Generalized \(q\)-Schur algebras and quantum Frobenius (Q2642045)

From the introduction: The quantum Frobenius map and it splitting are shown to descend to maps between generalized \(q\)-Schur algebras at a root of unity. We also define analogs of \(q\)-Schur algebras for any affine algebra, and prove the corresponding results for them. In their study of the quantum group \(U(\text{sl}_n)\), \textit{A. A. Beilinson, G. Lusztig} and \textit{R. MacPherson} [[BLM] Duke Math. J. 61, No. 2, 655--677 (1990; Zbl 0713.17012)] defined a family of finite-dimensional quotients known as \(q\)-Schur algebras. Subsequently Lusztig defined similar families of finite dimensional quotients for any quantum group of finite type, which here, following [\textit{S. Doty}, Represent. Theory 7, 196--213 (2003; Zbl 1030.17011)], we call ``generalized \(q\)-Schur algebras''. We show that, at a root of unity, the quantum Frobenius map and a splitting of it (which extends the splitting map defined by Lusztig) descend to a family of maps between generalized \(q\)-Schur algebras (Theorem 4.5). Note that in the general situation, the quantum Frobenius map is a homomorphism from the quantum group at a root of unity to the integral enveloping algebra of a different quantum group (the Langlands' dual in the case where the order of the root of unity is divisible by 6 for example). For the original \(q\)-Schur algebras we also give a simple characterization of the embedding in the geometric context of [BLM] (Lemma 5.2). Finally we define a family of algebras which are affine analogs of generalized \(q\)-Schur algebras (essentially an algebraic version of the algebras considered in the work of Nakajima), and establish the analogous results for these algebras (Theorem 6.2). One impetus for this work came from discovery by \textit{M. Fayers} and \textit{S. Martin} [J. Algebra 240, No. 2, 859-873 (2001; Zbl 0990.20029)] of an explicit embedding of the Schur algebra \(S_k(2, r)\) into the Schur algebra \(S_k(2, pr)\) when \(k\) is a field of characteristic \(p\). This paper grew out of a desire to find a quantum version of their embedding, which turns out to be provided by the splitting map for the quantum Frobenius. Our results show that the quantum Frobenius is compatible with certain families of based ideals in the modified quantum group, suggesting that there are compatibilities between this map and the canonical basis that have yet to be understood. We hope to return to this issue in future work (also related to [K. McGerty, Hall algebras and quantum Frobenius, preprint, arXiv:math.QA/0601150]).