Vous êtes sur la page 1sur 43

# AN INTRODUCTION TO II1 FACTORS

CYRIL HOUDAYER

Contents
1. Finite von Neumann algebras
1.1. Basics on von Neumann algebras
1.2. Finite von Neumann algebras
1.3. Orbit equivalence relations
2. Hilbert bimodules. Completely positive maps
2.1. Generalities
2.2. Dictionary between Hilbert bimodules and u.c.p. maps
2.3. Popas intertwining techniques
3. Approximation properties
3.1. Amenability
3.2. Property Gamma
3.3. Haagerup property
4. Rigidity of II1 factors
4.1. Rigid inclusions of von Neumann algebras
4.2. Applications to rigidity of II1 factors
4.3. Uniqueness of Cartan subalgebras
Appendix A. Polar decomposition of a vector
Appendix B. Von Neumanns dimension theory
Appendix C. Ultraproducts
Appendix D. On the geometry of two projections
References

2
2
3
6
13
13
15
17
22
22
27
30
32
32
34
37
39
39
40
41
42

## 2000 Mathematics Subject Classification. 46L07; 46L10; 46L54.

Key words and phrases. Finite von Neumann algebras; Equivalence relations; Hilbert bimodules; Amenability; Haagerup property; Relative property (T); Cartan subalgebras.
Research partially supported by ANR grant AGORA.
1

CYRIL HOUDAYER

## 1. Finite von Neumann algebras

1.1. Basics on von Neumann algebras. Let H be a separable complex Hilbert
space. We shall denote by h, i the inner product on H that we assumed to be
linear in the first variable (and conjugate linear in the second one). Let B(H) be
the algebra of all bounded linear maps T : H H. This is a Banach algebra for
the uniform norm:
kT k = sup kT k.
kk1

## We moreover have kST k kSk kT k , S, T B(H). The algebra B(H) is

naturally endowed with -operation called the adjonction defined as follows:
hT , i = h, T i, , H.
We have (T ) = T , kT k = kT k and
kT T k = kT T k = kT k2 .
Thus, B(H) is a C -algebra. We can define several weaker topologies on B(H) as
well, in the following way. Let (Ti ) be a net of operators in B(H).
Topology
Ti 0
norm
kT
k
0
i

P
2

2
ultra--strong
(kT

k
+
kT

k
)
0, (n ) `2 H
i
i
n
i
n
2
2
-strong
kTi kP+ kTi k 0, H
2
ultrastrong
0, (n ) `2 H
n kTi n k
strong
kTi k 0, H
P
2
ultraweak
hT

n i n , n i 0, (n ), (n ) ` H
weak
hTi , i 0, , H
For a non-empty subset S B(H), define the commutant of S in B(H) by
S 0 := {T B(H) : xT = T x, x S}
One can then define inductively S 00 = (S 0 )0 , S (3) = (S 00 )0 , S (k+1) = (S (k) )0 , for all
k 1. It is easy to see that
S

S 00

(k)

S (k+2) , k 1.

## Definition 1.1. Let M B(H) be a unital -subalgebra. We say that M is a von

Neumann algebra if M 00 = M .
Theorem 1.2 (Von Neumanns Bicommutant Theorem). Let M B(H) be a
unital -subalgebra. The following are equivalent:
(1) M 00 = M .
(2) M is strongly closed.
(3) M is weakly closed.
Proof. We only sketch the proof.
(1) = (2) is clear since commutants are always strongly closed.
(2) = (1). Let x M 00 . Let
V(x, 1 , . . . , n , ) := {y B(H) : kxi yi k < , i = 1, . . . , n}
be a strong neighborhood of x in B(H). Let K = `2n H and observe that B(K) =
Mn (C) B(H). Let = (1 , . . . , n ) K. Define V = (1 M ) K. Since

## M is strongly closed, V is a closed subspace of K. Denote by PV B(K) the

corresponding orthogonal projection. Since (1 a)PV = PV (1 a), a M , it
follows that 1 x commutes with PV , since x M 00 . Thus (1 x) V and we can
find y M such that (1 x) = (1 y), so in particular y N (x, 1 , . . . , n , ).
Then M 00 is contained in the strong closure of M and hence M = M 00 .
The fact that (2) and (3) are equivalent follows from Hahn-Banach Separation
Theorem (M is convex since it is a vector subspace!).

1.2. Finite von Neumann algebras. A von Neumann algebra M is said to be
finite, if every isometry v M is a unitary, i.e.
v v = 1 = vv = 1, v M.
One can show that a von Neumann algebra is finite if and only if it has a faithful
normal tracial state : M C:
is a positive linear functional with (1) = 1.
is faithful, i.e. x M , (x x) = 0 = x = 0.
is normal, i.e. is weakly continuous on (M )1 , the unit ball of M with
respect to the uniform norm k k .
is a trace, i.e. x, y M , (xy) = (yx).
An infinite dimensional finite von Neumann algebra M with trivial center, i.e.
M 0 M = C, is called a II1 factor.
The most simple examples of finite von Neumann algebras are the following:
(1) Abelian von Neumann algebra. Let (X, ) be a standard probability
space. Represent L (X, ) on the Hilbert space L2 (X, ) by multiplication
(f )(x) = f (x)(x), f L (x, ), L2 (X, ).
The von Neumann algebra M = L (X, ) comes equippedR with the trace given
by integration against the probability measure , i.e. = d.
(2) Group von Neumann algebra. Let be a countable discrete group. The
left regular representation : U(`2 ()) is defined as follows
s t = st .
The von Neumann algebra of is then defined by
L() = {s : s }00 .
The canonical trace on L() is = he , e i. One checks that L() is a II1 factor
if and only if the group has infinite conjugacy classes (icc), that is, t 6= e, the
set {sts1 : s } is infinite.
Exercise 1.3. Let T = [Tst ]s,t B(`2 ()), with Tst = hT t , s i. Show that
T L() if and only if T is constant down the diagonals, i.e. Tst = Txy whenever
st1 = xy 1 .
b is a second countable compact
Assume that is abelian. Then the dual
2
2 b
abelian group. Write F : ` () L () for the Fourier transform which is defined
by F(s ) = 7 hs, i. We get a canonical identification
b = FL()F .
L ()
(3) Group measure space construction. Let y (X, ) be a probability measure preserving (p.m.p.) action. Define an action : y L (X) by

CYRIL HOUDAYER

## (s (F ))(x) = F (s1 x), F L (X). We still denote by

: U(L2 (X))
the corresponding Koopman representation. We regard L (X) = L (X) 1
B(L2 (X) `2 ()). The unitaries us = s s U(L2 (X) `2 ()), for s ,
satisfy the following covariance relation:
us F us = s (F ).
Observe that by Fells absorption principle, the unitary representation (us )s is
simply a multiple of the left regular representation. The crossed product von Neumann algebra is then defined by
(
)00
X

L (X) o =
B(L2 (X) `2 ()).
as us : as L (X)
finite

Z
X
(
as us ) =
ae d.
X

## Recall that the action is said to be free if

({x X : sx = x}) = 0, e 6= s .
It is moreover said to be ergodic if
A = A = (A)(1 (A)) = 0, A X.
One can check that the action is free if and only if L (X) is maximal abelian in
L (X) o . In that case, we say that L (X) is a Cartan subalgebra, i.e. L (X)
L (X) o is maximal abelian and regular. Moreover, L (X) o is a II1 factor
if and only if the action is ergodic.
Observe that when the probability space X = {pt} is a point, then the group von
Neumann algebra and the group measure space construction coincide, i.e. L (X)o
= L().
Let M be a finite von Neumann algebra and fix a faithful normal trace. We
endow M with the following sesquilinear form
hx, yi = (y x), x, y M.
Denote by L2 (M, ) or simply by L2 (M ) the completion of M p
with respect to
h, i . The corresponding k k2 -norm on M is defined by kxk2 = (x x). Write
M 3 x xb
1=x
b L2 (M ) for the natural embedding. Note that the unit vector b
1
is cyclic (i.e. M b
1 is dense in L2 (M )) and separating (i.e. xb
1 = 0 = x = 0) for M .
For every x, y M ,
kxyk22

= (y x xy)

(y kx xk y)

kxk2 kyk22 ,

## so that we can represent M in a standard way on L2 (M ) by

(x)b
y=x
cy, x, y M.
This is the so-called GNS-representation. Observe that : M B(L2 (M )) is a
normal -representation and k(x)k = kxk . Abusing notation, we identify (x)

## with x M and regard M B(L2 (M )). Let J : L2 (M ) 3 xb

1
7 x b
1 L2 (M ) be
the canonical antiunitary.
Theorem 1.4. JM J = M 0 .
Proof. We first prove JM J M 0 . Let x, y, a, b M . We have
y , bb i = (bxa y ) = (y bxa )
\
hJxJyb
a, bbi = hxa
yi = hxJb
bi = hyJxJb
d
= hx
a , bc
a, J yc
a, bbi,

## so that JxJy = yJxJ.

Claim. The faithful normal state x 7 hxb
1, b
1i is a trace on M 0 .
Let x M 0 . We first show that Jxb
1 = x b
1. Indeed, for every a M , we have
1, b
1i
hJxb
1, ab
1i = hJab
1, xb
1i = hx ab
b b
b b
= ha x 1, 1i = hx 1, a1i.
Let now x, y M 0 . We have
1i = hyb
1, Jxb
1i = hxb
1, Jyb
1i
1, b
1i = hyb
1, xb
hxyb

= hxb
1, y b
1i = hyxb
1, b
1i.
Denote the trace x 7 hxb
1, b
1i on M 0 by 0 . Define the canonical antiunitary
2
2
0 0
K on L (M , ) = M 0b
1 = L (M ) by Kxb
1 = x b
1, x M 0 . The first part of
1, which is
the proof yields KM 0 K M 00 = M . Since K and J coincide on M 0b
dense in L2 (M ), it follows that K = J. Therefore, we have JM 0 J M and so
JM J = M 0 .

This Theorem shows in particular that the commutant of the (left) group von
Neumann algebra L() inside B(`2 ()) is the right von Neumann algebra R(),
that is, the von Neumann algebra generated by the right regular representation of
the group .
Exercise 1.5. Show that the strong operator topology on (M )1 is given by the
norm k k2 . Thus, strong and -strong topologies coincide on (M )1 since kxk2 =
kx k2 , x M .
Let B M be a von Neumann subalgebra. One can show that there exists
a unique -preserving faithful normal conditional expectation EB : M B (see
Section 2 for details)1. The map EB : M B is unital completely positive and
moreover satisfies
EB (b1 xb2 ) = b1 EB (x)b2 , x M, b1 , b2 B.
We say that EB is B-B bimodular. If we denote by eB : L2 (M ) L2 (B) the
orthogonal projection, we have eB (xb
1) = EB (x)b
1, for every x M .
Proposition 1.6 (Fourier coefficients). Let y (X, ) be a p.m.p. action. let A =
L (X) and M = L (X) o . Every x M has a unique Fourier decomposition
X
x=
xs us ,
s
1These notes are nonlinear!

CYRIL HOUDAYER

2
with
A (xus ). The convergence holds for the k k2 -norm. Moreover, kxk2 =
P xs = E
2
s kxs k2 .

## Proof. Define the unitary U : L2 (M ) L2 (X) `2 () by the formula

!
X
X
U
as us =
as s .
finite

finite

Then U b
1U = 1X e is a cyclic separating vector for M represented on the Hilbert
2
space L (X)`2 (). Abusing notation, we shall identify L2 (M ) with L2 (X)`2 ().
Under this identification eA is the orthogonal projection L2 (X) `2 () L2 (X)
Ce . Moreover
us eA us is the orthogonal projection L2 (X) `2 () L2 (X) Cs
P
and thus s us eA us = 1. Let x M . Regarding x(1X e ) L2 (X) `2 (),
we know that there exists as L2 (X) such that
X
X
x(1X e ) =
as s and kxk22 =
kas k22 .
s

Then we have
as s

= us eA us x(1X e )
= us eA us xeA (1X e )
= us EA (us x)(1X e )
= EA (xus )(1X s ).

P
It follows that as = EA (xus ). Therefore, we have x = s EA (xus )us and the
P
convergence holds for the k k2 -norm. Moreover, kxk22 = s kEA (xus )k22 .

Exercise 1.7. Let y (X, ) be a p.m.p. action. Let A = L (X) and M =
L (X) o .
(1) Show that y X is free if and only if A = A0 M (A is maximal abelian).
(2) Under the assumption that y X is free, show that M is a II1 factor if
and only if y X is ergodic.
(3) Assume that is icc. Show that M is a II1 factor if and only if y X is
ergodic.
Exercise 1.8. A von Neumann algebra M is diffuse if it has no minimal projection.
(1) Let N M be an inclusion of von Neumann algebras and let e N be a
projection. Show that e(N 0 M )e = (eN e)0 eM e.
(2) Let M be a finite von Neumann algebra. Show that M is diffuse if and only
if there exists a sequence of unitaries un U(M ) such that un 0 weakly.
For more on C -algebras and finite von Neumann algebras, we refer to the excellent book by Brown and Ozawa [2].
1.3. Orbit equivalence relations.
2The convergence does not hold in the strong topology.

## AN INTRODUCTION TO II1 FACTORS

1.3.1. Basic facts on measured equivalence relations. In this note, (X, ) will denote a nonatomic standard Borel probability space. A countable Borel equivalence
relation R is an equivalence relation defined on the space X X which satisfies:
(1) R X X is a Borel subset.
(2) For any x X, the class or orbit of x denoted by [x]R := {y X : (x, y)
R} is countable.
We shall denote by [R] the full group of the equivalence relation R, i.e. [R]
consists of all Borel isomorphisms : X X such that graph() R. The set of
all partial Borel isomorphisms : dom() range() such that graph() R will
be denoted by [[R]]. If is a countable group and (g, x) gx is a Borel action of
on X, then the equivalence relation given by
(x, y) R( y X) g , y = gx
is a countable Borel equivalence relation on X. Conversely, we have the following:
Theorem 1.9 (Feldman & Moore, [11]). If R is a countable Borel equivalence
relation on X, then there exist a countable group and a Borel action of on X
such that R = R( y X). Moreover, and the action can be chosen such that
(x, y) R g , g 2 = 1 and y = gx.
Given a countable Borel equivalence relation R on X, we say that is Rinvariant if
= , [R],
where (U) = (1 (U)), for any Borel U X. The following proposition is
useful and easy to prove:
Proposition 1.10. Let R be a countable Borel equivalence relation defined on X.
The following are equivalent:
(1) is R-invariant.
(2) is -invariant whenever is a countable group acting in a Borel way on
X such that R = R( y X).
(3) is -invariant for some countable group acting in a Borel way on X
such that R = R( y X).
(4) [[R]], (dom()) = (range()).
For any U X, we define the R-saturation of U by
[
[U]R =
[x]R
xU

{y X : x U, (x, y) R}.

We have U [U]R and [U]R is a Borel subset of X. We say that U X is Rinvariant if [U]R = U (up to null sets). The equivalence relation R is said to be
ergodic if any R-invariant Borel subset U X is null or co-null.
Exercise 1.11. Let R be a measured equivalence relation for which (almost) every
orbit is infinite.
Show that there exists a sequence (gn ) in [R] such that g0 = IdX
F
and R = n graph(gn ). In other words, we can write the equivalence relation R as
a countable disjoint union of graphs of elements in the full group [R].

CYRIL HOUDAYER

## Important Convention. In the rest of this paper, when we write an equivalence

relation R defined on (X, ), we always mean a countable Borel equivalence relation
such that the measure is R-invariant. When we write U X, we always mean a
Borel subset of X. From now on, we will neglect null sets, i.e. whenever a property
is true for every x X, we mean for -almost every x X. From now on, we will
always assume that (almost) every orbit of R is infinite, that is, R is a type II1
equivalence relation.
We define now a Borel measure on R. For W R a Borel subset, we define
Wx = {y X : (x, y) W} and W y = {x X : (x, y) W}. We define on R by
Z
(W) =
|Wx |d(x), W R.
X

## Lemma 1.12. Since is assumed to be R-invariant, we have:

Z
Z
|Wx |d(x) =
|W y |d(y), W R.
X

F
Proof. From Exercise 1.11, we know thatFR = n graph(gn ), for some gn [R]. Let
W R be a Borel subset. Then W = n (graph(gn ) W) and graph(gn ) W =
graph(n ) for n [[R]]. Thus, we can write W as a countable disjoint union
of graphs of [[R]]. Consequently, we just have to prove the equality when
W = graph(), for [[R]]. For any : dom() range() [[R]], we have
Z
(dom()) =
| graph()x |d(x)
ZX
(range()) =
| graph()y |d(y).
X

## We shall denote by D := {(x, x) : x X} R the diagonal. We have (D) = 1.

For i = 1, 2 let Ri be a countable Borel equivalence relation on (Xi , i ), and
assumed that i is Ri -invariant. We say that R1 and R2 are isomorphic and
denote R1 ' R2 if there exists a Borel isomorphism : X1 X2 , such that
1 = 2 and
(x, y) R1 ((x), (y)) R2 .
We will be using the following important fact. For : X1 X2 a Borel isomorphism such that 1 = 2 , we associate : L (X1 ) L (X2 ) defined by
( F )(x) = F (1 x), for any F L (X1 ), and any x X1 . Moreover the map
is an onto isomorphism (see [28, Proposition XIII.1.2]).
1.3.2. Construction of the von Neumann algebra of R. We define the left regular
representation of the equivalence relation R on the Hilbert space L2 (R, ). For
: dom() range() [[R]] and L2 (R, ), set
(u )(x, y) = 1range() (x)(1 (x), y), (x, y) R.
In other words, (u )(x, y) = (1 (x), y) if x range() and 0 otherwise. The
von Neumann algebra L (X) acts in the following way. If F L (X) and
L2 (R, ), we have
(F )(x, y) = F (x)(x, y), (x, y) R.

## For F L (X) and [[R]], define F L (X) by

F (x) = 1range() (x)F (1 (x)).
In other words, F (x) = F (1 (x)) if x range() and 0 otherwise.
Note that, we can also define the right regular representation of R on L2 (R, )
in the following way: for : dom() range() [[R]] and L2 (R, ), set
(v )(x, y) = 1range() (y)(x, 1 (y)), (x, y) R.
Exercise 1.13. Show that for any , [[R]], we have
(1) u u = u .
(2) u = u1 .
(3) u u = 1dom() and u u = 1range() .
(4) u F = F u , for any F L (X).
Definition 1.14 (Feldman & Moore, [12]). The von Neumann algebra L(R)
B(L2 (R, )) of the equivalence relation R is then defined as follows:
L(R) := W {u : [[R]]}.
Likewise, we define R(R) = W {v : [[R]]}. It is trivial to check that
L(R) R(R)0 . Define the canonical anti-unitary J : L2 (R, ) L2 (R, ) by
(J)(x, y) = (y, x), for any L2 (R, ), for any (x, y) R.
Proposition 1.15. Denote by 0 = 1D L2 (R, ) the characteristic function
corresponding to the diagonal D R.
(1) 0 is a cyclic separating vector for L(R).
(2) The vector state = h0 , 0 i is a faithful normal trace on L(R). In particular, L(R) is a finite von Neumann algebra.
(3) For any [[R]], Ju J = v . In particular, L(R) = R(R)0 .
Proof. (1) Write 0 = 1D , where D R is the diagonal. For [[R]], we have
1graph() = u1 0 . Recall that R can Fbe written as a countable disjoint union
of graphs
F of Borel isomorphisms R = n graph(gn ). Take any W R. Then
W = n Wn , with Wn = W graph(gn ). Since Wn is the graph of a partial Borel
isomorphism, it follows that 1W L(R)0 . Consequently, 0 is cyclic for L(R).
Exactly in the same way, 0 is cyclic for R(R), hence separating for L(R).
(2) It suffices to prove that (u u ) = (u u ), for every , [[R]]. Let
, [[R]]. We have
Z
(u u ) =
1{x=()1 (x)} (x)d(x)
ZX
=
1{x=1 1 (x)} (x)d(x)
X
Z
=
1{x0 =1 1 (x0 )} (x0 )d(x0 ) (x0 = 1 (x))
X
Z
=
1{x=()1 (x)} (x)d(x)
X

(u u ).

## (3) A straightforward computation shows that Ju J = v , for any [[R]].

It follows that JL(R)J = R(R). By the general theory of finite von Neumann
algebras (see Proposition 1.4), we get L(R) = R(R)0 .


10

CYRIL HOUDAYER

## Denote A = L (X) and M = L(R). Observe that A M . We know that there

exists a unique -preserving faithful normal conditional expectation EA : M A.
In order to know EA , it is sufficient to compute EA (u ) for any [[R]]. Denote
by eD : L2 (R) L2 (D, 0 ) the orthogonal projection.
Proposition 1.16. We have
(1) EA (ug ) = 1{x=g1 x} , g [R].
(2) eD (x0 ) = EA (x)0 , x L(R).
Proof. In order to prove (1) and (2), it suffices to show that EA (u ) = 1{x=1 (x)} ,
for every [[R]]. Let , [[R]]. Write f = 1{x=1 (x)} L (X). Then, we
have
Z
(EA (u )u ) =
EA (u )(x)1{x=1 (x)} (x)d(x)
X
Z
=
EA (u 1{x=1 (x)} )(x)d(x)
X

=
=
=

## (EA (u 1{x=1 (x)} )) = (u 1{x=1 (x)} )

Z
1{x=1 (x)} (x)f (x)d(x)
X
Z
f (x)1{x=1 (x)} (x)d(x) = (f u ).
X

## Thus, ((EA (u ) f )x) = 0, for any x M . Consequently, EA (u ) = f =

1{x=1 (x)} .

Proposition
1.17. Let (gn ) be a sequence in [R] such that g0 = IdX and R =
F
1
).
Then any x L(R) can be uniquely written
graph(g
n
n
X
x=
an ugn ,
n

where an A.
Proof. Since R =

graph(gn1 ), we have
M
L2 (R, ) =
L2 (graph(gn1 ), n ),
n

## where n is the restriction of to graph(gn1 ). Let x L(R). Define an =

EA (xugn ). Recall that 1graph(gn1 ) = ugn 0 . Denote by eD : L2 (R, ) L2 (D, 0 )
the orthogonal projection. It is easy to check that ugn eD ugn is the orthogonal
projection L2 (R, ) L2 (graph(gn1 ), n ). We have
ugn eD ugn (x0 )

## = ugn (eD ugn x0 )

= ugn EA (ugn x)0
= ugn EA (ugn x)ugn ugn 0
= EA (xugn )ugn 0 .

2
Therefore x0 =
n EA (xugn )ugn 0 in L (R, ) and so x =
where the convergence holds for the k k2 -norm.

EA (xugn )ugn


## The above proposition yields in particular L(R) = (L (X) {ug : g [R]})00 .

Proposition 1.18. Denote M = L(R) and A = L (X). Then

11

## (1) A = A0 M , i.e. A M is a maximal abelian -subalgebra.

(2) NM (A)00 = M , i.e. A M is regular.
P
Proof. (1) Let u U(A0 M ). As before we may write u = n an ugn for some
gn [R]. Fix F L (X). Since uF = F u, we get an (Fgn F ) = 0, for any n.
Thus, for any x supp(an ), F (gn1 (x)) = F (x).
P Using a previous remark, we get
gn1 (x) = x, for any x supp(an ). Thus, u = n an 1supp(an ) A.
(2) It is trivial once we noticed that M = (A {ug : g [R]})00 .

From this proposition, it follows that Z(M ) = M 0 M A0 M = A. Moreover,
for any U X, we have the following
U = [U]R gU = U, g [R].
Indeed, assume that U = [U]R and fix g [R]. For any x U, since (x, gx) R,
then gx U. Conversely, assume that gU = U, for any g [R]. Recall that there
exists a countable group and a p.m.p. action of on X such that R = R( y X).
If (x, y) R, with x U, there exists g such that y = gx. But then y U.
Thus we obtain:
Proposition 1.19. L(R) is a factor if and only if R is ergodic.
Then we summarize what we did so far in the following theorem:
Theorem 1.20. Let (X, ) be a nonatomic probability space. Let R be an ergodic
countable Borel equivalence relation on X such that is R-invariant. Then L(R)
is a II1 factor and L (X) L(R) is a Cartan subalgebra.
1.3.3. The full group of R and consequences. Denote A = L (X) and M = L(R).
We prove the following:
Theorem 1.21. We have
[R] = NM (A)/U(A).
P
Proof. Let u NM (A). As before, we may write u = n an ugn , for some an A,
gn [R]. We know that there exists a Borel isomorphism : X X such that
= and uF u = F , for any F A. Thus, uF = F u and so an (Fgn F ) =
0, for any n and any F A. Hence, for any x supp(an ), for any F A
F (gn1 (x)) = F (1 (x)).
Denote Y =
have

1X\Y an ugn = 0

## and so 1X\Y u = 0. Thus, (X\Y ) = 0. This finally proves that [R].

We have constructed a group morphism
:

NM (A) [R]
u 7

## which is onto since = u . Moreover, u ker() if and only if u A0 M . Thus,

ker() = U(A). This completes the proof.

Corollary 1.22. We have NM (A) = U(A) o [R].

12

CYRIL HOUDAYER

## Proof. We already know that the sequence

1 U(A) NM (A) [R] 1
is exact. It moreover splits with the following section
s:

[R] NM (A)
.
g 7 ug


## Theorem 1.23. For i = 1, 2, let Ri be a measured equivalence relation on (Xi , i ).

Denote Ai = L (Xi ) and Mi = L(Ri ). Then
R1 ' R2 (A1 M1 ) ' (A2 M2 ).
Proof. = First assume that R1 ' R2 . Then there exists a Borel isomorphism
: X1 X2 such that 1 = 2 and for any (x, y) X1 X1 , (x, y) R1 iff
((x), (y)) R2 . We define a unitary as follows:
U:

L2 (R1 , 1 ) L2 (R2 , 2 )

7 (x, y) 7 (1 (x), 1 (y)) .

## For any g [R1 ], F L (X1 ) and any L2 (R2 , 2 ), we have

(U ug U )(x, y)

= (g 1 1 (x), y)

(U F U )(x, y)

= F (1 (x))(x, y).

## Thus, U ug U = ug1 and U F U = F . Consequently, = Ad(U ) : M1 M2

is an onto -isomorphism such that (A1 ) = A2 .
= Assume now that there exists an onto -isomorphism : M1 M2 such
that (A1 ) = A2 . We know that there exists a Borel isomorphism : X1 X2
such that 1 = 2 and (F ) = F , for any F A1 . Fix g [R1 ]. Since ug
normalizes A1 inside M1 , it follows that (ug ) normalizes A2 inside M2 . Thus there
exist h [R2 ] and v U(A2 ) such that (ug ) = uh v.
Fix F A1 . From the one hand, we know that ug F ug = Fg . Thus we obtain
(ug F ug ) = (Fg ) = (Fg ) . On the other hand, we have
(ug F ug )

= uh vF v uh
= uh F uh
=

(F )h .

## Consequently, (Fg ) = (F )h , and so g 1 1 = 1 h1 on X2 . Equivalently,

g = h on X1 . For any x X1 , ((x), g(x)) = ((x), h(x)) R2 , since
h [R2 ].
Let now (x, y) R1 . We know that there exists g [R1 ] such that y = gx.
Thus, ((x), (y)) = ((x), g(x)) R2 . Reasoning exactly the same way with
1 , we obtain that is an isomorphism of equivalence relations.

1.3.4. Group actions and their orbit equivalence relations. Given a p.m.p. action
y (X, ), one can associate the orbit equivalence relation R( y X) defined by
(x, y) R( y X) s , y = sx.
When the action y X is free, the map
( X, counting ) 3 (s, x) 7 (x, sx) (R( y X), )

13

## is a p.m.p. Borel isomorphism.

Exercise 1.24. Let y (X, ) be a free p.m.p. action. Show that the von
Neumann algebra of the orbit equivalence relation L(R( y X)) and the group
measure space construction L (X) o are -isomorphic.
Definition 1.25. Let y (X, ) and y (Y, ) be two free p.m.p. actions. We
shall say that
(1) y (X, ) and y (Y, ) are conjugate if there exist a p.m.p. Borel
isomorphism : (X, ) ' (Y, ) and a group isomorphism : ' such
that (sx) = (s)(x), s , x X.
(2) y (X, ) and y (Y, ) are orbit equivalent (abbreviated OE) if there
exist a p.m.p. Borel isomorphism : (X, ) ' (Y, ) such that (x) =
(x), x X.
(3) y (X, ) and y (Y, ) are W -equivalent (abbreviated W E) if L (X)o
' L (Y ) o .
Let A = L (X) L (X) o = M and B = L (Y ) L (Y ) o = N .
Observe that Theorem 1.23 yields
y (X, ) OE y (Y, )

R( y X) ' R( y Y )
(A M ) ' (B N )

## Therefore the following implications are true

conjugacy = orbit equivalence = W -equivalence.
2. Hilbert bimodules. Completely positive maps
2.1. Generalities.
2.1.1. Hilbert bimodules. The discovery of the appropriate notion of representations
for von Neumann algebras, as so-called correspondences or bimodules, is due to
Connes [5].
Definition 2.1. Let M, N be finite von Neumann algebras. A Hilbert space H
is said to be an M -N -bimodule if it comes equipped with two commuting normal
-representations : M B(H) and : N op B(H). We shall intuitively write
x y = (x)(y op ), H, x M, y N .
We shall see that an M -M bimodule H is the analog of a unitary group representation : U(H ).
Example 2.2. The following are important examples of bimodules:
(1) The identity bimodule L2 (M ) with x y = xy.
(2) The coarse bimodule L2 (M ) L2 (N ) with x ( ) y = (x) (y). It
can be checked that as M -N -bimodules,
L2 (M ) L2 (N ) ' HS(L2 (M ), L2 (N ))
where HS(L2 (M ), L2 (N )) denotes the M -N -bimodule of Hilbert-Schmidt
operators on from L2 (M ) to L2 (N ).
(3) For any -preserving automorphism Aut(M ), we regard L2 (M ) with
the following structure: x y = x(y).

14

CYRIL HOUDAYER

## (4) Let B M be a von Neumann subalgebra and denote by eB : L2 (M )

L2 (B) the orthogonal projection. Consider the basic construction hM, eB i
which is the von Neumann subalgebra of B(L2 (M )) generated by M and eB .
We endow hM, eB i with the following semifinite trace: Tr(xeB y) = (xy),
for all x, y M (see Subsection 2.3). Then L2 (hM, eB i, Tr) is naturally endowed with a structure of M -M -bimodule. Moreover, as M -M -bimodules,
L2 (hM, eB i, Tr) ' L2 (M ) B L2 (M )
where B denotes Connes fusion tensor product (see [5, Appendix V.B]).
2.1.2. Unital completely positive maps. Let (M, M ), (N, N ) be finite von Neumann
algebras endowed with a fixed faithful normal trace. A linear map : M N is
said to be completely positive if the maps
n = Idn : Mn (C) M Mn (C) N
are positive for every n 1. We shall say that is unital if (1) = 1, and tracepreserving if moreover N ((x)) = M (x), for every x M .
Theorem 2.3 (Stinespring dilation). Let : M N be a (normal) u.c.p. map.
Then there exist a Hilbert space H, an isometry V : L2 (N ) H and a (normal)
-representation : M B(H) such that
(x) = V (x)V, x M.
Proof. Equip H0 = M alg L2 (N ) with the following sesquilinear form
X
X
X
h
ai i ,
bj j i =
h(bj ai )i , j iL2 (N )
i

i,j

and
P promote it to a Hilbert space H by separation and completion. Denote by
( j bj j ) the vector in H which it represents. Define now V : L2 (N ) H by
V = (1 ) . It is clear that V is an isometry, i.e. V V = 1L2 (N ) . For every
x M , we define a bounded linear operator (x) on H by
X
X
(x)(
bj j ) = (
xbj j ) .
j

## As expected, : M B(H) is a (normal) -representation such that (x) =

V (x)V , x M .

It follows that a u.c.p. map : M N satisfies for every x M ,
(x x)

V (x x)V

(x) (x).

## If is moreover assumed to be trace-preserving, the operator T : L2 (M ) L2 (N )

defined by
d x M.
T (b
x) = (x),
is bounded and kT k = 1.
Example 2.4. The following are important examples of -preserving u.c.p. maps:
(1) The trace : M M , the identity map Id : M M and more generally
all -automorphisms : M M which preserve the trace.

15

## (2) Let B M be a von Neumann subalgebra. Denote by eB : L2 (M ) L2 (B)

the orthogonal projection. Denote by EB : M B the unique -preserving
conditional expectation from M onto B which satisfies
\
E
x), x M.
B (x) = eB (b
It is easy to see that EB is indeed a u.c.p. map.
(3) Let M = L() be the von Neumann algebra of a countable group . Let
: C be a normalized positive definite function, i.e. for any finite
set F the matrix ((st1 ))s,tF is positive. Define the corresponding
-preserving u.c.p. map : L() L() by
X
X
(
as us ) =
(s)as us .
s

Exercise 2.5. Let : L() L() be a -preverving u.c.p. map. Show that
: C defined by (s) = ((us )us ) is a positive definite function.
2.2. Dictionary between Hilbert bimodules and u.c.p. maps.
2.2.1. From u.c.p. maps to Hilbert bimodules. Let : M N be a trace-preserving
u.c.p. map. Equip H0 = M L2 (N ) with the following sesquilinear form
X
X
X
h
ai i ,
bj j i =
h(bj ai )i , j iL2 (N )
i

i,j

## and promote it to a Hilbert space H by separation and completion. Observe that

H is a Stinespring Dilation of . Abusing notation, denote by b the vector in
H which it represents. The action is given by
a(b )x = (ab) (x),
for a M and x N . With the unit vector = 1 b
1 H , we have
hax, yiH = h(a)b
x, ybiL2 (N )
for every a M , x, y N . Since the u.c.p. map is assumed to be trace-preserving,
we get
h, i = M and h, i = N .
2.2.2. From Hilbert bimodules to u.c.p. maps. Let H be an M -N bimodule, with
a tracial unit vector , i.e. h, i = M and h, i = N . Then the linear operator
L : L2 (N ) H defined by L (b
x) = x is bounded and kL k = 1. For any
x, y N , we have
hb
x, L (y)iL2 (N )

hx, yiH

(xy )

hb
x, ybiL2 (N ) ,

## so that L (y) = yb. Therefore L is an isometry, i.e. L L = 1. Denote by J = JN

the canonical antiunitary. Moreover, for any a M , L aL N . Indeed for any
y, z1 , z2 N , we have
h(L aL )(Jy J)zb1 , zb2 iL2 (N )
h(Jy

haz1 y, z2 iH

## = hxL zb1 , L JyJ zb2 iH

= haz1 , z2 y iH .

16

CYRIL HOUDAYER

## Define the u.c.p. map : M N by (a) = L aL . Since

N ((a)) = hL aL b
1, b
1i = ha, i = M (x),
it follows that is trace-preserving. We moreover have
h(a)b
x, ybiL2 (N ) = hax, yiH ,
We can now prove the uniqueness and existence of the trace-preserving faithful
normal conditional expectation EB : M B.
Proposition 2.6. Let (M, ) be a finite von Neumann algebra with a fixed trace
and let B M be a von Neumann subalgebra such that B = |B . Then there exists
a unique normal faithful trace-preserving conditional expectation EB : M B.
Proof. Consider the M -B-bimodule H = M L2 (M )B . The vector b
1 H is obviously
a unit tracial vector. Denote by EB the corresponding normal trace-preserving
u.c.p. map EB : M B. Recall
hxb
1a, b
1biH = hEB (x)b
a, bbiL2 (B) , x M, a, b B.
Let x M , a, b, c, d B. We have
b L2 (B)
hEB (axb)b
c, di

= haxbb
1c, b
1diH
= hxb
1bc, b
1a diH
di 2
= hEB (x)bc,
L (B)
b
= haEB (x)bb
c, diL2 (B) ,

## hence EB (axb) = aEB (x)b. Assume now that EB (x x) = 0. Then

0 = hEB (x x)b
1, b
1iL2 (B) = hx xb
1, b
1iH = kxk22 ,
and so x = 0. Let E : M B be another trace-preserving conditional expectation.
Then
b L2 (B)
hEB (x)b
c, di

= hxb
1c, b
1diH
= (d xc)
= (E(d xc))
= (d E(x)c)
b L2 (B) ,
= hE(x)b
c, di

hence EB (x) = E(x). Therefore EB : M B is the unique normal faithful tracepreserving conditional expectation.

2.2.3. From unitary group representations to Hilbert bimodules. Let : U(H )
be a unitary representation of a countable discrete group . Let M = L() be the
group von Neumann algebra and denote by (us )s the canonical unitaries. Define
on K = H `2 () the following left and right commuting multiplications: for
every H and every s, t ,
us ( t )

(s s )( t ) = s st

( t ) us

(1H s1 )( t ) = ts .

## AN INTRODUCTION TO II1 FACTORS

17

It is clear that the right multiplication extends to the whole von Neumann algebra
M . Observe now that the unitary representations and 1H are unitarily
conjugate. Indeed, define U : H `2 () H `2 () by
U ( t ) = t t .
It is routine to check that U is a unitary and U (1H s )U = s s , for every
s . Therefore, the left multiplication extends to M . We have proven:
Proposition 2.7. The formulae above endow the Hilbert space K with a structure
of L()-L()-bimodule.
Observe that in the case = is the left regular representation of , the M -M bimodule K is nothing but the coarse bimodule L2 (M ) L2 (M ).
Exercise 2.8. Let : C be a normalized positive definite function. Let
(, H , ), with H unit vector, be the GNS-representation of , i.e., (s) =
hs , i, for all s . Show that the u.c.p. map associated to the bimodule K
satisfies
X
X
(
as us ) =
(s)as us .
s

Exercise 2.9. Prove the following dictionary between u.c.p. maps and Hilbert
bimodules:
u.c.p. maps
Hilbert bimodules
Id : M M
Identity bimodule L2 (M )
:M M
Coarse bimodule L2 (M ) L2 (M )
Automorphism : M M
L2 (M ) with x y = x(y)
EB : M M
L2 hM, eB i
2.3. Popas intertwining techniques.
2.3.1. The basic construction. Throughout this section, we will denote by M a finite
von Neumann algebra with a distinguished faithful normal trace . Let B M be
a unital von Neumann subalgebra. Let eB : L2 (M ) L2 (B) be the orthogonal
projection. We will denote by EB : M B the unique faithful normal -preserving
conditional expectation. It satisfies the following:
\
E
B (x)
EB (axb)

= eB (b
x),
= aEB (x)b, x M, a, b B.

The basic construction hM, eB i is the von Neumann subalgebra of B(L2 (M )) generated by M and the projection eB . Observe that JeB = eB J and eB xeB = EB (x)eB ,
x M .
Proposition 2.10. The following are true.
(1) hM, eB i = JB 0 J B(L2 (M )).
(2) The central support of eB in hM, eB i equals 1. In particular, the -subalgebra
span(M eB M ) is a -strongly dense in hM, eB i. The formula eB xeB =
EB (x)eB extends the conditional expectation EB : hM, eB i B.
(3) hM, eB i is endowed with a semifinite faithful normal trace defined by
Tr(xeB y) = (xy), x, y M
.

18

CYRIL HOUDAYER

Proof. (1) For x B, we clearly have xL2 (B) L2 (B) and xL2 (B) L2 (B) ,
hence xeB = eB x. If x M {eB }0 , then
EB (x)b
1 = eB (xb
1) = xeB (b
1) = xb
1.
Therefore x = EB (x) B. It follows that B = M {eB }0 . Thus,
JB 0 J = hJM 0 J, JeB Ji = hM, eB i.
(2) The map B 3 x 7 xeB BeB is a -isomorphism. Indeed, if xeB = 0,
then x = 0, for every L2 (B). Since x B, it follows that x = 0. Denote by
z(eB ) the central support of eB in B 0 . Then z(eB ) B and z(eB )eB = eB . Hence
z(eB ) = 1. Thus the central support of eB = JeB J in JB 0 J is equal to 1. It is
clear that I = span(M eB M ) is a -subalgebra of hM, eB i and a 2-sided ideal of
the -algebra generated by M and eB . Thus I is a closed 2-sided ideal of hM, eB i.
Moreover
IL2 (M ) = M eB L2 (M ) = M L2 (B) M b
1.
Since I is nondegenerate, we get I = hM, eB i.
(3) Since eB has central support 1 inP
hM, eB i, one can find partial isometries
(vi ) in hM, eB i such that vi vi eB and i vi vi = 1. It follows that
M
vi L2 (B) = L2 (M ).
i

## Define the following normal weight Tr on hM, eB i by

X
Tr(x) =
hxvib
1, vib
1i, x hM, eB i+ .
i

## Assume that Tr(x x) = 0. Then xvib

1 = 0, for every i. For every b B, we have
xvi bb
1 = xvi Jb J b
1 = Jb Jxvib
1 = 0.
Therefore x = 0 and Tr is faithful. For every x, y M , we have
X
X
Tr(xeB y) =
hxeB yvib
1, vib
1i =
heB yvi eB b
1, eB x vi eB b
1i
i

## hEB (yvi )eB b

1, EB (x vi )eB b
1i =

(EB (x vi ) EB (yvi ))

## hEB (vi x)eB b

1, EB (vi y )eB b
1i

heB vi xeB b
1, eB vi y eB b
1i

= h

hvi vi xb
1, y b
1i

vi vi xb
1, y b
1i

= hxb
1, y b
1i = (xy).

## We get that Tr is semifinite since span(M eB M ) is a -strongly dense -subalgebra

in hM, eB i. For every x, y, z, t hM, eB i, we have
Tr(xeB yzeB t)

=

19

## It follows from the previous proposition that

hM, eB i = {T B(L2 (M )) : T (b) = T ()b, L2 (M ), b B}.
Let HB be a right B-submodule of L2 (M )B . Write PH : L2 (M ) H for the
orthogonal projection. It is clear that PH hM, eB i. We define the von Neumann
dimension of HB by dim(HB ) := Tr(PH ).
Exercise 2.11 ([2]). Let (N, Tr) be a semifinite von Neumann algebra. Let =
{x N : kxk2,Tr 1}. Prove that the formal inclusion , L2 (N, Tr) is ultraweakweak continuous.
2.3.2. Intertwining subalgebras. In [23, 19], Popa introduced a very powerful tool
to prove the unitary conjugacy of two von Neumann subalgebras of a tracial von
Neumann algebra (M, ). If A (M, ) is a (possibly non-unital) von Neumann
subalgebras, denote by 1A the unit of A.
Theorem 2.12 (Popa, [23, 19]). Let (M, ) be a finite von Neumann algebra. Let
A M be a possibly non-unital von Neumann subalgebra and B M be a unital
von Neumann subalgebra. The following are equivalent:
(1) There exist n 1, a possibly non-unital -homomorphism : A Mn (C)
B and a non-zero partial isometry v M1,n (C) 1A M such that
xv = v(x), x A.
(2) There exists a nonzero A-B-subbimodule H of

2
A L (1A M )B

such that

dim(HB ) < .
(3) There exists a nonzero element d A0 1A hM, eB i+ 1A such that
Tr(d) < .
(4) There is no sequence of unitaries (uk ) in A such that
lim kEB (a uk b)k2 = 0, a, b 1A M.

## If one of the previous equivalent conditions is satisfied, we shall say that A

embeds into B inside M and denote A M B. For simplicity, we shall write
M n := Mn (C) M .
Proof. We first prove that (1), (2), (3) are equivalent. Then we show (1) = (4)
and (4) = (3).
(1) = (2). Take a nonzero component of v M1,n (C) 1A M that we may
Pn
assume to be v1 . Set w = v1 . We have Aw
j=1 vj B. Define H = AwB.
2
2
Therefore H v(`n L (B)) and dim(HB ) n.
(2) = (3). Write d = PH . Then d hM, eB i+ and Tr(d) = dim(HB ) < .
Since H is moreover a left A-module, we have ad = da, for every a A, hence
d A0 1A hM, eB i+ 1A such that 0 < Tr(d) < .
(3) = (1). Write q = 1[kdk /2,kdk ] (d) for the nonzero spectral projection of
d. We get that K = qL2 (M ) is a nonzero A-B-subbimodule of L2 (1A M ) such that
dim(KB ) = Tr(q) < . Thus, cutting down by a central projection of B (see [29,
Lemma C.1]), we get a nonzero A-B-subbimodule H L2 (1A M ) which is finitely
generated over B. Hence, we can take n 1, a projection p B n and a right
B-module isomorphism
: pL2 (B)n H.

20

CYRIL HOUDAYER

## Since H is a left A-module, we get a (unital) -homomorphism : A pB n p

satisfying x() = ((x)) for all x A, and pL2 (B)n . Define now ej
L2 (B)n as ej = (0, . . . , b
1, . . . , 0) and = (1 , . . . , n ) M1,n (C) H, with j =
(pej ). Let j {1, . . . , n}. For any x A, write (x) = (kl (x))kl pB n p. We
have
n
X
xj = x(pej ) = ((x)pej ) = (p(x)ej ) = (p
ij (x)ei )
i=1

=
=
=
=

n
X
i=1
n
X
i=1
n
X
i=1
n
X

## (p(0, . . . , ij (x), . . . , 0))

((pei )ij (x))
(pei )ij (x) ( is a right B-module isomorphism)
i ij (x).

i=1

## Consequently, for every x A, x = (x). In the von Neumann algebra M n+1

B(L2 (M ) L2 (M )n ), define


x
0
Xx =
, x A.
0 (x)
In the space L2 (M n+1 ), define

=

0
0

.
0

## Note that Xx 1An+1 M n+1 1An+1 , x A, and 1An+1 L2 (M n+1 )1An+1 . We

obtain Xx = Xx , for all x A. Denote by T the corresponding unbounded
operator affiliated with M n+1 . and write T = V |T | for its polar decomposition
(see Appendix A.1). We get Xx V = V Xx , for every x A, and V V 1An+1 .
Write


u v
V =
.
v0 w
It is straightforward to check that v M1,n (C) 1A M is a partial isometry from
ker w onto ker u such that xv = v(x), for every x A.
(1) = (4). By contradiction, assume that there exists a sequence of unitaries
(uk ) be a sequence of unitaries in A such that limk kEB (a uk b)k2 = 0 for all a, b
1A M . Then k(Idn EB )(v uk v)k2 0. But for every k N, v uk v = (uk )v v.
Moreover, (uk ) U(pB n p) and v v p. Thus,
k(Idn EB )(v v)k2

## = k(uk )(Idn EB )(v v)k2

= k(Idn EB )((uk )v v)k2
= k(Idn EB )(v uk v)k2 0.

## We conclude that (Idn EB )(v v) = 0 and so v = 0. Contradiction.

(4) = (3). We can take > 0 and K 1A M finite subset such that
max kEB (a ub)k2 , u U(A).

a,bK

21

Note that
kEB (a ub)k22

=

## Tr(eB (a ub) eB (a ub)eB ).

Define
now the element c = aK aeB a in 1A hM, eB i+ 1A . Note that Tr(c) =
P

## aK (aa ) < . Denote by C the ultraweak closure of the convex hull of {u cu :

2
u U(A)}. Observe that C hM, eB i+ L (hM, eB i) is bounded for both kk and
k k2,Tr and is closed in L2 (hM, eB i). Let d C be the unique element of minimal
k k2,Tr -norm. Since kudu k2,Tr = kdk2,Tr for all u U(A), we get u du = d, and
so d A0 1A hM, eB i+ 1A . We show now that d 6= 0. For all u U(A), we have
X
X
Tr(eB b (u cu)beB ) =
Tr(eB (a ub) eB (a ub)eB )
P

bK

a,bK

## (EB (a ub) EB (a ub))

a,bK

kEB (a ub)k22 2 .

a,bK

Consequently, using the facts that Tr(eB eB ) is a normal state on the basic construction hM, eB i and d C, we get
X
Tr(eB b dbeB ) 2 .
bK

## Assume that M = B o where y B is a trace-preserving action of a countable

group on a finite von Neumann algebra B. Denote by (vs )s the canonical
unitaries in M which implement the action. It is straightforward to see that A M
B if and only if there a sequence of unitaries un U(A) such that
lim kEB (un vs )k2 = 0, s .
n

In the case when A and B are maximal abelian in M , one can get a more precise
result (see [19, Theorem A.1]).
Proposition 2.13 (Popa, [19]). Let (M, ) be a finite von Neumann algebra. Let
A, B M be a maximal abelian von Neumann subalgebras. The following are
equivalent:
(1) A M B.
(2) There exists a nonzero partial isometry v M such that vv A, v v B
and v Av = Bv v.
Proof. We only need to prove (1) = (2). The proof follows the one of [19, Theorem
A.1]. We will use exactly the same reasoning as in the proof of [29, Theorem C.3].
Since A M B, we can find n 1, a nonzero projection q Mn (C) B, a
nonzero partial isometry w M1,n (C) pM and a unital -homomorphism :
A q(Mn (C) B)q such that xw = w(x), x A. Since we can replace q by an
equivalent projection in Mn (C) B, we may assume q = Diagn (q1 , . . . , qn ) (see for
instance second item in [29, Lemma C.2]). Observe now that Diagn (q1 B, . . . , qn B)
is maximal abelian in q(Mn (C)B)q. Since B is abelian, q(Mn (C)B)q is of finite
type I. Since A is abelian, up to unitary conjugacy by a unitary in q(Mn (C) B)q,
we may assume that (A) Diagn (q1 B, . . . , qn B) (see [29, Lemma C.2]). We can

22

CYRIL HOUDAYER

## now cut down and w by one of projections (0, . . . , qi , . . . , 0) and assume n = 1

from the beginning.
Write e = ww A (since A0 pM p = A) and f = w w (A)0 qM q. By
spatiality, we have
f ((A)0 qM q)f = ((A)f )0 f M f = (w Aw)0 f M f = w Aw,
which is abelian. Let Q := (A)0 qM q, which is a finite von Neumann algebra.
Since Bq Q is maximal abelian and f Q is an abelian projection, [29, Lemma
C.2] yields a partial isometry u Q such that uu = f and u Qu Bq. Define
now v = wu. We get
v Av = u w Awu = u f ((A)0 qM q)f u Bq.
Moreover vv = wuu w = wf w = e A. Since v Av and Bv v are both maximal
abelian, we get v Av = Bv v.

We can even go further if we moreover assume that A, B M are both Cartan
subalgebras and M is a II1 factor (see [19, Theorem A.1]).
Theorem 2.14 (Popa, [19]). Let M be a II1 factor. Let A, B M be Cartan
subalgebras. The following are equivalent:
(1) A M B.
(2) There exists u U(M ) such that uAu = B.
Proof. We only need to prove (1) = (2). By Proposition 2.13, there exists a
nonzero partial isometry v M such that vv A, v v B and v Av = Bv v.
Since A is diffuse, we may shrink vv A so that (vv ) = 1/n, for some n N.
Write p1 = vv , q1 = v v and take projections p2 , . . . , pn A, q2 , . . . , qn B
such that (pi ) = (qj ) = 1/n. Since NM (A)00 = NM (B)00 = M and M is a II1
factor, a classical exhaustion argument gives partial isometries ui , wj M such that
p1 = ui ui , pi = ui ui , ui Aui = Aui ui , ui Aui = Aui ui and likewise q1 = wj wj ,
qj = wj wj , wj Bwj = Bwj wj , wj Bwj = Bwj wj . Define
u=

n
X

ui v wi .

i=1

## It is now routine to check that u U(M ) and uAu = B.

3. Approximation properties
3.1. Amenability.
3.1.1. Noncommutative Lp spaces. We refer to [27, Chapter IX] for the details of
the following facts on noncommutative Lp spaces. Let (N , Tr) be a semifinite
von Neumann algebra endowed with a faithful, normal, semifinite trace Tr. For
1 p < , we define the Lp -norm on N by kxkp = Tr(|x|p )1/p . By completing
{x N : kxkp < } with respect to the Lp -norm, we obtain a Banach space Lp (N ).
We only deal with L1 (N ), L2 (N ), and L (N ) = N . The trace Tr extends to a
contractive linear functional on L1 (N ). We occasionally write x
b for x N when
regarded as an element in L2 (N ). For any 1 p, q, r , with 1/p + 1/q = 1/r,
there is a natural product map
Lp (N ) Lq (N ) 3 (x, y) 7 xy Lr (N )

## AN INTRODUCTION TO II1 FACTORS

23

which satisfies kxykr kxkp kykq , x, y. The Banach space L1 (N ) is identified with
the predual of N under the duality
L1 (N ) N 3 (, x) 7 Tr(x) C.
The Banach space L2 (N ) si identified with the GNS-Hilbert space L2 (N , Tr). Elements in Lp (N ) can be regarded as closed operators on L2 (N ) which are affiliated with N and hence in addition to the above-mentioned product, there are
well-defined notions of positivity, square root, etc... We shall use the generalized
Powers-Strmer Inequality (see [27, Theorem IX.1.2]):
k|| ||k22 k 2 2 k1 k + k2 k k2 , , L2 (N ).
The Hilbert space L2 (N ) is an N -N bimodule such that hxy, i = Tr(xy ),
x, y N , , L2 (N ). We also recall
the following formulae. Let fa be the
characteristic function of the interval ( a, +). For any , L2 (N )+ , we have
(see [7, Proposition 1.1] and [27, Theorem IX.2.14])
Z
2
kfa ()k2 da = kk22 ,
0
Z
2
kfa () fa ()k2 da k k2 k + k2 .
0

Let H be a complex separable Hilbert space and let (N , Tr) = (B(H), Tr),
where Tr is the canonical trace on B(H). Then, L1 (B(H)) can be identified with
the space of trace-class operators on H, denoted by S1 (H) in the sequel. In the
same way, L2 (B(H)) can be identified with the space of Hilbert-Schmidt operators
on H, denoted by S2 (H) in the sequel.
Let (M, ) be a finite von Neumann algebra, denote by H = L2 (M, ) its L2 space with respect to the finite trace . The Hilbert space H is endowed with a
c , x M . In the sequel, we shall simply
canonical anti-unitary J defined by J x
b=x

## denote H 3 7 H. We regard M B(H) through the GNS-construction.

The following linear map
U:

XH H
k k
k

S
2 (H)
X
7
h, k ik
k

## defines a unitary. We shall identify S2 (H) and H H through this unitary U .

Moreover, for any , H, for any x M ,
U (x )

= h, ix = xh, i = xU ( )

U ( x)

## Thus, U preserves the M -M -bimodule structure: S2 (L2 (M )) with its bimodule

structure, as a two-sided ideal of B(L2 (M )), is identified with the so-called coarse
correspondence L2 (M ) L2 (M ).
Finally, note that the symbol Lim will be used for a state on ` (N) or more
generally on ` (I) with I directed set.
3.1.2. Amenable finite von Neumann algebras. Recall that a countable discrete
group is amenable if one the following equivalent conditions holds:
There exists a -invariant state : ` () C, i.e. (s f ) = (f ), for all
s , f ` ().

24

CYRIL HOUDAYER

## There exists a sequence of almost invariant unit vectors n `2 (), i.e.

limn ks n n k2 = 0, for all s .
Let (M, ) be a finite von Neumann algebra with separable predual. Denote H =
L2 (M, ). We regard M B(H) through the GNS-construction. A state on
B(H) is said to be M -central if Ad(u) = , u U(M ).
Theorem 3.1 (Connes, [7]). Let M be a finite von Neumann algebra. The following
are equivalent:
(1) There exists a conditional expectation E : B(H) M .
(2) There exists an M -central state on B(H) such that |M = .
(3) There exists a net of unit vectors (n ) in S2 (H) such that hxn , n i (x),
x M , and k[n , u]k2 0, u U(M ).
Proof. (1) = (2). Let E be a conditional expectation from B(H) onto M . Denote
= E. Then x B(H), u M , one has
(uxu ) = (E(uxu )) = (uE(x)u ) = (E(x)) = (x).
Thus, the state is M -central and |M = .
(2) = (3). Let be an M -central state on B(H) such that |M = . Take a
net (n ) of positive norm-one elements in S1 (H) such that Tr(n ) converges to
pointwise. Then x B(H), u U(M ), one has
n

## = (x) (Ad(u )(x)) = 0

by assumption. It follows that the net (n Ad(u)n ) in S1 (H) converges to 0 in
the weak topology. By the Hahn-Banach Separation Theorem, one may assume,
by passing to finite convex combinations, that the net (n Ad(u)n ) in S1 (H)
converges to 0 in norm. Thus, k[u, n ]k1 0, u U(M ). Define the unit vectors
1/2
n = n S2 (H). Using the Powers-Strmer Inequality, u U(M ),
k[u, n ]k22

kun u n k22

kun u n k1

k[u, n ]k1 .

limhxn , n i =
n

n

n

## This proves (3).

(3) = (1). Assume that there exists a net of unit vectors (n ) in S2 (H) such
that hxn , n i (x), x M , and k[n , u]k2 0, u U(M ). Note that we also
have k[n , u]k2 0, u U(M ). Write n = wn n for the polar decomposition,
with n = (n n )1/2 0, kn k2 = 1, and wn partial isometry in B(H). Thus, n,
x B(H),
hxn , n i = Tr(xn n ) = Tr(xn2 ) = hxn , n i.

25

## Consequently, limn hxn , n i = (x), x M . Moreover, using the Powers-Strmer

Inequality, n, u U(M ),
= kun u n k22

k[u, n ]k22

kun2 u n2 k1

= ku(n n ) (n n )uk1

## kun n uk2 + kun n uk2 .

Therefore, limn k[u, n ]k2 = 0. So, we may moreover assume that n 0, n. Thus,
c M ,
lim kn ck22
n

## limhn c, n ci = lim Tr(n cc n )

lim Tr(n2 cc )
n

= (cc ) = kck22, .

## Define the following state

(x) = Limhxn , n i, x B(H).
n

## Note that |M = . Moreover, b, c M , y, z B(H), since k[n , bc]k2 0, we

get
|(bcyz)| = | Lim Tr(n2 bcyz)|
n

n

n

n

n

## = kbk2, kck2, kyk kzk .

Take now x B(H), a M and write a = v|a| its polar decomposition in M .
Thus, we have a = (v|a|1/2 )|a|1/2 and
|(ax)|

## k|a|1/2 k2, k|a|1/2 k2, kxk

= kak1, kxk .
It follows that |(ax)| kak1, kxk , a M , x B(H). By the duality
M = L1 (M ) , we know that there exists (x) M , such that (a(x)) = (ax),
a M , x B(H). It is straightforward to check that : B(H) M is a
conditional expectation.

The state in (2) of Theorem 3.1 is called a hypertrace. Condition (3) says
that the identity bimodule L2 (M ) is weakly contained in the coarse bimodule
L2 (M ) L2 (M ), that we usually denote by L2 (M ) weak L2 (M ) L2 (M ). This
is the analog of the notion of amenability in the case of groups since the identity
bimodule plays the r
ole of the trivial representation and the identity bimodule plays
the one of the left regular representation. For this reason, a finite von Neumann
algebra M which satisfies one of the equivalent conditions of Theorem 3.1 is said to
be amenable. Note that more generally for any M -M -bimodule H, L2 (M ) weak H

26

CYRIL HOUDAYER

## if there exists a net (n ) of unit vectors in H such that limn hn , n i = pointwise

and limn k[u, n ]k = 0.
Proposition 3.2. Let be a countable discrete group. Then, is amenable if and
only if the group von Neumann algebra L() is amenable.
Proof. Assume that is amenable. Denote by the left regular representation of
on `2 (). Then 1 , i.e. there exists a net of unit vectors (n ) in `2 () such
that limn kg n n k2 = 0, g G. We shall identify `2 () with the L2 -space of
L(). Denote the unit vector n = n n `2 () `2 () = S2 (`2 ()). Moreover,
we have for all g ,
kug n n ug k2

= kug n n n n ug k2

kug n n n n k2 + kn n n n ug k2

kg n n k2 + kg1 n n k2 .

## Therefore, limn kug n n ug k2 = 0, g . Thus Condition (3) in Theorem 3.1 is

satisfied, and L() is amenable.
Assume that L() is amenable. Let H = `2 (). By Condition (3) in Theorem
3.1, we know that there exists a sequence of unit vectors (n ) in H H such
that for any g , limn kug n n ug k2 = 0. Define the unitary representation
: U(H H) by
(g) = ug ug .
It is straightforward to see that is a mutiple of and that 1 . Consequently,
1 and is amenable.

Exercise 3.3. Let M be a diffuse amenable finite von Neumann algebra. Show
that a hypertrace given by Theorem 3.1 can never be normal on B(H).
Recall that a finite von Neumann algebra M is said to be approximately finite
dimensional (AFD) if there exists an
Sincreasing sequence of finite dimensional unital
-subalgebras Qn M such that n Qn is ultraweakly dense in M . Murray and
von Neumann showed in their seminal work the uniqueness of the AFD II1 factor.
The following is easy to prove.
Proposition 3.4. Let M be a finite AFD von Neumann algebra. Then M is
amenable.
Proof.S Let Qn M be a sequence of finite dimensional unital -subalgebras such
that n Qn is ultraweakly dense in M . Denote by n the (probability) Haar measure on the compact group U(Qn ). Fix (N)\N a free ultrafilter. For every
x B(L2 (M )) define
Z
E 0 (x) = lim
uxu dn (u).
n

U (Qn )

## It is clear that E : B(L (M )) M is a conditional expectation. Denote by J the

canonical antiunitary. Thus E : B(L2 (M )) M defined by E(x) = JE 0 (JxJ)J is
a conditional expectation.

The converse is Connes fundamental result.
Theorem 3.5 (Connes, [7]). Let M be a finite von Neumann algebra. Then M is
AFD if and only if M is amenable. There exists a unique amenable II1 factor. In
particular, all icc amenable countable discrete groups give the same II1 factor.

27

## Exercise 3.6. Let y P be a trace-preserving action of an amenable group on

an amenable finite von Neumann algebra P . Show that M = P o is amenable.
3.2. Property Gamma.
Definition 3.7. A II1 factor M is said to have the property Gamma if for every
> 0 and every x1 , . . . , xn (M )1 , there exists u U(M ) such that (u) = 0 and
k[u, xi ]k2 < , for all 1 i n.
Theorem 3.8 (Connes, [8]). Let M be a II1 factor. The following are equivalent:
(1) M does not have property Gamma.
(2) There exists a non-trivial central sequence in M , i.e. M 0 M 6= C.
(3) M 0 M is diffuse.
Proof. For a free ultafilter , we will denote by : ` (N, M ) M the quotient
map. For (1) = (2), fix a free ultrafilter . Since M does not have property
Gamma, there exists a sequence of unitaries un U(M ) such that (un ) = 0 and
limn k[un , x]k2 = 0, for all x M . Define u = ((un )) U(M 0 M ). We have
that (u) = 0 and so M 0 M 6= C. For (2) = (3), let e M 0 M be a nonzero
projection such that e 6= 1. Write = (e) (0, 1). We may find a sequence of
projections en which represents e such ((en )) = e and (en ) = , for all n N.
Observe that (en ) is a central sequence and since M is a II1 factor, we have that
en weakly (by weak compactness of the unit ball (M )1 ). Thus we construct a
subsequence ekn which satisfies for every n N,
1
k[ej , ekn ]k2 <
n
1
, 1 j n.
| (ej ekn ) 2 | <
n
Define f := ((en ekn )) M 0 M . The previous inequalities show that f
M 0 M is indeed a nonzero projection such that f e and f 6= e, since (f ) =
2 < = (e). Therefore, M 0 M cannot have any minimal projections and
hence is diffuse.
(3) = (1). Assume that M 0 M is diffuse. Let e M 0 M be a projection
such that (e) = 1/2. We can then represent e = ((en )) with projections
en U(M ) such that (en ) = 1/2, for all n N. Therefore, u = 2e 1 M 0 M ,
with u = ((un )) and (un ) = 0, for all n N. Hence M does not have property
Gamma.

The next Theorem, due to Connes, gives a spectral gap characterization of property Gamma.
Theorem 3.9 (Connes, [7]). Let M be a II1 factor. The following are equivalent:
(1) M has property Gamma.
(2) There exists a sequence of unit vectors n L2 (M ) Cb
1 such that
lim kxn n xk2 = 0, x M.
n

## (3) K(L (M )) C (M, M 0 ) = {0}.

2

Proof. (1) = (2) is clear. For (2) = (1), write = (n ) L2 (M ) . There are
two cases to consider.
Case (1): defines an element in L2 (M ). We have h, b
1i = 0 and x = x,
for all x M . Write = v|| for the polar decomposition of . We have that

28

CYRIL HOUDAYER

## v M 0 M is a partial isometry such that v 6= 0, 1. Thus M has property

Gamma.
Case (2): does not define an element in L2 (M ). We start by proving
the following claim.
Claim. For every finite subset F U(M ), for every > 0, there exists a projection
e M such that (e) < and k[u, e]k2 < kek2 , for all u F .
By Proposition C.2, we know that
c > 0, a > 0, lim kfa (|n |)|n |k2 > c.
n

We may then choose a subsequence (kn ) such that kfn (|kn |)|kn |k2 c, for all n
N. Then with n = kfn (|k 1|)|k |k2 fn (|kn |)|kn |, we still have that limn kxn
n
n
n xk2 = 0, for all x M . Observe that for all a > 0,

n fa (n )n afa (n )

and so (fa (n )) 1/ a.
4
Let F U(M ) be a finite
subset and > 0. Let = /(4|F |). Choose n N
large enough such that 1/ n < and = n satisfies
X
k uu k22 < .
uF

## It is clear that fa (uu ) = ufa ()u . Hence, we get

Z
kfa ()k22 da = kk22 = 1
0

and
XZ
uF

## kfa () ufa ()u k22 da

k uu k2 k + uu k2

uF

2(|F |

k uu k22 )1/2

uF

p
< 2 |F |
Z
p
= 2 |F |

## Therefore there exists a n such that

X
p
kfa () ufa ()u k22 < 2 |F |kfa ()k22 .
uF

Letting e = fa (), we have (e) (fa ()) < and kue euk2 < kek2 , for all
u F . The claim is proven.
The next claim uses a maximality argument.
Claim. For every finite subset F U(M ), for every > 0, there exists a projection
e M such that (e) = 1/2 and k[u, e]k2 < , for all u F .
Once the claim is proven, we are done. Indeed, we can construct a sequence of
projections en M such that (en ) = 1/2 and limn k[x, en ]k2 = 0, for all x M .
We the get a sequence of unitaries un = 2en 1, with (un ) = 0 and such that
limn k[x, un ]k2 = 0, for all x M . Hence, M has property Gamma. It only remains
to prove the claim.

29

## Let u1 , . . . , uk U(M ) and > 0. Let I be the set of families i = (E, U1 , . . . , Uk )

such that:
E M is a projection such that (E) 1/2.
Each Uj U(M ) is a unitary commuting with E.
kUj uj k1 (E), for all 1 j k.
We define a partial ordering on I in the following way: we write i i0 if
E E 0 and kUj Uj0 k1 (E 0 E), 1 j k.
It is easy to see that if i i0 and i0 i00 , then i i00 . The set (I, ) is moreover
inductive. By Zorns Lemma, there exists a maximal element i = (E, U1 , . . . , Uk )
I. Assume that (E) < 1/2. We will deduce a contradiction. Let > 0 such
that (E) + < 1/2 and 4 . Let vj = (1 E)Uj = Uj (1 E) U(N ) where
N = (1E)M (1E). By the previous claim (with N ), we know that there exists a
projection e N such that N (e) < and k[vj , e]k2,N kek2,N , for all 1 j k.
By Proposition D.1, there exists wj U(N ) such that wj vj evj wj = e and

kwj (1 E)k1,N 2kvj evj ek1,N 4kvj evj ek2,N kek2,N 4N (e).
Let E 0 = E + e M . It is a projection stricly larger than E and such that
(E 0 ) 1/2. Let Uj0 = Uj E + wj vj . It is easy to see that Uj0 U(M ) and Uj0 is
commuting with E 0 . Moreover, kwj vj vj k1 4 (e), kUj0 Uj k1 4 (e). Since
4 and kUj uj k1 (E) by assumption, we get
kUj0 uj k1 (E + e) = (E 0 ), 1 j k.
The element i0 = (E 0 , U10 , . . . , Uk0 ) satisfies i i0 and i 6= i0 , which contradicts the
maximality of i I. Therefore, we have that (E) = 1/2. We have thus shown for
each > 0, the existence of a projection E M such that (E) = 1/2, [E, Uj ] = 0
and kUj uj k1 . We finally get
k[uj , E]k2 = k[uj Uj , e]k2 2kuj Uj k2 2(2)1/2 .
As > 0 is arbitrary, we are done.
(1) = (3). Assume that K(L2 (M )) C (M, M 0 ) 6= {0}. Since C (M, M 0 )
is a simple C -algebra, we get K(L2 (M )) C (M, M 0 ). Let (xn ) be a (uniformly bounded) central sequence in M . We get that [y, xn ] 0 -strongly for all
y C (M, M 0 ). Denote by PC : L2 (M ) C the orthogonal projection. Since
K(L2 (M )) C (M, M 0 ), we get [PC , xn ] 0 -strongly. We have that
lim kxn (xn )1k2 = lim k(xn PC PC xn )b
1k = 0,
n

## and so (xn ) is trivial.

(3) = (2). Assume that it is impossible to find a sequence of unit vectors n
like in (2). Then then exists > 0, a finite subset F U(M ) such that
max k uu k2 kk2 , L2 (M ).
uF

## We may assume that F = F . Define the self-adjoint operator

1 X
T =
uJuJ C (M, M 0 ).
|F |
uF

We have kT k 1, T b
1=b
1, so that b
1 is an eigenvector for the eigenvalue 1. We
show that T 1 is invertible on L2 (M ) Cb
1, so that there is a spectral gap at 1

30

CYRIL HOUDAYER

## for the operator T . Otherwise, since T is selfadjoint, there would be a sequence of

unit vectors k L2 (M ) Cb
1 such that limk k(T 1)k k2 = 0. We have
1 X
k(1 T )k k2 = k
(1 uJuJ)k k2 .
|F |
uF

Recall that a Hilbert space is uniformly convex. For all k N, we have that
X 1
kk uJuJk k2 = 2 2<hk , T k i = 2<hk , k T k i 2kk T k k2 .
|F |
uF

## We have that limk kk uJuJk k2 = 0, for all u F and so limk kk k2 = 0, which

is absurd. Thus T 1 is invertible on L2 (M ) Cb
1 and there exists > 0 such
that Sp(T ) [1, 1 ] {1}. By continuous functional calculus, we get that
PC C (M, M 0 ) and so K(L2 (M )) C (M, M 0 ).

A closely related concept for groups is the one of inner amenability. A countable
discrete group G is said to be inner amenable if the adjoint representation Ad : G
U(`2 (G) Ce ) defined by Adg h = ghg1 contains a sequence of almost invariant
unit vectors. Examples of inner amenable groups include amenable groups, direct
product groups G H, where H is infinite amenable, Baumslag-Solitar groups.
Examples of groups which are not inner amenable include free groups Fn , n 2,
and property (T) groups. The following is easy to prove.
Proposition 3.10 (Effros, [10]). Then G be an icc countable discrete group. If
L(G) has property Gamma, then G inner amenable.
Proof. Assume L(G) has property Gamma. There exists a sequence of unitaries
vn U(M ) such that (vn ) = 0 and limn k[vn , x]k2 = 0, for all x M . Define
n := un e `2 (G) Ce . Since
k Adg n n k = kug vn vn ug k2 ,
we get that (n ) is a sequence of almost invariant unit vectors for the adjoint

The converse is false though, as it was recently discovered by Vaes [30]. We can
illustrate the subtle difference between the property Gamma of L(G) and the inner
amenability of G.
The group G is inner amenable if and only if there exists a sequence of unit
vectors n `2 (G) Ce such that
lim kxn n xk2 = 0, x C (G).
n

The von Neumann algebra L(G) has property Gamma if and only if there
exists a sequence of unit vectors n `2 (G) Ce such that
lim kxn n xk2 = 0, x L(G).
n

## 3.3. Haagerup property.

Definition 3.11 (Haagerup, [14]). A countable discrete group is said to have
the property (H) if there exists a sequence of positive definite functions n : C
such that limn n = 1 pointwise and n c0 (), for all n N.
We will often write Haagerup property for property (H).

## AN INTRODUCTION TO II1 FACTORS

31

Example 3.12. The following groups have the Haagerup property: amenable
groups, free groups and more generally all groups which act properly on a tree.
The Haagerup property is moreover stable under taking subgroups, amenable extentions, free products, wreath products.
We refer to the book by P.A. Cherix, M. Cowling, P. Jolissaint, P. Julg and A.
Valette [3] for a comprehensive of groups with the Haagerup property
Definition 3.13 (Choda, [4]). Let (N, ) be a finite von Neumann algebra endowed
with a fixed faithful normal trace. We say that N has the Haagerup property if
there exists a sequence n : N N of -preserving ucp maps which satisfies:
limn kn (x) xk2 = 0, for all x N .
Whenever wk (N )1 is a sequence such that wk 0 weakly, then we have
limk kn (wk )k2 = 0, for all n N.
Proposition 3.14 (Choda, [4]). Let be a countable discrete group. Then has
the Haagerup property if and only if L() has the Haagerup property.
Proof. Assume that has the Haagerup property. Then there exists a sequence
n : C such that limn n = 1 pointwise and n c0 (), for all n N. We
may assume that n (e) = 1, for all n N. Define n : L() L() by
X
n (
as us ) = n (s)as us .
s

## It is straightforward to check that n is a sequence of -preserving ucp maps which

satisfies conditions of Definition 3.13.
Conversely, assume that L() has the Haagerup property. Let n be a sequence
of -preverving ucp maps given by Definition 3.13. Define n (s) = (n (us )us ),
for all s . Then n is a sequence of positive definite functions that does the
job.

Theorem 3.15 (Haagerup, [14]). The free groups Fn have the Haagerup property.
Proof. Denote by Fn 3 g 7 |g| R+ the natural length function. We will show
that for all 0 < < 1, the function defined by
Fn 3 g 7 |g|
is a positive definite function on Fn . Since c0 (Fn ) and lim1 = 1, we get
that Fn has the Haagerup property.
We give a proof using Popas free malleable deformation [21, 24]. We may assume
e n ), where F
e n is a copy of Fn
f = L(Fn F
that n < . Let M = L(Fn ) and M
which is free from Fn . Denote by a1 , . . . , an the canonical generators of L(Fn )
e n )). For 1 k n, let
(resp. b1 , . . . , bn the ones of L(F
hk =

1
log(bk ),
1

## where log denotes the

principal branch of the logarithm. We get that hk is selfadjoint and bk = exp( 1hk ). For t R, define

e n )).
btk := exp( 1thk ) U(L(F

32

CYRIL HOUDAYER

## It is straightforward to check that

Z
t
(t) := (bk ) =

sin(t)
exp( 1tx)dx =
t
1

fM
f by
Define the following -automorphism t : M
t (ak ) = ak btk and t (bk ) = bk .
f and are -free from each
Since the unitaries {a1 , . . . , an , a1 bt1 , . . . , ak btk } generate M
other, we check easily that (t ) is a one-paramater family of trace-preserving e = 1 (N ). We see that N N
e = M
f
automorphisms. Write N = L(Fn ) and N
e
f
e
and N is -free from N . Consequently, we get M = N N and (t ) satisfies
1 (x 1) = 1 x, for all x N . Moreover, we have
(EN t )(ug ) = (t)2|g| ug , g Fn ,
Since EN t is u.c.p., it follows that t : Fn C defined by t (g) = ((EN
t )(ug )ug ) = (t)2|g| is a positive definite function. We are done.

4. Rigidity of II1 factors
4.1. Rigid inclusions of von Neumann algebras. The notion of property (T)
for a II1 factor was introduced by Connes and Jones in their seminal work [9]. Its
relative version for a inclusion of finite von Neumann algebras is due to Popa [19].
Definition 4.1 (Popa, [19]). Let (M, ) be a finite von Neumann algebra with a
fixed trace and B M be a subalgebra. The inclusion is said to be rigid if for every
> 0, there exist > 0 and a finite subset F M such that for every -preserving
u.c.p. map : M M , we have
sup k(x) xk2 = sup k(x) xk2 .
xF

x(B)1

The von Neumann algebra M has property (T) if the identity inclusion M M
is rigid. Note that we can relax the assumptions in Definition 4.1. Indeed, let
: M M be a completely positive map such that (1) 1 and . Then,
e : M M defined by
( )(x)
e
(1 (1))
(x)
= (x) +
( )(1)
is a -preserving u.c.p. map.
Theorem 4.2 (Popa, [19]). Let B M be an inclusion of finite von Neumann
algebras and let be a fixed trace on M . The following are equivalent:
(1) The inclusion B M is rigid;
(2) For every > 0, there exist > 0 and a finite subset F M such that
for any M -M bimodule H and any tracial unit vector H for which
kx xk , x F , there exists a B-central vector H such that
k k .
Proof. We prove both directions.
(1) = (2). Let 0 < < 1. Let F M be a finite subset and > 0 given by
Condition (1). Let H be an M -M -bimodule and a tracial unit vector H such
that kx xk , x F . Let : M M be the -preserving u.c.p. map

33

## associated with (H, ). Recall that hax, yiH = h(a)b

x, ybiL2 (N ) , a, x, y M .
Then we have
k(x) xk22

## = k(x)k22 + kxk22 2<hx, xi

2kxk22 2<hx, xi
= kx xk2 2

for every x F . It follows that k(x) xk2 , for every x (B)1 . We get for
every u U(B),
k uu k2

2 2< ((u)u )

2< (1 (u)u )

2k1 (u)u k2 2.
Denote by the circumcenter of the bounded set C = {uu : u U(B)}. Since
uCu = C, u U(B), by uniqueness it follows that uu = , u U(B). Thus
follows B-central and moreover k k (2)1/2 .
(1) = (2). Let 0 < < 1. Let F M be a finite subset and 0 < < 1 given by
Condition (2). Let : M M be a -preserving u.c.p. such that k(x) xk2
(2kxk2 )1 2 , x F . Let H be the M -M bimodule and be the tracial vector
associated with . We have hax, yiH = h(a)b
x, ybiL2 (N ) , a, x, y M . For every
x F , we get
kx xk22

=

## 2kx (x)k2 kxk2 2 .

Therefore, there exists a B-central vector H such that k k . For every
x (B)1 , we get
kx (x)k22

## = kxk22 + k(x)k22 2< ((x)x )

2kxk2 2<hx, xi
2kxkkx xk
=

2kxkkx( ) ( )xk

4kxk kxkk k 4.

Exercise 4.3. For i = 1, 2 let Bi Mi be an embedding of finite von Neumann
algebras. Show that the following are equivalent:
(1) For i = 1, 2, the inclusion Bi Mi is rigid.
(2) The inclusion B1 B2 M1 M2 is rigid.
Recall that a pair (, ) consisting of a countable with subgroup is said
to have the relative property (T) of Kazhdan-Margulis [16, 17] if every unitary
representation of which admits a sequence of almost invariant unit vectors, admits
a non-zero -invariant vector. Equivalently, for every > 0, there exist > 0 and
a finite subset F such that for any unitary representation : U(H )
which has a ((F ), )-invariant unit vector , there exists a non-zero ()-invariant
vector such that k k (see [15, Theorem 1.2(a3)]). A group is said to have
property (T) if the pair (, ) has the relative property (T).

34

CYRIL HOUDAYER

## Example 4.4. Here are a few classical examples.

(1) The example by excellence of a pair with relative property (T) is (Z2 o
SL(2, Z), Z2 ).
(2) For any n 3, SL(n, Z) has property (T).
(3) For any property (T) group and any group H, the pair (H , ) has
relative property (T).
We refer to the book by Bekka, de la Harpe and Valette [1] for a comprehensive
list of groups with (relative) property (T).
Theorem 4.5 (Popa, [19]). Let be an inclusion of countable groups. The
following are equivalent:
(1) The pair (, ) has the relative property (T);
(2) The inclusion L() L() is rigid.
Proof. Write B = L() L() = M .
(1) = (2). Assume that the pair (, ) has the relative property (T). Fix > 0.
We know that there exist > 0 and a finite subset F such that for any unitary
representation : U(H ) which has a ((F ), )-invariant unit vector , there
exists a non-zero ()-invariant vector such that k k . Let now H be a
M -M -bimodule and a unit tracial vector H such that kus us k , s .
Define the unitary representation : U(H) by s () = us us . The vector
is then ((F ), )-invariant. Thus there exists a ()-invariant vector H such
that k k . It is then clear that is B-central.
(1) = (2). Assume that the inclusion B M is rigid. Let = 1/2. We
know that there exist > 0 and a finite subset F M such that for any M -M bimodule H and any tracial vector H for which kx xk , x F , there
exists a B-central vector H such that k k . Let : U(H ) be a
unitary representation. Consider the M -M -bimodule K = H `2 () associated
with . Take a sequence (n ) H of almost invariant unit vectors and set n =
n e . It is then clear that (n ) K is a sequence of tracial vectors for which
limn kxn n xk = 0, x M . For n N large enough, we have kxn n xk ,
x F . Write = n . Therefore there exists a nonzero B-central
vector K
P
such that k k 1/2. Regard `2 (, H ) and write = s s s , where
s H . We have
X
ke k2 +
ks k2 = k k2 1/4.
s{e}

Since H is a unit vector, we have that e 6= 0. Since e is moreover ()invariant, the proof is complete.

The previous theorem shows in particular that the inclusion L (T2 ) L (T2 )o
SL(2, Z) is rigid.
4.2. Applications to rigidity of II1 factors. We use now the tools we introduced
in the previous sections to get structural properties for property (T) II1 factors.
Most of the proofs are based on a separability vs property (T) argument that
goes back to Connes [6].

## AN INTRODUCTION TO II1 FACTORS

35

4.2.1. Symmetry groups of property (T) factors are countable. For a II1 factor M ,
we endow the group Aut(M ) with the topology of pointwise k k2 -convergence: for
a sequence (n ) in Aut(M ), we have
n Id kn (x) xk2 0, x M.
Note that Aut(M ) is a polish group. We shall denote by Inn(M ) Aut(M ) the
subgroup of inner automorphisms. For any u U(M ), we shall write Adu (x) =
uxu , x M . We denote by Out(M ) the quotient group Aut(M )/ Inn(M ).
For the hyperfinite II1 factor R, the outer automorphisms group Out(R) is
huge. Indeed, one can embed any second countable locally compact group G
in Out(R). For property (T) II1 factors, the situation is dramatically different.
Theorem 4.6 (Connes, [6]). Let M be a property (T) II1 factor. Then Out(M ) is
countable.
Proof. We show that Inn(M ) Aut(M ) is an open subgroup. Thus, it follows that
Out(M ) = Aut(M )/ Inn(M ) is a Hausdorff discrete group, and by separability,
Out(M ) is necessarily countable. Fix = 1/2.
By property (T) of M , we know that there exists > 0 and F M finite subset
such that for every u.c.p. -preserving map : M M , we have
sup k(x) xk2 < = sup k(x) xk2 < 1/2.
xF

x(M )1

## Let V,F = { Aut(M ) : k(x) xk2 < , x F } be an open neighborhood

of Id in Aut(M ). Let V,F . Since k(x) xk2 < , x F , we know that
k(x) xk2 1/2, for every x (M )1 . Observe that k(u)u 1k2 1/2, for
every u U(M ). Consider
C = cow {(u)u : u U(M )}
the weak closure of the convex hull of all the (u)u s, for u U(M ). Observe that
C (M )1 is closed in L2 (M ).
Denote by a C the unique element of minimum k k2 -norm. It follows that
ka 1k2 1/2. We get a 6= 0. Observe that for every u U(M ), (u)au C and
k(u)au k2 = kak2 . By uniqueness, we get (u)au = a, for every u U(M ). So
we have a au = ua a, for every u U(M ). It follows that a a = R+ since M

## is a factor. Therefore v = a/ U(M ) and = Adv .


Observe that a property (T) factor cannot have property Gamma. Recall that
for a II1 factor M , the fundamental group of M is defined as follows:
F(M ) = { (p)/ (q) : pM p ' qM q}.
Murray & von Neumann showed that the unique AFD II1 factor R has full fundamental group, i.e. F(M ) = R+ . There is an alternative way of defining the
fundamental group of M . Denote by M = M B(`2 ) the corresponding II factor with semifinite trace Tr given by Tr = TrB(`2 ) . For any Aut(M ),
there exists a unique > 0 such that Tr = . We shall denote this by mod().
Moreover, the map mod : Aut(M ) R+ is a group homomorphism. It is then
easy to check that
F(M ) = {mod() : Aut(M )}.
Using property (T), Connes [6] gave the first example of II1 factor with countable
fundamental group.

36

CYRIL HOUDAYER

Theorem 4.7 (Connes, [6]). Let M be a property (T) II1 factor. Then F(M ) is
countable.
Proof. We construct a one-to-one map : F(M ) Out(M M ). Since M
has property (T), M M has property (T) as well (cf Exercise 4.3) and then
Out(M M ) is countable by Theorem 4.6. Therefore, F(M ) follows countable.
Claim. Let N be a II1 factor. Let Aut(N ) such that mod() = 1. Then
there exist u U(N ) and Aut(N ) such that
= Adu ( IdB(`2 ) ).
Therefore the group homomorphism
{ Aut(N ) : mod() = 1} 3 7 [] Out(N )
is well-defined.
Denote by (eij ) B(`2 ) the canonical matrix unit such that Tr(eii ) = 1, i
N. Let Aut(N ) such that mod() = 1. Write fij = (1 eij ). Since
Tr(f00 ) = Tr(1 e00 ) = 1, f00 and 1 e00 are equivalent projections in the factor
N , so that there exists P
a partial isometry v N such that vv = f00 and

v v = 1 e00 . Define u =
fj0 v(1 e0j ). It is routine to check that u U(N )

and u(1 eij )u = fij , which finishes the proof of the claim.
Let t F(M ) and choose t Aut(M ) such that mod(t ) = t. Since t t1
Aut((M M ) ) has modulus 1, the claim yields a unique t = [t ] Out(M M ),
so that the map
(1)

F(M ) 3 t 7 t Out(M M )

## is well-defined. If s 6= t, s1 t is outer, and so is (s s1 )1 (t t1 ). Thus

s 6= t . Since the map (1) is one-to-one, we are done.

4.2.2. Connes rigidity conjecture. Connes in the late 70s conjectured the following:
any countable icc property (T) groups , ,
' L() ' L().
Popa suggested the following strenghthening of Connes rigidity conjecture:
Conjecture 4.8 (Popa, [22]). If is an icc property (T) group and is a group,
then any -isomorphism : L() ' L()t forces t = 1 and there exist a group
isomorphism : ' and a character Hom(, T) such that
X
X
(
as us ) =
(s)as u(s) .
s

In particular,
F(L()) = {1} and Out(L()) = Out() Hom(, T).
Popa observed in [20] that Ozawas original result [18] could be used to prove
Connes rigidity conjecture up to countable classes.
Theorem 4.9 (Ozawa, [18]). For icc countable property (T) groups, the map
L() is countable-to-one.

## AN INTRODUCTION TO II1 FACTORS

37

Proof. The proof is an analog of the one of Theorem 2 in [18]. We prove the
result by contradiction and assume that there are uncountably many pairwise nonisomorphic icc property (T) groups (i )iI which give the same II1 factor M . By
Shaloms result [25], we know that each property (T) group is the quotient of a
finitely presented property (T) group. Since there are only countably many finitely
presented groups, we may assume that all the i s are quotients of the same property
(T) group .
We regard M B(L2 (M )) represented in its standard form and M = L(i ) for
every i I, so that i U(M ). Denote by i : U(M ) a group homomorphism
such that i () = i . Fix = 1/4. Since is a property (T) group, there exist
0 < < 1 and a finite subset E such that for any unitary representation
: U(H) and any ((E), )-invariant unit vector H, there exists a ()invariant vector H such that k k 1/4.
Since the finite von Neumann ` (E, M ) is k k2 -separable, there exist i 6= j
I such that maxsE ki (s) j (s)k2 . Let J be the canonical antiunitary
defined on L2 (M ) and define the unitary representation : U(M ) by (s) =
i (s)Jj (s)J. We have
k(s)b
1b
1kL2 (M ) = ki (s)j (s) 1k2 .
Since the vector b
1 is ((E), )-invariant, it follows that there exists a ()-invariant
2
vector L (M ) such that k b
1kL2 (M ) 1/4. Thus, for every s we have
ki (s) j (s)k2

= k(s)b
1b
1kL2 (M )
= k(s)(b
1 ) (b
1 )kL2 (M ) 1/2.

Exactly as in the proof of Theorem 4.6, denote by a the unique element of minimum
k k2 -norm in the weakly closed convex set C = cow {i (s)j (s) : s }. Since
ka 1k2 < 1, it follows that a 6= 0. Observe that i (s)aj (s) C and
ki (s)aj (s) k2 = kak2 . By uniqueness, we have i (s)aj (s) = a. Moreover,
a aj (s) = j (s)a a, for every s
. Since M is a factor and j ()00 = M , we

## have a a = R+ . Thus v = a/ U(M ) and i (s) = vj (s)v , for every

s . It follows in particular that i and j are isomorphic, contradiction.

4.3. Uniqueness of Cartan subalgebras.
Theorem 4.10 (Popa, [19]). Let y B be a trace-preserving action of a countable
group with the Haagerup property on a finite von Neumann algebra B. Denote
by M = B o the crossed product. Let A M be a rigid von Neumann subalgebra.
Then A M B.
Proof. Since has the Haagerup property, let n : C be a sequence of positive
definite functions such that limn n = 1 pointwise and n C0 (), for all n N.
Define n : M M the sequence of -preserving ucp maps as follows:
X
X
n (
as us ) =
n (s)as us .
s

It is straightforward to check that every n satisfies the following relative compactness property: if (wk ) is a sequence in (M )1 which satisfies limP
k kEB (awk b)k2 = 0,
for all a, b M , then limk kn (wk )k2 = 0. Indeed, write wk = s (wk )s us , where

38

CYRIL HOUDAYER

P
(wk )s = EB (wk us ). Then we have n (wk ) = s n (s)(wk )s us and thus
X
kn (wk )k22 =
|n (s)|2 k(wk )s k22 .
s

Assume that limk k(wk )s k2 = 0, for all s . Fix > 0. Since n c0 (),
V := {s : |n (s)| 2 /2} is a finite set. Then we have
X
X
kn (wk )k22 =
|n (s)|2 k(wk )s k22 +
|n (s)|2 k(wk )s k22
sV

/2

sV

k(wk )s k22

sV

/2 +

sV
2

## |n (s)| k(wk )s k22 .

sV

P
We can choose k0 N large enough so that sV |n (s)|2 k(wk )s k22 2 /2. Therefore, we have kn (wk )k2 , for all k k0 .
Since A M is rigid, there exists n N such that kn (w) wk2 1/4, for all
w U(A). By contradiction, assume that A M B. Then there exists a sequence
wk U(A) such that limk kEB (awk b)k2 = 0, for all a, b M . For k N large
enough, we get
1 = kwk k2 kn (wk ) wk k2 + kn (wk )k2 1/4 + 1/4 = 1/2,
which is absurd.


2

Corollary 4.11 (Popa, [19]). Consider the linear action SL2 (Z) y T . Then, up
to unitary conjugacy, L (T2 ) is the unique rigid Cartan subalgebra in L (T2 ) o
SL2 (Z).
Proof. Write M = L (T2 ) o SL2 (Z). Let A M be rigid Cartan subalgebra.
We get A M L (T2 ) by the previous Theorem. Since A, L (T2 ) M are both
Cartan subalgebras of the II1 factor M , Theorem 2.14 yields u U(M ) such that
uAu = L (T2 ).

We can now apply this last result to compute explicitely the fundamental group of
L(Z2 oSL2 (Z)). Gaboriau [13] showed that the group SL2 (Z) has fixed price and its
cost equals 13/12. This means that for every free ergodic p.m.p. SL2 (Z) y X, the
equivalence relation R(SL2 (Z) y X) has cost 13/12. It follows from the induction
formula [13, Proposition II.6] that R(SL2 (Z) y X) has trivial fundamental group.
Corollary 4.12 (Popa, [19]). We have F(L(Z2 o SL2 (Z))) = {1}.
Proof. Let M = L(Z2 o SL2 (Z)) = A o SL2 (Z). Let R be the equivalence relation
induced by the action SL2 (Z) y T2 . Let t 1 such that M ' M t . We can assume
that p A is a projection of trace t so that (pAp pM p) ' (At M t ). It follows
that At M is a rigid Cartan subalgebra and thus there exists u U(M ) such
that uAt u = A, by Corollary 4.11. This shows that R ' Rt (see Theorem 1.23)
and thus t = 1.

Popas result was the first explicit computation of a fundamental group of a II1
factor that was different from R+ , solving then a long-standing open problem of

39

## Appendix A. Polar decomposition of a vector

Let (N, ) be a finite von Neumann algebra. Since is fixed, we simply denote
L2 (N, ) by L2 (N ). We regard N B(L2 (N )). Let L2 (N ) such that 6= 0.
b L2 (N ) be the linear operator defined by T 0 (b
Let T0 : N
x) = x, for all x N .
Proposition A.1. The densily defined operator T0 is closable. Denote by T its
closure. The operator T is affiliated with N . Write T = v|T | for its polar decomposition. Then v N and |T | is affiliated with N .
Let B N be a von Neumann subalgebra such that x = x, for all x B. Then
we have xv = vx, for all x B.
Proof. First, we prove that the operator T0 is closable. It suffices to show that
(T0 ) is densily defined. Let y N and z N . Then,
hT0 (b
y ), zbi = hy, zbi = hJy , zbi
= hz Jy J, b
1i = hJy Jz , b
1i
0
= hb
y , Jz i = hb
y , (J)zi = hb
y , TJ
(b
z )i.
0
Then TJ
(T0 ) and so (T0 ) is densily defined. Thus T0 is closable and we
denote by T its closure. We prove now that T is affiliated with N . Let a, x N .
On the one hand,
T0 Ja J(b
x) = T0 (c
xa) = xa.

## On the other hand,

Ja JT0 (b
x) = xa.
Consequently, we have Ja JT0 T0 Ja J, for all a N . Since JN J = N 0 , it
follows that T is affiliated with N . Write T = v|T | for the polar decomposition
of T . Since v is bounded and affiliated with N , we have that v N . Moreover,
|T | is affiliated with N .
At last, let B N be a von Neumann subalgebra such that for any x B,
x = x. Fix x B. It is straightforward to check that xT T x. We also have
x(T ) (T ) x, and so x(T ) T (T ) T x. By functional calculus, it follows
that x|T | |T |x. Moreover, since N is a finite von Neumann algebra, since x N
and |T | is affiliated with N , it follows that x|T | and |T |x are closed, affiliated
with N and consequently the equality x|T | = |T |x holds. Thus,
xv|T | = xT T x v|T |x vx|T |.
It follows that xv and vx coincide on the range of |T |, and so xv = vx. Thus,
xv = vx, for every x B.

Appendix B. Von Neumanns dimension theory
Let (N, ) be a finite von Neumann algebra with a distinguished faithful normal
trace. Let H be a right Hilbert N -module, i.e. H is a complex (separable) Hilbert
space together with a normal -representation : N op B(H). For any b N ,
and H, we shall simply write (bop ) = b.
Proposition B.1. Let H be a right N -module. Then there exists an isometry
v : H `2 L2 (N ) such that v(b) = v()b, for all H, b N .

40

CYRIL HOUDAYER

## Proof. Let : N op B(H (`2 L2 (N ))) be the -representation defined by



(y op )
0
op
.
(y ) =
0
1`2 y op
Let Q = (N op ). It is clear that Q is a finite von Neumann algebra and the
projections




1 0
0
0
p=
and q =
0 0
0 11
belong to Q0 B(H(`2 L2 (N ))) which is a semifinite von Neumann algebra. Since
q is infinite and p is finite, [26, Theorem V.1.8] yields a nonzero central projection
z Z(Q0 ) such that zp - zq. There exists an isometry w Q0 such that w w = zp
and ww zq. If we write




z1
0
a b
z=
and w =
,
0 1 z2
c d
we get
a a + c c = z1
b b + d d =

aa + bb

cc + dd

1 z2 .

## Thus a = 0, b = 0, d = 0, c c = z1 and v = c : Hz1 `2 L2 (N ) is an isometry

(since u u = z1 ) which moreover satifies v(x) = (1 x)v (since w Q0 ). Then a
simple maximality argument finishes the proof.

Since p = vv commutes with the right N -action on `2 L2 (N ), it follows that
p B(`2 )N . Thus, as right N -modules, we have
HN ' p(`2 L2 (N ))N .
On B(`2 )N , we define the following faithful normal semifinite trace Tr (which
depends on ): for any x = [xij ]i,j (B(`2 )N )+ ,
X
Tr ([xij ]i,j ) =
(xii ).
i

We set dim(HN ) = Tr(vv ). Note that the dimension of H depends on but does
not depend on the isometry v. Indeed take another isometry w : H `2 L2 (N ),
satisfying w(b) = w()b, for any H, b N . Note that vw B(`2 )N and
w w = v v = 1. Thus, we have
Tr(vv ) = Tr(vw wv ) = Tr(wv vw ) = Tr(ww ).
We define dim(HN ) := Tr(vv ), for any isometry v as in Proposition B.1.
Appendix C. Ultraproducts
Let be a free ultrafilter on N and (N, ) a finite von Neumann algebra. Let
I be the norm closed ideal of ` (N, N ) defined by
n
o
I = (xn ) ` (N, N ) : lim kxn k2 = 0 .
n

Denote by : ` (N, N ) ` (N, N )/I the quotient map. The tracial ultraproduct of N is defined as N := (` (N, N )) with tracial faithful state

## AN INTRODUCTION TO II1 FACTORS

41

( ((xn ))) = limn (xn ). It is easy to check that N is indeed a von Neumann
algebra and is normal on N .
Exercise C.1. Let be a free ultrafilter and (N, ) a finite von Neumann algebra.
Prove that every projection e N lifts to a projection (en ) ` (N, N )
such that limn (en ) = (e).
Show that if N is a II1 factor, so is N .
2

## Regard L2 (N ) as a Hilbert subspace

of L (N ) . For a 0, denote by fa the
characteristic function of the interval ( a, +).

## Proposition C.2 ([7]). Let = (n ) L2 (N ) . Then L2 (N ) if and only if

satisfies the following equi-integrability condition:
> 0, a > 0, lim kfa (|n |)|n |k2 < .

(2)

## Proof. Assume that L (N ) satisfies (2). Simply write = (n ). Fix > 0.

Then for some a > 0, one has limn kfa (|n |)n k2 < so that the vector = (n )
with n = n (1fa )(|n |) satisfies kk2 = limn kn n k2 and kn k a.
Thus, L2 (N ).
Conversely assume that L2 (N ). We may assume that kn k2 1, for all
n N. Let (0, 1). We can then find a > 0 and xn N such that kn xn k2
and kxn k a, for all n N. Up to extracting, Powers-Strmer Inequality yields
in particular
k|n | |xn |k22 k|n |2 |xn |2 k1 kn xn k2 kn + xn k2 3, n N.
Using the same trick as in the proof of Theorem 3.9, we get
Z
kfb (|n |) fb (|xn |)k22 k|n |2 |xn |2 k1 3.
0

Since Eb (|xn |) = 0 for all b > a, we get kf2a (|n |)k2 (3)1/2 /a, for all n N.
Then, for all n N,
kf2a (|n |)|n |k2

## Appendix D. On the geometry of two projections

Proposition D.1 ([7]). Let M be a finite von Neumann algebra. If e and f are
equivalent projections in M , then there exists u U(M ) such that
ueu

= f

u|e f | = |e f |u

2|e f |.
|u 1|
Proof. Recall the analysis of two projections in [26, Chapter V]. Let e = 1 e
and f = 1 f . Set
e0

= e e f e f

f0

= f f e f e .

42

CYRIL HOUDAYER

## We know that e0 f0 = 0 and e0 e0 f0 = e0 f0 . We then represent



1 0
e = ef
e f 0 0,
0 0
 2

c cs
f = ef
0 e f 0,
cs s2
with c, s (e0 f0 )M (e0 f0 ) positive elements such that c2 + s2 = e0 f0 . Since
e f and e0 f0 , the finiteness of M yields e f e f . Consequently, e f
and e f are represented by matrices:


1 0
e f =
0 0


0 0
e f =
.
0 1
Therefore we come to the following situation:

 

1 0
1 0
e = ef

0,
0 0
0 0
 2
 

c cs
0 0
f = ef

0.
cs s2
0 1
and


s
|e f | = 0
0

 
0
1

s
0


0
0.
1

We set

u=ef

 
c s
0

s c
1


1
(e f ) .
0

## It is straightforward to check that u is a unitary, ueu = f , u|e f | = |e f |u and

that
(u 1) (u 1) 2(e f )2 .

References
[1] B. Bekka, P. de la Harpe, A. Valette, Kazhdans property (T). New Mathematical
Monographs, 11. Cambridge University Press, Cambridge, 2008. xiv+472 pp.
[2] N. Brown, N. Ozawa, C -algebras and finite-dimensional approximations. Grad. Stud.
Math., 88. Amer. Math. Soc., Providence, RI, (2008).
[3] P.A. Cherix, M. Cowling, P. Jolissaint, P. Julg, A. Valette, Groups with the
Haagerup Property. Progress in Mathematics 197. Birkh
auser Verlag, Basel, Boston, Berlin,
2001.
[4] M. Choda, Group factors of the Haagerup type. Proc. Japan Acad. 59 (1983), 174177.
[5] A. Connes, Noncommutative geometry. Academic Press. San Diego, California, 1994.
[6] A. Connes, A factor of type II1 with countable fundamental group J. Operator Theory 4
(1980), no. 1, 151153.
[7] A. Connes, Classification of injective factors. Ann. of Math. 104 (1976), 73115.
[8] A. Connes, Almost periodic states and factors of type III1 . J. Funct. Anal. 16 (1974),
415445.
[9] A. Connes, V.F.R. Jones, Property (T) for von Neumann algebras. Bull. London Math.
Soc. 17 (1985), 5762.
[10] E.G. Effros, Property and inner amenability. Proc. Amer. Math. Soc. 47 (1975), 483
486.

## AN INTRODUCTION TO II1 FACTORS

43

[11] J. Feldman, C. C. Moore, Ergodic equivalence relations, cohomology, and von Neumann
algebras. I. Trans. Amer. Math. Soc. 234 (1977), 289324.
[12] J. Feldman, C. C. Moore, Ergodic equivalence relations, cohomology, and von Neumann
algebras. II. Trans. Amer. Math. Soc. 234 (1977), 325359.
[13] D. Gaboriau, Co
ut des relations d
equivalence et des groupes. Invent. Math. 139 (2000),
4198.
[14] U. Haagerup, An example of non-nuclear C -algebra which has the metric approximation
property. Invent. Math. 50 (1979), 279293.
[15] P. Jolissaint, On property (T) for pairs of topological groups. Ens. Math. (2) 51 (2005),
3145.
[16] D. Kazhdan, Connection of the dual space of a group with the structure of its subgroups.
Funct. Anal. Appl. 1 (1967), 6365.
[17] G. Margulis, Finitely-additive invariant measures on Euclidean spaces. Ergodic Th. and
Dynam. Sys. 2 (1982), 383396.
[18] N. Ozawa, There is no separable universal II1 -factor. Proc. Amer. Math. Soc., 132 (2004),
487490.
[19] S. Popa, On a class of type II1 factors with Betti numbers invariants. Ann. of Math. 163
(2006), 809899.
[20] S. Popa, Deformation and rigidity for group actions and von Neumann algebras. International Congress of Mathematicians. Vol. I, 445477, Eur. Math. Soc., Zurich, 2007.
[21] S. Popa, On Ozawas property for free group factors. Int. Math. Res. Notices. Vol. 2007 :
article ID rnm036, 10 pages.
[22] S.
Popa,
Some
results
and
problems
in
W -rigidity.
Available
at
http://www.math.ucla.edu/popa/tamu0809rev.pdf.
[23] S. Popa, Strong rigidity of II1 factors arising from malleable actions of w-rigid groups I.
Invent. Math. 165 (2006), 369-408.
[24] S. Popa, Some rigidity results for non-commutative Bernoulli Shifts. J. Funct. Anal. 230
(2006), 273328.
[25] Y. Shalom, Rigidity of commensurators and irreducible lattices. Invent. Math. 141 (2000),
154.
[26] M. Takesaki, Theory of Operator Algebras I. EMS 124. Springer-Verlag, Berlin, Heidelberg, New-York, 2000.
[27] M. Takesaki, Theory of Operator Algebras II. EMS 125. Springer-Verlag, Berlin, Heidelberg, New-York, 2000.
[28] M. Takesaki, Theory of Operator Algebras III. EMS 127. Springer-Verlag, Berlin, Heidelberg, New-York, 2003.
[29] S. Vaes, Rigidity results for Bernoulli actions and their von Neumann algebras (after S.
Popa). S
eminaire Bourbaki, expos
e 961. Ast
erisque 311 (2007), 237-294.
[30] S. Vaes, An inner amenable group whose von Neumann algebra does not have property
Gamma. Acta Math., to appear. arXiv:0909.1485
CNRS-ENS Lyon, UMPA UMR 5669, 69364 Lyon cedex 7, France