ADM mass for \(C^0\) metrics and distortion under Ricci-DeTurck flow (Q6186150): Difference between revisions

This paper is concerned with defining and showing various properties of a \(C^0\) version of the ADM mass. This is largely possible due to the author's past work on a notion of scalar curvature when the metric is only \(C^0\), which uses the Ricci-De Turck flow as a mollifier and then takes a limit as one goes backward in time, leading to the notion of scalar curvature lower bounds ``in the sense of Ricci flow'' [\textit{P. Burkhardt-Guim}, Geom. Funct. Anal. 29, No. 6, 1703--1772 (2019; Zbl 1430.53042)]. The main connection between scalar curvature lower bounds and the ADM mass [\textit{R. A. Bartnik}, in: Selected works. Edited by Piotr T. Chruściel, James A. Isenberg and Shing-Tung Yau. Somerville, MA: International Press. 67--99 (2021; Zbl 1464.53042)] that one should keep in mind is the positive mass theorem. This tells us that the only way for an asymptotically locally euclidean manifold with \(\mathrm{scal}_g \geq 0\) to have vanishing mass is to be flat [\textit{R. Schoen} and \textit{S.-T. Yau}, Commun. Math. Phys. 65, 45--76 (1979; Zbl 0405.53045) ; \textit{Y. Li}, Geom. Topol. 22, No. 3, 1837--1891 (2018; Zbl 1387.53088)]. The contents of this paper seem very likely to serve as a major first step toward proving a \(C^0\) version of the positive mass theorem. As for how the mass is defined, the author first formulates a local version which, by using a test function depending on a radial coordinate \(r\) and integration by parts, is expressed in terms of quantities depending on the Riemannian metric only in a \(C^0\) sense. To show that the global mass exists for \(r \rightarrow \infty\), a detailed analysis of deformations of the metric under the Ricci-De Turck flow is conducted, provided the initial metric is sufficiently close to the Euclidean one in certain regions and senses. This is the natural setting when one keeps the positive mass theorem in mind. Starting the flow from such rough initial data relies on work due to \textit{H. Koch} and \textit{T. Lamm}, [Asian J. Math. 16, No. 2, 209--235 (2012; Zbl 1252.35159)]. The resulting estimates are used to control the distortion of the local mass under the flow, which acts as a monotonicity result and leads to the proof that a radial limit of the local mass exists. This then yields a mass which is defined over the entire manifold. A key ingredient for this distortion result is the choice of a suitable test function, which turns out to be one that evolves similarly to the radial derivative of a spherically symmetric solution to the backwards heat equation on Euclidean space. To show that the author's notion of a \(C^0\) mass is a suitable one, the following results are proved: \begin{itemize} \item[1.] Provided the metric is \(C^2\) and the ADM mass exists, then the two masses agree. \item[2.] Furthermore, if \(\mathrm{scal}_g \geq 0\) in the sense of Ricci flow on \(M \backslash K\) for a compact set \(K \subset M\) and \(g\) is sufficiently close to the Euclidean metric then the (global) \(C^0\) mass exists as a radial limit, is either finite or equal to \(+\infty\), and is independent of the choice of test function and chart. Moreover, finiteness holds if and only if a particular condition involving the scalar curvature of time slices of Ricci-DeTurck flows associated to g is satisfied. \end{itemize}