → 36+36+420+420.
This means that the T -tensor, which is a tensor with indices transforming under local
SU(8) transformations, can be decomposed in terms of two simpler complex tensors,
A1 = A1 and A2 i j kl = A2 i [j kl] , A2 i j ki = 0, which live in the representations 36
ij ji

and 420 of SU(8). This also implies that all modifications of the supersymmetry
transformations and of the Lagrangian can be written in terms of these tensors.
9.5 Gauging and the Embedding Tensor 251

Among the various constraints, we should single out a quadratic constraint called
supersymmetric Ward identity
 
1 3 nj 1 n 1 3 ij
A2 mj kl A2nj kl − A1 A1 mj = δm A2ij kl A2i j kl − A1 A1 ij . (9.137)
24 4 8 24 4

This identity is indeed necessary to close the supersymmetry algebra and, as we will
see shortly, relates the fermion O(g) shifts to the scalar potential.
As described in Sect. 9.1, we need to modify the supersymmetry transformations
of the fermions by
√ ij
δg ψμi = 2 g A1 γν j ,
(9.138)
δg χ ij k = −2g A2 l ij k  l .

This further requires the introduction of O(g) Yukawa terms4 for the fermions,

1 i 1 i
LYuk = e.g. √ A1 ij ψ μ γ μν ψνj + A2i j kl ψ μ γ μ χj kl
2 6
√ 
2 ij kpqrlm n
+  A2 pqr χ ij k χlmn + h.c., (9.139)
144

whose coefficients are precisely given by the irreducible components of the T -

tensor. Applying once more supersymmetry transformations to LYuk, one now
obtains O(g 2 ) terms, which can be cancelled by the variation of a new addition
to the Lagrangian in the form of a scalar potential,
 
2 1 3 ij
V =g A2i A2 j kl − A1 A1 ij .
j kl i
(9.140)
24 4

This cancellation happens, thanks to the supersymmetric Ward identity (9.137),

which assumes now a deeper meaning than just the consequence of a quadratic
constraint, because it relates the scalar potential to the squares of the shift of the
fermions, a general property we have seen already in Chap. 4, when dealing with
the minimal theory. It is interesting to notice that this potential can also be written
in terms of a real, symmetric and field-dependent (56 × 56)-matrix, MMN , defined
from the coset representatives,

MMN ≡ LM ij LN ij + LN ij LM ij . (9.141)

4Sometimes these terms are referred to as mass terms, because when the scalar fields pick a
vacuum expectation value they indeed generate masses for the fermion fields.
252 9 Gauged Supergravity

This is positive definite, with inverse M MN = Ω MP Ω NQ MP Q . The scalar

potential in terms of this matrix is especially interesting because it makes explicit
the quadratic dependence of the scalar potential on the embedding tensor:

1 2 
V = g XMN R XP Q S M MP M NQ MRS + 7 XMN Q XP Q N M MP .
672
(9.142)

Finally, to complete the construction, we have to replace the Abelian field

strengths in the Lagrangian by the field strengths H I , as described in the previous
section, and include the topological and Chern–Simons-like terms. Supersymmetry
is then restored by fixing an appropriate supersymmetry transformation for the
tensor fields.
The final Lagrangian, up to 4-fermi interactions, is described by
 
−1 1 1 μνρσ i ! i ←
−
!
e L = R−  ψ μ γν Dρ ψi σ − ψ μ D ρ γν ψi σ
2 2
 
1 ij k μ !
←−
ij k μ ! 1 !ij kl 2
− χ γ Dμ χij k − χ γ D μ χij k − |P |
12 12 μ

1√    
− 2 χ ij k γ ν γ μ ψν l P!μij kl + h.c. + H μν+Λ Oμν
+
Λ + h.c.
6
1  I + J +μν I − J −μν

− i NI J Hμν H − N I J Hμν H (9.143)
4
 
1 1
+ ig εμνρσ Θ I α Bμν α 2∂ρ Aσ I + g XMN I Aρ M Aσ N − g ΘI β Bρσ β
8 4
 
1 1
+ ig εμνρσ XMN I Aμ M Aν N ∂ρ Aσ I + gXP Q I Aρ P Aσ Q
3 4
 
1 1
+ ig εμνρσ XMN I Aμ M Aν N ∂ρ Aσ I + gXP QI Aρ P Aσ Q
6 4

+ e−1LYuk − V ,

+ + ij , and
where LΛij OμνΛ = 4i Oμν
√
+ ij 1 √ i [ρ 1 2 ij klnmnpq
Oμν = 2 ψ ρ γ γμν γ σ ] ψσj − ψ ρ k γμν γ ρ χ ij k −  χ klm γμν χnpq .
2 2 144
(9.144)
9.6 Classifying Gaugings 253

Looking at the Lagrangian (9.143), we notice all the structures we described in

our discussion so far. The first two lines contain the kinetic terms for the graviton,
the gravitini, the dilatini, and the scalar fields, where gauge covariant derivatives
have been introduced where necessary. The third line describes Noether couplings
between the scalars and the fermions as well as Pauli-like couplings between the
vectors and the fermions. The fourth and fifth lines describe the general structure of
the vector kinetic terms and their couplings to the fermions, now improved by the
couplings to tensor fields. The sixth line is the topological term required to be able
to properly relate the tensor fields to their scalar duals. The seventh and eighth lines
are the Chern–Simons-like couplings needed to ensure gauge invariance, and in the
last line, we have the Yukawa couplings and the scalar potential.

9.6 Classifying Gaugings

Our discussion on the gauging procedure made clear that even when we fix a choice
of gauge group, there is still the possibility that the linear and quadratic constraints
admit more than one solution, leading to gauged supergravities that are potentially
inequivalent even if they share the same set of gauge symmetry generators, tr ∈
ggauge ⊂ gU , because they differ in the choice of the (electric and magnetic) vector
fields that form the gauge connection. The aim of this section is to characterize
group-theoretically the space of these inequivalent theories, showing the relation
between the set of consistent choices of gauge connections for fixed tr ∈ ggauge ⊂
gU and symplectic transformations. In the following, we will argue heuristically for
the various requirements, but the interested reader can find a detailed proof in [8].
So, let us assume that we completely fixed once and for all the gauge group and
therefore fixed which among the various generators tα ∈ gU are selected as

tr ∈ ggauge ⊂ gU , r = 1, . . . , dim ggauge ≤ nV , (9.145)

where the last constraint is related to the existence of enough vector fields to gauge
Ggauge. To provide a consistent gauging, we therefore have to provide an embedding
tensor ΘM α , such that

ΘM α tα = θM r tr (9.146)

and obviously satisfying the necessary linear and quadratic constraints, which, for
instance, for N ≥ 3 supergravities are summarized by

[tr , ts ] = frs t tt , f[rs u ft ]u v = 0,

(9.147)
θM s frs t = −trM N θN t , θ(M r tr NP ) = 0.

The question we want to answer therefore is if there are multiple solutions of

(9.147) for θM r which are inequivalent. One can prove that all such solutions
254 9 Gauged Supergravity

are mapped onto each other by Sp(2nV , R) transformations. The answer is then
characterized group theoretically as the set of transformations of Sp(2nV , R) that
leave Ggauge invariant, which is the definition of the normalizer of Ggauge in
Sp(2nV , R): NSp(2nV ,R) (Ggauge ). In detail, if the original solution is specified by
θM r , any other solution is specified by
 
θ̂M r = NM N θN s gs−1 r , for NM N tr N P N −1 Q
= gr s ts M Q ,
P

with N ∈ NSp(2nV ,R) (Ggauge), g ∈ GL(dim Ggauge, R).

(9.148)

In particular, we are interested only in the set of transformations that change the
resulting gauge structure, which means the set of transformations that give rise to
XMN P differ from those generated by θM r . This means that we want to remove from
the previous set all the transformations in the stabilizer, S , of XMN P = θM r tr N P ,
which is represented by the quotient (here from the left)

S0 = SSp(2nV ,R) (X)NSp(2nV ,R) (Ggauge). (9.149)

The question that we are left to answer now is which among the symplectic
transformations in S0 give inequivalent theories. Before proceeding, though, we
should mention that, depending on the context, what we regard as inequivalent can
change. For instance, for our purposes it is more natural to regard as equivalent those
theories that differ from each other only in the value of the gauge coupling constant,
even if it is of course a physically relevant quantity. It is of course straightforward
to include it back. More importantly, we can decide to distinguish between theories
that have the same set of equations of motion and Bianchi identities but differ at the
quantum level or regard them as equivalent if we are only interested in the classical
regime. We will begin with the first option and therefore assume that we have fixed
the choice of an electric frame, so that we can quotient NSp(2nV ,R) (Ggauge) by the
action of local redefinitions of the physical fields only. The resulting set is also a
quotient space that we call S.

9.6.1 The Quotient Space S

As mentioned earlier, we assume for simplicity that the scalar manifold is homoge-
neous. Two ungauged Lagrangians of D = 4 supergravity are related by Sp(2nV , R)
transformations SM N acting on the GU /H coset as

L(ϕ)M N → SM P L(ϕ)P N . (9.150)

9.6 Classifying Gaugings 255

In the gauged models, the change of symplectic frame also acts on the embedding
tensor according to

XMN P → SM Q SN R XQR S (S −1 )S P . (9.151)

This ensures that the T -tensor, defined as

T (ϕ)MN P = L−1 (ϕ)M M L−1 (ϕ)N N XMN P L(ϕ)P P , (9.152)

and hence the fermionic shifts in the supersymmetry transformations as well as the
scalar potential are independent of the choice of symplectic frame. This in turn
guarantees that the combination of equations of motion and Bianchi identities is
invariant under symplectic transformations.
As it should be clear from our discussion above, any consistent embedding tensor,
θM r , can be mapped to the standard electric one, δM r , which selects the first dim
Ggauge vectors as gauge connections, by an element, N, of NSp(2nV ,R) (Ggauge). In
such reference symplectic frame, we therefore define
0 P
XMN ≡ δM r tr N P (9.153)

and call T 0 (ϕ)MN P the associated T-tensor. We then notice that if we apply an N −1
transformation only to the coset representatives, namely,

L(ϕ)M N → N −1 M P L(ϕ)P N , 0 P
XMN unchanged, (9.154)

then the T -tensor transforms as

0 P
TMN ≡ L−1 M M L−1 N N XMN
0 P
LP P

N −1
→ L−1 M M L−1 N N NM Q NN R XQR
0 S
N −1 S P LP P

= TMN
θ P
. (9.155)

As a result, the gauge kinetic functions and moment couplings transform accord-
ingly with the N −1 symplectic transformation. Clearly the equations of motion and
Bianchi identities are not invariant under (9.154), as is reflected by the fact that the
T -tensor changes. We then interpret (9.154) as a symplectic deformation, namely,
a map between two (potentially) inequivalent gauged models. The requirement
N ∈ NSp(2nV ,R) (Ggauge) ensures that tr are a good choice of gauge generators also
after the symplectic deformation, i.e., they belong to the gU algebra of both the old
and the new symplectic frame.
256 9 Gauged Supergravity

Let us then see which transformations give inequivalent models. One obvious
field redefinition we are allowed to make is a local redefinition of the vector
fields (i.e., a redefinition involving only the vector fields appearing in the original
Lagrangian). We will not change theory if by such a redefinition we rescale at most
the structure constants by an overall factor, i.e., if such transformations are in the
stabilizer of X0 , SGL(nV ,R) (X0 ). The other redefinitions that should not modify the
theory are those U-duality transformations that leave the gauge group invariant,5
namely, the transformations in the normalizer of Ggauge within GU : NGU (Ggauge ).
The combination of the two is indeed what we are looking for.
Take then two transformations N, N  ∈ NSp(2nV ,R) (Ggauge), related by

N = u N  s, u ∈ NGU (Ggauge), s ∈ SGL(nV ,R) (X0 ). (9.156)

Substituting in (9.155), we get

N −1
0 P
TMN → (L−1 u N  s)M M (L−1 u N  s)N N XMN
0 P
(s −1 N −1 u−1 L)P P ,
(9.157)

and at the same time the vector kinetic terms and moment couplings transform
with N −1 . The GU transformation uM N can be reabsorbed in the scalar fields, and
therefore it does not affect the physics. Since we have required that the action of s
on X0 is trivial up to an overall rescaling, so that
0 S −1 P
sM Q sN R XQR sS ∝ XRN
0 P
, (9.158)

we can reabsorb the rescaling in the gauge coupling constant, and similarly sM N can
be reabsorbed in a local field redefinition of the electric vectors AΛ
μ in the covariant
derivatives and in the non-minimal couplings:

Aμ I → Aμ J sJ I . (9.159)

We conclude that N and N  in (9.156) define the same gauged theory up to local
field redefinitions and rescalings of the gauge coupling constant.
We therefore arrive at the result that symplectic deformations are classified by
the space

S ≡ SGL(nV ,R) (X0 ) \ NSp(2nV ,R) (Ggauge) / NGU (Ggauge ), (9.160)

5Actually one should enlarge the possible transformations to all automorphisms of the U-duality
group. While this coincides with the U-duality group itself in most cases, there are instances
where additional discrete factors become relevant. A particularly relevant example is maximal
supergravity, where there is an additional Z2 parity symmetry that can be used.
9.6 Classifying Gaugings 257

where the quotients correspond to local field redefinitions. Notice that this definition
carries a dependence on the initial choice of electric frame, to the extent that
such choice affects the explicit form of XMN 0 P (for instance, it can affect the

Chern–Simons-like couplings in the gauge generators). Therefore we must specify

the explicit form of X0 that we use to compute S, or equivalently the specific
choice of electric frame in which we construct the gauged theory whose symplectic
deformations we want to compute.
Once again, we can make this explicit in the example of an N = 2 super-
gravity theory coupled to a vector multiplet, with homogeneous scalar manifold
SU(1,1)/U(1), parametrized as in Sect. 9.4.2. Let us look once again at the shift
symmetry of the imaginary part of the z field C → C + α. This is generated
by t+ ∈ su(1, 1), which we have seen becomes a perturbative symmetry in the
symplectic frame defined by the symplectic transformation S in (9.88). We therefore
analyze the sp(4, R) generators in this basis, which is related to the one in Sect. 9.4.2
by tI → S −1 tI S. In this basis, the generators of SU(1,1) take the form

1
{σ1 ⊗ σ1 , −i σ2 ⊗ 1, −σ3 ⊗ 1} (generators of su(1, 1)) , (9.161)
2
whereas the generators for the vector field redefinitions, GL(2,R), read

1
{−σ3 ⊗ 1, σ3 ⊗ σ3 , −σ3 ⊗ σ1 , −1 ⊗ iσ2 } (generators of gl(2, R)) .
2
(9.162)

Obviously, the first generator is in common between gl(2,R) and su(1, 1). The
remaining generators of the quotient,

GL(2, R)\Sp(4, R)/SU(1, 1), (9.163)

are
1
{σ1 ⊗ σ3 , i σ2 ⊗ σ3 , σ1 ⊗ σ1 , i σ2 ⊗ σ1 } . (9.164)
2
Clearly, there are no generators in su(1, 1) that commute with t+ and therefore
NSU(1,1)(U(1)t+ ) = U(1)t+ . On the other hand, there are generators in sp(4, R) that
commute with t+ , so that
 
NSp(4,R) U(1)t+ = t • , t↑ , t↓ , (9.165)
258 9 Gauged Supergravity

where
1
t• = − 1 ⊗ i σ2 , (9.166)
2
1
t↑ = (σ1 ⊗ σ1 − i σ2 ⊗ σ1 ) , (9.167)
2
1
t↓ = (σ1 ⊗ σ3 − i σ2 ⊗ σ3 ) . (9.168)
2
These generators satisfy the algebra of the isometries of the Euclidean plane, E2 ,

[t• , t↑ ] = t↓ , [t• , t↓ ] = −t↑ , [t↑ , t↓ ] = 0. (9.169)

We see that in this symplectic frame t• is in gl(2, R) and hence as generators

of inequivalent gaugings we are left with the generators t↑ and t↓ . As already
discussed at the end of Sect. 9.4.2, these are transformations that generate total
derivative terms, and hence we really have inequivalent gauged Lagrangians only
at the quantum level. There are also examples where the inequivalent gaugings
have a more sizeable modification to the Lagrangian, introducing new couplings
and modifying the scalar potential and its critical points. The most striking example
is given by the maximal supergravity with gauge group SO(8), which we therefore
discuss in the next section.

9.6.2 The S Space of SO(8) Maximal Gauged Supergravity

In this section we want to repeat the discussion of the S space for the SO(8)
gauging of maximal supergravity, to show that there actually is an infinite family
of inequivalent supergravity theories with the same gauge group [9].
Let us then consider the SO(8) gauged maximal supergravity, taken in its
standard electric frame with SL(8, R) as electric group. This is the standard form
of this theory as it is obtained, for instance, by reduction of M-theory on the seven
sphere S 7 [10]. The space of inequivalent theories is then

S = NGL(28,R) (SO(8)) \ NSp(56,R) (SO(8)) / NZ2 E7(7) (SO(8)), (9.170)

where in the last factor we used the full group of automorphisms of the U-duality
group, which for the maximal theory includes an additional Z2 parity. Moreover, we
replaced SGL(28,R) (X0 ) with NGL(28,R) (SO(8)) in this expression, because they
coincide and therefore we can remove any explicit reference to the embedding
tensor.
9.6 Classifying Gaugings 259

A detailed computation [8] shows that this reduces to the following quotient

S = GL(1, R) × S3 \ SL(2, R) × S3 / D8 , (9.171)

where S3 is the triality discrete automorphism of SO(8), and the SL(2, R) in

NSp(56,R) can be parameterized as follows:
 
λ
Gλ ≡ ⊗ 128 , λ ∈ R \ {0}, (9.172)
λ−1
 
1 −g 2 θ/2π
Wθ ≡ ⊗ 128 , θ ∈ R, (9.173)
1
 
cos ω − sin ω
Uω ≡ ⊗ 128 , ω ∈ [0, 2π], (9.174)
sin ω cos ω

in the basis where the SO(8) generators are given by

 
Λr
so(8) - tr = , (9.175)
Λr

with Λr the SO(8) generators in the adjoint representation.

In order to understand this, let us make contact with the embedding tensor
formalism. The consistency constraints on the embedding tensor require that it is a
singlet under SO(8). In fact, the 912 E7(7) representation in which ΘM α sits contains
two SO(8)-singlets in its manifestly triality-invariant decomposition:

912 → 36 + 36 + 420 + 420 (9.176)

→ 1θ + 1ξ + 2 · (35v + 35s + 35c + 350). (9.177)

The subscripts “θ ” and “ξ ” denote the relation to the symmetric tensors θij and ξ ij
that (when positive-definite) define the SO(8) generators inside SL(8, R) and that
we will always assume to be in the standard form θij ∝ ξ ij ∝ δij . The original
SO(8) gauged maximal supergravity [10] corresponds to θij ∝ δij , ξ = 0, and it is
electrically gauged in the standard formulation. What we call X0 corresponds to this
particular embedding tensor. The “ω-deformed” SO(8) gaugings are then defined
by turning on ξ = 0, and they are no longer electric in the original symplectic
frame. This is clearly achieved in the above parametrization by acting on X0 with the
matrix Uω . Following our analysis in Sect. 9.6.1, we prefer to regard the symplectic
deformations as leaving X0 unchanged, but acting on the coset representatives, thus
yielding the deformed theories in their respective electric frames.
260 9 Gauged Supergravity

In practice, most of the transformations in SL(2, R)×S3 either leave X0 invariant

up to a rescaling of the gauge coupling constant, g → λg, and their effect on the
kinetic terms can be reabsorbed in a local redefinition of the vector fields, or have
no effect on X0 altogether, but with a non-trivial effect on the vector kinetic term of
the form
θ
N (ϕ)ΛΣ → N (ϕ)ΛΣ + g 2 δΛΣ . (9.178)
2π

This transformation clearly represents a constant, SO(8) invariant shift in the θ -

angle of the gauge theory; hence, it has no effect on the (classical) equations of
motion and supersymmetry variations. In fact, it is clear that we can always add a
term ∝ δΛΣ F Λ ∧ F Σ to the gauged SO(8) electric action, and the analysis above
proves that there is no E7(7) transformation or local field redefinition that can remove
it. The transformation that has the most striking effect is instead the ω-deformation
of the SO(8) gauged maximal supergravity [9]. If we keep θ = 0, one indeed
reproduces the known parameter space for the ω-deformation of the SO(8) theory,
namely, S 1 /D8 , with identifications ω  ± ω + kπ/4, k ∈ Z and fundamental
domain ω ∈ [0, π/8]. If we include θ , the S-space of symplectic deformations of
SO(8) gauged maximal supergravity, in its standard electric frame, is a quotient of
a hyperboloid: (dS2 /Z8 )/Z2 . If we also impose periodicity in θ , the resulting space
has the topology of a two-sphere.
This space of inequivalent theories becomes evident if we analyze their couplings
and in particular their scalar potential. All such inequivalent models have an N = 8
vacuum with a negative cosmological constant, and obviously the quadratic spectra
around such vacua coincide. However, higher-order couplings change, as expected
for inequivalent models. We will now show explicitly some of the couplings and
compute their dependence on the parameter ω, with a special emphasis on the
scalar potential, which now shows a different spectrum of vacua according to the
parameter’s choice.
For the sake of clarity, we restrict the analysis of the potential to the G2 -invariant
sector of the theory. It is known that for ω = 0 one finds one N = 8 vacuum with
SO(8) symmetry, two parity conjugated vacua with N = 0 and SO(7)− residual
symmetry, another N = 0 vacuum with SO(7)+ symmetry, self-conjugate under
parity, and two parity conjugated N = 1 vacua with G2 symmetry [11]. The G2 -
invariant truncation contains two real scalar fields, ϕ = (ϕ1 , ϕ2 ), and the potential
can be written as the sum of three pieces,

V (ϕ)  − cos(2ω)f (ϕ1 , ϕ2 ) − sin(2ω)f (ϕ2 , ϕ1 ),

 = A(ϕ) (9.179)
9.6 Classifying Gaugings 261

where (in the following x ≡ e|ϕ|

)


(1+x 4 )3
 =
A(ϕ)  4 x 14
64|ϕ|
4(1 + x 4 )2 (1 − 5x 4 + x 8 )(ϕ14 + ϕ24 )

+ϕ12 ϕ22 (1 + 4x 4 − 106x 8 + 4x 12 + x 16) , (9.180)

which is an even function of ϕ1 and ϕ2 and symmetric in their exchange, and

(−1 + x 4 )5 ϕ13 
f (ϕ1 , ϕ2 ) = 4(1 + 5x 4 + x 8 )ϕ14 + (9.181)
 7 x 14
64|ϕ|

+ 7(1 + 6x 4 + x 8 )ϕ12 ϕ22 + 7(1 + x 4 )2 ϕ24 , (9.182)

which is odd in the first argument and even in the second. Three symmetry
operations leave the scalar potential invariant:
⎧
5 5 ⎪
⎪ ω ↔ ω − π4
ω ↔ −ω ω↔ω+ π ⎨
2
, , ϕ1 → ϕ2 . (9.183)
ϕ2 ↔ −ϕ2 ϕ ↔ −ϕ ⎪
⎪
⎩
ϕ2 → −ϕ1

The first one results from a parity-related symmetry, while the last two result
from E7(7)-duality transformations. Altogether this implies that we get inequivalent
potentials only in the expected range ω ∈ [0, π/8]. In fact, depending on the
parameter ω, the scalar potential exhibits a different number of vacua, as shown in
Fig. 9.3. The ω = 0 case corresponds to the usual truncation of the scalar potential
that keeps the SO(8) vacuum (although seemingly unstable, all the masses satisfy

Fig. 9.3 Scalar potential of the G2 truncation for ω = 0 (left) and for ω = π/8 (right). The red
dot is the SO(8) vacuum, the blue squares are vacua with SO(7) symmetry, and orange triangles
represent vacua with G2 residual gauge symmetry. New SO(7) and G2 vacua appear with respect
to the ω = 0 case
262 9 Gauged Supergravity

the Breitenlohner–Freedman bound), the SO(7)± vacua, and the G2 ones. When
ω = 0, a new SO(7) vacuum and new G2 vacua appear. In fact, not only the number
of vacua changes when ω = 0, but also the value of their cosmological constant,
as can be seen by looking at Fig. 9.3. In particular, the ratio of the value of the
cosmological constant of the various vacua in the two potentials with respect to that
of the N = 8 vacuum in the center is an ω-dependent function, different for each
one of the vacua.

References
1. M.K. Gaillard, B. Zumino, Duality rotations for interacting fields. Nucl. Phys. B193, 221
(1981). Dedicated to Andrei D. Sakharov on occasion of his 60th birthday
2. S. Helgason, Differential Geometry and Symmetric Spaces (Academic Press, New York, 1962)
3. H. Nicolai, H. Samtleben, Maximal gauged supergravity in three-dimensions. Phys. Rev. Lett.
86, 1686–1689 (2001) [arXiv:hep-th/0010076 [hep-th]]
4. H. Samtleben, Lectures on gauged supergravity and flux compactifications. Class. Quant. Grav.
25, 214002 (2008) [arXiv:0808.4076 [hep-th]]
5. G. Dall’Agata, R. D’Auria, S. Ferrara, Compactifications on twisted tori with fluxes and free
differential algebras. Phys. Lett. B 619, 149–154 (2005) [arXiv:hep-th/0503122 [hep-th]]
6. J. De Rydt, T.T. Schmidt, M. Trigiante, A. Van Proeyen, M. Zagermann, Electric/magnetic
duality for chiral gauge theories with anomaly cancellation. JHEP 12, 105 (2008)
[arXiv:0808.2130 [hep-th]]
7. B. de Wit, H. Samtleben, M. Trigiante, The maximal D = 4 supergravities. JHEP 06, 049 (2007)
[arXiv:0705.2101 [hep-th]]
8. G. Dall’Agata, G. Inverso, A. Marrani, Symplectic deformations of gauged maximal super-
gravity. JHEP 07, 133 (2014) [arXiv:1405.2437 [hep-th]]
9. G. Dall’Agata, G. Inverso, M. Trigiante, Evidence for a family of SO(8) gauged supergravity
theories. Phys. Rev. Lett. 109, 201301 (2012) [arXiv:1209.0760 [hep-th]]
10. B. de Wit, H. Nicolai, Phys. Lett. B108, 285 (1982); B. de Wit, H. Nicolai, Nucl. Phys. B208,
323 (1982)
11. N.P. Warner, Some new extrema of the scalar potential of gauged N = 8 supergravity. Phys.
Lett. B 128, 169 (1983)
12. M. Trigiante, Gauged supergravities. Phys. Rept. 680, 1–175 (2017) [arXiv:1609.09745 [hep-
th]]
Supergravity in Arbitrary Dimensions
10

While in the first nine chapters of this book, we focused on four-dimensional

theories, we use this final chapter to give a brief outlook on supergravity theories in
a number of dimensions other than four. Many details of their structure differ from
the four-dimensional case, but the main terms and relations remain valid. We begin
by analyzing the eleven- and ten-dimensional theories, as they are directly related to
the low-energy limit of string theory and then discuss the five-dimensional models,
which are very useful in the context of the gauge/gravity correspondence.

10.1 Higher-Dimensional Theories

N = 1 supersymmetry in 4D admits a short representation containing only the

graviton, gμν , and the gravitino, ψμ . On-shell, both fields have two degrees of
freedom, and we can use them to describe a pure supergravity theory. As we have
seen in Chap. 4, the presence of the gravitino provides a lot of new interesting
features and constraints to the standard gravity theory described by the Einstein–
Hilbert action. This model can then be further coupled to matter fields resulting in
even richer and more interesting theories. However, the minimal model can contain
just the two basic fields gμν and ψμ .
For supergravity models in arbitrary dimensions higher than D = 4, the situation
changes. The Lorentz group becomes SO(1,D − 1), and hence massless physical
states are classified by SO(D − 2) representations. This means that while in four
dimensions we can classify states by their helicities, in higher dimensions one has
generically many more states, and the representations of supersymmetry will no
longer admit short representations with only the graviton and gravitino fields. For a
generic D, the graviton field gμν describes a total of

(D − 2)(D − 2 + 1)
−1 (10.1)
2

© Springer-Verlag GmbH Germany, part of Springer Nature 2021 263

G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1_10
264 10 Supergravity in Arbitrary Dimensions

Table 10.1 The possible values for η and  together with the resulting minimal
spinor types, the minimal number of real supercharges, and the general form of the R-
symmetry groups (M = Majorana, SM = Symplectic Majorana, W = Weyl, MW = Majorana–
Weyl, SMW = Symplectic Majorana–Weyl)
D η  Min. spinor type Min. # of real supercharges R-symmetry group
2 +1 +1 MW 1 SO(NL ) × SO(NR )
−1 +1
3 +1 +1 M 2 SO(N)
4 +1 +1 M or W 4 U (N)
−1 −1
5 −1 −1 SM 8 U sp(2N)
6 +1 −1 SMW 8 U sp(2NL ) × U sp(2NR )
−1 −1
7 +1 −1 SM 16 U sp(2N)
8 +1 −1 16 U (N)
−1 +1 M or W
9 −1 +1 M 16 SO(N)
10 +1 +1 MW 16 SO(NL ) × SO(NR )
−1 +1
11 +1 +1 M 32 SO(N)
12 +1 +1 M or W 64 U (N)
−1 −1
... ... ... ... ... ...

on-shell degrees of freedom (the metric fluctuations are symmetric traceless matrix
states). At the same time, the number of degrees of freedom of a spinor representa-
tion grows even faster with D, being

# dof = k 2[
D/2]−1
, (10.2)

with k = 2 for Dirac spinors, k = 1 for Majorana spinors and for Weyl spinors,
and k = 1/2 for Majorana–Weyl spinors (cf. Table 10.1 and Appendix 10.A). This
implies that the gravitino ψμ has

k 2[
D/2]−1
(D − 2 − 1) (10.3)

states, where the (D − 2 − 1) factor comes from the vector index and the fact
that the Rarita–Schwinger action is invariant under δψμ = ∂μ λ and one has to
remove the auxiliary spinor γ μ ψμ . By a simple comparison of (10.1) with (10.3),
we immediately see that only in four dimensions, one can simply match bosonic
and fermionic degrees of freedom in a multiplet by using only the graviton and
gravitino fields. As soon as one moves to higher dimensions, one needs more fields,
10.2 Example: D = 11 Supergravity 265

both bosonic and fermionic ones. For instance, in five dimensions the graviton has
five degrees of freedom and the gravitino has eight (one has to take a Dirac spinor;
no Weyl or Majorana spinors are allowed.1 ) To complete the graviton multiplet,
the matching of bosonic and fermionic states thus requires three additional bosonic
degrees of freedom represented by a massless vector field, Aμ , the graviphoton. This
is just a special example of what is needed to construct a full supergravity multiplet
in higher dimensions: antisymmetric tensor fields. These are rank n fields Bμ1 ...μn
with complete antisymmetry of their indices and a tensor gauge invariance,

δBμ1 ...μn = n ∂[μ1 Λμ2 ...μn ] . (10.4)

The vector field is a special instance where n = 1, and we usually don’t see higher-
rank tensor fields in four-dimensional theories because for n = 2 they are equivalent
to scalar fields (as long as they are massless) and for n = 3, 4 they have no physical
states. However, in D dimensions the number of physical states of a rank-n tensor
is

(D − 2)!
(10.5)
(D − 2 − n)!n!

and they are further reducible into the self- and anti-self-dual parts when n = D/2.
This means they can play a fundamental role to provide the necessary bosonic
degrees of freedom needed to complete a supergravity multiplet.

10.2 Example: D = 11 Supergravity

In four and in any other dimension, the maximal number of real supercharges
allowed in constructing a theory with fields of spin ≤ 2 is 32. A Majorana spinor
in 11 dimensions has 25 = 32 components, and hence 11D supergravity is the
highest-dimensional supergravity model that can be constructed without introducing
higher-spin fields. The 11D supersymmetry algebra naturally contains a central
charge,

{Qα , Qβ } = (CΓ )m
αβ Pm + (CΓ )αβ Zmn ,
mn
(10.6)

which is associated with the existence of membrane-like objects in the theory.

1 In five dimensions, one can combine two Dirac spinors and impose a symplectic Majorana
condition, but the resulting number of degrees of freedom is as for a Dirac spinor. For more details,
see Appendix 10.A.
266 10 Supergravity in Arbitrary Dimensions

The massless 11D graviton has 44 degrees of freedom, while the gravitino has
128 physical states. This implies the need of additional higher-rank tensor fields. It
is actually easy to see that a three-form, Cμνρ , with a gauge transformation

δCμνρ = 3 ∂[μ Λνρ]

has exactly the 84 missing physical states needed to complete a supermultiplet

{eμa , ψμ , Cμνρ }. (10.7)

This theory clearly has no scalar potentials (there are no scalars at all), and it has
a unique parameter given by the 11D gravitational constant, κ11 .
The action of 11-dimensional supergravity was constructed first by Cremmer,
Julia, and Scherk [1] and consists of very few terms:
   
1 ω+! ω 1
S = 2
d 11
x e R(ω) − ψ μ Γ μνρ
D ν ψρ − Gμνρσ Gμνρσ
2κ11 2 24
√
2 2 μ1 ...μ11
−  Gμ1 ...μ4 Gμ5 ...μ8 Cμ9 μ10 μ11
(144)2
√   &
2 ρ στ η 3√
− ψ μΓ μνρσ τ η
ψν + 12 ψ Γ ψ 2 Gρσ τ η − 2κ ψ [ρ Γσ τ ψη] ,
192 2
(10.8)

where

Gμνρσ = 4 ∂[μ Cνρσ ] , (10.9)


1
ωμab = ωμab (e) − ψ Γμab αβ ψβ + 2ψ μ Γb ψa
8 α

−2ψ a Γμ ψb + 2ψ b Γa ψμ , (10.10)

1
ωμab = ωμab −
! ψ Γb ψa − ψ a Γμ ψb + ψ b Γa ψμ . (10.11)
4 μ
This action obviously includes the kinetic terms for the graviton, the gravitino,
and the three-form field C in the first line. The modified spin connection !
ω has been
introduced to take into account the four-Fermi interactions between the gravitino
fields. The second line is a special Chern–Simons-like term, which does not depend
on the metric. The last line, finally, contains further interaction terms between the
10.3 Dimensional Reduction and Ten-Dimensional Supergravities 267

three-form and the gravitino. The invariance under supersymmetry follows from the
application of

1
δeμa = Γ a ψμ , (10.12)
2
√ 
2
δψμ = Dμ (!
ω) + Γμ νρσ τ − 8δμν Γ ρσ τ Gνρσ τ
288

3√
+ 2 ψ [ν Γρσ ψτ ] , (10.13)
2

3√
δCμνρ = − 2 Γ[μν ψρ] . (10.14)
4
We don’t discuss here the proof of invariance under supersymmetry of this
action nor the peculiar new features appearing, thanks to the presence of a three-
form field. On the other hand, we are going to use this theory as a starting point
for a qualitative discussion of the features of the models that can be obtained by
dimensional reduction.

10.3 Dimensional Reduction and Ten-Dimensional

Supergravities

The 11-dimensional supergravity action we just presented should describe the low-
energy limit of M-theory, a supposedly consistent quantum theory of membranes
in 11 dimensions, which arises as the strong coupling limit of ten-dimensional type
IIA string theory. It is therefore conceivable that one could show a detailed relation
between supergravity in 11 dimensions and the low-energy limits of ten-dimensional
string theory models. Hence, we will now discuss the Kaluza–Klein reduction of 11-
dimensional supergravity to ten dimensions and less.
A simple way to dimensionally reduce a theory is to consider a spacetime metric
where one of the coordinates runs on a circle of fixed radius. In this way, the
fluctuations of the fields over this coordinate will be constrained by the geometry
and will result in effective masses and couplings for the fluctuations in the rest of the
spacetime. For instance, if we reduce 11D supergravity over a circle and keep only
the massless modes in the resulting effective 10D theory, we obtain a much richer
spectrum than the one we discussed in the previous section. If we split the 11D
coordinates, x M , into the non-compact ones, x μ , and the circle coordinate, θ ≡ x 10 ,
we see that the reduction of the 11D metric gives rise to three different fields, with
268 10 Supergravity in Arbitrary Dimensions

different transformation properties with respect to the residual 10D diffeomorphism

invariance:
⎧
⎪
⎪ gμν ten-dimensional metric,
⎨
gMN → gμ10 ∼ Aμ ten-dimensional vector, (10.15)
⎪
⎪
⎩
g10 10 ∼ φ ten-dimensional scalar.

From the 11-dimensional line element,

2
ds11 = e2φ(x)(dθ + dx μ Aμ (x))2 + ds10
2
, (10.16)

we see that the ten-dimensional photon field, Aμ , describes the fibration of the circle
on the ten-dimensional base spacetime and the dilaton field, φ, is associated with the
radius of the circle of the internal direction. In the same fashion, we can reduce the
11-dimensional rank-3 tensor,
⎧
⎪
⎪ Cμνρ ten-dimensional three-form,
⎨
CMNP → Cμν10 ≡ Bμν ten-dimensional two-form, (10.17)
⎪
⎪
⎩
Cμ10 10 = 0 vanishing because of antisymmetry,

and the 11-dimensional gravitino,

ψμ+ , ψμ− Two ten-dimensional gravitini

ΨM → (10.18)
+ −
ψ10 ≡ λ+ , ψ10 ≡ λ− Two ten-dimensional spin 1/2 fields.

The plus and minus signs on the fermions refer to the chirality of the resulting
ten-dimensional spinor fields. In fact, the 32-component Majorana spinor in 11D
reduces to a 32-component Majorana spinor in 10D, which, however, can be further
decomposed into two 10D Majorana–Weyl spinors of opposite chiralities, each
having 16 independent real components. The same is true for the supercharges,
Q11 → Q+ , Q− , which split into two Majorana–Weyl representations of
opposite chirality in ten dimensions. The resulting model is therefore a non-chiral
ten-dimensional supergravity theory with two supersymmetry generators: type IIA
supergravity.
The splitting of the 11-dimensional supercharge implies that one can construct
other supergravity models with only one supercharge (type I models) or with both
supercharges of the same chirality (type IIB supergravity). It is actually useful to
summarize the resulting spectrum of type IIA supergravity as follows (the numbers
10.3 Dimensional Reduction and Ten-Dimensional Supergravities 269

indicate the independent on-shell degrees of freedom of each field):

{ φ , λ− , Bμν , ψμ+ , gμν } Common Sector

1 8 28 56 35 = 64B + 64F
(10.19)
{ λ+ , Cμνρ , Aμ , ψμ− } IIA RR sector
8 56 8 56 = 64B + 64F .

The first line is a consistent 10D supermultiplet by itself when one restricts oneself
to a single supersymmetry, while the second line collects the fields that complete the
multiplet to the type IIA multiplet. The type IIA supergravity theory based on the
above fields arises as the low energy limit of type IIA string theory. In the context of
this string theory, the fields in the first line are referred to as the Neveu–Schwarz–
Neveu–Schwarz (NSNS) sector, whereas the fields in the second line are called the
Ramond–Ramond (RR) sector.
As the fields in the first line of (10.19) form a consistent supermultiplet of the
10D superalgebra with one supersymmetry generator, it is natural to suspect that
there is also a consistent supergravity action with this reduced field content. This
action indeed exists and is referred to as type I supergravity. Type I supergravity
by itself, however, has quantum anomalies, and in order to cancel these anomalies,
one has to add suitable super-Yang–Mills sectors based on 10D vector multiplets.
We will come back to this point and its relation to string theory at the end of this
section.
The Lagrangian of type IIA supergravity is obtained by dimensional reduction
of the one of the 11-dimensional theory presented in the previous section [2]. For
instance, from the kinetic term of the four-form, G, we obtain the kinetic terms of
the ten-dimensional four- and three-forms, F4 = dC3 + A ∧ H3 and H3 = dB,
respectively:

1 1 1
GMNP Q GMNP Q → Fμνρσ F μνρσ + e−2φ Hμνρ H μνρ . (10.20)
48 48 12
The 11D Ricci scalar reduces to the 10D Ricci scalar, plus kinetic terms for the
additional 10D degrees of freedom obtained from the 11D metric, i.e., for the 10D
vector and scalar fields. It is important to point out, however, that the presence of the
metric determinant, together with the direct reduction of the curvature term, gives
rise to dilaton factors in front of the various 10D terms, so that we cannot view the
resulting action as a standard Einstein gravity theory unless we perform a rescaling
of the metric. More concretely, the straightforward reduction of the Einstein–Hilbert
term gives
 
−2φ 1 2φ 2 2
e11 R11 → e10 e R10 − e F + 4(∂φ) . (10.21)
4
270 10 Supergravity in Arbitrary Dimensions

This has an apparent wrong sign for the kinetic term of the dilaton and at the same
time a non-trivial scalar factor in front of the ten-dimensional Einstein–Hilbert term.
To go to the Einstein frame, one performs a rescaling
φ
gμν → e 2 gμν , (10.22)

so that (10.21) gets mapped to the more standard form

 3

e2φ 2 1
e10 R10 − F − (∂φ)2 . (10.23)
4 2

Once we reduced the theory to ten dimensions, we see that we now have two
parameters: k11 and the vev of eφ . The field φ is the first modulus we meet, i.e.,
a massless scalar field whose expectation value is related to some geometrical
property of the internal spacetime and which affects the coupling constants of the
effective theory. In fact, the scalar potential of type IIA supergravity is trivial,2

V (φ) ≡ 0. (10.24)

The full bosonic sector of the type IIA supergravity action in the Einstein frame
reads:
  
1 10 1 1 −φ
SI I A = 2 d x e R − ∂μ φ∂ φ − μ
e Hμνρ H μνρ
2κ10 2 12
  
1 1 φ 1 3
− 2
d 10 x e e 2 Gμνρσ Gμνρσ − e 2 φ Fμν F μν (10.25)
2κ10 48 4

1
− 2 B2 ∧ dC3 ∧ dC3 ,
4κ10

where

G4 = dC3 − A1 ∧ H3 , H3 = dB2 , F2 = dA1 . (10.26)

In our simple reduction, the expectation value of φ is related to the radius of the S 1
we used to go from eleven to ten dimensions. If we interpret the IIA supergravity as
a low energy limit of type IIA string theory, we can identify the vev of the dilaton
with the string coupling constant as eφ  ∼ gs . It is then clear why taking a strong

2 There exists a mass-like deformation of type IIA supergravity, called massive type IIA supergrav-
ity, which does contain a scalar potential for the dilaton [3].
10.3 Dimensional Reduction and Ten-Dimensional Supergravities 271

coupling limit of type IIA strings, gs → ∞, implies to go from type IIA to M-

theory. In our supergravity setup, which, however, cannot be fully trusted when one
goes beyond the classical regime, this corresponds to send the radius of the S 1 to
infinity and therefore to go back to an 11-dimensional background.
In order to complete this quick survey of ten-dimensional supergravity models,
we just mention a couple of things. Minimal supergravity in ten dimensions (Type I)
can be coupled to matter. In fact ten-dimensional (N = 1) supersymmetry allows
for another multiplet: a super-Maxwell multiplet, containing a vector and a spin 1/2
field:

{ Aμ , λ± }.
(10.27)
8 8

While at the classical supergravity level we could consider coupling an arbitrary

number of such multiplets to the supergravity action, anomaly cancellation restricts
the allowed possibilities to well-defined gauge groups and therefore to a well-
defined and fixed number of vector multiplets. This is also reflected in the
corresponding string theories with 10D, N = 1 supersymmetry, namely, the type I
string theory and the two heterotic string theories. Their low energy limits consist of
type I supergravity coupled to N = 1 super-Yang–Mills theory with gauge group
SO(32) for the type I string theory and with gauge group SO(32) or E8 × E8 for the
two heterotic string theories.3
As we discussed before, we could also consider an N = 2 theory with both
supercharges of the same chirality. This is called type IIB supergravity [4,5], and its
field content can be obtained by substituting the RR sector in Eq. (10.19) with the
following:

{ C0 , λ−(2) , Cμν , Cμνρσ , ψμ+ (2) } IIB RR sector

(10.28)
1 8 28 35 56 = 64B + 64F .

We point out that the rank-4 tensor field appearing in the spectrum has only
35 degrees of freedom on shell, because its field strength is a five-form in ten
dimensions, and hence one can impose a self-duality constraint on it,

1
F5 = F5 , Fμ1 ...μ5 = μ ...μ F μ6 ...μ10 . (10.29)
5! 1 10

3Despite the presence of gauge interactions and half-maximal supersymmetry, these theories do
not have a scalar potential, which can be traced back to the gauge group being unrelated to the
R-symmetry group or any scalar field isometries and to the fact that there are no scalars in the
vector multiplets.
272 10 Supergravity in Arbitrary Dimensions

This constraint creates some problems when one wants to construct a Lorentz
covariant action because, as we can easily see, the standard kinetic term would be
identically zero

F5 ∧ F5 = F5 ∧ F5 = −F5 ∧ F5 = 0, (10.30)

where we first used the constraint and then swapped the order of the two five-
forms. A covariant and supersymmetric Lagrangian with a single scalar auxiliary
field, which is a pure gauge of a new symmetry of the action and a singlet under
supersymmetry, is provided in [6]. As a mnemonic tool, one can use the following
bosonic Lagrangian, where the self-duality equation is imposed on the resulting
equations of motion,
  
1 1 1 −φ
SI I A = 2
d 10 x e R − ∂μ φ∂ μ φ − e Hμνρ H μνρ
2κ10 2 12
  
1 1 2φ 1
− 2 10
d xe e ∂μ C0 ∂ μ C0 + Fμ1 ...μ5 F μ1 ...μ5
2κ10 2 2 · 5!

1 1 φ
− 2
d 10 x e e Fμνρ − C0 Hμνρ F μνρ − C0 H μνρ
2κ10 12

1
− 2 C4 ∧ H3 ∧ F3 ,
4κ10
(10.31)

where
1 1
H = dB2 , F3 = dC2 F5 = dC4 − C2 ∧ H3 + B2 ∧ F3 . (10.32)
2 2
This theory arises as the low energy limit of type IIB string theory.

10.4 Dimensional Reduction and the Origin of Gauged

Supergravities

In the same way as type IIA supergravity arises by reduction of 11-dimensional

supergravity on a circle, one could try to produce many different models in four
dimensions by compactifying ten- or eleven-dimensional supergravities on six- or
seven-dimensional spaces with various geometries. Obviously, the matter content
and the number of supersymmetries will depend on the vacuum expectation value
of the metric on the internal space. However, following what happened in the
IIA case, these reductions (when they preserve some supersymmetry) will usually
generate a number of massless scalar fields associated with the shape and volume
10.4 Dimensional Reduction and the Origin of Gauged Supergravities 273

of the internal space, just like the IIA dilaton was associated with the volume
(the radius) of the circle on which one compactifies M-theory. As discussed in
detail in Sect. 7.4.1, these moduli fields are a serious phenomenological problem
of the resulting effective theories, e.g., because they produce long range fifth
forces, modifying the behavior of Newtonian gravity in an unacceptable way. A
simple and efficient way out is given by assuming that also the other higher-
rank tensor fields appearing in the ten- and eleven-dimensional theories acquire
a non-trivial expectation value (the fluxes). As we will sketch in the following,
this results in the deformation of the four-dimensional supergravity models by the
gauging process and hence leads to gauged supergravity models, which then do
have a scalar potential that can make (at least some of) the moduli sufficiently
massive. The remarkable aspect of this type of reduction is that the fluxes can be
treated perturbatively and produce closed computable expressions for the lower
dimensional couplings and potentials.
We will give here an overview of the main features of flux compactifications
leading to gauged supergravities. For this reason we will focus on the simple
case of 11-dimensional supergravity reduced on the seven-dimensional torus, T7 ,
and discuss only its bosonic sector. The standard reduction, where one truncates
the spectrum to the massless modes, leads to N = 8 ungauged supergravity in
four dimensions. In detail, the reduction of the metric and three-form tensor field
produces the following massless spectrum: As we already discussed previously,

gμν gμI gI J

1 graviton 28

Cμνρ CμI J CμνI , CI J K

0 dof 7+21 vectors 7 35

= 63 scalars + 7 tensors

massless tensors can be dualized to massless scalar fields, and hence, after such
duality, one gets a total of 70 scalar fields, which is the standard scalar content of
the N = 8 supergravity multiplet.
An extremely important aspect of 4D supergravity is the U-duality group emerg-
ing as a generalization of the standard electric–magnetic duality (see Chap. 9). For
N = 8 supergravity, this is E7(7) ⊂ Sp(56,R) and implies that there is an underlying
symplectic action on 56 vector fields, 28 of which may appear simultaneously
in the action. Using the appropriate language, we can then consider the vectors
coming from the T7 compactification as the 28 electric ones: AΛ μ = {gμ , CμI J }.
I

However, one should also be able to identify the higher dimensional origin of the
28 dual vector fields, AμΛ . Twenty-one of them can be readily identified by the
reduction of the dual form of G4 = dC3 . This dual form is schematically obtained
as G7 ∼ G4 , so that its Bianchi identity reproduces the equation of motion for G4 :
dG7 = G4 ∧ G4 . The seven-form field strength then is the curvature of a six-form
274 10 Supergravity in Arbitrary Dimensions

potential that, upon reduction, produces

CμI J KLM = I J KLMNP ANP

μ . (10.33)

We also note that the seven scalars dual to the tensor fields above are also obtained
from the same potential as CI J KLMN . The remaining seven dual vector fields should
have an origin as dual metric fields, g̃μI , but it is not obvious how to achieve this yet,
though we expect them to correspond to non-geometric deformations. Anyway, at
the level of the 4D action, as we already discussed in Chap. 8, both descriptions are
equally valid, and as long as we don’t have gaugings we can (almost) freely dualize
the curvatures and the vectors.
For what concerns the scalar fields, we get a σ -model, but no scalar potential, as
expected for a standard Einstein–Maxwell extended supergravity.
The duality group is broken, and the 4D model gets deformed to a gauged
supergravity by introduction of “fluxes” for the four-form, i.e., an expectation value
for this field on the internal space:

GI J KL ≡ GI J KL  = 0. (10.34)

We now discuss how this is achieved, briefly noting that a similar situation occurs
when introducing a flux for the dual seven-form field, or a combination of the two.
A simple sketch of the reduction of the kinetic term of the four-form explains
how both the gauging and the scalar potential are generated. Without fluxes, the
reduction of the four-form kinetic term produces the expected kinetic terms for the
scalars and vectors:

gI K gJ L ∂ C ∂ [μ C ν] KL + g I A g J B g KC ∂μ CI J K ∂ μ CABC + . . .
   [μ ν]I J    (10.35)
gauge kin. function scalar σ -model

In the presence of a non-trivial flux, a new term emerges at the quadratic level,
containing only the scalar fields coming from the metric. This is an obvious scalar
potential term, and it is of the second order in the coupling constants dictated by the
fluxes G :

V (gI J ) = g I A g J B g KC g LD GI J KLGABCD . (10.36)

Another important coupling that emerges is linear in the flux,

∂[μ Cν]I J g μA g νB g I C g J D GABCD , (10.37)

which reconstructs the O(G ) couplings giving origin to non-Abelian covariant field
strengths

∂[μ Cν]I J + gμA gνB GABI J . (10.38)

10.5 Example: D = 5 275

From this same expression, it is also clear that GI J KL play the role of structure
constants. The gauging of the theory becomes even more evident if we look at
the kinetic terms of the scalar fields. Also these kinetic terms get modified to new
expressions involving O(G ) couplings reconstructing covariant derivatives:

∂μ CI J K + GI J KL gμL = !
∂μ CI J K . (10.39)

We therefore see that fluxes not only define the gauge couplings and the scalar
potential, but they also tell us the form of the embedding tensor, because they tell
us which vectors participate in the gauging and specify the couplings between the
scalar and vector fields defining the covariant derivatives.
Since the fluxes now play the role of gauge structure constants, one could expect
that they obey standard Jacobi identities. Although it may not be evident from the
example above, it can be shown that there is a one- to-one correspondence between
the consistency conditions on the gauge structure constants in four dimensions
and the Bianchi identities of the form fluxes in ten or 11 dimensions. There is,
however, an important subtlety that we want to emphasize here. When dealing with
flux compactifications (and especially in the case of non-trivial geometric fluxes,
i.e., globally defined torsion terms), the tensor fields appearing in the geometric
reduction transform under gauge transformations of the vector fields and vice versa.
This implies that generically the standard Jacobi identities coming from the gauge
algebra of the vector fields do not close. In fact, the gauge algebra is now really a
free differential algebra involving also the tensor fields, and therefore it is only the
closure of this structure that imposes all the necessary consistency conditions on the
structure constants (also to reproduce the higher dimensional Bianchi identities).
We conclude with some comments on the action of the duality group in relation
to the situation described in the previous paragraph. Although the natural setup
coming from a straightforward reduction of a flux background may result in a
free differential algebra that includes tensor fields, one may always use the duality
group to rotate the vector field basis so that no tensor fields are present in the final
Lagrangian. Obviously, in order to be consistent, the embedding of the gauge group
in the duality group will also be rotated accordingly, and the interpretation of the
resulting algebra in terms of the original theory may not be straightforward anymore.
Actually, this is a rough explanation of how non-geometric fluxes arise and why we
should expect them if we believe that the U-duality group of the effective theory
survives as a symmetry of the fully fledged higher dimensional fundamental theory
(maybe with some restrictions).

10.5 Example: D = 5

As a further illustration of the new structures emerging in higher dimensional

supergravity, we consider the case of five dimensions, where supergravity theories
have numerous applications, e.g., in the context of the AdS/CFT correspondence, the
study of black holes and domain walls, or phenomenological scenarios such as the
276 10 Supergravity in Arbitrary Dimensions

Randall–Sundrum or Hořava–Witten-type models. As we will see, the form of the

R-symmetry group provides useful constraints on the scalar manifold geometries,
which, for N ≥ 4, even fixes the target spaces completely once type and number
of multiplets in the theory are specified.
In five dimensions, one cannot impose Majorana or Weyl conditions,4 and the
possible numbers of real supercharges are 8, 16, 24, 32. One might call this N =
1, 2, 3, 4 supersymmetry, but in analogy with the counting in 4D, one usually refers
to these possible 5D supersymmetries as N = 2, 4, 6, 8, respectively, as we will
also do in these lecture notes. As explained in Appendix 10.A, in 5D, the parameter
 defined in (10.79) is  = −1, so that one may impose a symplectic Majorana
condition for the supersymmetry generators (i = 1, . . . , N ),

(Qi )∗ = Ωij BQj . (10.40)

This has the advantage of making the action of the R-symmetry group manifest.
More concretely, the linear rotations of the Qi that preserve the above symplectic
Majorana condition as well as the 5D supersymmetry algebra,

{Qi , QTj } = Ωij (CΓ a )Pa , (10.41)

form the R-symmetry groups U Sp(N ) ≡ U (N ) ∩ Sp(N , R).

10.5.1 N = 2 in 5D

For N = 2 ungauged supergravity in 5D, there are three important supermulti-

plets:

• Supergravity multiplet: (eμa , ψμi , Aμ ) (i = 1, 2)

Apart from the fünfbein, eμa , it contains two gravitini, ψμi , and a vector field, Aμ ,
the “graviphoton.”
• Vector multiplet: (Aμ , λi , ϕ)
The superpartners of the vector field, Aμ , are two gaugini, λi , and one real scalar
field, ϕ.
• Hypermultiplet: (ζ 1,2, q 1,2,3,4)
Here, ζ 1,2 are two spin-1/2 fermions (“hyperini”), and the q 1,2,3,4 denote four
real scalar fields.

4 A simple way to show that there are no Majorana spinors in 5D is to use the Majorana repre-
sentation for the first four gamma matrices, which are then manifestly real in this representation
and can thus naturally act on real spinors. The fifth gamma matrix, however, is then ±γ5 , which is
manifestly imaginary due to the i in its definition. There is thus no set of five gamma matrices with
the same reality properties.
10.5 Example: D = 5 277

In the above, the index i of the gravitini and the gaugini is a doublet index of the R-
symmetry group U Sp(2)R ∼ = SU (2)R . The hyperini, by contrast, are inert under
SU (2)R , which is the reason that we do not use i to label these two fermions.
The scalar ϕ is likewise SU (2)R -inert, whereas the hyperscalars q 1,2,3,4 form two
doublets under SU (2)R . All spinors are symplectic Majorana spinors.
In order to write down the general Lagrangian for N = 2 supergravity coupled
to nV vector multiplets and nH hypermultiplets, it is useful to group these fields as
follows:

(eμa , ψμi , Aμ ) ⊕ nV × (Aμ , λi , ϕ) ⊕ nH × (ζ 1,2, q 1,2,3,4)

= (eμa , ψμi , AIμ , λi , ζ A , ϕ x , q u ) (10.42)

where

I = 0, 1, . . . , nV (10.43)
x = 1, . . . , nV (10.44)
u = 1, . . . , 4nH (10.45)
A = 1, . . . , 2nH . (10.46)

Here we have combined the graviphoton and the nV vector fields of the vector
multiplets into an (nV + 1)-plet of vectors, AIμ . The indices x and u are curved
indices on the scalar manifolds, MV and MH , of the vector scalars and the
hyperscalars, respectively. The gaugini, λix , transform as tangent vectors of MV ,
and one may use curved indices, x, to label them, just as we did, or, as one often
also finds in the literature, one could use a flat tangent space index, a = 1, . . . , nV ,
instead of the curved tangent space index x. Both notations can be easily converted
into one another by contraction with vielbein on MV . As the holonomy group
of MV has no particular restrictions imposed by the supersymmetry algebra (see
below), the use of flat vs. curved indices x and a has no clear advantage. This is
different for the hyperini, as we will explain below.
In terms of the above fields, the bosonic Lagrangian of N = 2 matter-coupled
ungauged supergravity in 5D can be written as follows [7, 8]:

1 1 1
e−1 Lbos = R − ãI J (ϕ)FμνI
F μνJ − gxy (ϕ)∂μ ϕ x ∂ μ ϕ y
2 4 2
1 1
− huv (q)∂μ q u ∂ μ q v + √ CI J K  μνρσ λ Fμν
I
FρσJ
AK
λ. (10.47)
2 6 6

In this expression, CI J K is a constant, completely symmetric tensor. As was shown

by Sierra [8], the two sigma models associated with the hyperscalars and the vector
scalars do not mix, i.e., the scalar manifold metric is block diagonal in these two
278 10 Supergravity in Arbitrary Dimensions

sectors and the total scalar manifold decomposes into a direct product, MV × MH ,
just as in four dimensions.

10.5.1.1 The Geometry of MV

The scalars, ϕ x , of the vector multiplets are inert under the R-symmetry group
SU (2)R . The holonomy group of MV therefore does not receive any constraints
from the R-symmetry group, as we already mentioned above. However, as they are
connected by supersymmetry to the vector fields (or rather to nV of them), the scalar
manifold MV inherits part of the vector field structure. In fact, one finds that MV
is completely determined by the constants CI J K that define the Chern–Simons term
in the action. More precisely, the CI J K define a cubic polynomial [7],

N(XI ) ≡ CI J K XI XJ XK (10.48)

on an auxiliary space, RnV +1 , spanned by real coordinates XI (I = 0, 1, . . . , nV ).

On this auxiliary space, N then defines a (not necessarily positive definite) metric,

1 ∂ ∂
aI J (X) ≡ − log N(X). (10.49)
3 ∂X ∂XJ
I

The scalar manifold MV is then given as a cubic hypersurface in the auxiliary

RnV +1 :

MV = {XI ∈ RnV +1 |N(X) = 1}. (10.50)

MV can be parameterized by nV real coordinates, which are identified with the

physical scalar fields, ϕ x . The metric, gxy , on MV is given by the pull-back of the
auxiliary metric aI J , and the “gauge kinetic function” ãI J is the restriction of aI J
to the hypersurface MV :

3 ∂XI ∂XJ 

gxy (ϕ) = x y
aI J  (10.51)
2 ∂ϕ ∂ϕ N(X)=1


ãI J (ϕ) = aI J  . (10.52)
N(X)=1

The true scalar manifold is then actually the subspace of (10.50) for which gxy (ϕ)
and ãI J (ϕ) are positive definite. The above-described geometry of the scalar
manifold of the vector multiplet scalars is called very special (real) geometry. Upon
dimensional reduction to four dimensions, MV becomes a special Kähler manifold
of restricted type, namely, one for which the holomorphic prepotential is purely
cubic. Special Kähler manifolds that arise in this way from 5D are called very
special Kähler geometries, and the corresponding map is called the R-map [7].
10.5 Example: D = 5 279

10.5.1.2 The Geometry of MH

As the hyperscalars transform non-trivially under the R-symmetry group SU(2)R ,
we expect the holonomy group of MH to respect this structure and hence to contain
SU(2) as a factor. The largest group G such that SU(2) × G is still a subgroup
of (the maximal holonomy group) O(4nH ) is Usp(2nH ). The holonomy group of
MH should thus be contained in SU(2)× USp(2nH ), with the SU(2) part being
non-trivial. Manifolds of this type are called quaternionic Kähler, and we discussed
them in detail already in Chap. 8, because the hypermultiplet geometry in 5D is
really identical to the scalar field geometry of 4D hypermultiplets. The restricted
holonomy group also means that the tangent space group can be restricted to SU(2)
× USp(2nH ). Just as in four dimensions, this allows a natural split of the flat tangent
space index of MH into an SU(2) index i = 1, 2 and an USp(2nH ) index A =
1, . . . , 2nH . These indices are to be identified with the R-symmetry group index i
and the index A of the hyperini ζ A .

10.5.2 N = 4 in 5D

For 5D, N = 4 supersymmetry, the R-symmetry group is USp(4)R . USp(4)R is a

double cover of SO(5), which, by abuse of notation, we will also call SO(5)R . The
two relevant multiplets in ungauged supergravity are [9]:

• Supergravity multiplet: This multiplet consists of the graviton, four gravitini,

six vector fields, four spin-1/2 fields, and one real scalar field. This scalar is
necessarily SO(5)R -inert and parameterizes the real line:

MSG ∼
=R∼
= SO(1, 1) (10.53)

• Vector multiplet: This multiplet contains one vector, four spin-1/2 fields, and
five real scalars in the 5 of SO(5)R . The holonomy group of the scalar manifold
of nV such vector multiplets should thus contain SO(5) as a factor. The
largest remaining group factor that still allows the embedding into the maximal
holonomy group O(5nV ) is SO(nV ), i.e., Hol(MV ) ⊂ SO(5) × SO(nV ) with the
SO(5) part being non-trivial. According to Berger’s classification [10], the only
Riemannian manifold of dimension 5nV with this property is

SO(5, nV )
MV = . (10.54)
SO(5) × SO(nV )

10.5.3 N = 6 in 5D

In this case the R-symmetry group is USp(6)R . The only multiplet relevant for
supergravity is the supergravity multiplet. The scalars in this multiplet transform
non-trivially under the R-symmetry group, and the holonomy group of the scalar
280 10 Supergravity in Arbitrary Dimensions

manifold should contain USp(6) as a factor, which, together with the dimension
fixes it to be

SU∗ (6)
MSG ∼
= , (10.55)
USp(6)

where SU∗ (6) is a particular real form of SU(6).

10.5.4 N = 8 in 5D

In the maximally supersymmetric case, the R-symmetry group is USp(8)R , and the
supergravity multiplet contains the graviton, 8 gravitini, 27 vector fields, 48 spin-
1/2 fields, and 42 real scalar fields. The latter transform non-trivially under the R-
symmetry group, and we expect the holonomy group to contain a USp(8)-factor.
The only 42-dimensional space with this property is

E6(6)
MSG ∼
= , (10.56)
USp(8)

where E6(6) denotes the real form of E6 for which the difference of compact and
non-compact generators is 6.

10.5.5 Gaugings and Tensor Fields

As we discussed in Sect. 8.1, in four spacetime dimensions, a massless vector field

without gauge interactions can be equivalently described by its magnetic dual vector
field. In the presence of gauge interactions, on the other hand, this electric–magnetic
duality is broken, and one has to make sure that the gauging is performed in a
suitable duality frame.5
In five dimensions, there is no duality between electric and magnetic vector
fields, but if there are no gauge interactions, there is an analogous Poincaré duality
between vector fields, A = Aμ dx μ , and two-form fields, B = 12 Bμν dx μ ∧ dx ν .
In the simplest version, this is just the statement that the 5D source-free Maxwell
equations,

d  F = 0, dF = 0 (10.57)

5 As explained in Sect. 9.5, the embedding tensor formalism allows one to formally maintain
electric–magnetic duality at the expense of a redundant field content.
10.5 Example: D = 5 281

for the two-form field strength F = dA read

dH = 0, d  H = 0, (10.58)

when expressed in terms the dual three-form field strength, H := F , which imply
H = dB and d  dB = 0, i.e., the field equation of a massless two-form field, B.
If one now instead considers 5D theories with gauge interactions, the above
duality between 5D vector and tensor fields no longer holds, and one would naively
expect that a consistent gauging would require working exclusively with vector
fields. In many cases, this is also what happens, but there are also important
situations where some of the vector fields have to be converted to tensor fields in
a specific way in order to perform the gauging.
To understand this, let us assume we start from an ungauged 5D supergravity
theory in the standard form, as described in the above subsections, where all
potential tensor fields are dualized to vector fields. Suppose further the theory has
n vector fields and a global symmetry group, Gglobal,6 such that the n vector fields
transform in an n-dimensional representation of that global symmetry group (this
representation may be reducible or irreducible, depending on the theory).
If Gglobal has a subgroup, G, such that this n-dimensional representation of
Gglobal becomes the adjoint representation of G,

n(Gglobal) → adj(G), (10.59)

one can replace the Abelian field strengths, F I (I = 1, . . . , n), by the corresponding
non-Abelian field strengths, F I , and the partial derivatives, ∂μ , of charged matter
fields by gauge covariant derivatives, ! ∂μ , and gauge the group G. Just as in 4D,
this covariantization will break supersymmetry, which, however, can be restored by
introducing suitable Yukawa interactions and scalar potentials into the Lagrangian
as well as fermionic shifts to the supersymmetry transformation laws.
The above also holds true in the more general situation when the n-dimensional
representation of Gglobal decomposes into the adjoint of G plus singlets of G,

n(Gglobal) → adj(G) ⊕ singlets(G). (10.60)

If G has no Abelian factor, the singlet fields will just remain Abelian vector
fields, and they will have no gauge couplings to the matter fields, i.e., they will
be “spectator vector fields” with respect to the gauging. In case G has an Abelian
factor, on the other hand, the singlet vector fields will still remain Abelian, but they

6 In five dimensions, Gglobal is typically (a subgroup of) the isometry group, Iso(Mscalar ), of the
scalar manifold. If all scalars are inert under the respective R-symmetry group (as, e.g., for 5D,
N = 2 supergravity coupled to vector multiplets only, or for 5D, N = 4 pure supergravity), the
global symmetry group also has the R-symmetry group, GR = Usp(N ), as an additional factor.
282 10 Supergravity in Arbitrary Dimensions

might have minimal couplings to the matter fields, i.e., they might contribute to the
Abelian part of the gauge group G.
As a simple example, consider a theory with four vector fields in the fundamental
representation of a global symmetry group SO(4). With respect to the obvious
subgroup SO(3), the 4 of SO(4) decomposes into 3 ⊕ 1. As the 3 is the adjoint
of SO(3), one can use the three vector fields in the 3 to gauge SO(3), with the SO(3)
singlet vector field remaining a spectator vector field. Gaugings of this standard
type were investigated in 5D, N = 2 supergravity in [11] and in 5D, N = 4
supergravity in [9].
A more problematic situation arises, however, if the decomposition of the n-
dimensional representation of Gglobal with respect to the subgroup G also contains
non-singlets of G,

n(Gglobal) → adj(G) ⊕ singlets(G) ⊕ non-singlets(G). (10.61)

In this case, the gauging of G cannot be performed in the usual way, because vector
fields can only couple consistently to other vector fields if they sit in the adjoint of
the gauge group.
Historically, the first example of this situation occurred in 5D, N = 8
supergravity in the 1980s [12–14]. In the ungauged version, this theory has 27 vector
fields transforming in the 27-dimensional irreducible representation of the global
symmetry group, Gglobal = E6(6). E6(6) is the maximally non-compact real form
of the exceptional group E6 and forms the isometry group of the scalar manifold
Mscalar = E6(6)/USp(8) of this theory.
A particularly interesting subgroup of the global symmetry group E6(6) is the
subgroup SO(6), under which the 27 of E6(6) transforms as

27 → 15 ⊕ 6 ⊕ 6. (10.62)

Here, the 15 is the adjoint representation of SO(6), whereas the 6 denotes the
fundamental representation of SO(6), which is clearly a non-singlet representation.
Without the 6 ⊕ 6, one could gauge SO(6) with the 15 vector fields in the adjoint,
but due to the presence of the non-singlets, this is not possible in the standard way.
This by itself would not be a big deal, as not every subgroup of a global symmetry
group needs to be gaugeable, but in this particular case, there were very strong
arguments in favor of the existence of a gauging with the gauge group SO(6). These
arguments have to do with the compactification of type IIB supergravity on the
maximally supersymmetric background solution AdS5 × S 5 , which was expected to
admit a consistent truncation to the lowest lying Kaluza–Klein modes that should be
identical to 5D, N = 8 supergravity with gauge group SO(6) (the isometry group
of the five-sphere).
The resolution of this problem came from a closer inspection of the Kaluza–
Klein spectrum of this compactification [15, 16], which, apart from the 15 vector
fields, also revealed the presence of 12 tensor fields, which are not equivalent to
10.5 Example: D = 5 283

vector fields in an AdS5 background (they transform in a different representation of

the AdS5 isometry group as the vector fields).
This suggests that a consistent gauging of SO(6) might require treating the fields
in the 6 ⊕ 6 as antisymmetric tensor fields and not as vector fields. In fact, as tensor
fields, they could be treated as a special type of matter fields, so that their derivatives
could be covariantized by introducing minimal couplings to the 15 gauge fields
using the six-dimensional representation matrices of SO(6) (see Eq. (10.64)). This
approach turned out to be correct and led to the successful construction [12–14] of
5D, N = 8 gauged supergravity with gauge group SO(6), which, many years later,
also played a central role in the AdS/CFT correspondence.
The necessity of converting non-singlet vector fields to tensor fields in order to
perform certain gaugings also was found in 5D, N = 4 [17–19] and N = 2
theories [20–22]. In fact, the N = 2 cases allow one to isolate the contribution
to the scalar potential that arises due to the presence of charged tensor fields
from those contributions that come from the gauging of the R-symmetry group
or the presence of non-Abelian gauge fields [20], or from charged hypermultiplets
[21]. Interestingly, one finds that this scalar potential contribution is positive semi-
definite, i.e., it cannot by itself lead to AdS vacua [20]. In fact, the N = 2
theories with charged tensor fields were the first extended supergravity theories in
which a meta-stable de Sitter vacuum could be constructed [23, 24]. This implies
that the presence of tensor fields in 5D supergravity theories with non-singlet
representations outside the adjoint representation of the gauge group is not a
consequence of an AdS vacuum structure (although it is consistent with it), but
follows from more general considerations.
Another noteworthy feature of the 5D gaugings with charged tensor fields is that
the corresponding Lagrangians contain a first-order kinetic term for the tensor fields
of the schematic form
1 M! N
ΩMN εμνρσ κ Bμν ∂ρ Bσ κ , (10.63)
g

where g denotes the gauge coupling, ΩMN = −ΩNM is an antisymmetric constant

tensor, and M, N, . . . = 1, . . . , m label the m tensor fields. The gauge covariant
derivative, ! ∂μ , describes the minimal coupling of the n gauge fields, AIμ (I, J, . . . =
1, . . . , n), in the adjoint representation of the gauge group to the tensor fields and
takes the form

!
∂[ρ BσNκ] = ∂[ρ BσNκ] + gAI[ρ ΛN M
(10.64)
I M Bσ κ] ,

where ΛN I M denotes the representation matrices of the tensor fields with respect to
the gauge group.
Using this first-order form and taking into account also another, mass-like term
for the tensor fields not shown here, it is in principle possible to integrate out half
of the tensor fields so as to arrive at a Lagrangian with second-order kinetic term
for the remaining (massive) tensor fields (see [25] for a detailed discussion). Doing
284 10 Supergravity in Arbitrary Dimensions

this explicitly, however, yields fairly complicated expressions in general so that it is

usually easier to stick with the above first-order form. For more detailed discussions
of 5D gauged supergravity with tensor fields, we refer to the original literature.

10.A Appendix: Clifford Algebras and Spinors in Arbitrary D

As described in Sect. 10.1, the balance between bosonic and fermionic degrees
of freedom in supersymmetric field theories in general is implemented differently
for different spacetime dimension, D, because the degrees of freedom of the
corresponding fields have a different overall scaling with D. This is further
complicated by the strong D-dependence of the possible chirality and reality
conditions one can impose on Clifford algebra representations so as to generate
minimal spinor representations of the Lorentz group. The D-dependence of these
chirality and reality conditions also leads to D-dependent R-symmetry groups,
which in turn contribute to a rich variety of possible scalar manifold geometries in
the respective spacetime dimensions. It is the purpose of this subsection to classify
the representations of Clifford algebras, the minimal spinor representations of the
corresponding Lorentz groups as well as the resulting R-symmetry groups. This
generalizes the discussion of spinors in four dimensions given in Chap. 1.
Starting point is the Clifford algebra, Cliff(1, D − 1), in D Lorentzian dimen-
sions,

{Γa , Γb } = 2ηab (a, b, . . . = 0, 1, . . . , D − 1) (10.65)

ηab = diag(−1, +1, . . . , +1). (10.66)

Just as in 4D, the relation (10.65) implies that

1 1
!(Mab ) ≡ Σab ≡
ρ [Γa , Γb ] = Γab (10.67)
4 2
 ab 
form a representation of the Lorentz algebra. The exponentials exp ω 2Σab with
ωab being finite rotation angles or boost parameters then form a double-valued
representation of the Lorentz group SO0 (1, D − 1).

10.A.1 Irreducible Representations of Cliff(1, D − 1)

The structure of the irreducible representations of Cliff(1, D−1) is slightly different

for even and odd dimensions:

10.A.1.1 Even Dimensions

Up to equivalence, there is exactly one non-trivial irreducible representation (irrep)
of Cliff(1, D − 1) (see, e.g., [26]). It has complex dimension 2D/2 , i.e., the Γa are
10.A Appendix: Clifford Algebras and Spinors in Arbitrary D 285

complex (2D/2 × 2D/2 )-matrices, generalizing the (4 × 4)-matrices in 4D. Explicit

forms of these representations can be built up by successive tensor products of the
irreps of lower-dimensional Clifford algebras, starting with the case D = 2 (see,
e.g., [27]), but we do not need them for this book.

10.A.1.2 Odd Dimensions

If D is odd, irreps of Cliff(1, D −1) can be obtained from an irrep of Cliff(1, D −2)
by defining the analogue of the γ5 matrix in 4D:
D+1
Γ∗ ≡ (−i) 2 Γ0 Γ1 . . . ΓD−2 . (10.68)

This matrix satisfies

(Γ∗ )2 = 1 (10.69)
{Γ∗ , Γa } = 0 ∀a = 0, . . . , D − 2 (10.70)

so that either of
(±)
ΓD−1 ≡ ±Γ∗ (10.71)

can be used as the remaining gamma matrix to promote {Γ0 , . . . , ΓD−2 } to a

representation of Cliff(1, D − 1). One thus obtains two inequivalent representations
of Cliff(1, D − 1) for odd D, one for each sign in (10.71).

10.A.2 Irreducible Spinor Representations of SO0 (1, D − 1)

Thus far, we have discussed the irreps of Cliff(1, D − 1) and described how these
induce double-valued spinor representations of the corresponding Lorentz groups
SO0 (1, D−1). Just as in four dimensions, however, the spinor representations of the
Lorentz group so-obtained are in general not irreducible, even though they descend
from irreducible representations of Cliff(1, D − 1). In order to obtain an irreducible
spinor representation of SO0 (1, D − 1), one in general has to impose additional
constraints, which may be of the following type:

1. Chirality condition
2. Reality condition
3. Chirality and a reality condition

The possibilities to impose one of the above are strongly dimension dependent,
as we will now describe.
286 10 Supergravity in Arbitrary Dimensions

10.A.2.1 Chirality Conditions

For even D, we can always impose the following chirality condition to define a left-
or right-handed Weyl spinor:

Γ∗ ψ L = ±ψ L (10.72)
R R

Note that that this condition is Lorentz covariant because of [Σab , Γ∗ ] = 0.

For odd D, on the other hand, there is no non-trivial analogue of Γ∗ , because

Γ Γ ...Γ Γ (±) ∼ (ΓD−1

(±) 2
) ∼ 1. (10.73)
 0 1  D−2 D−1
(±)
∼ΓD−1

Thus a non-trivial chirality condition can only be imposed in even D.

10.A.2.2 Reality Conditions

It is again useful to distinguish between even and odd dimensions:

Even Dimensions
As discussed above, for even D there is only one equivalence class of irreps of
Cliff(1, D − 1) generated by matrices Γa . Hence, the complex conjugate matrices
±Γa∗ , which also satisfy the Clifford algebra, must be equivalent to the matrices Γa ,
i.e., there has to be a matrix, B, such that

Γa∗ = ηBΓa B −1 (10.74)

for both signs η = ±1.

Odd Dimensions
If D is odd, we can obviously find a matrix, B, that also satisfies (10.74) for the first
(D − 1) gamma matrices with η = ±1. What is non-trivial, however, is to extend
(±)
(10.74) also to the remaining gamma matrix ΓD−1 = ±Γ∗ (cf. Eq. (10.68)), i.e., to
have

(Γ∗ )∗ = ηBΓ∗ B −1 . (10.75)

Indeed, using the definition (10.68) and (10.74) for Γ0 , . . . , ΓD−2 , one easily shows

(Γ∗ )∗ = (−1) BΓ∗ B −1 ,

D+1
2 (10.76)

which is consistent with (10.75) only for one sign:


D+1 −1 for D=5 mod 4
η = (−1) 2 = (10.77)
+1 for D=3 mod 4
10.A Appendix: Clifford Algebras and Spinors in Arbitrary D 287

Obviously, the defining Eqs. (10.74) and (10.75) define B only up to an arbitrary
rescaling. We may thus choose the overall scaling such that

| det B| = 1 (choice) (10.78)

With this normalization, one has (cf. Exercise 10.1)

B ∗ B = 1 (10.79)
 = ±1. (10.80)

The important point now is that this parameter  is not arbitrary, but is instead
fixed by the values of η and D. Concretely, for D = 2n or D = 2n + 1, one finds7
√ π 
 = −η 2 cos (1 + η2n) , (10.83)
4
and one arrives at the possible values for  and η shown in Table 10.2.

The Majorana Condition

What makes the possible values of  so important is that it determines whether one
can impose a Majorana condition on a spinor, which, in terms of B, reads

ψ ∗ = αBψ (Majorana condition), (10.84)

where α is an arbitrary phase. This condition is consistent with Lorentz invariance,

because Γab∗ = BΓ B −1 . A Majorana spinor thus furnishes a complete represen-
ab
tation of the Lorentz algebra and has only half as many degrees of freedom as an
unconstrained complex Dirac spinor. The consistency of (10.84) with ψ ∗∗ = ψ,
however, imposes the consistency condition

 = +1 (for Majorana condition), (10.85)

7 This can be proven, e.g., with the help of the charge conjugation matrix C. In a friendly
representation (i.e., for Γa Γa† = 1 (no sum) and symmetric or anti-symmetric Γa ), C ≡ B T Γ0
satisfies, because of (10.74) and (10.79),

ΓaT = −ηCΓa C −1 (10.81)

C T
= −ηC. (10.82)
The matrices (CΓa1 ...ap ) then have a definite symmetry under transposition. This symmetry
depends on p, , and η. On the other hand, the set of all matrices Γa1 ...ap plus the unit matrix
form a complete basis of all complex (2[D/2] × 2[D/2] )-matrices. As the number of linearly
independent antisymmetric and symmetric of such matrices is fixed to be 2[D/2] (2[D/2] − 1)/2
and 2[D/2] (2[D/2] + 1)/2, respectively, one can determine the possible values of  as a function of
D and η (which, for odd dimensions, is itself fixed by D). (cf., e.g., [26]).
288 10 Supergravity in Arbitrary Dimensions

Table 10.2 The possible values for η and  together with the resulting minimal
spinor types, the minimal number of real supercharges, and the general form of the R-
symmetry groups (M = Majorana, SM = Symplectic Majorana, W = Weyl, MW = Majorana–
Weyl, SMW = Symplectic Majorana–Weyl)
D η  Min. spinor type Min. # of real supercharges R-symmetry group
2 +1 +1 MW 1 SO(NL ) × SO(NR )
−1 +1
3 +1 +1 M 2 SO(N)
4 +1 +1 M or W 4 U (N)
−1 −1
5 −1 −1 SM 8 U sp(2N)
6 +1 −1 SMW 8 U sp(2NL ) × U sp(2NR )
−1 −1
7 +1 −1 SM 16 U sp(2N)
8 +1 −1 16 U (N)
−1 +1 M or W
9 −1 +1 M 16 SO(N)
10 +1 +1 MW 16 SO(NL ) × SO(NR )
−1 +1
11 +1 +1 M 32 SO(N)
12 +1 +1 M or W 64 U (N)
−1 −1
... ... ... ... ... ...

limiting the possibility of Majorana spinors to certain dimensions, as indicated in

Table 10.2.

Symplectic Majorana Spinors

If  = −1, one can impose a symplectic Majorana condition. To this end, one needs
an even number of Dirac spinors ψi , (i, j, . . . = 1, . . . , 2N) and an antisymmetric
real matrix Ωij with Ω 2 = −12N and imposes

(ψi )∗ = Ωij Bψj . (10.86)

As one needs at least two Dirac spinors to impose the symplectic Majorana
condition, it does not lead to a reduction of the minimal number of degrees of
freedom relative to a single Dirac spinor. The symplectic Majorana condition is,
however, convenient, because it makes the action of the R-symmetry group (which
in these dimensions involve symplectic groups; see Table 10.2) manifest.
10.A Appendix: Clifford Algebras and Spinors in Arbitrary D 289

10.A.3 Majorana and Weyl Condition

In some dimensions, the Majorana and the Weyl condition can be imposed
simultaneously. This reduces the number of independent degrees of freedom to one
quarter relative to an unconstrained Dirac spinor. Imposing (we set the phase α = 1
for simplicity)

ψ ∗ = Bψ (10.87)
Γ∗ ψ = ±ψ, (10.88)

at the same time, obviously requires the consistency condition

(Γ∗ )∗ = BΓ∗ B −1 (10.89)

which is possible only if D = 4n − 2. But as there are no Majorana spinors in

D = 6, 14, . . ., Majorana–Weyl spinors can only exist for

D = 2 mod 8 (Condition for Majorana–Weyl spinors). (10.90)

Note, in particular, that in 4D, one can have Majorana spinors or Weyl spinors, but
not Majorana–Weyl spinors.
Analogously, in dimensions in which  = −1 allows a symplectic Majorana
condition, one can sometimes also simultaneously impose a Weyl condition, and the
corresponding spinors are then called symplectic Majorana–Weyl spinors. These are
the dimensions D = 6 mod 8
The minimal amount of supersymmetry in each spacetime dimension is gen-
erated by a spinor operator that corresponds to the minimal spinor representation
of the Lorentz group in the respective spacetime dimension. Extended supersym-
metries then correspond to multiples of such minimal spinors. The R-symmetry
group of the corresponding supersymmetry algebra has to respect these reality
and chirality conditions and thus depends on the minimal spinor type as shown in
Table 10.2. If the scalar fields of a given type of multiplet transform non-trivially
under the R-symmetry group (or a factor thereof), the holonomy group of the
scalar manifold typically contains this group (factor) as a factor. Especially for
large amounts of supersymmetry, this already strongly constrains the possible scalar
manifolds, as we described in detail for the theories in 4D and 5D.
For more than 32 real supercharges, one always has states with helicity |h| > 2
in the supermultiplets, which, for Lorentzian signature, limits supersymmetric field
theories to D ≤ 11.
We finally note that in spacetimes with non-Lorentzian signature, the possible
reality and chirality conditions for a given D are in general different. This is in
particular true for the Euclidean signature of the compactification manifolds in
string compactifications, so that the possible spinor type on these manifolds cannot
be read off from Table 10.2.
290 10 Supergravity in Arbitrary Dimensions

Exercises

10.1. Using (10.74) and Schur’s Lemma, show that (10.79) holds for some  ∈
C. Using the complex conjugate of (10.79) and the choice (10.78), show that this
implies (10.80).

References
1. E. Cremmer, B. Julia, J. Scherk, Supergravity theory in 11 dimensions. Phys. Lett. B76, 409–
412 (1978)
2. F. Giani, M. Pernici, N = 2 supergravity in ten-dimensions. Phys. Rev. D 30, 325–333 (1984)
3. L.J. Romans, Massive N = 2a supergravity in ten-dimensions. Phys. Lett. B 169, 374 (1986)
4. J.H. Schwarz, Covariant field equations of chiral N = 2 D = 10 supergravity. Nucl. Phys. B 226,
269 (1983)
5. P.S. Howe, P.C. West, The complete N = 2, D = 10 supergravity. Nucl. Phys. B 238, 181–220
(1984)
6. G. Dall’Agata, K. Lechner, M. Tonin, D = 10, N = IIB supergravity: lorentz invariant actions
and duality. JHEP 9807, 017 (1998) hep-th/9806140
7. M. Günaydin, G. Sierra, P.K. Townsend, The geometry of N = 2 Maxwell-Einstein supergravity
and Jordan algebras. Nucl. Phys. B 242, 244–268 (1984)
8. G. Sierra, N = 2 Maxwell matter Einstein supergravities in D = 5, D = 4 and D = 3. Phys. Lett.
B 157, 379–382 (1985)
9. M. Awada, P.K. Townsend, N = 4 Maxwell-Einstein supergravity in five-dimensions and its
SU(2) gauging. Nucl. Phys. B 255, 617–632 (1985)
10. M. Berger, Sur les groupes d’holonomie homogènes des variétés a connexion affines et des
variétés riemanniennes. Bull. Soc. Math. France 83, 279–330 (1953)
11. M. Günaydin, G. Sierra, P.K. Townsend, Gauging the d = 5 Maxwell-Einstein supergravity
theories: more on Jordan algebras. Nucl. Phys. B 253, 573 (1985)
12. M.Günaydin, L.J. Romans, N.P. Warner, Gauged N = 8 supergravity in five-dimensions. Phys.
Lett. B 154, 268–274 (1985)
13. M. Günaydin, L.J. Romans, N.P. Warner, Compact and noncompact gauged supergravity
theories in five-dimensions. Nucl. Phys. B 272, 598–646 (1986)
14. M. Pernici, K. Pilch, P. van Nieuwenhuizen, Gauged N = 8 D = 5 supergravity. Nucl. Phys. B
259, 460 (1985)
15. M. Günaydin, N. Marcus, The spectrum of the s**5 compactification of the chiral N = 2, D = 10
supergravity and the unitary supermultiplets of U(2, 2/4). Class. Quant. Grav. 2, L11 (1985)
16. H.J. Kim, L.J. Romans, P. van Nieuwenhuizen, The mass spectrum of chiral N = 2 D = 10
supergravity on S**5. Phys. Rev. D 32, 389 (1985)
17. L.J. Romans, Gauged N = 4 supergravities in five-dimensions and their magnetovac
backgrounds. Nucl. Phys. B 267, 433–447 (1986)
18. G. Dall’Agata, C. Herrmann, M. Zagermann, General matter coupled N = 4 gauged supergrav-
ity in five-dimensions. Nucl. Phys. B 612, 123–150 (2001) [arXiv:hep-th/0103106 [hep-th]]
19. J. Schon, M. Weidner, Gauged N = 4 supergravities. JHEP 05, 034 (2006) [arXiv:hep-
th/0602024 [hep-th]]
20. M. Günaydin, M. Zagermann, The Gauging of five-dimensional, N = 2 Maxwell-Einstein
supergravity theories coupled to tensor multiplets. Nucl. Phys. B 572, 131–150 (2000)
[arXiv:hep-th/9912027 [hep-th]]
21. A. Ceresole, G. Dall’Agata, General matter coupled N = 2, D = 5 gauged supergravity. Nucl.
Phys. B 585, 143–170 (2000) [arXiv:hep-th/0004111 [hep-th]]
References 291

22. E. Bergshoeff, S. Cucu, T. de Wit, J. Gheerardyn, S. Vandoren, A. Van Proeyen, N = 2 super-

gravity in five-dimensions revisited. Class. Quant. Grav. 21, 3015–3042 (2004) [arXiv:hep-
th/0403045 [hep-th]]
23. M. Günaydin, M. Zagermann, The Vacua of 5-D, N = 2 gauged Yang-Mills/Einstein tensor
supergravity: Abelian case. Phys. Rev. D 62, 044028 (2000) [arXiv:hep-th/0002228 [hep-th]]
24. B. Cosemans, G. Smet, Stable de Sitter vacua in N = 2, D = 5 supergravity. Class. Quant. Grav.
22, 2359–2380 (2005) [arXiv:hep-th/0502202 [hep-th]]
25. P.K. Townsend, K. Pilch, P. van Nieuwenhuizen, Selfduality in odd dimensions. Phys. Lett. B
136, 38 (1984)
26. P.C. West, Supergravity, brane dynamics and string duality [arXiv:hep-th/9811101 [hep-th]]
27. A. Van Proeyen, Tools for supersymmetry. hep-th/9910030
Index

Symbols Equivariance condition, 92, 93

η-problem, 168
1.5 order formalism, 40, 50
F
Fayet–Iliopoulos terms, 91, 94, 103
A Fermionic shifts, 75
Anomaly mediation, 149 Fierz identities, 17
Anti-de Sitter space, 51 First order formalism, 40, 50
Flux compactifications, 273, 274
Free differential algebra, 247
B F-term, 75, 118, 131
Breitenlohner–Freedman bound, 64, 167 F-term SUSY breaking, 99, 134, 137, 138, 141

C G
Cartan–Killing metric, 231 Gauge kinetic function, 87, 103, 211
Casimir for AdS, 61 Gauge mediation, 149
Chiral multiplet, 21, 74 Gauging, 60, 229, 241, 245, 249, 253, 280
Clifford algebra, 11, 284 Gaugino, 85, 154
Composite Kähler connection, 106, 113, 123 Goldstino, 136–138, 141, 142
Contorsion tensor, 41 Gradient flow relations, 117
Conventional constraint, 49, 67 Gravitino, 25
Coset manifold, 231 Gravitino mass, 57
Cosmological constant, 51, 58, 133 Gravity mediation, 149, 150
GUT scale, 3

D
de Sitter space, 51 H
Differential forms, 36 Higgs field, 100
Dirac conjugate, 14 Hodge dual, 37
Dirac spinor, 11 Holomorphic Killing vector, 90
D-terms, 92, 98, 115, 118, 131 Homogeneous space, 231
D-term SUSY breaking, 99, 116, 134, 137 Hypermultiplet, 177, 184, 197

E I
Einstein–Hilbert action, 38 Inflation, 167
Electric–magnetic duality, 178, 228 Isometry, 87, 164
Embedding tensor, 223, 242, 243, 246, 249, Iwasawa decomposition, 232
255
© Springer-Verlag GmbH Germany, part of Springer Nature 2021 293
G. Dall’Agata, M. Zagermann, Supergravity, Lecture Notes in Physics 991,
https://doi.org/10.1007/978-3-662-63980-1
294 Index

K Riemann tensor, 33
Kähler covariant derivative, 106, 109
Kähler form, 83
Kähler manifolds, 81 S
Kähler potential, 83, 103, 112 Second order formalism, 40, 49
Kähler transformations, 83 Shift term, 56
Killing prepotentials, 91, 103 Singleton, 63, 65
Killing vectors, 88, 103, 164, 222, 224 Slow-roll, 168
Soft SUSY breaking terms, 135, 152
Solvable algebra, 232
L Special Kähler geometry, 185, 191, 196
Levi–Civita connection, 32 Spin connection, 35, 38, 49
Lie derivative, 68 Supercovariant connection, 49
Super-Higgs, 137, 141
Superpotential, 74, 103, 109
M Supersymmetric Ward identity, 60
Majorana spinor, 15, 264, 287 Supersymmetry algebra, 18
Mass in AdS, 64 Supertrace, 99, 148
Maurer–Cartan form, 235 SUSY breaking, 136, 150
Moduli, 155, 158 SUSY breaking scale, 134, 148
Symmetric coset, 232
Symplectic frame, 230, 238, 241, 244, 255, 259
N Symplectic section, 190, 193, 194, 205, 207
Neveu–Schwarz–Neveu–Schwarz (NSNS)
sector, 269
Noether’s method, 23 T
No-scale models, 133, 161 Tensor fields, 265, 280
Tensor multiplets, 119
Torsion constraint, 49, 67
P Torsion tensor, 37
Palatini formalism, 40 Type I supergravity, 269
Planck mass, 33 Type IIA supergravity, 268
Poisson brackets, 92 Type IIB supergravity, 271
Polonyi model, 150
Prepotential, 188, 192, 194, 209, 213
U
U-duality, 227, 228
Q Unitary representations in AdS, 63
Quaternionic-Kähler geometry, 185, 200

V
R Vector multiplet, 85, 113
Ramond–Ramond (RR) sector, 269
Rarita–Schwinger action, 26
Reductive coset, 232 W
Ricci scalar, 33 Wess–Zumino model, 21, 74, 104
Ricci tensor, 33 Weyl spinor, 12, 264, 286