On energy functionals, Kähler-Einstein metrics, and the Moser-Trudinger-Onofri neighborhood (Q999862): Difference between revisions

Although existence of a Kähler-Einstein metric on a Fano manifold is usually proven by the continuity method, the assumptions under which this method works are most elegantly and conveniently expressed in terms of certain integral functionals on the set of potentials. The functionals \(I\) and \(J\) were introduced by \textit{Th. Aubin} [J. Funct. Anal. 57, 143--153 (1984; Zbl 0538.53063)], \(F\) by \textit{W. Ding} [Math. Ann. 282, No.~3, 463--471 (1988; Zbl 0661.53045)], \(E_0\) by \textit{T. Mabuchi} [Tôhoku Math. J., II. Ser. 38, 575--593 (1986; Zbl 0619.53040)] and \(E_n\) by \textit{S. Bando} and \textit{T. Mabuchi} [Algebraic geometry, Proc. Symp., Sendai/Jap. 1985, Adv. Stud. Pure Math. 10, 11--40 (1987; Zbl 0641.53065)]. The main impetus for this approach came from a paper of \textit{G. Tian} [Invent. Math. 130, No.~1, 1--37 (1997; Zbl 0892.53027)], where existence of a Kähler-Einstein metric was shown to be in fact equivalent to the ``properness'' of \(F\). (For simplicity, in the review, we restrict to the case of finite automorphism group.) More recently, other functionals \(E_1, \ldots E_{n-1}\) have been introduced by \textit{X. X. Chen} and \textit{G. Tian} [Invent. Math. 147, No.~3, 487--544 (2002; Zbl 1047.53043)]. This very well-written paper is a systematic study of all these functionals together. The main results are analogues of Tian's theorem quoted above for the other energy functionals \(E_k\). It is shown that for any \(k=0,1, \dots, n\) the properness of \(E_k\) is equivalent to that of \(F\), hence to the existence of a Kähler-Einstein metric. (In the case \(k=1\) this has been proven recently by \textit{Jian Song} and \textit{B. Weinkove} [Duke Math. J. 137, No.~1, 159--184 (2007; Zbl 1116.32018)], but the proof in the paper under review is simpler.) For the functionals \(E_k\) with \(k>1\), properness must be checked only on the potentials with positive Ricci curvature. In fact Cor. 5.5 asserts that these functionals are not even bounded below on general potentials, even if there is a Kähler-Einstein metric. The case where the functionals are not proper, but only bounded from below is considered in Thm. 1.4. Results for manifolds with nonzero holomorphic vector fields are given (see \S 4). In the last part of the paper, the author considers functionals with base a Kähler-Einstein metric. The properness of \(F\) yields an inequality for all Kähler potentials. This inequality in fact holds on a strictly larger set of functions, that is called the Moser-Trudinger-Onofri neighbourhood. The functions in this set admit a nice characterization in terms of the Ricci curvature operator and the functional \(E_n\) (see Thm. 5.4). In particular, the author gets a new purely Kählerian proof of the Moser-Trudinger-Onofri inequality on \(S^2\), which does not use symmetrization.