\(\ell^2\)-homology of right-angled Coxeter groups based on barycentric subdivisions (Q1882922)

The authors consider the vanishing of the reduced \(\ell^2\)-homology \(\mathcal H_*\) for a certain class of right-angled Coxeter groups. Their motivation comes from the Singer Conjecture: If \(M^n\) is a closed aspherical manifold, then \(\mathcal H_i(\widetilde M^n) = 0\) for all \(i \not=n/2\). A construction \textit{M. Davis} [Ann. Math. (2), 117, 293--324 (1983; Zbl 0531.57041)] associates to any finite flag complex \(L\) a right-angled Coxeter group \(W_L\) and a contractible cubical cell complex \(\Sigma_L\) on which \(W_L\) acts properly and cocompactly. If \(L\) is homeomorphic to the \((n-1)\)-sphere, then \(\Sigma_L\) is an \(n\)-manifold. If \(\Gamma\) is a torsion-free subgroup of finite index in \(W_L\) (such subgroups always exist), then \(\Gamma\) acts freely on \(\Sigma_L\) and \(\Sigma_L/\Gamma\) is an aspherical finite complex. Therefore the Singer Conjecture for such manifolds is natural. Conjecture: Suppose \(L\) is a triangulation of the \((n-1)\)-sphere as a flag complex. Then \(\mathcal H_i(\Sigma_L)=0\) for all \(i\not=n/2\). In their previous paper [Geom. Topol., 5, 7--74 (2001; Zbl 1118.58300)], the authors proved that the Conjecture holds true for \(n\leq 4\). The main result of this paper is to assert the vanishing of the reduced \(\ell^2\)-homology above middle dimensions for barycentric subdivisions of cellulations of manifolds. As a corollary, the authors obtain that the conjecture holds true for the barycentric subdivision of a PL-cellulation of the \((n-1)\)-sphere for all even \(n \leq 8\).