On crystal bases and Enright's completions (Q2370206)

The paper under review discusses the interplay between crystal bases of certain not necessarily integrable modules over quantized enveloping algebras of finite-dimensional complex simple Lie algebras in the generic case and Enright's completion functor. The quantized enveloping algebra \(U_q(\mathfrak{g})\) of any finite-dimensional complex simple Lie algebra \(\mathfrak{g}\) can be decomposed as \(U_q^-(\mathfrak{g})U_q^0(\mathfrak{g})U_q^+(\mathfrak{g})\) corresponding to the root space decomposition of \(\mathfrak{g}\). As in the original setting, Enright's completion functor is defined on a certain category \(\mathcal{I}\) of \(U_q(\mathfrak{g} )\)-modules which consists of all weight modules with respect to the action of \(U_q^0(\mathfrak{g})\) that are \(U_q^-(\mathfrak{g})\)-torsion free and \(U_q^+ (\mathfrak{g})\)-finite (in the sense that all the defining generators of \(U_q^+ (\mathfrak{g})\) act locally nilpotently). Enright's completion functor on \(\mathcal{I}\) was introduced and studied by \textit{Yi Ming Zou} [Commun. Algebra 23, No. 7, 2401--2428 (1995; Zbl 0827.17015)]. In particularly, Zou proved that every module in \(\mathcal{I}\) has a completion which is unique up to isomorphism and successive completions do not depend on the chosen reduced expression of an element of the Weyl group of \(\mathfrak{g}\). The latter result shows that completion depends essentially on the representation theory of \(U_q(\mathfrak{sl}_2)\) to which the author mainly restricts his attention. Moreover, every module in \(\mathcal{I}\) is a direct sum of (not necessarily finitely many) indecomposables and the latter are given (up to isomorphism) by quantized Verma modules \(M(n)\) with integer highest weights \(n\) and the projective covers \(T(n)\) of the Verma modules \(M(-n-2)\) for any non-negative integer \(n\) so that it is enough to define completions for these modules. \textit{M.\ Kashiwara} [Commun. Math. Phys. 133, No. 2, 249--260 (1990; Zbl 0724.17009)] defines a crystal basis for any integrable \(U_q(\mathfrak{g})\)-module \(M\) to be a pair \((L,B)\) consisting of a lattice \(L\) of \(M\) (the so-called crystal lattice) and a \(\mathbb{Q}\)-basis \(B\) of \(L/qL\) (the so-called crystal). A crystal basis is a certain parametrization of vector space bases of the module with several nice properties (e.g., the crystal basis of the direct sum is the union of the crystal bases of the summands and the crystal basis of the tensor product is the tensor product of the crystal bases of the factors) encoding the intrinsic structure of the module in a combinatorial way. Later the concept of a crystal basis was extended to not necessarily integrable modules. The author introduces the notions of a complete crystal basis and a completion of a crystal basis via the analogue notions for crystal lattices. Then he proves that this notion is compatible with Enright's completion of modules, i.e., a crystal basis is complete if and only if its corresponding module is complete. Moreover, it is shown that every module in a certain subcategory of \(\mathcal{I}\) has a crystal basis that can be completed and that a completion of the crystal lattice is unique. The author concludes the paper by giving two constructions of the completion of a crystal lattice.