Connectedness of the boundary in the AdS/CFT correspondence (Q5938764)

Let \(M\) be an \((n+1)\)-dimensional manifold with non-empty boundary \(N\). A metric \(g\) on \(M\) is conformally compact if it can be written as \(h/r^2\), where \(h\) is a compact metric on \(M\), and \(r\) is a function on \(M\) with a first-order zero on the boundary. \((M,g)\) is then complete, and its boundary has a well-defined conformal structure. The AdS-CFT correspondence is an important conjecture which states that, maybe after taking the tensor product of \(M\) with a compact manifold, the conformal field theory (CFT) on \(N\) should be related to the string theory in \(M\). More precisely, if \(c\) is a conformal class on \(N\), the partition function of the CFT on \((N,c)\) should be a sum of terms corresponding to the string theory on the \((M_i,g_i)\), where \(\partial M_i=N\) and the \(g_i\) are Einstein conformally compact metrics on \(M_i\) inducing \(c\) on \(N\). The CFT on \(N\) makes sense mostly when \((N,c)\) has positive scalar curvature, in the sense that there is a metric on \(N\) conformal to \(c\) which has positive scalar curvature. The main result of the paper is that if, in addition, \(\partial M=N\) and \(g\) is an Einstein, conformally compact metric on \(M\) inducing \(c\) on \(N\), then \(H_n(M,Z)=0\). The authors point out that this implies that \(N\) is connected, and also that \(M\) contains no ``wormhole''; it therefore removes some puzzling questions concerning the AdS-CFT correspondence. To prove the result, the authors introduce the ``brane action'' \(L\) on embedded hypersurfaces of \(M\), defined as the area minus \(n\) times a primitive of the volume form. Then they prove using a local computation that \(L\) has no minimum, while the behavior of the function at infinity shows that there is one in each non-zero class of \(H_n(M, Z)\). Since this paper appeared, several alternate proofs of its main result have been provided, in particular by \textit{M. T. Anderson} [Adv. Math. 179, No.2, 205-249 (2003; Zbl 1048.53032); see also math.DG/0104171], by \textit{M. Cai} and \textit{G. J. Galloway} [Adv. Theor. Math. Phys. 3, No. 6, 1769-1783 (2000; Zbl 0978.53084)] and by X. Wang (preprint).