Analysis on compact Lie groups

Semester Project Report

by

SHUBHAM GIRDHAR
1311041

to the

School of Mathematical Sciences

National Institute of Science Education and Research
Bhubaneswar

21 April 2017
 Abstract

The representations of compact Lie groups are studied. First, the basic results about
irreducible representations of compact group is considered. Then we consider the
group SU (2), the simplest non-commutative compact Lie group. Finally we study
the irreducible representations of SO(3).
Contents
1 Introduction 1
2 Topological Groups 2
2.1 Definition and Examples . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 Haar Measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Computation of Haar measure . . . . . . . . . . . . . . . . . . . . . . 4
3 Representation of Compact groups 7
3.1 Unitary representations . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Schurs orthogonality relations . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Peter-Weyl Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4 Characters and central functions . . . . . . . . . . . . . . . . . . . . . 15
4 The Groups SU (2) and SO(3) 17
4.1 Adjoint representation of SU (2) . . . . . . . . . . . . . . . . . . . . . 17
4.2 Irreducible representations of SU (2) . . . . . . . . . . . . . . . . . . . 18
4.3 Irreducible representations of SO(3) . . . . . . . . . . . . . . . . . . . 22
References 27

i
Chapter 1

Introduction
This report is concerned with the study of compact groups and their representations.

Chapter 2 discusses the basic results about topological groups. It introduces Haar
measure and also states its existence for locally compact groups. This chapter also

contains examples and method to compute Haar measure for certain class of locally
compact groups.
Chapter 3 is devoted to the Peter-Weyl theory for compact groups. By using spectral

theory for compact operators we see that an irreducible representation of a compact

group is finite dimensional. By the Peter-Weyl theory the classical Fourier analysis
is extended to compact groups.

Chapter 4 is concerned with the groups SU (2) and SO(3). The special othogonal
group SO(3) and its simply connected covering SU (2) are the simplest non commu-
tative compact linear groups. We study the irreducible representation of them.

1
Chapter 2

Topological Groups
2.1 Definition and Examples

Definition 1. A group G is topological group if it is equipped with Hausdorff topology

such that the maps (x, y) 7 xy from G G into G and x 7 x1 from G into g are
continuous.

Examples

1. Rn , with usual addition and topology.

2. R = R {0}, with usual multiplication and topology.

3. S 1 S 1 S 1 , with coordinate-wise multiplication and topology induces

from Cn .
Remark: The set g`(n, R) g`(n, C) of all n n matrices over R(C) can be
2 2
identified with Rn (Cn ) with the usual topology.

4. The general linear group: GL(n, R) = {A g`(n, R)|A is non singular }.

5. The special linear group: SL(n, R) = {A GL(n, R)|det(A) = 1}.

6. The orthogonal group: O(n) = {A GL(n, R)|A is an orthogonal matrix }.

7. The special orthogonal group: SO(n) = {A O(n)|det(A) = 1}.

2
 2 Topological Groups

2.2 Haar Measure

Every locally compact group admits a special measure, called the Haar measure,
which plays role similar to Lebesgue measure in R or S 1 .

Definition 2. A left Haar measure on a locally compact group G is a positive regular

Borel measure such that (xE) = (E), E B and x G.

A right Haar measure can be defined analogously. Also, if is left Haar measureon

G, then the measure defined by (E) = (E 1 ), E B, is the right Haar measure.

The following important theorem states the existence and uniqueness of left Haar
measure on any locally compact group G.

Theorem 1. Let G be locally compact group. There exists a left Haar measure on

G. If is any other left Haar measure on G, then there exists a constant c > 0 such
that = c.

If G is compact, then, note that (G) < . In this case, we normalise Haar
measure so that (G) = 1. Now, for fixed g G the linear form
Z
f 7 f (gxg 1 )d(x).
G

defines a left invariant measure. Hence, there exists a positive number M (g) such

that
Z Z
1
f (gxg )d(x) =M (g) f (x)d(x).
G G

Also, observe that

Z Z
1
f (xg )d(x) =M (g) f (x)d(x).
G G

The function M (g) is clearly multiplicative. In fact:

3
 2 Topological Groups

Lemma 1. The function M is a continuous group homomorphism,

M: G R+ .

Proof. Let f Cc (G) such that

Z
f (x)d(x) = 1.
G

Then
Z
M (g) = f (xg 1 )d(x).
G

Since, f is left uniformly continuous, it follows that M is continuous.

The function M is called the module of the group G. Observe that, for Borel
set E G, (Eg) =M (g)(E). If M= 1, the group is said to be unimodular. A

commutative group is unimodular.

Lemma 2. A compact group is unimodular. A discrete group is unimodular.

Proof. If G is compact then M (G) is compact subgroup of R+ , and {1} is the only
compact subgroup of R+ . On the other hand, if G is discrete, then a continuous
function with compact support is a function with finite support. the measure

defined on G by
Z X
f (x)d(x) = Gf (x)
G x

is left and right invariant.

2.3 Computation of Haar measure

1. Let G be a finite group. For any set E, |E| denotes the number of elements in
E. Let |G| = n. We define the measure on G by (E) = n1 |E|, E G. is

normalised Haar measure on G.

4
 2 Topological Groups

2. Let G be an infinite discrete group. The measure on G defined by (E) =

|E|, E G, is Haar measure on G.

3. The usual Lebesgue measure on Rn is the Haar measure on Rn .

4. The circle group S 1 can be identified with interval [0, 2]. The usual Lebesgue
d
measure 2
is the normalised Haar measure on S 1 .
Now, we discuss a method of computing Haar measures on a class of locally

compact groups. Let G be locally compact group which satisfies the following
three properties:

(i) As topological space, G is an open subset of some Rn .

(ii) If x = (x1 , . . . , xn ), y = (y1 . . . , yn ) the xy is of the form F (x1 , . . . , yn ),

where all partial derivatives of F exist and are continuous on G G.

(iii) For each a G, the Jacobians J(a ), J(a ) depend only on a, where

a (x) = ax and a (x) = xa, x G.

Define the left and right Haar integrals on G by

Z
f (x)
IL (f ) = dx
S(x)

and
Z
f (x)
IR (f ) = dx f Cc (G),
D(x)
where dx denotes the Lebesgue measure on Rn and S(x), D(x) denote the ab-

solute value of determinants J(x ), J(a ) respectively.

5. R : the multiplicative group of nonzero real numbers. In this case S(a) = |a| =
D(a), a R . Hence
Z
f (x)
I(f ) = dx, f Cc (R ).
R |x|

5
 2 Topological Groups

is a Haar integral on R

6. Let
  
x y
G= |x, y R, x 6= 0
0 1
with matrix multiplication as group operation. An element of G can be repre-

sented as (x, y). Then (a,b) (x, y) = (ax, ay + b) and (a,b) (x, y) = (ax, bx + y).
Therefore,
S(a, b) = a2 , D(a, b) = |a|

Thus the integrals defined by

Z Z
f (x, y)
IL (f ) = dx dy,
x2
Z Z
f (x, y)
IR (f ) = dx dy, f Cc (G),
|x|
are respectively the left Haar integral and the right Haar integral on G. This
provides an example where left and right Haar integrals are different.

6
Chapter 3

Representation of Compact groups

3.1 Unitary representations

Let G be a topological group and V be normed vector space over R or C (V 6= {0}).

Let L(V) denote the algebra of bounded operators on V.

Definition 3. A representation of G on V is a map : G L(V), such that

(1) (g1 g2 ) = (g1 )(g2 ), (e) = I,

(2) for every v V, the map

G V,

g 7 (g)(v),

is continuous.

Let H be a Hilbert space. An operator A on H is said to be unitary if it is

invertible and A1 = A . A representation of G on H is said to be unitary if, for

every g G, (g) is a unitary operator. This can be written as:

g G, v H, k(g)(v)k = kvk.

A subspace W V is said to be invariant if, for every g G, (g)W = W . Consider

0 (g) = (g)|W , the restriction of (g) to W , we get a representation of G on W .
This 0 is said to be subrepresentation of .

Definition 4. The representation is said to be irreducible if the only closed invariant

subspaces are {0} and V.

7
 3 Representation of Compact groups

Remark: By definition, a one dimensional representation is closed.

Let (1 , V1 ) and (2 , V2 ) be two representation of G. If a continuous linear map A

from V1 to V2 satisfies the relation A1 (g) = 2 (g)A, for every g G, Then A is

called the intertwinning operator.

Definition 5. The representations (1 , V1 ) and (2 , V2 ) are said to be equivalent if

there exists an isomorphism A : V1 V2 which intertwins the representation 1 and

2 .

Lemma 3. Let be a representation of compact group G on a finite dimensional

vector space V. There exists on V a Euclidean inner product for which is unitary.

Proof. Let us choose arbitrarily on V a Euclidean inner product (|)0 and put
Z
hu, vi = ((g)u|(g)v)0 d(g)
G

where is the Haar measure on G. Now, it is easy to check that h, i is a Euclidean

inner product and that the representation is unitary with respect to this inner

product.

Corollary 1. Let be a representation of compact group G on a finite dimensional

vector space V.

1. For every invariant subspace there is an invariant complementary subspace.

2. The vector space V can be decomposed into a direct sum of irreducible invariant
subspaces:

V = V1 VN .

Proof. By lemma 3 there exists on V a Euclidean inner product for which the repre-
sentation is unitary. If W is the invariant subspace, then the orthogonal subspace

8
 3 Representation of Compact groups

W is invariant and complementary to W .

Let V1 be a non zero invariant subspace with minimal dimension. Then V = V1 V1 .

If V1 6= 0, let V2 be a non zero invariant subspace of V1 with minimal dimension.

One continues the process till subspace Vk 6= 0. Since dimension of V is finite, the
process stops necessarily.

Theorem 2. (Schurs Lemma):

(i) Let (1 , V1 ) and (2 , V2 ) be two finite dimensional irreducible representation of

a topological group G. Let A : V1 V2 be a linear map which intertwins the

representations 1 and 2 :
A1 (g) = 2 (g)A

for every g G. Then either A = 0, or A is an isomorphism.

(ii) Let be an irreducible C-linear representation of a topological group G on a

finite dimensional complex vector space V. Let A : V V be a C-linear map
which commutes to the representation :

A(g) = (g)A

for every g G. Then there exists C such that

A = I.

One establishes in the same way similar statements for representations of Lie
algebras.

3.2 Schurs orthogonality relations

Let G be compact group, and the normalised Haar measure of G. Let (, H) be a

unitary representation of G. For v H one considers the operator Kv of H defined

9
 3 Representation of Compact groups

by
Z
Kv w = hw, (g)vi(g)v d(g)
G

This can also be written as

Z
0
hKv w, w i = hw, (g)vihw0 , (g)vid(g).
G

Lemma 4. 1. Kv is bounded, kKv k kvk2 .

2. Kv is self-adjoint.

3. Kv commute to the representation : for every g G, Kv (g) = (g) Kv .

4. Kv is compact operator.

Proof. Note that, for (i),

kKv wk kvk2 kwk

(iv) For v H let Pv be the rank one operator defined by Pv w = hw, viv. It is
compact operator and, for v fixed, the map

G L(H),

g 7 P(g)v ,

is continuous for the norm topology. The operator Kv can be written as

Z
Kv = P(g)v (dg).
G

Since the space of compact operators is closed for the norm topology, the operator

Kv is compact.

Observe that if v 6= 0, hKv v, vi 0, hence Kv 6= 0.

Theorem 3. 1. Every unitary representation of a finite group contains a finite

dimensional subrepresentation.

10
 3 Representation of Compact groups

2. Every irreducible unitary representation of a compact group is finite dimen-

sional.

Proof. Let (, H) be a unitary representation of a compact group. The operator Kv is

self-adjoint, compact, and non-zero if v 6= 0. Therefore, it has a non-zero eigenvalue,

and the corresponding eigenspace is finite dimensional. This subspace is invariant

under the representation .

Theorem 4. Let be an irreducible unitary C-linear representation of a compact

group G on a complex Euclidean vector space H with dimension d . Then, for u, v
H,
Z
1
|h(g)u, vi|2 (dg) = kuk2 kvk2 .
G d

Proof. For v H, the operator Kv commutes to the representation . By Schurs

Lemma there is a (v) C such that Kv = (v)I. Hence,

Z
|h(g)u, vi|2 (dg) = (v)kuk2
G

By permuting u and v we get, (u)kvk2 = (v)kuk2 , hence (u) = 0 kuk2 , where 0

is a constant. Let (e1 , . . . , en ) be an orthonormal basis of H(n = d ).
n
X
|h(g)u, ei i|2 = kuk2 .
i=1

By integration on G we get
n Z
X
2
kuk = |h(g)u, ei i|2 (dg) = n0 kuk2 ,
i=1 G

hence 0 = n1 . Finally
Z
1
|h(g)u, vi|2 (dg) = kuk2 kvk2 .
G d

11
 3 Representation of Compact groups

Let ij (g) denote the entries of the matrix (g) with respect to the basis {ei },

ij (g) = h(g)ej , ei i.

From the preceding theorem we get Schurs orthogonality relations:

Z
1
ij (g)kl (g)(dg) = ik jl .
G d
Let M denote the subspace of L2 (G) generated by the entries of the representation
, that is by the functions of the following form

g 7 h(g)u, vi (u, v H).

Theorem 5. Let (, H)and( 0 , H0 ) be two irreducible unitary representations of a

compact group G which are not equivalent. Then M and M0 are two orthogonal
subspace of L2 (G):
Z
h(g), vih 0 (g)u0 , v 0 i(dg) = 0 (u, v H, u0 , v 0 H0 ).
G

Proof. Let A be a linear map from H into H0 and put

Z
B= 0 (g 1 )A(g)(dg).
G

The B is a linear map from H into H0 which intertwines the representations and

0 , B (g) = 0 (g) B. By Schurs Lemma, B = 0. Hence

Z
hBu, u i = hA(g)u, 0 (g)u0 i(dg) = 0
0
G

Take for A the rank one operator defined by

Au = hu, viv 0

then A(g)u = h(g)u, viv 0 , and

Z
h(g), vih 0 (g)u0 , v 0 i(dg) = 0
G

12
 3 Representation of Compact groups

Remark: It follows that two irreducible representations 1 and 2 of a compact

group G are equivalent if and only if the spaces M1 and M2 agree.

3.3 Peter-Weyl Theorem

Let G be a compact group, and let R denote the right regular representation of G on

L2 (G):
(R(g)f )(x) = f (xg).

Let (, H) be an irreducible representation of G, and let (e1 , . . . , en ) be an orthonormal

(1)
basis of H(n = d ). We put ij (x) = h(x)ej , ei i. Let M be the subspace of M

generated by the entries of the first line, that is by the functions x 7 1j (x), for
j = 1, ..., n. Observe that
n
X
1j (xg) = 1k (x)kj (g).
k=1

(1)
This shows that the subspace M is invariant under R. Furthermore, the map
n
X n
X
A: cj ej 7 cj 1j (x)
j=1 j=1

(1)
from H into M is an isomorphism, and intertwins the representations and R. In

fact, if u = nj=1 cj ej , then

P

n n n
! n n
!
X X X X X
A(g)u = A cj (g)ej = A cj ij (g)ei = ij (g)cj 1i (x)
j=1 j=1 i=1 i=1 j=1
Xn
= cj 1j (xg) = R(g)Au.
j=1

1 (i)
Furthermore kAuk2 = n
kuk2 . Let M denote the subspace of M generated by

coefficients of i-th line. Then

M = M(1) (n)
M ,

13
 3 Representation of Compact groups

and the restriction to M of the representation R is equivalent to = n.

By considering the rows instead of the lines one gets the same statement with, instead

of representation R, the regular left representation L:

(L(g)f )(x) = f (g 1 x)

Theorem 6. (Peter-Weyl Theorem): Let G

b be the set of equivalence classes of irre-

ducible unitary representations of the compact group G, and, for each G,

b let M

be the space generated by the coefficients of a representation in the class . Then

[
M
L2 (G) = M
G
b

2
L L
Remark: \ G
b M denotes the closure in L (G) of b M , which is space of
G

finite linear combinations of coefficients of finite dimensional representations of G.

Proof. We saw that the subspaces M are two by two orthogonal (Theorem 5). Put

[
M
H= M ,
G
b

and H0 = H . We will show that H0 = {0}. On the opposite, let us assume that
this is not the case. The space H0 is invariant under the representation R and closed.

By theorem 3, it contains a closed subspace Y 6= 0 which is invariant under R and

irreducible. The restriction of R to Y belongs to one of the classes . Let f Y, f 6= 0,
and put
Z
F (g) = f (xg)f (x)(dx) = hR(g)f, f i.
G

The function F belongs to M . We will see that it is orthogonal to M . Let (, V)

be a representation of the class , and u, v V. Then
Z Z Z
F (g)h(g)u, vi(dg) = f (xg)f (x)h(g)u, vi(dg)(dx),
G G G

14
 3 Representation of Compact groups

and, by putting xg = g 0 ,
Z Z 
0 0 0
= f (x) f (g )h(g )u, (x)vi(dg ) (dx) = 0.
G G
R
Therefore, F = 0, and, since F (e) = G
|f (x)|2 (dx), it follows that f = 0. This
yields a contradiction.

The following theorem follows directly from Peter-Weyl Theorem and from the

Schurs orthogonality relations.

Theorem 7. (Plancherel Theorem): Let f L2 (G). Then f is equal to the sum of

its Fourier series:
P
1. f (g) = G
b d tr(fb() (g)).
It holds in mean square convergence, that is in L2 sense.
R
|f (g)|2 (dg) = d kfb()k2 .
P
2. G G
b

3. The map f 7 fb is a unitary isomorphism from L2 (G) onto the space of se-

quences of operators A = (A ) (A L(H )), for which

X
kAk2 = d kA k2 < ,
G
b

and equipped with this norm.

3.4 Characters and central functions

Let G be a compact group.

Definition 6. A function f defined on G is said to be central if

f (gxg 1 ) = f (x) (g, x G).

Let be a representation of G on a finite dimensional complex vector space V.

15
 3 Representation of Compact groups

Definition 7. The character of is the function defined on G by

(g) = tr(g).

It is a central function which only depends on the equivalence class of . It satisfies

the following properties:

(e) = dim(V),

1 2 (g) = 1 (g) + (g),

(g) = (g 1 ) = (g).

Theorem 8. Let 1 and 2 be two finite dimensional irreducible representation of G.

Then they are equivalent if and only if they have same character:

1 (g) = 2 (g).

A finite dimensional representation of G is irreducible if and only if

Z
| (g)|2 (dg) = 1.
G

16
Chapter 4

The Groups SU (2) and SO(3)

4.1 Adjoint representation of SU (2)

The group SU (2) is made of matrices

 

g= , , C, ||2 + ||2 = 1.

The inverse of this matrix is

 
1
g = .

The group SU (2) is homeomorphic to the unit sphere of C2 , therefore compact,
connected, and simply connected. The matrices
     
i 0 0 1 0 i
X1 = , X2 = , X3 = ,
0 i 1 0 i 0

forms a basis of its Lie algebra su(2): every matrix T in su(2) can be uniquely written

as
T = t1 X1 + t2 X2 + t3 X3 , t1 , t2 , t3 R

The commutation relations are the followings

[X1 , X2 ] = 2X3 , [X2 , X3 ] = 2X1 , [X3 , X1 ] = 2X2 .

With respect to this basis the adjoint representation can be written, if T = t1 X1 +

t2 X2 + t3 X3 , as
0 2t3 2t2
ad(T ) = 2t3 0 2t1 .
2t2 2t1 0

17
 4 The Groups SU (2) and SO(3)

Theorem 9. The map Ad is a surjective morphism from SU (2) onto SO(3). Its
kernel, which is centre of SU (2), is equal to {e}.

Hence, (SU (2), Ad) is a covering of order 2 of SO(3).

4.2 Irreducible representations of SU (2)

In order to study the irreducible representations of SU (2), we will first consider

irreducible finite dimensional representations of complex Lie algebra g = sl(2, C).

The reason is that g = sl(2, C) is the complexified Lie algebra of the real algebra
su(2). In fact every Z sl(2, C) can be uniquely written as Z = X + iY , with

X, Y su(2).
Let Pm be the space of polynomials in two variables, with complex coefficients, and
homogeneous of degree m. Note that dimPm = m + 1. Let m be the representation

of SL(2, C) on Pm defined by

(m (g)f )(u, v) = f (au + cv, bu + dv),

if
 
a b
g= .
c d
In order to study the derived representation m = dm of sl(2, C) on Pm , we will use
the following basis of sl(2, C):
     
1 0 0 1 0 0
H= ,E = ,F = ,
0 1 0 0 1 0

for which the commutation relations are as follows:

[H, E] = 2E, [H, F ] = 2F, [E, F ] = H.

18
 4 The Groups SU (2) and SO(3)

The derived representation is obtained as follows

(m (exptH)f )(u, v) = f (et u, e1 v),

f f
m (h)f = u v ,
u v
(m (exptE)f )(u, v) = f (u, tu + v),
f
m (E)f = u ,
v
(m (exptF )f )(u, v) = f (u + tv, v),
f
m (F )f = v .
u

The monomials fj ,
fj (u, v) = uj v mj , j = 0, . . . , m,

form a basis of Pm , and

m (H)fj = (2j m)fj ,

m (E)fj = (m j)fj+1 ,

m (F )fj = jfj1 ,

Lemma 5. The representation m is irreducible.

Proof. Let W be a non zero invariant subspace in P)m. The restriction to W of

the operator m (H) admits at least one eigenvalue. Therefore one of the vectors fj
belongs to W. Letting the powers of m (E) and of m (F ) act on this vector and

hence, all the vectors fj belong to W. Hence W = Pm .

Theorem 10. Every irreducible finite dimensional C-linear representation of sl(2, C)

is equivalent to one of the representations m .

19
 4 The Groups SU (2) and SO(3)

Proof. Let be an irreducible C-linear representation of sl(2, C) on a finite dimen-

sional complex vector space V. Let 0 be an eigenvalue of (H) with minimal real

part, and 0 an associated eigenvector:

(H)0 = 0 0 .

We will show that the vector 1 = (E)0 , if non-zero, is an eigenvector of (H):

(H)1 = (H)(E)0

= (E)(H)0 + ([H, E])0

= 0 (E)0 + 2(E)0 = (0 + 2)1 .

Hence one constructs a sequence of vectors k = (E)k 0 , and

(H)k = (0 + 2k)k .

If these vectors are non-zero, being eigenvectors of (H) for distinct eigenvalues, they
are linearly independent. There exists an integer m such that

k 6= 0 if k m, and m+1 = 0.

Let us determine the action of (F ) on the vectors k . We show first that (F )0 = 0.

(H)(F )0 = (F )(H)0 + ([H, F ])0

= 0 (F )0 2(F )0 = (0 2)(F )0

Since 0 is an eigenvalue with minimal real part, (F )0 = 0. Let us show recursively

that

(F )k = k k1 ,

with
k = k(0 + k 1).

20
 4 The Groups SU (2) and SO(3)

For k = 1,

(F )1 = (F )(E)0

= (E)(F )0 + ([F, E])0

= (H)0 = 0 0 .

Assume that (F )k = k(0 + k 1)k1 ,then

(F )k+1 = (F )(E)k

= (E)(F )k + ([F, E])k

= k(0 + k 1)(E)k1 (H)k

= (k + 1)(0 + k)k .

Let us show now that 0 = m. For that observe that

tr(H) = tr[(E), (F )] = 0.

But,

tr(H) = 0 + (0 + 2) + + (0 + 2m)

= (m + 1)0 + m(m + 1)

= (m + 1)(0 + m).

Finally

(H)k = (2k m)k ,

(E)k = k+1 ,

(F )k = k(m k + 1)k1 .

It follows that the representation is equivalent to m .

The representation m (or more precisely its restriction to SU (2)) is an irreducible

representation of SU (2). In fact, if a subspace is invariant under m , it is also invariant

under the derived representation m .

21
 4 The Groups SU (2) and SO(3)

Theorem 11. Let be an irreducible representation of SU (2) on a finite dimensional

complex vector space V. Then is equivalent to one of the representations m .

Proof. The derived representation d extends linearly as C-linear representation of

sl(2, C) on V. The representation is irreducible.In fact let W 6= {0} be a subspace

of V, invariant under . Then W is invariant under Exp (X) = (exp X), for
X su(2). The group SU (2), being connected, is generated by exp(su(2)), therefore

W is invariant under SU (2), and W = V since is irreducible. By Theorem 10 the

representation is equivalent to one of the representation m . Hence there exists an
isomorphism A from V onto Pm such that

A(X) = m (X)A,

for every X sl(2, C). It follows that, for every X su(2),

A Exp(X) = Exp m (X)A;

this can be written

A(exp X) = m (expX)A,

and since exp(su(2)) generates SU (2), for every g SU (2),

A(g) = m (g)A.

Remark: Natural representation of SU (2) on C2 is equivalent to the representa-

tion 1 and the adjoint representation is equivalent to 2 .

4.3 Irreducible representations of SO(3)

Let T be irreducible representation of SO(3) on a finite dimensional vector space

V. If Ad denotes the adjoint representation which maps SU (2) onto SO(3), then

22
 4 The Groups SU (2) and SO(3)

= T Ad is an irreducible representation of SU (2) on V, hence is equivalent to

m for some m N. Since Ad(e) = I, necessarily m (e) = I. This happens if and

only if m is even. Conversely, if m = 2l, then the representation 2l of SU (2) factors

to the quotient SU (2)/{e}
= SO(3), and there exists a representation
2l of SO(3)
such that

2l Ad = 2l .

Finally we can state:

Theorem 12. Every finite dimensional irreducible representation of SO(3) is equiv-

alent to one of the representation
2l .

Recall that
 
i 0
X1 = ,
0 i
and that
1 0 0
Ad(expX1 ) = exp(ad X1 ) = 0 cos2 sin2
0 sin2 cos2.
The eigenvalues of
2l (Ad(expX1 )) are the numbers

e2ij , l j l.

We will show that the representation

2l can be realised on the space of harmonic
polynomials in three variables, homogeneous of degree l. Let be the Laplace oper-

ator
n
X 2f
= .
i=1
x2i

Definition 8. A C 2 function F defined in an open set in Rn is said to be harmonic

if f = 0.

A transformation g of GL(n, R) acts on the functions as follows

(r(g)f )(x) = f (xg)

23
 4 The Groups SU (2) and SO(3)

(a vector x Rn is identified to a 1 n matrix). Let D = P ( x ) be differential
operator with constant coefficients.

Lemma 6.

r(g)P ( )r(g 1 ) = Q( )
x x
where Q is the polynomial defined by Q() = P ((g 1 )T ).

Proof. The operator r(g)P ( x )r(g 1 ) is a constant coefficient differential operator:

(r(g)P ( )r(g 1 f ) = Q( )f,
x x

Take f (x) = ehx,i ( Rn ):

 
1
r(g)P ( )r(g f (x) = r(g)P ( )ehxg ,i
1
x x
1 T
= r(g)P ( )ehx,(g ) i
x
1 )T i
= r(g)P ((g 1 )T )ehx,(g

= P ((g 1 )T ).

The statement follows.

Corollary 2. If g O(n), r(g) = r(g),

and, if f is harmonic, then r(g)f is harmonic too.

Let H` be the space of harmonic polynomials in three variables, homogeneous of

degree l (l N).

Lemma 7. dim H` = 2l + 1.

Proof. A polynomial f in H` can be written

l
X xk 1
f (x1 , x2 , x3 ) = fk (x2 , x3 ),
k=0
k!

24
 4 The Groups SU (2) and SO(3)

and
l2 k l
xk1 2 fk 2 fk
 
X x 1
X
f = fk+2 + + ,
k=0
k! k=0
k! x22 x23
hence
2 fk 2 fk
 
fk+2 = + .
x22 x23
Therefore f is determined by the polynomials f0 and f1 ; f0 is an arbitrary polynomial
in two variables of degree l, and f1 is arbitrary of degree l 1. Therefore dim
H` = (l + 1) + l = 2l + 1.

Let Tl be the representation of SO(3) on H` defined by

(Tl (g)f )(x) = f (xg).

Theorem 13. The representation Tl is irreducible. It is equivalent to

2l .

Proof. By Corollary 1 of Chapter 3, the space H` can be decomposed as a direct sum

of irreducible subspaces
(1) (N )
H` = H` H`

(k) (k) (k)

H` is invariant under Tl and the restriction Tl of Tl to H` is an irreducible
(k)
representation of SO(3). Hence there exists an integer lk such that Tl is equivalent
(k)
2lk . Therefore dim H`
to = 2lk + 1, and necessarily lk l. The eigenvalues of
(k)
Tl (Ad(expX1 )) are the numbers

e2ij , lk j lk , 1 k N.

Let us consider polynomial f (x1 , x2 , x3 ) = (x2 + ix3 )l . It belongs to H` and

(Tl (Ad(expX1 )) f ) (x1 , x2 , x3 ) = f (x1 , x2 cos2 + x3 sin2, x2 sin2 + x3 cos2)

= (x2 cos2 + x3 sin2, ix2 sin2 + ix3 cos2)l

= e2il f (x1 , x2 , x3 ).

25
 4 The Groups SU (2) and SO(3)

Hence e2il is one of the eigenvalues of Tl (Ad(expX1 )). It follows that one of the
(k)
numbers lk is equal to l, and that H` = H` . Therefor Tl is irreducible and is

equivalent to
2l .

26
References
[1] S. C. Bagchi, S. Madan, A. Sitaram, U. B. Tiwari, A first course on representation

theory and linear Lie groups, (University Press, 2000).

[2] J. Faraut, Analysis on compact Lie groups an introduction, Eleventh discussion

meeting on Harmonic analysis, Bhubaneswar, India (2010).

[3] M. Sugiura, Unitary representations and Harmonic analysis an introduction, 2

ed. (North Holland, 1990).

27