Connected sum construction for \(\sigma_k\)-Yamabe metrics (Q1948733): Difference between revisions

Given a compact Riemannian manifold \((M,g)\) of dimension \(n\geq 3\), we denote by \(Ric_g\), \(R_g\) the Ricci tensor and the scalar curvature of \((M,g)\), respectively. The Schouten tensor of \((M,g)\) is defined as follows \[ A_g:=\frac{1}{n-2}(\operatorname{Ric}_g-\frac{1}{2(n-1)}R_g g). \] If \(\lambda_1,\dots ,\lambda_n\) denote the eigenvalues of the symmetric endomorphism \(g^{-1}A_g\), then the \(\sigma_k\)-curvature of \((M,g)\) is defined as the \(k\)-th symmetric elementary function of \(\lambda_1,\dots ,\lambda_n\) namely \[ \sigma_k(g^{-1}A_g)=\sum_{i_1<\dots <i_k}\lambda_{i_1}\dots \lambda _{i_k}, \] for \(1\leq k\leq n\) and \(\sigma_0(g^{-1}A_g)=1\). The \(\sigma_k\)-Yamabe problem on \((M,g)\) consists in finding metrics with constant \(\sigma_k\)-curvature in the conformal class of \(g\). The case \(k=1\) is known as the ``Yamabe problem'', whose progressive resolution is due to \textit{H. Yamabe} [Osaka Math. J. 12, 21--37 (1960; Zbl 0096.37201)], \textit{N. S. Trudinger} [Ann. Sc. Norm. Super. Pisa, Sci. Fis. Mat., III. Ser. 22, 265--274 (1968; Zbl 0159.23801)], \textit{T. Aubin} [J. Math. Pur. Appl., IX. Sér. 55, 269--296 (1976; Zbl 0336.53033)], and \textit{R. Schoen} [J. Differ. Geom. 20, 479--495 (1984; Zbl 0576.53028)]. In the present paper, the authors produce families of Riemannian metrics with positive constant \(\sigma_k\)-curvature equal to \(2^{-k}\) \(n\choose k\) by performing the connected sum of two given compact nondegenerate \(n\)-dimensional solutions \((M_1,g_1)\) and \((M_2,g_2)\) of the (positive) \(\sigma_k\)-Yamabe problem, provided \(2\leq 2k<n\). The problem is equivalent to solving a second-order fully nonlinear elliptic equation. A metric \(g\) on \(M\) is said to be \(k\)-admissible if it belongs to the \(k\)-th positive cone \(\Gamma^+_k\), where \(g\in \Gamma^+_k\) \(\Leftrightarrow\) \(\sigma_j(g^{-1}A_g)>0\) for \(j=1,\dots ,k.\) The main result of the paper is the following theorem. Theorem 1. Let \((M_1,g_1)\) and \((M_2,g_2)\) be two compact \(n\)-dimensional \(k\)-admissible nondegenerate solutions to the positive \(\sigma_k\)-Yamabe problem, with \(2\leq 2k< n\). Then there exists a positive real number \(\varepsilon_0>0\) only depending on \(n,k\), and the \(C^2\)-norm of the coefficients of \(g_1\) and \(g_2\) such that for every \(\varepsilon \in (0,\varepsilon_0]\), the connected sum \(M_\varepsilon=M_1\sharp_\varepsilon M_2\) can be endowed with a \(k\)-admissible nondegenerate metric \(\widetilde{g}_\varepsilon\) with constant \(\sigma_k\)-curvature equal to \(2^{-k}\) \(n\choose k\). Moreover, \(\|\widetilde{g}_\varepsilon -g_i\|^{C^r(K_i)\rightarrow 0}\) for \(r> 0\) and any compact set \(K_i\subset M\backslash\{p_i\}\), the \(p_i\)'s, \(i=1,2\), being the points about which the connected sum is performed.