A von Neumann algebra framework for the duality of the quantum groups (Q1890545): Difference between revisions

The authors introduce and study what they call a Woronowicz algebra. They construct the dual and they prove a duality theorem. On could say that they obtain a duality for locally compact quantum groups, but within the von Neumann algebra framework. The notion is rather complicated and contains some technical assumptions. Therefore, in this review, let us restrict to a more descriptive formulation of the concepts and results (without going into the technical details). Consider a von Neumann algebra \(M\). A comultiplication on \(M\) is a unital, normal *-homomorphism \(\delta\) of \(M\) into the von Neumann tensor product \(M\otimes M\), satisfying coassociatvity \((\delta\otimes\iota) \delta= (\iota\otimes \delta)\delta\) (where \(\iota\) stands for the identity map). A Woronowicz algebra is a pair of a von Neumann algebra \(M\) and a comultiplication \(\delta\) on \(M\) such that there exists a left invariant weight (the left Haar weight) and such that there is an antipode admitting some kind of polar decomposition. This polar decomposition of the antipode is obtained by deforming a so-called unitary antipode (with square one) with the help of a deformation one-parameter group of automorphisms of \(M\). This notion is a natural extension of the concept of a Kac algebra as introduced by \textit{M. Enock} and \textit{J. M. Schwartz} [see e.g. Bull. Soc. Math. France, Suppl., Mémoir No. 44, 144 (1975; Zbl 0343.46044)]. For Kac algebras however, the square of the antipode is equal to one (or equivalently, the antipode is a *-anti-isomorphism) and the deformation automorphisms are trivial. This more restrictive case however turned out to be insufficient to cover the more recent examples due to \textit{V. G. Drinfeld} [see e.g. Proc. ICM Berkeley 1986, Vol. 1, 788-820 (1987; Zbl 0667.16003)] and Jimbo on the one hand and \textit{S. L. Woronowicz} on the other hand [see e.g. Commun. Math. Phys. 11, 117-181 (1987; Zbl 0628.58034)]. In these cases, the square of the antipode need not be one. So some non-trivial deformation automorphism turns out to be necessary to cover these interesting examples. The idea of deforming the antipode in order to cover also the recent examples was also formulated by E. Kirchberg at a conference in Copenhagen (1992). The authors however announced their results at a conference in Leipzig 1991 [An operator algebraic framework for the duality of quantum groups. Proc. AMP-91, Leipzig, Springer (1992)]. One of the main objects, used in the paper to construct the dual object is the Kac-Takesaki operator \(W\). It is a unitary operator on \({\mathcal H}\otimes {\mathcal H}\) where \(\mathcal H\) is the Hilbert space on which \(M\) acts in standard form. It satisfies the Pentagon equation [see e.g. \textit{S. Baaj} and \textit{G. Skandalis}, Ann. Sci. Éc. Norm. Sup. 26, No. 4, 452-488 (1993; Zbl 0804.46078)]. In fact, this operator is essentially the left regular representation if one thinks of a Woronowicz algebra as a locally compact quantum group. It is used to construct the dual von Neumann algebra \(\widehat M\). Essentially, \(\widehat M\) is the smallest von Neumann algebra such that \(W\in M\otimes \widehat M\). It is shown that \(\widehat M\) carries a natural dual comultiplication \(\widehat\delta\) and that \((\widehat M, \widehat\delta)\) is again a Woronowicz algebra. The dual of \((\widehat M, \widehat\delta)\) is shown to be canonically isomorphic to the original Woronowicz algebra \((M, \delta)\). Throughout the paper, left Hilbert algebras play an essential technical role (e.g. to construct the dual Haar weights). Recently, at a conference in Warsaw (1995), the first author announced a \(C^*\)-version of this work (in cooperation with Nakagami and Woronowicz). The two versions are of course closely related. The work should be considered as an attempt to define locally compact quantum groups and to extend Pontryagin's duality for Abelian locally compact groups to such locally compact quantum groups in such a way that also the recent examples fit into the scheme.