Free quantum groups

# The K-theory of free quantum groups

Roland Vergnioux and Christian Voigt Roland Vergnioux
Laboratoire de Mathématiques Nicolas Oresme
Université de Caen Basse-Normandie
BP 5186
14032 Caen cedex
France
Christian Voigt
Mathematisches Institut
Westfälische Wilhelms-Universität Münster
Einsteinstraße 62
48149 Münster
Germany
###### Abstract.

In this paper we study the -theory of free quantum groups in the sense of Wang and Van Daele, more precisely, of free products of free unitary and free orthogonal quantum groups. We show that these quantum groups are -amenable and establish an analogue of the Pimsner-Voiculescu exact sequence. As a consequence, we obtain in particular an explicit computation of the -theory of free quantum groups.
Our approach relies on a generalization of methods from the Baum-Connes conjecture to the framework of discrete quantum groups. This is based on the categorical reformulation of the Baum-Connes conjecture developed by Meyer and Nest. As a main result we show that free quantum groups have a -element and that .
As an important ingredient in the proof we adapt the Dirac-dual Dirac method for groups acting on trees to the quantum case. We use this to extend some permanence properties of the Baum-Connes conjecture to our setting.

###### 2000 Mathematics Subject Classification:
20G42, 46L80, 19K35

## 1. Introduction

A classical result in the theory of -algebras is the computation of the -theory of the reduced group -algebra of the free group on generators by Pimsner and Voiculescu [PVfree]. Their result resolved in particular Kadison’s problem on the existence of nontrivial projections in these -algebras. More generally, Pimsner and Voiculescu established an exact sequence for the -theory of reduced crossed products by free groups [PVsequences], [PVfree]. This exact sequence is an important tool in operator -theory.
The -theory of the full group -algebra was calculated before by Cuntz in a simple and elegant way, based on a general formula for the -theory of free products [Cuntzfreeproduct]. Motivated by this, Cuntz introduced the notion of -amenability for discrete groups and gave a shorter proof of the results of Pimsner and Voiculescu [Cuntzkam]. The fact that free groups are -amenable expresses in a conceptually clear way that full and reduced crossed products for these groups cannot be distinguished on the level of -theory.
The main aim of this paper is to obtain analogous results for free quantum groups. In fact, in the theory of discrete quantum groups, the rôle of free quantum groups is analogous to the rôle of free groups among classical discrete groups. Roughly speaking, any discrete quantum group can be obtained as a quotient of a free quantum group. Classically, the free group on generators can be described as the free product

 Fn=Z∗⋯∗Z

of copies of . In the quantum case there is a similar free product construction, but in contrast to the classical situation there are different building blocks out of which free quantum groups are assembled. More precisely, a free quantum group is of the form

for matrices and such that . Here we denote by and the free unitary and free orthogonal quantum groups introduced by Wang and Van Daele [Wangfree], [vDWuniversal]. The special case and of this family reduces to the classical free group on generators.
In order to explain our notation let us briefly review some definitions. Given a matrix , the full -algebra of the free unitary quantum group is the universal -algebra generated by the entries of an -matrix satisfying the relations that and are unitary. Here denotes the matrix obtained from by taking the transpose of its adjoint. The full -algebra of the free orthogonal quantum group is the quotient of by the relation . Finally, the full -algebra of the free product of two discrete quantum groups and is the unital free product of the corresponding full -algebras, see [Wangfree]. That is, the full -algebra of a free quantum group as above is simply the free product of the -algebras and .
We remark that one usually writes and for these -algebras, compare [Banicaunitary]. Following [Voigtbcfo], we use a different notation in order to emphasize that we shall view and as the full group -algebras of discrete quantum groups, and not as function algebras of compact quantum groups.
The approach to the -theory of free quantum groups in this paper is based on ideas and methods originating from the Baum-Connes conjecture [BC], [BCH]. It relies in particular on the categorical reformulation of the Baum-Connes conjecture developed by Meyer and Nest [MNtriangulated]. In fact, our main result is that free quantum groups satisfy an analogue of the strong Baum-Connes conjecture. The precise meaning of this statement will be explained in section LABEL:secbc, along with the necessary preparations from the theory of triangulated categories. Together with the results of Banica on the representation theory of free quantum groups [Banicafo], [Banicaunitary], the strong Baum-Connes property implies that every object of the equivariant Kasparov category for a free quantum group has a projective resolution of length one. Based on this, we immediately obtain our Pimsner-Voiculescu exact sequence by invoking some general categorical considerations [MNhomalg1]. As a consequence we conclude in particular that the reduced -algebras of unimodular free quantum groups do not contain nontrivial idempotents, extending the results of Pimsner and Voiculescu for free groups mentioned in the beginning.
This paper can be viewed as a continuation of [Voigtbcfo], where the Baum-Connes conjecture for free orthogonal quantum groups was studied. Our results here rely on the work in [Voigtbcfo] on the one hand, and on geometric arguments using actions on quantum trees in the spirit of [Vergniouxtrees] on the other hand.
To explain the general strategy let us consider first the case of free unitary quantum groups for satisfying . It was shown by Banica [Banicaunitary] that is a quantum subgroup of the free product in this case. Since both and satisfy the strong Baum-Connes conjecture [Voigtbcfo], [HKatmenable], it suffices to prove inheritance properties of the conjecture for free products of quantum groups and for suitable quantum subgroups. In the case of free products we adapt the construction of Kasparov and Skandalis in [KSbuildings] for groups acting on trees, and this is where certain quantum trees show up naturally. An important difference to the classical situation is that one has to work equivariantly with respect to the Drinfeld double of .
The quantum group associated to a general matrix does not admit an embedding into a free product as above. As in [Voigtbcfo] we use an indirect argument based on the monoidal equivalences for free quantum groups obtained by Bichon, de Rijdt, and Vaes [BdRV]. This allows us to reduce to matrices , and in this case one may even assume without loss of generality. We might actually restrict attention to -matrices throughout, however, this would not simplify the arguments.
Let us now describe how the paper is organized. In section 2 we collect some preliminaries on quantum groups and fix our notation. Section LABEL:secquantumsub contains basic facts about quantum subgroups of discrete quantum groups and their homogeneous spaces. In section LABEL:secdivisible we introduce and discuss the notion of a divisible quantum subgroup of a discrete quantum group. This concept appears naturally in the context of inheritance properties of the Baum-Connes conjecture. Roughly speaking, divisible quantum subgroups are particularly well-behaved from the point of view of corepresentation theory. In section LABEL:secdirac we define the Dirac element associated to a free product of discrete quantum groups acting on the corresponding quantum tree. Moreover we define the dual Dirac element and show that the resulting -element is equal to the identity. In section LABEL:secbc we review the approach to the Baum-Connes conjecture developed by Meyer and Nest. We explain in particular the categorical ingredients needed to formulate the strong Baum-Connes property in our context. Then, using the considerations from section LABEL:secdirac and [Voigtbcfo], we prove that free quantum groups have the strong Baum-Connes property. Finally, in section LABEL:secktheory we discuss the main consequences of this result. As indicated above, we derive in particular the -amenability of free quantum groups and establish an analogue of the Pimsner-Voiculescu sequence.
Let us make some remarks on notation. We write for the space of adjointable operators on a Hilbert -module . Moreover denotes the space of compact operators. The closed linear span of a subset of a Banach space is denoted by . Depending on the context, the symbol denotes either the tensor product of Hilbert spaces, the minimal tensor product of -algebras, or the tensor product of von Neumann algebras. We write for algebraic tensor products. For operators on multiple tensor products we use the leg numbering notation. If and are unital -algebras we write for the unital free product, that is, the free product of and amalgamated over .

## 2. Preliminaries on quantum groups

In this section we recall some basic definitions concerning quantum groups and their actions on -algebras. In addition we review the definition of the Drinfeld double and the description of its actions in terms of the underlying quantum group and its dual. For more information we refer to [BSUM], [Kustermansuniversal], [KVLCQG], [Vaesimprimitivity], [Woronowiczleshouches]. Our notation and conventions will mainly follow [NVpoincare], [Voigtbcfo].
Let be a normal, semifinite and faithful weight on a von Neumann algebra . We use the standard notation

 M+ϕ={x∈M+|ϕ(x)<∞},Nϕ={x∈M|ϕ(x∗x)<∞}

and write for the space of positive normal linear functionals on . If is a normal unital -homomorphism, the weight is called left invariant with respect to provided

 ϕ((ω⊗id)Δ(x))=ϕ(x)ω(1)

for all and . Similarly one defines right invariant weights.

###### Definition 2.1.

A locally compact quantum group is given by a von Neumann algebra together with a normal unital -homomorphism satisfying the coassociativity relation

 (Δ⊗id)Δ=(id⊗Δ)Δ

and normal semifinite faithful weights and on which are left and right invariant, respectively.

If is a locally compact group, then the algebra of essentially bounded measurable functions on together with the comultiplication given by

 Δ(f)(s,t)=f(st)

defines a locally compact quantum group. The weights and are given in this case by left and right Haar measures on , respectively. Of course, for a general locally compact quantum group the notation is purely formal.
An important tool in the study of locally compact quantum groups are multiplicative unitaries. If is a GNS-construction for the weight , then the multiplicative unitary is the operator on given by

 W∗(Λ(f)⊗Λ(g))=(Λ⊗Λ)(Δ(g)(f⊗1))

for all . This unitary satisfies the pentagonal equation and one can recover the von Neumann algebra as the strong closure of the algebra where denotes the space of normal linear functionals on . Moreover one has

 Δ(f)=W∗(1⊗f)W

for all . Let us remark that we will only consider quantum groups for which is a separable Hilbert space.
The group-von Neumann algebra of the quantum group is the strong closure of the algebra with the comultiplication given by

 ^Δ(x)=^W∗(1⊗x)^W

where and is the flip map. It defines a locally compact quantum group which is called the dual of . The left invariant weight for the dual quantum group has a GNS-construction , and according to our conventions we have .
The reduced -algebra of functions on the quantum group is

 Cr0(G)=[(id⊗L(HG)∗)(W)]⊂L∞(G),

and the reduced group -algebra of is

 C∗r(G)=[(L(HG)∗⊗id)(W)]⊂L(G).

Moreover we have . Together with the comultiplications inherited from and , respectively, the -algebras and are Hopf--algebras in the following sense.

###### Definition 2.2.

A Hopf -algebra is a -algebra together with an injective nondegenerate -homomorphism such that the diagram

You are adding the first comment!
How to quickly get a good reply:
• Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
• Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
• Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters