An extended Cheeger-Müller theorem for covering spaces (Q2568994)

\textit{D. B. Ray} and \textit{I. M. Singer} [Adv. Math. 7, 145--210 (1971; Zbl 0239.58014)] defined and studied an analytic version of the topological Reidemeister-Franz torsion. They conjectured that their analytic torsion was identical to the topological torsion, a fact later proved independently by \textit{J. Cheeger} [Ann. Math. (2) 109, 259--322 (1979; Zbl 0412.58026)] and \textit{W. Müller} [Adv. Math. 28, 233--305 (1978; Zbl 0395.57001)]. \textit{J.-M. Bismut} and the author [An extension of a theorem by Cheeger and Müller. Astérisque. 205. Paris: Société Mathématique de France (1992; Zbl 0781.58039)] reformulated the Cheeger-Müller theorem as a statement between Reidemeister and Ray-Singer metrics and extended the Cheeger-Müller theorem to the case when the metric on the auxiliary bundle is not flat. They also gave a new proof of the Cheeger-Müller theorem using the Witten Laplacian to compare the analytic torsion of the manifold to the torsion of the Morse-Smale complex associated to a Morse function. The paper under review generalizes this work to \(L^2\)-torsion on infinite Galois covering spaces of closed manifolds. An important contribution to this area is that the author does not assume the ``determinant class condition'', which has been assumed in various previous works on \(L^2\)-torsion. We remark that the first paper to remove the ``determinant class condition'' is [\textit{M. Braverman, A. Carey, M. Farber} and \textit{V. Mathai}, Commun. Contemp. Math. 7, No. 4, 421--462 (2005; Zbl 1079.55015)]. We also remark that the paper under review generalizes the result of Braverman et al to the non-unitary flat bundle case and thus is in some sense the strongest result so far. Finally, the proof of the main theorem (Theorem 4.2) is a direct adaption of the proof of the main result (Theorem 0.2) of the above cited paper of J.-M. Bismut and the author [Zbl 0781.58039]. In particular, the proof is purely analytical, very accessible to readers familiar with \(L^2\)-Witten complex, and it avoids the ideas of relative torsion.