On the \(E\)-polynomials of a family of \({\mathrm {Sl}}_n\)-character varieties (Q889940): Difference between revisions

Let \(n\) and \(g\) be positive integers, and \({\mathbb K}\) be a field that contains a primitive \(n\)th root of unity \(\zeta_n\). The \textit{twisted} \(\text{Sl}_n\)\textit{-character variety} \({\mathcal M}(\text{Sl}_n({\mathbb K}))\) is given by all the classes of \(2g\)-tuples \((A_1,B_1,\dots,A_g,B_g)\) in \(\text{Sl}_n({\mathbb K})^{2g}\) satisfying the matrix equation \[ [A_1,B_1]\cdots [A_g,B_g] = \zeta_n\text{Id} \] modulo simultaneous \(\text{PGL}_n({\mathbb K})\)-conjugation. The \(E\)-polynomial of a variety \(X\) is \[ E(q;X) := H_c(\sqrt{q},\sqrt{q},-1;X) \] where \[ H_c(x,y,t;X) := \sum h_c^{p,q;j}(X) x^py^qt^j \] and the \(h_c^{p,q;j}(X)\) are the mixed Hodge numbers with compact support of \(X\). The aim of the paper is to compute the \(E\)-polynomials of the \({\mathcal M}(\text{Sl}_n({\mathbb C}))\) for any genus \(g\) and any dimension \(n\). This is achieved by counting the number \(N_n(q)\) of points that these varieties have over the finite fields \({\mathbb F}_q\) (where \(n|q-1\)), and by showing (this is the main theorem of the paper) that the character varieties \({\mathcal M}(\text{Sl}_n({\mathbb C}))\) have a polynomial count, that is, that there exists a polynomial \(P(q)\in{\mathbb Z}[q]\) such that \(\#{\mathcal M}(\text{Sl}_n({\mathbb F}_q))=P(q)\) for sufficiently many prime powers \(q\). The author also finds explicit formulas for the counting functions and for the \(E\)-polynomials; as corollaries, he proofs that: {\parindent=6mm \begin{itemize}\item[(i)] the \(E\)-polynomial of \({\mathcal M}(\text{Sl}_n({\mathbb C}))\) is palyndromic and monic; \item [(ii)] the character varieties \({\mathcal M}(\text{Sl}_n({\mathbb C}))\) are connected; \item [(iii)] the Euler characteristic of \({\mathcal M}(\text{Sl}_n({\mathbb C}))\) is 1 for \(g=1\) and \(\mu(n)n^{4g-3}\) for \(n\geq 2\). \end{itemize}}