The critical asymptotics of scalar curvatures of the conformal complete metrics with negative curvature (Q1345232)

A number of people have considered the following problem (P) that is closely related to the Yamabe problem: Let \((M^ n,g)\) denote a complete, Riemannian manifold of dimension \(n \geq 3\) and let \(K : M^ n \to \mathbb{R}\) be a smooth function. Under what conditions is \(K\) the scalar curvature of some complete metric \(g^*\) that is conformal to \(g\)? In this article the authors restrict themselves to the special case that \(M^ n\) is simply connected and the sectional and Ricci curvatures of \(g\) satisfy \(\text{Sec}(g) \leq -B^ 2 < 0\) and \(\text{Ricc}(g) \geq - (n-1)A^ 2\) for some positive constants \(A\) and \(B\). For a fixed point \(x_ 0\) of \(M^ n\) let \(r(x) = d(x,x_ 0)\), where \(d\) denotes the Riemannian distance with respect to \(g\). The results here strengthen parts of work of Z. Jin, whose results may be stated as follows, assuming the same conditions on \(M^ n\). Theorem: (1) If there exist positive constants \(C\), \(\sigma\) such that \(K(x) \leq - {Ce}^{\sigma r(x)}\) for large \(r(x)\), then problem (P) has no solution. (2) If there exist positive constants \(C_ 1\), \(C_ 2\), \(\alpha\), \(k\) such that \(-C_ 1 \leq K(x) \leq -C_ 2 e^{-\alpha r^ k(x)}\) for all \(x\in M^ n\), then (P) has a solution. In contrast, the authors of this article prove the following Theorem: (1) If there exist positive constants \(C_ 1\), \(C_ 2\), \(\alpha\), \(k\) such that \(-C_ 1(1+r(x)^ 2) \leq K(x) \leq - C_ 2 e^{\alpha r^ k(x)}\) for all \(x \in M^ n\), then (P) has a solution. (2) If there exist positive constants \(C\), \(r_ 0\), \(\delta\) and a geodesic \(\{\gamma(t) : t \geq 0\}\) in \((M^ n,g)\) such that \(K(x) \leq - {Cr}^{2+\delta} (x)\) for all \(x\) in the geodesic tube \(T_ \varepsilon (\gamma) =\) \{open \(\varepsilon\)-neighborhood of \(\gamma\) in \(M^ n\}\) and for \(r(x) \geq r_ 0\), then (P) has no solution. The authors also state an improved version of the result above, which follows from a modification of the text proof. The existence proof of (1) uses the super and subsolution method.