Lessons for gravity from entanglementchep.iisc.ernet.in/Personnel/pages/asinha/arpan_thesis.pdf · Arpan Bhattacharyya Centre for High Energy Physics Indian Institute of Science Bangalore

Lessons for gravity from entanglement

A thesis submitted for the degree of

Doctor of Philosophy

in the Faculty of Sciences

Arpan Bhattacharyya

Centre for High Energy Physics

Indian Institute of Science

Bangalore - 560012. India.

June 2015

ii

Declaration

I hereby declare that the work presented in this thesis “Lessons for gravity from entangle-

ment” is based on the research work by me under the supervision of Prof. Aninda Sinha and

with my collaborators at the Centre for High Energy Physics, Indian Institute of Science,

Bangalore, India. It has not been submitted elsewhere as a requirement for any degree or

diploma of any other Institute or University. Proper acknowledgements and citations have

been made in appropriate places while borrowing research materials from other investiga-

tions.

Date : Arpan Bhattacharyya

Certified by :

Prof. Aninda. Sinha

Centre for High Energy Physics

Indian Institute of Science

Bangalore - 560012

India

iii

iv

List of publications

This thesis is based on the following publications :

1. “ Entanglement entropy in higher derivative holography ”

A. Bhattacharyya, A. Kaviraj and A. Sinha.

arXiv:1305.6694 [hep-th] JHEP 1308, 012 (2013)

2. “ On generalized gravitational entropy, squashed cones and holography ”

A. Bhattacharyya, M. Sharma and A. Sinha.

arXiv:1308.5748 [hep-th] JHEP 1401, 021 (2014)

3. “ Constraining gravity using entanglement in AdS/CFT ”

S. Banerjee, A. Bhattacharyya, A Kaviraj, K. Sen and A. Sinha.

arXiv:1401.5089 [hep-th] JHEP 1405, 029 (2014)

4. “ On entanglement entropy functionals in higher derivative gravity theories”

A. Bhattacharyya and M. Sharma.

arXiv:1405.3511 [hep-th] JHEP 1410, 130 (2014)

5. “ Renormalized Entanglement Entropy for BPS Black Branes ”

A. Bhattacharyya, S. S. Haque and A. Veliz-Osorio.

arXiv:1412.2568 [hep-th] Phy. Rev. D 91, 045026 (2015)

The following works were done during my PhD but is not included in this thesis :

1. “ On c-theorems in arbitrary dimensions ”

A. Bhattacharyya, L. Y. Hung, K. Sen and A. Sinha.

arXiv:1207.2333 [hep-th] Phy. Rev. D 86, 106006 (2012)

2. “ Entanglement entropy from the holographic stress tensor ”

A. Bhattacharyya and A. Sinha.

arXiv:1303.1884 [hep-th] Class. Quantum Grav 30, 235032 (2013)

3. “ Entanglement entropy from surface terms in general relativity ”

A. Bhattacharyya and A. Sinha.

arXiv:1305.3448 [hep-th] IJMPD 22 12, 1342020 (2013)

4. “ Attractive holographic c-functions ”

A. Bhattacharyya, S. S. Haque, V. Jejjala, S. Nampuri and A. Veliz-Osoio.

arXiv:1407.0469 [hep-th] JHEP 1411, 138 (2014)

v

vi

5. “ Viscosity bound for anisotropic superfluids in higher derivative gravity ”

A. Bhattacharyya and D. Roychowdhury.

arXiv:1410.3222 [hep-th] JHEP 1503, 063 (2015)

6. “ Lifshitz Hydrodynamics And New Massive Gravity ”

A. Bhattacharyya and D. Roychowdhury.

arXiv:1503.03254

Synopsis

One of the recent fundamental developments in theoretical high energy physics is the AdS/CFT

correspondence [1, 2, 3, 4] which posits a relationship between Quantum Field Theories (QFT)

in a given dimension and String Theory on a higher dimensional anti- de Sitter (AdS) space-

time. This has revolutionised our understanding of QFTs (more specifically conformal field

theories (CFTs)) and string theory/gravity, and has far reaching consequences for explo-

rations into a vast array of physical phenomena. Using the elegant formalism provided by

this powerful duality, often called “holography”, one can now use fundamental physical ob-

servables in QFT to better understand the nature of quantum gravity. The theoretical tools

provide a translation of calculable field theoretic observables into the language of gravity

thereby leading to the construction of holographic models for several interesting QFTs.

Entanglement is a fundamental physical property of all quantum systems. From models

of various condensed matter systems to its application as a tool for secure and fast communi-

cation in quantum information theory [5], it serves as an intersection point between different

subfields of physics [6]. From the AdS/CFT point of view quantum entanglement connects

geometry with quantum information, providing a window to understand how the bulk gravity

physics emerges from the holographic field theoretic viewpoint. Probing various aspects of

this connection in detail will be the broad theme of this thesis.

For extended, many-body systems, the most well known measure of quantum entangle-

ment is the “Entanglement Entropy” (EE) which is also the best understood measure within

the holographic framework. In early 2006, Ryu and Takayanagi (RT) gave a simple and

elegant prescription for computing this quantity using AdS/CFT duality within Einstein

gravity [7, 8]. They proposed that EE for a subsystem within an extended system (QFT),

is computed by the (proper) area of a static, codimension- 2, “extremal” surface inside the

dual AdS spacetime. The RT proposal has passed several non-trivial consistency checks, for

example strong sub-additivity, area law to name a few [9]. A remarkable aspect of the pro-

posal is the ease with which EE can now be calculated, while it is well known that obtaining

EE from first principles in QFT presents several technical challenges which have so far been

surmounted only in some 2d field theories using the “replica method” [10, 11, 12].

The most intriguing aspect of the RT proposal is its striking similarity to Bekenstein-

Hawking (BH) entropy which is proportional to the area of a black hole horizon, further

confirming an intimate relationship between entropy and geometry [13, 14, 15, 16]. This

leads to the natural question: what is the connection between EE and BH entropy? This

question has been sharpened recently by Lewkowycz and Maldacena (LM) via the concept of

Generalized Gravitational Entropy which extends the QFT replica trick to a replica symmetry

vii

for the dual space-time [17]. This was used to prove the RT conjecture successfully by deriving

the correct extremal surface equation for two derivative gravity theories. In this thesis I have

studied the generalization of LM method for higher derivative gravity theories [18, 19, 20,

21, 22, 23] describing holographic duals (of QFT’s with finite number of colours) and finite

’t Hooft coupling which takes the AdS/CFT correspondence beyond the usual supergravity

limit. If one wants to use AdS/CFT to study real life systems then it is absolutely necessary to

incorporate the finite coupling effect into the theory and hence the study of higher derivative

effects becomes very important. In these two papers [21, 22] I have formulated a proof for

the existence of the entropy functionals for certain higher derivative theories extending LM

method. We have shown that the for a certain special class of higher derivative theories

there exist well defined entropy functionals. To extend this proof for more general theories

of gravity opens up a possibility of breaking replica symmetry in the bulk space-time [24].

For higher derivative gravity, black hole entropy for a large class of stationary black

holes with bifurcate killing horizon is given by the well known Wald prescription [25, 26, 27]

which relates the concept of the Noether charge with the black hole entropy. Iyer and Wald

proposed a generalization for dynamical horizons. This throws up the question whether

there is a relation between these EE functionals and the Noether charge, and whether we can

derive them using the approach of Iyer and Wald. For a certain class of theories I have shown

that there exists a relation between these two [28] but a more rigorous proof is needed. This

somewhat firms up the area-entropy relation for arbitrary surfaces and proves the existence of

holographic EE functionals for higher curvature theories thereby extending the applicability

of Iyer-Wald formalism beyond the bifurcation surface.

Apart from this, it is well known that there exist several measures of quantum entangle-

ment, each satisfying a variety of mathematical inequalities and conditions [5]. Translating

these into the language of holography constrains the dual gravity theory and will lead to

general statements about the consistency of the theory. In this thesis I have discussed one

such measure namely Relative entropy [29], the positivity of which has led to constraints on

the underlying gravity theory [30]. Also entanglement entropy is a very useful tools for prob-

ing renormalization group (RG) flow from the holographic point of view [34, 31, 32, 35, 33].

We end with exploring the concept of renormalized entanglement entropy [36, 37] and its

application in probing RG flow in the context of N = 2 gauged supergravity [38].

References

[1] J. M. Maldacena, “The large N limit of superconformal field theories and supergravity,”

Adv. Theor. Math. Phys. 2, 231 (1998), [arXiv:hep-th/9711200]

[2] S. S. Gubser, I. R. Klebanov and A. M. Polyakov, “Gauge theory correlators from

noncritical string theory,” Phys. Lett. B428 (1998) 105, [arXiv:hep-th/9802109]

[3] E. Witten, “ Anti-de Sitter space and holography,” Adv. Theor. Math. Phys. 2 (1998)

253, [arXiv:hep-th/97802150]

[4] O. Aharony, S. S. Gubser, J. Maldacena H. Ooguri and Y. Oz, ” Large N field

theories, String theory and gravity,” Phys. Rept, 323 (2000) 183-386, [arXiv:hep-th/

9905111].

[5] Michael A. Nielsen and Isaac L. Chuang, “ Quantum Computation and Quantum

Information”, Cambridge University Press, 23-Oct-2000

[6] J. Eisert, M. Cramer and M. B. Plenio, “Area laws for the entanglement entropy - a

review,” Rev. Mod. Phys. 82 (2010) 277 [arXiv:0808.3773 [quant-ph]] and the references

there in.

[7] S. Ryu and T. Takayanagi, “Holographic derivation of entanglement entropy from

AdS/CFT,” Phys. Rev. Lett. 96 (2006) 181602 [hep-th/0603001].

[8] T. Nishioka, S. Ryu and T. Takayanagi, “Holographic Entanglement Entropy: An

Overview,” J. Phys. A 42 (2009) 504008 [arXiv:0905.0932 [hep-th]].

[9] S. Ryu and T. Takayanagi, “Aspects of Holographic Entanglement Entropy,” JHEP

0608 (2006) 045 [hep-th/0605073] and the references there in.

[10] C. Holzhey, F. Larsen and F. Wilczek, Nucl. Phys. B 424 (1994) 443 [hep-th/9403108].

[11] P. Calabrese and J. L. Cardy, “ Entanglement entropy and conformal field theory”,

Journal of Physics A: Mathematical and Theoretical, Volume 42, Issue 50, article id.

504005, 36 pp. (2009).

[12] P. Calabrese and J. L. Cardy, “Entanglement entropy and quantum field theory,” J.

Stat. Mech. 0406 (2004) P06002 [hep-th/0405152].

[13] Luca Bombelli, Rabinder K. Koul, Joohan Lee, and Rafael D. Sorkin, “Quantum

source of entropy for black holes”, Phys. Rev. D 34, 373

[14] M. Srednicki, “Entropy and area,” Phys. Rev. Lett. 71 (1993) 666 [hep-th/9303048].

[15] S. N. Solodukhin, “Entanglement entropy of black holes,” Living Rev. Rel. 14 (2011) 8

[arXiv:1104.3712 [hep-th]] and the references there in.

[16] E. Bianchi and R. C. Myers, “On the Architecture of Spacetime Geometry,” Class.

Quant. Grav. 31 (2014) 21, 214002 [arXiv:1212.5183 [hep-th]].

[17] A. Lewkowycz and J. Maldacena, “Generalized gravitational entropy,” JHEP 1308

(2013) 090 [arXiv:1304.4926 [hep-th]].

[18] L. Y. Hung, R. C. Myers and M. Smolkin, “On Holographic Entanglement Entropy and

Higher Curvature Gravity,” JHEP 1104 (2011) 025 [arXiv:1101.5813 [hep-th]].

[19] X. Dong, “Holographic Entanglement Entropy for General Higher Derivative Gravity,”

JHEP 1401 (2014) 044 [arXiv:1310.5713 [hep-th], arXiv:1310.5713].

[20] J. Camps, “Generalized entropy and higher derivative Gravity,” JHEP 1403 (2014) 070

[arXiv:1310.6659 [hep-th]].

[21] A. Bhattacharyya, A. Kaviraj and A. Sinha, “Entanglement entropy in higher derivative

holography,” JHEP 1308 (2013) 012 [arXiv:1305.6694 [hep-th]].

[22] A. Bhattacharyya and M. Sharma, “On entanglement entropy functionals in higher

derivative gravity theories,” JHEP 1410 (2014) 130 [arXiv:1405.3511 [hep-th]].

[23] R. X. Miao and W. z. Guo, “Holographic Entanglement Entropy for the Most General

Higher Derivative Gravity,” arXiv:1411.5579 [hep-th].

[24] J. Camps and W. R. Kelly, “Generalized gravitational entropy without replica symme-

try,” JHEP 1503 (2015) 061 [arXiv:1412.4093 [hep-th]].

[25] R. M. Wald, “Black hole entropy is the Noether charge,” Phys. Rev. D 48 (1993) 3427

[gr-qc/9307038].

[26] V. Iyer and R. M. Wald, “Some properties of Noether charge and a proposal for dynam-

ical black hole entropy,” Phys. Rev. D 50 (1994) 846 [gr-qc/9403028].

[27] V. Iyer and R. M. Wald, “A Comparison of Noether charge and Euclidean methods

for computing the entropy of stationary black holes,” Phys. Rev. D 52 (1995) 4430

[gr-qc/9503052].

[28] A. Bhattacharyya, M. Sharma and A. Sinha, “On generalized gravitational entropy,

squashed cones and holography,” JHEP 1401 (2014) 021 [arXiv:1308.5748 [hep-th]].

[29] D. D. Blanco, H. Casini, L. Y. Hung and R. C. Myers, “Relative Entropy and Hologra-

phy,” JHEP 1308 (2013) 060 [arXiv:1305.3182 [hep-th]].

T. Faulkner, M. Guica, T. Hartman, R. C. Myers and M. Van Raamsdonk, “Gravitation

from Entanglement in Holographic CFTs,” JHEP 1403 (2014) 051 [arXiv:1312.7856

[hep-th]].

[30] S. Banerjee, A. Bhattacharyya, A. Kaviraj, K. Sen and A. Sinha, “Constraining gravity

using entanglement in AdS/CFT,” JHEP 1405 (2014) 029 [arXiv:1401.5089 [hep-th]].

[31] H. Casini and M. Huerta, “A Finite entanglement entropy and the c-theorem,” Phys.

Lett. B 600 (2004) 142 [hep-th/0405111].

[32] H. Casini and M. Huerta, “A c-theorem for the entanglement entropy,” J. Phys. A 40

(2007) 7031 [cond-mat/0610375].

[33] R. C. Myers and A. Sinha, “Holographic c-theorems in arbitrary dimensions,” JHEP

1101 (2011) 125 [arXiv:1011.5819 [hep-th]].

[34] H. Casini and M. Huerta, “On the RG running of the entanglement entropy of a circle,”

Phys. Rev. D 85 (2012) 125016 [arXiv:1202.5650 [hep-th]].

[35] H. Casini and M. Huerta, “Positivity, entanglement entropy, and minimal surfaces,”

JHEP 1211 (2012) 087 [arXiv:1203.4007 [hep-th]].

[36] H. Liu and M. Mezei, “A Refinement of entanglement entropy and the number of degrees

of freedom,” JHEP 1304 (2013) 162 [arXiv:1202.2070 [hep-th]].

[37] H. Liu and M. Mezei, “Probing renormalization group flows using entanglement entropy,”

JHEP 1401 (2014) 098 [arXiv:1309.6935 [hep-th], arXiv:1309.6935].

[38] A. Bhattacharyya, S. Shajidul Haque and A. Veliz-Osorio, “Renormalized Entanglement

Entropy for BPS Black Branes,” Phys. Rev. D 91 (2015) 4, 045026 [arXiv:1412.2568

[hep-th]].

Acknowledgements

First and foremost, I would like to convey my sincere thanks to my advisor, Aninda Sinha for

his generous support as well as outstanding guidance during the entire tenure of my doctoral

research. Being his first PhD student was always a stimulating experience. I thank him

for guiding me throughout my doctoral work and helping me to complete my PhD in just

3 years. Apart from learning a great deal of physics from him, he has helped me a lot to

improve my soft skills.

I am also thankful to Menika Sharma, Ling-Yan (Janet) Hung, Shajid Haque, Alvaro Veliz

Osorio, Vishnu Jejjala, Suresh Nampuri and Dibakar Roychowdhury for useful collaborations

and numerous productive discussions which has added a lot towards my understanding of

the subject itself.

I would also like to thank Rajesh Gopakumar, Jose Edelstein, Axel Kleinschmidt, Johanna

Erdmenger, Tadashi Takayanagi, Heng-Yu Chen, Janet Hung, Shamik Banerjee, Vishnu

Jejjala for inviting me to give seminars. I am also thankful to them for numerous stimulating

discussions in various occasions. I am also thankful to Ashoke Sen, Rob Myers and Joan

Camps for valuable discussions.

I would also like to thank all the professors of the Department of Physics and the Centre

for High Energy Physics for providing beautiful courses. In particular, I would like to thank

Prof. Justin David for providing a beautiful course on QFT. I thank the Chairman of CHEP,

B Ananthanarayan for striving to maintain a vibrant and simulating atmosphere in the

department. I am indeed grateful to my Integrated PhD batchmates and colleagues at the

Centre for High Energy Physics for creating friendly and competitive atmosphere. During my

stay at IISc, I found various departmental activities like the weekly math-phys meets, journal

club sessions, seminars and colloquia etc. as quite stimulating and in particular playing a

very crucial role in developing the scientific mind. Finally, I would specially like to thank

Apratim and Shouvik for helping me enormously with all the diagrams and the Latex.

I thank the Indian Institute of Science for their generous financial support for attending

numerous conferences and visiting other research institutes in India and abroad during my

tenure.

Finally, I’m very thankful to my parents for giving me constant support and never giving

up hope on me.

Arpan Bhattacharyya

Bangalore, June 2015.

xiii

xiv

xv

To my parents.

xvi

xvii

“Somewhere, something incredible is waiting to be known”.

– Carl Sagan

xviii

Contents

1 Introduction 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1.1 Quantum Entanglement . . . . . . . . . . . . . . . . . . . . . . . . 3

1.1.2 Entanglement entropy and Holography . . . . . . . . . . . . . . . . 9

2 Holographic entanglement entropy functionals: A derivation 29

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

2.2 Entropy functional for general theories of gravity . . . . . . . . . . . . . . . 32

2.3 Test of the entropy functional for R2 theory . . . . . . . . . . . . . . . . . . 34

2.3.1 Minimal surface condition from the entropy functional . . . . . . . . 36

2.3.2 Minimal surface condition from the Lewkowycz-Maldacena method . 41

2.3.3 The stress-energy tensor from the brane interpretation . . . . . . . . 49

2.4 Quasi-topological gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

2.4.1 The entropy functional . . . . . . . . . . . . . . . . . . . . . . . . . 52

2.4.2 Universal terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2.4.3 Minimal surface condition . . . . . . . . . . . . . . . . . . . . . . . . 54

2.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3 Entanglement entropy from generalized entropy 69

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

3.2 Generalized entropy and Fefferman-Graham expansion . . . . . . . . . . . . 69

3.2.1 Four derivative theory . . . . . . . . . . . . . . . . . . . . . . . . . . 72

3.2.2 New Massive Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

3.2.3 Quasi-Topological Gravity . . . . . . . . . . . . . . . . . . . . . . . . 75

3.2.4 α′3 IIB supergravity . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

3.3 Comment about singularities in the metric . . . . . . . . . . . . . . . . . . . 76

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4 Connection between entanglement entropy and Wald entropy 81

xix

xx CONTENTS

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

4.2 Wald Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

4.3 Four derivative theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

4.3.1 Cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

4.3.2 Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

4.4 Quasi-Topological gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

4.5 α′3 IIB supergravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

4.6 Connection with Ryu-Takayanagi . . . . . . . . . . . . . . . . . . . . . . . . 87

4.7 Comments on the connection with the Iyer-Wald prescription . . . . . . . . . 89

4.8 Universality in Renyi entropy . . . . . . . . . . . . . . . . . . . . . . . . . . 90

4.9 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

5 Constraining gravity using entanglement entropy 95

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

5.2 Smoothness of entangling surface . . . . . . . . . . . . . . . . . . . . . . . . 96

5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

6 Relative entropy 101

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

6.2 Relative entropy considerations . . . . . . . . . . . . . . . . . . . . . . . . . 103

6.3 Relative entropy in Gauss-Bonnet holography . . . . . . . . . . . . . . . . . 109

6.3.1 Linear order calculations . . . . . . . . . . . . . . . . . . . . . . . . . 111

6.3.2 Quadratic corrections . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

6.3.3 Constant Tµν . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

6.3.4 Shockwave background . . . . . . . . . . . . . . . . . . . . . . . . . . 115

6.3.5 Correction from additional operators . . . . . . . . . . . . . . . . . . 117

6.4 Relative entropy for an anisotropic plasma . . . . . . . . . . . . . . . . . . . 118

6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

7 Coding holographic RG flow using entanglement entropy 131

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

7.2 Renormalized Entanglement Entropy . . . . . . . . . . . . . . . . . . . . . . 132

7.3 BPS black objects in AdS4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

CONTENTS xxi

7.4 REE for BPS black branes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

7.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

8 Conclusions 145

xxii

1 Introduction

1.1 Introduction

The concept of entanglement is a very old one and dates back to early 1930’s when quantum

mechanics was born. One of the measures of quantum entanglement is the entanglement

entropy. Although it has played a crucial role in understanding some aspects of quantum

mechanics, its application remains rather limited mainly due to its non local behaviour until

late 90’s.

In early 1970’s Bekenstein proposed that the black hole entropy (SBH) follows an area

law [1] .

SBH =A

4GN

(1.1)

where A is the area of the horizon and GN is the Newton constant. This formula is quite

counter-intuitive as entropy is usually an extrinsic quantity, depends on the volume of the

system. Later Stephen Hawking showed that a black hole emits radiation with a well defined

temperature [2], thereby establishing the concept of black hole entropy. It was observed that

the calculation of the entropy associated with the radiation emitted from the black hole is

plagued by the presence of ultraviolet divergences. These divergences can be associated with

the particles close to the horizon [3] and one has to regulate them to get a finite answer

for the entropy. Then in early 1980’s Bombelli, Koul, Lee and Sorkin in their seminal work

[4] showed that one can possibly understand that black hole entropy using the concept of

‘entanglement’. To the observers outside the black hole horizon there is no information about

the spacetime inside the horizon. They considered scalar fields in the black hole background

and traced out the spacetime inside the horizon, thereby defining a “reduced density” matrix

for the system. Using this, they computed the von-Neumann entropy and it was shown that

the entropy follows an area law. Later it was generalized by Srednicki [5] for massless scalar

fields in flat spacetime. He also showed that if one divides the spacetime in two parts, then

the entropy associate with the reduce density matrix for one of these two parts is proportional

to the area of the boundary between these two halves. Later this concept of “ entanglement

entropy”(EE) was made concrete by Callan, Holzhey, Larsen and Wilczek [6] and separately

by Susskind and Uglum [7]. From their work it is evident that the EE exhibits a universal

1

2 1.1. INTRODUCTION

behaviour, it goes as a logarithm of correlation length in 1 + 1 dimensions after suitably

regulating the ultraviolet divergences. But still its application remain rather limited as in

general it is very hard to compute this quantity for generic field theories.

But in early 2000, Calabrese and Cardy used the “ Replica Trick” formulated by Callan,

Holzey, Larson and Wilczek successfully to compute the EE for many cases in the context of

1+1 dimensional conformal field theory (CFT) [8], thereby increasing its physical importance.

After that EE has been calculated extensively not only for 1+1 dimensional CFT but also for

various other simple quantum field theories both analytically and numerically. Also in recent

times it has been successfully computed numerically using the lattice technique for quantum

many body systems [9]. Although the techniques employed for computing EE is very hard

and yet to be developed fully, still in recent times, we have lots of data regarding quantum

entanglement and EE coming from both analytical and numerical approaches [10, 11].

Figure 1.1: Entanglement and its diverese applications

in physics. (Picture courtesy- From the talk given by

Prof. Robert. C. Myers in the Conference “Entangle-

ment from gravity”, ICTS, 2014, Bangalore, India)

Recently EE has found many

applications in various branches of

physics like, quantum information,

condensed matter system, statis-

tical physics and in AdS/ CFT

[9, 10, 11, 12, 13, 14]. It serves as an

intersecting point between various

subfields of physics. In recent times

it has played a crucial role in un-

derstanding the nature of hologra-

phy (AdS/ CFT correspondence).

AdS/ CFT correspondence, com-

monly known as “ Holography” is

one of the most important dualities

in physics. It postulates that grav-

ity emerges from a certain class of

field theories. It is still not known

rigorously how holography works

from first principles. EE has the

merit to shed light on this problem

as it connects quantum information

of the system with geometry. There

exists an interesting connection be-

tween geometry and EE and our

goal in this thesis will be to explore some aspects of this connection. Recently apart from EE,

CHAPTER 1. INTRODUCTION 3

many different tools of quantum entanglement like, entanglement negativity, differential en-

tropy, quantum error coding, relative entropy, cMERA (continuous multiscale entanglement

renormalization ansatz) etc, have been used in an attempt to build geometry from the field

theory data, is some sense trying to prove the holographic principle [15, 16, 17, 18, 19, 20].

But still EE plays the central character in this program and we will use this to learn many

lessons of gravity in the holographic set up.

1.1.1 Quantum Entanglement

Let us first consider a quantum mechanical system consists of two spin 12

particles. We

denote the total Hilbert space by H . Now HA and HB denote the Hilbert spaces of the two

individual particles . Also,

H = HA ⊗HB.

Now consider the following state belonging to H,

|ψ1 >=1√2

(| ↑ + ↓>)A ⊗ (| ↓ + ↑>)B . (1.2)

This state is not an entangled state as there is no correlation between the two particles in

this state. On the other hand, if we consider the following state

|ψ2 >=1√2

(| ↑A↓B> −| ↑B↓A>), (1.3)

then one cannot factorize this state in terms of the individual particle states. So this state is

an example of “ entangled state”. In other words, entanglement is a property of a quantum

mechanical system that tells us, that one cannot describe the underlying pair of particles

belonging to this particular state independently. All the physical properties of the two

individual particles are correlated with each other. This argument can be extended for any

number of particles and interesting things shows up when one consider many body systems

due to the non local nature of the entanglement.

One of the measures of entanglement is the entanglement entropy (EE). Let us first see

how we can define this. First step is to define a “ density matrix” for the full system. If |ψ >is the wavefunction characterizing the total system then the density matrix can be defined

in the following way,

ρ = |ψ >< ψ|. (1.4)

Next step is to define the “ reduced density matrix”. Suppose we want to compute the EE for

the sub-system A. Then we first trace out the degrees of freedom corresponding the system

B.

ρA = TrBρ =< b|ρ|b >, (1.5)

4 1.1. INTRODUCTION

where ρA is the reduced density matrix for the subsystem A. Now we define EE by defining

the “von-Neumann entropy” which is,

SEE(A) = −TrAρA ln ρA . (1.6)

Here we have completed the trace over the subsystem A. Now this is a good time to cook

up some examples and elucidate the process of computing EE. First consider the state as

mentioned in Eq. (1.2). We trace out B and the corresponding reduced density matrix is ,

ρA =1

2

(1 1

1 1

). (1.7)

Now to compute the EE as defined in Eq. (1.6) one first diagonalize this and find the eigen-

values. In terms of the eigenvalues Eq. (1.6) becomes

SEE = −2∑i=1

λi lnλi. (1.8)

This in turn gives,

SEE = 0. (1.9)

So the entropy is zero and hence the state is not entangled.

Now consider the state as shown in Eq. (1.3). Corresponding reduced density matrix is,

ρA =1

2

(1 0

0 1

). (1.10)

The entropy is,

SEE = ln 2. (1.11)

It is an entangled state, in fact it is a maximally entangled state. So whenever SEE is nonzero

the state is entangled.

Now let us consider a more complicated system, a system of two coupled oscillators. This

type of systems are considered in [5] and we will review their calculation here to demonstrate

the increasing difficulty of computing this quantity when one consider quantum many body

systems.

Let us start by writing the hamiltonian that describes the two coupled oscillators.

H =1

2

[p2

1 + p22 + k1(x2

1 + x22) + k2(x1 − x2)2

], (1.12)

1 and 2 respectively denote the two oscillators. Then one defines the canonical coordinate

xA =x1 + x2√

2, xB =

x1 + x2√2

. (1.13)


In terms of these coordinates the ground state wave function can be written as

ψ0(x1, x2) = (ωAωBπ2

)1/4e−(ωAx2A+ωBx

2B)/2, (1.14)

where ωA = k1/21 and ωB = (k1 + 2k2)1/2 are the two frequencies corresponding to the two

normal modes. Now suppose we integrate out the oscillator 2. The reduced density matrix

is defined as,

ρ(x1, x′1) =

∫ ∞−∞

dx2ψ0(x1, x2)ψ0(x′1, x2)∗. (1.15)

This gives

ρ(x1, x′1) = (

δ

π)1/2e−α(x21+x′21 )/2+βx1x′1 , (1.16)

where β = 14

(ωA−ωB)2

(ωA+ωB)and δ = α − β = 2ωAωB

(ωA+ωB). Then we have to just compute the SEE

as defined in Eq. (1.6). To do that we have to find the eigenvalues of this reduced density

matrix. In this case we are fortunate, as one can easily solve this problem and eigenvalues

are given as,

λn =(

1− β

α + (α2 − β2)1/2

)( β

α + (α2 − β2)1/2

)n. (1.17)

Then

SEE = −∑n

λn lnλn. (1.18)

After performing this sum, which is somewhat tedious, we get

SEE(k1

k2

) = − ln(

1− β

α + (α2 − β2)1/2

)− β

α− β + (α2 − β2)1/2ln( β

α + (α2 − β2)1/2

).

(1.19)

Ultimately SEE is just a function of the ratio of k1 and k2.

Now the stage is prepared for us to generalize this concept for field theory. In the field

theory the problem becomes much more difficult and subtle. One key issue is to factorize the

Hilbert space. One way is to discretize the system over a lattice.1 However one can still use

the von-Neumann formula as defined in Eq. (1.6), but one has to deal with the ultraviolet

divergences that are present in the field theory.

As shown in the Fig. (1.2), we can consider a particular region in the field theory denoted

by A. To compute the EE for this region we trace out the remaining portion. The system is

discretized over the full space. The SEE(A) for the subsystem A is roughly proportional to

the number of links cut by the boundary of the region A. So it is telling us that, indeed EE

1 Several ambiguities might enter in the calculation because of the discretization, specially for the gauge

theory. But still one can extract a meaningful answer for EE in the field theory.

6 1.1. INTRODUCTION

Figure 1.2: System is discretized on a lattice and H =∑

iHi

is proportional to the area of the boundary dividing the two regions. If the total system is

in the pure state, then one can show

SEE(A) = SEE(B), (1.20)

where B denotes the remaining portion of the spacetime. From this one can intuitively guess

that SEE is proportional to the number of degrees of freedom that live at the boundary

between the two regions.

Although there is no formal proof of area law, it has been checked for many instances.

For almost all the cases when one considers a ground state of a local hamiltonian, one indeed

gets the area law. It is more or less robust, although the violation of it has been observed

for the excited states and also for non-local hamiltonians [21]. Let us close this section by

briefly sketching an argument for the area law. We will follow [5] and consider scalar field

theory. To show this let us go back to the oscillator case as almost all the field theory can

be described effectively using the coupled oscillators model. To start with let us write down

the Hamiltonian,

H =1

2

∫d3x[π2 + |∇φ(x)|2]. (1.21)

Here φ(x) denotes the scalar field and π is the canonical momentum. After this we express

this Hamiltonian in terms of the partial wave expansion of the scalar field,

φlm = x

∫dΩZlmφ(x). (1.22)

Zlm are the spherical harmonics. The above relation stems from the fact that we can expand

the scalar field in the basis of spherical harmonics. A similar relation can be written for the


conjugate momentum. We impose the canonical quantization relation,

[φlm(x), πl′m′(x′)] = iδll′δmm′δ(x− x′). (1.23)

In terms of this partial wave,

H =∑l,m

Hlm. (1.24)

Now,

Hlm =1

2

∫ ∞0

dx(π2lm(x) + x2[∂x(

φlmx

)]2 +l(l + 1)

x2φlm(x)

). (1.25)

Then we discretize the system. We put it on a lattice with a lattice spacing 1M

. The M plays

the role of the uv cutoff. The boundary condition imposed on φlm(x) is such that it vanishes

when x ≥ L where L is the length of the box in which the system is placed. Also

L = (N + 1)1

M, (1.26)

whereN is a integer and this relation shows that the system is discretized. So the Hamiltonian

becomes

Hlm =M

2

N∑j=1

[π2lm + (j +

1

2)2(φlm,j

j− φlm,j+1

j + 1

)2

+l(l + 1)

j2φ2lm,j

]. (1.27)

Now this looks exactly the same as the N coupled oscillators hamiltonian. We can proceed

as before extending the result of 2 coupled oscillators. We trace over the first n number of

sites to obtain the EE. Finally we get

SEE(n,N) =∑l

al(n)[− ln al(n) + 1], (1.28)

where al(n) = n(n+1)(2n+1)2

64l2(l+1)2+O( 1

l6). At this level we are only interested in the leading result

in l, as that will give after summing all values of l’s, the area like term. We perform the sum

over l numerically. We define a radius R midway between the outermost point which was

traced over and the innermost point which was not as,

R = (n+1

2)

1

M. (1.29)

Then it can be shown that the leading term of the EE is,

SEE = 0.30M2R2. (1.30)

From this it is clearly evident, that the EE corresponding to the ground state wavefunction

of a local hamiltonian of the scalar field satisfies the area law.

A more intuitive way to understand the area law [22] is to consider a particular entangling

region A as shown in Fig. (1.3). For simplicity let us stick to the massless scalar field model

8 1.1. INTRODUCTION

Figure 1.3: Modes straddling the boundary ∂A is responsible for SEE

in 3 + 1 dimensions. We expand the scalar field in terms of its modes. These modes are

quasi-localized and each has an momentum ~k. Now we know,

|k| = 2π

λ, (1.31)

where λ is the usual wavelength. The total number of modes inside the region A is given by,

N =

∫ kmax

kmin

dN =

∫ kmax

kmin

V d3k

(2π)3, (1.32)

where V is the volume of A. Now kmin = 2π2R

and kmin = 2πε

. As all the wavelengths are

localized inside A, maximum wavelength can atmost be equivalent to 2R and the minimum

wavelength is 0. But then kmin will be divergent, hence we have to put a uv cut-off ε. From

this the necessity of a cut-off becomes quite clear. Now we count the fraction of the modes

which resides at the boundary of A, responsible for the EE.

Ns ≈∫ kmax

kmin

αAdN

V, (1.33)

where Ns denotes the number of modes straddling the boundary and it is a fraction of the

total number of modes living inside A.

α denotes the thickness of the boundary (α << R) and A denotes its area. Now only the

mode localized near the boundary is responsible for SEE, so we can approximate α by 2πk

.

Also d3k = 4πk2dk and αd3k = 8π2kdk. We next perform the integration and the entropy is

proportional to Ns upto some phase-space factors. We get,

SEE ≈2πA

ε2. (1.34)

So EE is proportional to the area, hence proportional to number of degrees of freedom residing

at the boundary between the two region. It can be also shown in the same way that for 2 + 1

dimensions it is proportional to the circumference and in 1 + 1 it goes as logarithm.


Lastly, SEE satisfies one more important property, namely strong subadditivity. For a

bipartite system it tells us that,

SEE(A) + SEE(B) ≥ SEE(A ∪B) + SEE(A ∩B). (1.35)

This result can be extended for any arbitrary number of subsystems. This inequality provides

a non trivial constraints on SEE and any consistent holographic proposal should pass this

test.

1.1.2 Entanglement entropy and Holography

In this section we will review various facts about holographic entanglement entropy. The

main goal of this thesis is to understand the connection between EE and geometry, thereby

learning important lessons about underlying gravity theory in the holographic set up. By

holography we will mean AdS/ CFT correspondence. Two main character of this play is

Anti-de Sitter space (AdS) and conformal field theory (CFT). So before proceeding further

let us briefly comment on the structure of the conformal group and AdS spacetime [23].

Conformal group and structure of AdS

Conformal isometries keep the metric invariant upto a scale transformation. The conformal

transformations form a group by themselves. Poincare group comes as a subgroup under

the broad structure of the conformal group. The conformal transformation preserves the

angle between the two curves. These transformations consist of the following four kinds of

transformations.

Translation→ x′µ = xµ + aµ.

Lorentz → x′µ = Rµ

νxν , where infinitesimal matrix Rµν is antisymmetric.

Dilatation → x′µ = cxµ.

Special Conformal Transformation(SCT) → x′µ = xµ−cµx2

1−2 c.x+c2x2.

For SCT the conformal factor is (1−2 c.x+c2x2)2. SCT is nothing but a translation preceded

and followed by an inversion. The corresponding generators for the infinitesimal transforma-

tions are listed below. For a generic field

Translation(P µ) → −i∂µ.

Rotation (Jµν)→i(xµ∂ν − xν∂µ) + Sµν .

10 1.1. INTRODUCTION

Dilatation(D)→−i(d+ (x.∂)).

SCT(Kµ)→−i((2xµxν − 2gµνx2)∂ν + 2d.xµ) + 2xνSµν ,

where Sµν is an anti-symmetric spin matrix for a given field and satisfies the Lorentz

algebra. d is a real number that depends on the nature of the fields that are present in

the underlying theory.2 These generators satisfy the following commutation relations among

themselves.

[D,D] = 0 ,

[P a, P b] = 0 ,

[D,P a] = iP a,

[Jab, P c] = −i(gacP b − gbcP a) ,

[Jab, J cd] = −i(gadJ bc + gbcJad − gacJ bd − gbdJac) ,[Jab, D] = 0 ,

[D,Ka] = −iKa ,

[iKa, Kb] = 0 ,

[Ka, P b] = 2i(gabD − Jab) .

(1.36)

For example, in 2 + 1 dimensional flat spacetime we have the following 10 conformal genera-

tors,

J1 = ∂a = iPa ,

J2 = xb∂a − xa∂b = −iJab ,J3 = −(xa∂a) = −iD ,

J4 = (2xa(xd∂d)− (xdxd)∂a) = iKa.

(1.37)

a, b runs from 1 to 3. These generators satisfy the usual conformal commutation rules. Now

we will see what are the corresponding isometry generators of AdS4. We first write the AdS4

metric in poincare coordinates. This is the coordinate system we will often use throughout

this thesis.

ds2 =L2(dz2 + dx2 + dy2 + dt2)

z2. (1.38)

t denotes the Euclidean time. Then we do the following substitution

r =L

z. (1.39)

It gives

ds2 =L2

r2dr2 +

r2

L2(dt2 + dx2 + dy2). (1.40)

2 e.g for Fermion d = 32 and for Boson d = 1 .


Next we list all the 10 generators.

J1 = ∂t ,

J2 = ∂x ,

J3 = ∂y ,

J4 = x∂t − t∂x ,J5 = y∂x − x∂y ,J6 = t∂y − y∂t ,J7 = r∂r − t∂t − x∂x − y∂y ,

J8 = rt∂r −1

2t2∂t − tx∂x − ty∂y ,

J9 = rx∂r − tx∂t −1

2x2∂x − xy∂y ,

J10 = ry∂r − ty∂t − xy∂x −1

2y2∂y.

(1.41)

We make suitable identifications and t → it, such that the generators satisfy the usual

SO(3, 2) algebra [Jab, Jcd] = i[gadJbc+gbcJad−gacJbd−gbdJac] , where a, b, c, d ∈ 0, 1, 2, 3, 4,So basically they satisfy the same algebra as the CFT generators in one lower dimensions.

AdS/CFT

Now we describe what exactly this correspondence is. There are many dualities that exist

in the physics [24]. Among them AdS/CFT connects a strongly coupled field theory with a

weakly coupled gravity in Anti-de Sitter (AdS) space time [25]. It is a strong weak duality.

It has been observed that there exists an equivalence between a strongly coupled N = 4

supersymmetric SU(N) Yang-Mills (SYM) theory and Type IIB string theory on AdS5×S5

in the large N limit. Now consider a stack of N D3-branes. Open strings describe the

excitations of the D3-branes and the low energy dynamics is governed by N = 4 SYM gauge

theory. For this theory one can define a ’t-Hooft coupling λ = g2YMN = gsN. We can do a

perturbation theory when λ << 1 (also gs << 1). On the other hand we have closed string

excitations in the vacuum. This gives rise to the gravity multiplate in 10 dimensions, low

energy description of which is effectively given by Type IIB supergravity. One can construct

a metric solution for this theory for which the near horizon geometry looks like,

ds2 = α′[r2

√4πgsN

(−dt2 + dx21 + dx2

2 + dx23) +

√4πgsN

dr

r2+√

4πgsNdΩ25] (1.42)

We have assumed that α′ → 0 so that we can neglect stringy effects and work in the su-

pergravity regime. Identifying L2 = α′√

4πgsN, where L is the AdS radius we can see that


metric defined in Eq. (1.42) describes AdS5 × S5 geometry. Also string length ls =√α′ and

this description is valid when, (Lls

)4

= 4πgsN >> 1. (1.43)

This means that classical gravity description is valid when the AdS length scale is much

bigger than the string length and one can use this supergravity language when the ’t-Hooft

coupling becomes very large. Also we know g2YMN = 4πgsN. All these things point to the

fact that we have a classical gravity description when L >> ls in the bulk and it is equivalent

to a strongly coupled gauge theory at the boundary. Although this conjecture has not been

proved yet, it passes several important checks, for eg, matching of the spectrum of chiral

operators, correlation functions, supersymmetric indices etc. We obtain a precise dictionary

between field theory correlators and correlators of fields living inside the AdS space time.

One example is that, currents in the conformal field theory (CFT) side correspond to a gauge

field living inside the bulk spacetime.

One can easily see that the isometry group of AdS5 is SO(4, 2) and isometry group for S5

is SO(6). On the other hand the gauge theory remains invariant under the action of SO(4, 2)

conformal group and also possess an SO(6) R-symmetry. So we have obtained a geometric

realization of the field theory degrees of freedom.

Based on this, one can study systems described by strongly coupled field theories by using

equivalent classical gravity description. Holography is being used to study hydrodynamic

transports of quark gluon plasma, phase transitions in condensed matter systems etc [26]. 3

Holographic entanglement entropy

Although there exist several evidences supporting holographic principle, but it is still not

clear how gravity emerges from field theory. To understand this several tools have been

employed, EE is one of them. In AdS/CFT set up we will investigate EE and will see that it

will provide us with a nice geometrical problem. We will see that we can extract important

information about the underlying geometry, hence the gravity theory using this quantity.

Now as the AdS/CFT is a two way street, let us start by reviewing some basics about EE in

the CFT side of the story.

3 An analogous duality has been observed between the near horizon geometry of AdS3 × S3 ×M and

that of the low energy description on the branes in D1-D5 system in terms of 1+1 dimensional CFT. Also in

recent times holographic principles are being applied for other spacetimes, for eg, Lifshitz, de-Sitter etc [27].


Entanglement entropy in CFT

Most general structure of EE for a CFT in even spacetime dimensions is ,

SEE = c1Rd−2

εd−2+ · · ·+ c2 ln

(Rε

)+ · · · (1.44)

First term is the “ area term”. R denotes the radius of the entangling surface i.e the surface

for which the EE is computed. ε is the uv-cutoff. In even dimensions one gets a “logarithmic

term” known as universal term in EE. The coefficient proportional to this term is cut-off

independent and carries the information of the central charges of the underlying CFT. So

this term is also known as “ universal term” and we will be interested in computing this

term throughout this thesis for various theories of gravity. Also we will be considering time

independent scenarios and choose a particular time slice t = 0. Let us take an example. In

d = 4 dimensions, c2 takes the following form [28],

c2 = A

∫d2xR+ C

∫d2x(W abcdhachbd −K2

s +1

2KsabKsab). (1.45)

We have chosen t = 0 slice and rest of the 3 dimensional space is divided into two halves. So

the boundary between the entangling region and the rest of the space is a two dimensional

space. We will call it a “codimension-2” surface. The integration defined in Eq. (1.45) is

essentially over this boundary. c2 depends on the geometrical property of this codimension-2

surface. R is the Ricci scalar, Wabcd is the Weyl tensor and Ksab is the extrinsic curvature of

this surface. hac is the projection operator to the surface from the 4 dimensional spacetime.

hab = ηab − nsanbs. (1.46)

s denotes the two transverse directions and a, b are the surface indices. The extrinsic curva-

ture can be defined as,

Ksab = eαaeβb h

µαh

νβ∇µnsν . (1.47)

It has also an index (s) corresponding to the two transverse directions and two normals are

defined for that. Ks is the trace of Ksab. hµα is the bulk to surface projection operator and eαais the tangent vector 4. A and C are related to the two anomaly coefficient that are present

for the 4 dimensional CFT.

A =A

16π2, C =

C16π2

. (1.48)

A is known as Euler anomaly and C is know as Weyl anomaly. They show up in the non

vanishing part of trace of the stress energy tensor.

< T ii >=C

16π2W 2 − A

16π2E4 (1.49)

4eαa = ∂Xα

∂Xa , where µ and a denote respectively bulk and surface indices.


where, E4 = RijklRijkl−4RijR

ij +R2 is the 4 dimensional Euler tensor and W 2 is the square

of 4 dimensional Weyl tensor. i, j, k, l denote the 4 dimensional indices.

In general to compute EE form field theory one uses “ Replica Trick” [6, 8]. For that,

first step is to define Renyi-entropy (Sn) ,

Sn = − 1

n− 1ln trAρ

nA. (1.50)

Figure 1.4: Path integral formulation of

density matrix (“Aspects of holographic

entanglement entropy”, S. Ryu and

T. Takayanagi, arxiv:-:hep-th/0605073.)

Then one has to evaluate this quantity on a “

Replica space ”. First we have to write down the

reduced density matrix in path integral formalism

[6, 8]. For simplicity consider a 1 + 1 dimensional

space. We will closely follow the notations and con-

ventions of [13]. We choose a tE = 0 slice and

consider an interval (A) as our entangling region

as shown in the figure. We denote all the dynam-

ical fields that are present collectively as φ(tE, x)

where tE is the Euclidean time. We also impose

the following boundary condition on the fields liv-

ing inside A.

φ0(tE = 0+, x) = φ+(x). (1.51)

and for other fields living outside A we demand

that,

φ0(x) = φ′(x). (1.52)

The ground state wave function can be written as,

Ψ(φ0(x)) =

∫ φ0(tE=0+,x)=φ+(x)

tE=−∞Dφe−S(φ(x)). (1.53)

Complex conjugate of this is defined as,

Ψ(φ0(x))∗ =

∫ tE=∞

φ0(tE=0−,x)=φ−(x)

Dφe−S(φ(x)). (1.54)

From this one can construct the density matrix easily,

ρAφ+φ− =1

Z1

∫ tE=∞

tE=−∞Dφe−S(φ(x))Πx∈Aδ(φ(tE+ = 0, x− φ+(x))δ(φ(tE− = 0, x− φ−(x)).

(1.55)


We have taken care of the fact, that the boundary conditions imposed are obeyed by inserting

the delta functions. Then to evaluate the n’th Renyi-entropy we have to find the product of

the n copies of this density matrix glued with each other by making suitable identifications

which we do in the next step.

TrAρnA = ρAφ1+φ1− · · · ρAφn+φn− . (1.56)

Also we have imposed φi− = φi+1,+ for all i = 1, · · ·n. So this quantity gives rise to the

“Replica space ” which is similar to an n sheeted Riemann surface and has a discrete Z(n)

symmetry coming from permuting the n copies of the replica. So finally we have to compute

the path integral on this replicated manifold which is denoted by Zn.

T rAρn = (Z1)−n

∫replicaspace

Dφe−S(φ) =Zn

(Z1)n. (1.57)

This is an important formula for computing EE from field theory using the replica trick. We

will see in the later sections, what is its implication in the context of holography. Replica

method has been employed successfully for computing EE for 1 + 1 dimensional CFT ’s [8],

but for higher dimensions it is hard to apply this method [29].

Entanglement entropy in Holography

Now let us turn our attention to the holography, main hero of our story. Importance of EE

in this context is profound. We will see that it geometrize the problem in the bulk space

time. Before going into the details let us take an example, which will show the connection

between EE and geometry. We start by drawing an analogy with the quantum mechanics.

We consider CFTs on two spatially disconnected regions [30]. Next we consider a wave-

function for this system of the form,

|Ψ >= |ΨA > ×|ΨB > (1.58)

where A and B denote the individual wavefunctions of the two CFTs. So it is evident that

the state is not an entangled state as it is written as a direct product of two states. This

kind of state in holography corresponds to a disconnected geometry. Now let us consider two

disconnected CFTs placed on Sd. Ei is the energy corresponding to the ith eigenstate. Now

let us consider the following state,

|ψ >=∑i

e−βEi2 |Ei × |Ei > . (1.59)

This state is not a direct product state. So it is an “ entangled state”. From holography we

know that this corresponds to a thermofield double state and the dual geometry is an eternal


Figure 1.5: On the left hand side, we have shown a thermofield double state and its holo-

graphic dual eternal black hole is shown on the right hand side. (Picture courtesy- “ Building

up spacetime with quantum entanglement ” by Mark Van Raamsdonk, arXiv:-1005.3035)

black hole [31]. So this shows that quantum superposition of two states of two classically

disconnected CFTs corresponds to a classically connected geometry.

This analogy makes things more interesting as one can possibly hope to understand

geometry using EE. In the context of AdS/CFT one can ask, whether it is possible to associate

a concept of entropy for any arbitrary region of field theory sitting at the boundary of AdS?

If so, then the next obvious thing is to ask which portion of the bulk spacetime capture that

information? The answer comes in the form of Ryu-Takayanagi (RT) proposal for Einstein

gravity [32]. The proposal is very simple, it says that one can attach a notion of entanglement

entropy for any arbitrary region at the boundary of AdS at a constant time slice and the

corresponding entropy is given by the area of some special codimension 2 surface (γA) inside

the bulk spacetime. So consider a d dimensional bulk space time. Then the EE associated

with a region A at the boundary is given by,

SEE(A) =2π

`d−2p

Area(γA). (1.60)

`p is the planck length. To remind ourselves , this proposal is made for Einstein gravity.

S = − 1

2`d−1p

∫dd+1x[R− 2Λ]. (1.61)

The AdS metric is a solution of the action mentioned in Eq. (1.61) with a negative cosmo-

logical constant

Λ = −d(d− 1)

2L2.


Figure 1.6: holographic prescription for EE:

γA is the extremal surface

Now let us demystify this proposal and

see how it works. Let us for simplicity

consider a AdS5 metric and a spherical re-

gion (A) at the boundary as shown in the

Fig. (1.6).

ds2 =L2

z2(dz2+dt2+dr2+r2dθ2+r2 sin(θ)2dφ2).

(1.62)

z is the radial coordinate of the AdS space.

We consider a constant time slice and set

the Euclidean time t = 0. Now to evalu-

ate EE associated with the region A we will

employ RT method which tells us to find a

particular surface (γA) extending inside the

bulk spacetime which minimizes the area en-

closed. Area of that particular surface will give us the EE and γA is known as the “ minimal

surface”. To elucidate this further, we put t = 0 and r = f(z) in the (1.62) to obtain an

induce metric for the minimal surface. Then we evaluate the area for this.

SEE =8π2L3

`3p

∫dzf(z)2

√1 + f ′(z)2

z3. (1.63)

We minimize SEE defined in Eq. (1.63), thereby obtaining an equation for f(z). Solving that

equation with the following boundary condition ,

f(z = 0) = f ′(z = 0) = 0, (1.64)

where the prime denotes the derivative with respect to z, we get 5

f(z) =√R2 − z2. (1.65)

R is the radius of the spherical region (A) at the boundary. We plug this into Eq. (1.63) and

expand the resulting expression around z=0 which is divergent. We introduce a uv-cutoff ε

and finally we get,

SEE =4π2L3

`3p

[R2

ε2− ln(

R

ε)]. (1.66)

The leading term follows the area law and also we get an universal term.

One can consider more general surfaces and still get the expected results for the universal

term (1.45). RT proposal passes several consistency checks, it produces correct universal

5For any generic entangling surface, the extremization condition in the context of Einstein gravity can be

written as Ks = 0 where, Ks is the trace of the extrinsic curvature of this codimension-2 surface.


term for EE, which in the context of AdS3/CFT2 reproduces many know results coming

from CFT computations [13]. In higher dimensions all these holographic results give us some

insights about the field theory. Also RT method has been successfully applied for computing

EE in excited state [33, 34], for time dependent cases (covariant RT proposal [35]) and for

many other interesting cases 6. It also satisfies the strong subadditivity relation [36]. More

importantly it has provide us with a geometrical interpretation of boundary data which is

the first step towards understanding how gravity (geometry) emerges form the field theory.

We will also consider various higher derivative corrections to Einstein gravity. As we are

considering effective theories of gravity, these corrections naturally arise as we integrate out

higher momentum modes in Wilsonian RG flow. These higher derivative corrections describe

holographic duals of field theory with finite ’t-Hooft coupling ( they correspond to QFTs with

finite number of colour) and can arise either as 1N

corrections or when one considers string

loop corrections [38]. But these terms take the AdS/CFT description beyond supergravity

limit, thereby providing us with a good platform to understand the effect of finite coupling

on the underlying field theory.

We will explore EE in this context of higher derivative gravity, hoping to understand

some important lessons about these effective theories. For eg, one can add an Gauss-Bonnet

term to the Einstein Lagrangian [39, 40].

SEGB = − 1

2`3p

∫d5x[R− 2Λ + λL2(RµναβR

µναβ − 4RαβRαβ +R2)

](1.67)

where λ is the Gauss -Bonnet coupling. AdS is still a solution for this theory provided the

following relations hold,L

L= f 2

∞ (1.68)

and

1− f∞ + λf 2∞ = 0.7 (1.69)

To compute EE for this theory one has to start with a suitbale entropy functional. For this

case we have famous Jacobson-Myers functional (JM) [41].

SJM =2π

`3p

∫d3x√h[1 + λL2R

]. (1.70)

√h is the induce metric for the extremal surface and R is Ricci scalar of the extremal

surface. Now we will consider two different type of entangling region - sphere and cylinder.

We know that there are two types of anomaly coefficients in 3 + 1 dimensional CFT. For

generic higher curvature theories, unlike Einstein gravity, these two anomaly coefficients are

6Interested readers are referred to this thesis [37]7We choose the particular root which smoothly goes to one when λ→ 0.


numerically different from each other. From Eq. (1.45) it can be shown that, for a cylindrical

entangling surface the first term of Eq. (1.45) vanishes and one picks out the contribution of

Weyl anomaly in the universal part of the EE. On the other hand if one chooses a spherical

entangling surface the second term in Eq. (1.45) vanishes and one picks out the contribution of

Euler anomaly in the universal part. So we will consider these two different types of entangling

surface to explore the nature of the universal terms in the context of the higher derivative

theories. Extremizing the JM functional we get the following results for the universal parts

[41].

ScylinderEE = −CH2R

ln(R

ε) , SsphereEE = −4A ln(

R

ε), (1.71)

where, C = π2L3

f3/2∞ `3p

(1 − 2f∞λ) and A = π2L3

f3/2∞ `3p

(1 − 6f∞λ) are respectively Weyl and Euler

anomaly. So JM functional produces the expected universal terms. For any arbitrary higher

curvature theories one can construct an entropy functional which produces the correct uni-

versal part. But the question remains whether one can derive them or not. We will try to

formulate such a derivation and that will teach us some important lessons about underlying

gravity theories.

So we have now introduced all the characters of the story. We are now ready to explore the

interplay between gravity and entanglement and hoping to uncover some interesting physics.

Summary of the thesis

Before ending this section let us summarize the key points of the thesis at this stage.

1. In Chapter 2 we will try to derive the proposed entropy functionals for general theories

of gravity using the generalized gravitational entropy method proposed in [42]. We

will show that it is possible to derive the entropy functional only for a certain class of

gravity theories and discuss the implications of that.

2. In Chapter 3 we will demonstrate how to compute the universal terms of EE using the

generalized entropy method.

3. Identifying EE with the generalized entropy opens up the possibility of connecting

EE with the black hole entropy. For general theories of gravity black hole entropy is

given by the Wald formula [43]. We will explore the possibility of connecting EE with

the Wald entropy in Chapter 4, thereby opening up the possibility of deriving EE by

Noether charge method.

4. In Chapter 5 we will see how put constraints on the couplings of the higher derivative

terms using extremal surfaces.

20 REFERENCES

5. In Chapter 6 we will took a small step towards understanding how bulk equation of

motion arises from EE, which will give us some intuition about the emergence of gravity

from field theory. We will use the concept of relative entropy to understand this, which

is nothing but the change of entanglement entropy between vacuum and excited states.

Using relative entropy we will demonstrate how one can constrain underlying gravity

theory.

6. After that we will discuss briefly how to code holographic RG flow using EE in the

Chapter 7 and end with summarizing the main results and some open questions.

References

[1] J. D. Bekenstein, “Black holes and entropy,” Phys. Rev. D 7 (1973) 2333.

[2] S. W. Hawking, “Black hole explosions,” Nature 248 (1974) 30. S. W. Hawking, “Par-

ticle Creation by Black Holes,” Commun. Math. Phys. 43 (1975) 199 [Commun. Math.

Phys. 46 (1976) 206].

[3] G. ’t Hooft, “On the Quantum Structure of a Black Hole,” Nucl. Phys. B 256 (1985)

727.

[4] L. Bombelli, R. K. Koul, J. Lee, and R. D. Sorkin. A quantum source of entropy for

black holes,. Phys.Rev D34 (1986).

[5] M. Srednicki, “Entropy and area,” Phys. Rev. Lett. 71 (1993) 666 [hep-th/9303048].

[6] C. G. Callan, Jr. and F. Wilczek, “On geometric entropy,” Phys. Lett. B 333 (1994) 55

[hep-th/9401072]. C. Holzhey, F. Larsen and F. Wilczek, “Geometric and renormalized

entropy in conformal field theory,” Nucl. Phys. B 424 (1994) 443 [hep-th/9403108].

[7] L. Susskind and J. Uglum, “Black hole entropy in canonical quantum gravity and

superstring theory,” Phys. Rev. D 50 (1994) 2700 [hep-th/9401070].

[8] P. Calabrese and J. L. Cardy, “Entanglement entropy and quantum field theory,”

J. Stat. Mech. 0406, P06002 (2004) [hep-th/0405152].

P. Calabrese and J. Cardy, “Entanglement entropy and conformal field theory,” J. Phys.

A 42 (2009) 504005 [arXiv:0905.4013 [cond-mat.stat-mech]].

[9] There are huge number of papers. Interested readers may consult the following ones,

H. Casini, C. D. Fosco and M. Huerta, “Entanglement and alpha entropies for a massive

Dirac field in two dimensions,” J. Stat. Mech. 0507 (2005) P07007 [cond-mat/0505563].

REFERENCES 21

H. Casini and M. Huerta, “Entanglement and alpha entropies for a massive scalar field

in two dimensions,” J. Stat. Mech. 0512 (2005) P12012 [cond-mat/0511014].

H. Casini and M. Huerta, “Analytic results on the geometric entropy for free fields,”

J. Stat. Mech. 0801 (2008) P01012 [arXiv:0707.1300 [hep-th]].

H. Casini and M. Huerta, “Reduced density matrix and internal dynamics for multi-

component regions,” Class. Quant. Grav. 26 (2009) 185005 [arXiv:0903.5284 [hep-th]].

H. Casini and M. Huerta, “Entanglement entropy for a Maxwell field: Numeri-

cal calculation on a two dimensional lattice,” Phys. Rev. D 90 (2014) 10, 105013

[arXiv:1406.2991 [hep-th]].

H. Casini, M. Huerta and J. A. Rosabal, “Remarks on entanglement entropy for gauge

fields,” Phys. Rev. D 89 (2014) 8, 085012 [arXiv:1312.1183 [hep-th]].

W. Donnelly, “Decomposition of entanglement entropy in lattice gauge theory,” Phys.

Rev. D 85 (2012) 085004 [arXiv:1109.0036 [hep-th]].

W. Donnelly, “Entanglement entropy and nonabelian gauge symmetry,” Class. Quant.

Grav. 31 (2014) 21, 214003 [arXiv:1406.7304 [hep-th]].

L. Y. Hung and Y. Wan, “Revisiting Entanglement Entropy of Lattice Gauge Theo-

ries,” JHEP 1504 (2015) 122 [arXiv:1501.04389 [hep-th]].

Amico Luigi, Fazio. Rosario, Osterloh Andreas and Vedral Vlatko, “ Entanglement in

many-body systems”, 10.1103/RevModPhys.80.517

and all the citations of these papers.

[10] For example, interested readers are referred to,

“Entanglement in Quantum Critical Phenomena”, Vidal G., Latorre, J. I, Rico E. and

Kitaev.A , 10.1103/PhysRevLett.90.227902,

‘Diverging Entanglement Length in Gapped Quantum Spin Systems”

Verstraete, F., Martın-Delgado, M. A. and Cirac, J. I., 10.1103/PhysRevLett.92.087201

and the references and citations of these papers.

[11] A. Kitaev and J. Preskill, “Topological entanglement entropy,” Phys. Rev. Lett. 96

(2006) 110404 [hep-th/0510092].

“Detecting Topological Order in a Ground State Wave Function”, Levin, Michael and

Wen, Xiao-Gang, 10.1103/PhysRevLett.96.110405.

[12] There are huge number of works. In quantum information,

“Efficient Scheme for Two-Atom Entanglement and Quantum Information Processing

in Cavity QED”, Zheng, Shi-Biao and Guo, Guang-Can, 10.1103/PhysRevLett.85.2392

H. Casini, “Entropy inequalities from reflection positivity,” J. Stat. Mech. 1008 (2010)

P08019 [arXiv:1004.4599 [quant-ph]].

M. B. Plenio and S. Virmani, “An Introduction to entanglement measures,” Quant.

22 REFERENCES

Inf. Comput. 7 (2007) 1 [quant-ph/0504163].

“ Entanglement Theory and the Second Law of Thermodynamics”, Fernando G.S.L.

Brandao and Martin B. Plenio Nature Physics 4, 873 (2008)

and the citations and references of these papers.

[13] S. Ryu and T. Takayanagi, “Aspects of Holographic Entanglement Entropy,” JHEP

0608 (2006) 045 [hep-th/0605073].

[14] T. Nishioka, S. Ryu and T. Takayanagi, “Holographic Entanglement Entropy: An

Overview,” J. Phys. A 42 (2009) 504008 [arXiv:0905.0932 [hep-th]].

T. Takayanagi, “Entanglement Entropy from a Holographic Viewpoint,” Class. Quant.

Grav. 29 (2012) 153001 [arXiv:1204.2450 [gr-qc]].

“ Entanglement renormalization”, Guifre Vidal, Phys. Rev. Lett. 99, 220405 (2007).

[15] B. Swingle, “Entanglement Renormalization and Holography,” Phys. Rev. D 86 (2012)

065007 [arXiv:0905.1317 [cond-mat.str-el]]. M. Nozaki, S. Ryu and T. Takayanagi,

“Holographic Geometry of Entanglement Renormalization in Quantum Field Theo-

ries,” JHEP 1210 (2012) 193 [arXiv:1208.3469 [hep-th]].

A. Mollabashi, M. Nozaki, S. Ryu and T. Takayanagi, “Holographic Geometry of

cMERA for Quantum Quenches and Finite Temperature,” JHEP 1403 (2014) 098

[arXiv:1311.6095 [hep-th]].

M. Miyaji and T. Takayanagi, “Surface/State Correspondence as a Generalized Holog-

raphy,” arXiv:1503.03542 [hep-th].

M. Miyaji, T. Numasawa, N. Shiba, T. Takayanagi and K. Watanabe, “cMERA as

Surface/State Correspondence in AdS/CFT,” arXiv:1506.01353 [hep-th].

N. Bao, C. Cao, S. M. Carroll, A. Chatwin-Davies, N. Hunter-Jones, J. Pollack

and G. N. Remmen, “Consistency Conditions for an AdS/MERA Correspondence,”

arXiv:1504.06632 [hep-th].

[16] V. Balasubramanian, B. D. Chowdhury, B. Czech, J. de Boer and M. P. Heller,

“Bulk curves from boundary data in holography,” Phys. Rev. D 89 (2014) 8, 086004

[arXiv:1310.4204 [hep-th]].

V. Balasubramanian, B. Czech, B. D. Chowdhury and J. de Boer, “The entropy of a

hole in spacetime,” JHEP 1310 (2013) 220 [arXiv:1305.0856 [hep-th]].

B. Czech, X. Dong and J. Sully, “Holographic Reconstruction of General Bulk Sur-

faces,” JHEP 1411 (2014) 015 [arXiv:1406.4889 [hep-th]].

M. Headrick, R. C. Myers and J. Wien, “Holographic Holes and Differential Entropy,”

JHEP 1410 (2014) 149 [arXiv:1408.4770 [hep-th]].

R. C. Myers, J. Rao and S. Sugishita, “Holographic Holes in Higher Dimensions,” JHEP

REFERENCES 23

1406 (2014) 044 [arXiv:1403.3416 [hep-th]].

B. Chen and J. Long, “Strong Subadditivity and Emergent Surface,” Phys. Rev. D 90

(2014) 6, 066012 [arXiv:1405.4684 [hep-th]].

B. Czech and L. Lamprou, “Holographic definition of points and distances,” Phys. Rev.

D 90 (2014) 10, 106005 [arXiv:1409.4473 [hep-th]].

B. Czech, P. Hayden, N. Lashkari and B. Swingle, “The Information Theoretic Inter-

pretation of the Length of a Curve,” arXiv:1410.1540 [hep-th].

V. E. Hubeny, H. Maxfield, M. Rangamani and E. Tonni, “Holographic entanglement

plateaux,” JHEP 1308 (2013) 092 [arXiv:1306.4004, arXiv:1306.4004 [hep-th]].

[17] A. Almheiri, X. Dong and D. Harlow, “Bulk Locality and Quantum Error Correction

in AdS/CFT,” JHEP 1504 (2015) 163 [arXiv:1411.7041 [hep-th]].

E. Mintun, J. Polchinski and V. Rosenhaus, “Bulk-Boundary Duality, Gauge Invari-

ance, and Quantum Error Correction,” arXiv:1501.06577 [hep-th].

F. Pastawski, B. Yoshida, D. Harlow and J. Preskill, “Holographic quan-

tum error-correcting codes: Toy models for the bulk/boundary correspondence,”

arXiv:1503.06237 [hep-th].

[18] M. Rangamani and M. Rota, “Entanglement structures in qubit systems,”

arXiv:1505.03696 [hep-th].

E. Perlmutter, M. Rangamani and M. Rota, “Positivity, negativity, and entanglement,”

arXiv:1506.01679 [hep-th].

M. Kulaxizi, A. Parnachev and G. Policastro, “Conformal Blocks and Negativity at

Large Central Charge,” JHEP 1409 (2014) 010 [arXiv:1407.0324 [hep-th]].

M. Headrick, V. E. Hubeny, A. Lawrence and M. Rangamani, “Causality & holographic

entanglement entropy,” JHEP 1412 (2014) 162 [arXiv:1408.6300 [hep-th]].

M. Rangamani and M. Rota, “Comments on Entanglement Negativity in Holographic

Field Theories,” JHEP 1410 (2014) 60 [arXiv:1406.6989 [hep-th]].

S. A. Gentle and M. Rangamani, “Holographic entanglement and causal information

in coherent states,” JHEP 1401 (2014) 120 [arXiv:1311.0015 [hep-th]].

[19] D. D. Blanco, H. Casini, L. -Y. Hung and R. C. Myers, “Relative Entropy and Holog-

raphy,” JHEP 1308, 060 (2013) [arXiv:1305.3182 [hep-th]].

J. Bhattacharya, M. Nozaki, T. Takayanagi and T. Ugajin, “Thermodynamical Prop-

erty of Entanglement Entropy for Excited States,” Phys. Rev. Lett. 110, no. 9, 091602

(2013) [arXiv:1212.1164].

M. Nozaki, T. Numasawa, A. Prudenziati and T. Takayanagi, “Dynamics of En-

tanglement Entropy from Einstein Equation,” Phys. Rev. D 88, 026012 (2013)

[arXiv:1304.7100 [hep-th]].

24 REFERENCES

M. Nozaki, T. Numasawa and T. Takayanagi, “Holographic Local Quenches and En-

tanglement Density,” JHEP 1305, 080 (2013) [arXiv:1302.5703 [hep-th]].

D. Allahbakhshi, M. Alishahiha and A. Naseh, “Entanglement Thermodynamics,”

JHEP 1308, 102 (2013) [arXiv:1305.2728 [hep-th]].

G. Wong, I. Klich, L. A. Pando Zayas and D. Vaman, “Entanglement Temperature and

Entanglement Entropy of Excited States,” JHEP 1312, 020 (2013) [arXiv:1305.3291

[hep-th]].

P. Caputa, G. Mandal and R. Sinha, “Dynamical entanglement entropy with angular

momentum and U(1) charge,” JHEP 1311, 052 (2013) [arXiv:1306.4974 [hep-th]].

T. Faulkner, M. Guica, T. Hartman, R. C. Myers and M. Van Raamsdonk, “Gravitation

from Entanglement in Holographic CFTs,” arXiv:1312.7856 [hep-th].

[20] H. Casini, “Relative entropy and the Bekenstein bound,” Class. Quant. Grav. 25,

205021 (2008) [arXiv:0804.2182 [hep-th]].

R. Bousso, H. Casini, Z. Fisher and J. Maldacena, “Proof of a Quantum Bousso Bound,”

Phys. Rev. D 90 (2014) 4, 044002 [arXiv:1404.5635 [hep-th]].

R. Bousso, H. Casini, Z. Fisher and J. Maldacena, “Entropy on a null surface for

interacting quantum field theories and the Bousso bound,” Phys. Rev. D 91 (2015) 8,

084030 [arXiv:1406.4545 [hep-th]].

[21] N. Shiba and T. Takayanagi, “Volume Law for the Entanglement Entropy in Non-local

QFTs,” JHEP 1402 (2014) 033 [arXiv:1311.1643 [hep-th]].

[22] Unpublished notes of Prof. Rafael. Sorkin and his talk at RRI, Bangalore, India in

2014.

[23] Philippe Di Francesco, Pierre Mathieu, David Senechal,

“Conformal Field Theory”, Springer Science & Business Media, 1997.

[24] Robert Savit, “Duality in field theory and statistical systems,” Rev. Mod. Phys. 52,

453.

[25] J. M. Maldacena, “The Large N limit of superconformal field theories and supergravity,”

Int. J. Theor. Phys. 38 (1999) 1113 [Adv. Theor. Math. Phys. 2 (1998) 231] [hep-

th/9711200].

J. M. Maldacena, “The Large N limit of superconformal field theories and supergravity,”

Adv. Theor. Math. Phys. 2, 231 (1998) [hep-th/9711200].

E. Witten, “Anti-de Sitter space and holography,” Adv. Theor. Math. Phys. 2 (1998)

253 [hep-th/9802150].

S. S. Gubser, I. R. Klebanov and A. M. Polyakov, “A Semiclassical limit of the gauge

/ string correspondence,” Nucl. Phys. B 636 (2002) 99 [hep-th/0204051].

REFERENCES 25

[26] S. A. Hartnoll, “Lectures on holographic methods for condensed matter physics,” Class.

Quant. Grav. 26 (2009) 224002 [arXiv:0903.3246 [hep-th]].

J. McGreevy, “Holographic duality with a view toward many-body physics,” Adv. High

Energy Phys. 2010 (2010) 723105 [arXiv:0909.0518 [hep-th]].

C. P. Herzog, “Lectures on Holographic Superfluidity and Superconductivity,” J. Phys.

A 42 (2009) 343001 [arXiv:0904.1975 [hep-th]].

J. Erlich, “Recent Results in AdS/QCD,” PoS CONFINEMENT 8 (2008) 032

[arXiv:0812.4976 [hep-ph]].

J. Erdmenger, N. Evans, I. Kirsch and E. Threlfall, “Mesons in Gauge/Gravity Duals

- A Review,” Eur. Phys. J. A 35 (2008) 81 [arXiv:0711.4467 [hep-th]].

K. Peeters and M. Zamaklar, “The String/gauge theory correspondence in QCD,” Eur.

Phys. J. ST 152 (2007) 113 [arXiv:0708.1502 [hep-ph]].

J. Casalderrey-Solana, H. Liu, D. Mateos, K. Rajagopal and U. A. Wiedemann,

“Gauge/String Duality, Hot QCD and Heavy Ion Collisions,” arXiv:1101.0618 [hep-

th]

G. Policastro, D. T. Son and A. O. Starinets, “From AdS / CFT correspondence to

hydrodynamics,” JHEP 0209 (2002) 043 [hep-th/0205052].

and many more.

[27] A. Strominger, “The dS / CFT correspondence,” JHEP 0110 (2001) 034 [hep-

th/0106113]. V. Balasubramanian, J. de Boer and D. Minic, “Notes on de Sitter space

and holography,” Class. Quant. Grav. 19 (2002) 5655 [Annals Phys. 303 (2003) 59]

[hep-th/0207245].

D. T. Son, “Toward an AdS/cold atoms correspondence: A Geometric realization of

the Schrodinger symmetry,” Phys. Rev. D 78 (2008) 046003 [arXiv:0804.3972 [hep-th]].

M. Taylor, “Non-relativistic holography,” arXiv:0812.0530 [hep-th].

and many more.

[28] S. N. Solodukhin, “Entanglement entropy, conformal invariance and extrinsic geome-

try,” Phys. Lett. B 665 (2008) 305 [arXiv:0802.3117 [hep-th]].

[29] T. Faulkner, “Bulk Emergence and the RG Flow of Entanglement Entropy,” JHEP

1505 (2015) 033 [arXiv:1412.5648 [hep-th]].

[30] M. Van Raamsdonk, “Comments on quantum gravity and entanglement,”

arXiv:0907.2939 [hep-th].

[31] J. M. Maldacena, “Eternal black holes in anti-de Sitter,” JHEP 0304 (2003) 021 [hep-

th/0106112].

26 REFERENCES


AdS/CFT,” Phys. Rev. Lett. 96 (2006) 181602 [hep-th/0603001].

[33] J. Bhattacharya, M. Nozaki, T. Takayanagi and T. Ugajin, “Thermodynamical Prop-

erty of Entanglement Entropy for Excited States,” Phys. Rev. Lett. 110, 091602 (2013)

[arXiv:1212.1164 [hep-th]].

M. Nozaki, T. Numasawa, A. Prudenziati and T. Takayanagi, “Dynamics of Entangle-

ment Entropy from Einstein Equation,” arXiv:1304.7100 [hep-th].

G. Wong, I. Klich, L. A. P. Zayas and D. Vaman, “Entanglement Temperature and

Entanglement Entropy of Excited States,” arXiv:1305.3291 [hep-th].



[34] S. He, T. Numasawa, T. Takayanagi and K. Watanabe, “Quantum Dimension as En-

tanglement Entropy in 2D CFTs,” arXiv:1403.0702 [hep-th].

M. Nozaki, T. Numasawa and T. Takayanagi, “Quantum Entanglement of Local Op-

erators in Conformal Field Theories,” arXiv:1401.0539 [hep-th].

A. Mollabashi, N. Shiba and T. Takayanagi, “Entanglement between Two Interacting

CFTs and Generalized Holographic Entanglement Entropy,” arXiv:1403.1393 [hep-th].

P. Caputa, J. Simn, A. tikonas and T. Takayanagi, “Quantum Entanglement of Local-

ized Excited States at Finite Temperature,” JHEP 1501 (2015) 102 [arXiv:1410.2287

[hep-th]].

P. Caputa, J. Simn, A. tikonas, T. Takayanagi and K. Watanabe, “Scrambling time

from local perturbations of the eternal BTZ black hole,” arXiv:1503.08161 [hep-th].

[35] V. E. Hubeny, M. Rangamani and T. Takayanagi, “A Covariant holographic entangle-

ment entropy proposal,” JHEP 0707, 062 (2007) [arXiv:0705.0016 [hep-th]].

A. C. Wall, “Maximin Surfaces, and the Strong Subadditivity of the Covariant

Holographic Entanglement Entropy,” Class. Quant. Grav. 31 (2014) 22, 225007

[arXiv:1211.3494 [hep-th]].

[36] T. Hirata and T. Takayanagi, “AdS/CFT and strong subadditivity of entanglement

entropy,” JHEP 0702 (2007) 042 [hep-th/0608213].

M. Headrick and T. Takayanagi, “A Holographic proof of the strong subadditivity of

entanglement entropy,” Phys. Rev. D 76 (2007) 106013 [arXiv:0704.3719 [hep-th]].

H. Casini and M. Huerta, “On the RG running of the entanglement entropy of a circle,”


[37] “Various Aspects of Holographic Entanglement Entropy and Mutual Information” by

Jarkko Jarvela, Master Thesis, University of Helesinki, Department of Physics.

REFERENCES 27

[38] D. J. Gross and E. Witten, “Superstring Modifications of Einstein’s Equations,” Nucl.

Phys. B 277, 1 (1986).

M. B. Green and C. Stahn, “D3-branes on the Coulomb branch and instantons,” JHEP

0309, 052 (2003) [hep-th/0308061].

M. F. Paulos, “Higher derivative terms including the Ramond-Ramond five-form,”

JHEP 0810, 047 (2008) [arXiv:0804.0763 [hep-th]].

S. S. Gubser, I. R. Klebanov and A. A. Tseytlin, “Coupling constant dependence in

the thermodynamics of N=4 supersymmetric Yang-Mills theory,” Nucl. Phys. B 534,

202 (1998) [hep-th/9805156].

A. Buchel, J. T. Liu and A. O. Starinets, “Coupling constant dependence of the shear

viscosity in N=4 supersymmetric Yang-Mills theory,” Nucl. Phys. B 707, 56 (2005)

[hep-th/0406264].

R. C. Myers, M. F. Paulos and A. Sinha, “Quantum corrections to eta/s,” Phys. Rev.

D 79, 041901 (2009) [arXiv:0806.2156 [hep-th]].

A. Buchel, R. C. Myers, M. F. Paulos and A. Sinha, “Universal holographic hydrody-

namics at finite coupling,” Phys. Lett. B 669 (2008) 364 [arXiv:0808.1837 [hep-th]].

W. H. Baron and M. Schvellinger, “Quantum corrections to dynamical holographic ther-

malization: entanglement entropy and other non-local observables,” arXiv:1305.2237

[hep-th].

[39] X. O. Camanho, J. D. Edelstein and J. M. S. de Santos, “Lovelock theory and the

AdS/CFT correspondence,” arXiv:1309.6483 [hep-th].

[40] A. Buchel, J. Escobedo, R. C. Myers, M. F. Paulos, A. Sinha and M. Smolkin, “Holo-

graphic GB gravity in arbitrary dimensions,” JHEP 1003, 111 (2010) [arXiv:0911.4257

[hep-th]].

[41] L. -Y. Hung, R. C. Myers and M. Smolkin, “On Holographic Entanglement Entropy

and Higher Curvature Gravity,” JHEP 1104 (2011) 025 [arXiv:1101.5813 [hep-th]].

J. de Boer, M. Kulaxizi and A. Parnachev, “Holographic Entanglement Entropy in

Lovelock Gravities,” JHEP 1107, 109 (2011) [arXiv:1101.5781 [hep-th]].

A. Bhattacharyya, M. Sharma and A. Sinha, “On generalized gravitational entropy,


[42] A. Lewkowycz and J. Maldacena, “Generalized gravitational entropy,” JHEP 1308

(2013) 090 [arXiv:1304.4926 [hep-th]].

[43] R. M. Wald, “Black hole entropy is the Noether charge,” Phys. Rev. D 48, 3427 (1993)

[gr-qc/9307038].

V. Iyer and R. M. Wald, “A Comparison of Noether charge and Euclidean methods

28 REFERENCES


[gr-qc/9503052].

V. Iyer and R. M. Wald, “Some properties of Noether charge and a proposal for dy-

namical black hole entropy,” Phys. Rev. D 50 (1994) 846 [gr-qc/9403028].

2 Holographic entanglement entropy

functionals: A derivation

2.1 Introduction

It is has been observed that there exists a striking similarity between black hole entropy

and entanglement entropy [1]. In the context of AdS/CFT, the entanglement entropy1 for

a boundary field theory which is dual to Einstein gravity can be calculated using the well-

known Ryu-Takayanagi proposal (RT) [18, 19]. This proposal states that the entanglement

entropy SEE of any region on the boundary of AdS can be calculated by evaluating the area

of a minimal surface in the bulk which is homologous to this boundary region:

SEE =Area

4GN

. (2.1)

Building upon earlier attempts [20, 21, 22, 23], this proposal was recently proved in [24], for

a general entangling surface. The entanglement entropy formula in Eq. (2.1) is of the same

form as the formula for calculating the entropy of a black hole. In the black hole case, there

exists a simple generalization of this area law for calculating the entropy of a black hole in

any general higher-derivative gravity theory, known as the Wald entropy [25, 26, 27]. It is

natural to ask then if one can generalize the Ryu-Takayanagi prescription to higher-derivative

gravity theories by simply replacing the RHS of Eq. (2.1) with the Wald entropy. However,

this is known not to be the case [28, 29].

Recently, a general formula for calculating the holographic entanglement entropy (HEE)

in higher-derivative gravity theories was proposed in [30, 31]. It was also conjectured that

the minimal entangling surface can be determined by interpreting this formula as the entropy

functional for the higher derivative gravity theory and extremizing it. At present there exists

no general proof of this proposal. Main objective of this chapter is to carry out various tests

to determine the validity of this conjecture and formulate a general proof.

Lewkowycz and Maldacena (LM) [24] have proposed a derivation of the Ryu-Takayangi

(RT) prescription [18] for computing entanglement entropy (EE) [2] in holography [19]. A

1There exists a huge literature on entanglement entropy. For background and interesting applications see

[2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17].

29


generalization of black hole entropy is proposed in the context where there is no U(1) sym-

metry in the bulk. In the Euclidean theory, although there is no U(1) symmetry, one imposes

a periodicity condition of 2πn with n being an integer on the Euclidean time direction at

the boundary. This time direction shrinks to zero at the boundary. By suitably choosing

boundary conditions on the fields, LM propose to identify the on-shell Euclidean action with

a generalized gravitational entropy.

In calculations of entanglement entropy in quantum field theories, one frequently uses

the replica trick which entails introducing a conical singularity in the theory 2. An earlier

attempt to prove the RT formula was made by Fursaev [20]. In recent times, in the context

of AdS3/CFT2 there have been further developments in [8] towards a proof. In the context

of holography, this corresponds to taking the n→ 1 limit. In this case, LM suggest that the

time direction shrinks to zero on a special surface. The equation for this surface is derived

in Einstein gravity by showing that there is no singularity in the bulk equations of motion.

This surface has vanishing trace of the extrinsic curvature and corresponds to a minimal

surface–which is precisely what comes from minimizing the RT area functional. Next we will

try to generalize this for various higher derivative gravity theories.

We will first work with general four-derivative theory. It is also sufficient for our purpose

to consider only four-derivative theory as that will capture all the essential issues that we

like to bring up. The conjectured form of the holographic entropy functional for general

R2 theory first appeared in [33]. The formula of [30, 31] also reduces to this functional for

general R2 theory. For the purpose of this thesis, we will refer to this functional as the FPS

(Fursaev-Patrushev-Solodukhin) functional after the authors of the paper where it was first

proposed. In [34] it was shown that this entropy functional leads to the expected universal

terms in the entanglement entropy for cylindrical and spherical entangling surfaces, so the

FPS functional passes this basic first test. The obvious next step is to determine whether

the surface equation of motion derived from extremizing this functional is the same as that

derived using the generalized gravitational entropy method (which we will refer to as the LM

method) of [24].

General R2 theory depends on three parameters: λ1, λ2 and λ3. Gauss-Bonnet gravity is

a special point in this parameter space [35] and the FPS functional reduces to the Jacobson-

Myers functional at this point. For Gauss-Bonnet gravity, the question whether the surface

equation of motion one gets from the Jacobson-Myers functional matches with the surface

equation of motion derived using the LM method was addressed in [34, 36, 37]. We will look

2The only example where a derivation of EE exists without using the replica trick is for the spherical

entangling surface [4, 22] although in [32] it has been explained how this procedure is connected with the

replica trick. A proposal has been made in [7] for the equation for the entangling surface which does not

depend on the replica trick.

CHAPTER 2. HOLOGRAPHIC ENTANGLEMENT ENTROPY FUNCTIONALS: ADERIVATION 31

at the Gauss-Bonnet case first to emphasize several interesting points for this theory. For

this theory, the surface equation of motion that one gets from the Jacobson-Myers functional

matches with what one gets from the LM method, provided that terms cubic in the extrinsic

curvature are suppressed. We will find that for general R2 theory using a procedure similar to

the Gauss-Bonnet case leads to a match in the leading-order terms on both sides, where we

designate terms cubic in the extrinsic curvature as sub-leading. However, as we will show, in

the case of R2 theory, the LM method also yields an extra condition that cannot be satisfied

at arbitrary points of the parameter space. The conclusion is, therefore, that for a general

R2 theory the conditions that follow from the LM method do not correspond exactly to the

surface equation of motion derived from the FPS functional.

An alternative method to demonstrate that the FPS functional is the correct entropy

functional for R2 theory is to show that it can be interpreted as the action of a cosmic brane.

This method was employed in [30], where it was referred to as the cosmic brane method. We

will re-examine this procedure for R2 theory and show that the result we get is consistent

with what we get using the LM method.

What happens when we go to a six-derivative gravity theory? In this case, we consider

quasi-topological gravity [38] which is again a special point in the parameter space of R3

theories. We first construct the entropy functional for quasi-topological gravity using the

formula proposed in [30, 31]. We then show that this functional reproduces the expected

universal terms for this theory for the cylindrical and spherical entangling surfaces. This is

in agreement with the result of [39] that the entropy functional proposed in [30, 31] leads to

the correct universal terms for a general higher-derivative theory. We also find the minimal

surface condition for this theory using the LM method and show that it deviates from what

is expected from the HEE functional.

This chapter is organized as follows. In Sec. (2.2) we review the general entropy functional

proposed in [30, 31]. Our main focus in this chapter is general four-derivative gravity theory,

for which the entropy functional is the FPS functional. In Sec. (2.3) we find the surface

equation of motion for R2 theory by extremizing the FPS functional on the codimension-2

surface. We then compare it with what we obtain using LM prescription. We also make

some remarks on the Gauss-Bonnet case. We then investigate the cosmic-brane method of

[30]. In Sec. (2.4), we repeat our analysis for quasi-topological gravity. Lastly, in Sec. (2.5)

we summarize our findings and discuss their implications.

32 2.2. ENTROPY FUNCTIONAL FOR GENERAL THEORIES OF GRAVITY

2.2 Entropy functional for general theories of gravity

In this section we will review the general entropy formula proposed in [30, 31]. First we

summarize the argument leading up to this proposal, following [30]. For details the reader is

referred to [24, 30, 31, 33]. Some applications of this entropy formula are in [40].

In field theory, the entanglement entropy SEE = −Tr[ρ log ρ] can be calculated as the

n→ 1 limit of the Renyi entropy. The Renyi entropy in turn can be computed as

Sn = − 1

n− 1(logZn − n logZ1) . (2.2)

Here Zn is the partition function of the field theory on the manifold Mn which is the n-

fold cover of the original spacetime manifold M1. In the holographic dual theory one can

construct a suitable bulk solution Bn with boundary Mn. The manifold Mn at integer n has

a Zn symmetry that cyclically permutes the n replicas. In [24] it was proposed that this

replica symmetry extends to the bulk Bn. Orbifolding the bulk by this symmetry results in

a space Bn = Bn/Zn , that is regular except at the fixed points of the Zn action. The fixed

points form a codimension 2 surface with a conical defect in the bulk. This is the surface

that is ultimately identified with the minimal entangling surface in the n→ 1 limit. Further,

one can use gauge-gravity duality [41] to identify the field theory partition function on Mn

with the on-shell bulk action on Bn in the large-N limit

Zn ≡ Z[Mn] = e−S[Bn] . (2.3)

It is now straightforward to calculate the entanglement entropy. By construction, one can

identify

S[Bn] = nS[Bn] (2.4)

at integer n, where S[Bn] is the classical action for the bulk configuration Bn not including

any contribution from the conical defect. By analytically continuing Bn to non-integer n,

Eq. (2.4) can be used to define S[Bn]. Using Eqs. (2.2) and (2.3) and expanding around

n = 1, one gets

SEE = limn→1

n

n− 1

(S[Bn]− S[B1]

)= ∂εS[Bn]

∣∣∣ε=0

(2.5)

where ε ≡ n−1. The quantity S[Bn] can be calculated for the bulk theory by writing the bulk

metric locally around the surface in gaussian normal coordinates and introducing a conical

defect. It can be shown [24, 30] that ∂εS[Bn] gets a contribution entirely from the tip of the

cone. To compute it, therefore, one employs a metric regularized at the tip of the cone.

This calculation is similar to that employed in [42] for calculating the Wald entropy

from a regularized cone metric. Indeed, evaluating S[Bn] for a bulk theory with the cone

metric to linear order in ε and using Eq. (2.5) will result in two terms. The first is SWald:


the Wald entropy for the theory. However, as was noted in [33], there is a second way for

a linear contribution to arise. A term in the bulk lagrangian that is of order ε2 can get

enhanced to order ε after integrating over the transverse directions. Following [30], we label

the contribution of such terms as SAnomaly. At this point, the calculation of the form of

SEE is basically finished. Eq. (2.5) can be used to find the entanglement entropy for any

higher-derivative theory, including ones whose lagrangians involve derivatives of the Riemann

tensor. However, for a general higher-derivative theory it can be computationally difficult to

compute SAnomaly directly using Eq. (2.5).

In [30, 31] a simpler prescription for calculating the holographic entanglement entropy

for higher-derivative theories of gravity in d + 1 dimensions, for which the lagrangian (L)

contains only contractions of the Riemann tensor, was given. The formula is:

SEE =2π

`dp

∫ddy√h

∂L

∂Rzzzz

+∑α

(∂2L

∂Rzizj∂Rzkzl

)α

8KzijKzklqα + 1

. (2.6)

The notation used in the above equation and also in the rest of the chapter is as follows:

We use Greek Letters µ, ν, ρ, σ, · · · to label the bulk indices. We use the Latin letters

a, b, ......m, n to label the indices of the codimension 2 surface, while reserving the letters

p, q, r, s to denote the indices of the transverse directions. In these directions, we use the com-

plex coordinates z and z. The bulk metric is denoted by gµν .The metric on the codimension-2

entangling surface is denoted by hij while the surface itself is denoted by Σ. The bulk Rie-

mann tensor is denoted by Rµνρσ while the intrinsic Riemann tensor of the surface is denoted

by Rikjl. The extrinsic curvatures of the surface are denoted by Krij, where the first index la-

bels the extrinsic curvature in the transverse directions. We follow the curvature conventions

in [43].

The first term in Eq. (2.6) is the Wald entropy. The second term is the correction to the

Wald entropy and is evaluated in the following way: The second derivative of the lagrangian Lis a polynomial in components of the Riemann tensor. We expand the components Rpqij, Rpiqj

and Rikjl using

Rpqij = Rpqij +KpjkKkqi −KpikKkqj ,Rpiqj = Rpiqj +KpjkKkqi −Qpqij ,Rikjl = Rikjl +KpilKpjk −KpijKpkl . (2.7)

Here, Qpqij ≡ 12∂p∂qgij|Σ. Rpqij and Rpiqj can also be defined in terms of metric variables,

but the exact definition is not needed here. The variable α is used to label the terms in the

expansion. For each term labelled by α, qα is defined as the number of Qzzij and Qzzij, plus

one half of the number of Kpij, Rpqri, and Rpijk. The final step is to sum over α with weights

34 2.3. TEST OF THE ENTROPY FUNCTIONAL FOR R2 THEORY

1/(1 + qα). The quantities Rpqij, Rpiqj, and Rikjl can then be eliminated using Eq. (2.7),

resulting in an expression involving only components of Rµνρσ, Kpij and Qpqij.To yield the entanglement entropy, the formula in Eq. (2.6) should be evaluated on the

minimal entangling surface. This surface is supposed to be determined following the LM

method. [30, 31] also contain the proposal that the minimal surface can be determined

by extremizing SEE as given in Eq. (2.6) — SEE therefore being the entanglement entropy

functional for a general theory of gravity. Rest of this chapter is mainly based on the work

done with Prof. Aninda Sinha and Dr. Menika Sharma [34, 44].

2.3 Test of the entropy functional for R2 theory

In this section we consider general R2 theory in five dimensions. The lagrangian for this

theory is

L = L1 + L2 , (2.8)

where

L1 = R +12

L2(2.9)

is the usual Einstein-Hilbert lagrangian with the cosmological constant appropriate for five-

dimensional AdS space and

L2 =L2

2

(λ1RαβγδR

αβγδ + λ2RαβRαβ + λ3R

2)

(2.10)

is the R2 lagrangian.

The proposed entropy functional for this theory is

SEE,R2 = SWald,R2 + SAnomaly,R2 , (2.11)

where

SWald,R2 =2π

`3P

∫d3x√h

1 + L2

2(2λ3R + λ2Rµνn

νrn

rµ + 2λ1Rµνρσnµrn

νsn

rρnsσ),

(2.12)

and SAnomaly,R2 =2π

`3P

∫d3x√hL2

2

(− 1

2λ2KrKr − 2λ1KsijKsij

). (2.13)

As mentioned earlier, this entropy functional leads to the correct universal terms. To demon-

strate this, we will write down first the bulk AdS metric as,

ds2 =L2

z2(dz2 + dτ 2 + hijdx

idxj) (2.14)


where, L is the AdS radius and hij is a three dimensional metric given below. We will use

Greek letters for the bulk indices and Latin letters for the three dimensional indices. For the

calculation of EE for a spherical entangling surface we write the boundary hij in spherical

polar coordinates as,spherehijdx

idxj = dρ2 + ρ2dΩ22 , (2.15)

where dΩ22 = dθ2 + sin2 θdφ2 is the metric of a unit two-sphere and θ ∈ [0, π] and φ ∈ [0, 2π].

For a cylindrical entangling surface,

cylinderhijdxidxj = du2 + dρ2 + ρ2dφ2 . (2.16)

u is the coordinate along the direction of the length of the cylinder. For a cylinder of length

H, u ∈ [0, H]. Here L = L√f∞.

We put ρ = f(z), τ = 0 in the metric and minimize (2.11) on this codimension 2 surface

and find the Euler-Lagrange equation for f(z). Using the solution for f(z) we evaluate (2.11)

to get the EE.

For the sphere, we get f(z) =√f 2

0 − z2 which gives the universal log term,

SEE = −4a ln(f0

δ) . (2.17)

For the cylinder, f(z) = f0 − z2

4f0+ ... which gives,

SEE = −cH2R

ln(f0

δ) . (2.18)

a =π2L3

f3/2∞ `3

P

(1− 2f∞(λ1 + 2λ2 + 10λ3)) and c =π2L3

f3/2∞ `3

P

(1 + 2f∞(λ1 − 2λ2 − 10λ3)) . (2.19)

and δ is the UV cut-off comes from the lower limit of the z integral. f0 is the radius of the

entangling surface. These are the expected results [28, 35].

In this section, we will further test this entropy functional by determining whether the

surface equation of motion one gets from extremizing this functional is the same as the

surface equation of motion one gets following the LM method. In Sec. (2.3.1), we extremize

the functional for R2 theory. In this particular section, we will first find the surface equation of

motion for this functional in a general spacetime background. However, the Ryu-Takayanagi

proposal and its generalizations are most precisely formulated in the AdS/CFT context, so

eventually we will specialize to the AdS background. In Sec. (2.3.2) we find the surface

equation of motion using the LM method. In this case, we will always assume that the bulk

is AdS space. Since a variation of the LM method – called the cosmic-brane method – was

used in [30] to formulate a proof that the FPS functional is the correct entropy functional

for R2 theory, we also investigate this method in Sec. (2.3.3).


2.3.1 Minimal surface condition from the entropy functional

To extremize the functional in Eq. (2.11), we follow the methods of [45, 46, 47]. We denote

the surface we are going to extremize w.r.t the action in Eq. (2.11) by Σ. The induced metric

on Σ is

hij = eµi eνj gµν , (2.20)

where gµν is the bulk metric and eµi ≡ ∂iXµ are the basis vectors tangent to the surface

Σ, Xµ being the bulk coordinates. On the surface Σ, the gir component of the bulk metric

vanishes. The two normals to the surface are denoted by nµr where r = 1, 2 are the transverse

directions. The metric tensor in the tangent space spanned by the normal vectors (the metric

tensor of the normal bundle of the sub-manifold Σ) is the Kronecker delta:

δrs = ε nµrnνsgµν (2.21)

We work in Euclidean signature and set ε = +1. We use the inverse metric δrs, to raise

indices in the normal directions: nrµ = δrsnµs . Note that, repeated s indices always imply

summation over the transverse directions: nµsnνs ≡ nµ1n

ν1 + nµ2n

ν2. In this notation, the

completeness relation relating gµν , the inverse of the bulk metric, to hij, the inverse of the

induced metric, is

gµν = hijeµi eνj + nµsn

νs. (2.22)

The Gauss and Weingarten equations are

∇ieµj = ∂ie

µj + Γµνρe

ρi eνj − Γkije

µk = −Krijnµr

∇inµs = ∂in

µs + Γµρνe

ρin

νs − Γrsin

µr = Kjsie

µj . (2.23)

Here, ∇ is the Van der Waerden-Bortolotti covariant derivative [45] which acts on a general

tensor T s···ri···j as

∇kTs···ri···j = ∇kT

s···ri···j + ΓspkT

p···ri···j + · · ·+ ΓrpkT

s···pi···j , (2.24)

where ∇ is the usual covariant derivative associated with the surface Christoffel. This

Christoffel is related to the bulk Christoffel Γµσν as

Γijk = (∂jeµk + Γµσνe

σj eνk)e

iµ . (2.25)

The Chrisoffel Γris is the Christoffel induced in the normal bundle. It is related to the bulk

Christoffel as

Γris = (∂inµs + Γµσνe

σi n

νs)n

rµ . (2.26)

This Christoffel can be interpreted geometrically as the freedom to perform rotations of the

normal frame. It is, therefore, equivalent to a gauge field Ak, commonly referred to as a twist

potential. This field is defined as:

Ak ≡1

2εrs∂rgks , so that Γsjr = δpsεrpAj , (2.27)


where εrs is the Levi-Civita symbol.

The Gauss identity relating the bulk Riemann with all indices projected in the tangential

directions with the surface Riemann is

Rµνρσeµkeνi eρl eσj = Rkilj −KrklKrij +KrikKrjl . (2.28)

The Codazzi-Mainardi relation is

∇kKrij −∇iKrkj = Rµνρσeµkeνi eσj n

ρr . (2.29)

From Eq. (2.28) we get the Gauss-Codazzi identity

R = R− 2Rµνnνrnµr +Rµνρσn

µrnνsnρrnσs +KrKr −KsijKijs . (2.30)

We now consider an infinitesimal variation of the surface Σ given by Xµ −→ Xµ + δXµ.

The change δXµ is

δXµ = ξrnµr + ξieµi . (2.31)

where ξr and ξi are small parameters. For deriving the equation describing the minimal

surface we are only concerned with the variation in the normal direction, since the tangential

variation leads to a constraint equation. The variation then reduces to

δXµ = ξrnµr , (2.32)

The variation δXµ in the surface will induce a variation in the basis vectors eµi . This can

be computed by finding the basis vectors at Xµ + δXµ and parallel transporting them back

to Xµ. Taking the difference between the parallel-transported quantity and the original basis

vector at the coordinate Xµ, using the identities in Eq. (2.23) and then restricting to normal

variation results in

δeµi = nµs∇iξs + eµjK

jsiξ

s . (2.33)

The details of this calculation are in [45]. As stated in Eq. (2.24), the covariant derivative ∇acts on ξs as

∇iξs = ∂iξ

s + Γsirξr . (2.34)

The variation in any other tensor quantity can be calculated in a similar way, by parallel

transporting the quantity at the new coordinate back to the old coordinate and taking the

difference. This gives the variation in the bulk metric as zero. We write down the result

for other variations. For details the reader is referred to [45]. The variation of the induced

metric is

δhij = 2ξrKrij ,δ√h = ξr

√hKr . (2.35)


The variation of the extrinsic curvature is

δKsij = (−∇i∇jδsr +KsikKkrj +Rµνρσn

sµnσr eρi eνj )ξ

r ,

δKs = (−∇i∇iδsr −KsikKkir + hijRµνρσn

sµnσr eρi eνj )ξ

r . (2.36)

The covariant derivatives ∇ act all the way to the right.

Using these variations we can now compute the change in the action. For this we first

rewrite the Rµνnνrn

rµ and Rµνρσnµrn

νsn

rρnsσ terms in the action given in Eq. (2.12) as

Rµνnνrnµr = R− hijRµνe

νi eµj

Rµνρσnµrnνsnρrn

σs = R− 2hijRµνe

νi eµj + hikhjlRµνρσe

µi eνj eρkeσl (2.37)

using the completeness relation in Eq. (2.22). The variation of a term such as hijRµνeνi eµj is

given by

δ(hijRµνeνi eµj ) = (δhij)Rµνe

νi eµj + 2hijRµνδ(e

νi )e

µj + hijδ(Rµν)e

νi eµj (2.38)

The first two variations can be computed using Eqs. (2.35) and (2.33) respectively. For

evaluating the last term we need the variation of the bulk Ricci Tensor which is given by

δ(Rµν) = nσr ∇σRµνξr . (2.39)

The variation in the bulk Ricci scalar and Riemann tensor take a similar form. All these

variations are under the integral sign in Eq. (2.12) and we perform a integration by parts

where needed, discarding the term that is a total variation. Then using the variations given

above we obtain 3:

δ(√hR) =

√h KsRξs + nµs

√h∇µRξ

s ,

δ(√hRµνn

νrnµr ) =√h KsRµνn

νrnµr ξs + 2

√h∇i(Rµνn

νseµi )ξs −

√hnσsh

ijeµi eνj ∇σRµνξ

s + nµs√h ∇µRξ

s ,

δ(√hRµνρσn

µrnνsnρrnσs ) =√h KsRµνρσn

µrnνqnρrnσq ξ

s − 4√h∇i(Rµνρσn

µs eνj eρi eσkh

jk)ξs+

4√h∇i(Rµνn

νseµi )ξs +

√hhikhjleµi e

νj eρkeσl n

αs ∇αRµνρσξ

s −2nνsh

ijeµi eρj∇νRµρξ

s +√hnµs ∇µRξ

s . (2.40)

Similarly the variations for the terms present in the action in Eq. (2.13) are

δ(√hKsKs) =−2

√h∇i∇iKrξr +

√hKrKsKsξr − 2

√hKsKsijKijr ξr−

2√hKsRµνρσh

ijnµr eνi n

ρseσj ξ

r ,

δ(√hKsijKsij) =−2

√h∇i∇jKijr ξr +

√hKrKsijKsijξr − 2

√hKsijKsik Kkjr ξr−

2√hKsijRµνρσn

µr eνi n

ρseσj ξ

r . (2.41)

3We thank Joan Camps for valuable discussions


Adding these variations up with the appropriate factors will give us the equation for the

minimal surface for the action in Eq. (2.11) in a general spacetime background.

As a check of these equations we now demonstrate that the above results lead to the

correct surface equation of motion in the Gauss-Bonnet case. For Gauss-Bonnet the entropy

functional for general R2 theory reduces to the Jacobson-Myers functional

SJM =2π

`3P

∫d3x√h(1+λL2(R−2Rµνn

νrnµr +Rµνρσnµrnνsnρrn

σs +KsKs−KsijKsij) . (2.42)

This functional is valid in a general space-time background. Note that the integrand is equal

to√h(1 + λL2R) on using the Gauss-Codazzi identity Eq. (2.30). The surface equation of

motion for this theory using this form of the functional is [34],

K + λL2(RK− 2RijKij) = 0 . (2.43)

We now find the surface equation of motion by directly varying Eq. (2.42). Using the variation

equations Eqs. (2.40–2.41) and simplifying using the identities in Eqs. (2.28–2.29) we get

√h Ks ξs + λL2

[√h KsR ξs − 2

√hRjkKjks ξs

+√hhikhjleµi e

νj eρkeσl n

αs ∇αRµνρσξ

s − 2√h∇i(Rµνρσn

µs eνj eρi eσkh

kj)ξs

− 2√hKrRµνρσh

ijnµs eνi n

ρreσj ξ

s + 2√hKrijRµνρσn

µs eνi n

ρreσj ξ

s

+ 2√hRµνρσe

µj eνi eρkeσl h

ilKjks ξs]. (2.44)

The first three terms give precisely the equation of motion for Gauss-Bonnet theory. The

rest of the terms add up to zero, as we show in the following. We use the Bianchi identity

on the ∇αRµνρσ factor of the fourth term giving

∇αRµνρσ = −∇σRµναρ − ∇ρRµνσα (2.45)

and then rewrite each of these terms as

hikhjleµi eνj eρkeσl n

αs ∇σRµναρ = eσl ∇σ(hikhjleµi e

νj eρkn

αsRµναρ)− eσl ∇σ(hikhjleµi e

νj eρkn

αs )Rµναρ .

(2.46)

The expression in brackets in the first term of the R.H.S is a bulk scalar and therefore this

term can be written as

∂l(hikhjleµi e

νj eρkn

αsRµναρ) = −∇i(Rµνρσn

µs eνj eρi eσkh

jk) + Γrslhikhjleµi e

νj eρkn

αrRµναρ+

Γmjlhikhjleµi e

νme

ρkn

αsRµναρ , (2.47)

Inserting these expressions in Eq. (2.44) after expanding the second term on the R.H.S of

Eq. (2.46) and using the identities in Eq. (2.23) will lead to cancellation of all terms except

for the terms in the first line of Eq. (2.44).


AdS background

We now specialize to the case of AdS background which is the background we will use while

finding the equation of motion using the LM method. In AdS space the Riemann tensor is

Rµνρσ = −C(gµρgνσ − gµσgνρ) , (2.48)

where we have defined C = f∞/L2. Here, L is the length scale associated with the cosmo-

logical constant and is related to the AdS radius L as L = L√f∞. The variable f∞ satisfies

the following equation for R2 theory

1− f∞ +1

3f 2∞(λ1 + 2λ2 + 10λ3) = 0 . (2.49)

For ease of comparison with later results, we also rewrite the variation in√hR given in

Eq. (2.40) using the Gauss-Codazzi relation Eq. (2.30). The minimal surface equation is then

Kr + L2λ3(RKr − 2RijKrij + 2∇2Kr − 2∇i∇jKrij−KrK2 + 2KrijK

ij2 +KrK2 − 2Kr3 − 18CKr)+

λ2(12∇2Kr − 1

4KrK2 + 1

2KrijK

ij2 − 11

2CKr)+

λ1(2∇i∇jKrij −KrK2 + 2Kr3 − 4CKr) = 0 , (2.50)

where we have defined K2 = KsijKsij,Kij2 = KsKsij, K2 = KsKs and Kr3 = KsijKsik Krkj. Note

that these are a set of two equations one for each of the extrinsic curvatures K1,K2.

In AdS space we can make a further simplification using Eq. (2.29). The R.H.S of this

equation disappears on using Eq. (2.48). We then get the identity ∇k∇kKr = ∇i∇jKrij on

taking a further covariant derivative of the L.H.S. As explained in Appendix 2.5, in the LM

method for a time-independent metric, we can set K1 = K2 = K. We, therefore, also drop

the r index and Eq. (2.50) simplifies to

K + L2λ3(RK− 2RijKij −K3 + 3KK2 − 2K3 − 18CK)+

λ2(12∇2K − 1

4K3 + 1

2KK2 − 11

2CK)+

λ1(2∇2K −KK2 + 2K3 − 4CK) = 0 . (2.51)

We have also verified this equation by determining the bulk extremal surfaces for different

types of boundary entangling regions (sphere, cylinder and slab).

For the Gauss-Bonnet case: λ1 = λ, λ2 = −4λ and λ3 = λ, this equation reduces to the

known result in Eq. (2.43). Note that terms cubic in the extrinsic curvature as well as the

CK terms are not present in that equation. The Gauss-Bonnet case is special in this sense.


No such simplification occurs if we set the value for Weyl2 theory, λ1 = λ, λ2 = −4λ/3 and

λ3 = λ/6:

K +λL2

6(RK− 2RijKij + 8∇2K +K3 − 7KK2 + 10K3 + 2CK) = 0 . (2.52)

The CK term, in particular, stands out. If we trace the provenance of this term, it comes

from terms of the form KsRµνρσnµrn

νqn

ρrnσq and Kijs Rµνρσnµr eνi n

ρseσj in Eqs. (2.40) and (2.41)

— terms that have components normal to the surface. Nevertheless, for AdS background

these reduce to a term intrinsic to the surface. In fact, using the Gauss-Codazzi identity,

Eq. (2.30), we can rewrite this CK term as ∼ K3 +RK.

So far, we have only considered normal variations of the surface. Considering tangential

variations leads to a constraint equation. For R2 theory this constraint equation is indistin-

guishable from the condition in Eq. (2.29) which is the Codazzi-Mainardi relation.

2.3.2 Minimal surface condition from the Lewkowycz-Maldacena

method

We will now derive the surface equation for R2 using the LM method. As already mentioned,

the main idea of Ref. [24] is that one can obtain the minimal surface condition by extending

the replica trick to the bulk. The bulk will then have a Zn symmetry. Orbifolding by this

symmetry will lead to a spacetime in which the fixed points form a codimension-2 surface

with a conical deficit. In the n → 1 limit this surface can be identified with the entangling

surface. The metric of this surface can be parametrized in gaussian normal coordinates as

follows:

ds2 = e2ρ(z,z)dzdz + e2ρ(z,z)Ω(zdz − zdz)2+ (gij +Krijxr +Qrsijxrxs)dyidyj +

2e2ρ(z,z)(Ai + Brixr)(zdz − zdz)dyi + · · · . (2.53)

Here ρ(z, z) = − ε2

ln(zz) and ε = n−1, while x1 = z and x2 = z. This is the most general form

of the metric upto terms second order in z(z) [30, 31, 33, 48]. The · · · denote higher-order

terms. As we will see later, for R2 theory we also need to include third-order terms in the

metric expansion. The quantity Kij in this metric is identified with the extrinsic curvature,

while Ai is identified with the twist potential. Both of these are standard quantities that

characterize the embedding of the surface in the bulk. The quantities Ω,Bri and Q in the

second-order terms in the metric are not arbitrary, but can be written in terms of Krij,Aiand the components of the curvature tensors.

The bulk equation of motion will now contain divergences arising out of the conical

singularity of the form εz, εz2

. However, the matter stress-energy tensor is expected to be


finite. Therefore, we must set all divergences to zero. This condition fixes the location of the

entangling surface.

The bulk equation of motion for general four-derivative theory is [49]:

Rαβ −1

2gαβR−

6

L2gαβ −

L2

2Hαβ = 0 , (2.54)

where

Hαβ = λ1

(1

2gαβRδσµνR

δσµν − 2RασδµRβσδµ − 4∇2Rαβ + 2∇α∇βR +

4RδαRβδ + 4RδσRδαβσ

)+

λ2

(∇α∇βR + 2RδσRδαβσ − ∇2Rαβ +

1

2gαβRδσR

δσ − 1

2gαβ∇2R

)+

λ3

(− 2RRαβ + 2∇α∇βR +

1

2gαβR

2 − 2gαβ∇2R). (2.55)

Gauss-Bonnet theory

Our eventual goal is to find the surface equation of motion for general R2 theory, but it is

illuminating to look at the Gauss-Bonnet case first. The Gauss-Bonnet case was addressed

in [34, 36, 37] using a metric linear in z(z). In this section, we will find the surface equation

of motion for this theory using the metric in Eq. (2.53).

We first show that the second-order metric in Eq. (2.53) suffices for Gauss-Bonnet theory

and inclusion of higher-order terms in this conical metric will not affect the surface equation

of motion that we find for this theory from the LM method. The bulk equation of motion

for Gauss-Bonnet theory can be obtained from Eq. (2.54) by setting λ1 = λ, λ2 = −4λ and

λ3 = λ giving:

Hαβ = 4RδαRβδ − 4RδσRδαβσ − 2RRαβ − 2RασδµRβ

σδµ+

12gαβ(RδσµνR

δσµν − 4RδσRδσ +R2) . (2.56)

The surface equation of motion is derived by finding the divergences in this equation that

arise on using the conical metric in the limit z = z → 0. Terms higher than second-

order in the metric will not contribute to the curvature tensors to zeroeth-order in z(z),

although they might contribute at higher order. This is because the curvature tensors are of

dimension 1/Length2 while third-order terms in the metric will be of order 1/Length3. The

explicit values of the curvature tensors are listed in Appendix (2.5). These are calculated

using a conical metric which is third-order in z(z). Note also, that the curvature tensors

contain at most divergences of order 1/z. In the above equation of motion all terms are the

product of two curvature tensors. Since each curvature tensor can only contribute at most a


1/z divergence and no third(or higher)-order term occurs at zeroeth order in any curvature

tensor, third(and higher)-order terms will be absent in the divergence equations.

By the same logic one can see that second-order terms will contribute to the divergence

equations. However, in this case, cancellations between terms remove most of the second-

order quantities, leaving only the quantities Qzzij and Qzzij in the divergence equations.

In the z = z → 0 limit K1 = K2 as explained in Appendix 2.5, so we drop the index r on

Kr. The divergence in the zz component from Hαβ term in the bulk eom is

Hzz =ε

z

[λ(RK− 2KijRij)

]+ε

z

[e−2ρ(z,z)λ

−K3 + 3KK2 − 2K3

]. (2.57)

Setting this divergence to zero should yield the condition for the extremal surface. There is

no divergence in the zz component. The divergence in the zi component is

Hzi =ε

z

[e−2ρ(z,z)λ

2K∇iK − 2K∇jKji − 2Kji∇jK + 2Kij∇kKkj −

2Kkj∇iKkj + 2Kjk∇jKki]

. (2.58)

This divergence is equivalent to the constraint equation one gets for the entropy functional

(which doesn’t have to be necessarily the Jacobson-Myers functional) using tangential vari-

ations of the surface and vanishes similarly by Eq. (2.29). Finally, the divergence in the ij

component is

Hij =4ε

z

[e−4ρ(z,z)λ

2KikKklKlj + hijKK2 −KijK2 − hijK3 −KKikKkj − 4hijKQzz

+ 4hijKklQklzz − 8KkiQkzzj + 4KijQzz + 4KQzzij]

+

2ε2

z2

[e−4ρ(z,z)λ

2KijK − 2KikKkj − hijK2 + hijK2

]. (2.59)

In the above equation we have set Qzzij = Qzzij. Using the value of the Rzizj component

of the Riemann tensor from Appendix 2.5 and setting the background to AdS space, using

Eq. (2.48), we can show thatQzzij = 14KikKkj and as a result the ε

zdivergence exactly vanishes.

However the ε2

z2divergence will remain.

The final step is to take the ε, z → 0 limit. Depending on the ordering one chooses, there

are two ways to do this. One way is to take z → 0 limit first. Physically, this corresponds to

looking for a divergence in the bulk equation of motion while there is still a small but non-

zero conical deficit parameter ε. The second way is take ε→ 0 first. The limit is, therefore,

an iterated limit – the final result depends on the order in which the limit is taken, so there

is an inherent ambiguity in this procedure. In fact, this ambiguity can be made even larger

in scope if we take the limit simultaneously in ε and z. Mathematically, the divergence is a

function of the two variables: ε and z. In this ε-z plane there are an infinite number of paths


along which we can take the limit. At least on a mathematical level, there exists no reason

why the limit should only be taken along the z = 0 or ε = 0 path.

The path z = 0 is, however, the simplest way to take the limit so as to obtain εz→∞. In

this case, all terms containing εz

are leading divergences while terms containing e−2ρ(z,z)ε/z =

(zz)εε/z contribute to sub-leading divergences. Therefore, in this way of taking limits, setting

the Hzz divergence to zero will yield two different conditions for the minimal surface. The

first condition which, after adding the Einstein term, corresponds to the surface equation of

motion is

K + L2λ(RK− 2KijRij) = 0 . (2.60)

This agrees with the surface equation that comes from the Jacobson-Myers functional. How-

ever, there will also be an extra constraint [6] of the form

−K3 + 3KK2 − 2K3 = 0 , (2.61)

coming from the sub-leading divergence. The Hij divergence will also lead to a similar

constraint. The above condition can only be true for very special surfaces and therefore is an

over-constraint on the surface. In fact, if these two conditions were to be true simultaneously,

the surface equation of motion we would end up getting is:

cK + αλL2(K3 − 3KK2 + 2K3) = 0. (2.62)

To get this form of the equation, we have used the Gauss identities on AdS space. Here,

c = (1 − 2f∞λ) is proportional to the Weyl anomaly and α is a variable that can take any

arbitrary numerical value. The surface equation of motion corresponding to the Jacobson-

Myers functional can be recovered if α = 1. However, at present nothing within the LM

method sets the value of this parameter to one. Note that if α was zero, the minimal surface

that we would get is the same as in the Einstein case. It also the minimal surface that would

follow if one were minimizing just the Wald part of the entropy functional.

In the above paragraph we outlined one way in which the LM method could potentially

give rise to the correct surface equation of motion. Let us now explore if we can change

the limit-taking procedure itself to get the correct equation. This can be accomplished by

choosing a different path in the ε-z plane to take the limit. Taking the limit along the

path ε = 0 will simply kill off all divergences; this is not surprising since physically this

corresponds to turning off the conical deficit in the metric. However, we can pick a path in

the ε, z plane that will kill off the sub-leading divergence but preserve the leading divergence.

For example, as was shown in [36], taking any path of the form (z)2ε = ( zε)1+v, where v is a

number greater than one, will keep only the leading divergence. At this point, we can offer

no justification of why one should choose this particular way of taking limits. We are merely


demonstrating that there does exist a way to take limits in the ε, z plane that leads to the

correct surface equation of motion in the Gauss-Bonnet case. This way of taking limits is

equivalent to discarding terms suppressed by e−2ρ(z,z) and was also used in [30] to show that

the LM method leads to the same surface equation of motion in the Lovelock case as can

be derived from the entropy functional for Lovelock gravity [27, 30, 50]. This is the way of

taking limits that we will use. However, unless one can specify a mechanism or a physical

interpretation which reproduces this way of taking limits (which is possible if the metric itself

is re-defined), the argument that the LM method reproduces the correct surface equation of

motion for Gauss-Bonnet theory remains incomplete.

The same ambiguity in taking limits exists for general R2 theory. To remain consistent

with the Gauss-Bonnet point, for R2 theory we will continue to take limits as stated in the

paragraph above. However, in the general case this is not an ideal solution. As we will see,

∼ K3 terms always occur with the e−2ρ(z,z) factor in the divergence equations for R2 theory.

This means that if we use the above way of taking limits we will never get such terms at any

point in the parameter space. As we saw in Eq. (2.51), the surface equation of motion for R2

theory does contain such terms. However, our goal for general R2 theory is to see to what

extent we are able to reproduce the surface equation of motion in Eq. (2.51), while taking

the limit in such a manner that the result at the Gauss-Bonnet point agrees with what comes

from the Jacobson-Myers functional. It is clear, though, that the question of taking limits in

the LM method deserves more study.

The general case

We now work out the divergence equations for the R2 case. For general R2 theory all second-

order quantities will enter into the divergences. We can anticipate the effect that terms

containing Ω and B will have on the surface equation of motion coming from the LM method.

Consider the following components of the bulk Riemann tensor around z = z = 0:

Rpqrs

∣∣∣(z=0,z=0)

= −3e4ρ(z,z)εpqεrsΩ ,

Rpqri

∣∣∣(z=0,z=0)

= 3e2ρ(z,z)εpqBri , (2.63)

where, εab is defined as εzz = −εzz = e−2ρ(z,z)gzz. The quantity Ω is therefore equivalent

to −1/3Rµνρσnµrn

νsn

ρrn

σs evaluated at z, z = 0. We can determine a numerical value for the

quantities Bri and Ω in the metric by demanding that the bulk Riemann tensor be the AdS

solution at zeroeth order. Since for AdS space the Riemann tensor is given by Eq. (2.48), we

can write the components of the bulk Riemann tensor on the L.H.S of Eq. (2.63) in terms of

the components of the bulk metric. Expanding the metric using Eq. (2.53) and keeping only


the zeroeth-order terms in z, z we get

Ω = − 1

12C and Bri = 0 (2.64)

Therefore, Bri can be set to zero. In writing the divergences, we also ignore4 Qzzij and Qzz

ij ,

the remaining component being Qij = Qzzij = Qzz

ij .

For R2 theory the derivative order of the equation of motion is four. That means we

should include order z3 terms in the metric, since they can contribute to the divergences.

These terms can be parametrized as

ds2 = e4ρ(z,z)∆pqrstxpxqxrdxsdxt +Wrspijx

rxsxpdyidyj + 2e2ρ(z,z)Crsixrxs(zdz − zdz)dyi .

(2.65)

This is the most general form of the third-order terms in the metric. Here, we have written

the e2ρ(z,z) dependence of each term explicitly. As for the second-order quantities, the third-

order quantities ∆pqrst,Wrspij and Crsi can be found by calculating the curvature tensors, but

to linear order in z(z). Then, for example, e4ρ(z,z)∆pqrst ≡ −1/6∂p(Rµνρσnµqn

νrn

ρsn

σt ) evaluated

at z = z = 0. Note that the factor of e4ρ(z,z) will cancel from both sides on using the AdS

background. In fact, this particular term vanishes altogether in this background. On using

the metric with the third-order terms listed above to find the divergences in the equation

of motion we find that the Crsi, Wzzzij and Wzzzij do not contribute. The terms that are

relevant are Wzzzij and Wzzzij, because as will show below they will lead to unsuppressed

CK terms. Without loss of generality, we can set them to be equal and denote this term as

Wij.

4As we saw for the Gauss-Bonnet case, these terms will be present in the divergences, but they will not

change our conclusions for R2 theory.


For general R2 theory, the divergence in the zz component from the Hαβ term in the bulk

eom is

Hzz =ε

z

[− 1

2(λ2 + 4λ3)∇2K + (2λ1 + λ2 + 2λ3)∇i∇jKij + λ3(RK− 2KijRij) +

4(−2λ1 + 3λ2 + 14λ3)KijAiAj − 6(λ2 + 4λ3)KAiAi +

8(3λ1 + 2λ2 + 5λ3)KΩ]−

ε

z2

[e−2ρ(z,z)

(2λ1 − λ2 − 6λ3)K2 + 1

2(λ2 + 4λ3)K2 + 2(λ2 + 4λ3)Q

]+

ε

z

[e−2ρ(z,z)

− λ3K3 + (λ2 + 7λ3)KK2 − 2(3λ1 + 2λ2 + 6λ3)K3 +

(6λ1 + 5λ2 + 14λ3)KijQij − 32(λ2 + 4λ3)KQ −

4(λ2 + 4λ3)W]

. (2.66)

The divergences in the other components are

Hzz =2ε

z

[e−2ρ(z,z)

(λ3 + 1

4λ2)K3 + (λ1 − 3

2λ2 − 7λ3)KK2 +

2(−2λ1 + λ2 + 6λ3)K3 + (2λ1 − 3λ2 − 14λ3)KklQkl +

32(λ2 + 4λ3)KQ+ 8(λ2 + 4λ3)W

], (2.67)

Hzi =2ε

z

[e−2ρ(z,z)

− 1

2(2λ1 + λ2)Kki∇kK − (3λ1 − λ2 − 6λ3)Kkl∇iKkl −

14(3λ2 + 8λ3)K∇iK + (5λ1 + λ2)Kki∇lKlk − λ1K∇kKki +

(9λ1 + 2λ2)Kkj∇kKji − (λ2 + 4λ3)∇iQ− (4λ1 + λ2)∇kQki −(10λ1 − 2λ2 − 18λ3)AiK2 − 1

2(3λ2 + 12λ3)AiK2 +

8(4λ1 + λ2)KijKjkAk − 2(λ2 + 4λ3)AiQ]

, (2.68)

Hij =4ε

z

[e−4ρ(z,z)

(1

3λ1 + 1

4λ2 + 2

3λ3)hijK3 − (7λ1 + 2λ2 + 2λ3)KKikKkj +

2(16λ1 + 4λ2 + λ3)KikKklKlj − (λ1 + 3λ2 + 10λ3)hijKK2 −(3λ1 − 2λ3)KijK2 − 1

3(λ1 − 18λ2 − 70λ3)hijK3 +

2(4λ1 + λ2)QijK + 2(λ2 + 4λ3)hijKQ− 8(4λ1 + λ2)KikQkj −(λ2 + 4λ3)KijQ− 7(λ2 + 4λ3)hijKklQkl + 32(4λ1 + λ2)Wij+

32(λ2 + 4λ3)hijW]

. (2.69)

Whether or not the divergences in the ij, zi and zz components vanish before taking the

ε→ 0 limit will depend upon the exact values of the second-order terms. The zi divergence,


in particular, should be equivalent to the constraint equation coming from the tangential

variations and should vanish by the Codazzi relation in Eq. (2.29). As in the Gauss-Bonnet

case, the divergence in the ij component is not expected to fully vanish by itself. We therefore

take the limit as prescribed in the last section. This reduces the divergences in the zi and ij

components to zero. However, because of the presence of the W term there still remains an

unsuppressed divergence in the zz component. This divergence can only go to zero if K = 0

or the theory is at the Gauss-Bonnet point.

We now examine the divergences in the zz component, to be able to compare it with

the surface equation of motion derived using the FPS functional. First looking at the 1/z

divergence in that component, one can see that it contains the unsuppressed terms KijAiAj

and KAiAi which are not present in Eq. (2.51). However, these terms can be eliminated

in favor of other variables. Consider the Rzizj component of the Riemann tensor expanded

around z = 0, z = 0:

Rzizj

∣∣∣(z=0,z=0)

= 12e2ρ(z,z)Fij − 2e2ρ(z,z)AiAj + 1

4KzikKkzj − 1

2Qzzij . (2.70)

Using Eq. (2.48) again and multiplying both sides by Kij, we find that the AiAjKij term

can be written as ∼ CK + e−2ρ(z,z)K3 + e−2ρ(z,z)QK. The AiAiK terms can be written in a

similar fashion. Since only the CK term is unsuppressed we find

AiAjKij =CK4

+ · · · and

AiAiK =3CK

4+ · · · , (2.71)

where the dots denote the suppressed terms.

Next looking at the e−2ρ(z,z)/z divergence we find that the W term will contribute to

the surface equation of motion, since this term contains a e2ρ(z,z) factor that enhances the

divergence to 1/z. This term can be determined by using the following equation

∂zRzz

∣∣∣(z=0,z=0)

= −W + 2e2ρ(z,z)KijAiAj − 2e2ρ(z,z)ΩK + · · · . (2.72)

The R.H.S of this equation disappears in the AdS background. Using Eqs. (2.64) and (2.71)

we find

W =2e2ρ(z,z)CK

3+ · · · . (2.73)

The Q terms that are also present in this divergence do not contribute since as we show

below they are expected to contain only ∼ K2 terms and therefore remain suppressed.

Substituting these values in Eq. (2.66), and adding the Einstein term we find that the

1/z divergence of the zz component gives rise to the following surface equation of motion:

K + L2(2λ1 + 12λ2)∇2K + λ3(RK− 2KijRij) + λ1C1K + λ2C2K + λ3C3K = 0 (2.74)


where C1 = −4C, C2 = −11C/2 and C3 = −18C. The coefficients of the ∇2K,RK, KijRij

and CK terms in the above equation all match with those in Eq. (2.51). Because of the way

we are taking limits, the K3 terms that are present in Eq. (2.51) are not present here.

Finally we look at the ε/z2 divergence in the zz component. For this divergence to vanish,

we get the condition

(2λ1 − λ2 − 6λ3)K2 + 12(λ2 + 4λ3)K2 + 2(λ2 + 4λ3)Q = 0 . (2.75)

To satisfy this condition at arbitrary points of the parameter space, one has to demand thatQbe a function of ∼ K2 terms, and also λ1, λ2 and λ3. Demanding that Q be independent of λ1,

λ2 and λ3, will pick out a special point in the parameter space (apart from the Gauss-Bonnet

point where this condition is trivially satisfied).

To summarize the results for R2 theory:

1. Apart from the absence of ∼ K3 terms, Eq. (2.74) that we found using the LM method

is exactly the surface equation of motion that results from the FPS functional.

2. There are some problematic extra divergences. The zz component of the bulk equation

of motion has a divergence that can only disappear at the Gauss-Bonnet point. There

is also a second-order 1/z2 divergence in the zz component. This can be taken to fix the

value of the term Q; however, it is not possible to do this in a way that is independent

of the parameters of R2 theory.

2.3.3 The stress-energy tensor from the brane interpretation

In [24], it was noted that a equation of motion of a cosmic string is the same as the equation

for the minimal entangling surface. This is because a cosmic string produces a spacetime

with a conical defect with a metric of the form in Eq. (2.53). The equation of motion is

given by minimizing its action. For Einstein gravity this is just the Nambu-Goto action and

equation of motion of a cosmic string is

K = 0. (2.76)

This condition minimizes the surface area of the string as it sweeps through spacetime. The

same thing holds for a cosmic brane.

As was done in [30], where it was referred to as the cosmic brane method, this fact

can be exploited to construct the entropy functional from the bulk equation of motion. In

this section, we will check this construction of [30]. The idea is that the bulk equation of

motion in Eq. (2.54) should lead to the cosmic brane as a solution, to linear order in ε. In


particular, this means that L.H.S of Eq. (2.54) should be equal to the stress-energy tensor of

the brane. Since the brane is a localized source, the stress-energy tensor will contain delta

functions. Once we have found the stress-energy tensor we can identify the associated action

via Tαβ = δSδgαβ

.

Let us see how this works in the Gauss-Bonnet case. In the bulk equation of motion,

terms such as ∂z∂zρ(z, z) correspond to delta functions. We set δ(z, z) = e−2ρ(z,z)∂z∂zρ(z, z).

Note that δ(z, z) defined this way contains a factor of ε.

The delta divergences in the ij component of the bulk equation of motion to linear order

in ε are then:

Tij = δ(z, z)− 4λ (hijR− 2Rij)+

− 2λ e−2ρ(z,z)(hijK2 − hijK2 + 2KijK − 2KikKkj ). (2.77)

To identify this as the stress-energy tensor coming from the Jacobson-Myers functional (in-

terpreted as a cosmic brane action), the second term should go to zero. This term carries

a factor of e−2ρ(z,z) as compared to the first term and according to our way of taking limits

is suppressed. Our result is then in agreement with the claim in [30] that the cosmic-brane

method can be used to show that the Jacobson-Myers functional is the right entropy func-

tional for Gauss-Bonnet theory. However, as we will see there are problems for the general

four-derivative theory.

For R2 theory, the delta divergences in the ij component are


Tij = δ(z, z)[− 4λ3 (hijR− 2Rij)− 16(6λ1 + 11λ2 + 38λ3)hijΩ +

e−2ρ(z,z)− (λ2 + 2λ3)hijK2 + 2(λ2 + 3λ3)hijK2 −

2(12λ1 + 4λ2 + 4λ3)Qij − 2(λ2 + 4λ3)Qhij −2(4λ1 + λ2 + 2λ3)KijK + 2(14λ1 + 4λ2 + 4λ3)KkjKkj ) +

16(20λ1 + 11λ2 + 24λ3)AiAj ]

+

e−2ρ(z,z)∂zδ(z, z) + ∂zδ(z, z)

− 2(2λ1 + λ2 + 2λ3)Kij + (4λ3 + λ2)hijK

−

4e−2ρ(z,z) ∂z∂zδ(z, z)(λ2 + 4λ3) . (2.78)

Again, barring the term suppressed by e−2ρ(z,z), we have checked that the result for this

component is of the same form as that produced on calculating the stress-energy tensor from

an action equivalent to the FPS functional. The derivative of delta terms like ∂zδ(z, z) are

typical in the stress-energy tensor of actions containing terms that depend on the extrinsic

curvature [47]. However, the zz and zz components of the bulk equation of motion also

contain delta divergences that are not suppressed:

Tzz = − 4∂2zδ(z, z)(2λ1 + λ2 + 2λ3)− 2∂zδ(z, z)(4λ1 + λ2)K (2.79)

and

Tzz = − 2∂zδ(z, z) + ∂zδ(z, z)

(2λ1 + λ2 + 2λ3

)K +

4 ∂z∂zδ(z, z)(2λ1 + λ2 + 2λ3) . (2.80)

Taking the delta divergences in all components into account, the Tµν we have found does not

look like the stress-energy tensor for a cosmic brane corresponding to a three-dimensional

surface in the five-dimensional bulk. Note that the extra divergences all vanish for the Gauss-

Bonnet theory. The Gauss-Bonnet result therefore stands. However, any attempt to use this

method to show that the FPS functional is the correct entropy functional for R2 theory

should be able to account for these extra delta divergences.

2.4 Quasi-topological gravity

The lagrangian for quasi-topological gravity [38] contains terms cubic in the Riemann tensor.

It can be used to study a class of CFT’s involving three parameters in four dimensions. It

has many interesting features including the fact that its linearized equation of motion is

two-derivative order. Unitarity for this theory was studied in [51].

52 2.4. QUASI-TOPOLOGICAL GRAVITY

In Sec. (2.4.1), we find the HEE functional for quasi-topological theory using Eq. (2.6)

and compute the universal terms is Sec. (2.4.2). In Sec. (2.4.3), we find the surface equation

of motion for this theory using the LM method.

2.4.1 The entropy functional

The action for quasi-topological theory in five dimensions is

SQT = − 1

2`3P

∫d5x(L1 + L2 + ν Z5

), (2.81)

where L1 is the Einstein-Hilbert action given in Eq. (3.9) and L2 is the Gauss-Bonnet la-

grangian as in Eq. (3.10) with λ1 = λ3 = λ , λ2 = −4λ. The last term is the R3 lagrangian:

Z5 =µ0RαβγδRγδ

µνRµναβ + µ1Rα

βγδRβ

ηδζRα

ηγζ + µ2RαβγδR

αβγδR +

µ3RαβγδRαβγ

ηRδη + µ4RαβγδR

αγRβδ + µ5RαβRβ

γRγα + µ6R

βα R

αβ R + µ7R

3 . (2.82)

There are two different consistent R3 theories. For the first theory

µ0 = 0 , µ1 = 1 , µ2 = 38, µ3 = −9

7, µ4 = 15

7, µ5 = 18

7, µ6 = −33

14, µ7 = 15

56(2.83)

and the coupling constant is ν = 7µL4

4, while for the second theory

µ0 = 1 , µ1 = 0 , µ2 = 32, µ3 = −60

7, µ4 = 72

7, µ5 = 64

7, µ6 = −54

14, µ7 = 11

14(2.84)

and the coupling constant ν = 7µL4

8.

The R3 part for the HEE functional is

SEE,R3 =2πν

`3P

∫d3x√h (LWald,R3 + LAnomaly,R3) , (2.85)

where

LWald,R3 = 6µ0RzzαβRzzαβ + 3µ1

(Rzαz

βRzαzβ −Rzαz

βRzαzβ)

+ µ2

(RαβρσR

αβρσ −4RRzz

zz

)+ 2µ3

(Rα

zzzRα

z −RαzzzRα

z + 12Rαβρ

zRαβρz

)+ µ4(2Rz

αzβRαβ +

(Rzz)

2 −RzzRzz) + 3µ5RzαRzα + µ6

(RαβR

αβ + 2RRzz

)+ 3µ7R

2 . (2.86)

The symbols z and z in the above expression label the two orthogonal directions while the

indices α, β, ... are the usual bulk indices. The expression for the anomaly part is

LAnomaly,R3 =µ0(12K2ijQij − 6K4)− µ1(3

2K4 − 3

2K2

2 + 3KijKklRikjl)−µ2(6K2

2 − 2K2K2 − 8K2Q+ 4K2R)−µ3(2K4 + 1

2K2

2 −K2Q− 2K2ijQij − 2KijQij K + 2K2

ijRij)−µ4(2K3K −K2K2 − 2KijQij K + 2KijRij K)−µ5(3

4K2K2 − 3

2K2Q)− µ6(3

2K2K2 − 1

2K4 − 2K2Q+K2R) . (2.87)


where K4 = KijKjlKlkKki . In calculating the anomaly part from Eq. (2.6), we have used the

value of Bi = 0 that we found in Sec. (2.3.2). This is the reason that while terms involving

Bi are supposed to contribute to Eq. (2.6), the above equation does not contain any terms

containing Bi. The full HEE functional has contributions from the Einstein and R2 part also

which are given in Eqs.(2.12) and (2.13).

2.4.2 Universal terms

In this section, we will demonstrate that our HEE functional for the quasi-topological gravity

produces the correct universal terms. For the general structure and calculation of the uni-

versal term of the entanglement entropy in four dimensions, see [52]. These central charges

can be easily calculated using the technique of [53].

We follow the procedure given in [34] for R2 theory. Here we sketch the main steps of this

calculation. We will minimize Eq. (2.85) for a bulk surface with a spherical and cylindrical

boundary. We will carry out this procedure for the five-dimensional bulk AdS metric as

shown in the section (2.3).

Following the analysis of section (2.3) we set ρ = f(z), τ = 0 in the metric in Eq. (2.14)

and minimize the entanglement entropy functional (whose R3 part is given in Eq. (2.85)) on

this codimension 2 surface to find the Euler-Lagrange equation for f(z). Using the solution

for f(z) we evaluate the entropy functional to get the EE.

For the spherical boundary, we get f(z) =√f 2

0 − z2 which gives the EE as

SEE = −4a ln(f0

δ) . (2.88)

Here, δ is the UV cut-off that comes from the lower limit of the z integral and f0 is the radius

of the entangling surface. The value of a is

a =π2L3

f3/2∞ `3

P

(1− 6f∞λ+ 9f 2∞µ) . (2.89)

For this case, the entire contribution comes from the Wald entropy as the extrinsic curvatures

are identically zero.

For the cylindrical boundary, we find f(z) = f0 − z2

4f0+ ... leading to

SEE = −cH2R

ln(f0

δ) . (2.90)

The value of c corresponding to the theory in Eq. (2.83) is

c =π2L3

f3/2∞ `3

P

1− 2f∞λ+ 9f 2

∞µ+ f 2∞µ(42µ1 − 336µ2 − 56µ3)

, (2.91)


while that corresponding to the theory in Eq. (2.84) is

c =π2L3

f3/2∞ `3

P

1− 2f∞λ+ 9f 2

∞µ− f 2∞µ(168µ2 + 28µ3)

. (2.92)

The 1 + 9f 2∞µ part is the usual Wald entropy contribution, while the remaining part comes

from the anomaly part. After putting in the values of µ’s given in Eqs.(2.83) and (2.84) we

obtain

c =π2L3

f3/2∞ `3

P

(1− 2f∞λ− 3f 2∞µ) (2.93)

for both theories.

These results for the universal terms agree with those calculated in [34] for the two quasi-

topological theories. Note from Eqs. (2.91) and (2.92) that only a few terms from LAnomaly,R3

have contributed to the universal term. Terms of the form ∼ K4 do not contribute to this

calculation at all. Since Q ∼ K2, terms of the form ∼ K2Q also do not contribute.

2.4.3 Minimal surface condition

We now find the surface equation of motion for quasi-topological gravity using the LM

method. For ease of calculation, we set all second-order quantities and cross-components

in the metric in Eq. (2.53) to zero. The bulk equation of motion for this theory is [49]:

Rαβ −1

2gαβR−

6

L2gαβ −

L2

2Hαβ − νFαβ = 0 , (2.94)

where Fαβ is defined in [38, 49]. The εz

divergence in the zz component of the equation of

motion coming from the Fαβ term is

F 1zz =

ε

z

[(3

2µ1 − µ2 − µ3 − 3

2µ5 − 4µ6 − 12µ7)Rij∇2Kij − (1

2µ2 + µ6 + 6µ7)RijKijR +

(µ2 + 16µ6 + 3µ7)RijRijK + 1

2(µ6 + 1

2µ4)K∇2R+ (µ4 + µ3 + 4µ2)∇i∇iK −

(3µ1 − 8µ− 2− 3µ3 − µ4 + 32µ5)∇lRlijk∇kKij − 1

2(µ4 + 3µ1)Kkl∇i∇jRkijl −

(34µ1 − 5

2µ2 − µ3 − 1

2µ4 − 3

4µ5 − 2µ6 − 6µ7)R∇i∇jKij +

14(µ4 + 2µ3 + 8µ2)Kij∇i∇jR

]. (2.95)

While we haven’t computed the surface equation of motion that one gets on minimizing the

functional in Eq. (2.85), this is not very hard to do using the methods of Sec. (2.3.1) and

Mathematica5. The main point is, however, that the surface equation of motion that one

5We have used the Xact package for a number of calculations in this chapter


will get from the entropy functional will contain K4 terms that are absent in the above

divergence.

Other divergences are also present in the zz component:

F 2zz =

ε

z2

[e−2ρ(z,z)

12(3µ1 + 19µ2 + 2µ3 + 14

3µ6)RijklKikKjl − (7

2µ6 + 18µ7)RK2 +

(2µ2 + 32µ6 + 6µ7)RK2 + (µ4 − 3

2µ5)K∇2K −

(43µ2 − µ4 + 3

2µ5 + 4

3µ6)Kij∇i∇jK − (µ4 − 3

2µ5)K∇i∇jKij +

(3µ1 − 23µ2 − µ3 + µ4 − 3

2µ5 − 2

3µ6)∇kKij∇kKij −

(3µ1 + 23µ2 + µ3 + 3µ4 − 3

2µ5 + 2

3µ6)∇iKij∇kKjk −

(8µ2 − µ4 + 32µ5)RijKijK + (3µ1 + 8µ2 − µ4 − 2µ6)RijKjkKik

]+

ε

z3

[e−4ρ(z,z)

(3µ1 − 2µ2 − 2µ3)K3 + (µ4 − 3µ5 − 2µ6)KK2

]. (2.96)

As for R2 theory, these divergences can be used to determine second and higher-order terms

in the metric. At linear-order in the metric, divergences in all other components of the

equation of motion go to zero if we take the limit as mentioned in Sec. (2.3.2).

2.5 Discussion

We found the surface equation of motion for general R2 theory and quasi-topological gravity

using the generalized gravitational entropy method of [24]. We found that these do not match

exactly with what can be derived by extremizing the HEE functional for these theories – the

HEE functional being calculated using the formula proposed in [30, 31].

Let us summarize our findings regarding R2 theory. First, the leading-order terms on both

sides do match. In fact, barring ∼ K3 terms, the surface equation of motion that follows

from the LM method is precisely the surface equation of motion that follows from the FPS

functional.

The main problem with the LM method is that there are divergences in components other

than the zz component, for a general higher-derivative theory. In the Gauss-Bonnet case,

there are ways we can take the limit to set these divergences to zero. However, the effect of

taking the limit in this way is to remove all ∼ K3 divergences from all components of the

equation of motion. This means that we do not get any ∼ K3 terms in the surface equation

of motion using the LM method. No matter what the HEE functional for R2 theory is, it

is unlikely that no ∼ K3 terms will occur in its surface equation of motion at any point

in its parameter space. Even after taking the limit as prescribed, for general R2 theory,

there remain extra divergences in the bulk equation of motion. It is impossible to set these

56 2.5. DISCUSSION

divergences to zero at all points of the parameter space, although this can be done for specific

points like the Gauss-Bonnet point.

As we discussed, the absence of ∼ K3 terms is the R2 equation of motion is an artifact

of the way limits have to be taken in the LM method for the Gauss-Bonnet case. The limit

can also be taken in such a way so as to preserve ∼ K3 terms. It is worth recapitulating the

results this way of taking the limit gives for Gauss-Bonnet theory. As we showed, using the

second-order conical metric, the bulk equation of motion for Gauss-Bonnet theory, before

taking the limit, has divergences only in the zz and ij components. There is no divergence

in the zz component, while the divergence in the zi component turns out to be a constraint

equation that vanishes by itself on using the Codazzi-Mainardi relation on AdS space. This

same constraint equation results from the Jacobson-Myers functional, as well, on taking

tangential variations of the surface. It is not clear what the relevance of the divergence in the

ij component is in the LM method. Were we to ignore this divergence, the surface equation

of motion that would result from the zz component for Gauss-Bonnet theory, after taking

the limit, is cK = 0, where c is proportional to the Weyl anomaly. This equation is clearly

not what comes from the Jacobson-Myers functional. However, the resulting minimal surface

is what one obtains on extremizing just the Wald entropy part of the functional. It would

be interesting to check whether the zz component of R2 theory also leads to the same result.

One of the pending issues with the LM method is to fix the ambiguity present in the limit-

taking procedure. However, fixing this by itself does not seem enough to simultaneously cure

the two problems present for R2 theory: the absence of ∼ K3 terms and the presence of extra

constraints; although, it can remove one of these problems from the list. The ambiguity in

the limit-taking procedure is not unique to the LM method. Similar, though not exactly the

same, issues occur in studies of co-dimension two branes in the context of brane-world gravity

[54]. It is possible that a further modification to the LM method will fix these problems;

on the contrary, it may be that one cannot get rid of it in any way. The problem of extra

divergences is related to the derivative order of the bulk equation of motion and seems to

spring from the pathology of the general R2 theory itself. In this sense, it is not surprising

that we encounter it for general higher-derivative theories. Higher-derivative theories are

known to suffer from problems regarding unitarity [55, 56, 57]. These problems seem to be

manifested in the LM method in the inability to remove all divergences, that occur on using

the conical metric, from the bulk equation of motion.

What does our analysis say about the validity of the formula proposed in [31, 30] as the

entanglement entropy functional? For general R2 theory as we demonstrated the leading-

order terms match on both sides, which stops short of being a validation of the proposal

for this theory. This test, at present, is similar in scope in refining conjectured entropy

functionals for higher-derivative theories as the test whether the entropy functional leads to


the correct universal terms. Also we showed, for quasi-topological theory the universal terms

are not sensitive to terms of the form ∼ K4 in the entropy functional (similar statement

applies for other higher-derivative theories) and one can change these terms and still have

the universal terms come out to be correct.

The LM method, therefore, in its current form has limitations that make it ineffective in

testing proposed entropy functionals for generic higher-derivative gravity theories. The fact

that the LM method only works for specific theories may indicate one of two things. One

possibility is that entropy functionals only exist for specific theories such as Lovelock theories,

for which the result of the surface equation of motion from the existing entropy functional

and the LM method coincide. The other possibility, as mentioned before, is that the LM

method needs some modification. In this context, it is also desirable that alternate methods

to test entropy functionals be developed. Our argument can be extended for other Lovelock

theories in general dimensions and can be shown that extremization condition can be derived

consistently. But for for general theories it fails. In short LM method in its current form

only picks out some particular theories of gravity.

After our work [44], it was shown in the [58], that one can possibly modify the LM

method by breaking replica symmetry inside the bulk spacetime, but keeping the symmetry

preserved at the boundary. This generates new terms in the conical metric and enhances the

Zn symmetry group inside the bulk. Using this proposal one can show that, it is possible

to cancel the offending K3 terms for the Gauss-Bonnet theory without taking the limits as

shown in this chapter. This methods works only when one treats the Gauss-Bonnet coupling

perturbatively. But still it remains to be investigated, whether it is possible to cure the

problem for general four-derivative theories.

Appendix

A: Conical Metric

Near the conical singularity the bulk metric can also be written as

ds2 = ρ(x, y)−2ε(dx2 + dy2) + ρ(x, y)−2εapj dxpdxj + gij dx

idxj . (2.97)

The two-dimensional part is written in cartesian coordinates x and y and ρ(x, y) =√x2 + y2.

We have written the metric upto terms first order in x(y). The co-dimension two surface (Σ)

is located at x = 0 and y = 0. The metric gij can be written down order by order in x(y)

after expanding around the surface Σ as

gij = hij + x ∂xgij∣∣Σ

+ y ∂ygij∣∣Σ

+ · · · . (2.98)

58 2.5. DISCUSSION

The surface tensor hij is independent of x and y. The variable apj ∼ O(x).

The extrinsic curvatures for the co-dimension two surface (Σ) are defined as

Ksij = eβj∇insβ∣∣Σ

= eβj (∂insβ − Γδαβeαi nsδ)

∣∣Σ. (2.99)

Expanding the Christoffel in terms of the metric and using the fact that the first term eβj ∂insβ

vanishes it follows that

Kxij =1

2∂x gij

∣∣Σ, Kyij =

1

2∂ygij

∣∣Σ. (2.100)

We now make the simplifying assumption that the metric gij is independent of the co-ordinate

y. Under this assumption, the extrinsic curvature Kyij vanishes as ∂ygij vanishes.

The complex coordinates z and z used in the metric in Eq. (2.14) are related to x and y

as

z = x+ iy, z = x− iy . (2.101)

In these coordinates the metric is

ds2 = e2ρ(z,z)(dzdz) + gijdxidxj + 2e2ρ(z,z)Ai(zdz − zdz)dyi , (2.102)

where

gij = hij + zKzij + zKzij + · · · . (2.103)

The extrinsic curvatures in this coordinate system are related to Kxij and Kyij as

Kzij =Kxij + iKyij

2, Kzij =

Kxij − iKyij2

. (2.104)

Since Kyij = 0 we have

Kzij = Kzij . (2.105)

Similar considerations apply to the second-order quantities Q.

B. Curvature Tensors

In this appendix, we list components of the curvature tensors for the metric in Eq. (2.53),

that do not appear in the main text. We retain only terms uptil zeroeth-order in z, z.

The components of the Christoffels are

Γzzz = − εz, Γz zz = − ε

z, Γzij = −e−2ρ(z,z) Kzij , Γz ij = −e−2ρ(z,z) Kzij ,

Γizj =1

2Kizj , Γizj =

1

2Kizj , Γijk =

1

2gil(∂jglk + ∂kglj − ∂lgjk) ,

Γzzi = −2Ai , Γzzi = 2Ai . (2.106)

REFERENCES 59

The components of the Riemann tensor are

Rpqij = 2e2ρ(z,z)εpqFij + (KpjkKkqi −KpikKkqj) ,Rzizj = 1

4KzjkKkzi −Qzzij − ε

2zKzij ,

Rzizj = 12e2ρ(z,z)Fij − 2e2ρ(z,z)AiAj + 1

4KzjkKkzi − 1

2Qzzij ,

Rpijk = 12(∇kKpij −∇jKpik) ,

Rikjl = Rikjl + 12e−2ρ(z,z)(KzilKzjk +KzilKzjk −KzijKzkl −KzijKzkl) , (2.107)

where Fij ≡ ∂iAj − ∂jAi.The components of the Ricci tensor are

Rzi = 12(∇jKzji −∇iKz) ,

Rzz = 14KzijKijz − 1

2Qzz − ε

2zKz ,

Rzz = 14Kzij Kijz − 1

2Qzz − 2e2ρ(z,z)(AiAi − 3Ω) ,

Rij = e−2ρ(z,z)(KkzjKzik +KkzjKzik − 1

2KzijKz − 1

2KzijKz − 2Qzzij

)+Rij − 8AiAj . (2.108)

As in the main text, ∇ used in the above equations is the Van der Waerden-Bortolotti

covariant derivative [45] defined in Eq. (2.24).

The Ricci scalar is

R = R+ 24Ω− 16AiAi − e−2ρ(z,z)(KzKz − 3KzijKijz + 4Qzz

). (2.109)

References



M. Srednicki, “Entropy and area,” Phys. Rev. Lett. 71 (1993) 666 [hep-th/9303048].

L. Susskind and J. Uglum, “Black hole entropy in canonical quantum gravity and

superstring theory,” Phys. Rev. D 50 (1994) 2700 [hep-th/9401070].

E. Bianchi and R. C. Myers, “On the Architecture of Spacetime Geometry,” Class.

Quant. Grav. 31 (2014) 21, 214002 [arXiv:1212.5183 [hep-th]].

H. Casini, F. D. Mazzitelli and E. T. Lino, “Area terms in entanglement entropy,”


S. N. Solodukhin, “Newton constant, contact terms and entropy,” Phys. Rev. D 91

(2015) 8, 084028 [arXiv:1502.03758 [hep-th]].

S. N. Solodukhin, “One loop renormalization of black hole entropy due to nonminimally

coupled matter,” Phys. Rev. D 52 (1995) 7046 [hep-th/9504022].

60 REFERENCES

S. N. Solodukhin, “Entanglement entropy of black holes,” Living Rev. Rel. 14 (2011)

8 [arXiv:1104.3712 [hep-th]].

and the references there in.



C. Holzhey, F. Larsen and F. Wilczek, “Geometric and renormalized entropy in con-

formal field theory,” Nucl. Phys. B 424 (1994) 443 [hep-th/9403108].

P. Calabrese and J. L. Cardy, “Entanglement entropy and quantum field theory,”

J. Stat. Mech. 0406, P06002 (2004) [hep-th/0405152].

H. Casini and M. Huerta, “Entanglement entropy in free quantum field theory,”

J. Phys. A 42, 504007 (2009) [arXiv:0905.2562 [hep-th]].


ment entropy proposal,” JHEP 0707, 062 (2007) [arXiv:0705.0016 [hep-th]].

[4] R. C. Myers and A. Sinha, “Seeing a c-theorem with holography,” Phys. Rev. D 82,

046006 (2010) [arXiv:1006.1263 [hep-th]].

R. C. Myers and A. Sinha, “Holographic c-theorems in arbitrary dimensions,”

JHEP 1101, 125 (2011) [arXiv:1011.5819 [hep-th]].

Y. Bea, J. D. Edelstein, G. Itsios, K. S. Kooner, C. Nunez, D. Schofield and J. A. Sierra-

Garcia,“Compactifications of the Klebanov-Witten CFT and new AdS3 backgrounds,”

JHEP 1505 (2015) 062 [arXiv:1503.07527 [hep-th]].

R. C. Myers and A. Singh, “Comments on Holographic Entanglement Entropy and RG

Flows,” JHEP 1204 (2012) 122 [arXiv:1202.2068 [hep-th]].

T. Albash and C. V. Johnson, “Holographic Entanglement Entropy and Renormaliza-

tion Group Flow,” JHEP 1202 (2012) 095 [arXiv:1110.1074 [hep-th]].

[5] For some interesting properties of extremal surfaces one may consult

S. Banerjee, A. Bhattacharyya, A. Kaviraj, K. Sen and A. Sinha, “Constraining gravity

using entanglement in AdS/CFT,” arXiv:1401.5089 [hep-th].

S. S. Pal, “Extremal Surfaces And Entanglement Entropy,” arXiv:1312.0088 [hep-th].

[6] J. Erdmenger, M. Flory and C. Sleight, “Conditions on holographic entangling surfaces

for black hole geometries in higher derivative gravity,” arXiv:1401.5075 [hep-th].

[7] For an alternative proposal for calculating entanglement entropy without using replica

trick

A. Bhattacharyya and A. Sinha, “Entanglement entropy from the holographic stress

tensor,” Class. Quant. Grav. 30, 235032 (2013) [arXiv:1303.1884 [hep-th]].

REFERENCES 61

A. Bhattacharyya and A. Sinha, “Entanglement entropy from surface terms in general

relativity,” Int. J. Mod. Phys. D 22, 1342020 (2013) [arXiv:1305.3448 [gr-qc]].

[8] T. Faulkner, “The Entanglement Renyi Entropies of Disjoint Intervals in AdS/CFT,”


T. Hartman, “Entanglement Entropy at Large Central Charge,” arXiv:1303.6955 [hep-

th].

[9] For quantum corrections to Ryu-Takayanagi proposal

T. Barrella, X. Dong, S. A. Hartnoll and V. L. Martin, “Holographic entanglement

beyond classical gravity,” JHEP 1309, 109 (2013) [arXiv:1306.4682 [hep-th]].

T. Faulkner, A. Lewkowycz and J. Maldacena, “Quantum corrections to holographic

entanglement entropy,” JHEP 1311, 074 (2013) [arXiv:1307.2892]

B. Swingle, L. Huijse and S. Sachdev, “Entanglement entropy of compressible holo-

graphic matter: loop corrections from bulk fermions,” arXiv:1308.3234 [hep-th].


erty of Entanglement Entropy for Excited States,” Phys. Rev. Lett. 110, 091602 (2013)

[arXiv:1212.1164 [hep-th]].

M. Nozaki, T. Numasawa, A. Prudenziati and T. Takayanagi, “Dynamics of Entangle-

ment Entropy from Einstein Equation,” arXiv:1304.7100 [hep-th].

G. Wong, I. Klich, L. A. P. Zayas and D. Vaman, “Entanglement Temperature and

Entanglement Entropy of Excited States,” arXiv:1305.3291 [hep-th].




momentum and U(1) charge,” JHEP 1311 (2013) 052 [arXiv:1306.4974 [hep-th]].

[11] S. He, T. Numasawa, T. Takayanagi and K. Watanabe, “Quantum Dimension as En-

tanglement Entropy in 2D CFTs,” arXiv:1403.0702 [hep-th].

M. Nozaki, T. Numasawa and T. Takayanagi, “Quantum Entanglement of Local Op-

erators in Conformal Field Theories,” arXiv:1401.0539 [hep-th].

A. Mollabashi, N. Shiba and T. Takayanagi, “Entanglement between Two Interacting

CFTs and Generalized Holographic Entanglement Entropy,” arXiv:1403.1393 [hep-th].

P. Caputa, J. Simn, A. tikonas and T. Takayanagi, “Quantum Entanglement of Local-

ized Excited States at Finite Temperature,” JHEP 1501 (2015) 102 [arXiv:1410.2287

[hep-th]].

P. Caputa, J. Simn, A. tikonas, T. Takayanagi and K. Watanabe, “Scrambling time

from local perturbations of the eternal BTZ black hole,” arXiv:1503.08161 [hep-th].

62 REFERENCES

W. Z. Guo and S. He, “Rnyi entropy of locally excited states with thermal and bound-

ary effect in 2D CFTs,” JHEP 1504 (2015) 099 [arXiv:1501.00757 [hep-th]].

M. Headrick, A. Maloney, E. Perlmutter and I. G. Zadeh, “Renyi Entropies, the Ana-

lytic Bootstrap, and 3D Quantum Gravity at Higher Genus,” arXiv:1503.07111 [hep-

th].

S. He, T. Numasawa, T. Takayanagi and K. Watanabe, “Notes on Entanglement En-

tropy in String Theory,” JHEP 1505 (2015) 106 [arXiv:1412.5606 [hep-th]].

[12] J. Maldacena and G. L. Pimentel, “Entanglement entropy in de Sitter space,”

JHEP 1302 (2013) 038 [arXiv:1210.7244 [hep-th]]. K. Narayan, T. Takayanagi and

S. P. Trivedi, “AdS plane waves and entanglement entropy,” JHEP 1304 (2013) 051

[arXiv:1212.4328 [hep-th]].

K. Narayan, “Non-conformal brane plane waves and entanglement entropy,” Phys. Lett.

B 726 (2013) 370 [arXiv:1304.6697 [hep-th]].

D. Mukherjee and K. Narayan, “AdS plane waves, entanglement and mutual informa-

tion,” Phys. Rev. D 90, no. 2, 026003 (2014) [arXiv:1405.3553 [hep-th]].

K. Narayan, “de Sitter extremal surfaces,” arXiv:1501.03019 [hep-th].

K. Narayan, “Lightlike limit of entanglement entropy,” Phys. Rev. D 91 (2015) 8,

086010 [arXiv:1408.7021 [hep-th]].

K. Narayan, “de Sitter space and extremal surfaces for spheres,” arXiv:1504.07430

[hep-th].

M. Miyaji and T. Takayanagi, “Surface/State Correspondence as a Generalized Holog-

raphy,” arXiv:1503.03542 [hep-th].

Y. Sato, “Comments on Entanglement Entropy in the dS/CFT Correspondence,” Phys.

Rev. D 91 (2015) 8, 086009 [arXiv:1501.04903 [hep-th]].

S. Chakraborty, P. Dey, S. Karar and S. Roy, “Entanglement thermodynamics for an

excited state of Lifshitz system,” JHEP 1504 (2015) 133 [arXiv:1412.1276 [hep-th]].

[13] H. Liu and S. J. Suh, “Entanglement Tsunami: Universal Scaling in Holographic Ther-

malization,” Phys. Rev. Lett. 112 (2014) 011601 [arXiv:1305.7244 [hep-th]].

H. Liu and S. J. Suh, “Entanglement growth during thermalization in holographic sys-

tems,” Phys. Rev. D 89 (2014) 6, 066012 [arXiv:1311.1200 [hep-th]].

S. R. Das, D. A. Galante and R. C. Myers, “Smooth and fast versus instantaneous

quenches in quantum field theory,” arXiv:1505.05224 [hep-th].

S. R. Das, D. A. Galante and R. C. Myers, “Universality in fast quantum quenches,”

JHEP 1502 (2015) 167 [arXiv:1411.7710 [hep-th]].

S. R. Das, D. A. Galante and R. C. Myers, “Universal scaling in fast quantum quenches

in conformal field theories,” Phys. Rev. Lett. 112 (2014) 171601 [arXiv:1401.0560 [hep-

REFERENCES 63

th]].

A. Buchel, R. C. Myers and A. van Niekerk, “Nonlocal probes of thermalization in

holographic quenches with spectral methods,” JHEP 1502 (2015) 017 [arXiv:1410.6201

[hep-th]].

C. T. Asplund, A. Bernamonti, F. Galli and T. Hartman, “Holographic Entanglement

Entropy from 2d CFT: Heavy States and Local Quenches,” JHEP 1502 (2015) 171

[arXiv:1410.1392 [hep-th]].

T. Hartman and J. Maldacena, “Time Evolution of Entanglement Entropy from Black

Hole Interiors,” JHEP 1305 (2013) 014 [arXiv:1303.1080 [hep-th]].

T. Albash and C. V. Johnson, “Evolution of Holographic Entanglement Entropy

after Thermal and Electromagnetic Quenches,” New J. Phys. 13 (2011) 045017

[arXiv:1008.3027 [hep-th]].

V. Balasubramanian, A. Bernamonti, J. de Boer, N. Copland, B. Craps, E. Keski-

Vakkuri, B. Muller and A. Schafer et al., “Holographic Thermalization,” Phys. Rev. D

84 (2011) 026010 [arXiv:1103.2683 [hep-th]].

[14] P. Bueno, R. C. Myers and W. Witczak-Krempa, “Universality of corner entanglement

in conformal field theories,” arXiv:1505.04804 [hep-th].

P. Bueno and R. C. Myers, “Corner contributions to holographic entanglement en-

tropy,” arXiv:1505.07842 [hep-th].

[15] O. Ben-Ami, D. Carmi and M. Smolkin, “Renormalization group flow of entanglement

entropy on spheres,” arXiv:1504.00913 [hep-th].

V. Rosenhaus and M. Smolkin, “Entanglement Entropy for Relevant and Geometric

Perturbations,” JHEP 1502 (2015) 015 [arXiv:1410.6530 [hep-th]].

V. Rosenhaus and M. Smolkin, “Entanglement entropy, planar surfaces, and spectral

functions,” JHEP 1409 (2014) 119 [arXiv:1407.2891 [hep-th]].

V. Rosenhaus and M. Smolkin, “Entanglement Entropy Flow and the Ward Identity,”

Phys. Rev. Lett. 113 (2014) 26, 261602 [arXiv:1406.2716 [hep-th]].

M. Smolkin and S. N. Solodukhin, “Correlation functions on conical defects,” Phys.

Rev. D 91 (2015) 4, 044008 [arXiv:1406.2512 [hep-th]].

X. Huang, L. Y. Hung and F. L. Lin, “OPE of the stress tensors and surface operators,”

arXiv:1502.02487 [hep-th]. S. Banerjee, “Trace Anomaly Matching and Exact Results

For Entanglement Entropy,” arXiv:1405.4876 [hep-th].

A. Allais and M. Mezei, “Some results on the shape dependence of entanglement and

Rnyi entropies,” Phys. Rev. D 91 (2015) 4, 046002 [arXiv:1407.7249 [hep-th]].

A. F. Astaneh, G. Gibbons and S. N. Solodukhin, “What surface maximizes entangle-

ment entropy?,” Phys. Rev. D 90 (2014) 8, 085021 [arXiv:1407.4719 [hep-th]].

64 REFERENCES

S. N. Solodukhin, “Conformal a-charge, correlation functions and conical defects,”

Phys. Lett. B 736 (2014) 283 [arXiv:1406.5368 [hep-th]].

[16] A. F. Astaneh and S. N. Solodukhin, “The Wald entropy and 6d conformal anomaly,”

arXiv:1504.01653 [hep-th].

A. F. Astaneh, A. Patrushev and S. N. Solodukhin, “Entropy discrepancy and total

derivatives in trace anomaly,” arXiv:1412.0452 [hep-th].

Y. Huang and R. X. Miao, “A note on the resolution of the entropy discrepancy,”

arXiv:1504.02301 [hep-th].

R. X. Miao, “Universal Terms of Entanglement Entropy for 6d CFTs,”

arXiv:1503.05538 [hep-th].

R. X. Miao and W. z. Guo, “Holographic Entanglement Entropy for the Most General

Higher Derivative Gravity,” arXiv:1411.5579 [hep-th].

[17] F. M. Haehl, T. Hartman, D. Marolf, H. Maxfield and M. Rangamani, “Topological

aspects of generalized gravitational entropy,” JHEP 1505 (2015) 023 [arXiv:1412.7561

[hep-th]].

A. Hamma, L. Y. Hung, A. Marciano and M. Zhang, “Area Law from Loop Quantum

Gravity,” arXiv:1506.01623 [gr-qc].

P. Caputa, V. Jejjala and H. Soltanpanahi, “Entanglement entropy of extremal BTZ

black holes,” Phys. Rev. D 89 (2014) 4, 046006 [arXiv:1309.7852 [hep-th]].

C. V. Johnson, “Large N Phase Transitions, Finite Volume, and Entanglement En-

tropy,” JHEP 1403 (2014) 047 [arXiv:1306.4955 [hep-th]].

T. Albash and C. V. Johnson, “Holographic Studies of Entanglement Entropy in Su-

perconductors,” JHEP 1205 (2012) 079 [arXiv:1202.2605 [hep-th]].

N. Shiba and T. Takayanagi, “Volume Law for the Entanglement Entropy in Non-local

QFTs,” JHEP 1402 (2014) 033 [arXiv:1311.1643 [hep-th]].[arXiv:1403.0702 [hep-th]].

N. Ogawa, T. Takayanagi and T. Ugajin, “Holographic Fermi Surfaces and Entangle-

ment Entropy,” JHEP 1201 (2012) 125 [arXiv:1111.1023 [hep-th]].

T. Takayanagi, “Holographic Dual of BCFT,” Phys. Rev. Lett. 107 (2011) 101602

[arXiv:1105.5165 [hep-th]].

M. Fujita, Y. Hatsuda and T. Takayanagi, “Probing AdS Wormholes by Entanglement

Entropy,” JHEP 1106 (2011) 141 [arXiv:1104.4907 [hep-th]].

W. Li and T. Takayanagi, “Holography and Entanglement in Flat Spacetime,” Phys.

Rev. Lett. 106 (2011) 141301 [arXiv:1010.3700 [hep-th]].

A. Castro, S. Detournay, N. Iqbal and E. Perlmutter, “Holographic entanglement en-

tropy and gravitational anomalies,” JHEP 1407 (2014) 114 [arXiv:1405.2792 [hep-th]].

M. Ammon, A. Castro and N. Iqbal, “Wilson Lines and Entanglement Entropy in

REFERENCES 65

Higher Spin Gravity,” JHEP 1310 (2013) 110 [arXiv:1306.4338 [hep-th]].

J. de Boer and J. I. Jottar, “Thermodynamics of higher spin black holes in AdS3,”

JHEP 1401 (2014) 023 [arXiv:1302.0816 [hep-th]].


AdS/CFT,” Phys. Rev. Lett. 96, 181602 (2006) [hep-th/0603001].

[19] S. Ryu and T. Takayanagi, “Aspects of Holographic Entanglement Entropy,”

JHEP 0608, 045 (2006) [hep-th/0605073].

T. Nishioka, S. Ryu and T. Takayanagi, “Holographic Entanglement Entropy: An

Overview,” J. Phys. A 42, 504008 (2009) [arXiv:0905.0932 [hep-th]]

[20] D. V. Fursaev, “Proof of the holographic formula for entanglement entropy,” JHEP

0609, 018 (2006) [hep-th/0606184].

[21] M. Headrick, “Entanglement Renyi entropies in holographic theories,”

Phys. Rev. D 82, 126010 (2010) [arXiv:1006.0047 [hep-th]].

[22] For spherical entangling surface Ryu-Takayanagi proposal has been proved in

H. Casini, M. Huerta and R. C. Myers, “Towards a derivation of holographic entangle-

ment entropy,” JHEP 1105, 036 (2011) [arXiv:1102.0440 [hep-th]].

[23] D. V. Fursaev, “Entanglement Renyi Entropies in Conformal Field Theories and Holog-


[24] A. Lewkowycz and J. Maldacena, “Generalized gravitational entropy,” JHEP 1308,

090 (2013) [arXiv:1304.4926 [hep-th]].


[gr-qc/9307038].



[gr-qc/9503052].

[26] V. Iyer and R. M. Wald, “Some properties of Noether charge and a proposal for dy-


[27] using Hamiltonian approach

T. Jacobson and R. C. Myers, “Black hole entropy and higher curvature interactions,”

Phys. Rev. Lett. 70, 3684 (1993) [hep-th/9305016].

T. Jacobson, G. Kang and R. C. Myers, “On black hole entropy,” Phys. Rev. D 49

(1994) 6587 [gr-qc/9312023].

66 REFERENCES



[29] J. de Boer, M. Kulaxizi and A. Parnachev, “Holographic Entanglement Entropy in



JHEP 1401, 044 (2014) [arXiv:1310.5713 [hep-th]].

[31] J. Camps, “Generalized entropy and higher derivative Gravity,” JHEP 1403, 070 (2014)

[arXiv:1310.6659 [hep-th]].

[32] L. Y. Hung, R. C. Myers, M. Smolkin and A. Yale, “Holographic Calculations of Renyi

Entropy,” JHEP 1112 (2011) 047 [arXiv:1110.1084 [hep-th]].

L. Y. Hung, R. C. Myers and M. Smolkin, “Some Calculable Contributions to Holo-

graphic Entanglement Entropy,” JHEP 1108 (2011) 039 [arXiv:1105.6055 [hep-th]].

[33] D. V. Fursaev, A. Patrushev and S. N. Solodukhin, “Distributional Geometry of

Squashed Cones,” arXiv:1306.4000 [hep-th].





[hep-th]].

X. O. Camanho, J. D. Edelstein and J. M. Snchez De Santos, “Lovelock theory

and the AdS/CFT correspondence,” Gen. Rel. Grav. 46 (2014) 1637 [arXiv:1309.6483

[hep-th]].


holography,” JHEP 1308, 012 (2013), [arXiv:1305.6694 [hep-th]].

[37] B. Chen and J. -j. Zhang, “Note on generalized gravitational entropy in Lovelock grav-

ity,” JHEP 1307, 185 (2013) [arXiv:1305.6767 [hep-th]].

[38] R. C. Myers and B. Robinson, “Black Holes in Quasi-topological Gravity,”

JHEP 1008, 067 (2010) [arXiv:1003.5357 [gr-qc]].

R. C. Myers, M. F. Paulos and A. Sinha, “Holographic studies of quasi-topological

gravity,” JHEP 1008, 035 (2010) [arXiv:1004.2055 [hep-th]].

[39] R. -X. Miao, “A Note on Holographic Weyl Anomaly and Entanglement Entropy,”

Class. Quant. Grav. 31 (2014) 065009 [arXiv:1309.0211 [hep-th]].

REFERENCES 67

[40] M. Alishahiha, A. F. Astaneh and M. R. M. Mozaffar, “Entanglement Entropy for

Logarithmic Conformal Field Theory,” Phys. Rev. D 89 (2014) 065023 [arXiv:1310.4294

[hep-th]].

M. Alishahiha, A. F. Astaneh and M. R. M. Mozaffar, “Holographic Entanglement

Entropy for 4D Conformal Gravity,” JHEP 1402 (2014) 008 [arXiv:1311.4329 [hep-

th]].

D. K. O’Keeffe and A. W. Peet, “Electric hyperscaling violating solutions in Einstein-

Maxwell-dilaton gravity with R2 corrections,” arXiv:1312.2261 [hep-th].

R. Pourhasan, “Spacetime entanglement with f(R) gravity,” JHEP 1406 (2014) 004

[arXiv:1403.0951 [hep-th]].

[41] J. M. Maldacena, “The Large N limit of superconformal field theories and supergravity,”

Adv. Theor. Math. Phys. 2, 231 (1998) [hep-th/9711200].

[42] D. V. Fursaev and S. N. Solodukhin, “On the description of the Riemannian geometry

in the presence of conical defects,” Phys. Rev. D 52, 2133 (1995) [hep-th/9501127].

[43] R. M Wald, “General Relativity”,The University of Chicago Press, 1984.



[45] K. Yano , “Infinitesimal variations of submanifolds,” Kodai Mathematical Journal 1

(1978), no. 1, 30–44. doi:10.2996/kmj/1138035439.

Chen, Bang-yen; Yano, Kentaro, “On the theory of normal variations”, Journal of

Differential Geometry 13 (1978), no. 1, 1–10.

[46] K. S. Viswanathan and R. Parthasarathy, “String theory in curved space-time,”

Phys. Rev. D 55 (1997) 3800 [hep-th/9605007].

R. Capovilla and J. Guven, “Geometry of deformations of relativistic membranes,”

Phys. Rev. D 51, 6736 (1995) [gr-qc/9411060].

[47] B.Boisseau and P.S.Letelier, Phys.Rev. D46, 1721

[48] V. Rosenhaus and M. Smolkin, “Entanglement Entropy: A Perturbative Calculation,”


[49] Y. Kats and P. Petrov, “Effect of curvature squared corrections in AdS on the viscosity

of the dual gauge theory,” JHEP 0901, 044 (2009) [arXiv:0712.0743 [hep-th]].

A. Sinha, “On higher derivative gravity, c-theorems and cosmology,” Class. Quant.

Grav. 28 (2011) 085002 [arXiv:1008.4315 [hep-th]].

68 REFERENCES

[50] S. Sarkar and A. C. Wall, “Generalized second law at linear order for actions that

are functions of Lovelock densities,” Phys. Rev. D 88, 044017 (2013) [arXiv:1306.1623

[gr-qc]].

[51] T. C. Sisman, I. Gullu and B. Tekin, “All unitary cubic curvature gravities in D di-

mensions,” Class. Quant. Grav. 28 (2011) 195004 [arXiv:1103.2307 [hep-th]].

[52] A. Schwimmer and S. Theisen, “Entanglement Entropy, Trace Anomalies and Holog-

raphy,” Nucl. Phys. B 801, 1 (2008) [arXiv:0802.1017 [hep-th]].

S. N. Solodukhin, “Entanglement entropy, conformal invariance and extrinsic geome-

try,” Phys. Lett. B 665, 305 (2008) [arXiv:0802.3117 [hep-th]].

[53] S. ’i. Nojiri and S. D. Odintsov, “On the conformal anomaly from higher derivative

gravity in AdS / CFT correspondence,” Int. J. Mod. Phys. A 15, 413 (2000) [hep-

th/9903033].

[54] For representative papers see: P. Bostock, R. Gregory, I. Navarro and J. Santiago,

“Einstein gravity on the codimension 2-brane?,” Phys. Rev. Lett. 92, 221601 (2004)

[hep-th/0311074].

C. Charmousis and R. Zegers, “Matching conditions for a brane of arbitrary codimen-

sion,” JHEP 0508, 075 (2005) [hep-th/0502170].

[55] S. Deser, H. Liu, H. Lu, C. N. Pope, T. C. Sisman and B. Tekin,

“Critical Points of D-Dimensional Extended Gravities,” Phys. Rev. D 83 (2011) 061502

[arXiv:1101.4009 [hep-th]].

[56] J. Z. Simon, “Higher Derivative Lagrangians, Nonlocality, Problems and Solutions,”

Phys. Rev. D 41 (1990) 3720.

E. A. Bergshoeff, O. Hohm, J. Rosseel and P. K. Townsend, “Modes of Log Gravity,”


M. Alishahiha and R. Fareghbal, “D-Dimensional Log Gravity,” Phys. Rev. D 83 (2011)

084052 [arXiv:1101.5891 [hep-th]].

I. Gullu, M. Gurses, T. C. Sisman and B. Tekin, “AdS Waves as Exact Solutions to

Quadratic Gravity,” Phys. Rev. D 83 (2011) 084015 [arXiv:1102.1921 [hep-th]].

[57] M. Alishahiha and A. Naseh, “Holographic renormalization of new massive gravity,”


N. Johansson, A. Naseh and T. Zojer, “Holographic two-point functions for 4d log-

gravity,” JHEP 1209 (2012) 114 [arXiv:1205.5804 [hep-th]].

[58] J. Camps and W. R. Kelly, “Generalized gravitational entropy without replica symme-

try,” JHEP 1503 (2015) 061 [arXiv:1412.4093 [hep-th]].

3 Entanglement entropy from

generalized entropy

3.1 Introduction

In the Chapter.(2) we have shown that the identification of the entanglement entropy with the

generalized gravitational entropy has opened up the avenue for systematically generalizing

holographic entanglement entropy (EE) for more general bulk theories of gravity other than

Einstein gravity. This understanding is crucial in order to understand systematics of how

finite coupling effects in the field theory modify entanglement entropy. In this chapter we will

only focus on those particular theories which arise from “classical” and local higher derivative

corrections to the bulk theory1. Now our main objective will be to compute the universal

term for EE for various higher derivative gravity theories using generalized entropy method.

Rest of this chapter is mainly based on the work done with Prof. Aninda Sinha and Dr.

Menika Sharma [2].

3.2 Generalized entropy and Fefferman-Graham expan-

sion

Let us start by recapitulating generalized entropy method one more time. We will see fol-

lowing [3], the generalized gravitational entropy is defined as,

S = −n∂n(ln[Z(n)]− n ln[Z(1)])n=1 , (3.1)

where ln[Z(1)] is identified with the Euclidean gravitational action for which the period of

the Euclidean time is 2π and the boundary condition for other fields collectively denoted as φ

present in the action is φ(0) = φ(2π) . ln[Z(n)] is identified with the Euclidean gravitational

action In for which the period of the Euclidean time is 2πn and the boundary condition for φ

is still φ(0) = φ(2π) . This is the usual replica trick. Translating this fact for the holographic

1there are other type of corrections which arise from “quantum” or loop corrections to the effective action

which would include non-local effects. Interested readers are referred to [1].

69

70 3.2. GENERALIZED ENTROPY AND FEFFERMAN-GRAHAM EXPANSION

case we can define In for a regularized geometry on a cone whose opening angle is 2π/n. We

can analytically continue this for non integer n and then can compute the entropy. Also while

evaluating ln[Z(n)] we can perform the time τ integral from 0 to 2π and multiply it by n so

that ln[Z(n)] = n ln[Z]2π . The entropy calculated using this method is equal to the area of

some codimension 2 surface where the time circle shrinks to zero which can be shown to be

the minimal surface in Einstein gravity [3]. In this section we will show that this procedure

also gives the correct entanglement entropy for higher curvature gravity theories. To compute

the EE we have to start with some specific boundary geometry for the nth solution. Then we

can construct our bulk spacetime using the Fefferman-Graham expansion. We will consider

the following two 4-dimensional metrics following [4],

ds2cylinder = f(r, b)dr2 + r2dτ 2 + (f0 + rnd1−n cos(τ))2dφ2 + dz2

ds2sphere = f(r, b)dr2 + r2dτ 2 + (f0 + rnd1−n cos(τ))2(dθ2 + sin2 θdφ2)

(3.2)

where, f(r, b) = r2+b2n2

r2+b2. For b→ 0 and n→ 1 limit these two metrics reduce to the cylinder

and the sphere. This metric is known as “squashed cone” metric. The key point in Eq. (3.2)

as compared to earlier regularizations e.g., [3] is the introduction of a regulator in the extrinsic

curvature terms. This is needed since otherwise the Ricci scalar would go like (n− 1)/r and

would be singular. Another important point is that b is a regulator which at this stage does

not have an restriction except that f(0, b) = n2. In AdS/CFT we do not expect an arbitrary

parameter to appear in the metric. b is here a dimensionful quantity having the dimension

of r . So b must be proportional to f0(n− 1)α>0 such that it goes to zero as n→ 1. We can

take the metrics in Eq. (3.2) as boundary metrics and construct the bulk spacetime using the

Fefferman-Graham expansion. Notice that our starting point is a smooth metric. At the end

of the calculation, when we remove the regulators and compute EE, we will separately check

what the contribution from the singularities is going to be. In the best case scenario, although

the boundary metric will be singular once the regulator is removed, the bulk metric will at

most be mildly singular, namely the on-shell bulk action will not be singular, following the

terminology used in [3]. As in [3] we could have done a conformal transformation to pull out

a factor of r2 such that the r, τ part of the metric looks like dτ 2 + dr2

r2which would make the

time-circle non-shrinking. We can use this form of the metric with a suitable regularization

and do the calculation after verifying that there are no singularities in the bulk. Since this is

a conformal transformation of a smooth metric, the results for the universal part of the EE

will remain unchanged. One can write the bulk metric as,

ds2 = L2dρ2

4ρ2+

(g(0)ij + ρg

(2)ij + .....)

ρdxidxj . (3.3)

CHAPTER 3. ENTANGLEMENT ENTROPY FROM GENERALIZED ENTROPY 71

To evaluate the log term we will need the g(2)ij coefficient and here we will use Eq. (3.2) as

g(0)ij . We will consider here a 5 dimensional bulk lagrangian. In this case,

g(2)ij = − L

2

2(R

(0)ij −

1

6g

(0)ij R

(0)) ,

where R(0)ij and R(0) are constructed using g

(0)ij . Note that in all subsequent calculations g

(2)ij

will play an important role. The structure of g(2)ij is independent of the form of the higher

derivative terms present in the action. Only terms proportional to n − 1 in the on-shell

bulk action contributes to the SEE. The calculation is similar in spirit to the way that Weyl

anomaly is extracted in AdS/CFT, e.g., [5] except that the n−1 dependence comes from the

neighbourhood of r = 0 in the bulk action. In the next section we proceed to give details of

this.

Regularization procedure

To illustrate the regularization procedure in some detail, we start with some simple examples

involving curvature polynomials 2. We calculate g(2)ij and evaluate the following integral ,

I1 =

∫d5x√g RµνR

µν . (3.4)

Following3 [4], in the integrand, we put r = bx then expand around b = 0 and pick out the

O(b0) term. The r integral is between 0 < r < r0. This makes the upper limit of the x

integral to be r0/b which goes to infinity. We will be interested in the log term so we extract

first the coefficient of 1ρ

term which has the following form,

I1 = b

∫dρ

ρdτ d2y

∫ ∞0

dx (n− 1)2ζ(x, n)(bx)2n−3 +O((n− 1)3) . (3.5)

We have here shown only the leading term. Note that at this stage the integrand is propor-

tional to (n− 1)2 whereas we need get something proportional to (n− 1). The integral over

x will give a factor of 1/(n − 1). We will now expand ζ(x, n) around n = 1 and then carry

out the integral over x. After expanding around n = 1 this leads to

I1 = (n− 1)ζ1 +O(n− 1)2 + · · · . (3.6)

Note that the rn factor in the cylindrical and the spherical parts in Eq. (3.2) were crucial in

reaching this point. ζ1 is just a quantity independent of the regularization parameters b, d,

2We thank Sasha Patrushev for discussions on this topic.3Alternatively we could have done the expansion around x = 0 first, since it was assumed in [4] that the

metric is valid between 0 < r < b f0. Then we could have integrated x in the neighbourhood of x = 0.

The results are identical.


ε, ε′. The same procedure is applied for other curvature polynomial integrals. For example,

I2 =

∫√g d5xRµνρσR

µνρσ = (n− 1)ζ2 +O(n− 1)2 + · · · ,

I3 =

∫√g d5xR2 = O(n− 1)2 + · · · .

(3.7)

3.2.1 Four derivative theory

Let us now consider the general R2 theory lagrangian as shown below.

L = L1 + L2 , (3.8)

where

L1 = R +12

L2(3.9)

is the usual Einstein-Hilbert lagrangian with the cosmological constant appropriate for five-

dimensional AdS space and

L2 =L2

2

(λ1RαβγδR

αβγδ + λ2RαβRαβ + λ3R

2)

(3.10)

is the R2 lagrangian.

Also we will henceforth consider only a 5 dimensional bulk spacetime unless mentioned

otherwise. The boundary of this spacetime is at ρ = 0 . We then evaluate the total action

and extract the 1ρ

term and carry out the τ integral. We put r = b x and expand Eq. (3.8)

around b = 0 . Then we pick out the O(b0) term.

S = − 1

2`3P

∫dρ

ρdx d2y (n− 1)2a1

(bx)2n

x3+O((n− 1)3) , (3.11)

where

a1 =A(x)

18 b2f5/2∞ f0 (1 + x2)4

. (3.12)

A(x) is a function of x . For the cylinder we get,

A(x) = πL3(f 2∞(λ1

(4x8 + 16x6 + 43x4 + 36x2 + 9

)− 2

(20x8 + 80x6 + 161x4 + 108x2 + 27

)(λ2 + 5λ3)

)+ 6f∞

(5x8 + 20x6 + 38x4 + 24x2 + 6

)− 3

(8x8 + 32x6 + 59x4 + 36x2 + 9

) ).

(3.13)

We then carry out the x integral.

S = − 1

2`3P

∫dρ

ρd2y

A1(x, n)

36 b2 f5/2∞ (n2 − 1) f0 x2

∣∣∣∣∞0

, (3.14)


where

A1(x, n) = πL3(n− 1)2(bx)2n

[(n− 1)x4

2F1

(2, n+ 1;n+ 2;−x2

) (f 2∞(5λ1 − 14(λ2 + 5λ3))

+ 6f∞ − 3)

+ 2(n− 1)x42F1

(3, n+ 1;n+ 2;−x2

) (f 2∞(5λ1 − 14(λ2 + 5λ3))

+ 6f∞ − 3)

+ 2F1

(4, n+ 1;n+ 2;−x2

) (4f 2∞λ1x

4(1− n)− 40f 2∞λ2x

4(1− n)

− 200f 2∞λ3x

4(1− n) + 30f∞x4(1− n)− 24x4(1− n)

)− 9f 2

∞λ1(1 + n)

+ 54f 2∞λ2(1 + n) + 270f 2

∞λ3(1 + n)− 36f∞(1 + n) + 27(1 + n)

].

(3.15)

For the cylinder after doing the expansion around n = 1 and the remaining integrals (note

that ρ = z2 in the coordinates used in [6] and so ln δρ = 2 ln δ),

SEE = −cH2R

ln(f0

δ) . (3.16)

Here we have used 1 = f∞− 13f 2∞(λ1 +2λ2 +10λ3) and c is given in Eq. (2.19). For the sphere

we proceed similarly. In this case, expanding Eq. (3.11) around b = 0 we get ,

S = · · · − 1

2`3P

∫dρ

ρdx d2y (n− 1)2a1

(bx)2n

x3+O((n− 1)3) , (3.17)

where

a1 =A(x)

72 b2f5/2∞ f 4

0 (1 + x2)4. (3.18)

A(x) is a function of x . For the sphere we get,

A(x) = −πL3 sin(θ)

[300 b4λ3x

10f 2∞ − 45 b4x10f∞ + 600 b4λ3x

8f 2∞ − 90 b4x8f∞

+ 300 b4λ3x6f 2∞ − 45 b4x6f∞ + 36 b4x10 + 72 b4x8 + 36 b4x6 − 680 b2λ3R

2x10f 2∞

+ 84 b2R2x10f∞ − 1920 b2λ3R2x8f 2

∞ + 216 b2R2x8f∞ − 120 b2λ3R2x6f 2

∞ + 36 b2R2x6f∞

+ 1120 b2λ3R2x4f 2

∞ − 96 b2R2x4f∞ − 60 b2R2x10 − 144 b2R2x8 − 36 b2R2x6 + 48 b2R2x4

+ 2λ1f2∞(− 3b4

(x2 + 1

)2x6 + 2b2R2

(7x6 + 24x4 − 3x2 − 20

)x4 + 4R4

(x8 − 73x6 + 242x4

+ 361x2 + 54))

+ 4λ2f2∞(15b4

(x2 + 1

)2x6 − 2b2R2

(17x6 + 48x4 + 3x2 − 28

)x4

+ 4R4(13x8 − 13x6 + 230x4 + 301x2 + 54

))+ 4320λ3R

4f 2∞ + 1040λ3R

4x8f 2∞ − 192R4x8f∞

− 1040λ3R4x6f 2

∞ − 168R4x6f∞ + 18400λ3R4x4f 2

∞ − 2760R4x4f∞ + 24080λ3R4x2f 2

∞

− 3144R4x2f∞ − 576R4f∞ + 168R4x8 + 264R4x6 + 2208R4x4 + 2328R4x2 + 432R4)

].

(3.19)


After doing the x integral,

S = − 1

2`3P

∫dρ

ρd2y

A1(x, n)

144 b2 f5/2∞ n (n+ 1) f 4

0 x2

∣∣∣∣∞0

, (3.20)

where A1(x, n) is a function of x and n .

A1(x, n) = πL3(n− 1) sin(θ)(bx)2n

[− 8(n+ 1)R4

(f 2∞(λ1(145x2(n− 1) + 54n) + 2(λ2 + 5λ3)

(85x2(n− 1) + 54n))− 3f∞(n(35x2 + 24)− 35x2) + n(75x2 + 54

)− 75x2

)+ 2F1

(4, n+ 1;n+ 2;−x2

)(−72(n− 1)nR4x4

((λ1 + 2 (λ2 + 5λ3)) f 2

∞ − 3f∞ + 3))

+ 2F1

(3, n+ 1;n+ 2;−x2

)(8(n− 1)nR2x4(f 2

∞(λ1(15b2 + 328R2)− 2 (λ2 + 5λ3)(21b2 − 232R2

)) + 6

(3b2 − 46R2

)f∞ − 9b2 + 192R2)) + 2F1

(1, n+ 1;n+ 2;−x2

)((n− 1)nx4(−36b4 + 60b2R2 + f∞(45b4 − 84b2R2 + 2f∞(λ1(3b4 − 14b2R2 + 580R4)

− 2(λ2 + 5λ3)(15b4 − 34b2R2 − 340R4))− 840R4) + 600R4)) + 2F1

(2, n+ 1;n+ 2;−x2

)(−3(n− 1)nx4(2f 2

∞(λ1(b4 + 2b2R2 − 264R4)− 2(λ2 + 5λ3)(5b4 − 2b2R2 + 168R4))

+ 3(5b4 − 4b2R2 + 136R4)f∞ − 12(b4 − b2R2 + 24R4)))

].

(3.21)

For the sphere after doing the expansion around n = 1 and the remaining integrals ,

SEE = −4 a ln(f0

δ) , (3.22)

where we have used 1 = f∞ − 13f 2∞(λ1 + 2λ2 + 10λ3) and a is given in Eq. (2.19). Thus we

get the expected universal terms using the regularization proposed in [4].

3.2.2 New Massive Gravity

As an example for a calculation of generalized gravitational entropy in other dimensions, we

consider the New Massive Gravity action in three dimensions [7] and use the notation in [8]

S = − 1

2`P

∫d3x√g[R +

2

L2+ 4λL2(RµνR

µν − 3

8R2)].

Here 1− f∞ + f 2∞λ = 0. The entropy functional for this is not intrinsic as compared to the

three dimensional Einstein gravity and is given by

SEE =2π

`P

∫dx√gxx[1 + 4λL2([Rµνn

µsn

νs −

1

2KsKs]−

3

4R)]. (3.23)


The integral is over the one dimensional entangling region. s denotes the two transverse

direction. We calculate the generalized gravitational entropy following the same procedure

as used above. The two dimensional squashed cone metric is given by

ds2 = f(r, b)dr2 + r2dτ 2 .

f0 in this case also corresponds to the radius of the entangling surface.

In 3 dimensions [9, 10]

g(2)ij = − L

2

2R(0)g

(0)ij + tij (3.24)

Only divergence and trace of tij are known.

g(0)ij t

ij = R(0) , ∇itij = 0 .

R(0) = − 2b2 (n2 − 1)

(b2n2 + r2)2 . (3.25)

Using 1− f∞ + f 2∞λ = 0 and we get,

S = · · ·+ 1

2`P

∫dρ

ρ

∫ 2π

0

dτ

∫ r=f0

r=0

drL(rb2(n2 − 1)(1 + 2f∞λ)

f1/2∞√b2 + r2(b2n2 + r2)3/2

+ · · · . (3.26)

Note that tij does not enter in the calculation of the universal term. After doing the integrals

we get

S = · · ·+∫dρ

ρ

[πL (1 + 2f∞λ)

`P√f∞

1

n−

√b2 + f0

2

b2n2 + f02

]+ · · · . (3.27)

Then expanding around b = 0 and n = 1 we get the correct universal term

SEE =c

3ln(

f0

δ) , (3.28)

where, c3

= 2πL(1+2f∞λ)

f1/2∞ `P

.

3.2.3 Quasi-Topological Gravity

The six-derivative action for quasi-topological gravity is given below [11],

S = − 1

2`3P

∫d5x√g[R +

12

L2+L2λ

2GB +

L47µ

4Z5

](3.29)

where,

GB = RµνρσRµνρσ − 4RµνR

µν +R2 and

Z5 = RµνρσRν

ασβRα

µβρ +

3

8RµνρσR

µνρσR− 9

7RµνρσR

µνραR

σα +15

7RµνρσR

µρRνσ

+18

7RµσR

σαRµα −

33

14RαβR

αβR +15

56R3 .

(3.30)

76 3.3. COMMENT ABOUT SINGULARITIES IN THE METRIC

Following exactly the same procedure we can derive the holographic entanglement entropy

for this six derivative gravity theory and obtain the correct universal terms..

For the sphere we get,

SEE = − 4π2L3

f3/2∞ `3

P

(1− 6f∞λ+ 9f 2∞µ) ln(

f0

δ) . (3.31)

For the cylinder

SEE = − π2L3H

2f3/2∞ `3

PR(1− 2f∞λ− 3f 2

∞µ) ln(f0

δ) . (3.32)

3.2.4 α′3 IIB supergravity

The action for this follows from [12]

S = − 1

2`3P

∫d5x√g[R +

12

L2+ L6γκ5

](3.33)

where,

κ5 = CαβµνCρβµσCαδγ

ρCνδγσ −

1

4CαβµνC

αβρσC

µρδγC

νσδγ .

Cαβµν is the Weyl tensor in 5 dimensions. In the context of IIB string theory, γ = 18ζ(3)α′3/L6.

For this theory we find that the universal parts of EE do not get corrected compared to the

Einstein case. This is expected since from the perspective of the AdS/CFT correspondence,

the C4 correction correspond to 1/λ corrections and the anomalies are not expected to receive

such corrections. Recently the effect of the C4 correction on Renyi entropy was analysed in

[13].

3.3 Comment about singularities in the metric

There are singularities in the five dimensional metric coming entirely from g(2)ij . We expand

the metric around r = 0 . Upto the leading order the metric is shown below.

For the sphere (diagonal components are gρρ, grr, gττ , gθθ, gφφ ),

L2

4f∞ρ20 0 0 0

0 (n−1) cos(τ)L2

f0 rf∞+ 1

ρ0 0 0

0 0 r2

ρ− L2(n−1) r cos(τ)

f0f∞0 0

0 0 0 f02

ρ0

0 0 0 0 f02 sin2(θ)ρ

. (3.34)

REFERENCES 77

For the cylinder,

L2

4f∞ρ20 0 0 0

0 (n−1) cos(τ)L2

2f0 rf∞+ 1

ρ0 0 0

0 0 r2

ρ− L2(n−1) r cos(τ)

2f0f∞0 0

0 0 0 f02

ρ0

0 0 0 0 1ρ

. (3.35)

The grr component is singular in r. The other components are non singular. However it is

easy to see that the determinant does not have a singularity at r = 0. The singularity in the

metric gives rise to singularities in the components of the Riemann tensor. We have explicitly

checked that these singularities do not enter in the higher derivative actions considered in

this paper. Hence these are mild singularities in the sense used in [3]. Note that in order to

calculate the universal part of EE in four dimensions only g(2)ij is important.

3.4 Discussion

The newly proposed regularization in [4] yields the expected universal terms in the EE

in higher derivative gravity theories dual to four dimensional CFTs. We considered the

Fefferman-Graham metric with the regularized metrics in [4] as the boundary metric. Then

we computed the generalized gravitational entropy as proposed in [3]. The universal log

terms worked out to be as expected. We showed that upto the order we are interested in, the

singularities in the metric are mild. As pointed out in [3] we could also have done a conformal

transformation of the boundary metric with conical singularity such that it is non-singular

and then done the calculation. We expect the results to be identical. One can possibly

use other regularization scheme and compute EE using generalized entropy to obtain correct

results, but we will demonstrate the importance of this particular regularization in the next

chapter when we will try to connect EE with the Wald entropy formula.

References

[1] T. Barrella, X. Dong, S. A. Hartnoll and V. L. Martin, “Holographic entanglement

beyond classical gravity,” arXiv:1306.4682 [hep-th].

T. Faulkner, A. Lewkowycz and J. Maldacena, “Quantum corrections to holographic

entanglement entropy,” arXiv:1307.2892 [hep-th].

B. Swingle, L. Huijse and S. Sachdev, “Entanglement entropy of compressible holo-

graphic matter: loop corrections from bulk fermions,” arXiv:1308.3234 [hep-th].

78 REFERENCES



[3] A. Lewkowycz and J. Maldacena, “Generalized gravitational entropy,” arXiv:1304.4926

[hep-th].





th/9903033].

A. Buchel, J. Escobedo, R. C. Myers, M. F. Paulos, A. Sinha and M. Smolkin, “Holo-


[hep-th]].



[7] E. A. Bergshoeff, O. Hohm and P. K. Townsend, “Massive Gravity in Three Dimen-

sions,” Phys. Rev. Lett. 102, 201301 (2009) [arXiv:0901.1766 [hep-th]].

[8] A. Sinha, “On the new massive gravity and AdS/CFT,” JHEP 1006, 061 (2010)

[arXiv:1003.0683 [hep-th]].

[9] S. de Haro, S. N. Solodukhin and K. Skenderis, “Holographic reconstruction of space-

time and renormalization in the AdS / CFT correspondence,” Commun. Math. Phys.

217 (2001) 595 [hep-th/0002230].

[10] K. Skenderis and S. N. Solodukhin, “Quantum effective action from the AdS / CFT

correspondence,” Phys. Lett. B 472 (2000) 316 [hep-th/9910023].

[11] R. C. Myers and B. Robinson, “Black Holes in Quasi-topological Gravity,” JHEP 1008,

067 (2010) [arXiv:1003.5357 [gr-qc]].

R. C. Myers, M. F. Paulos and A. Sinha, “Holographic studies of quasi-topological

gravity,” JHEP 1008, 035 (2010) [arXiv:1004.2055 [hep-th]].

[12] D. J. Gross and E. Witten, “Superstring Modifications of Einstein’s Equations,” Nucl.

Phys. B 277, 1 (1986).

M. B. Green and C. Stahn, “D3-branes on the Coulomb branch and instantons,” JHEP

0309, 052 (2003) [hep-th/0308061].

M. F. Paulos, “Higher derivative terms including the Ramond-Ramond five-form,”

REFERENCES 79

JHEP 0810, 047 (2008) [arXiv:0804.0763 [hep-th]].

S. S. Gubser, I. R. Klebanov and A. A. Tseytlin, “Coupling constant dependence in

the thermodynamics of N=4 supersymmetric Yang-Mills theory,” Nucl. Phys. B 534,

202 (1998) [hep-th/9805156].

A. Buchel, J. T. Liu and A. O. Starinets, “Coupling constant dependence of the shear

viscosity in N=4 supersymmetric Yang-Mills theory,” Nucl. Phys. B 707, 56 (2005)

[hep-th/0406264].

R. C. Myers, M. F. Paulos and A. Sinha, “Quantum corrections to eta/s,” Phys. Rev.

D 79, 041901 (2009) [arXiv:0806.2156 [hep-th]].

A. Buchel, R. C. Myers, M. F. Paulos and A. Sinha, “Universal holographic hydrody-

namics at finite coupling,” Phys. Lett. B 669 (2008) 364 [arXiv:0808.1837 [hep-th]].

W. H. Baron and M. Schvellinger, “Quantum corrections to dynamical holographic ther-

malization: entanglement entropy and other non-local observables,” arXiv:1305.2237

[hep-th].

[13] D. A. Galante and R. C. Myers, “Holographic Renyi entropies at finite coupling,” JHEP

1308 (2013) 063 [arXiv:1305.7191 [hep-th]].

80 REFERENCES

4 Connection between entanglement

entropy and Wald entropy

4.1 Introduction

As discussed in the Chapter.(2), there is a striking similarity between black hole entropy

and entanglement entropy (EE) in Einstein gravity. In the black hole case, there exists a

simple generalization of the area law for calculating the entropy of a black hole in any general

higher-derivative gravity theory, known as the Wald entropy [1, 2, 3]. It is by now well known,

in an arbitrary theory of gravity, taking the Wald entropy functional in AdS space will give

rise to the wrong universal terms in EE [4, 5]. A prescription is given for evaluating EE

in [6, 7], suitably modifying the Wald entropy functional for general surfaces based on the

generalized entropy principle. In the Chapter.(2) of this thesis we have shown that, for any

arbitrary theory of gravity it is not possible to derive those conjectured entropy functionals

[8]. On the other hand, based on the Noether charge prescription one can derive the Wald

entropy functionals, there by increasing its geometrical significance. In this chapter our main

objective is to explore the possibility of connecting EE with the Wald entropy as that we will

give us a concrete holographic derivation of EE based on Noether charge method, enabling

us to avoid the conflicts as pointed out in Chapter.(2).

In this chapter we will consider the background constructed out of the squashed cone

metric in Eq. (3.3) and evaluate the Wald entropy for various theories of gravity in this

background on a particular surface. This will produce the correct universal terms in EE.

After that we will comment on the connection with Iyer-Wald prescription for computing

entropy for black holes with dynamical horizon, opening up a possibility of connecting EE

rigorously with the Noether charge.

In the Chapter.(3) we have shown that if we evaluate generalized entropy [9] in the

squashed cone background we will get the correct universal terms for EE. In this chapter it

will be clear why that particular background is important, as that will play a crucial role in

our quest to connect EE with the Wald entropy.

81

82 4.2. WALD ENTROPY

4.2 Wald Entropy

Before proceeding further, let us mention the key steps in Wald’s derivation of the entropy

functionals for black holes for general theories of gravity [2]. Let us consider a diffeomorphism

invariant lagrangian L. We denote all the dynamical fields collectively by φ. In general φ

carries indices depending on the nature of the fields present. Varying L with respect to φ we

get, ∫ddxδL =

∫ddxE. δφ+

∫ddx∇µΦµ (4.1)

First piece gives the equation of motion and the second piece is a surface term. Now one can

construct a Noether current out of this surface term as follows

Jµ = Φµ(φ,Lζφ)− ζµL, (4.2)

where ζµ denotes the killing vectors associated with the diffeomorphism invariance. Using

the equation of of motion one can write down,

J = dQ (4.3)

where Q is the Noether charge. At this stage several ambiguities enter in the calculation.

For example, one can add a closed form to Q to define a new charge Q→ Q+ dΨ such that,

J = dQ (4.4)

still holds. Also one can add a total derivative term to L which will not affect the equation

of motion but it will change the expression for Q such that Eq. (4.3) still holds. For more

detailed discussion of these ambiguities interested readers are referred to the original reference

[2].

But it was shown by Iyer and Wald in the [2] that all these ambiguities vanish on a

bifurcation surface and one can write down a first law like relation for the entropy.

δ

∫Q =

κ

2πδSwald (4.5)

where κ is the surface gravity related to the temperature of the black hole and the charge is

integrated over the horizon of the black hole. From this we one can read of the Wald entropy

solely in terms of geometrical quantities.

Swald =

∫dd−2x

√h

∂L

∂Rαβγδ

εαβ εγδ . (4.6)

This expression is evaluated on the black hole horizon, which is a codimension-2 surface.

Here εαβ = n1αn

2β − n2

αn1β is the binormal corresponding to the two transverse directions 1, 2 .

CHAPTER 4. CONNECTION BETWEEN ENTANGLEMENT ENTROPY AND WALDENTROPY 83

One crucial point is that, Wald formula is applicable only on a bifurcation surface. For

a static black hole, the horizon is a bifurcation surface for which the extrinsic curvatures

vanish. For surfaces with non zero extrinsic curvature one cannot apply this formula. Now

let us see what it gives for our case. We will evaluate Eq. (4.6) on the background constructed

in Eq. (3.3) on a codimension-2 surface r = 0 , τ = 0. Rest of this chapter is mainly based

on the work done with Prof. Aninda Sinha and Dr. Menika Sharma [10].

4.3 Four derivative theory

The Wald entropy is calculated from Eq. (3.8). We have to first evaluate the following,

∂L

∂Rαβγδ

=1

2(gαγgβδ − gαδgβγ) + L2

[λ1R

αβγδ +1

4λ2

(gβδRαγ − gβγRαδ − gαδRβγ + gαγRβδ

)+

1

2λ3R

(gαγgβδ − gαδgβγ

) ](4.7)

and after some simplifications we get,

Swald =2π

`3P

∫d3x√h(1 +

L2

2(2λ3R + λ2Rµνn

νrn

rµ + 2λ1Rµνρσnµrn

νsn

rρnsσ)). (4.8)

In this section we will show that starting with the boundary metrics in Eq. (3.2) we can

construct a bulk spacetime on which Swald will produce the expected universal parts for the

entanglement entropy for both cylindrical and spherical region. Note that (4.8) differs from

(2.11) by the O(K2) terms.

4.3.1 Cylinder

As we will show, a particular form of the regularization b = α(n − 1)1/2 appearing in the

Eq. (3.2), where α is some number which we will determine later (it will turn out to be surface

dependent but theory independent), is needed to get the correct universal term. Recall that

the only restriction on b was that f(r, b) has to be n2 in the r = 0 limit. However, in

holographic calculations we expect that the bulk metrics will only depend on the AdS radius,

the radius of the entangling region and n. As such we can expect that the only way that

b→ 0 would arise in holographic calculations is such that b is some positive power of (n−1).

Now we will evaluate Eq. (4.8) using Eq. (3.3) using the cylinder metric to be its boundary.

Then we extract the coefficient of the 1ρ

term. We set τ = 0 . There is no integral over r

84 4.3. FOUR DERIVATIVE THEORY

in the Wald entropy as the entangling surface is located at r = 0, τ = 0 . We put r = b x .

After that we expand around x = 0 and then expand around n = 1 . We retain only the n

independent part as other terms vanish in n → 1 limit. Below we quote some intermediate

steps after expanding in ρ, r and n respectively. It is important to take the limits in r, n in

that particular in order to get the correct result [11]. After doing the ρ expansion we pick

out the 1ρ

term of (4.8) which is shown below.

Swald = · · ·+ 2π

`3P

∫dρdφdz

A(x, n)

ρ+O(ρ) + · · · , (4.9)

where

A(x, n) =L3 (n2 − 1) d−n ((4λ2 + 20λ3 − 2λ1) f∞ − 1) (2f0d

n − d (n2 + n+ x2 − 2) (bx)n)

24b2f3/2∞ (n2 + x2)2

.

Then expanding A(x, n) around x = 0 we get,

A(x, n) =L3 (n2 − 1) f0 ((4λ2 + 20λ3 − 2λ1) f∞ − 1)

12 b2n4f3/2∞

+ · · · . (4.10)

If

b =2f0√

3

√n2 − 1β(n) ,

where β(1) = 1 we get upon further expanding A(x, n) around n = 1

A(x, n) = −L3 (1 + 2 (λ1 − 2 (λ2 + 5λ3)) f∞)

16f0f3/2∞

+O(n− 1) + · · · . (4.11)

Notice that the choice for b was independent of the theory, i.e., in this case of λi’s. Finally

we get,

Swald = −π2L3H(1 + 2f∞(λ1 − 2λ2 − 10λ3))

2f3/2∞ `3

Pf0

ln(f0

δ) . (4.12)

This is precisely what is expected.

4.3.2 Sphere

We proceed similarly for the sphere case. First we expand in ρ and pick out the 1ρ

term.

Swald = · · ·+ 2π

`3P

∫dρdθdφ

A(x, n)

ρ+O(ρ) + · · · . (4.13)


Here

A(x, n) =L3d−2n sin(θ)

12b2f 3/2x2(n2 + x2)2

[4λ1f∞(b2x2d2n(n2 + x2)2 − d2 (n4 − n3x2 + 3n2x2 + nx2

+ x4 − x2)(bx)2n + d (n2 − 1)Rx2(n2 + n+ x2 − 2)(b dx)n − (n2 − 1)R2x2d2n)

− (2(λ1 + 2(λ2 + 5λ3))f∞ − 1)(−2b2x2d2n (n2 + x2)2 + d2(n4(3x2 + 2) + n3x2

+ 3n2x4 − nx2 − x4 + x2)(bx)2n + d (n2 − 1)Rx2(n2 + n+ x2 − 2)(b d x)n

− (n2 − 1)R2x2 d2n)]

(4.14)

Then expanding A(x, n) around x = 0 we get 1,

A(x, n) =L3 sin(θ)

(2b2n4 (4 (λ1 + λ2 + 5λ3) f∞ − 1) + (n2 − 1) f0

2 ((−2λ1 + 4λ2 + 20λ3) f∞ − 1))

12b2f3/2∞ n4

.

(4.15)

Only the x independent term is shown. If (for consistency checks see below)

b = f0

√n2 − 1β(n) (4.16)

where β(1) = 1, expanding around n = 1 we get,

A(x, n) = −L3 sin(θ) (1− 2 (λ1 + 2 (λ2 + 5λ3)) f∞)

4f3/2∞

+O(n− 1) + · · · . (4.17)

As in the cylinder case, notice that the choice for b is theory independent. Finally we get,

Swald = −4π2L3(1− 2f∞(λ1 + 2λ2 + 10λ3))

f3/2∞ `3

P

ln(f0

δ) (4.18)

We have fixed b for both the cylinder and the sphere case. In all the subsequent calculations

of Wald entropy we will use these same values for b.

1Remember that at this stage n = 1 + ε. Thus we will drop x2n compared to x2.


4.4 Quasi-Topological gravity

The Wald entropy is calculated for (3.29) using (4.6) . For this case,

∂L

∂Rαβγδ

=1

2(gαγgβδ − gαδgβγ) + L2

[λ1R

αβγδ +1

4λ2

(gβδRαγ − gβγRαδ − gαδRβγ + gαγRβδ

)+

1

2λ3R

(gαγgβδ − gαδgβγ

) ]+

7µL4

4

[(3µ1(RαργσRβ δ

ρ σ −RαρδσRβ γρ σ)) +

µ2

2[(gαγgβδ − gαδgβγ)

RµνρσRµνρσ + 4RRαβγδ] +

µ3

4[gβδRαρσµRγ

ρσµ − gβγRαρσµRδρσµ − gαδRβρσµRγ

ρσµ + gαγRβρσµRδρσµ

− 2RγρRαβδρ + 2RδρRαβγ

ρ + 2RβρRαργδ − 2RαρRβ

ργδ] +

µ4

2(Rρσ[gβδRα

ργσ − gβγRα

ρδσ

− gαδRβργσ + gαγRβ

ρδσ] + [RαγRβδ −RαδRβγ]) +

3µ5

4[gβδRασRγ

σ − gβγRασRδσ

− gαδRβσRγσ + gαγRβσRδ

σ] +µ6

2

[R(gβδRαγ − gβγRαδ + gαγRβδ − gαδRβγ

)+ (gαγgβδ − gαδgβγ)RµνR

µν]

+3

2µ7(R2[gαγgβδ − gαδgβγ])

].

(4.19)

Now the coefficients are,

µ1 = 1 , µ2 =3

8, µ3 = −9

7, µ4 =

15

7, µ5 =

18

7, µ6 = −33

14, µ7 =

15

56,

and λ2 = −4λ1, λ3 = λ1 = λ. Proceeding similarly as mentioned for the R2 theory we get

the expected universal terms.

For the cylinder 2,

Swald = − π2L3H

2f3/2∞ `3

PR(1− 2f∞λ− 3f 2

∞µ) ln(f0

δ) . (4.20)

For the sphere,

Swald = − 4π2L3

f3/2∞ `3

P

(1− 6f∞λ+ 9f 2∞µ) ln(

f0

δ) . (4.21)

Again note that the choice for α did not depend on the theory.

2The c and a coefficients for an arbitrary higher derivative theory can be easily calculated using the

short-cut mentioned in the appendix of [12].


4.5 α′3 IIB supergravity

The Wald entropy is calculated for (3.33) using (4.6) . For this case,

∂L

∂Rαβγδ

=1

2(gαγgβδ − gαδgβγ) + L6γ

[13

(gβγCαµδνCνρσηCµρση − gβδCαµγνCνρσηCµ

ρση

+ gαδCβµγνCνρσηCµρση − gαγCβµδνCνρσηCµ

ρση) +1

6(gαγgβδ − gαδgβγ)(CσµνρCσηνζCηρζµ

− 1

2Cµν

ρσCµνηζCηρζσ) +1

6(gβδCαρζσCρσµνC

γζµν − gαδCβρζσCρσµνC

γζµν

− gβγCαρζσCρσµνCδζµν + gαγCβρζσCρσµνC

δζµν) +

1

6(gβδCαρζσCγµ

ρνCζσµν

− gαδCβρζσCγµρνCζσµν − gβγCαρζσCδµ

ρνCζσµν + gαγCβρζσCδµ

ρνCζσµν)

+ (CαρµσCβµδηCγ

ρησ − CβρµσCαµδηCγ

ρησ − CαρµσCβµγηCδ

ρησ + CβρµσCαµγηCδ

ρησ)

− 1

2(CγδσζCβ

ζµρCασµρ + CαβσζCδ

ζµρCγσµρ) +

2

3(gαδCβρζνCρσνµC

γµζσ

− gβδCαρζνCρσνµCγµζσ + gβγCαρζνCρσνµC

δµζσ − gαγCβρζνCρσνµC

δµζσ)].

(4.22)

Proceeding similarly as mentioned for the R2 theory we get the expected universal terms.

For the cylinder,

Swald = −π2L3H

2`3PR

ln(f0

δ) . (4.23)

For the sphere,

Swald = −4π2L3

`3P

ln(f0

δ) . (4.24)

As expected, for this case the universal terms are independent of the higher derivative cor-

rection.

4.6 Connection with Ryu-Takayanagi

The Ryu-Takayanagi calculation involves the minimization of an entropy functional3. For

both the sphere and the cylinder, one can check that minimizing the Wald area functional in

the Fefferman-Graham background for squashed cones leads to the correct universal terms

provided we choose b as mentioned above. Recall that the Wald entropy functional in AdS

spacetime was not the correct one [4, 5]. However, our background is not AdS and it turns

out that the Wald entropy functional leads to the correct universal terms. We show this for

3We thank Rob Myers for discussions on this section.

88 4.6. CONNECTION WITH RYU-TAKAYANAGI

the cylinder, the sphere case working similarly. Putting r = R(ρ) = r0 + r1ρα around ρ = 0

leads to r0 = 0 and the equation

cnrn1ραn+1 − 4r2

1Rcnα(α− 2)ρ2α = 0 ,

where we have shown the leading terms which would contribute around n = 1. If we set

n = 1 we recover the result α = 1, r1 = −1/(4f0) for a cylinder–this is expected. The n = 1

boundary geometry is just flat space with the dual bulk being AdS. Hence we expect to

recover the RT result. However if n = 1 + ε, then it is easy to see that either r1 = 0 or α = 2

or r1 = −1/(4f0) and α = 1 + ε. As in the RT case, only the linear term in R(ρ) would have

affected the universal term–since α 6= 1 if n = 1 + ε we find that there is no linear term.

For n 6= 1 the minimal surface is at r = 0 = τ. This is the reason why the Wald entropy on

the r = 0 = τ surface and the RT entropy functional approach give the same result for the

universal terms in the squashed cone background. We now point out a direct comparison

between the calculation done in AdS spacetime and that in the squashed cone background

for the sphere in what follows.

The Ryu-Takayanagi prescription was implemented in the following way for a spheri-

cal entangling surface. Consider the AdS5 metric with the boundary written in spherical

coordinates

ds2 =L2

z2(dz2 + dt2 + dr2 + r2dθ2 + r2 sin2 θdφ2) . (4.25)

Now put r = f(z) = f0 + f2z2 + · · · and t = 0 and minimize the relevant entropy functional.

Implicitly our analysis says that this surface and the r = 0 = τ surface in the coordinate

system we have been using are related. Since in both cases the extrinsic curvatures vanish

we can attempt to make a direct comparison. In order to do this we make a coordinate

transformation:dz

z

√1 + f ′(z)2 =

dρ

2ρ. (4.26)

Around ρ = 0 we will find z2 = ρ− 2f 22ρ

2 + · · · and f(z)2/z2 = f02/ρ+ 2f0f2(1 + f0f2) + · · · .

Now around ρ = 0, the metric on the r = 0 = τ surface takes the form

ds2 = L2[dρ2

4ρ2+K(ρ)(dθ2 + sin2 θdφ2)] , (4.27)

where

K(ρ) =f0

2

ρ− L2

6b2n4(2b2n4 + (n2 − 1)f0

2) .

This also shows that for n 6= 1 minimal surface is at r = 0 = τ . Now choosing b as in

Eq. (4.16), expanding upto O((n − 1)0) and comparing with the RT calculation we find

f2 = −1/(2f0). This is exactly what we would have got if we minimized the RT area

functional (or the relevant higher derivative entropy functional) in AdS space. This also

serves as a consistency check for the choice of b.


4.7 Comments on the connection with the Iyer-Wald

prescription

Why does the Wald entropy functional lead to the correct result in our case? Wald’s formula

in Eq. (4.6) is valid for a surface which is a local bifurcation surface on which the Killing field

vanishes. For a bifurcation surface, the extrinsic curvatures vanish. SEE mentioned in (2.11)

differs from Swald only by the extrinsic curvature terms. The Noether charge method of [1]

needs a bifurcation surface to remove various ambiguities [2, 3]. According to the prescription

of Iyer and Wald [2], in order to compute the entropy for horizons which are not bifurcate,

e.g., dynamical horizons, the curvature terms in ∂L∂Rabcd

are replaced by their boost invariant

counterparts [2]. To do this we have to construct a boost invariant metric from our original

metric. Let gab be our starting d dimensional metric with the two normals n1a, n

2b . The boost

invariant part of gab will only have terms with the same number of n1, n2. We then consider

a d − 2 dimensional surface and find a neighbourhood of it O such that for any points x

belonging to this neighbourhood, we can find a point P which lies on a unit affine distance

on a geodesic with a tangent vector va on the d− 2 dimensional plane perpendicular to this

surface under consideration. Now we assign a coordinate system U, V, x1, ...xd−2 for the point

x where U, V are the components of va along n1a and n2

a. A change of normals under the

boosts na1 → αna1, nb2 → α−1na2 will change the coordinates as follows U → αU, V → α−1V .

Now we Taylor expand gab around Uand V ,

gab = g(0)ab + U∂g + V ∂g + UV ∂∂g + ........... . (4.28)

We have shown the expansion schematically. Under boosts, the terms linear in U, V do not

remain invariant. The prescription in [2] is to drop these terms. The UV term is invariant

under the boost. One important point to note is that , ψa = U( ∂∂U

)a−V ( ∂∂V

)a is a Killing field

of the metric. This means that Lie derivative of gab with respect to ψ is zero. Effectively,

we have constructed a new spacetime in which the original dynamical horizon becomes a

bifurcate Killing horizon.

The evidence for the existence of this bifurcation surface would be that extrinsic curva-

tures for this surface in the bulk background vanishes. Our entangling surface is a codimension-

2 surface. Now we calculate the extrinsic curvatures for this surface in the bulk Fefferman-

Graham metric. There will be two of them—one along the direction of the normal (τ)n for

τ = 0 and the other one along the normal (r)n for r = 0. We start with the 5 dimensional

metrics given in Eq. (3.3). The non-zero components of the normals are

(τ)nτ =1√gττ

, (r)nr =1√grr

.

With these we calculate the two extrinsic curvatures (τ)Kµν and (r)Kµν . Then we put r = b x

90 4.8. UNIVERSALITY IN RENYI ENTROPY

and τ = 0 as before. As the entangling surface is located at r = 0, τ = 0 we further do an

expansion around x followed by an expansion in n. Now (τ)Kab = 0 whereas (r)Kab = A(x, n, ρ)

is some function of x , n and ρ . First we expand it around x = 0 and then we do an expansion

around n = 1 . We find that (r)Kab = 0 .

Thus effectively the Fefferman-Graham construction is the same as the Iyer-Wald pre-

scription, provided we take the limits in the manner prescribed in [11]. The replacement of

rKijdxidxj by rnKijdx

idxj plays a key role in this construction. Recall that this was needed

to keep the boundary Ricci scalar finite. Also another important point to notice that for

the squashed cone metric there is no time like killing vector as the metric components are

dependent on τ . The Wald-Iyer prescription calls for calculating the Wald functional in the

context of black hole entropy where there exists a time like killing vector. But in the metric

(3.2) the cos(τ) factor which breaks the time translational symmetry is accompanied by a

factor of rn . In our calculation we have taken the r → 0 limit first and then the n→ 1 limit.

Thus the cos(τ) multiplied by rn is suppressed in this way of taking limits. For this reason

we have an approximate time-translational symmetry in our new space time.

Upto this point the discussion is independent of the choice of b. Now when one wants to

evaluate the Wald entropy functional with this squashed cone metric one needs to specify b

as mentioned in the previous sections for the sphere and the cylinder to obtain the correct

universal terms. As there is no integral over r in the Wald entropy functional, the final result

obtained will be b dependent as we have found and hence we have to choose b accordingly.

4.8 Universality in Renyi entropy

Before closing out this chapter let us mention an interesting application of this generalized

entropy. In [13, 14, 15, 16] it was shown that for spherical entangling surfaces in four dimen-

sions the Renyi entropy has a universal feature. Namely

∂nSn|n=1 ∝ cT .

In four dimensions cT ∝ c, the Weyl anomaly. If we use Eq. (4.15) and identify it as the

expression for Sn with the choice for b given below it4, then we indeed find that this is true!

This also works for the six and eight derivative examples. Thus this approach enables us to

check some information away from n = 1. Further, as a bonus, we can predict what happens

in the case of a cylindrical entangling surface where holographic results for the Renyi entropy

are not available. If we use Eq. (4.10) or its analog for the six and eight derivative examples,

we find that ∂nSn|n=1 ∝ cT still holds. It will be interesting to explicitly verify this in field

theory.

4In order to get the proportionality constant to work out, we will need to adjust ∂nβ(n)|n=1 in b.

REFERENCES 91

4.9 Discussion

In this chapter we computed the Wald entropy on the r = 0 = τ co-dimension 2 surface in

the Fefferman-Graham metric and found that it gives the correct universal terms for both

spherical and cylindrical surfaces. In order to get the expected results, we needed to choose

a surface dependent but theory independent regularization parameter. Recall that in bulk

AdS space, from the entropy functional way of computing EE in general theories of gravity,

one needed to use the entropy functional proposed in [6, 7] , which differed from the Wald

entropy functional by extrinsic curvature terms as shown in the Chapter.(2). These extrinsic

curvature terms are important to get the correct universal piece for any entangling surface

with extrinsic curvature.

Whether EE can be thought of as a Noether charge needs further investigation. Our

findings in this chapter seems to suggest that this may indeed be true. The Fefferman-Graham

metric is the analog of the Iyer-Wald metric used to compute the entropy for dynamical

horizons in [2]. Our conjecture then is that the Wald entropy (after appropriately fixing the

regularization) evaluated on the r = 0 = τ co-dimension two surface in the Fefferman-Graham

metric is going to capture the expected universal terms for any entangling surface.

Recently it was pointed out in the [17], that the ambiguities in the Nother charge method

can be fixed by demanding that the resulting entropy functional satisfies a generalized sec-

ond law [18] when evaluated on any arbitrary surface. Also it has been shown that, those

ambiguities correspond to the extrinsic curvature dependent terms of the proposed entropy

functionals as shown in Eq. (2.6) upto O(K2). This opens up the possibility of deriving these

holographic entropy functionals [6, 7] directly from the Noether charge method and that will

also solidify our conjecture of connecting EE with the Noether charge.

References


[gr-qc/9307038].



[gr-qc/9503052].

[2] V. Iyer and R. M. Wald, “Some properties of Noether charge and a proposal for dy-


[3] T. Jacobson and R. C. Myers, “Black hole entropy and higher curvature interactions,”


92 REFERENCES

T. Jacobson, G. Kang and R. C. Myers, “On black hole entropy,” Phys. Rev. D 49

(1994) 6587 [gr-qc/9312023].






JHEP 1401, 044 (2014) [arXiv:1310.5713 [hep-th]].

[7] J. Camps, “Generalized entropy and higher derivative Gravity,” JHEP 1403, 070 (2014)

[arXiv:1310.6659 [hep-th]].



[9] A. Lewkowycz and J. Maldacena, “Generalized gravitational entropy,” JHEP 1308,

090 (2013) [arXiv:1304.4926 [hep-th]].





[12] K. Sen, A. Sinha and N. V. Suryanarayana, “Counterterms, critical gravity and holog-

raphy,” Phys. Rev. D 85 (2012) 124017 [arXiv:1201.1288 [hep-th]].

[13] D. A. Galante and R. C. Myers, “Holographic Renyi entropies at finite coupling,” JHEP

1308, 063 (2013) [arXiv:1305.7191 [hep-th]].

[14] L. -Y. Hung, R. C. Myers, M. Smolkin and A. Yale, “Holographic Calculations of Renyi

Entropy,” JHEP 1112, 047 (2011) [arXiv:1110.1084 [hep-th]].

[15] E. Perlmutter, “A universal feature of CFT Renyi entropy,” arXiv:1308.1083 [hep-th].

[16] L. Y. Hung, R. C. Myers and M. Smolkin, “Twist operators in higher dimensions,”

JHEP 1410 (2014) 178 [arXiv:1407.6429 [hep-th]].

[17] S. Bhattacharjee, S. Sarkar and A. Wall, “The holographic entropy increases in

quadratic curvature gravity,” arXiv:1504.04706 [gr-qc].

A. C. Wall, “A Second Law for Higher Curvature Gravity,” arXiv:1504.08040 [gr-qc].

REFERENCES 93

[18] A. C. Wall, “Ten Proofs of the Generalized Second Law,” JHEP 0906 (2009) 021

[arXiv:0901.3865 [gr-qc]].

S. Sarkar and A. C. Wall, “Generalized second law at linear order for actions that

are functions of Lovelock densities,” Phys. Rev. D 88 (2013) 044017 [arXiv:1306.1623

[gr-qc]].

94 REFERENCES

5 Constraining gravity using

entanglement entropy

5.1 Introduction

One can constrain gravity in several interesting ways by using quantum entanglement. In

this chapter we focus on one of them and in the next chapter we discuss another one. From

Chapter.(2) and Chapter.(3) it is evident that the extremal surface plays a crucial role in

evaluating entanglement entropy (EE), in fact the entropy functionals have to be evaluated on

the extremal surface to get the correct universal terms. In this chapter we will demonstrate

that in the context of higher derivative gravity theories, one can constrain the coupling

of the higher derivative terms by demanding the smoothness of the extremal surface. More

specifically, we will derive constraints on the Gauss-Bonnet (GB) coupling by demanding that

the entangling surface for sphere, cylinder and the slab close off smoothly in the bulk. The

slab case was considered before in the Ref. [1]. The GB action is given in the Eq. (2.8) with

λ1 = λ3 = λ and λ2 = −4λ and the corresponding entropy functional is the Jacobson-Myers

(JM) functional as shown in the Eq. (2.42). At the onset note that treating the truncated

GB gravity on its own leads to problems with entanglement entropy as was pointed out in

the Ref. [1]. In particular if we consider an entangling surface that topologically looks like

M2 × R, then the R term in the JM entropy functional becomes topological. Adding more

handles to the entangling surface will allow us to lower the entanglement entropy arbitrarily

if λ > 0. Since this particular sign of λ happens to arise in many consistent examples

in string theory (see for eg.[2]), this hints at a problem in interpreting GB gravity on its

own as a model for theories describing c 6= a–of course, there is no reason to suspect any

inconsistencies if this is just the first perturbative correction in an infinite set of higher

derivative corrections. We will not have anything to add to this observation. We will simply

focus on what constraints arise on the GB coupling by demanding smoothness and compare

the result with the causality/positive energy constraints [3, 4, 5] as shown in Eq.(5.1).

− 7

36≤ λ ≤ 9

100. (5.1)

To elaborate a bit more, we note at this point that the GB coupling λ is bounded. Following

the Refs. [3, 4] for the calculation of the three point correlation function of stress tensor one

95

96 5.2. SMOOTHNESS OF ENTANGLING SURFACE

needs to compute a energy flux which comes form the insertion of εijTij , where εij and Tij

are the polarization tensor and stress tensor respectively. Demanding the positivity of this

energy flux in the holographic set up we get the following three constraints and from those

we obtain bounds on λ . These coincide with the bounds arising from micro-causality [6].

Tensor channel : 1− 10f∞λ ≥ 0⇒ λ ≤ 9

100

Vector channel : 1 + 2f∞λ ≥ 0⇒ −3

4≤ λ ≤ 1

4

Scalar channel : 1 + 6f∞λ ≥ 0⇒ − 7

36≤ λ ≤ 1

4

(5.2)

From this we get

− 7

36≤ λ ≤ 9

100.

This is the same as the condition 1− 4f∞λ− 60f 2∞λ

2 ≥ 0.

Rest of this chapter is based on the work done with Dr. Shamik Banerjee, prof, Aninda

Sinha, Apratim Kaviraj and Kallol sen [7].

5.2 Smoothness of entangling surface

The general strategy we will adopt is the following. The entangling surface equation follows

from Eq.(2.43), coming from minimizing JM functional. We will consider 5 dimensional AdS

spacetime.

ds2 =L2(dz2 + dt2 + dΣ2)

z2(5.3)

and choose a constant time slice t = 0.

When we consider a spherical entangling , we parametrize dΣ in terms of spherical polar

coordinates.

dΣ2 = dr2 + r2(dθ2 + sin(θ)2dφ2) (5.4)

and the extremal surface is given by r = f(z). We will consider a cylindrical surface also.

For that we parametrize dΣ in terms of cylindrical polar coordinates,

dΣ = du2 + dr2 + r2dθ (5.5)

where u is the coordinate along the length of the cylinder and the extremal surface is charac-

terized again by r = f(z). For the slab, we simply write dΣ in cartesian coordinate and the

corresponding extremal surafce is given by x = f(z), where x is one of the coordinates of Σ.

Now let us assume that the surface f(z) closes off smoothly at z = zh inside the bulk

spacetime. Around this point, let us assume

f(z) =∞∑i=0

ci(zh − z)α+i . (5.6)

CHAPTER 5. CONSTRAINING GRAVITY USING ENTANGLEMENT ENTROPY 97

We need to determine α and ci’s. At z = zh, f′(z) → +∞ since the tangent to the surface

will be perpendicular at that point. This means that 0 < α < 1 and c0 > 0. Using these two

conditions, we will find that λ will be bounded.

Cylinder

Consider the cylinder case first. In cylindrical coordinates, assume the required hypersurface

to have the form r = f(z). From Eq. (6.72), we get the following equation,[zf ′′(z)

(6f∞λzf

′(z) + f(z)((4f∞λ+ 1)f ′(z)2 − 2f∞λ+ 1

))−(f ′(z)2 + 1

)(f ′(z)

(z(4f∞λ+ 1)f ′(z) + 3f(z)

(f ′(z)2 − 2f∞λ+ 1

))− 2f∞λz + z

)]= 0 .

(5.7)

We take the trial solution Eq. (5.6) and determine an appropriate α. We obtain α = 1/2, 3/2.

We will drop the second solution since this will lead to a conical tip. Expanding the eom in

powers of (zh − z) and setting the leading order term to 0, we get 4 roots of c0. We take the

two positive ones, √2

3

√zh(1 + 4f∞λ±

√1− 10f∞λ+ 16f 2

∞λ2) . (5.8)

With f∞ = (1−√

1− 4λ)/2λ, this puts some constraints on λ. Since the bottom sign vanishes

in the λ→ 0 limit, we will ignore this solution. For the other case, we have

λ ≤ 7

64. (5.9)

The quantities inside the square root have to be positive to make the root real. If we look

carefully we will find that 1 − 10f∞λ + 16f 2∞λ

2 has to be positive. This is almost same as

that of the tensor channel constraint except for the extra additional factor of 16f 2∞λ

2 . That

is why we get a bigger bound instead of λ< 9100

.

Sphere

The eom reads,[zf ′′(z)

(12f∞λzf(z)f ′(z) + f(z)2

((4f∞λ+ 1)f ′(z)2 − 2f∞λ+ 1

)+ 6f∞λz

2)−(f ′(z)2 + 1

)(6f∞λz

2f ′(z) + 2zf(z)((4f∞λ+ 1)f ′(z)2 − 2f∞λ+ 1

)+ 3f(z)2f ′(z)

(f ′(z)2 − 2f∞λ+ 1

)) ]= 0 .

(5.10)

We get only α = 1/2 as a solution to the indicial equation. We get six roots of c0 from the

leading order of eom. Three of them are positive:

√2zh ,

√4f∞zhλ± 2

√2zh√f∞λ(−1 + 2f∞λ) . (5.11)

The positivity of the first root cannot give us any constraint on λ. The other two roots go

to zero as λ goes to zero so we will ignore them.

98 5.3. DISCUSSION

Slab

The eom reads,

− 3(1− 2f∞λ+ f ′(z)2)(f ′(z) + f ′(z)3) + z(1− 2f∞λ+ (1 + 4f∞λ)f ′(z)2)f ′′(z) = 0 (5.12)

We get α = 1/2, 1 which give non-zero c0. Arguing as before we will only consider α = 1/2.

Here we get the following positive solution for c0:

c0 =

√2

3

√zh + 4f∞zhλ . (5.13)

Demanding this to be positive, we get

− 5

16≤ λ ≤ 1

4. (5.14)

This agrees with [1]. Thus together with the constraints from the cylinder we have

− 5

16≤ λ ≤ 7

64. (5.15)

We can recast this inequality as one for a/c where a, c are the Euler and Weyl anomaly

coefficients respectively for a 4d CFT. This gives us

1

3≤ a

c≤ 5

3. (5.16)

Quite curiously, the lower bound 1/3 is precisely what appears in non-supersymmetric

theories in the Refs. [3, 4, 5], in particular for a free boson. The upper bound of 5/3 cor-

responds to a free theory with one boson and two vector fields. For a non-supersymmetric

theory, the bound on a/c worked out1 in the Refs. [3, 4] was 1/3 ≤ a/c ≤ 31/18. Just to

point out in words, the 1/3 came from the cylinder calculation while the 5/3 came from the

slab. The causality constraints on the other hand translates into 1/2 ≤ a/c ≤ 3/2.

We compare the different bounds on λ in Fig. (5.1). As is clear, the causality constraints

are the tightest.

5.3 Discussion

We have considered different entangling surfaces and demanded that these close off smoothly

in the bulk. In Gauss-Bonnet gravity, this led to the coupling being constrained. The spher-

ical entangling surface did not lead to any constraints on the coupling while the cylindrical

and slab entangling surfaces did. It will also be interesting to find if there are other en-

tangling surfaces which lead to a tighter bound and if the bounds are stronger than the

causality constraints. Moreover one can generalize this for arbitrary R2 theory and try get

the constraints on the coupling.

1Note 31/18 ≈ 1.72 while 5/3 ≈ 1.67.

REFERENCES 99

Figure 5.1: Comparison between the various constraints on the GB coupling. The length of

the line represents the range of allowed λ.

References

[1] N. Ogawa and T. Takayanagi, “Higher Derivative Corrections to Holographic Entan-

glement Entropy for AdS Solitons,” JHEP 1110 (2011) 147 [arXiv:1107.4363 [hep-th]].

[2] A. Buchel, R. C. Myers and A. Sinha, “Beyond eta/s = 1/4 pi,” JHEP 0903, 084

(2009) [arXiv:0812.2521 [hep-th]].

[3] D. M. Hofman and J. Maldacena, “Conformal collider physics: Energy and charge

correlations,” JHEP 0805 (2008) 012 [arXiv:0803.1467 [hep-th]].

[4] D. M. Hofman, “Higher Derivative Gravity, Causality and Positivity of Energy in a UV

complete QFT,” Nucl. Phys. B 823 (2009) 174 [arXiv:0907.1625 [hep-th]].

[5] A. Zhiboedov, “On Conformal Field Theories With Extremal a/c Values,”


[6] M. Brigante, H. Liu, R. C. Myers, S. Shenker and S. Yaida, “Viscosity Bound Violation

in Higher Derivative Gravity,” Phys. Rev. D 77 (2008) 126006 [arXiv:0712.0805 [hep-

th]].

M. Brigante, H. Liu, R. C. Myers, S. Shenker and S. Yaida, “The Viscosity Bound and

Causality Violation,” Phys. Rev. Lett. 100 (2008) 191601 [arXiv:0802.3318 [hep-th]].

A. Buchel and R. C. Myers, “Causality of Holographic Hydrodynamics,” JHEP 0908

(2009) 016 [arXiv:0906.2922 [hep-th]].

[7] S. Banerjee, A. Bhattacharyya, A. Kaviraj, K. Sen and A. Sinha, “Constraining gravity

using entanglement in AdS/CFT,” JHEP 1405 (2014) 029 [arXiv:1401.5089 [hep-th]].

100 REFERENCES

6 Relative entropy

6.1 Introduction

In this chapter we will discuss one more way of constraining gravity based on quantum

entanglement. Certain entanglement measures such as relative entropy [1], which roughly

speaking tells us how distinguishable two states are, need to be positive in a unitary theory.

The positivity of this quantity was studied in holographic field theories with two derivative

gravity duals in [2]–related work include [3]. In the context of quantum field theories with

holographic dual gravity descriptions, one can ask what this inequality translates into.

Let us begin by discussing relative entropy. Relative entropy between two states ρ and σ

is defined as

S(ρ|σ) = tr (ρ log ρ)− tr (ρ log σ) . (6.1)

As reviewed in appendix A, in quantum mechanics, this quantity is positive for a unitary

theory. In [2], relative entropy was discussed in the holographic context. The state σ was

chosen to be the reduced density matrix for a spherical entangling surface. In this case,

σ ≡ e−H/tr e−H with H being the modular hamiltonian. It can be easily shown that (see

Refs. [2, 4])

S(ρ|σ) = ∆H −∆S , (6.2)

where ∆H = 〈H〉1 − 〈H〉0 and ∆S = S(ρ) − S(σ) with S(ρ) = −tr ρ log ρ being the von

Neumann entropy for ρ and is the entanglement entropy for a reduced density matrix ρ.

Then the positivity of S(ρ|σ) would require,

∆H ≥ ∆S . (6.3)

Now we can calculate the modular hamiltonian for the sphere [5], from the formula,

H = 2π

∫r<R

dd−1xR2 − r2

2RT00 . (6.4)

Here Tµν is the d-dimensional field theory stress tensor and 00 is the time-time component.

We know how to compute Tµν in holography. The Ryu-Takayanagi prescription (and known

generalizations [6]) gives us a way to compute ∆S. Thus we can check if and how the

inequality ∆H ≥ ∆S is satisfied. In [2], many examples were considered and in each case

101


it was shown that this inequality is respected in Einstein gravity. If we consider a small

excitation around the vacuum state then to linear order in the perturbation ∆H = ∆S. This

can be shown to be equivalent with the linearized Einstein equations [7, 8]. This equality has

been recently shown to hold for a general higher derivative theory of gravity in [9]. It is thus

very interesting to ask what constraints we get at the nonlinear order. We will address this

question for the special case of a constant stress tensor for the case where the holographic

entanglement entropy is given by the Ryu-Takayanagi prescription—in other words, we will

ask if even at non-linear level we get Einstein gravity. We find that the constraints arising

from relative entropy give us a larger class of models than just Einstein gravity. However,

we show that there exists matter stress tensor for which the bulk null energy condition is

violated everywhere except at the Einstein point. This in turn implies that relative entropy

can continue to be positive although the bulk null energy condition is violated. In fact we

can ask the question the other way round: are there examples where the relative entropy

is negative but the bulk null energy condition still holds? We will give an example where

this happens. Thus the connection between energy conditions and the positivity of relative

entropy, which in some sense is reminiscent of the connection between energy conditions and

the laws of thermodynamics, appears to be less direct than what one would have expected.

In order to get some intuition about what feature of gravity ensures the positivity of

relative entropy, we extend the calculations in [2] to higher derivative theories. In particular

we focus on Gauss-Bonnet gravity in 5 bulk dimensions [10, 11, 12] since in this context there

is a derivation [13, 14] of the corresponding entropy functional [15, 16, 17]. We find that in

all examples that we consider, the positivity of the two point function of the stress tensor

guarantees that the relative entropy is positive. In particular we show this for a constant

field theory stress tensor as well as for a disturbance that is far from the entangling surface.

At this point we should emphasize that, the inequality for relative entropy can only be

explicitly checked when the modular hamiltonian is known. Unfortunately, currently this is

not known for cases when the entangling region is a cylinder or a slab.

This chapter is organized as follows. In section (6.2), we consider constraints arising from

the positivity of relative entropy in a holographic set up where the entanglement entropy

is given by the Ryu-Takayanagi entropy functional. These constraints arise at a quadratic

order in a perturbation with a constant field theory stress tensor. In section (6.3), we turn

to the study of relative entropy in Gauss-Bonnet holography. In section (6.4), we investigate

the relative entropy for an anisotropic plasma which breaks conformal invariance. We find

that the relative entropy in this case is negative and we suggest some possible explanations

for this. We conclude in section (6.5). The appendix contains further calculations relevant

for the rest of the chapter. We will use capital latin letters to indicate bulk indices and greek

letters to indicate boundary indices. Lower case latin letters will indicate an index pertaining

CHAPTER 6. RELATIVE ENTROPY 103

to the co-dimension 2 entangling surface. 1

6.2 Relative entropy considerations

In this section we will use the results in [2] to derive certain constraints at nonlinear order

that arise due to the positivity of relative entropy. In Fefferman-Graham coordinates, the

bulk metric can be written as

ds2 =L2

z2dz2 + gµνdx

µdxν . (6.5)

For Einstein gravity, the bulk equations of motion allow us to systematically solve for gµν as

an expansion around the boundary z = 0 (see eg.[19]). The idea here is to see what mileage

we get if we do not know what the bulk theory is but we demand that the relative entropy

calculated using the Ryu-Takayanagi entropy functional is positive. We want to calculate the

quadratic correction to the entanglement entropy for the following form of boundary metric,

gµν =L2

z2

[ηµν + azdTµν + a2z2d(n1TµαT

αν + n2 ηµνTαβT

αβ) + · · ·], (6.6)

where a = 2d

`d−1P

Ld−1. This form is consistent with Lorentz invariance for a constant Tµν . We will

treat Tµν as a small perturbation to the vacuum. At linearized order, it has been shown in

Refs. [7, 8, 9] Einstein equations arise from the condition ∆H = ∆S. We wish to investigate

what happens at the next order. We will keep n1 and n2 arbitrary and derive constraints on

them arising from the inequality ∆H ≥ ∆S. Our analysis follows [2] very closely, the only

change being that we will not specify n1 and n2 to be the Einstein values. Since at linear

order (the argument will be reviewed in the next section) we have the equality ∆H = ∆S

and since T00 from the holographic calculation is just given by the coefficient of the zd term

in the metric, the inequality implies ∆S ≤ 0 at quadratic order. Thus our task is to calculate

∆(2)S, the quadratic correction to ∆S, as a function of n1, n2. The analysis below is valid

for d > 2.

We start with the Ryu-Takayanagi prescription for calculating entanglement entropy in holog-

raphy,

S =2π

`d−1P

∫dd−1x

√h . (6.7)

From Taylor expansion one can show that the quadratic correction to√h is,

δ(2)√h =

1

8

√h(hijδhij)

2 +1

4

√h δhijδhij +

1

4

√h hijδ(2)hij . (6.8)

1 The paper by Erdmenger et al [18] deals with a related idea of looking for pathological surfaces in certain

higher derivative theories of gravity and is of some relevance in this context. Interested readers are referred

to that.

104 6.2. RELATIVE ENTROPY CONSIDERATIONS

The induced metric is,

hij = gij +L2

z2∂iz∂jz . (6.9)

This is evaluated at the extremal surface z = z0 + εz1 =√R2 − r2 + εz1. Hence, at 0-th

order, the metric and its inverse are,

hij =L2

z20

(ηij +

xixjz2

0

)and hij =

z20

L2

(ηij − xixj

R2

). (6.10)

In ∆(2)S, we get 3 kinds of second order contributions. To be systematic, we write,∫dd−1x δ(2)

√h = A(2,0) + A(2,1) + A(2,2) , (6.11)

where schematically, these are the (δg)2, z1δg and z21 contributions respectively. To calculate

the first term, we can set z1 = 0. Then

δhij = aL2zd−2Tij andδ(2)hij

2= a2L2z2d−2

0 (n1TiαTαj + n2 ηijTαβT

αβ) . (6.12)

This gives,

A(2,0) = Ld−1a2

∫dd−1x Rzd0

(Ti0T

i0

(n1

2+ (d− 1)n2 − n2

r2

R2

)+ (T00)2

(n2

2(d− 1)− n2r

2

2R2

)+TijT

ij

(n1

2+n2

2(d− 1)− n2r

2

2R2− 1

4

)− n1

2R2xixjTi0T

0j + xixjTikT

kj

(1

2R2− n1

2R2

)+

1

8

(T 2 − T 2

x − 2TTx))

,

where Tx = xixjTijR2 and T = T ii . The last two terms in (6.11) are same as they appear in [?]

. Quoting the result,

A(2,1) = Ld−1a

∫dd−1x

R

2z0

[T

(z1 −

z20

R2xi∂iz1

)+TijR2

(2z2

0xi∂jz1 − z1x

ixj − z20x

ixjxk∂kz1

R2

)],

(6.13)

A(2,2) = Ld−1

∫dd−1x

R

zd0

[d(d− 1)z2

1

2z20

+z2

0(∂z1)2

2R2− z2

0(xi∂iz1)2

2R4+

(d− 1)xi∂iz21

2R2

]. (6.14)

We can find z1 by minimizing A(2,1) + A(2,2), which gives,

z1 = −aR2zd−1

0

2(d+ 1)(T + Tx) . (6.15)

Plugging this and summing we get from Eq. (6.11),∫dd−1x δ(2)

√h = Ld−1a2

∫dd−1x

(c1T

2 + c2T2x + c3T

2ij + c4Ti0T

i0 + c5

xixjTikTkj

R2

+ c6

xixjTi0T0j

R2+ c7TTx

),

(6.16)


where unlike [2]2, the coefficients c1 · · · c7 are dependent on n1 and n2,

c1 =(R2 − r2)(d−4)/2

8(1 + d)2R

(−4(1 + d)2n2(r2 −R2)2(r2 − (d− 1)R2) (6.17)

+R2(2(d2 + 2d− 1)r4 + (1− 5d2)r2R2 + (2d2 − d− 1)R4)), (6.18)

c2 =(−r2 +R2)

12

(−4+d)((1− 5d2) r2R3 + (−3 + d(3 + 4d))R5)

8(1 + d)2, (6.19)

c3 =(−r2 +R2)

d/2(−2n2r

2 + (−1 + 2n1 + 2(−1 + d)n2)R2)

4R, (6.20)

c4 =(−r2 +R2)

d/2(n1R

2 − 2n2 (r2 − (−1 + d)R2))

2R, (6.21)

c5 =(d2 − (1 + d)2n1)R (−r2 +R2)

d/2

2(1 + d)2, (6.22)

c6 = −n1

2R(−r2 +R2

)d/2, (6.23)

c7 =(−1 + d)R3 (−r2 +R2)

12

(−4+d)((1− 3d)r2 + (1 + 2d)R2)

4(1 + d)2. (6.24)

Now we integrate the expression (6.16) over the (d− 2)-sphere on the boundary. We use the

trick,∫dd−1x f(r)xixjxkxl · · ·n pairs = N(δijδkl · · ·+ permutations)

∫dd−1x f(r)r2n , (6.25)

where N is some normalization constant. For n = 1, N = 1/(d − 1); and for n = 2,

N = 1/((d− 1)2 + 2(d− 1)). The final result comes out in the form 3,∫dd−1x

√h = a2Ld−1Ωd−2

(C1T

2 + C2T2ij + C3T

2i0

), (6.26)

with

C1 =2−3−dd (1 + 4 (d2 − 1)n2)

√πR2dΓ[d+ 1]

(d2 − 1) Γ[

32

+ d] , (6.27)

C2 =2−3−dd

√πR2dΓ[1 + d]

(d2 − 1) Γ[

32

+ d] (

−1− 2d+ 4(d+ 1)n1 + 4(d2 − 1

)n2

), (6.28)

C3 = −2−1−dd(n1 + 2(d− 1)n2)√πR2dΓ[1 + d]

(d− 1)Γ[

32

+ d] . (6.29)

2There appears to be an overall sign missing for c6 in [2].3The expression for C3 in [2] after substituting for n1, n2 is off by a factor of d/(d+2) although the overall

sign is correct. This appears to be related to the opposite sign used for c6. We have cross-checked our results

on mathematica for various cases and the notebook may be made available on request.


Now we must demand that ∆(2)S ≤ 0. We can write ∆(2)S = V TMV with V being a

(d−1)(d+2)/2 dimensional vector with the independent components of Tµν as its components.

Then demanding that the eigenvalues of M are ≤ 0 will ensure ∆(2)S ≤ 0. This leads to

n1 + 2(d− 1)n2 ≥ 0 , (6.30)

2d+ 1− 4(d+ 1)n1 − 4(d2 − 1)n2 ≥ 0 , (6.31)

d+ 2− 4(d+ 1)n1 − 4d(d2 − 1)n2 ≥ 0 . (6.32)

Figure 6.1: (colour online) For d > 2 we get the allowed n1, n2 region to be the blue triangle

above for a generic stress tensor. The region above the blue solid line and below the blue

dashed and dotted lines are allowed from the relative entropy positivity. For d→∞ the region

collapses to a line 0 ≤ n1 ≤ 1 indicated in green. The Einstein value (n1, n2) = (12,− 1

8(d−1)) is

shown by the black dot. The region below the solid red line and above the dashed and dotted

red lines are allowed by the null energy condition. By turning on a generic component of

the stress tensor only the Einstein value is picked out. By switching off certain components

of the stress tensor, various bands bounded by the solid, dashed and dotted lines are picked

out.

We get the region indicated in Fig. (6.1) allowed by this set of inequalities. One interesting


observation is that when d → ∞, then the allowed region becomes the interval 0 ≤ n1 ≤ 1

with n2 = 0 coinciding with the Einstein result. The area of the triangle is given by

Aread =d2

8(d+ 1)2(d− 2). (6.33)

Notice that the (extrapolated) Aread=2 is infinity. This makes sense since in d = 2 we expect

constraints on only 2 eigenvalues (since T 2 and T 2ij are no longer independent) which will give

us an unbounded region. Further Aread→∞ → 0 which leads to a line interval for d→∞ as

shown in Fig. (6.1).

At this stage, we have a wider class of theories that are allowed by the inequality than

the Einstein theory. The other theories need extra matter in addition to Einstein gravity to

support them. As such we could ask if the matter needed satisfies the null energy condition.

As an example consider turning on a constant T01 in d = 4. Then we find

RAB −1

2gAB(R +

12

L2) = T bulkAB , (6.34)

with T bulkAB working to be

T bulkAB = 16z6T 201

[3

2(δn1 + 4δn2)δzAδ

zB + (δn1 + 6δn2)δ0

Aδ0B − (δn1 + 6δn2)δ1

Aδ1B

− 2(δn1 + 3δn2)∑i=2,3

δiAδiB

].

(6.35)

m Here δn1 = n1−1/2 and δn2 = n2 +1/24. Using this we find that the null energy condition

T bulkAB ζAζB ≥ 0 leads to

T bulk00 + T bulk22 = −δn1 ≥ 0 , (6.36)

T bulk00 + T bulkzz =5

2δn1 + 12δn2 ≥ 0 , (6.37)

with T bulk00 + T bulk11 = 0. These simplify to n1 ≤ 1/2 and n2 ≥ −1/24. Thus the region in

fig.1 that respects the null energy condition is smaller than that allowed by the positivity of

relative entropy.

For a general constant stress tensor in general d we proceed as follows. We note that for

a metric of the form in Eq. (6.5), with gµν a function of z only, we have [20]

Rµν = R′µν − (z∂zKµν +KKµν − 2KµκKκν ) , (6.38)

Rµz = 0 , (6.39)

z2Rzz = −gµνz∂zKµν +KµνKµν , (6.40)

R = R′ − (2zgµν∂zKµν +K2 − 3KµνKµν) , (6.41)


where Kµν = 12z∂zgµν . Here ′ denotes a quantity computed with gµν . Using these it is

straightforward (but tedious) to compute (setting L = 1 for convenience, defining Sµν =

n1TµαTαν + n2ηµνTαβT

αβ and aborbing the factors of a into Tµν ; the raising and lowering of

indices on Tµν , Sµν are done with ηµν . Also we have used T µµ = 0.)

gµν = z2[ηµν − T µνzd + (T µαT να − Sµν)z2d] , (6.42)

Kµν = − 1

2z2(2ηµν − (d− 2)zdTµν − 2(d− 1)z2dSµν) , (6.43)

Kνµ = −1

2[2δνµ − dzdT νµ + dz2d(TµαT

αν − 2Sνµ)] , (6.44)

K = −1

2[2d+ dz2d(TαβT

αβ − 2Sαα)] , (6.45)

Kµν = −z2

2[2ηµν − (d+ 2)zdT µν + 2(d+ 1)z2d(T µαT

να − Sµν)] , (6.46)

zgµν∂zKµν =1

2[4d+ z2d(4d(d− 2)Sαα − d(d− 4)TαβT

αβ)] , (6.47)

KµνKµν = d+

z2d

4[d(d+ 4)TαβT

αβ − 8dSαα ] , (6.48)

KµκKκν =

1

4z2[4ηµν − 4(d− 1)zdTµν + z2d(d2T κµTκν − 4(2d− 1)Sµν)] . (6.49)

Using these we find

T bulkzz = −d(d− 1)z2d−2TαβTαβ(δn1 + dδn2) , (6.50)

T bulkµν = d2z2d−2[−δn1TµκT

κν + ηµνTαβT

αβ(δn1 + (d− 1)δn2)]. (6.51)

Here δn1 = n1−1/2 and δn2 = n2 +1/(8(d−1)) i.e., the deviations from the Einstein values.

Now we are in a position to ask if the matter supporting this bulk stress tensor satisfies the

null energy conditions or not. First we note that T bulk00 + T bulk11 ≥ 0 immediately leads to

− d2(−T 200 + T 2

ij)δn1 ≥ 0 . (6.52)

This leads to a definite sign for δn1 if and only if (−T 200 + T 2

ij) has a definite sign. But in

general, there is no reason for this combination to have a definite sign. So we are led to

suspect that for a generic stress tensor, δn1 = 0. To confirm this suspicion let us look at

T bulkzz + T bulk00 .

T bulkzz + T bulk00 =

−d[(d− 1)T 2

00(δn1 + 2dδn2) + T 2ij[(2d− 1)δn1 + 2d(d− 1)δn2]

+T 20i[(2− 3d)δn1 − 4d(d− 1)δn2]

]. (6.53)

As in the relative entropy analysis, we write the RHS as V TMV where V is a (d−1)(d+2)/2

dimensional vector whose non-zero independent components are the T00, Tij, T0i’s. Then we


demand that the eigenvalues of M are positive for the null energy condition to hold for a

generic constant traceless stress tensor Tµν . This yields

(3d− 2)δn1 + 4d(d− 1)δn2 ≥ 0 , (6.54)

(2d− 1)δn1 + 2d(d− 1)δn2 ≤ 0 , (6.55)

δn1 + 2(d− 1)δn2 ≤ 0 . (6.56)

Only for δn1 = δn2 = 0 are these inequalities satisfied for d > 2. Thus the null energy

condition picks out the Einstein value if we ask if for a generic constant stress tensor the

O(T 2) terms are supported by matter. Of course as we saw for d = 4 we can turn on T0i and

set everything else to zero, there would be a region in the n1, n2 parameter space where the

null energy condition and the positivity of the relative entropy would hold (this corresponds

to the region between the red and blue solid lines in Fig. (6.1). For the generic case, only

the Einstein value is picked out. To emphasis, that the Einstein value was picked out for

the generic case, relied only on the null energy condition analysis and did not rely on the

positivity of the relative entropy. To summarize, we found that there exists a larger class of

theories in the (n1, n2) parameter space than just the Einstein theory. However, except at the

Einstein point, we found that there always exists some matter stress tensor which violates

the bulk null energy condition.

6.3 Relative entropy in Gauss-Bonnet holography

In this section we will calculate relative entropy for excited states in Gauss-Bonnet gravity.

For definiteness, we will consider d = 4 or 5-dimensional bulk. We will follow the conventions

in [11]. The total action is given by

I = Ibulk + IGH + Ict , (6.57)

where

Ibulk =

∫d5x√g

[R +

12

L2+λ

2L2(RABCDR

ABCD − 4RABRAB +R2)

]. (6.58)

The generalized Gibbons-Hawking term is given by [21]

IGH = − 1

`3P

∫d4x√γ[K − λL2(2GµνK

µν +1

3(K3 − 3KK2 + 2K3)

]. (6.59)

Here Gµν = Rµν−1/2γµνR made from the boundary γµν , K2 = KµνKµν and K3 = Kα

βKβγK

γα.

Kµν is the extrinsic curvature and K = Kαα . The counterterm action Ict is needed for the

cancellation of the power law divergences in Itot. For our case this works out to be [22, 23]

(L and f∞ are defined below)

Ict =1

`3P

∫d4x√γ[c1

3

L+ c2

L

4R], (6.60)

110 6.3. RELATIVE ENTROPY IN GAUSS-BONNET HOLOGRAPHY

where R is the four dimensional Ricci scalar and c1 = 1− 23f∞λ and c2 = 1 + 2f∞λ .

The equations of motion are given by

RAB −1

2gABR−

6

L2gAB −

λL2

2HAB = 0, (6.61)

where

HAB =1

2gAB(R2 − 4RCDR

CD +RCDEFRCDEF )− 2RRAB + 4RA

CRCB

− 2RACDERBCDE − 4RCDRCABD .

(6.62)

AdS5 given by

ds2 =L2

z2

(dz2 − dt2 + dx2

1 + dx22 + dx2

3

)(6.63)

where L = L/√f∞ with 1 − f∞ + λf 2

∞ = 0. The dual CFT is characterized by the central

charges c, a appearing in the trace anomaly[11, 24]:

c =π2L3

`3P

(1− 2λf∞) , a =π2L3

`3P

(1− 6λf∞) . (6.64)

The CFT stress tensor two point function is given by

〈Tµν(x)Tρσ(0)〉 =40c

π2(x2)4Iµν,ρσ(x) , (6.65)

where I is a function of x and the positivity of the two point function leads to c > 0.

We will need the formula for the holographic stress tensor (see [25])

Tµν =1

`3p

[Kµν − gµνK + λL2(qµν −1

3gµνq)]−

3

Lc1γµν +

L

2c2[Rµν(γ)− 1

2γµνR(γ)] , (6.66)

where

q = hµνqµν

qµν = 2KKµαKαν − 2KµαK

αβKβν + Kµν(KαβKαβ − K2) + 2KRµν + RKµν − 2KαβRαµνβ −

4Rα(µKν)α . The terms proportional to c1, c2 come from Ict.

We also note that the GB coupling λ is bounded. Following [26] for the calculation of the

three point correlation function of stress tensor one needs to compute a energy flux which

comes form the insertion of εijTij , where εij and Tij are the polarization tensor and stress

tensor respectively. Demanding the positivity of this energy flux in the holographic set up

we get the following three constraints and from those we obtain bounds on λ . These coincide

with the bounds arising from micro-causality [27].

Tensor channel : 1− 10f∞λ ≥ 0⇒ λ ≤ 9

100

Vector channel : 1 + 2f∞λ ≥ 0⇒ −3

4≤ λ ≤ 1

4

Scalar channel : 1 + 6f∞λ ≥ 0⇒ − 7

36≤ λ ≤ 1

4

(6.67)


From this we get

− 7

36≤ λ ≤ 9

100.

This is the same as the condition 1− 4f∞λ− 60f 2∞λ

2 ≥ 0.

6.3.1 Linear order calculations

We are interested in considering the excited state to be a perturbative excitation of the

ground state. At linear order in the perturbation ∆H = ∆S. Let us review the argument of

[2] why. Let ρ0 be a reference state. Now let ρ(α) be a continuous family of states dependent

on a parameter α that runs over all possible values. We choose the parametrization such

that ρ(α = 0) = ρ0. Now relative entropy vanishes for two states that are equal. So we must

have S(ρ(0)|ρ0) = 0 and also S(ρ(α → ε±)|ρ0)→ 0+ > 0 where ε is a small positive valued

number denoting a small perturbation from the reference state ρ0. This means at α = 0 we

must have, d(S(ρ(α)|ρ0))/dα = 0. Or equivalently at the linear order of the perturbation ε,

∆H = ∆S , (6.68)

which follows from Eq. (6.2). We can demonstrate this with a simple example4. Let ρ0 to be

the vacuum of the CFT4 whose holographic dual is the empty AdS5 (our linearized results

are a sub-case of the more general case worked out in [9]),

ds2 =L2

z2

(dz2 − dt2 + dx2

1 + dx22 + dx2

3

)(6.69)

We choose ρ1 to be the dual of a metric which is being perturbed around the empty AdS.

Following [2], we take the perturbation to be of the form,

δgµν =`3P

2L3z2∑n

z2nT (n)µν . (6.70)

To keep track of the perturbation we keep the components of T(n)µν proportional to a small

number ε. We compute the entanglement entropy from the Jacobson-Myers functional,

S =2π

`3P

∫d3x√h(1 + λL2R) +

4π

`3P

λL2

∫d2x√h K . (6.71)

Here, hab is the induced metric on the minimal surface and R and K are respectively the

intrinsic ricci scalar and extrinsic curvature evaluated on that surface. To simplify notation,

we will set L = 1 . The minimal surface equation is given by

K + λL2(RK− 2RijKij) = 0 , (6.72)

4The change in entanglement entropy for excited states in GB holography has been considered in [28].


which was derived in [13, 29] following [30]. For the spherical entangling surface in the

unperturbed metric the following continues to be an exact solution

z = z0 =√R2 − r2 . (6.73)

In the perturbed case, it changes to

z = z0 + ε z1 . (6.74)

However note that we obtained z0 by extremization. Hence z1 can only contribute to a

quadratic order in ε and not at linear order. Thus at linear order we can set z1 = 0. Using

z = z0 to compute (6.71) and then extracting the terms proportional to ε gives us ∆S.

Now we can calculate the modular hamiltonian, from the formula in Eq. (6.4), where T00 is

obtained in holography using Eq. (6.66). Since T00 = 0, for empty AdS, this directly gives

∆H. Now we will demonstrate the equality in Eq. (6.68) by considering a special case (we

have checked that this holds in the other examples considered below as well).

Using Gauss-Bonnet eom, we can determine T(n)µν in terms of the lowest mode T

(0)µν . It

turns out they are all derivatives of T(0)µν . To keep it simple we take T

(0)µν to be a constant.

Also note that to satisfy GB eom, we must have traceless and divergenceless conditions on

T(0)µν ,

T (0)µ

µ = 0 and ∂µT(0)µ

ν = 0 . (6.75)

Consider an isotropic perturbation

T (0)µν =

(E , E

3,E3,E3

)(6.76)

Note that this satisfies the conditions in Eq. (6.75). However the holographic dual tensor Tµν

is not same as T (0)µν . We compute it from Eq. (6.66) as,

Tµν = (1− 2f∞λ)

(E , E

3,E3,E3

)(6.77)

Now using (6.4) one gets5,

∆H =8π2L3ER4

15`3P

(1− 2f∞λ) .

As discussed before, we can compute ∆S from (6.71) with z =√R2 − r2, and then take out

the ε order coefficients. We obtain,

√h(1 + λL2R) = −E (R2 (3 + 30f∞λ)− r2 (1 + 58f∞λ))

6f3/2∞ R

(6.78)

5There is a typo in Eq. (6.29) in [9] for 〈Tµν〉. There is a factor of 2 missing in front of the term proportional

to a1 in that expression. Taking this into account our expression agrees with their both for GB and for the

general R2 theory discussed in appendix C.


from which we calculate,

∆S =8π2L3ER4

15`3P

(1− 2f∞λ) . (6.79)

This demonstrates ∆H = ∆S for an isotropic perturbation.

6.3.2 Quadratic corrections

Now we turn to the more interesting case of quadratic corrections which lead to inequalities.

We take the following form for the boundary metric,

z2gµν = ηµν + zdTµν + z2d(n1TµαTαν + n2 ηµνTαβT

αβ) + · · · (6.80)

where compared to Eq. (6.6) we have absorbed a factor of a into the stress tensor. We need to

fix the numbers n1 and n2. By plugging into the GB equations of motion given by Eq. (6.61),

we find that

− 3(n1 + 4n2) + f∞(1 + 6n1λ+ 24n2λ) = 0 (6.81)

n1(9− 17f∞ + 25f 2∞λ)− 4(4f 3

∞λ− 3n2(1− 9f∞ + 17f 2∞λ)) = 0 (6.82)

Solving the two equations and using the relation 1− f∞ + f 2∞λ = 0 we get

n1 =1

2

1 + 2f∞λ

1− 2f∞λand n2 = − 1

24

1 + 6f∞λ

1− 2f∞λ. (6.83)

These results match with the λ = 0 case given in [2]6.

6.3.3 Constant Tµν

The next step is to calculate the second order change in ∆S. For a general but constant

stress tensor we can guess the following form of the second order correction of entropy from

Lorentz invariance,

∆(2)S = C1T2 + C2TijT

ij + C3T0iT0i (6.84)

where T denotes the trace of the spatial part of the stress tensor Tµν . The latin indices run

from 1 to 3, and denote the spatial part of a tensor. They are raised with ηij. Our task is to

identify the constants Ci’s for a non-zero λ. The only condition on the stress tensor is that it

6Notice a curious fact. If we demanded that n1 ≥ 0 and n2 ≤ 0, or in other words even in GB gravity

they have the same sign as in Einstein gravity then with c > 0, we would get

1 + 2f∞λ ≥ 0 , 1 + 6f∞λ ≥ 0 .

But these are nothing but the scalar and vector channel constraints in Eq. (6.67)! These leads us to wonder

if entanglement entropy knows about the causality constraints.


is symmetric and traceless. To do the perturbative analysis we assume that the components

of the stress tensor are proportional to a perturbative parameter ε. Also we have absorbed

a parameter a in the stress tensor. The background metric will be changed in the quadratic

order as given in (6.6). Now, assume that the minimal surface z0 =√R2 − r2 is modified as

z = z0 + εz1 . (6.85)

z1 contributes at the quadratic order in the JM functional (6.71). So it is sufficient to

consider only the first order fluctuation to the entangling surface. Next we expand the

entropy functional upto quadratic order and then extract the terms proportional to ε2 which

gives the quadratic correction to the entropy. We vary it with respect to z1. This gives us

the equation of motion for z1. We find the solution and put it back to ∆(2)S. Since it was

shown that at linear order, ∆S = ∆H, at second order we must have ∆(2)S > 0. We get the

following equation for z1 ,

(1− 2f∞λ)

[∂2(z0z1)− xixj

R2∂i∂j(z0z1)− (R2 − r2)2 (T + 3Tx)

]= 0 , (6.86)

with the solution,

z1 = −R2z3

0

10(T + Tx) . (6.87)

Notice that the equation is the same as what appears in the Einstein case upto the overall

factor of (1− 2λf∞). The Gibbons Hawking term doesn’t contribute to the action when we

put in the solution. Alternatively, we could have taken the action and integrated all terms

involving z′′1 (x)’s by part and cast it in the conventional form. The surface term resulting

from this will cancel with the appropriate Gibbons-Hawking term. We have checked both

approaches and have got the same result. Integrating the resulting action over the volume

of the entangling region, we obtain the second order correction to the entropy,

∆(2)S = −8π3L3(1− 2f∞λ)

`3P

(C1T

2 + C2T2ij + C3T

2i0

), (6.88)

C1, C2, C3 are same as the Einstein values obtained in section 2.

Note that this is just a factor of (1−2f∞λ) times what is obtained in the Einstein gravity

(the Einstein result was manifestly negative). This can be cross-checked easily on a computer

by suitably turning on various components of the stress tensor and identifying various tensor

structures.

Now from the discussions in the previous sections, it is clear that this quantity has to be

negative. The only constraint to ensure ∆(2)S < 0 is

1− 2f∞λ > 0 . (6.89)


This is equivalent to saying the central charge c > 0 which also is the condition needed for the

positivity of the two point function of the field theory stress tensor. The condition λ < 1/4

ensures that this holds. If this inequality on λ did not hold, the corresponding vacuum would

have ghosts [11].

6.3.4 Shockwave background

Up to this point we have only considered constant stress-tensor. It is interesting to ask if we

get non-trivial constraints for Tµν not constant. To explore a nontrivial case of non-constant

Tµν , consider the following 5 dimensional metric

ds2 =L2

z2(dz2 + dxµdx

µ + f(t+ x3)W (z, x1, x2)(dt+ dx3)2) (6.90)

where µ = 1, 2.

The above metric solves the GB equation exactly, given that W (z, x1, x2) satisfies the

following differential equation,

∂2zW + ∂2

x1W + ∂2

x2W = −3

z∂zW , (6.91)

with no constraint on f(t+x3). If f = δ(t+x3) then this is the shockwave metric considered

for example in [26] to derive constraints on higher derivative gravity theories. We will set

f = 1 and in a slight abuse of terminology continue to refer the metric as a shockwave.

W (z, x1, x2) is taken as

W (z, x1, x2) =L2z4

(z2 + (x1 − x′1)2 + (x2 − x′2)2)3. (6.92)

Here (x′1, x′2) represent the point where the disturbance is peaked. Since in our calculations

we perturb the background metric, we should choose x′1 and x′2 to be outside the entangling

region. With this in mind we proceed with the second order calculation. Next we consider

a shockwave disturbance localized just outside the entangling surface. We will set x′2 = 0 in

(6.92). We start with the following metric which is obtained by expanding W around z = 0

and retaining the first two terms in the expansion,

ds2 =L2

z2(dz2 + dxµdx

µ + (z4L2ε3

(x21 + (x2 − x′2)2)3

− 3z6L2ε4

(x21 + (x2 − x′2)2)4

)(dt+ dx3)2) (6.93)

The ε factors have been inserted to keep track of the order of the expansion and matches with

the power appear in the denominator. If we write the entangling surface as z = z0 + ε3z1

then the quadratic terms in z1 will involve ε6 which is at a higher order than the second

order term in the metric above. Thus we expect to see an inequality ∆H > ∆S with the


above metric setting z1 = 0. We thus evaluate the entropy functional considering only the

unperturbed entangling surface and expand it upto ε4 and pick out the ε4 term which gives

the first leading order change in the relative entropy. The integrand is shown below,

∆(2)S =2π

`3P

∫dx3dx1dx2

3L5

2Rf5/2∞ (x2

3 + (x2 − x′2)2)6

[(x2

3 + x22 + x2

1 −R2)(40f∞(x23 + x2

2 + x21 −R2)

(4R2(x23 + (x2 − x′2)2)− 4(x4

3 + x23(x2

1 + 2x2(x2 − x′2)) + (x21 + x2

2)(x2 − x′2)2))λ+ 16f∞

(R2 − x23 − x2

2 − x21)(x2

3 + (x2 − x′2)2)(2R2 − 13x23 − 2x2

1 − 13x22 + 12x2x

′2)λ

− (x23 + (x2 − x′2)2)2(60f∞(x2

3 + x− 22)λ−R2(1 + 18f∞λ) + x21(1 + 18f∞λ))) .

(6.94)

Then we perform the integration over x3 which goes from −√R2 − r2 to

√R2 − r2 and

x1 = r cos(θ) , x2 = r sin(θ) . Now after some algebraic manipulation we can write the

integrand as,

∆(2)S =2πL5

2`3Pf

5/2∞ R(r2 + x′22 − 2rx′2 sin(θ))6

(f1 + f2 sin(θ) + f3 sin(θ)2) (6.95)

where f1, f2, f3 are some function of r and λ . Integral over θ goes from 0 to 2π and integral

over r goes from 0 to R. We first perform the θ integral. To perform the θ integral we have

used the following integral identity:∫ 2π

0

dθ

a+ b sin(θ)=

2π√a2 − b2

,

Finally we get,

∆(2)S =2π

`3P

∫ R

0

dr[− L5

240f5/2∞ R

(f1(−30 (8a5 + 40a3b2 + 15ab4) π

(a2 − b2)11/2)

+ f2(90b (8a4 + 12a2b2 + b4) π

(a2 − b2)11/2)− f3(

30 (4a5 + 41a3b2 + 18ab4) π

(a2 − b2)11/2))],

(6.96)

where, a2 = r2 + x′22 and b = −2rx′2 . Next we perform the r integration. The leading

contribution in ∆(2)S comes form the lower limit of the r integral which is shown below.

∆(2)S =π2L5

96`3Pf

5/2∞ R2

(1− 2f∞λ)f(x′2) , (6.97)

where, f(x′2) is a negative valued function given by

f(x′2) =

(√x′22 − 1 (−136 + 72x′22 − 56x′42 + 15x′62 )− 3 (32− 16x′22 + 36x′42 − 22x′62 + 5x′82 )Csc−1(x′2)

)(x′22 − 1)

9/2

(6.98)


and plotted in Fig. (6.2). To satisfy, ∆S ≤ ∆H we will get, 1− 2f∞λ ≥ 0 or in other words

c > 0. Note that in order for us to be able to expand in small z, the perturbation needs to

be located far away from the entangling surface. This is because in the denominator in W

we had z2 + x21 + (x2 − x′2)2. When we plug in z = z0, the maximum value for z is R and

this happens when x1 = x2 = 0. Thus we will need R x′2 for the expansion to be valid. It

will be interesting to see what happens as we move the perturbation closer and closer to the

entangling surface. However this appears to be a very hard problem.

Figure 6.2: Negative of the function f(x′2) is plotted which is a positive valued function

6.3.5 Correction from additional operators

In this section we consider perturbed states in which certain additional operators acquire

nontrivial vacuum expectation value. Our analysis will follow [2]. The holographic dual of

these operators will involve additional massive fields in the bulk. We will show that even for

such cases in Gauss-Bonnet gravity, the relation ∆H > ∆S will hold. Again we are in AdS5

with the bulk action given by,

I =

∫d5x√−G

(R +

12

L2+λL2

2

(R2 − 4R2

AB +R2ABCD

)− 1

2(∂φ)2 − 1

2m2φ2

), (6.99)

where we have added a massive scalar field which acts as a bulk dual of a scalar operator of

dimension ∆. When m2 = ∆(4−∆), the field φ behaves asymptotically as,

φ = γOz∆ . (6.100)

118 6.4. RELATIVE ENTROPY FOR AN ANISOTROPIC PLASMA

Now we can work out the stress tensor corresponding to this from the formula,

TAB =1

2∂Aφ∂Bφ−

1

4gAB((∂φ)2 +m2φ2) . (6.101)

This will result in the following change to the boundary metric boundary metric,

z2δgµν = azd∑n

z2nT (n)µν + z2∆

∑n

z2nσ(n)µν (6.102)

where we must have,

σ(0)µν = − γ2

12(1− 2f∞λ)ηµνO2 . (6.103)

in order to satisfy Gauss-Bonnet eom. The higher modes, namely σ(n)µν (n > 0) are composed

of derivatives of σ(0)µν . As in [2], we consider O to be slowly varying and, hence, neglect the

higher modes.

It is not necessary to find any correction to the entangling surface. There are two different

perturbations, both in their first orders, and using the z0 minimal surface to compute ∆S

will suffice. The correction to entropy will have two parts,

∆S = ∆TS + ∆OS . (6.104)

The first part, ∆TS comes from the holographic boundary stress tensor Tµν , and its the same

as what we calculated before for the linear order. The second part comes from the scalar

field and is obtained by calculating the area functional with the metric of Eq. (6.102).

∆OS = −π3/2R2∆γ2(−2 + 3∆)Γ[−1 + ∆]Ωd−2

48 aΓ[

12

+ ∆] O2 . (6.105)

Note that the result is independent of λ. Since the result is negative it seems the met-

ric already knows of the positivity of relative entropy even for Gauss-Bonnet provided the

unitarity bounds are respected.

6.4 Relative entropy for an anisotropic plasma

We now want to turn our attention to a holographic anisotropic plasma–there is going to be

a surprise in store. We consider the holographic dual of the deformed N = 4 SYM where

the deformation is generated by anisotropy along one spatial direction viz.

S = SN=4 +1

8π2

∫θ(z) Tr F ∧ F, (6.106)


θ is the field generating anisotropy along the z direction. The holographic dual is the Einstein-

dilaton-axion system given by

Sbulk =1

2`3P

∫M

√−g(R +

12

L2− 1

2(∂φ)2 − 1

2e2φ(∂χ)2) +

1

2`3P

∫∂M

√−γ2K, (6.107)

where φ is the dilaton and at the level of the solution is taken to be a function of the AdS

radius only and χ is the axion dual to the gauge theory θ-term, responsible for inducing

anisotropy, which is taken to be χ = ρx3. This model was proposed and studied in detail in

[31]. The low anisotropy regime corresponding to ρ/T 1 in this model is unstable [31].

The metric equations are given by (L = 1)

RMN −1

2RgMN − 6gMN = TMN , (6.108)

where the bulk matter stress tensor is given as

TMN =1

2∂Mφ∂Nφ−

1

4(∂φ)2gMN +

1

2e2φ∂Mχ∂Nχ−

1

4e2φ(∂χ)2gMN . (6.109)

The metric, φ and χ equations can be written as

RMN + 4gMN −1

2∂Mφ∂Nφ−

1

2e2φ∂Mχ∂Nχ = 0,

∇2φ− e2φ(∂χ)2 = 0,

∇2χ = 0 .

(6.110)

The metric in the FG coordinates is given by

ds2 =dz2

z2+

1

z2γµν(z, x

i)dxµdxν , (6.111)

where

γtt = −1 +ρ2

24z2 + . . . ,

γx1x1 = γx2x2 = 1− ρ2

24z2 + . . . ,

γx3x3 = 1 +5ρ2

24z2 + . . . ,

(6.112)

If we introduce a temperature, the modification to the metric will start at O(z4). Further,

the scalar field introduces a new scale which breaks scale invariance explicitly and the trace

of the boundary stress tensor is now non-zero. It needs to be checked if the null energy

condition is satisfied by the bulk stress tensor TMN given by Eq. (6.109). Contracting the

above with the null vectors ξµ we have

TMNξMξN =

1

2[(∂ξφ)2 + e2φ(∂ξχ)2], (6.113)

120 6.4. RELATIVE ENTROPY FOR AN ANISOTROPIC PLASMA

where ∂ξ(φ, χ) = ξM∂M(φ, χ) and ξMξNgMN = ξ2 = 0. Since the bulk scalar axion follows

the profile χ = ρx3 then

ξM∂Mχ = ρξx3 , (6.114)

whereas the dilaton field φ depends on the radial coordinate. The NEC for the bulk stress

tensor becomes by contracting with the null vectors Tµνξµξν as

Tx3x3 =ρ2

2e2φ(ξx3)2 =

ρ2

2e2φ ≥ 0,

Tuu =1

2(∂ξφ)2 ≥ 0 .

(6.115)

Thus we have explicitly verified that the bulk stress tensor satisfies the null energy condition.

We now want to verify the calculation for the relative entropy in this low anisotropy

regime. As mentioned before, the low anisotropy phase is thermodynamically unstable. We

can thus try to see what happens to the relative entropy in such a phase. Also note that we

are considering Einstein gravity for which the entropy functional is the Ryu-Takayanagi one.

Further in the low anisotropy regime, we are interested in, since we are expanding γµν upto

O(z2) (assuming a small entangling surface Rρ 1) and the stress tensor appears at O(z4),

we have ∆H = 0. Here the state σ is the vacuum state which corresponds to ρ = 0 and is

conformally invariant. Thus the modular hamiltonian will be the same as in Eq. (4). Thus

we only need to compute the change in the entanglement entropy. Furthermore, at leading

order in ρ we expect to see an inequality and as such we do not need to evaluate the change

in the entangling surface.

Putting in the solution for the entangling surface f(x1, x2, x3) =√R2 − x2

1 − x22 − x3

3 we

have

√h =

1

48(R2 − x21 − x2

2 − x33)2R

[48R2 + (R2− x21− x2

2− x23)(3R2− 5x2

3 + x21 + x2

2)ρ2] +O(ρ4) .

(6.116)

The entanglement entropy then becomes

S =2π

`3p

∫dx1dx2dx3

1

48(R2 − x21 − x2

2 − x33)2R

[48R2+(R2−x21−x2

2−x23)(3R2−5x2

3+x21+x2

2)ρ2] .

(6.117)

In spherical polar coordinates x3 = r cos θ, x1 = r sin θ sinφ, x2 = r sin θ cosφ where (θ, φ) are

spherical polar coordinates we have

∆1S =2πρ2

`3p

∫(3R2 − 2r2 − 3r2 cos 2θ)

48(R3 − r2R)r2 sin θdθdφdr . (6.118)

Carrying out the (θ, φ, r) integrals we find (on reinstating L factors)

∆1S =π2ρ2R2L3

6`3p

(−5

3− log[

ε

2R]) . (6.119)


Here ε is a cutoff and r = R − ε (since ε → 0 corresponds to z → 0 it is related to the UV

cutoff). The log-divergence is due to the breaking of conformal invariance by the excited

state. However, notice that in the limit of ε→ 0, the result leads to ∆1S > 0 and hence the

positivity of relative entropy is violated.

Since the positivity of relative entropy in quantum mechanics depends on unitarity (re-

viewed in appendix A), this leads to the following possible interpretations:

1. There are additional contributions which we are missing and they are required for

the positivity of the relative entropy to hold in this case. One could speculate that

there are additional saddle points of the bulk gravity theory which contribute to the

entanglement entropy. It will be interesting to find out those saddle points and see if

they “unitarize” the problem7.

2. Holographic relative entropy positivity needs further conditions than just bulk uni-

tarity. It could be that the derivation of the positivity does not go through in any

straightforward manner to quantum field theory.

3. In the low anisotropy regime, may be there is a loss of bulk unitarity that is not

immediately apparent.

All possibilities need further investigation. Let us first briefly comment on the third possi-

bility. Expanding the linearized equations near the boundary and upto linear order in ρ we

have

( +2

L2)hij = 0, ( +

2

L2)hMx3 +

ρ

2LMx3χ1 = 0, (6.120)

φ1 − 2ρ∂x3χ1 = 0, χ1 = 0, (6.121)

where hMN , φ1 and χ1 are metric, φ and χ fluctuations respectively and i, j take values

apart from x3. ∇A is evaluated using the AdS5 metric. Here LMN ≡ δMx3∂N + δNx3∂M

is a linear operator. The coupling between the metric and χ fluctuation is of the form

Hh + Lχ = 0, Hχ = 0. But this form is similar to what arises in the context of logarithmic

conformal field theories which are non-unitary [33]. Thus one should check if there are log

modes in the fluctuations. We can do this following [34]. According to the arguments in [34]

log modes arise if the form of the equations is ( + a)2hµν = 0. Let us check what the form

of the equations are when we decouple them. Using 8 ∇Aχ1 = ∇Aχ1 − 4L2∇Aχ1, we find

7This is very similar to the resolution of information loss paradox in case of eternal AdS Black Holes as

formulated by Maldacena [32]. The exponentially small correlation as required by the unitarity arises form

the periodically identified Euclidean AdS, although this is not the dominant contribution to the canonical

ensemble.8Useful identities can be found for eg. in the appendices of [35]

122 6.5. DISCUSSION

that the decoupled equation for hMx3 takes the form

( +2

L2)( +

4

L2)hMx3 = 0 ,

while for φ we get

( +4

L2)φ = 0 .

Neither of the four derivative linear operator is of the form ( + a)2 and hence following

the arguments in [34] there are no log modes so the dual field theory is not a log CFT.

Naively it may appear that the propagator for say the φ field will look like 1/(p2(p2 +m2)) =

1/m2(1/p2 − 1/(p2 + m2)), and hence the theory is non-unitary. However, this is not true

since in addition to the decoupled form of the equations above, the relations in Eq. (6.121)

still have to hold–any loss of unitarity would have shown up in the asymptotic fall offs in the

field. Thus it appears that the other two possibilities become plausible.

In closing this section, we note that in [2] it was argued that the relative entropy should

increase as the radius of the entangling surface increases. In our case since ∂RS(ρ1|ρ0) =

−∂R∆1S ≈ π2ρ2RL3

3`3p(log[ ε

2R]) < 0 and hence this monotonicity would also appear violated.

6.5 Discussion

In this chapter we used holographic entanglement to constrain gravity in interesting ways.

First, we started with the Ryu-Takayanagi entropy functional (which holds for Einstein grav-

ity) and considered what constraints arise at nonlinear order on the metric by demanding

that relative entropy is positive. At linearlized level, it is now known that for the spherical

entangling surface ∆H = ∆S leads to linearized equations for any higher derivative theory

of gravity [9]. We considered a constant field theory stress tensor. At the next order, we

found interesting constraints on the terms allowed by the positivity of relative entropy. These

were more general than what arises from Einstein gravity. We analysed energy conditions for

matter that could support these additional theories. We showed that the additional theories

could be supported by matter that violates the null energy condition. In other words, holo-

graphic relative entropy can be positive although the bulk null energy condition is violated.

It is an important open problem to understand if this feature persists for a more general stress

tensor. We also gave an example of a model which corresponds to an anisotropic plasma,

where for small anisotropy, the relative entropy is negative. This occurred even though the

bulk stress tensor satisfied the null energy condition. We gave some possible explanations for

this. We will leave further investigations of similar models as an open problem.

Second, we analysed the inequality in Gauss-Bonnet gravity for a given class of small

perturbations around the vacuum state. We found that for all our examples, the positivity


of the stress tensor two point function ascertained that this inequality was respected. On the

bulk side this corresponds to metric fluctuations having positive energy. The simplicity of the

final result does cry out for a simpler explanation for our findings. Although the intermediate

integrals involved appeared very complicated, the final result was simply proportional to the

Weyl anomaly c. It would be nice to find a simple explanation for this finding. Some

preliminary studies of the general four derivative theory has been made in appendix B.

Another interesting open problem is to consider a disturbance close to the entangling surface.

We were able to consider a disturbance that was localized far from the entangling surface

and show that the relative entropy is positive. Whether the constraints change as one moves

the disturbance closer to the entangling surface is an open problem.

In the previous chapter we considered different entangling surfaces and demanded that

these close off smoothly in the bulk. In Gauss-Bonnet gravity, this led to the coupling

being constrained. Now, suppose we knew how to extend the relative entropy results for the

spherical entangling surface to other surfaces. Then the smoothness criteria above seems to

constrain the coupling of the higher derivative interaction. This suggests that implicitly the

relative entropy inequality knows about this. Since the positivity of relative entropy seems

to rely only on the unitarity of the field theory, this raises the question if there is any conflict

with unitarity if one is outside the allowed region for the coupling. It will be interesting to

investigate this question since apriori there does not appear to be any such conflict in the

dual gravity.

Appendix

A: Positivity of Relative entropy

Here we review the proof in quantum mechanics leading to the positivity of relative entropy.

This can be found in Nielsen and Chuang’s book listed in [1]. We define relative entropy as,

S(ρ|σ) = Tr(ρ ln ρ)− Tr(ρ lnσ) , (6.122)

where ρ and σ are the density matrices of two different states. Now consider their orthonormal

decomposition,

ρ =∑i

pi |i〉〈i| and σ =∑j

qj |j〉〈j| (6.123)

124 6.5. DISCUSSION

where |i〉 and |j〉 may not be the same set of eigenvectors. We can write,

S(ρ|σ) = Tr(ρ ln ρ)− Tr(ρ lnσ) =∑i

〈i| ρ ln ρ |i〉 −∑i

〈i| ρ lnσ |i〉

=∑i

〈i| ρ ln ρ |i〉 −∑i

∑j

〈i| ρ lnσ |j〉〈j|i〉 =∑i

pi ln pi −∑i,j

pi 〈i| lnσ |j〉〈j|i〉

=∑i

pi ln pi −∑i,j

pi ln qj 〈i|j〉〈j|i〉 =∑i

pi ln pi −∑i,j

Pij pi ln qj . (6.124)

In the second line we just inserted 1 =∑

j |j〉〈j|, and in the last line we have used the

notation Pij = 〈i|j〉〈j|i〉. Note that we must have,∑i

Pij =∑j

Pij = 1 . (6.125)

Till here, all that we have used is the unitarity of the theory. Now, lnx is a concave function;

which means we must have,

ln(tx+ (1− t)y) ≥ t ln(x) + (1− t) ln(y) for 0 ≤ t ≤ 1 . (6.126)

It is easy to generalize this to,

ln (x1t1 + x2t2 + ...+ xmtm) ≥ t1 ln(x1) + t2 ln(x2) + ...+ tm ln(xm) (6.127)

wherem∑i=1

ti = 1 and 0 ≤ ti ≤ 1 ∀i ∈ [1,m] .

The equality follows if for some p, we have tp = 1. Using this, and (6.125) we can write,

−∑j

Pij pi ln qj ≥ −pi ln ri where ri =∑j

Pijqj . (6.128)

Hence we get,

S(ρ|σ) ≥∑i

pi ln

(piri

)= −

∑i

pi ln

(ripi

). (6.129)

Now note that, lnx ≥ x− 1. This gives

S(ρ|σ) ≥ −∑i

pi ln

(ripi

)≥ −

∑i

pi

(1− ri

pi

),

=∑i

(pi − ri) = 0 . (6.130)

Hence, S(ρ|σ) ≥ 0 and the equality follows when ρ = σ. To repeat, the only assumption that

went in the proof was the unitarity of the quantum theory. So, whenever we have a unitary

theory we can expect relative entropy to be positive.


B: Relative entropy for R2 theory in shockwave back-

ground

In this section we want to sketch the calculation for the relative entropy in shockwave back-

ground for a general R2 theory9 where the disturbance is located very far away from the

entangling surface. The action for this theory is shown below,

I =

∫d5x√G

(R +

12

L2+L2

2

(λ3R

2 + λ2RABRAB + λ1RABCDR

ABCD))

. (6.131)

In this case, f∞ satisfies 1− f∞+ 13f 2∞(λ1 + 2λ2 + 10λ3) = 0. We start of with the shockwave

metric as given in eq.(6.90) . We have explicitly checked that this is still a solution for the

R2 theory. Next we quote the area functional for this theory [6, 14, 36],

SEE =2π

`3P

∫d3x√h(1+

L2

2(2λ3R+λ2(RABn

Ai n

Bi −

1

2KiKi)+2λ1(RABCDn

Ai n

Bj n

Ci n

Dj −KiabKabi ))

).

(6.132)

Here i denotes the two transverse directions to the co-dimension 2 surface z = f(x1, x2, x3)

and t = 0 and Ki’s are the two extrinsic curvatures along these two directions pulled back

to the surface and a, b are three dimensional indices. Then we proceed in the same way as

before. We set z = z0 =√R2 − r2 . Also as before we set x′1 = 0 and without loss of any

generality and we will expand the integrand around x′2 = ∞ . First we expand upto O(ε3)

which is the linearized term and hence should yield ∆H = ∆S. The expression for ∆(1)S is

∆(1)S =16π2L5R4

15f5/2∞ `3

Px′62

(1 + 2f∞(λ1 − 2(λ2 + 5λ3)) . (6.133)

The λi dependence has packaged into being proportional to c for the general theory [?].

Using the results of [9] (eq.(6.29) in that paper with the typo mentioned in footnote 4 taken

into account), we find that ∆H = ∆S at this order as expected. Then we expand (6.132)

upto ε4 order and pick out the ε4 term which gives us the ∆(2)S . Note that for a general R2

theory the surface term is not known. So we can only do this calculation for the disturbance

located very far away from the entangling surface such that we do not have to consider the

perturbation to the entangling surface as this will contribute to some order higher than ε4.

Further since the extrinsic curvatures are each proportional to ε3 and hence the O(K2) terms

would be proportional to O(ε6), they will not contribute. The result before carrying out the

9The corresponding entropy functional will be useful in studying relative entropy in non-unitary log

CFTs–for recent applications for entanglement entropy in these theories, see [37].

126 REFERENCES

integrations is shown below,

∆(2)S =2π

`3P

∫dx3dx1dx2

[ 3L5

2f5/2∞ Rx′82

((x21 + x2

2 + x23 −R2)[R2(2f∞(23λ1 + λ2 − 10λ3) + 1)

− 60f∞λ1

(x2

1 + x22

)− x2

3(2f∞(23λ1 + λ2 − 10λ3) + 1)])].

(6.134)

Then we perform the integration over x3 which goes from −√R2 − r2 to

√R2 − r2 and

x1 = r cos(θ) , x2 = r sin(θ) . Now after some algebraic manipulation we can write the

integrand as,

∆(2)S = −48π2L5R6(1 + 2f∞(13λ1 + λ2 − 10λ3))

35f5/2∞ `3

Px′82

. (6.135)

Note that this is not proportional to c for this theory. Since for generic values of the couplings

λi, the bulk theory is non-unitary this may not be surprising. This may be indicative of the

fact that rather than depending only on the two point function of the stress tensor, the

higher point functions also contribute as in the second reference in [26]. The bulk theory will

make sense as an effective theory where the couplings are small. In this circumstance, we

can use field redefinitions to make the theory equivalent to Gauss-Bonnet with λ ∝ λ1. For

the Gauss-Bonnet value λ1 = λ3 = λ , λ3 = −4λ it reduces to,

∆(2)S = −48π2L5R6(1− 2f∞λ)

35f5/2∞ `3

Px′82

, (6.136)

which is proportional to c for the GB theory.

References

[1] M. A. Nielsen, I. L. Chuang. Quantum Computation and quantum Information, Cam-

bridge Univ. Press., Cambridge

V. Vedral. Introduction to quantum information science, Oxford University Press, New

York (2006).

[2] D. D. Blanco, H. Casini, L. -Y. Hung and R. C. Myers, “Relative Entropy and Holog-



erty of Entanglement Entropy for Excited States,” Phys. Rev. Lett. 110, no. 9, 091602

(2013) [arXiv:1212.1164].

M. Nozaki, T. Numasawa, A. Prudenziati and T. Takayanagi, “Dynamics of En-

tanglement Entropy from Einstein Equation,” Phys. Rev. D 88, 026012 (2013)

REFERENCES 127

[arXiv:1304.7100 [hep-th]].

M. Nozaki, T. Numasawa and T. Takayanagi, “Holographic Local Quenches and En-

tanglement Density,” JHEP 1305, 080 (2013) [arXiv:1302.5703 [hep-th]].


JHEP 1308, 102 (2013) [arXiv:1305.2728 [hep-th]].

G. Wong, I. Klich, L. A. Pando Zayas and D. Vaman, “Entanglement Temperature and

Entanglement Entropy of Excited States,” JHEP 1312, 020 (2013) [arXiv:1305.3291

[hep-th]].


momentum and U(1) charge,” JHEP 1311, 052 (2013) [arXiv:1306.4974 [hep-th]].

J. Bhattacharya, V. E. Hubeny, M. Rangamani and T. Takayanagi, “Entanglement

density and gravitational thermodynamics,” Phys. Rev. D 91 (2015) 10, 106009

[arXiv:1412.5472 [hep-th]].

[4] H. Casini, “Relative entropy and the Bekenstein bound,” Class. Quant. Grav. 25,

205021 (2008) [arXiv:0804.2182 [hep-th]].

R. Bousso, H. Casini, Z. Fisher and J. Maldacena, “Proof of a Quantum Bousso Bound,”


R. Bousso, H. Casini, Z. Fisher and J. Maldacena, “Entropy on a null surface for

interacting quantum field theories and the Bousso bound,” Phys. Rev. D 91 (2015) 8,

084030 [arXiv:1406.4545 [hep-th]].

[5] R. Haag. Local quantum physics: Fields, particles, algebras - 1992. Springer. Berlin,

Germany: (1992) (Texts and monographs in physics).



J. Camps, “Generalized entropy and higher derivative Gravity,” arXiv:1310.6659 [hep-

th].

[7] N. Lashkari, M. B. McDermott and M. Van Raamsdonk, “Gravitational Dynamics

From Entanglement ”Thermodynamics”,” arXiv:1308.3716 [hep-th].

[8] J. Bhattacharya and T. Takayanagi, “Entropic Counterpart of Perturbative Einstein

Equation,” JHEP 1310, 219 (2013) [arXiv:1308.3792 [hep-th]].

[9] T. Faulkner, M. Guica, T. Hartman, R. C. Myers and M. Van Raamsdonk, “Gravitation

from Entanglement in Holographic CFTs,” arXiv:1312.7856 [hep-th].

128 REFERENCES



th/9903033].



[hep-th]].

[12] X. O. Camanho, J. D. Edelstein and J. M. S. de Santos, “Lovelock theory and the

AdS/CFT correspondence,” arXiv:1309.6483 [hep-th].


holography,” JHEP 1308, 012 (2013), [arXiv:1305.6694 [hep-th]].


squashed cones and holography,” JHEP 1401, 021 (2014), [arXiv:1308.5748 [hep-th]].

[15] T. Jacobson and R. C. Myers, “Black hole entropy and higher curvature interactions,”



and Higher Curvature Gravity,” JHEP 1104, 025 (2011) [arXiv:1101.5813 [hep-th]].



[18] J. Erdmenger, M. Flory and C. Sleight, “Conditions on holographic entangling surfaces

for black hole geometries in higher derivative gravity,” arXiv:1401.5075 [hep-th].

[19] S. de Haro, S. N. Solodukhin and K. Skenderis, “Holographic reconstruction of space-

time and renormalization in the AdS / CFT correspondence,” Commun. Math. Phys.

217 (2001) 595 [hep-th/0002230].

[20] http://people.brandeis.edu/ headrick/HeadrickCompendium.pdf

[21] T. S. Bunch, “Surface terms in higher derivative gravity,” J. Phys. A: Math. Gen. 14

L139 (1981).

R. C. Myers, “Higher Derivative Gravity, Surface Terms And String Theory,” Phys.

Rev. D 36, 392 (1987).

[22] A. Yale, “Simple counterterms for asymptotically AdS spacetimes in Lovelock gravity,”

Phys. Rev. D 84, 104036 (2011) [arXiv:1107.1250 [gr-qc]].

Y. Brihaye and E. Radu, “Black objects in the Einstein-Gauss-Bonnet theory with

REFERENCES 129

negative cosmological constant and the boundary counterterm method,” JHEP 0809,

006 (2008) [arXiv:0806.1396 [gr-qc]].

[23] A. Bhattacharyya, L. -Y. Hung, K. Sen and A. Sinha, “On c-theorems in arbitrary

dimensions,” Phys. Rev. D 86, 106006 (2012) [arXiv:1207.2333 [hep-th]].

D. P. Jatkar and A. Sinha, “New Massive Gravity and AdS4 counterterms,” Phys. Rev.

Lett. 106, 171601 (2011) [arXiv:1101.4746 [hep-th]].

[24] R. C. Myers and A. Sinha, “Seeing a c-theorem with holography,” Phys. Rev. D 82,

046006 (2010) [arXiv:1006.1263 [hep-th]].

R. C. Myers and A. Sinha, “Holographic c-theorems in arbitrary dimensions,”JHEP

1101, 125 (2011) [arXiv:1011.5819 [hep-th]].

[25] J. T. Liu and W. A. Sabra, “Hamilton-Jacobi Counterterms for Einstein-Gauss-Bonnet

Gravity,” Class. Quant. Grav. 27, 175014 (2010) [arXiv:0807.1256 [hep-th]].

[26] D. M. Hofman and J. Maldacena, “Conformal collider physics: Energy and charge

correlations,” JHEP 0805, 012 (2008) [arXiv:0803.1467 [hep-th]].

D. M. Hofman, “Higher Derivative Gravity, Causality and Positivity of Energy in a UV

complete QFT,” Nucl. Phys. B 823, 174 (2009) [arXiv:0907.1625 [hep-th]].

[27] M. Brigante, H. Liu, R. C. Myers, S. Shenker and S. Yaida, “Viscosity Bound Violation

in Higher Derivative Gravity,” Phys. Rev. D 77 (2008) 126006 [arXiv:0712.0805 [hep-

th]].

M. Brigante, H. Liu, R. C. Myers, S. Shenker and S. Yaida, “The Viscosity Bound and

Causality Violation,” Phys. Rev. Lett. 100 (2008) 191601 [arXiv:0802.3318 [hep-th]].

A. Buchel and R. C. Myers, “Causality of Holographic Hydrodynamics,” JHEP 0908

(2009) 016 [arXiv:0906.2922 [hep-th]].

[28] W. -z. Guo, S. He and J. Tao, “Note on Entanglement Temperature for Low Thermal

Excited States in Higher Derivative Gravity,” JHEP 1308, 050 (2013) [arXiv:1305.2682

[hep-th]].

[29] B. Chen and J. -j. Zhang, “Note on generalized gravitational entropy in Lovelock grav-

ity,” JHEP 07, 185 (2013) [arXiv:1305.6767 [hep-th]].


ment entropy proposal,” JHEP 0707, 062 (2007) arXiv:0705.0016 [hep-th]].

[31] D. Mateos and D. Trancanelli, “The anisotropic N=4 super Yang-Mills plasma and its

instabilities,” Phys. Rev. Lett. 107, 101601 (2011) [arXiv:1105.3472 [hep-th]].

130 REFERENCES

D. Mateos and D. Trancanelli, “Thermodynamics and Instabilities of a Strongly Cou-

pled Anisotropic Plasma,” JHEP 1107, 054 (2011) [arXiv:1106.1637 [hep-th]].

[32] J. M. Maldacena, “Eternal black holes in anti-de Sitter,” JHEP 0304, 021 (2003)

[hep-th/0106112].

[33] D. Grumiller, W. Riedler, J. Rosseel and T. Zojer, “Holographic applications of loga-

rithmic conformal field theories,” J. Phys. A 46, 494002 (2013) [arXiv:1302.0280 [hep-

th]].

[34] E. A. Bergshoeff, O. Hohm, J. Rosseel and P. K. Townsend, “Modes of Log Gravity,”


[35] K. Sen, A. Sinha and N. V. Suryanarayana, “Counterterms, critical gravity and holog-

raphy,” Phys. Rev. D 85, 124017 (2012) [arXiv:1201.1288 [hep-th]].



[37] M. Alishahiha, A. F. Astaneh and M. R. M. Mozaffar, “Entanglement Entropy for

Logarithmic Conformal Field Theory,” arXiv:1310.4294 [hep-th].

M. Alishahiha, A. F. Astaneh and M. R. M. Mozaffar “Holographic Entanglement

Entropy for 4D Conformal Gravity,” arXiv:1311.4329 [hep-th].

7 Coding holographic RG flow using

entanglement entropy

7.1 Introduction

In this chapter we will briefly discuss an application of holographic entanglement entropy

(EE). We have observed throughout this thesis that, this quantity is sensitive to UV physics

and the leading divergence obeys the area law. This indicates that the EE is not a well

defined observable in the continuum limit.

In 2d, Casini and Huerta [1] devised a method to extract the universal contribution to the

entanglement entropy. Liu and Mezei [2, 3] generalized this prescription to higher dimensions.

The resulting quantity, known as the renormalized entanglement entropy (REE), is UV finite

and local on the scale of the entangling region. The evolution of REE with respect to the size

of the entangling region can be used as a probe to realize the RG flow. Moreover, for vacuum

states the REE for a spherical entangling region provides a c-function parametrizing the RG

flow 1. We want to compute the REE for states that break Lorentz invariance due to the

presence of a finite charge density. While holographic RG flows for vacuum states correspond

to domain walls (i.e. solutions interpolating between same dimensional AdS) [12, 13, 14], the

flow for charged states can be described in terms of black holes/branes.

Specifically, we want to study the REE for BPS black solutions in N = 2, 4d FI gauged

supergravity. Starting with the work of Cacciatori-Klemm [15] these models have been stud-

ied extensively over the last few years [16, 17, 18], culminating with the full construction of

static BPS solutions for all symmetric models by [19, 20]. These are solutions that interpo-

late between AdS4 and AdS2 × Σk, where Σk is the surface of constant sectional curvature

with k = −1, 0, 1. Since these objects interpolate between different AdS spaces they are

interesting from the holographic perspective. Morever, for the STU model these solutions

have a M-theory realization through an embedding into the de Wit-Nicolai N = 8 theory

[21]. This chapter is organized as follows. In section (7.2) we outline the computation for the

REE for the black brane solutions. In section (7.3), we summarize the BPS black objects in

1 For more applications of REE in the context of holographic RG flows, interested readers are referred to

some of these references [4, 5, 6, 7, 8, 9, 10, 11] and the references there in.

131

132 7.2. RENORMALIZED ENTANGLEMENT ENTROPY

AdS4, followed by the computation for the REE in section (7.4). Finally, in Appendix (7.5),

we discuss the symplectic invariant in N = 2. This chapter is based on the work [22] done

with Dr. Shajid Haque and Dr. Alvaro veliz-Osorio.

7.2 Renormalized Entanglement Entropy

In this section we outline a general procedure to obtain the universal contributions to the

entanglement entropy for quantum systems that can be described holographically by a metric

of the form

ds2 = −a2dt2 + a−2dr2 + b2dΩ2k. (7.1)

In the above expression, dΩ2k is the line element of a surface of constant sectional curvature

k. Clearly we must demand AdS asymptotics. Hence, as r →∞ the metric takes the form

a→ r

l4b→ r

l4. (7.2)

We wish to compute the entanglement entropy for a subsystem A consisting of a disk Σ(R) of

radius R. From the Ryu-Takayanagi prescription [28], we know that this quantity corresponds

to the area of an extremal surface attached to ∂Σ(R) going into the bulk (see Fig. 7.1). For

the metric (7.1) this problem corresponds to the Plateau problem for the functional

S(R) =2π

l2p

∫ R

0

dρ ρ b2√

1 + e−2ψr2 eψ ≡ ab. (7.3)

In the subsequent calculations we will absorb the factor 2πl2p

into S(R). From the above

functional, it can be showed that the profile r(ρ) of the minimal surfaces can be found from

the ODE

r − ψ′ r2 +

(r

ρ− 2

b′

be2ψ

)(1 + e−2ψ r2

)= 0, (7.4)

where ˙ = ∂ρ and ′ = ∂r. This equation is supplemented with the boundary conditions

r(0) = r0 > 0 r(0) = 0. (7.5)

Solutions of equations (7.4) and (7.5) correspond to extremal surfaces attached to the bound-

ary of a disk at infinity and whose tip is at r = r0. Moreover, the depth of the tip can be

related to the size R of the entangling disk at the boundary, R corresponds to the value

of ρ for which r(ρ) → ∞. Therefore, each surface can be labeled either by the size of the

entangling region R or by the depth it reaches in the bulk r0 (see Fig. 7.1).

Once we have found the profile r(ρ) of the minimal surface, we are instructed to plug it

into the functional (7.3) in order to obtain the holographic entanglement entropy. However,

CHAPTER 7. CODING HOLOGRAPHIC RG FLOW USING ENTANGLEMENTENTROPY 133

Boundary

r

R3 R2 R1> >

r3

r2

r1Horizon

Ρ

Figure 7.1: Minimal surfaces in AAdS

one must be careful since the resulting quantity is divergent. We should, therefore, regularize

this integral first. We introduce a UV cut-off in the following way–let ε 1 and restrict the

values of r such thatl4b(r)

> ε. (7.6)

Using r(ρ), we can translate this bulk cut-off into a boundary cut-off, i.e., we must consider

only ρ < Rε, wherel4

b(r(Rε))= ε. (7.7)

Then we can compute the finite quantity

S(R, ε) =

∫ Rε

0

dρA(ρ,R), (7.8)

where A(ρ,R) stands for the integrand of (7.3) evaluated on the solution r(ρ).

In order to systematically obtain the universal contribution to (7.8), which we call here-

after renormalized entanglement entropy, we use the operator introduced in [1, 2]. In four

bulk space-time dimensions, the renormalized entanglement entropy is given by

S(R) ≡(Rd

dR− 1

)S(R, ε). (7.9)

This quantity can be alternatively written as

S(R) = A(R,Rε)

(R∂Rε

∂R

)+

∫ Rε

0

(R∂A∂R−A

). (7.10)

134 7.3. BPS BLACK OBJECTS IN ADS4

As an illustration let us briefly discuss the application of this procedure for pure AdS4.

For this geometry equation (7.4) yields

r(ρ) =l24√

R2 − ρ2. (7.11)

This corresponds to an extremal surface reaching into the bulk until r = l24R−1. Using equa-

tion (7.7) we find

Rε =

√R2 − ε2

2. (7.12)

For this simple case we can compute equation (7.10) explicitly. First we find

A(R, ρ) =Rρ

(R2 − ρ2)3/2, (7.13)

and then ∫ Rε

0

dρA(R, ρ) =R√

R2 −R2ε

− 1, (7.14)

which diverges as we take ε → 0. However, the additional terms in (7.10) contribute as

follows:

A(R,Rε)∂Rε

∂R=

R2

(R2 −R2ε )

3/2∫ Rε

0

dρ∂A∂R

= − R2ε

(R2 −R2ε )

3/2. (7.15)

So we find that for AdS4

S(R) = 1. (7.16)

This result is consistent with the interpretation of the REE as a c-function probing the

holographic renormalization group flow [14, 29, 30, 31, 32].

7.3 BPS black objects in AdS4

We wish to apply the techniques presented in the previous section to an interesting class of

solutions of the form (7.1), namely 14-BPS black objects in N = 2, FI gauged supergravity.

These correspond to zero temperature solutions supported by scalars and abelian gauge fields.

They are parametrized by two vectors of 2nv + 2 real quantities

Γ =

(pI

qI

)and G =

(gI

gI

), (7.17)

where pI and qI are the magnetic and electric charges of the gauge fields while gI and gI are

the parameters of the Fayet-Iliopoulos potential. In (7.17) the index I = 1, . . . , nv + 1, where


nv is the number of vector multiplets considered. These parameters are not independent. In

fact, they are subject to the symplectic constraint

〈G,Γ〉 = k, (7.18)

where just as in (7.1), k labels the horizon topology, e.g., k = −1, 0, 1 indicates spherical,

flat and hyperbolic respectively.

In the following discussion we focus only on the properties of the spacetime metric and

leave the behavior of the scalars and gauge fields aside. In terms of the quantities (7.17), the

warp factors are given by [19, 20]

eψ = a b =(I4(G)1/4 r + 〈G,B〉

)r b =

1

2I4(H)1/4, (7.19)

where I4(V ) is defined in equation (7.33) in the Appendix, and H is a symplectic vector of

linear functions

H = Ar +B. (7.20)

Here, A and B are constant symplectic vectors. The former can be obtained directly from

the FI parameters

A =1

2I4(G)−3/4dI4(G), (7.21)

while the latter is given by a combination of the charges and FI parameters, dictated by the

algebraic equation1

4dI4(B,B,G) = Γ, (7.22)

where dI4 is defined in equation (7.35) in the Appendix. Moreover, physical consistency

requires to choose solutions of (7.22) that fulfill the constraints

〈G,B〉 > 0 I4(B) > 0. (7.23)

Therefore the construction of a BPS solution is reduced to a purely algebraic problem.

The warp factors (7.19) correspond to a metric that interpolates between AdS4 at infinity

(UV) and an AdS2×Σ2k near horizon (IR) geometry as r → 0. The AdS radii corresponding

to these UV and IR geometries are given by

l4 = I4(G)−1/4 l2 =1

2I4(B)1/4〈G,B〉−1. (7.24)

Moreover, the entropy (density for k=0) is proportional to

b2 IR−−−→ σ20 =

1

4

√I4(B). (7.25)

At this point we want to remind the reader that these solutions are also accompanied by

flowing scalars. Due to the attractor mechanism [16, 17] the scalars flow from constant to

constant. These scalars can be thought of as coupling constants and giving rise to the notion

of an attractive RG flow [23].

136 7.4. REE FOR BPS BLACK BRANES

7.4 REE for BPS black branes

In this section we compute the renormalized entanglement entropy as discussed in section

(7.2) for solutions of the kind presented in section 7.3. In the following computation we

restrict ourselves to black objects with flat horizons (k = 0), i.e., black branes. Furthermore,

we will consider solutions of the STU-model. This model captures the essential features

of extremal black holes in N ≥ 2, d = 4 theories [26]. In the STU-model, the structure

constants for the prepotential (see Appendix 7.5) are given by cijk = |εijk|. By plugging the

warp factors (7.19) into equation (7.4), we obtain an explicit, albeit complicated, differential

equation for r in terms of ρ. Now we need to set the parameters that will support the

solution i.e. charges/FI parameters. To each such charge/FI configuration we can associate

three symplectic invariant combinations, which correspond to the AdS length scales and

the entropy density. As we will see in the following our results depend only on these three

quantities. In order for the solution to be regular these quantities must not vanish and we

choose the charges/FI parameters accordingly.

Hereafter we will consider solutions supported by non-vanishing charges/FI parameters

(q0, pi; g0, gi) i = 1, . . . , nv + 1 (7.26)

or

(p0, qi; g0, gi) i = 1, . . . , nv + 1. (7.27)

In the following discussion we will display the results for the first configuration (7.26). For the

second configuration we have verified explicitly that we get completely analogous results. It is

straightforward to extend the following discussion to other configurations as well. Moreover,

the reader must keep in mind that the results that follow are invariant under symplectic

transformations of the kind discussed in [25, 26].

Given the intricacy of the resulting ODE describing the extremal surfaces’ profile, we

proceed to solve it numerically. Moreover, in order to realize the program outlined in section

7.2 we are compelled to produce a large sample of such minimal surfaces ri(ρ) with i being an

index for the sample (see Fig. 7.2). Now by introducing a cut-off ε it is possible to compute

numerically (7.8), creating thus a list of regularized areas Si(ε) corresponding to each of

the extremal surfaces ri(ρ). Hence we are left with a list of points (Ri, Si(ε)), which can be

interpolated to find S(R, ε). Finally, from this function we construct the desired renormalized

entanglement entropy S(R) for a given set of charges/FI parameters. We must point out that

by construction S(R) is a cut-off independent quantity. This behavior is exhibited by our

numerical computations as we tune ε to ever smaller values.

The resulting REE is depicted in Fig. 7.3 for a particular example. However the observed

behavior is generic regardless of the values chosen for the parameters (7.26). First of all, as


0.0 0.2 0.4 0.6 0.80.0

0.1

0.2

0.3

0.4

0.5

r

Ρ

R4 R3 R2 R1

r4

r1

Horizon

Asymptopia

r(Ρ)

:

Figure 7.2: Some extremal surfaces for a p1 = 2, p2 = p3 = 1 and −g0 = g1 = g2 = g3 = 1

black brane

S(R)→ 1 as R→ 0 in agreement with (7.16), the REE then decreases monotonically until it

reaches a minimum S∗ when the entangling disk has a radius R∗. After reaching that critical

value, S(R) starts to increase and approaches the value σ0 as we get closer to the horizon.

0.2 0.4 0.6 0.8

0.5

0.6

0.7

0.8

0.9

UV IR

S(R)

Σ0

RRmin

Smin

Figure 7.3: Renormalized entanglement entropy for a p1 = 2, p2 = 1, p3 = 1/2 and −g0 =

g1 = g2 = g3 = 1 black brane

138 7.5. DISCUSSION

We wish to explore how the values of S∗ and R∗ depend upon the charges/FI parameters

Γ and G. First of all, it is clear that these parameters must enter only through symplectic

invariant combinations. The warp factors (7.19) can be specified in terms of the invariant

quantities I4(G), 〈G,B〉, I4(A), I4(B). From equation (7.24, 7.25) we can identify them as

l2, l4, σ0, I4(A). But a closer inspection shows that I4(A) can be expressed in terms of l4.

Hence, there are three independent invariants upon which S∗ depends, namely the AdS radii

and the entropy density. In order to find a pattern, we start by identifying a subclass of

parameters for which one of the symplectic invariants is held fixed. One such family is given

by

g1 = g2 = g3 p1 = p2 = p3. (7.28)

The crucial point here is that for black branes with these kind of parameters the near-

horizon AdS2 radius l2 is independent of the value of p1. Let’s see how this comes about. For

charge/FI combinations of the form (7.28) the solution of equation (7.22), consistent with

the conditions (7.23), reads

B1 = B2 = B3 = λ1 sgn(g0)

√p1

g1

, B0 = λ2 |g0|−1√p1 g1. (7.29)

Here λ1 and λ2 are known positive constants, and the components of B that are omitted

vanish. The upshot is that B is proportional to√p1. Therefore, since l2 in (7.24) is invariant

under rescalings of B, it is clear that the p1 dependence drops out.

Now we fix the overall scale l4, and study the behavior of S∗. In this context, changing p1

gives rise to a one parameter family of solutions with constant l2 and varying σ0. Interestingly,

we find that S∗ is constant along this family, which implies that S∗ is a function of l2 only.

Notice that in the regime R = 0 to R = R∗ the REE decreases monotonically from the

pure AdS4 value to a constant which can be determined solely from the AdS2 radius. This

is reminiscent of the c-function discussed in [23]. It would be interesting to investigate this

connection further.

Then we explore how S∗ varies with l2. The variation is displayed in Fig (7.4). This plot

clearly shows that S∗ increases with l2. On the other hand, R∗ depends on both l2 and σ0.

Moreover, it increases with l2 and decreases with σ0.

7.5 Discussion

In this chapter we have computed the renormalized entanglement entropy, S for 4d, N = 2

BPS black brane solutions. These solutions interpolate between AdS4 in the UV and a space

with an AdS2 factor in the IR. Specifically, we have investigated the behavior of S as a


0.07 0.08 0.09 0.10 0.11 0.12

0.30

0.35

0.40

0.45

S*

l2

Figure 7.4: Variation of S∗ with l2

function of the size of the entangling region. We have found that in this context S first

decreases monotonically with R, reaches an extremum and then increases again. This is the

key finding of our investigation.

We have already mentioned in the main text that all the parameters of the brane solutions

enter by three independent symplectic invariant combinations that can be identified as the

three independent scales of the system, namely l4, l2 and σ0. In our inspection we found

that starting from the UV the S monotonically decreases until it reaches a minimum that is

determined completely by the radius of AdS2. Then it starts to increase again and approaches

the black brane entropy density. Furthermore, as pointed out in the previous section, when

we increase the entropy of the brane this transition occurs closer to the UV. The fact that

the S decreases monotonically for that region in R resembles the behavior of a c-function for

vacuum states. Then the chemical potential starts to dominate once we go deeper in the IR.

Entanglement entropy measures the entropy due to tracing out part of the total system.

If the total system is in a mixed state this quantity receives contributions both from en-

tanglement and from the mixedness of the original system. Since black branes correspond

to mixed states we expect our computation to be influenced by both of these factors. In

light of that, we are inclined to interpret our result in the following way–the REE is driven

predominantly by entanglement close to the UV before reaching R∗, where contributions due

to the mixedness of the branes take over.

140 7.5. DISCUSSION

Appendix

A: Duality transformations for N = 2 gauged supergrav-

ity

Supergravity solutions can be easily written in terms of symplectic vectors. These are vectors

with 2nv + 2 components of which the first nv + 1 components are labeled with an upper

index and the remaining ones with a lower index, e.g.,

V =

(V I

VI

). (7.30)

I = 0, . . . , nv. These vectors are acted upon by symplectic transformations(V I

VI

)=

(U Z

W V

)(V I

VI

), (7.31)

where

UTV −W TZ = V UT −WZT = 1

UTW = W TU , ZTV = V TZ. (7.32)

We refer to these reparametrizations as duality transformations [25].

In N = 2 models, physical quantities must be invariant under duality transformations.

Duality invariant quantities can be succinctly expressed in terms of the symplectic quartic

invariant I4 [27]

I4 (V ) =1

4!tMNPQVM VN VP VQ

= −(V IVI

)2+

2

3V0 cijkV

iV jV k − 2

3V 0 cijkViVjVk

+ cijk clmnV iV jVlVm. (7.33)

In this expression and the ones to follow, the lower-case indices run from 1 to nv only. The

constant coefficients cijk encode the underlying special geometry prepotential

F = −1

6cijk

X iXjXk

X0. (7.34)

For future convenience, we also define

dI4(V ) = ΩMN∂I4(V )

∂VN, (7.35)

REFERENCES 141

with

ΩMN =

(0 I−I 0

), (7.36)

the canonical symplectic matrix. Moreover, given four symplectic vectors we define

I4

(V (1), V (2), V (3), V (4)

)= tMNPQ V

(1)M V

(2)N V

(3)P V

(4)Q . (7.37)

Notice the absence of the overall symmetrization factor with respect to (7.33). In practice,

we can obtain the t-tensor by hitting (7.33) with four derivatives. The black hole solution is

determined by a set of electric/magnetic charges and fluxes.

References

[1] H. Casini and M. Huerta, “On the RG running of the entanglement entropy of a circle,”


[2] H. Liu and M. Mezei, “A Refinement of entanglement entropy and the number of

degrees of freedom,” JHEP 1304, 162 (2013) [arXiv:1202.2070 [hep-th]].

[3] H. Liu and M. Mezei, “Probing renormalization group flows using entanglement en-

tropy,” JHEP 1401 (2014) 098 [arXiv:1309.6935 [hep-th], arXiv:1309.6935].

[4] R. C. Myers and A. Singh, “Comments on Holographic Entanglement Entropy and RG

Flows,” JHEP 1204 (2012) 122 [arXiv:1202.2068 [hep-th]].

[5] M. Ishihara, F. L. Lin and B. Ning, “Refined Holographic Entanglement Entropy for

the AdS Solitons and AdS black Holes,” Nucl. Phys. B 872 (2013) 392 [arXiv:1203.6153

[hep-th]].

[6] A. Lewkowycz, R. C. Myers and M. Smolkin, “Observations on entanglement entropy

in massive QFT’s,” JHEP 1304 (2013) 017 [arXiv:1210.6858 [hep-th]].

[7] K. Kontoudi and G. Policastro, “Flavor corrections to the entanglement entropy,”

JHEP 1401 (2014) 043 [arXiv:1310.4549 [hep-th]].

[8] K. K. Kim, O. K. Kwon, C. Park and H. Shin, “Renormalized Entanglement En-

tropy Flow in Mass-deformed ABJM Theory,” Phys. Rev. D 90 (2014) 4, 046006

[arXiv:1404.1044 [hep-th]].

[9] T. Nishioka, “Relevant Perturbation of Entanglement Entropy and Stationarity,” Phys.

Rev. D 90 (2014) 4, 045006 [arXiv:1405.3650 [hep-th]].

142 REFERENCES

[10] K. K. Kim, O. K. Kwon, C. Park and H. Shin, “Holographic entanglement entropy of

mass-deformed Aharony-Bergman-Jafferis-Maldacena theory,” Phys. Rev. D 90 (2014)

12, 126003 [arXiv:1407.6511 [hep-th]].

[11] A. Singh, “Holographic Entanglement Entropy: RG Flows and Singular Surfaces,”

CITATION = INSPIRE-1263730 and references there in.

[12] L. Girardello, M. Petrini, M. Porrati and A. Zaffaroni, “The Supergravity dual of N=1

superYang-Mills theory,” Nucl. Phys. B 569, 451 (2000) [hep-th/9909047].

[13] D. Z. Freedman, S. S. Gubser, K. Pilch and N. P. Warner, “Renormalization group

flows from holography supersymmetry and a c theorem,” Adv. Theor. Math. Phys. 3,

363 (1999) [hep-th/9904017].

[14] R. C. Myers and A. Sinha, “Holographic c-theorems in arbitrary dimensions,” JHEP

1101, 125 (2011) [arXiv:1011.5819 [hep-th]].

[15] S. L. Cacciatori and D. Klemm, “Supersymmetric AdS(4) black holes and attractors,”

JHEP 1001, 085 (2010) [arXiv:0911.4926 [hep-th]].

[16] G. Dall’Agata and A. Gnecchi, “Flow equations and attractors for black holes in N =

2 U(1) gauged supergravity,” JHEP 1103, 037 (2011) [arXiv:1012.3756 [hep-th]].

[17] S. Barisch, G. Lopes Cardoso, M. Haack, S. Nampuri and N. A. Obers, “Nernst branes

in gauged supergravity,” JHEP 1111, 090 (2011) [arXiv:1108.0296 [hep-th]].

[18] S. Barisch-Dick, G. L. Cardoso, M. Haack and A. Veliz-Osorio, “Quantum corrections

to extremal black brane solutions,” JHEP 1402, 105 (2014) [arXiv:1311.3136 [hep-th]].

[19] S. Katmadas, “Static BPS black holes in U(1) gauged supergravity,” JHEP 1409 (2014)

027 [arXiv:1405.4901 [hep-th]].

[20] N. Halmagyi, “Static BPS Black Holes in AdS4 with General Dyonic Charges,”


[21] B. de Wit and H. Nicolai, “The Consistency of the S**7 Truncation in D=11 Super-

gravity,” Nucl. Phys. B 281, 211 (1987).

[22] A. Bhattacharyya, S. Shajidul Haque and A. Veliz-Osorio, Phys. Rev. D 91 (2015) 4,

045026 [arXiv:1412.2568 [hep-th]].

[23] A. Bhattacharyya, S. S. Haque, V. Jejjala, S. Nampuri and A. Veliz-Osorio, “Attractive

holographic c-functions,” arXiv:1407.0469 [hep-th].

REFERENCES 143

[24] S. Bellucci, S. Ferrara, A. Marrani and A. Yeranyan, “stu Black Holes Unveiled,”

Entropy 10, 507 (2008) [arXiv:0807.3503 [hep-th]].

[25] B. de Wit, “Electric magnetic dualities in supergravity,” Nucl. Phys. Proc. Suppl. 101,

154 (2001) [hep-th/0103086].

[26] S. Bellucci, S. Ferrara, A. Marrani and A. Yeranyan, “d=4 Black Hole Attractors

in N=2 Supergravity with Fayet-Iliopoulos Terms,” Phys. Rev. D 77, 085027 (2008)

[arXiv:0802.0141 [hep-th]].

[27] S. Ferrara and M. Gunaydin, “Orbits of exceptional groups, duality and BPS states in

string theory,” Int. J. Mod. Phys. A 13, 2075 (1998) [hep-th/9708025].


AdS/CFT,” Phys. Rev. Lett. 96, 181602 (2006) [hep-th/0603001].

[29] D. L. Jafferis, I. R. Klebanov, S. S. Pufu and B. R. Safdi, “Towards the F-Theorem:

N=2 Field Theories on the Three-Sphere,” JHEP 1106 (2011) 102 [arXiv:1103.1181

[hep-th]].

[30] I. R. Klebanov, S. S. Pufu and B. R. Safdi, “F-Theorem without Supersymmetry,”

JHEP 1110 (2011) 038 [arXiv:1105.4598 [hep-th]].

[31] I. R. Klebanov, T. Nishioka, S. S. Pufu and B. R. Safdi, “Is Renormalized Entanglement

Entropy Stationary at RG Fixed Points?,” JHEP 1210 (2012) 058 [arXiv:1207.3360

[hep-th]].

[32] I. R. Klebanov, T. Nishioka, S. S. Pufu and B. R. Safdi, “On Shape Dependence and

RG Flow of Entanglement Entropy,” JHEP 1207 (2012) 001 [arXiv:1204.4160 [hep-th]].

144 REFERENCES

8 Conclusions

In this thesis we have tried to understand the interplay between quantum information and

geometry through entanglement entropy (EE) in the framework of AdS/CFT correspondence.

It has provided us with a nice geometrical problem and has the merit to shed light on how

holography works.

Ryu-Takayanagi (RT) proposal has provided us a way to compute EE from holography

for Einstein gravity. In this thesis we have discussed various generalization of RT proposal

for higher derivative gravity theories. These higher derivative theories are interesting as they

extend the AdS/CFT dictionary beyond supergravity limit and incorporate the effect of finite

coupling. Recently Lewkowycz and Maldacena (LM) have proposed to set this calculation

in a self consistent way, solidifying its geometrical connection. Their idea stems from the

fact that there exists a striking similarity between EE and the black hole entropy. Using LM

technique one can formulate a proof for RT proposal and but unfortunately this cannot be

generalized for more general theories of gravity, it seems that it only works for certain class

of gravity theories. It will be nice find a way to overcome this difficulty in future.

On the other hand there is a well defined way of deriving the black hole entropy for any

diffeomorphism invariant theory based on the Nother charge method proposed by Iyer and

Wald. Based on this Noether charge construction Wald has proposed a general formula for

black hole entropy for any diffeomorphism invariant theory. For Einstein gravity both the

black hole entropy and EE follows area law. Now the question is, can one simply use Wald

entropy functional to compute EE for any general theories of gravity? The answer is no,

as Wald entropy functional is valid only for bifurcation surface and for general surface one

has to systematically modify it to get correct EE as predicted form the holography. Now it

seems that one can connect EE with this Noether charge method and in this thesis we have

partially achieved that, although still a rigorous proof is needed. Recently it has been shown

that indeed one can connect EE with the Noether charge construction by demanding that the

entropy functionals used to evaluate EE for general theories of gravity satisfy second law of

thermodynamics (a generalized version of it). So it will be very nice to make this connection

more rigorous in the future as it will provide a solid geometrical interpretation for EE .

Further we have used holographic entanglement to constrain gravity in interesting ways.

First, we have used positivity of relative entropy and analysed the inequality ∆S ≤ ∆H

145

146

for Gauss-Bonnet gravity for a given class of small perturbations around the vacuum state.

We found that for all our examples, the positivity of the stress tensor two point function

ascertained that this was respected. On the bulk side this corresponds to metric fluctuations

having positive energy. The simplicity of the final result does cry out for a simpler explanation

for our findings. Although the intermediate integrals involved appeared very complicated,

the final result was simply proportional to the Weyl anomaly c. It would be nice to find a

simple explanation for this finding. Unfortunately the relative entropy calculation is limited

to the spherical entangling surface as for this kind of surface only, the modular hamiltonian

(H) is known.

Finally, we also considered other entangling surfaces and demanded that these close off

smoothly in the bulk. In Gauss-Bonnet gravity, this led to the coupling being constrained.

The spherical entangling surface did not lead to any constraints on the coupling while the

cylindrical and slab entangling surfaces did. This leads to an interesting question. Suppose

we knew how to extend the relative entropy results for the spherical entangling surface to

other surfaces. Then the smoothness criteria above seems to constrain the coupling of the

higher derivative interaction. This suggests that implicitly the relative entropy inequality

knows about this. Since the positivity of relative entropy seems to rely only on the unitarity

of the field theory, this raises the question if there is any conflict with unitarity if one is

outside the allowed region for the coupling. It will be interesting to investigate this question

since apriori there does not appear to be any such conflict in the dual gravity. It will also be

interesting to find if there are other entangling surfaces which lead to a tighter bound and if

the bounds are stronger than the causality constraints.

Then at the end we have discussed how to code holographic RG flow using EE. We have

investigated the behaviour of renormalized entanglement entropy (REE) in the context of a

lorentz violating RG flow. We have found that REE first monotonically decreases and then

increases smoothly along the RG flow, thus exhibiting a minima. It would be interesting to

see if we can use the behavior of REE as a function of entangling surface in order to establish

an order parameter for the phase transition between the vacuum to vacuum flow and vacuum

to charged state flow. This is because the existence of the extremum for the REE tells us

that the system is transiting from its vacuum behavior at that point.

Form all this analysis it is quite evident that entropy functionals used to evaluate EE

play a central role. In recent times, many different tools of quantum entanglement like,

entanglement negativity, differential entropy, quantum error coding, relative entropy, cMERA

etc, have been used in an attempt to build geometry from the field theory data, is some

sense trying to prove the holographic principle. In the light of these recent advances, these

holographic entropy functionals play a crucial role and a thorough understanding of these

entropy functionals will teach us many important lessons about gravity.

Documents

Lessons for gravity from entanglementchep.iisc.ernet.in/Personnel/pages/asinha/arpan_thesis.pdf · Arpan Bhattacharyya Centre for High Energy Physics Indian Institute of Science Bangalore