Hilbert space representations of cross product algebras (Q1395797)

The structures under consideration in this paper are particular cases of the cross product (aka the smash product) algebra \(\mathcal{U}\ltimes\mathcal{X}\) of a Hopf *-algebra, \( \mathcal{X}\) and a unital right \(\mathcal{U}\)-module *-algebra. (The cross product contains each factor as a subalgebra.) The authors suggest the following interpretation of this: view \(\mathcal{X}\) as an algebra of functions on a quantum space and the elements of \(\mathcal{U}\) as acting on them analogously to differential operators acting on functions in the usual sense. Then \(\mathcal{U}\ltimes\mathcal{X}\) is an algebra of `differential operators' with `coefficients in \(\mathcal{X}\)'. Hence, as for algebras of differential operators in the usual sense, one should consider Hilbert space *-representations of these objects. Mainly, the authors consider the cross product of \(\mathcal{U}_q(gl_2)\) with \(\hat{\mathcal{O}}(\mathbb{C}_q^2)\) or \(\hat{\mathcal{O}}(\mathbb{R}_q^3)\); and of \(\mathcal{U}_q(su_2)\) with \(\mathcal{O}(SU_q(2))\) or \(\mathcal{O}(S_q^2)\). The background material necessary to understand this work is a familiarity with the theories of the representation of quantum groups and of unbounded operator algebras. A significant class of representations of the particular cross product algebras is given in this paper. Of particular importance is the representation they term the Heisenberg representation, briefly described as follows. Let \(h\) be a \(\mathcal{U}\)-invariant state on the algebra \(\mathcal{X}\), which is to be a right \(\mathcal{U}\)-module, with module action \(\triangleleft\), let \(S\) be the antipode of \(\mathcal{U}\), and let \(\mathcal{U}_0\) be a unital *-subalgebra of \(\mathcal{U}\) which is also a right coideal. Let \(\pi_h\) be the GNS *-representation for \(h\) acting on the Hilbert space \(\mathcal{H}\), with cyclic vector \(v_h\). Then for all \(f\in \mathcal{U}_0\), \(x\in \mathcal{X}\), the Heisenberg representation of \(\mathcal{U}_0\ltimes \mathcal{X}\) (with respect to \(h\)) is the closure of the unbounded *-representation obtained from \[ \tilde{\pi}_h(f)(\pi_h(x)v_h)=\pi_h(x\triangleleft S^{-1}(f))v_h. \] One knows that there is a unique invariant linear functional \(h\) on \(\mathcal{O}(SL_q(2))\) satisfying \(h(1)=1\), analogous to the Haar measure on a compact group. From \(h\) one can construct a state, also denoted \(h\), on \(SU_q(2)\) and obtain thereby a Peter-Weyl theorem for \(\mathcal{O}(SU_q(2))\); \(h\) is called the Haar state. The Heisenberg representation for \(h\) is shown to be the only closed irreducible *-representation of \(\mathcal{U}_q(su_2)\triangleleft \mathcal{O}(SU_q(2))\) whose restriction to the first factor is integrable (in the author's sense). More generally, if we replace \(\mathcal{O}(SU_q(2))\) by a general \(\mathcal{X}\) as above, the authors prove that a closed *-representation of \(\mathcal{U}_q(su_2)\triangleleft \mathcal{X}\) is unitarily equivalent to the Heisenberg representation \(\pi_h\) (for the Haar state) if and only if it is irreducible and its restriction to the first factor is integrable. Loosely speaking, the rest of the paper is devoted to obtaining constructive expressions for invariant states on the cross product algebras noted above. Here ``constructive'' means an expression as a quantum trace, and two different representation approaches are developed.