FRT-duals as quantum enveloping algebras (Q1398142)

Let \((A, r)\) be a coquasitriangular Hopf algebra. The FRT-dual of \(A\) is the Hopf subalgebra \(U(A)\) of the Hopf dual generated by the l-functionals \(r(. \otimes a)\) and \(\overline r(a \otimes .)\), \(a \in A\), where \(\overline r\) denotes the convolution inverse of \(r\). By results of \textit{S. Majid} [Foundations of quantum group theory, Cambridge: Cambridge Univ. Press (1995; Zbl 0857.17009)] and \textit{T. J. Hodges} [J. Algebra 192, 303-325 (1997; Zbl 0879.17003)] there exist an injective algebra map \(\iota: U(A \bowtie A) \to U(A) \otimes U(A)\) and a surjective Hopf algebra map \(\zeta: A^{\text{op}} \otimes U(A) \to U(A \bowtie A)\); here, \(A \bowtie A\) is the quantum double of \((A, r)\). This paper contributes to the study of these maps in the case where \(A = \mathbb C_q[G]\) is the quantum coordinate algebra associated to a connected complex semisimple Lie group \(G\), and \(0 \neq q \in \mathbb C\) is not a root of unity. In this case one has \(U(A) \simeq U_q(\mathfrak g)\) is (a certain version of) the quantized enveloping algebra of the Lie algebra \(\mathfrak g\) of \(G\). The main results of the paper under review state that \(\zeta\) is an isomorphism, as conjectured by Hodges in loc. cit., and on the other hand, that there exists no bialgebra structure on \(U_q({\mathfrak g}) \otimes U_q({\mathfrak g})\) such that \(\iota\) is a bialgebra map. For the classical groups, a certain choice of root vectors for \(U_q(\mathfrak g)\) is expressed in terms of l-functions, which allows the author to give an explicit formula for their coproduct.