配色: 字号:
Black Hole Entropy and Viscosity Bound in Horndeski Gravity
2016-04-24 | 阅:  转:  |  分享 
  
arXiv:1509.07142v2[hep-th]4Oct2015

MI-TH-1533

BlackHoleEntropyandViscosityBound

inHorndeskiGravity

Xing-HuiFeng1,Hai-ShanLiu2,3,H.L¨u1andC.N.Pope3,4

1CenterforAdvancedQuantumStudies,DepartmentofPhysics,

BeijingNormalUniversity,Beijing100875,China

2InstituteforAdvancedPhysics&Mathematics,

ZhejiangUniversityofTechnology,Hangzhou310023,China

3GeorgeP.&CynthiaWoodsMitchellInstituteforFundamentalPhysicsandAstronomy,

TexasA&MUniversity,CollegeStation,TX77843,USA

4DAMTP,CentreforMathematicalSciences,CambridgeUniversity,

WilberforceRoad,CambridgeCB3OWA,UK

ABSTRACT

Horndeskigravitiesaretheoriesofgravitycoupledtoascalarfield,inwhichtheac-

tioncontainsanadditionalnon-minimalquadraticcouplingofthescalar,throughitsfirst

derivative,totheEinsteintensorortheanalogoushigher-derivativetensorscomingfrom

thevariationofGauss-BonnetorLovelockterms.Inthispaperwestudythethermody-

namicsofthestaticblackholesolutionsinndimensions,inthesimplestcaseofaHorndeski

couplingtotheEinsteintensor.WeapplytheWaldformalismtocalculatetheentropyof

theblackholes,andshowthatthereisanadditionalcontributionoverandabovethosethat

comefromthestandardWaldentropyformula.Theextracontributioncanbeattributed

tounusualfeaturesinthebehaviourofthescalarfield.Wealsoshowthataconventional

regularisationtocalculatetheEuclideanactionleadstoanexpressionfortheentropythat

disagreeswiththeWaldresults.Thisseemslikelytobeduetoambiguitiesinthesub-

tractionprocedure.WealsocalculatetheviscosityinthedualCFT,andshowthatthe

viscosity/entropyratiocanviolatetheη/S≥1/(4π)boundforappropriatechoicesofthe

parameters.

xhfengp@mail.bnu.edu.cnhsliu.zju@gmail.commrhonglu@gmail.compope@physics.tamu.edu

Contents

1Introduction2

2BlackHolesinHorndeskiGravity5

2.1Thetheory....................................5

2.2Staticblackholesolutions............................6

2.3UniquenessoftheHorndeskiblackholesolutions...............9

3BlackHoleEntropyandThermodynamics11

3.1Waldentropyformula..............................11

3.2Waldformalism..................................12

3.3FurthercommentsontheentropyfromWaldformalism...........16

3.4NoetherchargeandmassofAdSplanarblackholes..............18

3.5Euclideanaction.................................19

4Viscosity/EntropyRatio22

5Conclusion24

1Introduction

Inthedictionaryofgravity/gaugedualitymappingsintheAdS/CFTcorrespondence[1–3],

perturbationsofthemetricarerelatedtotheenergy-momentumtensorofthefieldtheory

intheboundaryoftheAdSspacetime[2–4].Inthispicture,anAdSplanarblackholeis

thegravitationaldualofacertainidealfluid.Awidelyvalidrelationbetweentheshear

viscosityandtheentropydensitywasestablished,namely[5–8]

η

S=

1

4π.(1.1)

Onewaytounderstandthisratioisthatitcanbeshownthattheviscosityisproportionalto

thecross-sectionoftheblackholeforlow-frequencymasslessscalarfields[8].Alternatively,

theshearviscosityisdeterminedbytheeffectivecouplingconstantofthetransversegraviton

onthehorizon,byemployingthemembraneparadigm[9].(Thiswasconfirmedbyusingthe

Kuboformulain[10,11].)In[12],itwasshownthattheblackholeentropyisdetermined

bytheeffectiveNewtoniancouplingatthehorizon,andthatitisthusnotsurprisingthat

theratiooftheshearviscositytotheentropydensityisuniversal,inthesensethatthe

dependenceofthequantitiesonthehorizoniscanceled.Recently,itwasestablishedthat

2

therelation(1.1)oftheboundarytheoryisdualtoageneralisedSmarrrelationobeyedby

thebulkAdSplanarblackholes,therebyprovidinganewunderstandingofitsuniversality,

anditsconnectiontotheblackholethermodynamics[13].Therehavebeenanumberof

papersinliteratureestablishingtheuniversalityoftheratio(1.1)[14–17].(See[18]fora

review.)

Theviscosity/entropyratio(1.1)can,however,beviolatedwhenthebulkgravitytheory

isextendedbytheadditionofhigher-ordercurvatureterms[19,20].1(Seealso,forfurther

examples,[25–27].)

Thisleadsustooneofthemotivationsforthispaper,whichistoinvestigatewhether

onecanviolatetheratio(1.1)withoutintroducinghigher-ordercurvaturetermsinthebulk

theory.InatypicaltheoryofEinsteingravity,matterfieldscoupletogravityminimally

throughthemetric.Ascalarfieldcanalsocoupletogravitynon-minimally,suchasin

Brans-Dicketheory[28],wheretheeffectiveNewtonconstantvariesinspacetime.However,

itwasestablishedin[13]thattheratio(1.1)holdsingeneralinsuchatheory.Scalar

fieldscan,however,alsocouplenon-minimallytogravityinotherways.Inparticular,

theirderivativescancoupletothecurvaturetensor.Horndeskiconsideredawideclass

ofsuchgravity/scalartheoriesintheearlyseventies[29],focusinghisattentiononcases

wherethefieldequations,bothforgravityandthescalarfield,involvenohigherthan

secondderivatives.TheHorndeskitheorieswererediscoveredrecentlyinstudiesofthe

covariantisationofGalileontheories[30].

TheHorndeskitermstaketheform

H(k)=E(k)μν?μχ?νχ,(1.2)

wheretheE(k)tensorsare“energy-momentumtensors”associatedwiththeEulerintegrands

ofvariousorder,namely

E(k)νμ≡δνρ1···ρ2kμσ1···σ2kRσ1σ2ρ1ρ2···Rσ2k?1σ2kρ2k?1ρ2k.(1.3)

TheH(k)termsareanalogoustoEulerintegrands,inthattheyhavethepropertythat

eachfieldcarriesnomorethanasinglederivativeandhencethelinearizedequationsof

motioninvolveatmostsecondderivatives.Thusalthoughthetheoryinvolveshigher-order

derivatives,itcontainsnolinearghostexcitations.Inthispaper,weshallconsiderEinstein

gravitywithacosmologicalconstant,togetherwithjustthetwolowest-orderHorndeski

1Weshallnotbeconcernedinthispaperwithothertypesofviolation,duetothebreakingoflocal

rotationalsymmetry;see,forexample,[21–24].

3

terms,namely

H(0)=gμν?μχ?μχ,H(1)=?4Gμν?μχ?νχ,(1.4)

whereGμνistheEinsteintensor.Wefindthatalthoughthetheorycontainsthecurvature

tensoronlylinearly,theviscosity/entropyratio(1.1)nolongerholds.

Itisworthcommentingthattheviscositycanbecomputedbystandardprocedures

usingtheAdS/CFTcorrespondence,involvingthestraightforwardtechniqueofstudying

linearisedperturbationsaroundthebackgroundbulksolution.Thecalculationofthevis-

cosity/entropyratiothenhingesupontheproperdefinitionoftheentropyoftheblack

hole.SinceHawkingestablishedthethermalradiationofablackhole[31,32],therehas

beennoambiguityinestablishingtheblackholeentropyinagenerally-covarianttheory.

Inparticular,inEinsteingravityminimallycoupledtomatter,theentropyisgivenbyone

quarteroftheareaofthehorizon.ThisarealawhasbeengeneralizedtotheWaldentropy

formulawhenmorecomplicatedcouplingsorhigher-ordercurvaturetermsareinvolved,

namely[33,34]

SW=?18

integraldisplay

+

dn?2x



h?L?Rabcd?ab?cd.(1.5)

whereLisdefinedbytheactionI=integraltextdnx√?gL.Applyingthisformulatostaticblackholes

withspherical,toricorhyperbolicisometries,theHorndeskiterms(1.4)donotcontribute

totheWaldentropySW,andhenceonemightexpectthattheentropywouldstillbejust

onequarterofthehorizonarea.However,wefindthatthisisinfactnotthecase.By

examiningtheWaldprocedure[33,34]indetail,wefindthatinatheorysuchasHorndeski

gravitythereisanadditionalcontributiontotheentropythatisnotencompassedbythe

usualWaldformula(1.5).Itarisesbecausethederivativeofthescalarfielddivergeson

thehorizonintheblack-holesolutions(althoughthereisnophysicaldivergence,sinceall

invariants,suchasgμν?μχ?νχ,remainfinite).

Thepaperisorganisedasfollows.Insection2weintroducetheHorndeskitheorythat

weshallbeconsidering,andwereviewthestaticblackholesolutions.Theseareknown

forallthecasesofspherical,toroidalandhyperbolichorizongeometries.Ourfocuswill

beonthesphericalandthetoroidalhorizons.Wealsoincludeademonstrationofthe

uniquenessoftheknownstaticsolutions.Insection3weaddresstheproblemofcalculating

theentropy,andalsothemass,ofthestaticblackholes.Webeginbycalculatingthe

entropyusingthestandardWaldformula(1.5),andthenweconsidertheapplicationofthe

Waldformalisminmoredetail,showingthatthereisanothercontributiontotheentropy

thatisnotcapturedby(1.5).Weshowthatinthecaseoftheplanarblackholes(with

toroidalhorizons),theentropyexpressionweobtainisconsistentwiththecomputation

4

oftheNoetherchargeassociatedwithascalingsymmetryoftheblackholes.Wealso

considerthecalculationoftheEuclideanaction,showingthat,atleastwhenfollowinga

naiveregularisationprocedure,thisyieldsyetanotherresultfortheentropy,andthemass,

thatdisagreeswiththosefromtheWaldformalism.Insection4wecalculatetheshear

viscosityinthedualboundarytheoryusingtheAdS/CFTcorrespondence,andhencewe

obtainanexpressionfortheviscosity/entropyratio.Thisisdifferentfrom1/(4π)onaccount

oftheHorndeskiterm,andweshowthatforanappropriatechoiceoftheparametersitcan

violatetheη/S≥1/(4π)bound.Thepaperendswithconclusionsinsection5.

2BlackHolesinHorndeskiGravity

2.1Thetheory

Aswehavediscussedintheintroduction,Horndeskigravityrepresentsaclassofhigher-

derivativetheoriesinvolvinggravitywithanon-minimallycoupledscalar.Thecouplings

differfromthoseintheBrans-Dicketheory,sinceintheHorndeskitheoriesthescalar

couplesthroughitsderivativetothecurvaturetensors.WeshallfocusontheHorndeski

theorywhoseLagrangianinvolvesatmostonlylinearcurvatureterms.Asweshallshow,

theviscosity/entropyratio(1.1)canbeviolatedeveninsuchatheory.Theactionisgiven

by

I=116π

integraldisplay

dnx√?gL,L=κ(R?2Λ)?12(αgμν?γGμν)?μχ?νχ,(2.1)

whereκ,αandγarecouplingconstants,andGμν≡Rμν?12RgμνistheEinsteintensor.

Notethatthetheoryisinvariantunderaconstantshiftofχ.Inatypicalgravitytheory

withascalarfield,suchasBrans-Dicketheory,onecandefinedifferentmetricframesby

meansofconformalscalingsusingthescalarfield.However,fortheHorndeskitheory(2.1),

thiswouldleadtothebreakingofthemanifestconstantshiftsymmetryofthescalar,and

henceitwouldnotbeanaturalfieldredefinitiontomakehere.

Thevariationoftheaction(2.1)givesriseto

δI=116π

integraldisplay

dnx√?g(Eμνδgμν+Eδχ+?μJμ).(2.2)

where

Eμν=κ(Gμν+Λgμν)?12α

parenleftBig

?μχ?νχ?12gμν(?χ)2

parenrightBig

?12γ

parenleftBig

1

2?μχ?νχR?2?ρχ?(μχRν)

ρ

??ρχ?σχRμρνσ?(?μ?ρχ)(?ν?ρχ)+(?μ?νχ)squareχ+12Gμν(?χ)2

?gμνbracketleftbig?12(?ρ?σχ)(?ρ?σχ)+12(squareχ)2??ρχ?σχRρσbracketrightbig

parenrightBig

,

5

E=?μparenleftbig(αgμν?γGμν)?νχparenrightbig.(2.3)

Thetotalderivativetermin(2.2)playsnoroleintheequationsofmotion

Eμν=0,E=0.(2.4)

However,itdoesplayanimportantroleintheWaldformalism,whichweshallpresentin

section3.2.

2.2Staticblackholesolutions

Wenowconsiderstaticblackholes,withtheansatz

ds2n=?h(r)dt2+dr

2

f(r)+r

2d?2

n?2,?,χ=χ(r),(2.5)

whered?2n?2,?with?=1,0,?1isthemetricfortheunitSn?2,then-torusortheunit

hyperbolicn-space.Itisconvenienttotaked?2n?2,?=ˉgijdyidyjforgeneralvaluesof?to

bethemetricofconstantcurvaturesuchthatitsRiccitensorisgivenbyˉRij=(n?3)?ˉgij.

Wemay,forexample,taked?2n?2,?tobegivenby

d?2n?2,?=du

2

1??u2+u

2d?2

n?3,(2.6)

whered?2n?3isthemetricoftheunit(n?3)-sphere.

Itisclearfromtheequationsofmotionthatχ=χ0(constant)isasolution,inwhich

case,theHorndeskigravityreducestoEinsteingravitywithacosmologicalconstantΛ0.It

followsthattheSchwarzschild-AdSblackholeisasolutionofthetheory.Weshallregard

thissolutionasbeing“trivial,”inthesenseofnotyieldinganythingnew.Inaddition,a

one-parameterfamilyofblackholesolutionsforwhichthescalarfieldisnotaconstantwas

constructedin[35].(Seealso,[36,37].)Inthissection,wewouldliketoprovethatthese

aretheonlyblackholesolutionsfromtheansatz(2.5)inwhichthescalarisr-dependent.

First,wereviewtheconstructionin[35].

ThescalarequationofmotionE=0yields

parenleftBig

rn?4

radicalbigg

f

h

parenleftBig

γparenleftbig(n?2)rfh′+(n?2)(n?3)(f??)hparenrightbig?2αr2h

parenrightBig

χ′

parenrightBig′

=0.(2.7)

TherearetwomoreequationsthatfollowfromEμν=0:



parenleftBig

(n?2)rf′+(n?2)(n?3)(f??)+2Λ0r2

parenrightBig

+2αr2fχ′2

+γ(n?2)

parenleftBig

4rfχ′′+parenleftbig3rf′+(n?3)(f+?)parenrightbigχ′

parenrightBig

fχ′=0,

6



parenleftBig

(n?2)rfh′+(n?2)(n?3)h(f??)+2Λ0r2h

parenrightBig

?2αr2fhχ′2

+γ(n?2)

parenleftBig

3rfh′+(n?3)(3f??)h

parenrightBig

fχ′2=0.(2.8)

In[35],aclassofblackholesolutionwasobtainedbysolving(2.7)bytaking

γparenleftbig(n?2)rfh′+(n?2)(n?3)(f??)hparenrightbig?2αr2h=0.(2.9)

(Inotherwords,theintegrationconstantinthefirstintegralof(2.7)wastakentobezero,

andχ′wasallowedtobenon-zero,thusimplyingthatitsco-factor,givenin(2.9),mustbe

equaltozero.)Thisleadstothesolution

h=?μrn?3+8κ[g

2r2(2κ+βγ)+2?κ]

(4κ+βγ)2

+(n?1)

2β2γ2g4r4

?(n+1)(n?3)(4κ+βγ)22F1

bracketleftBig

1,12(n+1);12(n+3);?n?1(n?3)?g2r2

bracketrightBig

,

f=(4κ+βγ)

2bracketleftbig(n?1)g2r2+(n?3)?bracketrightbig2

bracketleftbig(n?1)(4κ+βγ)g2r2+4(n?3)?κbracketrightbig2h,χ′2=βf

bracketleftBig

1+(n?3)?(n?1)g2r2

bracketrightBig?1

,(2.10)

whichisvalidforallvaluesof?.Inpresentingthesolution,wehaveintroducedtwoparam-

eters(g,β)inplaceoftheoriginalparameters(α,Λ0)intheLagrangian,with

α=12(n?1)(n?2)g2γ,Λ0=?12(n?1)(n?2)g2

parenleftBig

1+βγ2κ

parenrightBig

.(2.11)

Notethatthesolutioncontainsonlyoneintegrationconstant,μ.Allotherparameters

arethoseofthetheoryitself.Notealsothatsincethedimensionnisaninteger,the

hypergeometricfunctionreducestopolynomialswithanarctanfunctioninevendimensions,

andwithalogfunctioninodddimensions.Tobeexplicit,wehave

n=even:(2.12)

2F1[1,12(n+1);12(n+3);?x]=

(?1)n/2(n+1)

xn/2

braceleftBigarctan√x

√x?

bracketleftBigarctan√x

√x

bracketrightBig

n

2?1

bracerightBig

,

n=odd:(2.13)

2F1[1,12(n+1);12(n+3);?x]=

(?1)n?12(n+1)

2xn?12

braceleftBiglog(1+x)

x?

bracketleftBiglog(1+x)

x

bracketrightBig

n?3

2

bracerightBig

,

whereweusethenotation[F(x)]mtodenotethetruncatedpowerseriesexpansionofF(x)

aroundx=0,inwhichonlythetermsuptoandincludingxmareretained.Thus

bracketleftBigarctan√x

√x

bracketrightBig

n

2?1

=

n/2?1summationdisplay

p=0

(?x)p

2p+1,

bracketleftBiglog(1+x)

x

bracketrightBig

n?3

2

=

n?3

2summationdisplay

p=0

(?x)p

p+1,(2.14)

fornevenandnodd,respectively.

7

Forstaticsolutionsofthiskind,itisinfactalwayssufficienttoconstructthesolution

with?=1.Thesolutionsforallothervaluesof?,whichwepresentedabove,canthenbe

obtainedfromthe?=1solutionbymeansoftherescalings

r?→r√?,t?→√?t,d?2n?2?→?d?2n?2,?,μ?→??(n?1)/2μ(2.15)

Fromnowon,weshallpresentresultsforthetwospecificcases?=0and?=1.

?=0solution:

When?=0,thesolutionreducestotheverysimpleform

h=f=g2r2?μrn?3,χ′2=βf.(2.16)

Notethatinthis?=0case,χcanbesolvedforexplicitly,giving

χ=2

√β

(n?1)glog

parenleftbigradicalbig(gr)n?1+radicalbig(gr)n?1?μgn?3parenrightbig+χ

0.(2.17)

Thusthe?=0solutiondescribesanAdSplanarblackhole,withtherequirementsthat

μ>0andβ≥0.Thehorizonradiusr=r0isgivenbyμ=g2rn?10.TheHawking

temperatureisgivenby

T=(n?1)g

2

4πr0.(2.18)

?=1solution:

For?=1,thesolutiondescribesaspherically-symmetricandstaticblackhole.Ina

large-rexpansion,ifniseventhefunctionshandfhavetheasymptoticforms

h=g2r2?μrn?3+

summationdisplay

k=0

ck

r2k=g

2r2+4κ?βγ

4κ+βγ?+···,

f=g2r2?μrn?3+

summationdisplay

k=0

dk

r2k=g

2r2+4(n?1)κ+(n?5)βγ

(n?1)(4κ+βγ)?+···,(2.19)

where(ck,dk)areconstants,whicharefunctionsoftheparameters(κ,g,β)butindependent

ofμ.Ifnisodd,thenfork=(n?3)/2,thequantityckhasanadditionaltermproportional

tologr.Thisamountstoalogarithmicallydivergingadditiontothemasscoefficientμat

order1/rn?3.Thisinturnimpliesthatdkhasadditionallogrtermsforallk≥(n?3)/2.

Notethatallthe(ck,dk)vanishfor?=0.

ThemetricisasymptoticlocallytoAdSspacetime,anditcannotbecomepureAdS

spacetime,regardlessofthechoiceoftheparameterμ.Toseethatthesolutiondescribesa

blackhole,wenotethathispositiveasrgoestoinfinity,butbecomesoforder?μ/rn?3

8

asr→0,wherethereisaspacetimecurvaturesingularity.Thuswhenμ>0,theremust

existsomeintermediatevalueofr,beaneventhorizonr=r0,forwhich

h(r0)=0=f(r0).(2.20)

Thisimpliesthattheparameterμcanbeexpressedintermsofthehorizonradiusr0inthis

?=1caseas

μ=8κr

n?3

0

(4κ+βγ)2

parenleftBig

2κ+(2κ+βγ)g2r20

+(n?1)

2β2γ2g4r40

8κ(n?1)(n?3)2F1[1,

1

2(n+1);

1

2(n+3);?

n?1

(n?3)g

2r2

0]

parenrightBig

.(2.21)

Notethatthisrelationbetweenμandr0isfarmorecomplicatedthanthesimpleexpression

μ=g2rn?10thatholdsinthe?=0case.Thetemperatureofthe?=1blackholeisgiven

by

T=

radicalbigh′(r

0)f′(r0)

4π=

(n?1)g2

4πr0+

(n?3)κ

π(4κ+βγ)r0.(2.22)

Notethatifwesetμ=0,thenthesolutionhasnoeventhorizon,andnearr=0the

functionsh,fandχhavetheforms

h=16κ

2

(4κ+βγ)2

parenleftBig

1+(2κ+βγ)g

2r2

2κ+···

parenrightBig

,

f=1+((n?3)κ?βγ)g

2r2

(n?3)κ+···,

χ=χ0+(n?1)β2(n?3)gr2+···.(2.23)

Thustheμ=0solutionisasmoothspherically-symmetricsoliton,withoutanyfreeparam-

eters,thatisasymptoticlocallytoAdSspacetime.Therealsoexistsasolutionfor?=1in

thelimitof4κ+βγ=0,butitdoesnotdescribeablackhole.

2.3UniquenessoftheHorndeskiblackholesolutions

Weshallleavethediscussionofthemassandentropyoftheblackholestothenextsection.

Toclosethissection,weshallshowthatthesolutionsdiscussedaboveareinfacttheonly

blackholeswithnon-constantχthatarecontainedwithintheansatz(2.5)inthetheory.

Toshowthis,wereturntotheequationofmotion(2.7)forthescalarfield.Onecan

immediatelywritedownthefirstintegral

χ′=qr

4?nradicalbigh/f

γparenleftbig(n?2)rfh′+(n?2)(n?3)(f??)hparenrightbig?2αr2h,(2.24)

whereqisanintegrationconstant.Thesolutionswediscussedabovewereobtainedby

takingq=0.Itwaspossibletofindsuchsolutionswithχ′negationslash=0byimposingtherelation

9

(2.9),whichinfactrenderedthescalarequationofmotion(2.7)trivial.Ifinsteadwetake

theintegrationconstantqtobenon-zero,thenχ′isnowdeterminedby(2.24).

Ifasolutionwithqnegationslash=0istodescribedescribeablackhole,theremustbeanevent

horizonatsomeradiusr=r0.ThefunctionshandfnearthehorizonwillhaveTaylor

expansionsoftheform

f=f1(r?r0)+f2(r?r0)2+···,h=h1(r?r0)+h2(r?r0)2+···.(2.25)

Itfollowsfrom(2.24)thatχ′nearthehorizonhastheexpansion

χ′=?χ?1r?r

0

+?χ0+?χ1(r?r0)+···.(2.26)

Substitutingtheseexpansionsintotheotherequationsofmotion,wefindthatnosuch

solutionscanexist.Inotherwords,theassumptionthatthereexistsahorizon,nearwhich

theexpansions(2.25)wouldhold,isinconsistentwiththeequationsofmotionwhenqnegationslash=0..

Inordertohaveasolutionwithahorizon,wemustthereforesetq=0,whichthenreduces

tothepreviouscasediscussedabove.However,asmentionedalready,inorderforthis

solutionnottobetrivial,i.e.forχ′tobenon-vanishing,wemustthenalsoimposethe

condition(2.24).Thisleadsthetotheblackholesolution(2.10).

Inthenear-horizonregion,thefunctionχintheblack-holesolutions(2.10)hasan

expansionoftheform

χ=?χ0+?χ1(r?r0)12+?χ2(r?r0)32+···.(2.27)

Substitutingbackintotheequationsofmotion,wefindthatallthecoefficientsinthe

expansionscanbeexpressedintermsoftwoparameters,h1andr0.Forexample,

f1=(n?2)(n?3)γ?+2αr

20

(n?2)γr0,χ1=

2radicalbig(n?1)βgr

3

20

(n?1)g2r20+(n?3)?,···.(2.28)

Thusthesolutionhasthreeintegrationconstants(?χ0,h1,r0).However,theparameters

(?χ0,h1)aretrivial.Itfollowsthattheonlynon-trivialparameterisr0,whichisdetermined

byμinthefinalsolution.

Finallywewouldliketoemphasizeagainthatβisnotanintegrationconstant,buta

parameterofthetheory.Forβnegationslash=0,therearetwoblackholes,buteachassociatedwitha

differentvacuum.Whenβ=0,thereisonlytheSchwarzschild-AdSblackholesolutionin

thetheory.

10

3BlackHoleEntropyandThermodynamics

Intheprevioussection,wereviewedtheHorndeskigravitytheory,anditsstaticblackhole

solutions.Weidentifiedthehorizonandcomputedthetemperatureoftheseblackholes.

Inthissection,weconsidervariouspossiblemethodsforcalculatingtheirentropy.Itturns

outthatdifferentwell-establishedmethodsyielddifferentanswers.Acorrectanswerofthe

entropyisimportantforstudyingtheblackholethermodynamics,anditisparamountfor

determiningtheη/Sratio,aswediscussedintheintroduction.

3.1Waldentropyformula

Firstletusconsiderthewell-knownWaldentropyformula(1.5).Itisstraightforwardto

seethatfortheHorndeskiLagrangianLgivenin(2.1),onehas

Tμνρσ≡?L?R

μνρσ

=12κ(gμρgνσ?gνρgμσ)(3.1)

+18γ[gμρχνχσ?gνρχμχσ+gνσχμχρ?gμσχνχρ?(gμρgνσ?gνρgμσ)χλχλ],

wherewehavedefinedχμ=?μχ.ForthestaticblackholesintheHorndeskitheory,

describedinsection2,wefindfrom(3.1)thattheWaldentropyformula(1.5)forthe

entropygivesthesameresultasinstandardEinsteingravity,namelyonequarterofthe

areaoftheeventhorizon,

SW=14κrn?20ωn?2,(3.2)

whereωn?2isthevolumeofaunitSn?2inthe?=1case.For?=0,correspondingto

atoroidalhorizon,theperiodsofthecirclesformingthetoruscanbechosenarbitrarily,

andweshall,forconvenience,thentakeωn?2=1inthispaper,andsocorrespondinglyS

shouldthenbeviewedastheentropydensity.

Sincethestaticblackholesolutionsarecharacterisedbyonlyoneparameter(i.e.one

integrationconstant),itisguaranteedthatonecanobtainanexpressionfora“thermody-

namicmass”byintegratingthefirstlawofblackholethermodynamics2

dM=TdS.(3.3)

2Inamoregeneralsituationwheretherearefurtherintensive/extensivepairsofthermodynamicvariables

contributingontheright-handsideofthefirstlawformulti-parametersolutions,theintegrabilityofthe

right-handsidecanprovideanon-trivialcheckonthecorrectnessofthethermodynamicquantities.Nosuch

consistencycheckarisesinthecaseofaone-parameterfamilyofsolutions,sinceall1-formsareexactinone

dimension.

11

Ifweusetheexpression(3.2)fortheentropy,thenfromtheresultfortheHawkingtem-

peratureobtainedintheprevioussectionwethereforefind

?=0:M=κ(n?2)16πμ,(3.4)

?=1:M=

parenleftBigκ(n?2)

16πg

2rn?1

0+

κ2(n?2)

4π(4κ+βγ)r

n?3

0

parenrightBig

ωn?2.(3.5)

Notethatinthe?=0caseitwasstraightforwardtoexpressthemassintermsofthe“mass

parameter”μ,becauseofthesimplerelationμ=g2rn?10fortheseplanarblackholes.On

theotherhand,therelationbetweenμandr0ismuchmorecomplicatedinthe?=1

case,andisgivenin(2.21).Thuswhen?=1theexpression(3.5)forMwouldbecomea

complicatedtranscendentalfunctionofthemassparameterμ.

Onthefaceofit,themassformula(3.4)forthe?=0caselooksnotunreasonable.In

factthethermodynamicalquantitiessatisfyalsotheexpectedgeneralisedSmarrrelation

M=n?2n?1TSW.(3.6)

However,forthe?=1case,themassformula(3.5)lookslessreasonable.Asmentioned

above,itwouldbeacomplicatedtranscendentalfunctionofthe“massparameter”μ.Whilst

thisfact,ofitself,doesnotconclusivelyshowthatitmustbeincorrect,itdoesperhapsraise

doubtsaboutitslikelyvalidity,sinceitwouldbeaveryunusualkindofrelationthatisnot

normallyseeninotherblackholesolutions.Furthermore,ifthe?=1massformulaiscalled

intoquestionthenthisalsoraisesquestionsaboutthevalidityofthe?=0massformula.

Inordertoexploretheseissuesingreaterdepth,weshallmakeamoredetailedinves-

tigationoftheWaldprocedure,inordertoseewhethertherearenewsubtletiesthatcan

ariseinatheorysuchasthatofHorndeski.

3.2Waldformalism

Waldhasdevelopedaprocedureforderivingthefirstlawofthermodynamicsbycalculating

thevariationofaHamiltonianderivedfromaconservedNoethercurrent.Thegeneral

procedurewaspresentedin[33,34].TheWaldentropyformula(1.5)isaconsequenceof

applyingthisprocedureinrathergenerichigher-derivativetheories.TheWaldformalism

hasbeenusedtostudythefirstlawofthermodynamicsforasymptotically-AdSblackholes

invarietyoftheories,includingEinstein-scalar[39,40],Einstein-Proca[41],Einstein-Yang-

Mills[42],ingravitiesextendedwithquadratic-curvatureinvariants[43],andalsoforLifshitz

blackholes[44].However,theratherunusual-lookingresultsthatitledtoforthemassof

the?=1blackholesinsection3.1raisedthepossibilitythattheformula(1.5)mightnot

12

bevalidforHorndeskigravity.Forthisreason,weshallnowstudyindetailtheapplication

oftheWaldformalismfortheaction(2.1).

Ageneralvariationofthefieldsintheaction(2.1)wasgivenin(2.2).Thesurfaceterm

Jμisgivenby

Jμ=2?L?R

ρσμν

?σδgρν?2?ν?L?R

ρμνσ

δgρσ+?L?(?

μχ)

δχ

=

parenleftBig

κJμg+αJμχ+γ(Jμgc+Jμχc)

parenrightBig

,(3.7)

with

Jμg=gμρgνσ(?σδgνρ??ρδgνσ),Jμχ=?gμν?νχδχ,Jμχc=Gμν?νχδχ,

Jμgc=?14(?χ)2Jμg+14gμρgνσ[?σ(?χ)2δgνρ??ρ(?χ)2δgνσ]

+12gμλ?ρχ?σχ?ρδgσλ?12?ρ(?μχ?σχ)gρλδgσλ

?14gμλ?ρχ?σχ?λδgρσ+14?λ(?ρχ?σχ)gλμδgρσ

?14gρλ?μχ?σχ?σδgρλ+14?σ(?σχ?μχ)gρλδgρλ.(3.8)

FollowingtheWaldprocedure,wecannowdefinea1-formJ(1)=JμdxμanditsHodgedual

Θ(n?1)=(?1)n+1?J(1).(3.9)

Wenowspecialisetoavariationthatisinducedbyaninfinitesimaldiffeomorphism

δxμ=ξμ.Onecanshowthat

J(n?1)≡Θ(n?1)?iξ?L0=?d?J(2),(3.10)

aftermakinguseoftheequationsofmotion.Hereiξdenotesacontractionofξμonthe

firstindexofthen-form?L0.Onecanthusdefinean(n?2)-formQ(n?2)≡?J(2),such

thatJ(n?1)=dQ(n?2).Notethatweusethesubscriptnotation“(p)”todenoteap-form.

Tomakecontactwiththefirstlawofblackholethermodynamics,wetakeξμtobethe

time-likeKillingvectorthatisnullonthehorizon.Waldshowsthatthevariationofthe

Hamiltonianwithrespecttotheintegrationconstantsofaspecificsolutionisgivenby

δH=116πδ

integraldisplay

c

J(n?1)?116π

integraldisplay

c

d(iξΘ(n?1))=116π

integraldisplay

Σ(n?2)

parenleftBig

δQ(n?2)?iξΘ(n?1)

parenrightBig

,(3.11)

wherecdenotesaCauchysurfaceandΣ(n?2)isitsboundary,whichhastwocomponents,

oneatinfinityandoneonthehorizon.ThusaccordingtotheWaldformalism,thefirstlaw

ofblackholethermodynamicsisaconsequenceof

δH∞=δH+.(3.12)

13

FortheHorndeskigravityconsideredinthispaper,wefind

Jα1···αn?1=E.O.M+2?α1···αn?1μ?ν

braceleftBig

κ?[νξμ]?14γ(?χ)2?[νξμ]+12γ?[ν(?χ)2ξμ]

+12γ?σχ?[νχ?σξμ]?12γ?σ(?σχ?[νχ)ξμ]?12γ?[ν(?μ]χ?σχ)ξσ

bracerightBig

,

Qα1···αn?2=?α1···αn?2μν

braceleftBig?L

?Rμνρσ?ρξσ?2ξ[σ?ρ]

parenleftBig?L

?Rμνρσ

parenrightBigbracerightBig

=?i1···in?2μνbraceleftbigκ?μξν?14γ(?χ)2?μξν+12γ?σχ?μχ?σξν

+12γparenleftbig?μ(?χ)2parenrightbigξν?12γ?σ(?σχ?μχ)ξν?12γ?μ(?νχ?σχ)ξσbracerightbig,

(iξΘ)α1···αn?2=?α1···αn?2μλ

parenleftBig

2?L?R

ρσμν

?σδgρν?2?ν?L?R

ρμνσ

δgρσ+?L?(?

μχ)

δχ

parenrightBig

ξλ.(3.13)

Tospecialisetoourstaticblackholeansatz(2.5),theresultfortheLagrangianwithγ=0

iswellestablished(see,forexample,[39,40]),andisgivenby

Qκ,α=rn?2κ

radicalbigg

f

hh

′?(n?2),

iξΘκ,α=?rn?2

radicalBigg

h

f

parenleftBig

κ(?fhδh′+fh



2h2δh?

h′

2hδf?

n?2

rδf)?αfχ

′δχ

parenrightBig

?(n?2),

(δQ?iξΘ)κ,α=?rn?2

radicalBigg

h

f

parenleftBig

κn?2rδf+αfχ′δχ

parenrightBig

?(n?2),(3.14)

Wefindthatthecontributionsassociatedwiththeγtermintheactionaregivenby

Qγ=?12(n?2)γrn?3

radicalBigg

h

ff

2χ′2?(n?2),

iξΘγ=?12(n?2)γrn?3

radicalBigg

h

ff

2parenleftbigχ′2δh

2h+(

n?3

r(1?

?

f)+

h′

h)χ

′δχparenrightbig?(n?2),

(δQ?iξΘ)γ=12(n?2)γrn?3

radicalBigg

h

ff

2

parenleftBig

?32χ′2δff?δ(χ′2)

+(n?3r(1??f)+h



h)χ

′δχ

parenrightBig

?(n?2),(3.15)

WenowapplytheWaldformalismtotheblackholesolutions.First,wenotethatasa

consequenceofequation(2.9),whenweaddthecontributionsin(3.14)and(3.15)theχ′δχ

inthetotalexpressioncancel,givingtheresult

δQ?iξΘ=?(n?2)rn?3

radicalBigg

h

f

bracketleftBig

(κ?34γfχ′2)δf+γf2δ(χ′2)

bracketrightBig

.(3.16)

Infact,ascanbeseenfromtheexpressionforχ′2infortheblackholesolutionsin(2.10),

wehaveδ(fχ′2)=0,andso(3.16)canbefurthersimplified,togive

δQ?iξΘ=?(n?2)rn?3

radicalBigg

h

f

parenleftBig

κ+14γfχ′2

parenrightBig

δf.(3.17)

14

Wefirstconsiderthesimplercaseofthe?=0AdSplanarblackholes,forwhich

fχ′2=β.Wefind

δH∞=(n?2)κ16π

parenleftBig

1+βγ4κ

parenrightBig

δμ,

δH+=(n?1)(n?2)g



16π

parenleftBig

1+βγ4κ

parenrightBig

rn?20δr0.(3.18)

ThusweseeindeedthatδH∞=δH+,sinceμ=g2rn?10.Thisimpliesthatthatwecan

definethemassandentropyas

M=(n?2)κ16π

parenleftBig

1+βγ4κ

parenrightBig

μ,S=14κ

parenleftBig

1+βγ4κ

parenrightBig

rn?20,(3.19)

suchthat

δH∞=δM,δH+=TδS.(3.20)

Thefirstlawofblackholethermodynamics(3.3)thenfollowsstraightforwardlyfromthe

Waldidentity(3.12).Howeverthefactor1+βγ/(4κ)inboththeentropyandthemass

disagreeswiththeresultsin(3.2)and(3.4)thatweobtainedinsection3.1fromadirect

applicationoftheWaldentropyformula(1.5)andtheintegrationofthefirstlawdM=TdS.

Thecaseofthespherically-symmetricblackholeswith(?=1)ismorecomplicated.We

findthatδHevaluatedonthehorizontakesthegeneralform

δH+=(n?2)ωn?2T64(16κ+γf1?χ21)rn?30δr0,(3.21)

wheref1and?χ1arecoefficientsinthenear-horizonexpansionsdefinedin(2.25)and(2.27).

Forourspecific?=1solution,wehave

f1=(n?1)g2r0+n?3r

0

,?χ1=2

radicalbig(n?1)βgr3

20

(n?1)g2r20+n?3,

h1=

parenleftBig

(n?1)(4κ+βγ)g2r20+4(n?3)κ

parenrightBig2

(4κ+βγ)parenleftbig(n?2)g2r20+n?3parenrightbigr0.(3.22)

ThusifwedefineδH+=TdS,withTgivenin(2.22),wefindthattheentropyisgivenby

S=ωn?2

bracketleftBig

1

4κr

n?2

0+

(n?1)(n?2)βγg2rn0

16n(n+2)(n?3)2

parenleftBig

(n+2)(n?3)

?n(n?1)g2r202F1[1,12(n+2);12(n+4);?n?1n?3g2r20]

parenrightBigbracketrightBig

.(3.23)

Notethatthefirstterminsidethesquarebracketsgivespreciselytheresultwesawearlier

(3.2)forWaldentropySW,derivedusingtheformula(1.5).Theremainingcontributionin

thesquarebracketsisproportionaltoγ,thecoefficientoftheHorndeskitermintheaction

(2.1).

15

Toderivethefirstlaw,weevaluatetheδHatasymptoticinfinity,andwefind

δH∞=(n?2)κωn?216π

parenleftBig

1+βγ4κ

parenrightBig

δμ,(3.24)

Thisimpliesthatthemassisgivenby

M=(n?2)κωn?216π

parenleftBig

1+βγ4κ

parenrightBig

μ.(3.25)

Thisturnsouttobetheexactlythesameformasthatinthe?=0AdSplanarblackhole.

Itisnowstraightforwardtoverifythatthefirstlaw(3.3)isindeedsatisfied.Notethatχ0,

beingaconstantshiftintegrationconstantofχ,playsnoroleinthefirstlaw.

Itisworthcommentingthatforthe?=0solutions,themassesweobtainedin(3.4)and

in(3.19)bythetwodifferentmethodsarebothproportionaltoμ.Theonlydifferenceisin

theconstantprefactorcoefficient.Thisonitsownmakesitdifficulttojudgewhichisthe

morereasonableresult.However,when?=1,thedifferencebecomesmorestriking.The

result(3.25)fromthedetailedWaldprocedurethatwepresentedinthispaperisseemingly

moreplausible,fortworeasons.Firstly,themassissimplyproportionaltotheparameterμ,

insteadofbeingaconvolutedtranscendentalfunctionofμ.Secondly,themassdependence

onμisthesameforboththe?=0and?=1solution.Insolutionswithnoadditional

scalarhair,andsincethe?=0solutioncanbeobtainedasascalinglimitofthe?=1

solution,thisconclusionwouldseemtobereasonable.

3.3FurthercommentsontheentropyfromWaldformalism

Havingderivedthefirstlawofthermodynamicsandalsotheentropyinsection3.2,using

thegeneralWaldformalism,wenowexaminethesomewhatunusualfeaturesoftheblack

holesinHorndeskigravitythatleadtothebreakdownofthestandardWaldentropyformula

(1.5).Itfollowsfrom(3.13)thatforthestaticansatz(2.5)that

Q(n?2)=2h′

radicalbigg

f

hS

?0?1?0?1

(n?2)?4hT

0101;1?(n?2),(3.26)

wherethehattedindicesaretangent-spaceindices,thesemicolondenotesacovariantderiva-

tiveand

Tμνρσ≡?L?R

μνρσ

,S?0?1?0?1(n?2)=T?0?1?0?1rn?2?(n?2).(3.27)

Notethat0isthetimedirectionand1istherdirection.TheexpressionforTμνρσfor

theHorndeskigravityisgivenby(3.1).Typically,oneevaluatesQ(n?2)onthehorizonat

r=r0,withh=h1(r?r0)+···andf=f1(r?r0)+···,andsothesecondtermonthe

16

right-handsideof(3.26)vanishesandhence,aswasobservedin[33,34],wefind

1

16π

integraldisplay

r=r0

Q(n?2)=TSW,(3.28)

whereSWisthestandardWaldentropy,givenby(1.5).

Establishingthevariationalidentity(3.12)ismoresubtle,evenforthestandardcaseof

Einsteingravity.ItrequiresthatweevaluateδQonthehorizon.Naively,onewouldsimply

obtainδTSW+TδSWfrom(3.28),andthenonewouldexpectthattheδTSWtermwould

becancelledbytheiξΘcontributionin(3.11),leadingto

δH+=TδSW.(3.29)

However,inordertoevaluatethevariationproperly,weneedtoexpand(3.28)uptoorder

(r?r0),sinceδ(r?r0)=?δr0andsoitisnon-zeroeveninthelimitwhenonesetsr=r0

onthehorizon.TheneteffectisthatallthetermsinδQ(n?2)arecancelledoutbyterms

iniξΘ,andinfacttheTδStermarisesfromtheremainingtermsiniξΘalone.

Tobespecific,letusexamineδQ?iξΘforaspherically-symmetricblackholeinpure

Einsteingravitycoupledtoamasslessscalar,asgivenby(3.14).IfwefirstperformTaylor

expansionsofQandiξΘ,asgiveninthefirsttwoequationsin(3.14),aroundthehorizonat

r=r0,thenindeedtheabovestatementcanbeverified.Thefinalequationin(3.14)gives

analternativebutequivalentevaluationwiththevariationδQ,whichmakestheobservation

moreapparent.WemayevaluateδQfirst,andthensetr=r0.Inthiscase,thern?2factor

inQκ,αjustdependsonthecoordinater,andhenceisnotvaried.Withthisprocedure,

wefindthatallthetermsinδQκ,αarecancelledoutbytermsiniξΘκ,α,leadingtothe

thirdequationof(3.14).Thususingthisprocedure,wefindthattheδH+=TδStermfor

theusualEinsteingravityarisesfromthe(n?2)δf/rterminiξΘin(3.14).Thisterm

correspondsto

rn?2

radicalBigg

h

f

2

rfgijT

1i1jδf?(n?2).(3.30)

ItisratherintriguinghowthistermisultimatelyrelatedtoSWwhichinvolvesonlyT0101.

Indeed,weseefrom(3.1)thatinvielbeincomponents,T?0?1?0?1=?12κandT?1?i?1?j=12κδijfor

theHorndeskiblackholesolutions.Inparticular,theγtermdoesnotcontributeineither

case.

IntheblackholesofHorndeskigravitytherearefurthersubtleties.Firstly,theαterm

iniξΘκ,αin(3.14)doesnotvanishforthesesolutions,andcancontributeatermtothe

entropythatisnotcontainedinSW.Furthermore,althoughthesecondtermin(3.26)

17

vanishesonthehorizon,itsvariationdoesnot.Thisextratermcanbeseenintheformof

Qγin(3.15).Thus(δQ?iξΘ)γin(3.15)willgiveanadditionalcontributiontotheentropy

thatisoverandabovethatofthestandardWaldcontributionSW.Thuswenowhave

δH+=TδS,withSnegationslash=SW.(3.31)

However,theWaldidentity(3.12),aswehaveseen,continuestohold.Thenon-vanishing

contributionsfromboththeαandtheγtermshavethesameessentialorigin,namelythat

thescalarfieldχisnotregularonthehorizon,butrather,ithasabranchcutsingularity,

asshownin(2.27).

Onemightquestionwhetherthisiscompatiblewiththeinterpretationofthesolutionsas

blackholes.However,aswehaveremarkedinsection2.1,thescalarχinHorndeskigravityis

likeanaxion,inthesensethatitentersthetheoryonlythroughitsderivative.Inparticular,

therefore,itwouldnotbenaturaltodefinedifferentconformally-scaledmetricframes(in

themannerthatonedoeswiththedilatoninstringtheory),sincethatwouldbreakthe

manifestaxionicshiftsymmetryofχ.Furthermore,allinvariantpolynomialsconstructed

from?μχwiththemetricandtheRiemanntensorareregularonthehorizon.Forexample,

gμν?μχ?νχisfiniteandnon-zeroonthehorizon.(Thesepropertiescanbeseenfromthe

factthatthevielbeincomponentsofthegradientofχarefiniteeverywhere,includingon

thehorizon,sinceonejusthasEμ?1?μχ=√fχ′=√β[1?(n?3)?/((n?2)g2r2)]?1/2,with

allothercomponentsvanishing,whereEμ?aistheinversevielbein.)Thissupportstheidea

thatthesesolutionsadmitavalidblackholeinterpretation,butatthepricethattheWald

entropyformula(1.5)nolongerprovidesthecompleteexpressionfortheentropy.However,

theidentity(3.12),andhencethefirstlawofblackholethermodynamics,continuetohold,

withtheentropybeingderivedfromthestrictapplicationoftheWaldformalism.

3.4NoetherchargeandmassofAdSplanarblackholes

Intheprevioussubsections,wedescribedtwodifferentmethodsforcalculatingtheentropy

andmassoftheHorndeskiblackholes,onebasedontheuseoftheWaldformula(1.5)for

theentropy,andtheotherbasedonamoredetailedconsiderationoftheWaldformalism.

Inboththeseapproaches,wedidnotuseindependentprocedurestocalculatethemass

andentropy,butrather,wereliedontheuseofthefirstlawofthermodynamicstoobtain

onefromtheother.Sincetheblack-holesolutionsarecharacterisedbyonlyoneparameter,

thereisnonon-trivialintegrabilitycheck,inthesensethattheright-handsideofthefirst

lawdM=TdSwouldbeintegrableregardlessofwhethertheexpressionfortheentropy

18

wascorrectornot.Thefactthatthetwoapproachesledtodifferentresultscallsforan

independentcheckonthecalculationofthemass,ortheentropy.Eventhoughthemass

andentropyobtainedfromtheWaldformalisminsection3.2seemstobemorereasonable,

themassisdeterminedthroughanintegrationofthefirstlaw,ratherthandirectly,inthis

case.Aquestiononecanaskiswhetherthemassisindeedaconservedquantity.

FortheAdSplanarblackholes(i.e.the?=0solutions),thisquestioncanbeanswered

bymeansofasimpleNoethercalculation.For?=0,werewritetheansatzas

ds2=dρ2?a(ρ)2dt2+b(ρ)2d?2?,χ=χ(ρ).(3.32)

Theeffectiveone-dimensionalLagrangianbecomes

L=116πabn?2

parenleftBig

κ(R?2Λ0)?12αχ′2+12γG11χ′2

parenrightBig

,

R=?2a

′′

a?

2(n?2)b′′

b?

2(n?2)a′b′

ab?

(n?2)(n?3)b′2

b2+

?(n?2)(n?3)

b2,

G11=(n?2)a

′b′

ab+

(n?2)(n?3)b′2

2b2?

?(n?2)(n?3)

2b2.(3.33)

whereaprimedenotesaderivativewithrespecttoρ.TheLagrangianisinvariantunder

theglobalscaling

a→λ2?na,b→λb.(3.34)

ThisglobalsymmetryyieldsaconservedNoethercharge

QN=116π(n?2)bn?3(ba′?ab′)(4κ+γχ′2).(3.35)

Intermsofthecoordinatesoftheoriginalansatz(2.5),wehave

QN=n?232πrn?3

radicalbigg

f

h(rh

′?2h)(4κ+γfχ′2).(3.36)

SubstitutingtheAdSplanarblackholesolutionintothisNoetherchargeformula,wefind

QN=(n?1)(n?2)κ8π

parenleftBig

1+βγ4κ

parenrightBig

μ=2(n?1)M.(3.37)

ThusweseethatQNisthesameasthemassobtainedfromtheWaldformalisminsection

3.2,uptosomepurelynumericalconstants.Thissupportstheconclusionthatthemass

andentropyobtainedinsection3.2arevalid,whilsttheresultsinsection3.1arenot.

3.5Euclideanaction

Analternativemethodthathasbeenusedforcalculatingthermodynamicquantitiesfor

blackholesolutionsisbymeansofthequantumstatisticalrelation

Φthermo≡M?TS=IT,(3.38)

19

firstproposedforquantumgravityin[38].HereΦthermodenotesthethermodynamicpoten-

tial,orthefreeenergy,andIistheEuclideanaction.TheregularisedEuclideanactionwas

calculatedforthe?=1Horndeskiblackholeinfourdimensionsin[35].Wehaverepeated

thatcalculation,andobtainedthesameresult(saveforanoverallfactorof2discrepancy).

However,theresultingexpressionsforMandentropyarequitedifferentfromthosein

sections3.1or3.2,andaregivenby

M=12κ

parenleftBig

1+βγ4κ

parenrightBig

μ?3βγg

2r30parenleftbig4κ(3g2r20+1)+3βγg2r20parenrightbig

8(4κ+βγ)(1+3g2r20)parenleftbig4κ(3g2r20?1)+3βγg2r20parenrightbig,

S=κπr20+3πβγg

2r40parenleftbig4κ+3(4κ+βγ)g2r20parenrightbig

2(1+3g2r20)parenleftbig4κ?3(4κ+βγ)g2r20parenrightbig.(3.39)

Notethatwhenβ=0,forwhichtheblackholereducestothestandardSchwarzschild-AdS

one,wegetM=12μandS=κπr20,asonewouldexpect.Itisclearthatthemasssuffers

fromthesameshortcomingastheoneweobtainedfromtheWaldentropyformulain(3.5),

inthatitbecomesaconvolutedtranscendentalfunctionofμfornon-vanishingβ.(Itisa

differenttranscendentalfunctionfromtheonefollowingfrom(3.5),however.)

Thecalculationforthe?=0AdSplanarblackholes(2.16)ismucheasier,andcan

bestraightforwardlycarriedoutforageneralspacetimedimensionn.Theregularised

Euclideanactioncanbedefinedbysubtractingtheactionofthebackgroundμ=0vacuum

fromtheactionfortheblackholeitself,namely

Ireg=IE[gμν,χ]?IE[g(0)μν,χ(0)],(3.40)

whereg(0)μνandχ(0)arethebackgroundfieldobtainedbysettingμ=0intheblackhole

solution(2.16).Wefind

Ireg=?κ16(n?1)

parenleftBig

1?(n?2)βγ4κ

parenrightBig

rn?20.(3.41)

Notethatinthiscalculation,wehavesetωn?2=1,sothattheresultingextensivequantities

aredensities.Usingthequantumstatisticalrelation(3.38)andthethermodynamicfirst

law(3.3),wethenfindthatthefreeenergy,mass,temperatureandentropyforthe?=0

blackholesaregivenby

F=?κμ16π

parenleftBig

1?(n?2)βγ4κ

parenrightBig

,M=?(n?2)F,

T=g

2(n?1)r0

4π,S=

1

4κr

n?2

0?

1

16(n?2)βγr

n?2

0.(3.42)

Theseexpressionsalsodisagree,inthiscasebyconstantoverallfactors,withthe?=0

resultsobtainedinsections3.1and3.2.Takeninisolation,itwouldbehardtomake

20

anyjudgmentastowhethertheseexpressionsweretrustworthyornot.Interestinglythe

generalizedSmarrrelation(3.6)isalsosatisfied.However,the?=1results(3.39)forthe

massandtheentropycertainlyraisequestionsaboutthevalidityofthiscalculationusing

theEuclideanaction.

ThereisanothermethodthathasbeenusedinordertoobtainafiniteEuclideanaction,

byaddingasurfacetermandacounterterm.Takingn=4dimensionsasanexample,the

wholeactionisthengivenby

I=Ibulk?2IGH?Ict,(3.43)

whereIGHisthestandardGibbons-Hawkingsurfaceterm,andfor?=0,thecounterterm

isgivenby

Ict=κ

integraldisplay

dx3√γc1g,withc1=4+βγκ,(3.44)

Theγinthesquarerootisthedeterminantofinducedmetricγμν.Withthesecombinations,

thetotalactionisthesameastheresultofregularization.For?=1,thecountertermis

Ict=κ

integraldisplay

dx3√γ(c1g+c2R[γ]g),withc1=4+βγκ,c2=1?βγ4κ(3.45)

andthevalueoftheactionhasanadditionaltermlineartheimaginary-timeperiod(i.e.

inverselyproportionaltothetemperature),incomparisontothatoftheregularizedcalcu-

lationabove:

Irenorm=Ireg+

√3πβ2γ2

12g(4κ+βγ)

?

T.(3.46)

Theeffectonthethermodynamicsisthattheentropyisunchanged,butthemassacquires

anadditivecontributioninthespherically-symmetric?=1solutions,independentofthe

parameterinthesolutions.Thisisnotsurprising,sincewhen?=1,theμ=0solutionis

notvacuumAdSspacetime,butinsteadasmoothsoliton,whichhasaconstantmass.Inthe

earlierregularisationbysubtractingthebackground,thisconstantenergywassubtracted

out.

Thequestionremainsastohowonemightreconciletheresultsfortheentropyandthe

mass,ascalculatedfromtheregularisedEuclideanaction,withourprevious,anddifferent,

resultsobtainedusingtheWaldformalism.Wedonothaveadefinitiveresolutiontothis

puzzle,otherthantosuggestthatbecauseoftheratherunusualfeaturesoftheblack-hole

solutionsinHorndeskigravity,itmaybethatthenaiveapplicationofasubtractionproce-

duretoobtainaregularisedEuclideanactionmaybeinherentlyambiguous.Inasomewhat

relatedcontext,itwasfoundin[45]thatattemptstoemploytheAbbott-Desermethod[46]

tocalculatethemassofasymptotically-AdSblackholesfounderedonambiguitiesinthesub-

tractionprocedureinsomecases,forsolutionsingaugedsupergravitieswherescalarfields

21

wereinvolved.Intheabsenceofarigorousderivationofavalidsubtractionschemeforthe

calculationoftheEuclideanaction,itseemsthatonecouldengineerdifferentschemesthat

gavedifferentresults,withnoguideastowhichresultshouldberegardedasthecorrect

one.

4Viscosity/EntropyRatio

Oneofthemotivationsforthispaperwastostudytheviscosity/entropyratioinHorndeski

gravity.Havingobtainedaformulafortheentropyoftheblackholes,wearenowin

apositiontoproceed.Tocalculatetheshearviscosityoftheboundaryfieldtheory,we

consideratransverseandtracelessperturbationoftheAdSplanarblackhole,namely

ds2=?fdt2+dr

2

f+r

2parenleftbigdxidxi+2Ψ(r,t)dx1dx2parenrightbig,(4.1)

wherethebackgroundsolutionisgivenby(2.11),(2.16)and(2.17).Wefindthatthemode

Ψ(r,t)satisfiesthelinearisedequation

r(4κ+βγ)(g2rn?1?μ)2Ψ′′+(4κ+βγ)(g2rn?1?μ)(ng2rn?1?μ)Ψ′

?r2n?5(4κ?βγ)¨Ψ=0.(4.2)

Foraninfallingwavewhichispurelyingoingatthehorizon,thesolutionforawavewith

lowfrequencyωisgivenby

Ψ=e?iωtψ(r),ψ(r)=exp(?iωKlogf(r)g2r2)+O(ω2),

K=14πT

radicalBigg

4κ?βγ

4κ+βγ.(4.3)

NotethattheconstantparameterKisdeterminedbythehorizonboundarycondition.The

overallintegrationconstantisfixedsothatΨisunimodularasymptotically,asr→∞.

InordertostudytheboundaryfieldtheoryusingtheAdS/CFTcorrespondence,we

substitutetheansatzwiththelinearisedperturbationintotheaction.Thequadraticterms

intheLagrangian,afterremovingthesecond-derivativecontributionsusingtheGibbons-

Hawkingterm,canbewrittenas

L2=P1Ψ′2+P2ΨΨ′+P3Ψ2+P4˙Ψ2,(4.4)

with

P1=?18(4κ+βγ)(g2rn?1?μ)r,P2=12g2rn?1[4κ?(n?2)βγ]?μ(2κ?n?34βγ),

22

P3=n?14g2rn?2[4κ?(n?2)βγ],P4=r

2n?5(4κ?βγ)

8(g2rn?1?μ)(4.5)

NotethatP3=12P′2.WethenfindthatthetermsquadraticinΨintheLagrangianare

givenby

L2=ddr(P1ΨΨ′+12P2Ψ2)+ddt(P4Ψ˙Ψ)?Ψ

bracketleftBig

P1Ψ′′+P′1Ψ′+P4¨Ψ

bracketrightBig

.(4.6)

Thelastterm,enclosedinsquarebrackets,vanishesbyvirtueofthelinearisedperturbation

equation(4.2),andsothequadraticLagrangianisatotalderivative.Theviscosityis

determinedfromtheP1ΨΨ′term,followingtheproceduredescribedin[6,20].Usingthis,

wefindthattheviscosityisgivenby

η=κ(n?1)μ64π2T

radicalbigg

1?β

2γ2

16κ2.(4.7)

Wehave,fortheplanarblackholes,

μ=g2rn?10,T=(n?1)g

2r0

4π,(4.8)

andtheentropythatwederivedinsection3.2usingtheWaldformalismisgivenby

S=14κ

parenleftBig

1+βγ4κ

parenrightBig

rn?20.(4.9)

Wethereforefindthattheviscosity/entropyratioisgivenby

η

S=

1



radicalBigg

4κ?βγ

4κ+βγ(4.10)

fortheHorndeskiblackholes.3Notethatκandβarebothpositive.Forreality,wemust

have

?4κβ<γ<4κβ.(4.11)

Whenβ=0,whichturnsoffthescalarfield,theratiogoesbacktotheuniversalvalueof

1/(4π).Whenγ>0,theratioislessthan1/(4π)andhencetheboundisviolated.For

γ<0,theratioisgreaterthan1/(4π).

Finally,wenotethatintermsoftheoriginalparametersofthetheory(2.1),theviscos-

ity/entropyratioisgivenby

η

S=

1



radicalBigg

3α+γΛ0

α?γΛ0.(4.12)

Interestingly,theratioisindependentoftheparameterκ.

3Intriguingly,althoughtheratioiscalculatedfortheAdSplanarblackhole(?=0),thesameratio

(4κ?βγ)/(4κ+βγ)appearsinthesub-leadingconstantterminthelarge-rexpansionofh=?gttgivenin

(2.19),butonlyforthespherically-symmetric(?=1)solutions(itvanishesforthe?=0solutions).

23

5Conclusion

MotivatedbyapplicationsfortheAdS/CFTcorrespondence,westudiedtheblackholesin

atheoryofEinsteingravitycoupledtoascalarfield,includinganon-minimalHorndeski

termwherethegradientofthescalarcouplestotheEinsteintensor.Therearetwotypes

ofstaticblackholesinthisHorndeskigravity.OneoftheseistheusualSchwarzschild-AdS

blackhole,forwhichthescalarfieldisconstant.Ourfocusisontheothernon-trivial

one-parameterfamilyofstaticblackholes,forwhichthescalardependsnon-triviallyon

theradialcoordinate.Althoughthescalarhasabranch-cutsingularityonthehorizon,it

isaxion-likeandentersthetheoryonlythroughaderivative.Furthermore,inanorthonor-

malframe,?aχisregulareverywhere,bothonandoutsidethehorizon,andallinvariants

involvingthescalarfieldarefiniteeverywhere.Wealsodemonstratedtheuniquenessof

thesestaticblackholesolutionsinthetheory.

Westudiedthethermodynamicsoftheblackholesandfoundthreesurprises.Thefirst

isthatthestandardWaldentropyformula(1.5)doesnotgivethecompleteexpressionfor

theentropyoftheseblackholes.Thiscanbeattributedtothefactthatthederivationofthe

Waldentropy(1.5)requiresthatthescalarberegularonthehorizon.Infact,thebranch

cutsingularityofthescalaronthehorizonimpliesthatthereisanextracontributionto

theentropy.WestudiedtheWaldformalismindetail,andexhibitedthenewcontribution

explicitly.ItturnsoutthattheWaldidentity(3.12)continuestoholdfortheseblackholes,

andsodoesthefirstlawofblackholethermodynamics.Theentropy,however,isnolonger

givenby(1.5),butcanbedeterminedfromtheimplementationoftheWaldprocedure.We

furtherestablished,usingasimpleconstructionoftheNoetherchargederivablefromthe

scalingsymmetryoftheplanarblackholes,thatthemassoftheAdSplanarblackhole,as

wederivedfromtheWaldprocedure,isindeedaconservedquantity.

ThesecondsurpriseconcernstheuseofthequantumstatisticalrelationE?TS=TIto

calculatethethermodynamicparametersoftheblackholesolutions.Inordertoapplythis

method,itisnecessarytocalculatetheEuclideanactionIoftheblackholesolution.The

problemisthatadirectintegrationoftheEuclideanisedactionyieldsaresultthatdiverges

attheupperendoftheradialintegration,andsoitisnecessarytoadoptsomeregularisation

procedure.Wetriedtoapplytwodifferentsuchprocedures.Thefirstinvolvedsubtracting

thedivergingcontributionofabackgroundwherethemassissettozerofromthediverging

contributionfromtheblackholewithnon-zeromass.Theotherprocedureinvolvedaddinga

boundarycounterterm.Thetwomethodsgavethesameresultsforthemassandtheentropy,

buttheseresultsdifferedfromthosethatweobtainedbyusingtheWaldformalism.The

24

originofthismismatchisnotcleartous;itmayberelatedtointrinsicambiguitiesinthe

subtractionschemesthatweusedinordertoregularisethedivergences.Suchambiguities

arepossiblymorelikelyinatheorysuchasHorndeskigravity,withitssomewhatunusual

features,andsoregularisationschemesforcalculatingtheEuclideanactionthatusually

workinlessexactingsituationsmayneedtobescrutinisedmorecarefullyhere.

Thethirdsurpriseconcernstheresultsinsection4fortheviscosity/entropyratio.In

wideclassesofconventionaltheorieswithnohigher-derivativetermsintheLagrangian,one

findsaratheruniversalresultthatη/S=1/(4π).Counter-examplestotheuniversalityof

theratiohavebeenfound,butforisotropicsituationssuchaswehaveconsideredtheyare

alwaysassociatedwithhigher-derivativegravities,suchasGauss-Bonnetormoregeneral

Lovelockgravities.Asfarasweareaware,ourfindingsfortheblackholesintheHorndeski

theorywestudiedinthispaperprovidethefirstexampleoftheviolationoftheη/S=1/(4π)

resultinatheorywhoseLagrangianisatmostlinearincurvaturetensor.

AwordofcautionabouttheuseoftheWaldformalismtocalculatetheentropyis

perhapsappropriatehere.IfweconsiderEinstein-Maxwelltheoryasanexample,thefirst

lawdM=TdS+ΦdQforReissner-Nordstr¨omblackholescanbederivedfromtheWald

formalismbycalculatingδH∞andδH+,andusingthefactthatδH∞=δH+.TheΦdQ

contributioncaneitherenterinδH+alone,ifoneusesthegaugewherethepotentialvanishes

atinfinity,orinδH∞alone,ifoneusesthegaugewherethepotentialvanishesonthe

horizon,orelseinbothδH∞andδH+,ifoneusessomeintermediategaugewherethe

potentialvanishesneitheratinfinitynoronthehorizon.Inthefirstlaw,onlythepotential

differenceΦ≡Φ+?Φ∞contributes.Ifthegaugewherethepotentialvanishesonthe

horizonischosen,thenδH+=TδSandsoδH+/Tisanexactdifferential,whichcanbe

integratedtogivetheentropy,whileδH∞=dM+Φ∞dQ,andisnotexact.Inthegauge

wherethepotentialinsteadvanishesatinfinity,δH∞=dM,whichisanexactdifferential,

whileδH+=TdS+Φ+dQ,andsoδH+/Tisnotexact.

Morecomplicatedsituationswereencounteredrecentlywhereasymptotically-AdSdy-

onicallychargedblackholeswereconstructedinafour-dimensionalgaugedsupergravity

involvingascalarandaMaxwellfield[47,48].ItwasfoundthatδH∞wasnon-exact,

andhencenon-integrable,evenwhenagaugewheretheelectricandmagneticpotentials

vanishedatinfinitywaschosen,becauseofavaryingcontributionfromtheasymptotic

coefficientsinthelarge-distanceexpansionofthescalarfield.Thefirstlawofblackhole

(thermo)dynamics,involvingthescalarcontribution,couldneverthelessbederivedusing

thestrictWaldformalism[47].Theresultswerelatergeneralisedtoblackholesingeneral

25

Einstein-scalartheories[39,40],Einstein-Procatheories[41],andgravityextendedwith

quadraticcurvatureinvariants[43].

AnalogousissuescouldinprinciplearisewhenconsideringδH+:itiscommonlythecase

thatδH+onthehorizoncanbeexpressedasTδS.InatheorysuchasEinstein-Maxwell,

thisisagauge-dependentpropertyaswediscussedabove,andinordertohaveδH+/Tbe

anexactdifferentialinthiscaseonewouldneedtoworkinthegaugewheretheelectric

potentialvanishedonthehorizon.Inmosttheoriesthathavebeenstudied,theentropyis

simplygivenbySWdefinedbytheWaldentropyformula(1.5).Thewidespreadvalidityof

theWaldentropyformulaisrelatedtothefactthattypically,matterfieldsvanishonthe

horizonofablackhole(andMaxwellpotentialscanbesettozerobymeansofappropriate

gaugechoices).IntheHorndeskigravityconsideredinthispaper,however,theaxion-like

scalarχhasanunusualbehaviournearthehorizonandnearinfinity,andindeedwehave

alreadyseenthatδH+negationslash=TδSW.Weneverthelessassumedthatitwasstillthecasethat

δH+=TδS,i.e.thatδH+/Tcouldbeintegratedtodefineanentropyfunction.That

δH+/Tisintegrableisguaranteedintheone-parameterfamilyofsolutionsconsideredin

thispaper,sinceall1-formsinonedimensionareexact.Inamultiple-parameterblackhole

solution,however,theredoesnotappeartobeanyguarantee,apriori,thatδH+/Tmustbe

atotaldifferentialinatheorysuchasHorndeskigravity.Thenon-integrabilityofthesort

thatoccursinδH∞inthedyonicasymptotically-AdSblackholeswediscussedabovemight

also,inprinciple,occurforδH+/Tonthehorizon,ifnotallthefieldsarestrictlyvanishing

onthehorizon.Itwouldbeinterestingtostudythisfurtherinmoregeneralsolutionsin

theoriessuchasHorndeskigravities.

ThefindingsinthispaperindicatethatHorndeskigravity,anditsblackholesolutions

inparticular,deservefurtherinvestigationbothintheirownright,andalsointhecontext

oftheAdS/CFTcorrespondence.

Acknowledgements

WearegratefultoSeraCremoniniforhelpfuldiscussions.H-S.L.issupportedinpart

byNSFCgrants11305140,11375153and11475148,SFZJEDgrantY201329687andCSC

scholarshipNo.201408330017.C.N.P.issupportedinpartbyDOEgrantDE-FG02-

13ER42020.TheworkofX-H.FengandH.L.aresupportedinpartbyNSFCgrantsNO.

11175269,NO.11475024andNO.11235003.

26

References

[1]J.M.Maldacena,ThelargeNlimitofsuperconformalfieldtheoriesandsupergravity,

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

[2]S.S.Gubser,I.R.KlebanovandA.M.Polyakov,Gaugetheorycorrelatorsfromnon-

criticalstringtheory,Phys.Lett.B428,105(1998),hep-th/9802109.

[3]E.Witten,Anti-deSitterspaceandholography,Adv.Theor.Math.Phys.2,253(1998),

hep-th/9802150.

[4]O.Aharony,S.S.Gubser,J.M.Maldacena,H.OoguriandY.Oz,LargeNfieldtheories,

stringtheoryandgravity,Phys.Rept.323,183(2000)[hep-th/9905111].

[5]G.Policastro,D.T.SonandA.O.Starinets,Theshearviscosityofstronglycoupled

N=4supersymmetricYang-Millsplasma,Phys.Rev.Lett.87,081601(2001),

hep-th/0104066.

[6]D.T.SonandA.O.Starinets,MinkowskispacecorrelatorsinAdS/CFTcorrespondence:

Recipeandapplications,JHEP0209,042(2002),hep-th/0205051.

[7]P.Kovtun,D.T.SonandA.O.Starinets,Holographyandhydrodynamics:Diffusionon

stretchedhorizons,JHEP0310,064(2003),hep-th/0309213.

[8]P.Kovtun,D.T.SonandA.O.Starinets,Viscosityinstronglyinteractingquantumfield

theoriesfromblackholephysics,Phys.Rev.Lett.94,111601(2005),hep-th/0405231.

[9]N.IqbalandH.Liu,UniversalityofthehydrodynamiclimitinAdS/CFTandthe

membraneparadigm,Phys.Rev.D79,025023(2009),arXiv:0809.3808[hep-th].

[10]R.G.Cai,Z.Y.NieandY.W.Sun,Shearviscosityfromeffectivecouplingsofgravitons,

Phys.Rev.D78,126007(2008),arXiv:0811.1665[hep-th].

[11]R.G.Cai,Z.Y.Nie,N.OhtaandY.W.Sun,ShearviscosityfromGauss-Bonnetgravity

withadilatoncoupling,Phys.Rev.D79,066004(2009),arXiv:0901.1421[hep-th].

[12]R.Brustein,D.GorbonosandM.Hadad,Wald’sentropyisequaltoaquarterofthe

horizonareainunitsoftheeffectivegravitationalcoupling,Phys.Rev.D79,044025

(2009),arXiv:0712.3206[hep-th].

27

[13]H.S.Liu,H.L¨uandC.N.Pope,GeneralisedSmarrformulaandtheviscositybound

forEinstein-Maxwell-Dilatonblackholes,Phys.Rev.D92,no.6,064014(2015),

arXiv:1507.02294[hep-th].

[14]A.BuchelandJ.T.Liu,Universalityoftheshearviscosityinsupergravity,Phys.Rev.

Lett.93,090602(2004),hep-th/0311175.

[15]A.Buchel,Onuniversalityofstress-energytensorcorrelationfunctionsinsupergravity,

Phys.Lett.B609,392(2005),hep-th/0408095.

[16]P.Benincasa,A.BuchelandR.Naryshkin,Theshearviscosityofgaugetheoryplasma

withchemicalpotentials,Phys.Lett.B645,309(2007),hep-th/0610145.

[17]K.LandsteinerandJ.Mas,Theshearviscosityofthenon-commutativeplasma,JHEP

0707,088(2007),arXiv:0706.0411[hep-th].

[18]S.Cremonini,Theshearviscositytoentropyratio:Astatusreport,Mod.Phys.Lett.

B25,1867(2011),arXiv:1108.0677[hep-th].

[19]Y.KatsandP.Petrov,EffectofcurvaturesquaredcorrectionsinAdSontheviscosity

ofthedualgaugetheory,JHEP0901,044(2009),arXiv:0712.0743[hep-th].

[20]M.Brigante,H.Liu,R.C.Myers,S.ShenkerandS.Yaida,Viscosityboundviolation

inhigherderivativegravity,Phys.Rev.D77,126006(2008),arXiv:0712.0805[hep-th].

[21]M.NatsuumeandM.Ohta,Theshearviscosityofholographicsuperfluids,Prog.Theor.

Phys.124,931(2010),arXiv:1008.4142[hep-th].

[22]J.Erdmenger,P.KernerandH.Zeller,Non-universalshearviscosityfromEinstein

gravity,Phys.Lett.B699,301(2011),arXiv:1011.5912[hep-th].

[23]O.OvdatandA.Yarom,Amodulatedsheartoentropyratio,JHEP1411,019(2014),

arXiv:1407.6372[hep-th].

[24]X.H.Ge,Y.Ling,C.NiuandS.J.Sin,Holographictransportsandstabilityin

anisotropiclinearaxionmodel,arXiv:1412.8346[hep-th].

[25]F.W.Shu,ThequantumviscosityboundinLovelockgravity,Phys.Lett.B685,325

(2010),arXiv:0910.0607[hep-th].

[26]J.deBoer,M.KulaxiziandA.Parnachev,AdS7/CFT6,Gauss-Bonnetgravity,and

viscositybound,JHEP1003,087(2010),arXiv:0910.5347[hep-th].

28

[27]X.O.CamanhoandJ.D.Edelstein,CausalityconstraintsinAdS/CFTfromconformal

colliderphysicsandGauss-Bonnetgravity,JHEP1004,007(2010),arXiv:0911.3160

[hep-th].

[28]C.BransandR.H.Dicke,Mach’sprincipleandarelativistictheoryofgravitation,

Phys.Rev.124,925(1961).

[29]G.W.Horndeski,Second-orderscalar-tensorfieldequationsinafour-dimensional

space,Int.J.Theor.Phys.10,363(1974).

[30]A.Nicolis,R.RattazziandE.Trincherini,TheGalileonasalocalmodificationof

gravity,Phys.Rev.D79,064036(2009),arXiv:0811.2197[hep-th].

[31]S.W.Hawking,Blackholeexplosions,Nature248,30(1974).

[32]S.W.Hawking,Particlecreationbyblackholes,Commun.Math.Phys.43,199(1975)

[Erratum-ibid.46,206(1976)].

[33]R.M.Wald,BlackholeentropyistheNoethercharge,Phys.Rev.D48,3427(1993),

gr-qc/9307038.

[34]V.IyerandR.M.Wald,SomepropertiesofNoetherchargeandaproposalfordynamical

blackholeentropy,Phys.Rev.D50,846(1994),gr-qc/9403028.

[35]A.Anabalon,A.CisternaandJ.Oliva,AsymptoticallylocallyAdSandflatblackholes

inHorndeskitheory,Phys.Rev.D89,084050(2014),arXiv:1312.3597[gr-qc].

[36]M.Rinaldi,Blackholeswithnon-minimalderivativecoupling,Phys.Rev.D86,084048

(2012),arXiv:1208.0103[gr-qc].

[37]E.BabichevandC.Charmousis,Dressingablackholewithatime-dependentGalileon,

JHEP1408,106(2014),arXiv:1312.3204[gr-qc].

[38]G.W.GibbonsandS.W.Hawking,Actionintegralsandpartitionfunctionsinquantum

gravity,Phys.Rev.D15,2752(1977).

[39]H.S.LiuandH.L¨u,ScalarchargesinasymptoticAdSgeometries,Phys.Lett.B730,

267(2014),arXiv:1401.0010[hep-th].

[40]H.L¨u,C.N.PopeandQ.Wen,ThermodynamicsofAdSblackholesinEinstein-scalar

gravity,JHEP1503,165(2015),arXiv:1408.1514[hep-th].

29

[41]H.S.Liu,H.L¨uandC.N.Pope,ThermodynamicsofEinstein-ProcaAdSblackholes,

JHEP1406,109(2014),arXiv:1402.5153[hep-th].

[42]Z.Y.FanandH.L¨u,SU(2)-Colored(A)dSblackholesinconformalgravity,JHEP

1502,013(2015),arXiv:1411.5372[hep-th].

[43]Z.Y.FanandH.L¨u,Thermodynamicalfirstlawsofblackholesinquadratically-extended

gravities,Phys.Rev.D91,no.6,064009(2015),arXiv:1501.00006[hep-th].

[44]H.S.LiuandH.L¨u,ThermodynamicsofLifshitzblackholes,JHEP1412,071(2014),

arXiv:1410.6181[hep-th].

[45]W.Chen,H.L¨uandC.N.Pope,Massofrotatingblackholesingaugedsupergravities,

Phys.Rev.D73,104036(2006),hep-th/0510081.

[46]L.F.AbbottandS.Deser,Stabilityofgravitywithacosmologicalconstant,Nucl.Phys.

B195,76(1982).

[47]H.L¨u,Y.PangandC.N.Pope,AdSdyonicblackholeanditsthermodynamics,JHEP

1311,033(2013)[arXiv:1307.6243[hep-th]].

[48]D.D.K.ChowandG.Compere,DyonicAdSblackholesinmaximalgaugedsupergravity,

Phys.Rev.D89,no.6,065003(2014)[arXiv:1311.1204[hep-th]].

30

献花(0)
+1
(本文系lapshing23首藏)