A variational characterization of calibrated submanifolds (Q6098495)

From MaRDI portal
scientific article; zbMATH DE number 7695760
Language Label Description Also known as
English
A variational characterization of calibrated submanifolds
scientific article; zbMATH DE number 7695760

    Statements

    A variational characterization of calibrated submanifolds (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    14 June 2023
    0 references
    The typical setting in calibrated geometries consists of a manifold and a calibration compatible with (or inducing) a Riemannian metric on the manifold, and then possibly the search for the calibrated submanifolds, which are in particular of minimal volume due to the properties of the calibration being both of comass 1 and a closed differential form. In this article, the authors begin with a Riemannian manifold \((\overline{M},\overline{g})\), consider any given closed embedded \(k\)-submanifold \(M\subset \overline{M}\), and study the volume functional \(\overline{g}\longmapsto\mathcal{V}(\overline{g})= \mathrm{vol}(M,g)= \int_M\mathrm{vol}_{(M,g)}\) where \(g=\overline{g}_{|M}\). The functional clearly has no critical points on the space of all metrics on \(\overline{M}\), so a novel kind of calibration theory is designed in this paper. The problem was also studied in [\textit{C. Arezzo} and \textit{J. Sun}, J. Reine Angew. Math. 709, 171--200 (2015; Zbl 1331.53105); \textit{J. Sun}, J. Math. Anal. Appl. 434, 1474--1488 (2016; Zbl 1328.53096); \textit{Q. Tan}, Ann. Global Anal. Geom. 53, 217--231 (2018; Zbl 1468.53021)]. The manifold is now assumed to be endowed with a set of \(k\)-forms \(\mu\), not necessarily closed. Then the variational calculus of \(\mathcal V\) is restricted to those metrics which live in a subclass \(\mathcal G\) determined by \(\mu\), where a fortiori it is required that \(\overline{g}\) keeps \(\mu\) with comass 1. A general and main assertion is the so-called ``meta-theorem''. Roughly, condensed in Theorems A and B in the paper, it says that the metric \(\overline{g}\) on \(\overline{M}\) is a critical point of \(\mathcal V\) if and only if \(M\) is calibrated by \(\mu\), this is, \(\mu_{|M}=\mathrm{vol}_{(M,g)}\). The result is established rigorously and developed \textit{only} in four specific contexts or geometries: (1) The first case is that of a Hermitian structure \((\overline{M},\overline{g},J,\omega)\). Here, the family \(\mathcal G\) is given as the metrics on \(\overline{M}\) of the form \(\overline{g}_t(X,Y) = \omega_t(X,JY)\), where \(\omega_t =\omega + \mathrm{d}\alpha_t^{(1,1)}\) and \(\alpha_t\) is a one-parameter family of compactly supported smooth 1-forms on \(\overline{M}\) with \(\alpha_{0}=0\), for \(|t|\) sufficiently small so as to assure positive definiteness. All metrics define Hermitian structures (eventually with torsion) compatible with the same \(J\). The \(2k\)-forms \(\mu\) are now the typical forms \(\frac{1}{k!}\omega_t^k\) and the fixed oriented closed \(2k\)-submanifold \(M\) is a complex submanifold of \((\overline{M},J)\). (2) The second case is a similar variation of a metric induced by a \(\mathrm{G}_2\)-structure \(\varphi\) on a given 7-dimensional manifold \((\overline{M},\varphi)\). As it is well known, the structure 3-form \(\varphi\) induces a metric \(\overline{g}\). Hence the authors may consider small deformations via \(\varphi_t=\varphi+\mathrm{d}\beta_t\) with \(\beta_{0}=0\). The submanifold \(M\) is a 3-manifold, and if calibrated it is associative. Equivalently, \(\overline{g}\) is critical for the functional \(\mathcal V\) in the directions of \(\mathrm{d}\dot{\beta}={\frac{d}{dt}}_{\mid_{t=0}}{\varphi}_t\). Indeed the main theorem holds true for the starting \(\mathrm{G}_2\)-structure. (3) The third case is that of the calibration \(\star\varphi\) and the corresponding calibrated and coassociative submanifolds. (4) The fourth case is for \(\mathrm{Spin}(7)\) structures on 8-manifolds and the Cayley submanifolds. Here the family of variations of the metric has to be carefully chosen and theorems A and B reveal marked differences from the first three cases. A difficult task in all four contexts seems to be proving Theorem B, i.e., to see that \(M\) is calibrated by \(\mu\) if \(\overline{g}\) is a critical point of \(\mathcal V\), and moreover to see that \(M\) is of the required type, respectively, complex, associative, coassociative or Cayley. The last section is devoted to observations and generalisations of the variational characterization of submanifolds. Firstly, the more abstract proof of Theorem A in all cases, i.e., to see that \(\overline{g}\) is a critical point of \(\mathcal V\) if \(M\) is calibrated by \(\mu\). Secondly, considering the possible role of the Berger-Ebin theorem in the study of such variations. Thirdly, the framework of variations of the metric on \(\overline{M}\) induced by the \(k\)-forms \(f_t^*\mu\) defined by the flow of diffeomorphisms \(f_t\) of a given vector field on \(\overline{M}\). Fourthly, an application of the theory to the variational characterization of the so-called Smith maps, providing further a kind of Theorems A and B about the equivalence between such maps and critical points of the \(k\)-energy.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    calibration
    0 references
    volume functional
    0 references
    Hermitian manifolds
    0 references