return

Celebratio Mathematica

Yvonne Choquet-Bruhat

Maximal hypersurfaces in
asymptotically anti-de Sitter spacetimes

by Piotr T. Chruściel and Gregory J. Galloway

Introduction

It is a great pleas­ure to ded­ic­ate this work to Yvonne Cho­quet-Bruhat, who wrote pi­on­eer­ing pa­pers on this sub­ject [1], [2], and on many oth­ers.

In­deed, Yvonne has writ­ten sev­er­al pa­pers con­cern­ing ex­ist­ence and unique­ness prop­er­ties of max­im­al (and CMC) hy­per­sur­faces in space­time man­i­folds. In [1] she states, “The ex­ist­ence of a max­im­al sub­man­i­fold (with re­spect to area) is an im­port­ant prop­erty for a space time, hy­per­bol­ic rieman­ni­an man­i­fold sat­is­fy­ing Ein­stein equa­tions.” She goes on to de­scribe the im­port­ance of this for solv­ing the Ein­stein con­straint equa­tions, and also the rel­ev­ance of this (at the time) for prov­ing pos­it­iv­ity of mass.

Mean­while the ex­ist­ence of max­im­al hy­per­sur­faces has been found to be use­ful for oth­er pur­poses as well. Of par­tic­u­lar rel­ev­ance here is a clas­sic­al res­ult of Bart­nik [e6] that es­tab­lishes, un­der cer­tain con­di­tions, the ex­ist­ence of max­im­al hy­per­sur­faces in asymp­tot­ic­ally flat space­times. This and re­lated res­ults have been use­ful for a num­ber of pur­poses in the study of such space­times. Giv­en the role of asymp­tot­ic­ally anti-de Sit­ter space­times in the­or­et­ic­al phys­ics (e.g., via the AdS/CFT cor­res­pond­ence), one is nat­ur­ally led to con­sider max­im­al hy­per­sur­faces in this set­ting, where now the no­tion of renor­m­al­ised volume also be­comes rel­ev­ant.

The ques­tion then arises of ex­ist­ence of max­im­al hy­per­sur­faces in \( (n{+}1) \)-di­men­sion­al, \( n\ge 2 \), asymp­tot­ic­ally loc­ally hy­per­bol­ic space­times, with con­trolled asymp­tot­ic be­ha­viour at the con­form­al bound­ary at in­fin­ity so that the renor­m­al­ised volume is defined. The simplest case where an af­firm­at­ive an­swer can be giv­en is in a per­turb­at­ive set­ting, when the met­ric, or the asymp­tot­ic data, or both, are per­turbed away from a max­im­al sli­cing, and when an en­ergy con­di­tion is sat­is­fied. We prove this in The­or­em 9.3 be­low, gen­er­al­ising a re­lated res­ult of [e14].

However, simple ex­amples show that some glob­al reg­u­lar­ity con­di­tions need to be sat­is­fied by the space­time for ex­ist­ence in gen­er­al. Here two con­di­tions arise nat­ur­ally: that of ex­ist­ence of bar­ri­ers, and that of com­pact­ness of the do­main of de­pend­ence of a fi­du­cial Cauchy sur­face. Un­der these two con­di­tions we show solv­ab­il­ity of an asymp­tot­ic Di­rich­let prob­lem for the max­im­al hy­per­sur­face equa­tion; see The­or­ems 9.1 and 9.5. This gen­er­al­ises pre­vi­ous res­ults of [e7], [e15], es­tab­lished un­der re­strict­ive con­di­tions on the di­men­sion or on the class of space­times con­sidered.

As such, these last two the­or­ems do not guar­an­tee that a well-defined no­tion of renor­m­al­ised volume ex­ists without fur­ther con­di­tions; see for in­stance The­or­em 9.2 where both “good bar­ri­ers” and time­like con­ver­gence is as­sumed. This is re­lated to the ques­tion of the be­ha­viour of the max­im­al hy­per­sur­face at the con­form­al bound­ary at in­fin­ity, which turns out to be del­ic­ate. In­deed, while max­im­al hy­per­sur­faces are typ­ic­ally space­like and smooth in the in­teri­or, they might be­come asymp­tot­ic­ally null when the con­form­al bound­ary is ap­proached. We show that max­im­al hy­per­sur­faces which are uni­formly space­like at in­fin­ity, so that the last pos­sib­il­ity does not oc­cur, are uniquely defined by their asymp­tot­ic Di­rich­let data, meet the con­form­al bound­ary or­tho­gon­ally, and have a full poly­ho­mo­gen­eous ex­pan­sion at in­fin­ity, in par­tic­u­lar a well-defined renor­m­al­ised volume; see The­or­ems 6.4 and 8.1 for pre­cise state­ments.

Our ana­lys­is shows that a pos­sible ob­struc­tion to reg­u­lar­ity of max­im­al hy­per­sur­faces at the con­form­al bound­ary at in­fin­ity is that the hy­per­sur­face be­comes asymp­tot­ic­ally null. The key re­main­ing ques­tion is to show that asymp­tot­ic­ally null max­im­al hy­per­sur­faces do not ex­ist, or to con­struct a counter­example.

For the full art­icle, click on the PDF ver­sion at the up­per right of this page.

Works

[1] Y. Cho­quet-Bruhat: “Max­im­al sub­man­i­folds and sub­man­i­folds with con­stant mean ex­trins­ic curvature of a Lorent­zi­an man­i­fold,” Ann. Scuola Norm. Sup. Pisa Cl. Sci. (4) 3 : 3 (1976), pp. 361–​376. MR 423405 Zbl 0332.​53035 article

[2] Y. Cho­quet-Bruhat: “Prop­er­ties of max­im­al sub­man­i­folds in space-times with com­pact or open space-like sec­tions,” Rep. Math­em­at­ic­al Phys. 12 : 1 (1977), pp. 9–​17. MR 461389 Zbl 0363.​53029 article

[3] Y. Cho­quet-Bruhat: Gen­er­al re­lativ­ity and the Ein­stein equa­tions. Ox­ford Math­em­at­ic­al Mono­graphs. Ox­ford Uni­versity Press (Ox­ford, UK), 2009. MR 2473363 Zbl 1157.​83002 book