Contents lists available at ScienceDirect

Journal of Functional Analysis

www.elsevier.com/locate/jfa

The spatial Rokhlin property for actions of compact quantum groups

SelQuk Barlaka, Gabor Szabo b'*, Christian Voigtc

a Department of Mathematics and Computer Science, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark

b Fraser Noble Building, Institute of Mathematics, University of Aberdeen, Aberdeen AB24 3UE, United Kingdom

c School of Mathematics and Statistics, University of Glasgow, 15 University Gardens, Glasgow G12 8QW, United Kingdom

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 June 2016 Accepted 29 September 2016 Available online xxxx Communicated by Stefaan Vaes

primary 46L55, 20G42 secondary 46L35

Keywords: Rokhlin property Quantum groups Classification of C*-algebras Classification of dynamical systems

We introduce the spatial Rokhlin property for actions of coexact compact quantum groups on C*-algebras, generalizing the Rokhlin property for both actions of classical compact groups and finite quantum groups. Two key ingredients in our approach are the concept of sequentially split *-homomor-phisms, and the use of braided tensor products instead of ordinary tensor products.

We show that various structure results carry over from the classical theory to this more general setting. In particular, we show that a number of C*-algebraic properties relevant to the classification program pass from the underlying C*-algebra of a Rokhlin action to both the crossed product and the fixed point algebra. Towards establishing a classification theory, we show that Rokhlin actions exhibit a rigidity property with respect to approximate unitary equivalence. Regarding duality theory, we introduce the notion of spatial approximate representability for actions of discrete quantum groups. The spatial Rokhlin property for actions of a coexact compact quantum group is shown to be dual to spatial approximate

* Corresponding author.

E-mail addresses: barlak@imada.sdu.dk (S. Barlak), gabor.szabo@abdn.ac.uk (G. Szabo), christian.voigt@glasgow.ac.uk (C. Voigt).

http://dx.doi.org/10.1016/j.jfa.2016.09.023

0022-1236/© 2016 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

2 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

representability for actions of its dual discrete quantum group, and vice versa.

© 2016 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Contents

0. Introduction..............................................................................................................2

1. Preliminaries..............................................................................................................5

1.1. Quantum groups ..............................................................................................5

1.2. Actions, crossed products and Takesaki—Takai duality............................................11

1.3. Braided tensor products....................................................................................15

2. Induced actions on sequence algebras............................................................................16

2.1. Sequence algebras..............................................................................................16

2.2. Induced actions — discrete case............................................................................18

2.3. Induced actions — compact case..........................................................................19

3. Equivariantly sequentially split *-homomorphisms..........................................................25

4. The Rokhlin property and approximate representability..................................................29

4.1. The spatial Rokhlin property..............................................................................29

4.2. Spatial approximate representability....................................................................33

4.3. Duality............................................................................................................36

5. Rigidity of Rokhlin actions..........................................................................................39

6. Examples ..................................................................................................................48

References..........................................................................................................................51

0. Introduction

The Rokhlin property for finite group actions on unital C*-algebras was introduced and studied by Izumi in [15,16], building on earlier work of Herman-Jones [12] and Herman-Ocneanu [13]. Since the very beginning it has proven to be a useful tool in the theory of finite group actions. The Rokhlin property was subsequently generalized by Hirshberg-Winter [14] to the case of compact groups, and studied further by Gardella [9]; see also [11,10]. For finite quantum group actions, Kodaka-Teruya introduced and studied the Rokhlin property and approximate representability in [21].

The established theory, which we shall now briefly summarize in the initial setting of finite group actions, has three particularly remarkable features: The first is a multitude of permanence properties; it is known that many C*-algebraic properties pass from the coefficient C*-algebra to the crossed product and fixed point algebra. This was in part addressed by Izumi in [15], and studied more in depth by Osaka-Phillips [27] and Santiago [30]. The second feature is rigidity with respect to approximate unitary equivalence; a result of Izumi [15] asserts that two Rokhlin actions of a finite group on a separable, unital C*-algebra are conjugate via an approximately inner automorphism if and only if the two actions are pointwise approximately unitarily equivalent; see [11] for the non-unital case and [10] for the case of compact groups. As demonstrated by Izumi in [15], this rigidity property is useful for classifying Rokhlin actions on classifiable C*-algebras

via ^-theoretic invariants. The third feature is duality theory; a result of Izumi [15] shows that an action of a finite abelian group on a separable, unital C*-algebra has the Rokhlin property if and only if the dual action is approximately representable, and vice versa. This has been generalized to the non-unital case by Nawata [25], and to actions of compact abelian groups by the first two authors [4]; see also [8].

In the present paper, we introduce and study the spatial Rokhlin property for actions of coexact compact quantum groups, generalizing and unifying the work mentioned above. In particular, we carry over various structure results from the classical to the general case. Firstly, this allows us to remove all commutativity assumptions in the study of duality properties for Rokhlin actions. This is relevant even for classical group actions. Indeed, the Pontrjagin dual of a nonabelian group is no longer a group, but can be viewed as a quantum group. Accordingly, a natural way to fully incorporate nonabelian groups into the picture is to work in the setting of quantum groups from the very beginning. Secondly, it turns out that some results can be given quite short and transparent proofs in this more abstract setup, simpler than in previous accounts. Finally, our results are also of interest from the point of view of quantum group theory. Indeed, they provide examples of quantum group actions that either allow for classification, or the systematical analysis of structural properties of crossed product C*-algebras, in particular whether they fall within the scope of the Elliott program.

Let us highlight two comparably new ingredients in our approach. The first is the notion of (equivariantly) sequentially split *-homomorphisms introduced by the first two authors in [4]. It has already been demonstrated in [4] that many structural results related to the Rokhlin property can be recast and conceptually proved in the language of sequentially split *-homomorphisms, and some new ones could be proved as well. The second ingredient is a purely quantum feature, namely the braided tensor product construction. This provides the correct substitute for tensor product actions in the classical theory, and it also gives a conceptual explanation of the fact that the central sequence algebra is no longer the right tool in the quantum setting. Being widely known in the algebraic theory of quantum groups, braided tensor products in the operator algebraic framework were first introduced and studied in [26].

As already indicated above, for most of the paper we will assume that our quantum groups satisfy exactness/coexactness assumptions. This may appear surprising at first sight; it is essentially due to the fact that we have chosen to work in a reduced setting, that is, with reduced crossed products and minimal (braided) tensor products. We shall indicate at several points in the main text where precisely exactness enters. Our setup yields the strongest versions of conceivable definitions of Rokhlin actions and approximately representable actions, however, at the same time our examples are restricted to the amenable/coamenable case. On the other hand, the reduced setting matches best with the existing literature on quantum group actions, see for instance [2,3,31,26] and references therein. The necessary modifications to set up a full version of our theory are mainly of technical nature; we have refrained from carrying this out here.

4 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Let us now explain how the paper is organized. In Section 1, we gather some preliminaries and background on quantum groups, including a review of Takesaki-Takai duality and braided tensor products. Section 2 deals with induced actions of discrete and compact quantum groups on sequence algebras. Already for classical compact groups these actions typically fail to be continuous, and in the quantum setting this leads to a number of subtle issues. In Section 3, we define and study equivariantly sequentially split *-homomorphisms. We show that, as in the case of group actions, this notion behaves well with respect to crossed products and fixed point algebras. We also establish a general duality result for equivariantly sequentially split *-homomorphisms. In Section 4, we introduce the spatial Rokhlin property for actions of coexact compact quantum groups, and spatial approximate representability for actions of exact discrete quantum groups. We verify that various C*-algebraic properties pass to crossed products and fixed point algebras. Moreover, we show that the spatial Rokhlin property and spatial approximate representability are dual to each other. In Section 5, we present some steps towards a classification theory for actions with the spatial Rokhlin property. Among other things, we prove that two such G-actions on a C*-algebra A are conjugate via an approximately inner automorphism if and only if the actions are approximately unitarily equivalent as *-homomorphisms from A to Cr(G) < A. This generalizes a number of previous such classification results, in particular those of Izumi [15], Gardella-Santiago [10] and Kodaka-Teruya [21]. In this section, we also generalize a ^-theory formula for the fixed-point algebra of a Rokhlin action, first proved for certain finite group cases by Izumi [15] and for compact group actions by the first two authors in [4]. Finally, in Section 6 we present some examples of Rokhlin actions. In particular, we show that any coamenable compact quantum group admits an essentially unique action with the spatial Rokhlin property on O2.

The work presented here was initiated while the first two authors participated in the conference CSTAR at the University of Glasgow in September 2014. Substantial parts of this work were carried out during research visits of the authors to Oberwolfach in June

2015, of the first two authors at the Mittag-Leffler Institute from January to March

2016, and of the third author to the University of Southern Denmark in April 2016. The authors are grateful to all these institutions for their hospitality and support.

The first author was supported by GIF Grant 1137-30.6/2011, ERC AdG 267079, SFB 878 'Groups, Geometry, and Actions' and the Villum Fonden project grant 'Local and global structures of groups and their algebras' (2014-2018), 7423. The second author was supported by SFB 878 'Groups, Geometry, and Actions' and the Engineering and Physical Sciences Research Council Grant EP/N00874X/1 'Regularity and dimension for C*-algebras'. The third author was supported by the Engineering and Physical Sciences Research Council Grant EP/L013916/1 and the Polish National Science Centre grant no. 2012/06/M/ST1/00169.

The authors would like to thank the referee for a thorough reading of the manuscript and for helpful suggestions.

1. Preliminaries

In this preliminary section we collect some definitions and results from the theory of quantum groups and fix our notation. We will mainly follow the conventions in [26] as far as general quantum group theory is concerned. For more detailed information and background we refer to [34,24,22,23].

Let us make some general remarks on the notation used throughout the paper. We write L(E) for the space of adjointable operators on a Hilbert A-module, and K(E) denotes the space of compact operators. The closed linear span of a subset X of a Banach space is denoted by [X]. If x, y are elements of a Banach space and £ > 0 we write x =£ y if ||x — y|| < £.

Depending on the context, the symbol < denotes either the tensor product of Hilbert spaces, the minimal tensor product of C*-algebras, or the tensor product of von Neumann algebras. We write © for algebraic tensor products.

If A and B are C*-algebras then the flip map A < B ^ B < A is denoted by a. That is, we have a(a < b) = b < a for a G A, b G B.

If H is a Hilbert space we write £ G L(H < H) for the flip map £(£ < rf) = V < For operators on multiple tensor products we use the leg numbering notation. For instance, if W G L(H < H) is an operator on H<H, then W12 = W < id G L(H <H<H) and W23 = id <W. Moreover, W13 = S12W23S12.

If B is a C*-algebra we write B for the smallest unitarization of B.

1.1. Quantum groups

Although we will only be interested in compact and discrete quantum groups, let us first recall a few definitions and facts regarding general locally compact quantum groups.

Let p be a normal, semifinite and faithful weight on a von Neumann algebra M. We use the standard notation

M+ = {x G M+ | p(x) < N = {x G M | p(x*x) < <»}

and write M+ for the space of positive normal linear functionals on M. Assume that A : M ^ M < M is a normal unital *-homomorphism. The weight p is called left invariant with respect to A if

p((w < id)A(x)) = p(x)w(1)

for all x G M+ and w G M+. Similarly one defines the notion of a right invariant weight.

Definition 1.1. A locally compact quantum group G is given by a von Neumann algebra L^(G) together with a normal unital *-homomorphism A : LTO(G) ^ LTO(G) < L^(G) satisfying the coassociativity relation

6 S. Barlak et al. / Journal of Functional Analysis ••• (••••) ••• — •••

(A < id)A = (id <A)A

and normal semifinite faithful weights y and ^ on LTO(G) which are left and right invariant, respectively. The weights y and ^ will also be referred to as Haar weights of G.

Remark 1.2. Our notation should help to make clear how ordinary locally compact groups can be viewed as quantum groups. Indeed, if G is a locally compact group, then the algebra LTO(G) of essentially bounded measurable functions on G together with the comultiplication A : L~(G) ^ L~(G) < L~(G) given by

A(f )(s,t) = f (st)

defines a locally compact quantum group. The weights y and ^ are given in this case by left and right Haar measures, respectively. Of course, for a general locally compact quantum group G the notation LTO(G) is purely formal. Similar remarks apply to the C*-algebras Cf*(G), C*(G) and Cj (G), Cr0(G) associated to G that we discuss below. It is convenient to view all of them as different appearances of the quantum group G.

Remark 1.3 (cf. [22]). Let G be a locally compact quantum group and let A : ^ L2(G) be the GNS-construction for the Haar weight y. Throughout the paper we will only consider second countable quantum groups, that is, quantum groups for which L2(G) is a separable Hilbert space.

One obtains a unitary WG = W on L2(G) < L2(G) by

W*(A(x) < A(y)) = (A < A)(A(y)(x < 1))

for all x, y € Nv. This unitary is multiplicative, which means that W satisfies the pentagonal equation

W12W13 W23 = W23W12.

From W one can recover the von Neumann algebra LTO(G) as the strong closure of the algebra (id<L(L2(G))*)(W), where L(L2(G))* denotes the space of normal linear functionals on L(L2(G)). Moreover one has

A(x) = W*(1 < x)W

for all x € M. The algebra LTO(G) has an antipode, which is an unbounded, a-strong* closed linear map S given by S(id®w)(W) = (id®w)(W*) for <x € L(L2(G))*. Moreover, there is a polar decomposition S = Rr_i/2 where R is an antiautomorphism of LTO(G) called the unitary antipode and (rt) is a strongly continuous one-parameter group of

automorphisms of LX(G) called the scaling group. The unitary antipode satisfies a o (R < R) o A = A o R.

The group-von Neumann algebra L(G) of the quantum group G is the strong closure of the algebra (L(L2(G))* < id)(W) with the comultiplication A : L(G) ^ L(G) < L(G) given by

A(y) = W *(i < y)W

where W = £W*£ and £ G L(L2(G) < L2(G)) is the flip map. It defines a locally compact quantum group G which is called the dual of G. The left invariant weight p for the dual quantum group has a GNS-construction A : N^ ^ L2(G), and according to our conventions we have L(G) = LTO(G).

Remark 1.4. Since we are following the conventions of Kustermans and Vaes [22], there is a flip map built into the definition of A. As we will see below, this is a natural choice when working with Yetter-Drinfeld actions; however, it is slightly inconvenient when it comes to Takesaki-Takai duality. We will write G for the quantum group corresponding to L(G)cop. That is, LX(G) is the von Neumann algebra L(G) equipped with the opposite comultiplication A = Acop = a o A, where a denotes the flip map. By slight abuse of terminology, we shall refer to both G and G as the dual of G. According to Pontrjagin duality, the double dual of G in either of these conventions is canonically isomorphic to G.

Remark 1.5 (cf. [22]). The modular conjugations of the left Haar weights p and p are denoted by J and J, respectively. These operators implement the unitary antipodes in the sense that

R(x) = Jx*J, R(y) = Jy* J

for x G L~(G) and y G L(G). Note that L~(G)' = JL'X(G)J and L(G)' = JL(G)J for the commutants of LTO(G) and L(G). Using J and J one obtains multiplicative unitaries

V = (J < J )W (J < J), V = (J < J )W (J < J)

which satisfy V G L(G)' < L~(G) and V G L~(G)' < L(G), respectively. We have

A(f ) = V(f < 1)V*, A(x) = V(x < 1)t>*

for f G LTO(G) and x G L(G). We also record the formula

(J < J)W (J < J) = W*, which is equivalent to saying (R < R)(W) = W.

8 S. Barlak et al. / Journal of Functional Analysis ••• (••••')

We will mainly work with the C*-algebras associated to the locally compact quantum group G. The algebra [(id<L(L2(G))*)(W)] is a strongly dense C*-subalgebra of LTO(G) which we denote by C0(G). Dually, the algebra [(L(L2(G))*<id)(W)] is a strongly dense C*-subalgebra of L(G) which we denote by Cr* (G). These algebras are the reduced algebra of continuous functions vanishing at infinity on G and the reduced group C* -algebra of G, respectively. One has W € M(Cr0(G) << Cr*(G)).

Restriction of the comultiplications on LTO(G) and L(G) turns C0(G) and Cr*(G) into Hopf C* -algebras in the following sense.

Definition 1.6. A Hopf C*-algebra is a C*-algebra H together with an injective nonde-generate *-homomorphism A : H ^ M(H < H) such that the diagram

M (H <g> H )

M (H <g> H )

id ® A

M (H <g> H <g> H )

is commutative and [A(H)(1 <g> H)] = H <g> H = [(H <g> 1)A(H)

Definition 1.7. A compact quantum group is a locally compact quantum group G such that C0 (G) is unital. Similarly, a discrete quantum group is a locally compact quantum group G such that Cr*(G) is unital.

We will write C'(G) instead of C0 (G) if G is a compact quantum group. A finite quantum group is a compact quantum group G such that C'(G) is finite dimensional. This is the case if and only if C* (G) is finite dimensional.

Definition 1.8. A unitary corepresentation of a Hopf C*-algebra H on a Hilbert B-module E is a unitary U G M(H <g> K(E)) = L(H <g> E) such that

(A < id)(^) = U13U23.

A unitary representation of a locally compact quantum group G is a unitary corepresen-tation of C0 (G).

Remark 1.9. This terminology is compatible with the classical case, more precisely, for a classical locally compact group G there is a bijective correspondence between unitary corepresentations of C0(G) on Hilbert spaces and strongly continuous unitary representations of G in the usual sense, see [24, Section 5] and [31, 5.2.5].

Example 1.10. The (left) regular representation of a locally compact quantum group G is the representation of G on the Hilbert space L2(G) given by the multiplicative unitary

W G M(C0(G)gK(L2(G))), see Remark 1.3. In fact, the relation (A(id)(W) = W13W23 is equivalent to the pentagon equation for W.

Remark 1.11. Let G be a locally compact quantum group. The full group C*-algebra of G is a Hopf C*-algebra C*(G) together with a unitary representation W G M(C0(G) ( Cf* (G)) satisfying the following universal property: for every unitary representation U G M(C0 (G) ( K(E)) of G there exists a unique nondegenerate *-homomorphism n : Cf*(G) ^ L(E) such that (id (n)(W) = U.

Similarly, one obtains the full C*-algebra C0(G) of functions on G.

Definition 1.12. Let G be a locally compact quantum group. We say that G is amenable if the canonical quotient map C*(G) ^ Cr*(G) is an isomorphism. Similarly, G is called coamenable if the dual G is amenable.

If G is coamenable we will simply write C0(G) for C0(G). By slight abuse of notation, we will also write C0(G) if a statement holds for both C0(G) and C0(G). In particular, if G is compact and coamenable we will simply write C(G) instead Cr(G). Similarly, if G is amenable we will write C*(G) for Cr*(G). We remark that every compact quantum group is amenable, and equivalently every discrete quantum group is coamenable.

Remark 1.13 (cf. [34,24]). Let G be a compact quantum group. In analogy with the theory for compact groups, every unitary representation of G is completely reducible, and all irreducible representations are finite dimensional. We write Irr(G) for the set of equivalence classes of irreducible unitary representations of G. Our general second countability assumption amounts to saying that the set Irr(G) is countable.

A unitary representation of G on a finite dimensional Hilbert space is given by a unitary ux G Cr(G) (gK(Ha), so it can be viewed as an element ux = (uj) G Mn(Cr(G)) if dim(H*) = n. Moreover, the corepresentation identity translates into the formula

ij ) = n uik ®

for the comultiplication of the matrix coefficients ui.

By Peter-Weyl theory, the linear span of all matrix coefficients ui for A G Irr(G) forms a dense *-subalgebra O(G) C C(G). In fact, together with the counit e : O(G) ^ C given by

e(uij) = sij

and the antipode S : O(G) ^ O(G) given by

S(uj) = (u* )*

the algebra O(G) becomes a Hopf *-algebra.

10 S. Barlak et al. / Journal of Functional Analysis

We will use the Sweedler notation A(f) = f(1) < f(2) for the comultiplication of general elements of O(G). This is useful for bookkeeping of coproducts, let us emphasize however that this notation is not meant to say that A(f) is a simple tensor. For higher coproducts one introduces further indices, for instance f(1) <f(2) <f(3) is an abbreviation for (A < id)A(f) = (id<A)A(f).

Again by Peter-Weyl theory, one obtains a vector space basis of O(G) consisting of the matrix coefficients uj where A ranges over Irr(G) and 1 < i, j < dim(A). Here we abbreviate dim(A) = dim(Hx). If we write O(G)x for the linear span of the elements uj for 1 < i, j < dim(A), we have a direct sum decomposition

O(G) = alg- 0 O(G)x.

xeirr(G)

Note that the coproduct of O(G) takes a particularly simple form in this picture; from the above formula for A(uj) we see that it looks like the transpose of matrix multiplication.

Let ux be an irreducible unitary representation of G, and let uj be the corresponding matrix elements in some fixed basis. The contragredient representation ux is given by the matrix (ux )ij = S(uxi) where S is the antipode of O(G). In general ux is not unitary, but as any finite dimensional representation of G, it is unitarizable. The representations ux and ux°c are equivalent, and there exists a unique positive invertible intertwiner Fx : Hx ^ HXcc satisfying tr(Fx) = tr(F-1). The trace of Fx is called the quantum dimension of Hx and denoted by dimg(A).

With this notation, the Schur orthogonality relations are

¥(uxj(uki)*) = SxnSik dim^(A) (Fx)ij

where A, n € Irr(G) and <p : Cr(G) ^ C is the Haar state of G. In the sequel we shall fix bases such that Fx is a diagonal operator for all A € Irr(G).

Remark 1.14. Let again G be a compact quantum group. While the comultiplication for the C*-algebra Cr(G) looks particularly simple in terms of matrix coefficients, dually the multiplication in the C*-algebra C*(G) is easy to describe. More precisely, we have

C *(G)= 0 K(Hx),

xelrr(G)

where the right hand side denotes the c0-direct sum of the matrix algebras K(Hx). If uj are the matrix coefficients in O(G), then the dual basis vectors uj, that is, the linear functionals on O(G) given by

uijj (ukl) = Sxn Sik Sjl

form naturally a vector space basis of matrix units for the algebraic direct sum

D(G)=alg- 0 K (Hi)

Aeirr(G)

inside Cr*(G).

Let us note that D(G) should not be confused with a quantum double, but there will be no conflicting notation in this regard appearing in this paper.

Let us also note that according to the Schur orthogonality relations, see Remark 1.13, the functionals wi extend continuously to bounded linear functionals on Cr(G).

1.2. Actions, crossed products and Takesaki-Takai duality

Let us now consider actions on C*-algebras.

Definition 1.15. A (left) coaction of a Hopf C*-algebra H on a C*-algebra A is an injective nondegenerate *-homomorphism a : A ^ M (H < A) such that the diagram

M (H < A)

, , id CO)a , ,

M (H < A) -> M (H < H < A)

is commutative and [a(A)(H ( 1)] = H <g A.

If (A, a) and (B, ¡3) are C*-algebras equipped with coactions of H, then a *-homo-morphism p : A ^ M(B) is called H-colinear if 3<p = (id®p)a.

An action of a locally compact quantum group G on a C*-algebra A is a coaction a : A ^ M(C0(G) (g A) of C0(G) on A. We will also say that (A, a) is a G-C*-algebra in this case. A *-homomorphism <p : A ^ M(B) between G-C*-algebras is called G-equivariant if it is C0 (G)-colinear.

If H is a Hopf C*-algebra and a : A ^ M(H ( A) a coaction, then the density condition [a(A)(H ( 1)] = H ( A implies in particular that the image of a is contained in the H-relative multiplier algebra of H ( A, defined by

MH(H ( A) = {m G M(H ( A) | m(H ( 1), (H ( 1)m C H ( A} .

In contrast to the situation for ordinary multiplier algebras, the tensor product map id (g>p : H ( A ^ H ( B of a (possibly degenerate) *-homomorphism <p : A ^ B admits a unique extension MH(H( A) ^ MH(H ( B) to the relative multiplier algebra, which will again be denoted by id (p. If p is injective, then this also holds for id (g>p : MH(H (g A) ^ MH(H (g B). We refer to [6, Appendix A] for further details on relative multiplier algebras.

12 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Remark 1.16. For the construction of examples of Rokhlin actions we shall consider inductive limit actions of Hopf C*-algebras. Assume H is a Hopf C*-algebra and that A1 ^ A2 ^ • •• is an inductive system of H-C*-algebras with coactions aj : Aj ^ MH(H < Aj) and injective equivariant connecting maps. Then the direct limit A = lim Aj becomes an H-C*-algebra in a canonical way. Firstly, we have lim MH (H < Aj) C MH (H < A) naturally since our system (Aj has injective connecting maps. The maps aj define a compatible family of *-homomorphisms aj : Aj ^ MH (H < Aj) ^ MH(H < A), and induce a *-homomorphism a : A ^ MH(H < A). Coassociativity of a and the density condition [(H < 1)a(A)] = H < A follow from the corresponding properties of the coactions aj. Injectivity of the map a : A ^ MH (H < A) follows from injectivity of the maps aj and of the connecting maps in the system H < Aj.

Remark 1.17 (cf. [28]). Let G be a compact quantum group. Further below we will use the fact that any G-C*-algebra A admits a spectral decomposition. In order to discuss this we review some further definitions and results.

Let G be a compact quantum group and let (A, a) be a G-C*-algebra. Since G is compact, the coaction is an injective *-homomorphism a : A ^ Cr(G) << A satisfying the coassociativity identity (A < id) o a = (id <a) o a and the density condition [(Cr(G) < 1)a(A)] = Cr(G) < A. For A € Irr(G) we let

Ax = {a € A | a(a) € O(G)x © A}

be the A-spectral subspace of A. Here we recall that O(G)x C O(G) denotes the span of the matrix coefficients for Hx, see Remark 1.13.

The subspace Ax is closed in A, and there is a projection operator px : A ^ Ax defined by

px(a) = (6x < id)a(a)

dim(x)

0x(x) = dim,(A) £ (Fx)-j1^(x(uxj)*). j=1

By definition, the spectral subalgebra S(A) C A is the *-subalgebra

S (A) = SG(A) = alg- 0 Ax.

xelrr(G)

From the Schur orthogonality relations and [(Cr(G) << 1)a(A)] = Cr(G) << A it is easy to check that S(A) is dense in A.

For u € D(G) and a € A let us define

a • u = (u < id) o a(a).

Then the Schur orthogonality relations imply that A •D(G), the linear span of all elements a • u as above, is equal to S(A). Moreover, from the coaction property of a it follows that A becomes a right D(G)-module in this way.

Definition 1.18. Let G be a locally compact quantum group and let A be a G-C* algebra with coaction a : A ^ M(CQ(G) < A). The reduced crossed product G Kar A is the C*-algebra

G ka r a = [(CQ(G) < 1)a(A)] C M(K(L2(G)) < A).

Recall here that CQ (G) = Cr*(G) as a C*-algebra, but equipped with the opposite comultiplication.

The reduced crossed product is equipped with a canonical dual action of G, which turns it into a G-C*-algebra. More precisely, the dual action is given by comultiplication on the copy of CQ(G) and the trivial action on the copy of A inside M(G Ka r A). By our second countability assumption, the crossed product G Ka r A is separable provided A is separable.

For the purpose of reduced duality, we have to restrict ourselves to regular locally compact quantum groups [2]. All compact and discrete quantum groups are regular, so this is not an obstacle for the constructions we are interested in further below.

Remark 1.19. If G is a regular locally compact quantum group then the regular representation of L2(G) induces an action of G on the algebra K(L2(G)) of compact operators by conjugation. More generally, if A is any G-C*-algebra we can turn the tensor product K(L2(G)) << A into a G-C*-algebra by equipping K(L2(G)) << A ^ K(L2(G) << A) with the conjugation action arising from the tensor product representation of G on the Hilbert A-module E = L2 (G) < A. Explicitly, following the notation in [26], we consider

A(£ < a) = X*2S12(id ®a)(£ < a)

where X = XV£ and V is as in Remark 1.5. This determines a coaction A : E ^ M(CQ (G) < E), which in turn corresponds to a unitary corepresentation

V* € L(CQ(G) <E) = M(CQ(G) < K(E)).

The conjugation action Adx : K(E) ^ M(CQ(G) <<K(E)) = L(CQ(G) <<E) is then defined by

Ad x(T ) = Vx (1 < T )V*,

and K(E) becomes a G-C*-algebra in this way. Under the isomorphism K(L2(G)) << A = K(L2(G) < A), this G-action is given by

14 S. Barlak et al. / Journal of Functional Analysis ••• (••••) ••• — •••

aK(T g a)=Adx(T g a) = X*2(1 g T g 1)a(a)^Xi2, T G K(L2(G)),a G A.

If G is a classical group then the resulting action is nothing but the diagonal action on the tensor product K(L2(G)) ( A. For further details we refer to [1,26].

Let us now state the Takesaki-Takai duality theorem for regular locally compact quantum groups [2], see [31, chapter 9] for a detailed exposition.

Theorem 1.20. Let G be a regular locally compact quantum group and let A be a G-C*-algebra. Then there is a natural isomorphism

(G k^ G Kft|[ A, a) = (K(L2(G)) ( A,aK)

of G-C*-algebras.

Let us give a brief sketch of the proof of Theorem 1.20 for the sake of convenience. Recall from Remark 1.5 that J is the modular conjugation of the left Haar weight of (G), and similarly J the modular conjugation of the left Haar weight of L(G). We shall write U = J J. By definition, we have

G Ka,r G Ka,[ A = [(UC0(G)U ( 1 ( 1)(Acop(Cr*(G)) ( 1)a(A)23].

Conjugating by W*2 and using that C0(G) and JC0(G)J commute, this is isomorphic to

[(UC0(G)U ( 1 (g 1)(Cr*(G)) ( 1 (g 1)(id(a)a(A)] = [(UC^(G)UC*(G)) ( 1)a(A)]

since a is injective. Moreover, [UC^(G)UCr*(G)] identifies with K(L2(G)). Hence the right hand side is isomorphic to K(L2(G)) g> A, taking into account that

[(K(L2(G)) g 1)a(A)] = [(K(L2(G))C0(G) g 1)a(A)] = [K(L2(G))C0(G) g A]

by the density condition [(C0(G) g 1)a(A)] = C0(G) g A.

Let us also identify the bidual coaction. By construction, the bidual coaction a maps UfU g 1 G K(L2(G)) g A for f G C0 (G) to (1 g U )A(f )(1 g U) and leaves Cr*(G) g 1 and a(A) fixed. Using the relations from Remark 1.5 one can show

a(T g a) = X*2(1 g T g 1)a(a) 13X12,

where X = (1 g U)W(1 g U).

At some points we will also need the full crossed product G Ka,f A of a G-C * -algebra (A, a); we refer to [26] for a review of its definition in terms of its universal property for covariant representations.

1.3. Braided tensor products

Finally, let us discuss Yetter-Drinfeld actions and braided tensor products. We refer to [26] for more information.

Definition 1.21. Let G be a locally compact quantum group. A G-YD-C*-algebra is a C*-algebra A together with a pair of actions a : A ^ M(CQ (G) < A) and 7 : A ^ M(Cr*(G) < A) satisfying the Yetter-Drinfeld compatibility condition

(a < id) o (id <a) o 7 = (Ad( W) < id) o (id <7) o a.

Here W € M(CQ(G) < Cr*(G)) is the multiplicative unitary.

The braided tensor product, which we review next, generalizes at the same time the minimal tensor product of C*-algebras and reduced crossed products. Roughly speaking, it allows one to construct a new C*-algebra out of two constituent C*-algebras, together with some prescribed commutation relations between the two factors.

Definition 1.22. Let G be a locally compact quantum group. Given a G-YD-C*-algebra A and a G-C*-algebra (B, 3), one defines the braided tensor product A Kg B = A K B by

A K B = [y(A)12/(B)13] C M(K(L2(G)) < A < B).

The braided tensor product A K B becomes a G-C*-algebra with coaction a K /3 : A K B ^ M(CQ(G)< AKB) in such a way that the canonical embeddings iA : A ^ M(AKB) and : B ^ M(A K B) are G-equivariant. We write a K 1 and 1 K b for the images of elements a € A and b € B in M(A K B), respectively.

Example 1.23. A basic example of a G-YD-action is given by the C*-algebra A = CQ (G) for a regular locally compact quantum group G, equipped with a = A and the adjoint action Y(f) = W*(1 < f )W.

We will mainly be interested in the special case of the braided tensor product construction where the first factor is equal to CQ (G) with the Yetter-Drinfeld structure from Example 1.23.

Lemma 1.24. Let G be a regular locally compact quantum group. For any G-C*-algebra A, there exists a G-equivariant *-isomorphism

Ta : (CQ(G) K A, A K a) ^ (CQ(G) < A, A < id)

such that Ta(1 K a) = a(a) for all a € A C M(CQ(G) K A) and Ta(f K 1) = f < 1 for f € CQ(G) C M(CQ(G) K A).

16 S. Barlak et al. / Journal of Functional Analysis ••• (••••) ••• — •••

Proof. The map T is obtained from the identifications

C0(G) H A = [W*2(1 g C0(G) g 1)W12a(A)13]

- [(1 g C0 (G) g 1) W12 a (A) 13 W* ] = [(C0(G) g 1 g 1)W*2a(A)23W12] = [(C0(G) g 1 g 1)(id ga)a(A)]

- [(C0(G) g 1)a(A)] = C0(G) g A,

where we use W = EW* E in the third step and the fact that a is injective in the penultimate step.

It is straightforward to check that Ta(a) = a(a) for a G A C M(C0(G) H A) and Ta(f) = f g 1 for f G C0(G) C M(C0(G) H A).

In particular, the canonical action of G on C0(G) H A corresponds to the translation action on the first tensor factor in C0 (G) g A under this isomorphism. □

2. Induced actions on sequence algebras

The theory of Rokhlin actions for compact quantum groups relies on the possibility of obtaining induced actions on the level of sequence algebras. In this section we shall recall a few facts on sequence algebras, and then discuss the construction of induced actions, separately for the case of discrete and compact quantum groups.

2.1. Sequence algebras

Let us first recall some notions related to sequence algebras, see [4] and [17]. If A is a C -algebra we write

£w(A) = {(a„)„ | an G A and sup ||a„|| <

for the C * -algebra of bounded sequences with coefficients in A. Moreover denote by

co(A) = {(an)n G £~(A) | lim ||aj = 0) C £w(A) the closed two-sided ideal of sequences converging to zero. Definition 2.1. Let A be a C * -algebra. The sequence algebra Aw of A is

= (A)/co(A).

Given a bounded sequence (an)neN G £w(A), the norm of the corresponding element in Aw is given by

||[(a„)„eN]|| = limsup ||a„||.

Note moreover that A embeds canonically into Aw as (representatives of) constant sequences. We will frequently use this identification of A inside Aw in the sequel.

Notation 2.2. We denote by

Annœ(A) = {x G Aœ | xa = ax = 0 for all a G A}

the two-sided annihilator of A inside Aœ. Moreover, we write

= [A • • A] C

for the hereditary subalgebra of Aw generated by A. Note that the embedding A — DWaa is clearly nondegenerate. Finally, consider also the normalizer of DWaa inside Aa

N(Dm,a,Aoo) = {x G | xD^,A + D^,ax C Dw,a} .

We remark that Annw(A) sits inside N(Dw, a, Aw) as a closed two-sided ideal.

The multiplier algebra of Dœ>A admits the following alternative description.

Proposition 2.3 (cf. [4, Proposition 1.5(1)] and [17, 1.9(4)]). Let A be a a-unital C *-algebra. The canonical *-homomorphism N(Dœ<A, Aœ) ^ M(DTOja) given by the universal property of the multiplier algebra is surjective, and its kernel coincides with AnnTO(A).

Let us also note that the construction of Dœ>A is compatible with tensoring by the compacts.

Lemma 2.4 (cf. [4, 1.6]). Let A be a C *-algebra and let K(H) be the algebra of compact operators on a separable Hilbert space H. The canonical embedding K(H) g Aw ^ (K(H) g A)w induces an isomorphism K(H) g DWia — Dto,k(H)®a.

Given a C * -algebra equipped with an action of a quantum group G, we shall now discuss how to obtain induced actions on the sequence algebras introduced above.

18 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

2.2. Induced actions - discrete case

In the case of discrete quantum groups the situation is relatively simple. In fact, if G is a discrete quantum group then the C*-algebra C0(G) of functions on G is a C*-direct sum of matrix algebras. Explicitly, it is of the form

Cq(G) ^ 0 K(H a)

where A = Irr(G) is the set of equivalence classes of irreducible representations of the dual compact quantum group, see Remark 1.13.

If (A, a) is a G-C*-algebra, then one of the defining conditions for the coaction a : A ^ M(C0(G) < A) is that it factorizes over the Co(G)-relative multiplier algebra MCo(g)(Co(G) < A). With the notation as above, we have

Mc0(G)(Co(G) < A) = H K(H ) < A,

that is, we can identify the relative multiplier algebra with the l™-product of the algebras K(HA) < A. In other words, we have

a : A ^ Yi K(H a) < A C M(Co(G) < A).

It follows that applying a componentwise induces a *-homomorphism a™ : l™(A) ^ M(C0(G) < l™(A)) by considering the composition

l™(A) ^ n„eN n AeA K(HA) < A

= nAeA n„eN K(Ha) < A

= nA eA K(Ha ) < n„eN A C M (Co(G) < l™(A)), here we use that K(H A) is finite dimensional for all A € A in the second isomorphism.

Lemma 2.5. Let G be a discrete quantum group and let (A, a) be a G-C*-algebra. Then the map a™ : l™(A) ^ M(C0(G) < l™(A)) constructed above turns l™(A) into a G-C*-algebra.

Proof. Injectivity and coassociativity of a™ follow immediately from the corresponding properties of a. For the density condition

[(Cq(G) < 1)a™(l™(A))] = Cq(G) < l™(A)

notice that it suffices to verify

[(pCo(G) <g> 1)a~(i~(A))] = pCo(G) <g> (A)

for all finite rank (central) projections p G C0 (G). This in turn follows from the density condition for a, combined with fact that tensoring with finite dimensional algebras commutes with taking direct products. □

The map constructed above induces an injective *-homomorphism : ^ M(C0(G) <g) ATO) that fits into the following commutative diagram

£x (A)

Mco{g)(C0(G] <g> £~(A))

Mco{g)(C0(G) <g> A«,)

Indeed, it suffices to observe that a™ maps c0(A) into ^Aeirr(G) K(HA) <8> c0(A). Coas-sociativity and the density conditions for a™ are inherited from a™. We therefore obtain the following result.

Lemma 2.6. Let G be a discrete quantum group and let (A, a) be a G-C*-algebra. Then the map a™ : A™ ^ M(Co(G) <g) A™) turns A™ into a G-C*-algebra.

2.3. Induced actions - compact case

Whereas for discrete quantum groups the extension of actions to sequence algebras always yields genuine actions, the situation for compact quantum groups is more subtle. Already classically, a strongly continuous action of a compact group G on a C*-algebra A induces an action on and Aœ, but these induced actions typically fail to be

strongly continuous, compare [4].

We shall address the corresponding problems in the quantum case by using an ad-hoc notion of equivariant *-homomorphisms into sequence algebras. Our discussion also requires the technical assumption of coexactness.

Definition 2.7. A locally compact quantum group G is exact if the functor of taking reduced crossed products by G is exact. We say that G is coexact if the dual G is exact.

It is well-known that a discrete quantum group G is exact if and only if C*(G) is an exact C*-algebra, see [33, 1.28]. In other words, a compact quantum group G is coexact if and only if Cr(G) is an exact C*-algebra.

Remark 2.8. Let A and B be C*-algebras. If B is exact, then there exists a canonical injective *-homomorphism B®5—> (B®A)œ coming from the following commutative diagram with exact rows

S. Barlak et al. / Journal of Functional Analysis

B < co(A)

B < iœ(A)

c0(B < A)-^ £œ(B < A)

(B < A)c

For similar reasons, we have a chain of natural inclusions B g M(DTO ,a) C M(B g , a) C M(Dto ,b^a) provided B is exact.

Now let G be a compact quantum group and let (A, a) be a G-C*-algebra. The *-homomorphisms a : A ^ Cr(G) g A induces a *-homomorphism

a~ : £™(A) ^ £^(Cr(G) g A),

obtained by applying a componentwise, and also a *-homomorphism

: ^ (Cr(G) g A)TO.

The maps and aTO, despite not being coactions in the sense of Definition 1.15 in general, turn out to be good enough to obtain a tractable notion of equivariance and suitable crossed products, at least when G is coexact. The reason for this is Remark 2.8, which is used in the definition below.

Definition 2.9. Let G be a coexact compact quantum group and let (A, a) and (B, /) be G-C * -algebras. A *-homomorphism p : A ^ is said to be G-equivariant if o p = (id gp) o a, where both sides are viewed as maps from A into (Cr(G) g B)TO. If p is G-equivariant, then we also write p : (A, a) ^ (BTO,

Note in particular that if p : A ^ is a G-equivariant *-homomorphism, then we automatically have o p(A) C Cr(G) g BTO.

Remark 2.10. As indicated above, if G is a compact group and a : G ^ A is a strongly continuous action on a C * -algebra, there always exists a (not necessarily strongly continuous) induced action of G on ATO. If (B, /) is another G-algebra, then it is easy to see that any *-homomorphism p : A ^ is G-equivariant in the sense 2.9 if and only if it is G-equivariant in the usual sense.

Indeed, the equality (id gp) o a = o p clearly implies equivariance the usual sense. For the converse implication, one notes that if p is equivariant in the usual sense, it maps A automatically into the continuous part of the action on BTO. Therefore, for a G A the equality of (id gp) o a(a) = o p(a) in C(G) g = C(G, BTO) can be checked by evaluating both sides at the points of G.

In the quantum setting, we need a substitute of the continuous part of an action in order to define crossed products. We shall rely on the structure of compact quantum groups to obtain a construction suitable for the situation at hand.

Recall from Remark 1.13 that the dense Hopf *-algebra O(G) C Cr(G) has a linear basis of elements of the form ui where A G Irr(G) and 1 < i, j, < dim(A). As explained in Remark 1.14, the linear functionals wi G Cr(G)* given by

wij = iSikSjl

span the space D(G), which can be viewed as the dense *-subalgebra of C0(G) given by the algebraic direct sum of matrix algebras K(Ha) for A G Irr(G). Moreover, in this picture the elements wi are matrix units in K(Ha), that is,

wij wh = i Sjkwil

Let (A, a) be a G-C * -algebra. Recall from Remark 1.17 that A becomes a right D(G)-module with the action

a ■ w = (w g id) o a(a).

It is crucial for our purposes that such module structures also exist on l«(A) and A«. Indeed, note that applying w g id in each component we obtain slice maps £«(C'(G) g A) ^ £«(A) and (Cr(G) g A)« ^ A«, which, by slight abuse of notation, will again be denoted by w g id in the sequel. We will also continue to use the notation a ■ w for the module structures obtained in this way.

In analogy with the constructions in Remark 1.17 we shall now define spectral subspaces of £«(A) and A«, and use this to define corresponding continuous parts. Although the settings differ somewhat, this is similar to Kishimoto's definition of equicontinuous sequences for flows, cf. [20,19].

Definition 2.11. Let G be a coexact compact quantum group and let (A, a) be a G-C * -algebra. The spectral subspaces of £«(A) and A« with respect to a are defined by

S(£«(A)) = £«(A) ■ D(G), S(A«) = A« ■ D(G), respectively. The continuous parts of £«(A) and A« with respect to a are defined by

goo,a(A) = [S(f«(A))] C £«(A), A^a = [S(A«,)] C A«,,

respectively.

At this point it is not immediately obvious that the subspaces in Definition 2.11 are closed under multiplication. We will show this further below.

Lemma 2.12. The canonical map S(£«(A)) ^ S(A«) is surjective.

22 S. Barlak et al. / Journal of Functional Analysis

Proof. Let x € S(ATO). Then we can write x = x • p = (pOid)aTO(x) for some finite rank idempotent p € D(G). If X € (A) is any lift of x, then X • p is a lift of x as well, which in addition is contained in S (£TO(A)). □

It follows from Lemma 2.12 that the canonical map £TO'a(A) ^ is surjective.

Proposition 2.13. Let G be a coexact compact quantum group and let (A, a) be a G-C*-algebra. Then we have

S(£~(A)) = {x € £~(A) | a~(x) € O(G) © £~(A)}; S(ATO) = {x € | aTO(x) €O(G) © ATO}.

Moreover, both £TO'a(A) and are G-C*-algebras in a canonical way.

Proof. Let us consider first the assertions for S(£TO(A)). By construction, for an element x € S(£~(A)) satisfying a~(x) € O(G) © £~(A) we can write

a~(x) = Y, O xj

\eF i,j

for some finite set F C Irr(G) and elements xj € £TO(A). By the definition of the counit e : O(G) ^ C, we see that applying w = £ aeF E^ * ] € D(G) C C'(G)* in the first tensor factor gives

x • w = (w O id) o (x) = (e O id) o aTO(x) = x.

This means {x € £~(A) | a~(x) € O(G) © £~(A)} C S(£~(A)).

Conversely, write x € S(£TO(A)) as a finite sum x = ^i ^ • wi of some elements wi € D(G) and y1 € £TO(A). We may assume without loss of generality that each w% is contained in K(H) for some Xi € Irr(G). Let F C Irr(G) be the finite subset consisting of all Xj. Writing x = (xn)„eN and y = (yn)neN this means that each xn = ^i yln • w% is contained in the subspace S(A)F = A • D(G)F C A • D(G), where D(G)F = i D(G)a i. Since D(G)f is finite dimensional, we conclude

a~(x) € O(G) © £~(A).

Here we use that the construction of £TO-products is compatible with tensoring by finite dimensional spaces. We conclude S(£TO(A)) C{x € £~(A) | a~(x) € O(G) © £~(A)}.

The corresponding assertion for S(ATO) is obtained in a similar way. According to Lemma 2.12, we know that x € S(ATO) is represented by an element x € S(£TO(A)), so the above argument shows aTO(x) € {z € | aTO(z) € O(G) © ATO} by construction of aTO. Conversely, if x satisfies aTO(x) € O(G) © then the same argument as in the case £TO(A) above shows x € S(ATO).

As a consequence of these considerations we obtain in particular that S(¿«(A)) and S(A«) are *-algebras, and hence ¿«,a(A) C ¿«(A) and A«,a C A« are C * -subalgebras.

It remains to show that these C * -algebras are G-C * -algebras in a canonical way, with coactions induced by a« and a«, respectively.

Let us again first consider the case ¿«,a(A). From coassociativity of a we obtain that a« maps S(¿«(A)) to O(G) ©S^«(A)). Therefore it induces a *-homomorphism ¿«a(A) ^ Cr(G) g ¿«a(A) C Cr(G) g ¿«(A), which we will again denote by a«. Injectivity of the latter map is clear. Similarly, the coaction identity (id ga«)a« = (Agid)a« follows immediately from the coaction identity for a. For the density condition note that we can write 1 g x = S(x(_2))x(_i) g x(0) for x G S^«(A)), using the Hopf algebra structure of O(G), and the Sweedler notation a«(x) = x(-1) g x(0). Hence

(O(G) © 1)a«(S(¿«(A))) = O(G) ©S^«(A)),

which implies [(Cr(G) g 1)a«(¿«a(A))] = Cr(G) g a«^«a(A)) upon taking closures.

The case A«,a is analogous. The considerations for S(¿«(A)) above and Lemma 2.12 imply that a« maps S(A«) to O(G) ©S(A«). Therefore it induces a *-homomorphism A«,a ^ Cr(G)gA«,a C Cr(G)gA«, which we will again denote by a«. Coassociativity and density conditions are inherited from the corresponding properties of the coaction a« : ¿«a(A) ^ Cr(G) g ¿«a(A). □

Using Proposition 2.13 we obtain an alternative way to describe the notion of equiv-ariance introduced in Definition 2.9

Proposition 2.14. Let G be a coexact compact quantum group and let (A, a) and (B, (3) be G-C *-algebras. For a *-homomorphism p : A ^ Bthe following are equivalent:

a) p : A ^ B« is G-equivariant;

b) p(a ■ w) = p(a) ■ w for all w G D(G) and a G A;

c) p(A) C B«,^ and p : (A, a) ^ (B«,^, («) is G-equivariant in the usual sense.

Proof. a) ^ b): Let w G D(G) and a G A, and recall D(G) C Cr(G)*. Applying w g id to both sides of the equality (id gp) o a(a) = (3« o p(a), we obtain

p(a ■ w) = (id gp)(w g id)a(a) = (w g id)(idgp)a(a) = (w g id)(« o p(a) = p(a) ■ w.

b) ^ c) : As p(a ■ w) = p(a) ■ w for all w G D(G) and a G A, it follows that p(S(A)) C S(B«). Hence, o p maps S(A) into O(G) © S(B«). For a G S(A) and w G D(G) we therefore compute

24 S. Barlak et al. / Journal of Functional Analysis ••• (••••')

(w O id)(TOy(a) = y(a) • w = y(a • w) = (w O id)(idOy)a(a).

It is now straightforward to check that o y(a) = (idOy) o a(a) for all a € S(A). As S(A) C A is dense, we conclude that y : A ^ is G-equivariant.

c) ^ a): This implication follows immediately from the definitions. □

We shall now define crossed products of induced actions on sequence algebras.

Definition 2.15. Let G be a coexact compact quantum group and let (A, a) be a G-C*-algebra. We define

G ^a^,' = G ^a^,,' Aто,a,

that is, G is defined to be the crossed product of the continuous part of

with respect to the coaction : ATO,a ^ C'(G) O ATO,a. In a similar way we define

G Ka-,r £~(A) = G Ka-,r £~a(A).

Remark 2.16. The notation introduced in Definition 2.15 will allow us to unify our exposition of several results in subsequent sections. Remark that the crossed products G and G £TO(A) carry honest G-C*-algebra structures given by the dual

actions.

At a few points we will need a notion of equivariance for *-homomorphisms with target D^bB or M(Dto,b).

Let G be a coexact compact quantum group and (B, ¡3) be a G-C*-algebra. Note that nondegeneracy of the *-homomorphism ¡3 : B ^ C'(G) O B implies that

3^(Dto,b) C D^c'g^b = (C'(G) O B)(C'(G) O B)TO(C(G) O B) is a nondegenerate C*-subalgebra. Hence induces a *-homomorphism

M(D^) ^ M(D^c(g)®B), which we will again denote by

Definition 2.17. Let G be a coexact compact quantum group and let (A, a), (B, ¡3) be G-C*-algebras. A *-homomorphism y : A ^ M(DTO,B) is called G-equivariant if oy = (idOy) o a, where both sides are viewed as maps from A into M(DTO,C(g)®b). If y is G-equivariant, then we also write y : (A, a) ^ (M(DTO,B), ¡3TO).

Remark 2.18. Note in particular that if y : A ^ M) is G-equivariant in the sense of Definition 2.17 then ¡3^ o y(A) c Gr(G) <g> M(D^s) C M(D^c'{g)®b).

It is immediate from the definitions that a *-homomorphism p : A ^ is

G-equivariant as a *-homomorphism A ^ B« if and only if it is G-equivariant as a *-homomorphism A ^ M(D«,B).

3. Equivariantly sequentially split *-homomorphisms

In this section we discuss the notion of sequentially split *-homomorphisms between G-C * -algebras, which was studied in [4] in the case of actions by groups.

Definition 3.1 (cf. [4, 2.1, 3.3]). Let G be a quantum group which is either discrete or compact and coexact. Moreover let (A, a), (B, (3) be G-C * -algebras. We say that an equivariant *-homomorphism p : (A, a) ^ (B, (3) is equivariantly sequentially split if there exists a commutative diagram of G-equivariant *-homomorphisms of the form

{A, a) -^ (A

oo, aoo /

where the horizontal map is the standard embedding. If ^ : (B, ¡3) ^ (A«, a«) is an equivariant *-homomorphism fitting into the above diagram, then we say that ^ is an equivariant approximate left-inverse for p.

An important feature of the theory of sequentially split *-homomorphisms is that it is compatible with forming crossed product C -algebras. The proof makes use of the following fact.

Lemma 3.2. Let G be a quantum group which is either discrete and exact or compact and coexact. Moreover let (A, a) be a G-C *-algebra. Then there exists a G-equivariant *-homomorphism G Ka^,r A« ^ (G Ka,r A)«, compatible with the natural inclusions of G K a r A on both sides.

Proof. Assume first that G is discrete and exact. Since taking reduced crossed products with G is exact, the canonical map GKa»,r¿«(A) ^ ¿«(GKa,rA) induces a commutative diagram with exact rows

0 -^ G KC0 (a),r Co(A) -^ G K„»r ¿«(A) -G K„„,r A« -^ 0

c0(G Ka,r A)-^ i°(G Ka,r A) -^ (G Ka,r A)o ->■ 0.

Here co(a) : co(A) ^ MCo(G)(C0(G) < c0(A)) denotes the restriction of ao to co(A).

S. Barlak et al. / Journal of Functional Analysis

It is clear from the construction that the *-homomorphism GKa^,rA^ (GKa,rA)TO is G-equivariant and compatible with the canonical inclusions of G Ka r A.

Assume now that G is compact and coexact. Let us abbreviate K = K(L2 (G)). Then the canonical map K O £TO(A) ^ £TO(K O A) induces the following commutative diagram with exact rows

K <g> co(A)

co(K <g> A)

K <g> £œ(A)

->• K <g> Ac

£~(K ® A)

(K <g> A)c

The middle vertical arrow restricts to a *-homomorphism G r £TO(A) -—> £x(G Ka, A). By Lemma 2.12, the canonical G-equivariant map G xa~,r £TO(A) ^ G is surjective. Observe moreover that the canonical surjection £TO(K O A) ^

(KoA)to restricts to the canonical surjection £m(GKrA) ^ (Gk,A)TO. It follows that the embedding KoAto -—> (KoA)to restricts to an embedding GKa^,r-—> (GKa,A)TO. This map is clearly compatible with the canonical embeddings of G Ka, A. Moreover, as G ,r £TO(A) ^ (G Ka,r A)TO is (G-equivariant, this also holds for G ^

(G K a r A)TO. This finishes the proof. □

Proposition 3.3. Let G be quantum group which is either discrete and exact or compact and coexact. Moreover let (A, a) and (B, (3) be G-C*-algebras. Assume that p : (A, a) ^ (.B, (3) is an equivariantly sequentially split *-homomorphism.

Then the induced *-homomorphism G Kr p : G Ka,r A ^ G B between the crossed products is G-equivariantly sequentially split.

Proof. Let ^ : (B, (3) ^ (ATO, aTO) be an approximate left-inverse for p. Passing to crossed products, we obtain a commutative diagram of G-equivariant *-homomorphisms

(G Ka_r A, a)

(G Ka<x,r Ato,7)

^^ GKr^

(G B,j3)

where 7 = aTO. Composing G Kr ^ with the G-equivariant *-homomorphism G

^ (G Ka,r A)TO from Lemma 3.2 yields an equivariant approximate left-inverse for G Kr p. □

Let us next recall the definition of the fixed point algebra of an action of a compact quantum group.

Definition 3.4. Let G be a compact quantum group. For a G-C*-algebra (A, a) we denote by

Aa = (a G A | a(a) = 1 < a}c A the C * -subalgebra of fixed points.

Lemma 3.5. Let G be a coexact compact quantum group and let (A, a) be a G-C*-algebra. Then the canonical inclusion (Aa)° ^ (A°,a)a~ is an isomorphism.

Proof. Let a G (A°,a)a~ be represented by (an)neN G 1°(A). Then

lim ||a(a„) - 1 < an|| = 0

by the fixed point condition. Applying the Haar state p : Cr(G) ^ C in the first tensor factor gives

lim ||(p < id) o a(an) - a„|| = 0

Since (p < id) o a maps A into Aa we conclude that a is represented by an element of £°(Aa). □

As we show next, a naturality property as in Proposition 3.3 also holds for fixed point algebras of actions of compact quantum groups.

Proposition 3.6. Let G be a coexact compact quantum group, and let (A, a), (B, 3) be G-C *-algebras. Assume that p : (A, a) ^ (B, 3) is a G-equivariantly sequentially split *-homomorphism.

Then the induced *-homomorphism p : Aa ^ BP is a sequentially split.

Proof. Let ^ : (B, 3) ^ (A°, a°) be an equivariant approximate left-inverse for p. By equivariance of we have

a° o ^(b) = (id o 3(b) = 1 < ^(b)

for all b G

Hence ^ maps into (Ao a)a~. According to Lemma 3.5 the latter identifies with (Aa)°, and therefore ^ : BP ^ (Aa)° is an approximate left-inverse for p : Aa ^ BP. □

The following stability result is an important feature for the theory of sequentially split *-homomorphisms.

Proposition 3.7. Let G be a quantum group which is either compact and coexact or discrete and exact. Moreover let p : (A, a) ^ (B, 3) be a nondegenerate G-equivariant *-homomorphism between G-C *-algebras.

Then p is G-equivariantly sequentially split if and only if id<p : (K(L2(G)) < A, as) ^ (K(L2(G)) < B, 3k) is G-equivariantly sequentially split.

Please cite this article in press as: S. Barlak et al., The spatial Rokhlin property for actions of compact quantum groups, J. Funct. Anal. (2016), http://dx.doi.org/10.1016/jjfa.2016.09.023

28 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Proof. Let us first consider the case that G is compact and coexact.

Assume that p is G-equivariantly sequentially split, and let ^ : (B, (3) ^ (ATO, aTO) be a G-equivariant approximate left inverse of p. Then the map id O^ : K(L2(G)) O B ^ K(L2(G)) O C (K(L2(G)) O A)to is G-equivariant, and yields an equivariant approximate left-inverse for id Op.

Conversely, assume that id Op : (K(L2(G)) O A, aK) ^ (K(L2(G)) O B, (K) is G-equivariantly sequentially split. Let ^ : (K(L2 (G)) O B, (K) ^ ((K(L2(G)) O A)TO, (ai)OT) be a G-equivariant approximate left-inverse. As p is assumed to be non-degenerate, the image of ^ is contained in Dtok(l2(g))®a. Using the isomorphism DTOjk(l2(g))®a = K(L2(G)) O D^a from Lemma 2.4, we see that ^ defines a non-degenerate *-homomorphism from K(L2(G)) O B into K(L2(G)) O DTOja. Let us denote the extension M(K(L2(G)) O B) ^ M(K(L2(G)) O DTOjA) to multiplier algebras again by

We shall write for the restriction of ^ to B = 1 O B C M(K(L2(G)) O B). Then ^B : B ^ M(DTO ,k(L2(G)®a) is a *-homomorphism whose image is contained in the relative commutant of K(L2(G)) O 1. According to [4, 1.8], its image im(^B) is therefore contained in 1 O M(DTO ,a). Using again nondegeneracy of p we see that im(^B) is in fact contained in 1 O , a. From these observations and the sequential split property we conclude that ^ can be written in the form ^ = id^(L2(G)) O^ for a non-degenerate *-homomorphism ^ : B ^ It is easy to check that ^ is an approximate left-inverse

for p.

We claim that ^ : B ^ is G-equivariant. For this consider a simple tensor

T O b e K(L2(G)) O B and compute

(id O^)3k(T O b) = (id O^)(X*2(1 O T O 1)3(b)isXi2)

= (id O id OV0(X*2(1 O T O 1)3(b)isXi2) = X*2(1 O T O 1)((idO^)3(b))isXi2

(ok)to(*(T O b)) = (aK)TO(T O ^(b)) = X*2(1 O T O 1)aTO(V>(b))isXi2.

Here all expressions are viewed as elements of (Cr(G) O K(L2(G)) O A)TO. Equivariance of ^ means that the above expressions are equal. We conclude (id(b) = aTO(^(b)) for all b e B as desired.

In the case that G is discrete and exact we can follow the above arguments almost word by word, in this case the situation is even slightly easier since all algebras involved are honest G-C * -algebras. □

Proposition 3.8. Let G be a quantum group which is either compact and coexact or discrete and exact. Moreover assume that (A, a), (B, (3) are separable G-C *-algebras, and let p : (A, a) ^ (B, (3) be a non-degenerate equivariant *-homomorphism.

Then p is G-equivariantly sequentially split if and only if

p = G Kr p : (G Kar A, a) ^ (G K^r B,j3) is G-equivariantly sequentially split.

Proof. If p is G-equivariantly sequentially split, then Proposition 3.3 shows that p is G-equivariantly sequentially split. On the other hand, if p is (G-equivariantly sequentially split, then Proposition 3.3 implies that p is G-equivariantly sequentially split. Under the G-equivariant isomorphism given by Takesaki-Takai duality, see Theorem 1.20, the map p corresponds to id^ <p : (K < A, ai) ^ (K < B, (k). Here we abbreviate K = K(L2(G)). Given an equivariant approximate left-inverse (G K^ r G B, ¡3) ^ ((G Ka r G Ka r A)«, ¿3«) for p, componentwise application of Takesaki-Takai duality yields a G-equivariant approximate left-inverse (K < B, (k) ^ ((K < A)«, (¿k)«) for idi <p. Hence, idi <p is G-equivariantly sequentially split. The claim now follows from Proposition 3.7. □

4. The Rokhlin property and approximate representability

In this section we introduce the key notions of this paper, namely the spatial Rokhlin property and spatial approximate representability for actions of quantum groups. Moreover we prove that these notions are dual to each other.

4.1. The spatial Rokhlin property

Let us start by defining the spatial Rokhlin property.

Definition 4.1. Let G be a coexact compact quantum group and (A, a) a separable G-C * -algebra. We say that a has the spatial Rokhlin property if the second-factor embedding

a = 1 m idA : (A, a) ^ (C'(G) IE A, A № a) is G-equivariantly sequentially split.

Remark 4.2. Definition 4.1 is indeed a generalization of the classical notion of the Rokhlin property, see [4, 4.3]. Indeed, for a classical compact group G, the braided tensor product C(G) I A agrees with the ordinary tensor product C(G) < A. The term spatial in Definition 4.1 refers to the fact that we have chosen to work with minimal (braided) tensor products; we will comment further on the implications of this choice in Remark 4.9 below.

30 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

In special cases the Rokhlin property can be recast in the following way. Recall that if G is a compact quantum group and (A, a) a G-C * -algebra we write S(A) for the spectral subalgebra of A. We shall use the Sweedler notation a(a) = «(-i) O a(0) for the coaction a : S(A) ^O(G) ©S(A).

Proposition 4.3. Let G be a coexact compact quantum group and (A, a) a separable G-C -algebra.

a) If a has the spatial Rokhlin property, then there exists a unital and G-equivariant *-homomorphism k : (Cr(G), A) ^ (M(DTOja), aTO) satisfying

aK(/) = K(a(_2)/S(a(_i)))a(0)

for all / e O(G) C Cr(G) and a e S(A) C A. Moreover 'we have ||K(S(a(_i)))a(0)y < ||a|| for all a e S(A).

b) Assume G is coamenable. If a *-homomorphism k : Cr(G) ^ M(DTOja) as in a) exists such that ||k(S(a(_i)))a(0) || < ||a|| for all a e S(A), then a has the spatial Rokhlin property.

Proof. a) Assume first that a has the spatial Rokhlin property. Let

$ : (Cr(G) H A, A H a) ^ (ATO,aTO)

be an equivariant approximate left-inverse for La as required by Definition 4.1. Since ia is non-degenerate, the image of this *-homomorphism is contained in DTO a, and $ : Cr(G) H A ^ is again nondegenerate. Let us also denote the unique strictly

continuous extension of ^ to multipliers by the same letter, so that we have

$ : M(Cr(G) H A) ^ M(Dto,a).

By equivariance of the unital *-homomorphism

K = $ O icr(G) : Cr(G) ^ M(Dto,a)

is also equivariant. According to the definition of the braided tensor product, we have

1A(a)lC,(G)(/) = 1C,(G) (a(_2)/S(a(_i))) 1A(a(0))

for all / e O(G) and a e S(A). The desired twisted commutation relation for k then follows by applying $. Moreover, the norm condition ||K(S(a(_i)))a(0)|| < ||a| is a consequence of the formula K(S(a(_i)))a(0) = $(T__i(1 Oa)) for a e S(A) and the fact that $ and the isomorphism T_i from Lemma 1.24 are *-homomorphisms between C * -algebras.

b) Consider the map i : S (A) ^ D°aa given by ¿(a) = K(S(a(_1)))a(0). Then the commutation relation for k gives

l(a)K(f) = K(S(a(_3))a(_2)fS(a(_1)))a(0) = k(i )l(a)

for any f G O(G), using the antipode relation for the Hopf algebra O(G). Using that S is antimultiplicative we therefore obtain

i(ab) = K(s(a(-I)b(-1) ))a(0)b(0)

= k(s (b(-1)))K(S(a(-1)))a(0)b(0) = k(S (b(-1)))i(a)b(0) = l(a)K(S(b(-1)))b(0) = l(a)l(b)

for all a, b G S(A). Moreover, using S(h*) = S-1(h)* for all h G O(G) we have

¿(a* ) = K(S(a*_1)))a*0) = k(S 1(a(_1))*)a,

(-iy^a(o) _ k[<S \a(-i)J Ja(o) = (a(o)K(S-1(a(_i))))* = (K(a(-2)S-1(a(-3))S(a(-i)))a(o))* = (k(S ("(-i) ))"(o))* = i(a)*

for a € S(A). It follows that i is a *-homomorphism. By assumption i is bounded, so that it extends to a *-homomorphism i : A ^ D«aa.

Combining k and i we obtain a *-homomorphism ^ = k < i : C(G) << A = C(G) ®max A ^ D«,a, using the universal property of the maximal tensor product and nuclearity of C(G). Since k is equivariant one checks that i maps into the fixed point algebra of A« ,a, and together with equivariance of k it follows that ^ : (C(G) << A, A <<id) ^ (D«aa, a«) is G-equivariant. Using the isomorphism from Lemma 1.24 we see that it defines an approximate left-inverse for iA : A ^ C(G) № A = C(G) < A. □

Remark 4.4. Let us point out that the norm condition in part b) of Proposition 4.3 is automatically satisfied if G is a finite quantum group. Indeed, in this case the spectral subalgebra S(A) is equal to A, and the claim follows from the fact that *-homomorphisms between C -algebras are contractive. It can also be shown that the norm condition always holds if G is a classical compact group, but it seems unclear whether it is automatic in general.

Classically, a Rokhlin action of a compact group on an abelian C * -algebra C0(X) induces a free action of G on X. In the quantum case, an analogue of the notion of freeness has been formulated by Ellwood in [7]. Namely, an action a : A ^ Cr(G) << A of a compact quantum group G on a C * -algebra A is called free if [(1<A)a(A)] = Cr(G)<A. It is shown in [7, Theorem 2.9] that this generalizes the classical concept of freeness.

32 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Let us verify that the spatial Rokhlin property implies freeness also in the quantum case.

Proposition 4.5. Let G be a coexact compact quantum group and let (A, a) be a separable G-C*-algebra. If a has the spatial Rokhlin property, then it is free.

Proof. Let $ : Cr(G) H A ^ be an equivariant approximate left-inverse for the inclusion map iA : A ^ Cr (G) H A. Notice that the action of G on (Cr (G) H A, A H a) = (Cr(G) O A, A O id) is free, so that

[(A H a)(Cr(G) H A)(1 O Cr(G) H A)] = Cr(G) O (Cr(G) H A).

In fact, for any / e O(G) C Cr(G) and a e A we find finitely many elements x1, yi e Cr(G) H A such that / O (1 H a) = ^(A H a)(xi)(1 O yi) in Cr(G) O (Cr(G) H A). Applying id O$ to this equality and using equivariance, we obtain

/ O $(1 H a) = / O a = Y a^$(xi)(1 O $(yi))

in (Cr(G) O A)TO. Consider lifts (X^)„eN and (y^)„eN in l^(A) for $(xi), $(yi), respectively. Then

/ O a = lim V a(xin)(1 O fn),

n—Y^O f

which implies C r(G) © A C [a(A)(1 O A)], and hence also Cr (G) O A = [a(A)(1 O A)]. □

Let us point out that the Rokhlin property is strictly stronger than freeness; this is already the case classically. For instance, the antipodal action of G = Z2 on Si does not have the Rokhlin property.

Here comes the first main result of this paper:

Theorem 4.6. Let G be a coexact compact quantum group and let (A, a) be a separable G-C *-algebra. If a has the spatial Rokhlin property, then the two canonical embeddings

Aa A and G Ka,r A '—> K(L2(G)) O A

are sequentially split. In particular, if A has any of the following properties, then so do the fixed-point algebra Aa and the crossed product G Ka_r A:

• being simple;

• being nuclear and satisfying the UCT;

• having finite nuclear dimension or decomposition rank;

• absorbing a given strongly self-absorbing C *-algebra D.

Proof. Let ^ : Cr(G) № A ^ A« be an equivariant approximate left-inverse for the embedding ia : A ^ Cr(G) № A. The resulting commutative diagram of equivariant *-homomorphisms

Cr(G) IE A

induces a commutative diagram of *-homomorphisms

Aa-(Aa)c

by Proposition 3.6. Notice here that Lemma 1.24 implies that (Cr (G) № A)AHa = A in such a way that the canonical embeddings of Aa on both sides are compatible.

For the statement about the crossed product GKa,r A, observe that the spatial Rokhlin property for a means that a : A ^ Cr(G) < A is sequentially split, taking into account the isomorphism from Lemma 1.24. According to Proposition 3.3, it follows that the map

G Kr a : G Ka,r A <-> G ><A®ldA,r (C r(G) < A) = (G X A,r C r(G)) < A

is sequentially split. Moreover, by the Takesaki-Takai duality Theorem 1.20, we have G KA,r Cr(G) = K(L2(G)), and the resulting map G Ka,r A ^ K(L2(G)) < A is the standard embedding.

The asserted permanence properties are then a consequence of [4, Theorem 2.9]. □

4-2. Spatial approximate representability

Let us now define spatial approximate representability.

Let G be a discrete quantum group and (A, a) a separable G-C * -algebra. Denote by Wa = (id<iG)(W) € M(C0(G) < (G ka r A)) where iG : Cr*(G) ^ M(G Ka,r A) is the canonical embedding. The unitary Wa implements the inner action of G on the crossed product, more precisely Ad(Wa) turns M(G Ka,r A) into a G-C * -algebra such that

Wa (1 < a)Wa = a(a)

for all a € A C M(G Ka,r A).

Definition 4.7. Let G be a discrete quantum group and (A, a) a separable G-C * -algebra. We say that a is spatially approximately representable if the natural embedding

34 S. Barlak et al. / Journal of Functional Analysis ••• (••••)

ja : (A, a) ^ (G Ka,r A, Ad(W*)) is G-equivariantly sequentially split.

Proposition 4.8. Let G be an exact discrete quantum group and (A, a) a separable G-C *-algebra.

a) If a is spatially approximately representable, then there exists a unitary representation V e M(C0(G) <g> DœA) of G such that

(id ®aTO)(V ) = V23V13V23

V* (1 O a)V = a(a)

for all a e A.

b) If G is amenable and there exists a unitary representation V e M(Co(G) O DTOja) as in a), then a is spatially approximately representable.

Proof. a) Assume that a is spatially approximately representable and let

$ : G Ka.r A ^

be a G-equivariant approximate left-inverse for the embedding A ^ G Ka,r A. Since this embedding is nondegenerate, the image of $ is contained in D^,a, and $ : G Ka,r A ^ is a nondegenerate *-homomorphism. Let us denote the unique strictly continuous extension of $ by the same letter, so that

$ : M(G Ka.r A) ^ M(D^a).

We let

V = (id<g>V0(Wa) e M(Co(G) < D^a)

be the unitary representation corresponding to the restriction of ^ to C* (G). Equivari-ance of ^ : G Kar A ^ means

(id ®^)(W*(1 < x)Wa) = o ^(x) for all x e G K a r A. In particular, for every a e A we obtain

a(a) = o ^(a) = V* (1 < a)V.

Moreover, if y = (w <g> id)(Wa) € M(G Kar A) for w € L(L2(G))*, then

(w <g> id)(id <g>a«)(V) = a« o tf(y)

= (id®^)(Wa(1 ® y)Wa) = (w <g> id <g> id)(id <g> id ®tf)(W2*3W13W23 ) = (w <g> id <g> id)(V2*3Vi3V23).

Since this holds for all w € L(L2(G))* we conclude (id<8>a«)(V) = V2*3V13V23.

b) Suppose that V € M(C0(G) ® D«a) is a unitary satisfying the conditions in a). Clearly, the canonical map i : A ^ D« a is equivariant and nondegenerate, and the formula

(id <g>i) o a(a) = V*(1 <g> i(a))V for all a € A

means that i and V define a covariant pair. Hence they combine to a nondegenerate *-homomorphism

tf : G Ka,f A ^ M(D«,a)

such that tf o i(a) = a for all a € A. Since G is amenable, we can identify the full crossed product G Ka _f A with the reduced crossed product G Ka _r A. To verify that tf : G Ka_ r A ^ M(D« _a) is G-equivariant, it suffices to check this separately on the copies of Cr*(G) and A inside M(G Ka_r A).

For a € A C Gxa_ rA, the equivariance condition follows immediately from the relation tf o i(a) = a. On Cr* (G) it is obtained by slicing the equation

(id ®^«)(id ®tf)(Wfi) = (id ®^«)(V) = V23V13V23

= (id ® id ®tf)(W2*3 W13W23),

in the first tensor factor and using Cr*(G) = [(L(L2(G))* ® id)(W)]. We conclude that tf determines a G-equivariant approximate left-inverse for the inclusion A ^ G Ka_ r A. □

Remark 4.9. Definition 4.7 generalizes approximate representability for actions of discrete amenable groups, see [4, 4.23]. In the same way as already indicated in Remark 4.2, the term spatial in our definition is included since we work with minimal (braided) tensor products and reduced crossed products. In fact, approximate representability for classical discrete groups is defined in terms of the full crossed product instead, see [4, 4.23]. Notice that the trivial action of the free group F2 on C is clearly approximately representable, but it is easily seen not to be spatially approximately representable in the sense of Definition 4.7.

ARTICLE IN PRESS

YJFAN:7618

36 S. Barlak et al. / Journal of Functional Analysis ••• (••••')

It would therefore be more natural to develop the theory with maximal tensor products and full crossed products instead. However, this would mean in particular that one would have to work with full coactions taking values in maximal tensor products, which is technically less convenient.

Let us point out that all the above mentioned issues disappear for coamenable compact quantum groups and amenable discrete quantum groups, respectively; in these cases, we may omit the term spatial, and speak of the Rokhlin property and approximate representability.

4.3. Duality

We shall now show in several steps that the spatial Rokhlin property and spatial approximate representability are dual to each other.

Proposition 4.10. Let G be a compact quantum group and (A, a) a separable G-C*-algebra. Consider the G-equivariant *-homomorphism

ia = 1 H idA : (A, a) ^ (Cr(G) H A, A H a).

Then there exists a (G-equivariant *-isomorphism

* a : (G KAHa ,r (Cr(G) H A), (A H a)~) ^ (G k â,r (G k a ,, A), Ad(Wâ* )) that makes the following diagram commutative:

(G K a,r A, a)

Ok, ia

(G k ah a ,r (C r(G) H A), (A H a)

(G kâ,r (G ka ,r A), Ad(WÏ))

Proof. Using Lemma 1.24 and Theorem 1.20, we obtain as the composition of the following identifications:

G k ah a ,r (C r(G) H A) = G KA0idA,r (Cr (G) < A)

^ [(C*(G) < 1 < 1)A(Cr(G))i2(1 < 1 < A) ^ [C* (G)Cr(G) < A] = K(L2(G)) < A = [(UCr(G)UC*(G) < 1)a(A)] = G k â,r G k a,r A.

Note that the copy of A inside M(G xAHa,r (Cr(G) M A)) identifies with a(A) C M(K(L2(G)) ® A), and that the same holds for the copy of A inside M(G Xa,r G Xa,r A). Similarly, the copies of Cr* (G) on both sides identify with Cr* (G)<g>1 C M(K(L2 (G))<g>A).

Moreover, the above identifications are compatible with the action of G on K(L2(G)) <g) A implemented by conjugation with £V"£. More precisely, the coaction

T ^ Ad(Si2l>12Si2)(1 <g> T)

on K(L2(G)) ® A corresponds to the dual coaction on G XABa,r (Cr(G) M A) and to the conjugation coaction 7 = Ad( WJ) : G xa,r G Xa,r A ^ M(Cr*(G)cop ® G xa,r G Xa,r A), given by y(T) = WU2(1 ® T)(WU)1 2 where WJ = WU = (1 <g> U)W(1 <g> U). For the latter observe SUE = (1 ® U)W(1 ® U) and take into account the passage from Cr*(G) to Cr* (G)cop. □

Proposition 4.11. Let G be a discrete quantum group and (A, a) a separable G-C*-algebra. Consider the G-equivariant inclusion

ja : (A, a) ^ (G Xa ,r A, Ad(Wa*)). Then there exists a G-equivariant *-isomorphism

: (G ),r (G X a ,r A), Ad(WaT) ^ (Cr (G) M (G X a,r A), A M a)

that makes the following diagram commutative:

A G^rja

(G xa,r A, a) ->-

(G XAd(wa),r (G xa ,r A), Ad(W*)")

(cr(G) m (g xa ,r A), a m a)

Proof. We obtain $ a as the composition of the following identifications:

G XAd(W) (G Xa,r A) = [(C*(G) < 1 < 1)W2(1 < C*(G) < 1)a(A)2sWi2] = [Wi*2(^cop(C;(G)) < 1)(1 < c;(G) < 1)a(A)23Wi2] = [(Acop(C*(G)) < 1)(1 < C*(G) < 1)a(A)23] = [(C*(G) < C*(G) < 1)a(A)23] = C* (G)cop < (G X a,r A) = C'(G) < (G X a ,r A)

= c '(G) m (g x a ,r A).

38 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Under these identifications, the copy of Cr* (G) inside M(G Ka ,r A) on the left hand side gets identified with 1 I (Cr*(G) << 1) inside Cr(G) I (G Ka ,r A), and the copy of A in Gka ,r A is mapped to 1Ia(A). In other words, we indeed obtain a commutative diagram as desired.

Moreover, it is not hard to check that the dual action on G KAd(w*),r (G Ka,r A) corresponds to the action of Acop = A on the first tensor factor of Cr(G) < (G Ka,r A). It follows that $ a is G-equivariant. □

As a consequence, we obtain the duality between the spatial Rokhlin property and spatial approximate representability.

Theorem 4.12. Let G be a coexact compact quantum group and let (A, a) be a separable G-C*-algebra. Then a has the spatial Rokhlin property if and only if a is spatially approximately representable.

Dually, let G be an exact discrete quantum group and let (A, a) be a separable G-C *-algebra. Then a is spatially approximately representable if and only if a has the spatial Rokhlin property.

Proof. Let us first consider the case of compact quantum groups. By the general duality result from Proposition 3.8, we know that

ia : (A, a) <—> (Cr(G) I A, A I a)

is G-equivariantly sequentially split if and only if the induced *-homomorphism

G Kr la : (G ka ,r A, a) ^ (G KA№l,r (C(G) I A), (A I a)~)

is (G-equivariantly sequentially split. By Proposition 4.10, there exists a commutative diagram of G-equivariant *-homomorphisms

(G k a,r a, a) -GK'1a , (G k ah a ,r (C (G) I A), (A I a)~)

(G ka,r (G ka ,r A), Ad(W-J))

We conclude that ia is G-equivariantly sequentially split if and only if j& is G-equi-variantly sequentially split. This means that a has the spatial Rokhlin property if and only if aa is spatially approximately representable.

The claim in the discrete case is proved in an analogous fashion. Again by Proposition 3.8, the G-equivariant *-homomorphism

j a : (A, a) ^ (G k a,r A, Ad(^a*))

Please cite this article in press as: S. Barlak et al., The spatial Rokhlin property for actions of compact quantum groups, J. Funct. Anal. (2016), http://dx.doi.org/10.1016/jjfa.2016.09.023

is G-equivariantly sequentially split if and only if the induced *-homomorphism

G Kr ja : (G Xa ,r A, a) ^ (G KAd(W,)>r (G xa>r A),

is G-equivariantly sequentially split. By Proposition 4.11, there exists a commutative diagram of G-equivariant *-homomorphisms

(G xa,r A, a)

GKr ja

(G XAd(^a),r (G xa ,r A), Ad(W*

(Gr(G) E (G xa ,r A), A E a)

We conclude that ja is G-equivariantly sequentially split if and only if is G-equi-variantly sequentially split. Hence a is spatially approximately representable if and only if aa has the spatial Rokhlin property.

5. Rigidity of Rokhlin actions

In this section we provide a classification of actions of coexact compact quantum groups with the spatial Rokhlin property on separable C -algebras. This type of result was first obtained by Izumi in [15]. Our basic approach follows Gardella-Santiago [11, Section 3], who proved corresponding results for finite group actions. We note that Gardella-Santiago have also announced the results of this section for actions of classical compact groups, see [10].

Recall that if (B, 3) is a G-C * -algebra for a compact quantum group G then B& C B denotes the fixed point subalgebra. We shall also write /3 for the induced coaction on the minimal unitarization B of B; note that 3(1) = 1 < 1.

Definition 5.1. Let G be a compact quantum group. Let a : A ^ Cr(G) < A and 3 : B ^ Cr(G) < B be two G-actions on C * -algebras, and assume that A is separable. Let pi, p2 : (A, a) ^ (B, /) be two equivariant *-homomorphisms. We say that pi and p2 are approximately G-unitarily equivalent, written p1 «u,G p2, if there exists a sequence of unitaries vn G W(B&) such that

p2(x) = lim vnp1(x)vn for all x G A.

Remark 5.2. For the trivial (quantum) group G, the above definition recovers the usual notion of approximate unitary equivalence between *-homomorphisms. We write simply pi «u p2 instead of p1 «u,G p2 in this case.

Proposition 5.3. Let G be a compact quantum group. Let a : A ^ Cr(G) < A and 3 : B ^ Cr(G) < B be two G-actions on separable C *-algebras. Let p : (A, a) ^ (B,/)

40 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

and ^ : (B, 3) ^ (A, a) be two equivariant *-homomorphisms such that ^ o p ~U,G idA and po^ ~u,g ids. Then there exists an equivariant *-isomorphism $ : (A, a) ^ (B, 3) such that $ ~u,g P and $-1 ~u,g ^.

Proof. This follows from a straightforward adaptation of the proof of [29, Corollary 2.3.4 to the setting of G-equivariant *-homomorphism. For this, one requires the approximate intertwinings from [29, Definition 2.3.1] to be (approximately) G-equivariant in the obvious way. The resulting *-isomorphism $ : A ^ B then turns out to be a-to-3-equivariant. Moreover, the approximate unitary equivalences $ «U p and $-1 «U ^ that come out of the proof are indeed implemented by unitaries in U(B3) and U(Aa), respectively. □

Let us now consider a series of partial results that will lead to the classification of Rokhlin actions.

Lemma 5.4. Let G be a compact quantum group. Let a : A ^ Cr(G) ® A and 3 : B ^ Cr(G) <g) B be two G-actions on separable C *-algebras. Let p : A ^ B be a *-homomorphism that is equivariant modulo approximate unitary equivalence, i.e. 3 ◦ p «U (id ®p) o a as *-homomorphisms between A and Cr(G) <g) B. Then for every finite set FCCA and every s > 0, there exists a unitary v G (Cr(G) M B)~ such that

(A M 3) o Ad(v) o (1 M p)(x) = (id ®(Ad(v) o (1 M p))) o a(x)

|| [(1 M p)(x),v] || < s + ||3 o p(x) - (id<g>p) o a(x)||

for all x G F.

Proof. For convenience, the term id will always denote the identity map on Cr(G) in this proof. Identity maps on other sets are decorated with the corresponding set.

Using our assumptions on p, we may choose unitaries un G (Cr(G) ® B)~ such that

Ad(un) o 3 o p n—^ (id ®p) o a (e5.1)

in point-norm. By Lemma 1.24, we have the equivariant isomorphism

T3 : (Cr(G) M B, A M 3) ^ (Cr(G) <g> B, A <g> ids) (e5.2)

satisfying

T3 (1 M x) = 3(x) for all x G B. (e5.3)

Let us also denote by T3 the obvious extension to the unitarizations. Set vn = T-1(un). We calculate

lim (A M 3) o Ad(vn) o (1 M p)

n—^^O

= lim (A M 3) o Ad(T-1(un)) o (1 M p)

n—TO P

(e=3) lim (A M 3) o T-1 o Ad(un) o / o p

n—TO P

(e=1) (A M 3) o T-1 o (id ©p) o a

(e=2) (id ©Tp-1) o (A © idB) o (id ©p) o a

= (id ©T-1) o (id © id ©p) o (A © idA) o a

= (id ©T-1) o (id © id ©p) o (id ©a) o a

= (id©(T-1 o (id©p) o a)} o a

( = ) lim ( id ©(T-1 o Ad(un) o / o p)) o a

n—TO P '

= lim (id©(Ad(vn) o (1 M p))\ o a.

n—TO

One should note that even though the existence of all these limits is a priori not clear at the beginning of this calculation, it follows a posteriori from the steps in this calculation. Moreover, we calculate for all x G A that

II [1 M p(x),vn]\\

= I|[(1 M p)(x),T-1(un)]\

(e=3) || [(3 o p)(x),Un]|| = ||(Ad(Un) o 3 o p)(x) - (3 o p)(x)||

(e5.1)

—> ||(id©p) o a(x) — 3 o p(x)||.

From these two calculations, it is clear that for given FCCA and £ > 0, any of the unitaries vn satisfies the desired property for sufficiently large n. □

Lemma 5.5. Let G be a coexact compact quantum group. Let a : A ^ Cr(G) © A and 3 : B ^ Cr(G) © B be two G-actions on separable C *-algebras. Assume that 3 has the spatial Rokhlin property. Let p : A ^ B be a *-homomorphism that is equivariant modulo approximate unitary equivalence, i.e. 3 o p ~u (id ©p) o a as *-homomorphisms between A and Cr(G) ©B. Then for every finite set FCA and £ > 0, there exists a unitary v G B such that

3 o Ad(v) o p(x) =e (id ©(Ad(v) o p)} o a(x)

|| [p(x),v] || < £ + ||3 o p(x) — (1 © p) o a(x)||

for all x G F.

42 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Proof. As 3 is assumed to have the spatial Rokhlin property, let

^ : (Cr(G) M B, A M 3) ^ (Bw,3w)

be an equivariant *-homomorphism satisfying

^(1 M b) = b for all b G B. (e5.4)

We also denote by ^ the canonical extensions to the smallest unitarizations on both sides. Now let FCA and s > 0 be given. Apply Lemma 5.4 and choose a unitary w G (Cr(G) M B)~ such that

(A M 3) o Ad(w) o (1 M p)(x) =/2 (id <g>(Ad(w) o (1 M p))) o a(x) (e5.5)

||[(1 M p)(x),w]|| < s/2 + ||3 o p(x) - (id<p) o a(x)|| (e5.6)

for all x G F. Set v = ^(w) G (Bw)~ C Bw. Combining the equivariance of ^ with (e5.4), (e5.5) and (e5.6), we obtain

3w o Ad(v) o p(x) =e/2 (id <(Ad(v) o p)) o a(x)

||[p(x),v]|| < s/2 + ||3 o p(x) - (id<p) o a(x)||

for all x G F. Now represent v by some sequence of unitaries vn G B. Then these equations translate to the conditions

limsup ||3 o Ad(vn) o p(x) — (id <(Ad(vn) o p)) o a(x) || < s/2

limsup ||[p(x),vn]| < s/2 + ||3 o p(x) - (id<p) o a(x)||

for all x G F. It follows that for sufficiently large n, any of the unitaries vn satisfies the desired inequalities with respect to s in place of s/2. This finishes the proof.

Proposition 5.6 (cf. [11, 3.2]). Let G be a coexact compact quantum group. Let a : A ^ Cr(G) < A and 3 : B ^ Cr(G) < B be two G-actions on separable C *-algebras. Assume that 3 has the spatial Rokhlin property. Let p : A ^ B be a *-homomorphism that is equivariant modulo approximate unitary equivalence, i.e. 3 o p ~U (id<p) o a as *-homomorphisms between A and Cr(G) < B. Then there exists an equivariant *-homomorphism ^ : (A, a) ^ (B, 3) with ^ «U p.

Proof. Let

Fl C F2 C F3 C . . . (CA

be an increasing sequence of finite subsets with dense union. Let (en)„eN be a decreasing sequence of strictly positive numbers with Xm=i £n < Using Lemma 5.5 we find a unitary v1 € B satisfying

ß o Ad(v1) o p(x) =El (id <gi(Ad(vi) o p)) o a(x)

for all x € F1. Applying Lemma 5.5 again (but now for Ad(v1) o p in place of p), we find a unitary v2 € B satisfying

ß o Ad(v2v1) o p(x) =£2 (id ®(Ad(v2v1) o p)) o a(x)

||[(Ad(v1) o p)(x),v2]|| < £2 + llß o Ad(v1) o p(x) - (id®(Ad(vO o p)) o a(x)||

for all x € F2. Applying Lemma 5.5 again (but now for Ad(v2v1) o p in place of p), we find a unitary v3 € B satisfying

ß o Ad(v3v2v1) o p(x) =£3 (id ®(Ad(v3v2v1) o p)) o a(x)

||[(Ad(v2v1) o p)(x),v3]|| < £3 + ||ß o Ad(v2v1) o p(x) - (1 ® (Ad(v2v1) o p)) o a(x)||

for all x € F3. We inductively repeat this process and obtain a sequence of unitaries vn € B satisfying

ß o Ad(v„ • • • v1) o p(x) =£n (id®(Ad(v„ • • • v1) o p)) o a(x) (e5.7)

for all n > 1 and

|[(Ad(v„_1 • • • v1) o p)(x),vn]|| (e5 8) < £n + ||ß o Ad(v„_1 • • • v1) o p(x) - (id®(Ad(v„_1 • • • v„) o p)) o a(x)||

for all x G Fn and n > 2. For m > n > k and x G Fk this implies

44 S. Barlak et al. / Journal of Functional Analysis ••• (••••)

|| Ad(vm • •• v{) o p(x) — Ad(v„ ••• V1) o p(x)||

m — 1

< ^^ || Ad(vj+1 • • • V1) o p(x) — Ad(vj • • • V1) o p(x)

j=n m—1

Y1 ll[(Ad(vJ •••^1)0 p)(x),vj+1]

(e5.7),(e5.8) ^1 m

< £j+1 + £j < 2 •z, £j ■

j=n j=n

As the sn were chosen as a 1-summable sequence and the union of the Fn is dense, this estimate implies that the sequence Ad(vn ••• v1) o p(x) is Cauchy for every x G A. In particular, the point-norm limit ^ = limn^TO Ad(vn • • • v1) o p exists and yields a well-defined *-homomorphism from A to B. By construction we have ^ «u p, and the equivariance condition

3 o ^ = (id o a follows from (e5.7). This finishes the proof. □

Lemma 5.7. Let G be a compact quantum group. Let a : A ^ Cr(G) < A and 3 : B ^ Cr(G) < B be two G-actions on separable C *-algebras. Let p1, p2 : (A, a) ^ (B, 3) be two equivariant *-homomorphisms. If p1 «u p2, then 1 IE p1 1 IE p2 as equivariant

*-homomorphisms from A to C(G) I B.

Proof. Let un G U(B) be a sequence of unitaries satisfying

Ad(un) o p1 n—p2 ■ (e5.9)

Using Lemma 1.24, we consider the equivariant isomorphism

Tp : (Cr(G) I B, A I 3) ^ (Cr(G) < B, A < idB)

that satisfies condition (e5.3). We shall also denote by Tp the obvious extension to the unitarizations. Set

T-1(1 <g> un) G (Cr(G) H B)

As 1 << un is in the fixed-point algebra of A <g> id_g, it follows that vn is in the fixed-point algebra of A I 3. We have

Ad(w„) o (1 H px) (e=3) To Ad(1 <g> un) o p o px

= T—1 o Ad(1 ® un) o (1 <g) px) o a

T- 1 o (1 ® p2 ) o

(e5.9) ^-1

= T-1 o 3 o p2

(e5.3) = 1 M p2.

This shows our claim. □

Proposition 5.8. Let G be a coexact compact quantum group. Let a : A ^ Cr(G) © A, 3 : B ^ Cr(G) © B and 7 : C ^ Cr(G) © C be three G-actions on separable C *-algebras. Let

p1,p2 : (A, a) ^ (B,3), ^ :(B,3) ^ (C,7)

be equivariant *-homomorphisms. Assume that ^ is equivariantly sequentially split. Then p1 ~u,G p2 if and only if ^ o p1 «u,G ^ o p2.

Proof. If p1 «u,G p2, then clearly ^ o p1 «u,G ^ o p2. For the converse, assume that ^op1 «u,G ^op2. Let k : (C, 7) ^ (BTO, 3TO) be an equivariant approximate left-inverse for Then k o ^ o p1 «uG k o ^ o p1, or in other words, p1 «u,G p2 as equivariant *-homomorphisms from A to BTOp. Given FCA and £ > 0, we therefore find some u G U((BTOp)) such that

Ad(u) o p1(a) =e p2 (a) for all a G F.

According to Lemma 3.5, we may choose a representing sequence (un)neN C U(BP) for u. Picking a suitable member of this sequence, we find a unitary v G U(BP) such that

Ad(w) o px(a) =2e p2(a) for all a G F.

This shows that p1 «u,G p2 as equivariant *-homomorphisms from (A, a) to (B, 3). □

Corollary 5.9 (cf. [11, 3.1]). Let G be a coexact compact quantum group. Let a : A ^ Cr(G) © A and 3 : B ^ Cr(G) © B be two G-actions on separable C *-algebras. Assume that 3 has the spatial Rokhlin property. Let p1, p2 : (A, a) ^ (B, 3) be two equivariant *-homomorphisms. Then p1 «u p2 if and only if p1 ~u,g p2.

Proof. We have to show that p1 «u p2 implies p1 «u,G p2. By Lemma 5.7, the amplified *-homomorphisms are G-approximately unitarily equivalent, that is, 1 M p1 «u,G 1 M p2. As 3 has the Rokhlin property, the canonical embedding 1 M idB : (B, 3) ^ (Cr (G) M B, A M 3) is equivariantly sequentially split. Writing 1 M pi = (1 M idB) o pi for i = 1, 2, an application of Lemma 5.8 yields p1 «u,G p2. This finishes the proof. □

46 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Here comes the main result of this section, which generalizes analogous results for finite group actions due to Izumi [15, 3.5], Nawata [25, 3.5] and Gardella-Santiago [11, 3.4]. It also generalizes the corresponding results for finite quantum groups by Osaka-Teruya [21, 10.7] and for classical compact groups by Gardella-Santiago [10].

Theorem 5.10. Let G be a coexact compact quantum group. Let a, 3 : A ^ Cr (G) < A be two G-actions on a separable C *-algebra. Assume that both have the spatial Rokhlin property. Then a «u 3 as *-homomorphisms if and only if there exists an equivariant isomorphism 6 : (A, a) ^ (A, 3) which is approximately inner as a *-automorphism

Proof. First assume that 6 : (A, a) ^ (A, 3) is an equivariant *-isomorphism which is approximately inner as a *-automorphism. Then

3 «u 3 o 6 = (id <6) o a «u a. Now assume that a and 3 are approximately unitarily equivalent. Then clearly

3 o id^ = 3 ~u a = (id < id^) o a,

and analogously a o id^ ~u (id < id^) o 3. Since both a and 3 have the spatial Rokhlin property, it follows from Proposition 5.6 that there exist equivariant *-homomorphisms p1 : (A, a) ^ (A, 3) and p2 : (A, 3) ^ (A, a) that are both approximately inner as *-homomorphisms. Hence Corollary 5.9 implies p1 op2 «u,G id^ and p2 op1 «u,G id^. According to Proposition 5.3 we conclude that there exists an equivariant *-isomorphism 6 : (A, a) ^ (A, 3) with 6 p1. In particular, 6 is also approximately inner as a

*-automorphism. □

To conclude this section, we generalize the K-theory formula for fixed-point algebras of Rokhlin actions, which is originally due to Izumi and was recently extended by the first two authors.

Theorem 5.11 (cf. [15, 3.13] and [4, 4.9]). Let G be a coexact compact quantum group. Let a : A ^ Cr(G) < A be an action on a separable C *-algebra with the spatial Rokhlin property. Then the inclusion Aa c—> A is injective in K-theory, and its image coincides with the subgroup

K* (Aa) = {x G K* (A) | K* (a)(x) = K* (1 < idA)(x)} ■

Proof. If x G im(K*(Aa) ^ K*(A)), then clearly K*(a)(x) = K*(1 < idA)(x). For the converse, let x = [p] — [1^] be an element of K0(A), where p G Mn(A) and G Mk(A) C Mn(A) for some k < n such that p — 1k G Mn(A). Let us write

Mn(a) : Mn(A) ^ Mn((C(G) © A)~)

for the canonical extension of a to unitarizations and matrix amplifications.

Similarly, we write Mn((1©idA)~) for the extension of 1©idA. If x satisfies K0(a)(x) = K0(1 © idA)(x) then

[Mn(a)(p)] — [Mn(a)(1fc)] = [Mn((1 © idA)~)(p)] — [Mn((1 © idA)~)(1fc)]

in K0(Cr(G) © A). Notice that Mn(4)(1k) = Mn((1 © idA)~)(1k) by definition of a, so that we get

[Mn(a)(p)] = [Mn((1 © idA)~)(p)]. By definition of K0, we therefore find natural numbers m, l such that

Mn(a)(p) e 1m e 0i -mvn Mn((1 © idA)~)(p) e 1™ e 0i

in Mn+™+i((Cr(G) © A)~). Using the equivariant isomorphism

T = T-1 : (Cr(G) © A, A © idA) ^ (Cr(G) M A, A M a)

from Lemma 1.24, we can view this as a relation in Mn+™+l((G(G) M A)~). More precisely, using that a(a) and 1 © a in Cr(G) © A for a G A correspond to the elements 1 M a and T(1 © a) in C(G) M A, respectively, we get

Mn((1 m idAr)(p) e 1™ e 0i -mvn Mn(T) o Mn((1 © idA)~)(p) e 1™ e 0i

in Mn+™+l((Cr(G) M A)~). Write n + m + l = r. Since a has the Rokhlin property, let

tf : (Cr(G) M A, A M a) ^ ^a«,)

be an equivariant *-homomorphism with ^'(1 M a) = a for all a G A. We then consider

^ = Mr(i>>) : Mr((Cr(G) M A)~) ^ Mr(A)to,

which is an approximate left-inverse for Mr((1 M idA)~). Then

p e 1™ e 0i ~myn ^ o Mn(T) o Mn((1 © idAD(p) e 1™ e 0i

in Mr(A)TO. Note that Mn((1 © idA)~)(p) is contained in the invariant part of Mn((Cr(G) © A)~). That is,

Mn((A © idA)~) o Mn((1 © idA)~)(p) = Mn((1 © 1 © idA)~)(p)

48 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

in Mn((Cr(G) < Cr(G) < A)~). By equivariance of T and the same applies to

= ^ o Mn(T) o Mn((1 < idA)^)(p),

that is, the latter element satisfies

Mn(a^)(q^) = Mn((1 < idA^)~)(qTO).

Now the invariant part of Mn(A)TO equals Mn((Aaby Lemma 3.5. Since the relation of being a partial isometry with a fixed range projection is well-known to be weakly stable, this shows that there exists a projection q G Mr((Aa)~) such that

p © 1m © 0; ~MvN q

in Mr(A). Hence

x = [p] — [1fc] = [p © 1m] — [1m+k] = [q] — [1m+k]

is contained in im(K0(A a) ^ K0(A)) as desired.

For the statement about the K1-group, one uses suspension to reduce matters to K0, see the proof of [4, 4.9]. □

6. Examples

In this final section we present some examples of Rokhlin actions.

Example 6.1. Let G be a coamenable compact quantum group acting on A = C(G) by the regular coaction a = A. Then a has the spatial Rokhlin property.

Indeed, in this case the embedding ia : A ^ C(G)IA = C(G)<A is given by ia = A. Since G is coamenable, the counit e : O(G) ^ C extends continuously to C(G) = Cr(G), and id<e is an equivariant left-inverse for la. Hence composition with the canonical embedding of C(G) = A into yields an equivariant approximate left-inverse.

The *-homomorphism k : C(G) ^ corresponding to this Rokhlin action according to Proposition 4.3 is induced by the canonical embedding of C(G) into its sequence algebra.

Remark 6.2. Let G be a finite quantum group and a : A ^ C(G) < A an action on a separable, unital C * -algebra. In [21], Kodaka-Teruya introduce and study the Rokhlin property and approximate representability in this setting; in fact they also allow for twisted actions in their paper. It follows from Proposition 4.8 that a is spatially approximately representable in the sense of Definition 4.7 if and only if it is approximately representable in the sense of Kodaka-Teruya [21, Section 4]. As a consequence, Theorem 4.12 shows that a has the spatial Rokhlin property in the sense of Definition 4.1, if

and only if it has the Rokhlin property in the sense of Kodaka-Teruya [21, Section 5]. In particular, our definitions recover Kodaka-Teruya's notions of the Rokhlin property and approximate representability and extend them to the non-unital setting. A substantial difference between our approach and [21] is that the duality of these two notions becomes a theorem rather than a definition.

Example 6.3. Let G be a finite quantum group of order n = dim(C(G)). Then B = Mn = Mn(C) = K(L2(G)) is a G-YD-C * -algebra with the coactions 3 : B ^ C(G) < B, y : B ^ C* (G) << B given by

3(T ) = w * (1 < t )W, y (t ) = W *(1 < t )w *,

respectively.

Let us write Bmk = B H • • • H B for the k-fold braided tensor product. Note that the embeddings M^k ^ Mn H M^k = M^k+1 given by T ^ 1 H T are G-equivariant. As explained in Remark 1.16, we may therefore form the inductive limit action a : A ^ C(G) <g> A of G on the corresponding inductive limit A.

We remark that A can be identified with the UHF-algebra Mn^. Indeed, the braided tensor product Mn K Mn is easily seen to be isomorphic to the ordinary tensor product Mn < Mn as a C * -algebra, using that

Mn H Mn = [W*2(1 < Mn < 1)WW123(Mn)13]

^ [(Mn < 1 < 1)(id <3)3(Mn)] = [(Mn < 1)3(Mn)] = Mn < Mn ■

An analogous statement holds for iterated braided tensor products.

We obtain a G-equivariant *-homomorphism k : C(G) ^ by setting k(/) = [(ik(/))kew], where Lk : Mn ^ A is the embedding into the k-th braided tensor factor of A. Moreover, for a G im(Mn) C A and / G C(G) we have

a1k (/) = a(-2)1k (/)S(a(-1))a(0)

provided k > m. It follows that k satisfies the commutation relations required by Proposition 4.3, and the norm condition in Proposition 4.3 b) is automatic since G is a finite quantum group. Hence (A, a) has the spatial Rokhlin property.

Proposition 6.4. Let G be a coexact compact quantum group and D a strongly self-absorbing C *-algebra. Then there exists at most one conjugacy class of G-actions on D with the spatial Rokhlin property.

Proof. By [32, Corollary 1.12], any two unital *-homomorphisms from D to Cr(G)<D are approximately unitarily equivalent. Therefore the claim follows from Theorem 5.10. □

50 S. Barlak et al. / Journal of Functional Analysis ••• (••••') —

Remark 6.5. As a consequence of Proposition 6.4 we see that the action of a finite quantum group G of order n = dim(C(G)) constructed in Example 6.3 is the unique Rokhlin action of G on Mn& up to conjugacy. In particular, it is conjugate to the action constructed by Kodaka-Teruya in [21, Section 7].

Finally, we shall construct a Rokhlin action of any coamenable compact quantum group on 02. As a preparation, recall that an element a in a C * -algebra A is called full if the closed two-sided ideal generated by a is equal to A.

Lemma 6.6. Let G be a coamenable compact quantum group and let i : C(G) ^02 be a unital embedding. Then (id ©i) o A(f) is a full element in C(G) © 02 for any nonzero f G C(G).

Proof. Let f G C(G) = Cr(G) be nonzero. To show that h = (id©i)(A(f)) is full it is enough to verify that (n © id)(h) is nonzero for all irreducible representations n of C(G). Indeed, if the ideal generated by h is proper, there must exist a primitive ideal of C(G) © 02 containing h. Since 02 is nuclear and simple these ideals are of the form I ©02 for primitive ideals I C C(G), see [5, Theorem 3.3]. Now if n : C(G) ^ L(Hn) is any ^representation, then

(n © id)(A(f)) = (n © id)(W)*(1 © f )(n © id)(W) is nonzero in L(Hn) © C(G), and hence

(id©i)((n © id)(A(f))) = (n © i)(A(f)) = (n © id)(h) is nonzero as well since i is injective. □

Theorem 6.7. Let G be a coamenable compact quantum group. Then up to conjugacy, there exists a unique G-action on the Cuntz algebra 02 with the spatial Rokhlin property.

Proof. According to Proposition 6.4 it suffices to construct some G-action on 02 with the spatial Rokhlin property. Since Cr(G) = C(G) is nuclear, hence in particular exact, there exists a unital embedding i : C(G) 5—> 02. For every n > 0, consider the unital *-homomorphism $„ : C(G) © 0fn ^ C(G) © 0f"+1 given by

$n(x © y) = ((id ©i) o A)(x) © y for all x G C(G), y G 0f". Notice that $„ = $0 © for all n > 1. We have

(A © ido2) o $0 = (A © ido2) o (id ©i) o A = (id © id ©i) o (A © id) o A = (id © id ©i) o (id ©A) o A = (id ©$0) o A.

This means that $0 is an injective equivariant *-homomorphism $0 : (C(G), A) ^ (C(G) ©02, A © ido2). We thus also have that each

$„ : (C(G) ©0fn, A © idO®„) ^ (C(G) ©0®n+1, A © idO®„+x) is injective and equivariant. Define the inductive limit

(A, a) = lim | (C(G) ©0fn, A © idO®„), $n } ,

where a denotes the inductive limit coaction, compare Remark 1.16.

Notice that each building block in this inductive limit has the Rokhlin property, and moreover the first-factor embedding of C(G) into C(G) © 0®n satisfies the required conditions from Proposition 4.3 b) on the nose; see Example 6.1. Similarly to what happens in Example 6.3, the sequence

Kn = $n,TO o (id ©1O|m) : C(G) ^ A

yields a *-homomorphism k : C(G) ^ ATO satisfying the conditions in Proposition 4.3 b). Hence a has the Rokhlin property.

By Lemma 6.6, we know that $0, and thus also each $n is a full *-homomorphism. It follows that the inductive limit A is simple, and it is clearly separable, unital and nuclear. Moreover A is 02-absorbing by [32, 3.4]. This implies A = 02 due to Kirchberg-Phillips [18]. □

References

[1] S. Baaj, G. Skandalis, C*-algèbres de Hopf et théorie de Kasparov équivariante, K-Theory 2 (6) (1989) 683-721, http://dx.doi.org/10.1007/BF00538428.

[2] S. Baaj, G. Skandalis, Unitaires multiplicatifs et dualité pour les produits croisés de C*-algèbres, Ann. Sci. Éc. Norm. Supér. (4) 26 (4) (1993) 425-488, http://www.numdam.org/item?id=ASENS_ 1993_4_26_4_425_0.

[3] S. Baaj, G. Skandalis, S. Vaes, Non-semi-regular quantum groups coming from number theory, Comm. Math. Phys. 235 (1) (2003) 139-167, http://dx.doi.org/10.1007/s00220-002-0780-6.

[4] S. Barlak, G. Szabo, Sequentially split *-homomorphisms between C*-algebras, Int. J. Math 27 (12) (2016), 48 pp., http://dx.doi.org/10.1142/S0129167X16501056.

52 S. Barlak et al. / Journal of Functional Analysis ••• (••••')

[5] B.E. Blackadar, Infinite tensor products of C'-algebras, Pacific J. Math. 72 (2) (1977) 313-334.

[6] S. Echterhoff, S. Kaliszewski, J. Quigg, I. Raeburn, A Categorical Approach to Imprimitivity Theorems for C*-Dynamical Systems, Mem. Amer. Math. Soc., vol. 180, American Mathematical Society, 2006.

[7] D.A. Ellwood, A new characterisation of principal actions, J. Funct. Anal. 173 (1) (2000) 49-60, http://dx.doi.org/10.1006/jfan.2000.3561.

[8] E. Gardella, Compact group actions on C'-algebras: classification, non-classifiability, and crossed products, and rigidity results for Lp-operator algebras, Ph.D. thesis, University of Oregon, 2015.

[9] E. Gardella, Crossed products by compact group actions with the Rokhlin property, J. Noncommut. Geom. (2016), in press, http://arxiv.org/abs/1408.1946.

10] E. Gardella, L. Santiago, Classification of actions of compact groups with the Rokhlin property, 2016, in preparation.

11] E. Gardella, L. Santiago, Equivariant *-homomorphisms, Rokhlin constraints and equivariant UHFabsorption, J. Funct. Anal. 270 (7) (2016) 2543-2590.

12] R. Herman, V. Jones, Period two automorphisms of UHF C*-algebras, J. Funct. Anal. 45 (2) (1982) 169-176.

13] R. Herman, A. Ocneanu, Stability for integer actions on UHF C*-algebras, J. Funct. Anal. 59 (1984) 132-144.

14] I. Hirshberg, W. Winter, Rokhlin actions and self-absorbing C*-algebras, Pacific J. Math. 233 (1) (2007) 125-143.

15] M. Izumi, Finite group actions on C*-algebras with the Rohlin property I, Duke Math. J. 122 (2) (2004) 233-280.

16] M. Izumi, Finite group actions on C*-algebras with the Rohlin property II, Adv. Math. 184 (1) (2004) 119-160.

17] E. Kirchberg, Central sequences in C*-algebras and strongly purely infinite algebras, in: Operator Algebras: The Abel Symposium 2004, vol. 1, Springer, Berlin, 2006, pp. 175-231.

18] E. Kirchberg, N.C. Phillips, Embedding of exact C*-algebras in the Cuntz algebra O2, J. Reine Angew. Math. 525 (2000) 17-53.

19] A. Kishimoto, Rohlin property for flows, in: Advances in Quantum Dynamics, South Hadley, MA, 2002, in: Contemp. Math., vol. 335, Amer. Math. Soc., Providence, RI, 2003, pp. 195-207.

20] A. Kishimoto, A Rohlin property for one-parameter automorphism groups, Comm. Math. Phys. 179 (1996) 599-622.

21] K. Kodaka, T. Teruya, The Rohlin property for coactions of finite dimensional C*-Hopf algebras on unital C'-algebras, J. Operator Theory 74 (2) (2015) 329-369.

22] J. Kustermans, S. Vaes, Locally compact quantum groups, Ann. Sci. Éc. Norm. Supér. (4) 33 (6) (2000) 837-934, http://dx.doi.org/10.1016/S0012-9593(00)01055-7.

23] J. Kustermans, L. Vainerman, A. Van Daele, S.L. Woronowicz, Locally compact quantum groups, in: Conference on Noncommutative Geometry and Quantum Groups, Lecture Notes, in: Contemp. Math., vol. 335, 2001.

24] A. Maes, A. Van Daele, Notes on compact quantum groups, Nieuw Arch. Wiskd. (4) 16 (1-2) (1998) 73-112.

25] N. Nawata, Finite group actions on certain stably projectionless C'-algebras with the Rohlin property, Trans. Amer. Math. Soc. 368 (2016) 471-493.

26] R. Nest, C. Voigt, Equivariant Poincaré duality for quantum group actions, J. Funct. Anal. 258 (5) (2010) 1466-1503, http://dx.doi.org/10.1016/jjfa.2009.10.015.

27] H. Osaka, N.C. Phillips, Crossed products by finite group actions with the Rokhlin property, Math. Z. 270 (1-2) (2012) 19-42.

28] P. Podles, Symmetries of quantum spaces. Subgroups and quotient spaces of quantum SU(2) and SO(3) groups, Comm. Math. Phys. 170 (1) (1995) 1-20, http://projecteuclid.org/ getRecord?id=euclid.cmp/1104272946.

29] M. R0rdam, E. Stormer, Classification of Nuclear C*-Algebras. Entropy in Operator Algebras, Encyclopaedia of Mathematical Sciences, Springer, 2001.

30] L. Santiago, Crossed products by actions of finite groups with the Rokhlin property, Int. J. Math. 26 (7) (2015).

31] T. Timmermann, An Invitation to Quantum Groups and Duality: From Hopf Algebras to Multiplicative Unitaries and Beyond, EMS Textbooks in Mathematics, European Mathematical Society (EMS), Zürich, 2008.

[32] A.S. Toms, W. Winter, Strongly self-absorbing C*-algebras, Trans. Amer. Math. Soc. 359 (8) (2007) 3999-4029.

[33] S. Vaes, R. Vergnioux, The boundary of universal discrete quantum groups, exactness, and factori-ality, Duke Math. J. 140 (1) (2007) 35-84, http://dx.doi.org/10.1215/S0012-7094-07-14012-2.

[34] S.L. Woronowicz, Compact quantum groups, in: Symétries Quantiques, Les Houches, 1995, North-Holland, Amsterdam, 1998, pp. 845-884.