Cohomology for infinitesimal unipotent algebraic and quantum groups. (Q692788): Difference between revisions

Let \(G\) be a simple, simply connected algebraic group over an algebraically closed field \(k\) of prime characteristic \(p>2\) with \(G\) being split over \(\mathbb F_p\). Let \(B\subset G\) denote a (positive) Borel subgroup and \(U\subset B\) denote the unipotent radical of \(B\). Further, let \(\mathfrak g\) and \(\mathfrak u\) denote the (restricted) Lie algebras of \(G\) and \(U\), with similar notation for other Lie algebras. Given a subset \(J\) of simple roots for the associated (simple) root system, there is an associated Levi subgroup \(L_J\subset G\) and parabolic subgroup \(P_J=L_J\ltimes U_J\), where \(U_J\) is the unipotent radical of \(P_J\). The first main goal of the paper is to study the cohomology of the first Frobenius kernels \((U_J)_1\) and \((P_J)_1\) with coefficients in a simple \(G\)-module \(L(\lambda)\) of highest weight \(\lambda\) when the prime \(p\) is not too small. Recall that this is equivalent to the cohomology of the (finite-dimensional) restricted enveloping algebra \(u(\mathfrak u_J)\) or \(u(\mathfrak p_J)\). Using Kostant's theorem for the ordinary Lie algebra cohomology of \(\mathfrak u\), the authors improve slightly on work of \textit{E. M. Friedlander} and \textit{B. J. Parshall} [Math. Ann. 273, 353-374 (1986; Zbl 0586.20021)] to identify \(H^\bullet((U_J)_1,L(\lambda))\) as an \(L_J\)-module for any dominant weight in the fundamental \(p\)-alcove. More significantly, in the case of the trivial module (\(\lambda=0\)), the authors identify the ring structure of \(H^\bullet((U_J)_1,k)\) when the prime \(p\) is not too small, and show by example that the claim may fail for small primes. The authors then use the \(U_J\)-results to identify \(H^\bullet((P_J)_1,L(\lambda))\) as a \(P_J\)-module for a dominant weight \(\lambda\) in the fundamental \(p\)-alcove. The authors then turn to quantum analogues of these objects. Here one works over the complex numbers with an \(\ell\)th root of unity \(\zeta\) for an odd positive integer \(\ell\). Associated to \(G\) (or more precisely, to its underlying root system), one has the associated quantized enveloping algebra \(\mathbb U_q(\mathfrak g)\) for an indeterminate \(q\). One can specialize to \(\zeta\) in two ways, giving a Lustzig quantum group \(U_\zeta(\mathfrak g)\) and a De Concini-Kac quantum group \(\mathcal U_\zeta(\mathfrak g)\). The small quantum group \(u_\zeta(\mathfrak g)\) arises as a (finite-dimensional) subalgebra of the former and a quotient of the latter and is an analogue of the ordinary restricted enveloping algebra. One analogously has quantum groups corresponding to \(B\), \(U\), \(P_J\), and \(U_J\). Parallel to the results above, the authors provide a description of the cohomology of \(u_\zeta(\mathfrak u)\) and \(u_\zeta(\mathfrak p_J)\) with coefficients in certain simple \(U_\zeta(\mathfrak g)\)-modules. Further, they give a description of the ring structure of \(H^\bullet(u_\zeta(\mathfrak u),\mathbb C)\) for \(\ell\) not too small. The techniques in the quantum setting are somewhat different from those used in the algebraic group setting and follow, in part, work of the reviewer with \textit{D. Nakano, B. Parshall} and \textit{C. Pillen} [Mem. Am. Math. Soc. 1077 (2014; Zbl 1344.20066)]. Also, in this case, information on ordinary Lie algebra cohomology is replaced by the cohomology of \(\mathcal U_\zeta(\mathfrak u_J)\) due to the \textit{University of Georgia VIGRE Algebra Group} [Proc. Am. Math. Soc. 138, No. 1, 85-99 (2010; Zbl 1197.17009)].