A class of Einstein submanifolds of Euclidean space (Q2074693)

Since the Riemann-Koebe uniformization theorem, the use of ``canonical metrics'' to understand the landscape of geometric structures has been at the forefront of mathematical research. One of the leading candidates for the notion of a canonical metric is given by those Riemannian metrics \(g\) of constant Ricci curvature: \[\text{Ric}_g = \lambda g, \] where \(\lambda \in \mathbb{R}\) is some constant. Such metrics are called Einstein, and a Riemannian manifold is said to be Einstein if it supports an Einstein metric. For surfaces and threefolds, Einstein metrics have constant sectional curvature; therefore, the notion is of less interest in these low dimensions. In dimensions four and up, however, Einstein metrics deviate from metrics of constant sectional curvature. By an observation of Cartan (communicated by \textit{T. Y. Thomas} [Am. J. Math. 59, 793--794 (1937; JFM 63.0701.02)]) and the work of \textit{A. Fialkow} [Ann. Math. (2) 39, 762--785 (1938; JFM 64.1361.03)], an Einstein hypersurface \(f : M^n \to \mathbb{R}^{n+1}\), \(n \geq 3\), is either flat or an open subset of a round sphere. In particular, if \(M^n\) is a complete manifold, then it is either a cylinder over a complete plane curve or a round sphere. Structure results concerning Euclidean Einstein hypersurfaces are lacking. As remarked by the authors, it seems that the aforementioned result is the only classification theorem under purely intrinsic assumptions. There are, however, results under extrinsic assumptions. For instance, \textit{C.-R. Onti} [Arch. Math. 110, No. 5, 523--531 (2018; Zbl 1408.53025)] obtained a classification theorem by assuming that the submanifold has parallel mean curvature vector and flat normal bundle. In [\textit{M. Dajczer} et al., Proc. Am. Math. Soc. 146, No. 9, 4035--4038 (2018; Zbl 1402.53014)] it was shown that any Einstein Euclidean submanifold with flat normal bundle is locally holonomic, i.e., the manifold supports a system of orthogonal coordinates such that the coordinate vector fields diagonalize the second fundamental form of the immersion at every point. To state the main theorem of the present article, let us recall that an \((n-2)\)-rotational submanifold \(f : M^n \to \mathbb{R}^{n+p}\), \(n \geq 3\), with axis \(\mathbb{R}^{p+1}\) over a surface \(g : L^2 \to \mathbb{R}^{p+2}\) is the \(n\)-dimensional submanifold generated by the orbits of the points of \(g(L)\) (disjoint from \(\mathbb{R}^{p+1}\)) under the action of the subgroup \(\text{SO}(n-1)\) of \(\text{SO}(n+p)\) which keeps \(\mathbb{R}^{p+1}\) invariant, pointwise. Let us further denote by \(F^m(\varepsilon)\), \(m \geq 2\), an Einstein manifold with Ricci curvature \((m-1)\varepsilon\), where \(\varepsilon \in \{-1, 1, 0 \}\). The main theorem is then the following: Theorem. Let \(M^n = L^2 \times_{\varphi} F^{n-2}(\varepsilon)\), \(n \geq 5\), be a complete simply connected Einstein manifold whose sectional curvature is not constant on any open subset. If \(F^{n-2}(\varepsilon)\) has constant sectional curvature and \(f : M^n \to \mathbb{R}^{n+2}\) is an isometric immersion, then one of the following holds: \begin{itemize} \item[(i)] \(M^n = \mathbb{S}^2(r_1) \times \mathbb{S}^{n-2}(r_2)\) is a Clifford torus and \(f\) is the product of the inclusions \(\mathbb{S}^{n_j}(r_j) \subset \mathbb{R}^{n_j+1}\), where \(n_j=2, n-2\). \item[(ii)] \(M^n = \mathbb{R}^2 \times_{\varphi} \mathbb{S}^{n-2}(1)\) is endowed with the generalized Schwarzchild metric and \(f\) is an \((n-2)\)-rotational submanifold. \end{itemize}