Some isolation results for \(f\)-harmonic maps on weighted Riemannian manifolds with boundary (Q1633537)

Let \((M,g)\) be a Riemannian manifold and \(f : M \rightarrow \mathbb{R}\) a smooth function. The Bakry-Emery Ricci tensor is given by \(\mathrm{Ric}_{f} = \mathrm{Ric} + \mathrm{Hess}(f)\). Many results involving Ricci curvature from Riemannian geometry (the case \(f \equiv 0\)) generalize to the setting involving \(\mathrm{Ric}_{f}\). For example, Myer's theorem extends in the following way. Let \(H\) and \(k\) be positive constants. If \(\mathrm{Ric}_{f} \geq (n-1)H\) and \(|f| \leq k \), then \(M\) is compact and \(\mathrm{diam}_{M} \leq \frac{\pi}{\sqrt{H}} + \frac{4k}{(n-1)\sqrt{H}}\). There is also an analogue of the Bishop-Gromov volume comparison theorem, where one considers the weighted volume \(\mathrm{vol}_{f}(B(p,r)) = \int_{B(p,r)} e^{-f}~\mathrm{dvol}_{g}\) of a metric ball \(B(p,r)\) of center \(p\) and radius \(r\). For further information see for example [\textit{G. Wei} and \textit{W. Wylie}, J. Differ. Geom. 83, No. 2, 377--405 (2009; Zbl 1189.53036)]. In this article the authors pursue a different direction, introduced by Lichnerowicz, where one generalizes harmonic functions to \(f\)-harmonic and \((p,f)\)-harmonic functions. In the unweighted case let \((M,g)\) and \((N,h)\) be compact Riemanian manifolds. For an arbitrary smooth function \(\varphi : M \rightarrow N \) let \(e(\varphi) = |d \varphi|^{2} : M \rightarrow \mathbb{R}\), the energy density function. Define the energy \(E(\varphi) = \int_{M} e(\varphi)~ d vol_{g}\). The function \(\varphi\) is \(\mathit{harmonic}\) if it is a critical point of the energy functional \(E\). Now let \((M,g)\) and \((N,h)\) be as before. Let \(f : M \rightarrow \mathbb{R}\) be a smooth function, and let \(d\sigma = e^{-f}~ \mathrm{dvol}_{g}\) be the corresponding smooth volume form on \(M\). For a smooth function \(\varphi : M \rightarrow N\) define \(E_{f}(\varphi) = \int_{M} e(\varphi)~d\sigma\). The function \(\varphi\) is said to be \textit{\(f\)-harmonic} if it is a critical point of the functional \(E_{f}\). For a positive number \(p\) we say that \(\varphi : M \rightarrow \mathbb{R}\) is \textit{\((p,f)\)-harmonic} if it is a critical point of the functional \(E_{(p,f)}\) given by \(E_{(p,f)}(\varphi) = \int_{M} |d \varphi|^{p}~ d\sigma\). A classical result of Eells-Sampson states that if \(\mathrm{Ric}\geq 0\) on \(M\) and for the sectional curvature holds \(K \leq 0\) on \(N\), then any harmonic function \(\varphi : M \rightarrow N\) is totally geodesic. Moreover if \(\mathrm{Ric}\) is positive at one point of \(M\), then \(\varphi\) is a constant function. Lichnerowicz generalized this result to the setting of \(f\)-harmonic functions for manifolds without boundary, and Ilias generalizes it to the case of manifolds with boundary. The Ricci tensor \(\mathrm{Ric}\) in these versions is replaced by \(\mathrm{Ric}_{f}\). The authors consider another generalization of the Eells-Sampson result in which \(f\)-harmonic maps \(\varphi\) must be constant if the curvatures of \(M\) and \(N\) are suitably bounded and the energy density norm \(|e(\varphi)|\) is suitably small. In the \(f\)-harmonic case replace \(\mathrm{Ric}\) by \(\mathrm{Ric}_{f}\) and let \(k_{1}\), \(k_{2}\) be positive numbers such that \(\mathrm{Ric}_{f} \geq k_{1}\) on \(M\) and \(K \leq k_{2}\) on \(N\). Then an \(f\)-harmonic function \(\varphi\) is constant if \(|e(\varphi)|_{L^{\infty}} < \frac{n k_{1}}{(n-1)k_{2}}\). More generally, the authors replace the Ricci tensor Ric by \(\mathrm{Ric}^{m}_{f} = \mathrm{Ric}_{f} - \frac{df \otimes df}{m-n}\), where \(m > n = \dim M\). A representative of the results obtained is the following. Let \(k_{1}\), \(k_{2}\) be positive numbers such that \(\mathrm{Ric}_{f}^{m} \geq k_{1}\) on \(M\) for some \(m > n\) and \(K \leq k_{2}\) on \(N\). Let \(m \geq 6\) and let \(\varphi : M \rightarrow N\) be an \(f\)-harmonic map. Then \(\varphi\) is constant if \(|e(\varphi)|_{L^{m/2}~d\sigma} \leq C\), where \(C\) is a small positive constant depending on \(k_{1}\), \(k_{2}\), \(m\) and \(\mathrm{vol}_{f}(M) = \int_{M} e^{-f}~\mathrm{dvol}_{g}\). In the final part of the paper the authors extend the results above from the \(f\)-harmonic setting to the \((p,f)\)-harmonic setting.