On the quantum group \(SL_ q(2)\) (Q1918116)

Consider the group \(SL(2)\) of \(2 \times 2\) complex matrices with determinant 1. The algebra \(A\) of polynomial (regular) functions on the group can be characterized as the algebra over \(\mathbb{C}\) with identity, generated by four mutually commuting elements \(a,b,c,d\) satisfying the condition \(ad- bc =1\). These elements are the matrix elements, considered as functions on the group. Dual to the group multiplication is the coproduct on \(A\) given by \[ \Delta (a) = a \otimes a+ b \otimes c, \quad \Delta(b) = a \otimes b+ b\otimes d, \quad \Delta (c) = c\otimes a+d \otimes c, \quad \Delta (d)= c \otimes b+ d \otimes d. \] There is a standard way to quantize \(SL(2)\). Take any non-zero complex number \(q\). Now let \(A_q\) be the algebra over \(\mathbb{C}\) with identity generated by elements \(a,b,c,d\) satisfying the commutation rules \[ ba= qab,\;ca= qac, \quad db= qbd,\;dc= qcd, \quad bc= cb,\;da - ad= (q-q^{-1})bc \] with quantum determinant \(ad-q^{-1} bc=1\). This algebra is again made into a Hopf algebra by using the same form for the comultiplication above. For \(q=1\), clearly we get the algebra \(A\) of polynomial functions on the group \(SL(2)\). Therefore, the algebra \(A_q\) is generally thought of as the algebra of `polynomial functions on the quantum group' \(SL_q(2)\). See [\textit{L. Faddeev}, \textit{N. Reshetikhin} and \textit{L. Takhtadzhyan}, Leningr. Math. J. 1, 193-225 (1990); translation from Algebra Anal. 1, No. 1, 178-206 (1989; Zbl 0715.17015)] for more details. In the classical case, the subgroup \(H\) of diagonal matrices yields a quotient \(S=A/I\) of \(A\) where \(I\) is the ideal generated by the elements \(b\) and \(c\). This quotient \(S\) is generated by the elements \(\widetilde a\) and \(\widetilde d\) (the images under the quotient map of \(a\) and \(d)\) and they satisfy \(\widetilde a \widetilde d= \widetilde d \widetilde a =1\). This ideal is also a two-sided Hopf ideal and the comultiplication \(\Delta\) is well-defined on the quotient and we have \(\Delta (\widetilde a) = \widetilde a\otimes \widetilde a\) and \(\Delta (\widetilde d) = \widetilde d \otimes \widetilde d\). Similarly, in the quantum case, the ideal \(I_q\) of \(A_q\) generated by \(b\) and \(c\) is a Hopf ideal and the quotient is a Hopf algebra. In fact, it is easily seen that \(A_q/I_q\) is isomorphic with \(S\) as a Hopf algebra (for any \(q)\). In the paper under review the authors try to construct other (but similar) quantizations of \(SL(2)\) by looking at other possible Hopf algebras \(A\) with a Hopf ideal \(I\) such that the Hopf algebra \(A/I\) is isomorphic with the Hopf algebra \(S\) considered above. The starting point in the general picture is as follows. Let \(A\) be any Hopf algebra and suppose that \(I\) is a Hopf ideal. Denote by \(S\) the quotient Hopf algebra \(A/I\). Consider the graded algebra \(\text{gr} (A) = \oplus^\infty_{n=0} \text{gr}_n(A)\) where \(\text{gr}_n(A) = I^n/I^{n+1}\) (and so \(\text{gr}_0 (A) = S)\). The product in \(A\) induces a natural \(S\)-bimodule structure on \(\text{gr} (A)\). On the other hand, we also have an \(S\)-bicomodule structure on \(A\) given by the maps \((\pi \otimes \text{id}) \Delta: A \to S \otimes A\), \((\text{id} \otimes \pi) \Delta: A \to A \otimes S\) where \(\pi\) denotes the quotient map from \(A\) to \(S\). Also this bicomodule structure induces a natural \(S\)-bicomodule structure on the graded Hopf algebra \(\text{gr} (A)\). Moreover, the module and the comodule structures are compatible in the standard way, making \(\text{gr} (A)\) into a Hopf bimodule [see e.g. \textit{E. Abe}, Hopf algebras. Cambridge University Press (1980; Zbl 0476.16008) and \textit{M. Sweedler}, Hopf algebras. Benjamin (1969; Zbl 0194.32901)]. The authors use the term \(S\)-tetramodule (cf. definition 2.5 in the paper). Also \(\text{gr}_1\) is such an \(S\)-tetramodule, further denoted by \(T\). Looking again at the case where \(S\) is the Hopf algebra corresponding to polynomial functions on the group \(H\) of diagonal matrices of \(SL(2)\), the authors investigate the different possibilities for \(\text{gr} (A)\) by imposing extra conditions on \(T\). One of the basic observations is that for this case, it is rather easy to characterize the \(S\)-tetramodules. In the paper, together with the direct approach, also the dual approach is followed in the general setting. Remark that there is a natural Hopf algebra \(U_q\) [cf. \textit{V. Drinfel'd}, Proc. Int. Congr. Math., Berkeley 1986, Vol. 1, 798-820 (1987; Zbl 0667.16003) and \textit{M. Jimbo}, Lett. Math. Phys. 11, 247-252 (1986; Zbl 0602.17005) which is dually paired with the Hopf algebra \(A_q\) described above (see e.g. the reviewer, Bull Lond. Math. Soc. 25, 209-230 (1993; Zbl 0796.16034)]. All these observations in the general abstract situation and by imposing natural conditions, yield in the end dual pairs \((A_q (n, \varepsilon)\), \(U_q (n, \varepsilon))\), generalizing the pair \((A_q, U_q)\) for any \(n\in \mathbb{N}\) and for corresponding \(n\)-th roots of unity \(\varepsilon\). The algebras \(U_q (n, \varepsilon)\) turn out to be genuine Hopf algebras. For the algebras \(A_q (n, \varepsilon)\), the natural candidate for the comultiplication, takes values in some completed tensor product, while the associated graded algebras \(\text{gr} (A_q (n, \varepsilon))\) are graded Hopf algebras.