Scribd
Upload
8 views

notes3

GRB Physics For Competitions

Uploaded by

Shakir Khattak
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

notes3

GRB Physics For Competitions

Uploaded by

Shakir Khattak
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Topics in Representation Theory: The

Adjoint Representation

1 The Adjoint Representation


Besides the left and right actions of G on itself, there is the conjugation action

c(g) : h → ghg −1

Unlike the left and right actions which are transitive, this action has fixed points,
including the identity.
Definition 1 (Adjoint Representation). The differential of the conjugation
action, evaluated at the identity, is called the adjoint action

Ad(g) = c∗ (g)(e) : Te G → Te G

Identifying g with Te G and invoking the chain rule to show that

Ad(g1 ) ◦ Ad(g2 ) = Ad(g1 g2 )

this gives a homomorphism

Ad(g) : G → GL(g)

called the adjoint representation.


So, for any Lie group, we have a distinguished representation with dimension
of the group, given by linear transformations on the Lie algebra. Later we will
see that there is an inner product on the Lie algebra with respect to which these
transformations are orthogonal.
For the matrix group case, the adjoint representation is just the conjugation
action on matrices
Ad(g)(y) = gY g −1
since one can think of the Lie algebra in terms of matrices infinitesimally close
to the unit matrix and carry over the conjugation action to them.
Given any Lie group representation

π : G → GL(V )

taking the differential gives a representation

dπ : g → End(V )

defined by
d
dπ(X)v = (π(exp(tX))v)|t=0
dt

1
for v ∈ V . Using our previous formula for the derivative of the differential
of the exponential map, we find for the adjoint representation Ad(g) that the
associated Lie algebra representation is given by
d d
ad(X)(Y ) = (c(exp(tX))∗ (Y ))|t=0 = (Ad(exp(tX))(Y ))|t=0 = [X, Y ]
dt dt
For the special case of matrix groups we can check this easily since expanding
the matrix exponential gives

etX Y e−tX = Y + t[X, Y ] + O(t2 )

So associated to Ad(G), the adjoint representation of the Lie group G on g,


taking the derivative we have ad(g), a Lie algebra representation of g on itself

ad(g) : X ∈ g → ad(X) = [X, ·] ∈ End(g)

An important property of the adjoint representation is that there is an invari-


ant bilinear form on g. This is called the “Killing form”, after the mathematician
Wilhelm Killing (1847-1823). Killing was responsible for many important ideas
in the theory of Lie algebras and their representations, but not for the Killing
form. Borel seems to have been the first to use this terminology, but now says
he can’t remember what inspired him to use it[1].
Definition 2 (Killing Form). The Killing form on g is the bilinear form

K(X, Y ) = T r(ad(X) ◦ ad(Y ))

Here X ∈ g, Y ∈ g and T r is the trace.


The Killing form has the following important properties:
Theorem 1. (i) It is symmetric

(X, Y ) = K(Y, X)

(ii) It is invariant under the adjoint action, i.e. for all g ∈ G

K(Ad(g)X, Ad(g)Y ) = (X, Y )

(iii) For each Z ∈ g, the endomorphism ad(Z) is skew-symmetric with re-


spect to the Killing form, i.e.

K(ad(Z)X, Y ) = −K(X, ad(Z)Y )

Proof (i) This follows from the symmetry of the trace

T r(AB) = T r(BA)

(ii) We have
ad(Ad(g)X) = Ad(g)ad(X)(Ad(g))−1

2
so

K(Ad(g)X, Ad(g)Y ) = T r(Ad(g)ad(X)(Ad(g))−1 ◦ Ad(g)ad(Y )(Ad(g))−1 )


= T r(Ad(g)ad(X)ad(Y )(Ad(g))−1 )
= T r((Ad(g))−1 Ad(g)ad(X)ad(Y ))
= T r(ad(X)ad(Y ))
= K(X, Y )

by cyclicity of the trace.


(iii)If we take
g = exp(tZ)
then differentiating (ii) at t = 0 gives

K(ad(Z)X, Y ) + K(X, ad(Z)Y ) = 0

We will see later that for semi-simple compact Lie groups, the Killing form
is non-degenerate and its negative is a positive definite inner product on g.
As usual, the simplest example to keep in mind is G = SU (2). In this
case the Lie algebra su(2) has a basis of skew-hermitian 2 by 2 matrices, these
span the tangent space R3 to the group at the identity, which is that tangent
space to S 3 . The adjoint group action on this R3 is an action by orthogonal
tranformations in SO(3). The Killing form is just the negative of the standard
inner product on
Using Pauli matrices, a standard basis is:
   
0 i/2 0 −1/2
S1 = −iσ1 /2 = , S2 = −iσ2 /2 = ,
−i/2 0 1/2 0
 
−i/2 0
S3 = −iσ3 /2 =
0 i/2
and these satisfy
[Sj , Sk ] = jkl Sl
where jkl is a symbol antisymmetric in its indices and such that it is 1 for 123
and all cyclic permutations of the indices (123).
Writing an element X ∈ su(2) as

X = x1 S1 + x2 S2 + x3 S3

the adjoint group action on X by an element g ∈ SU (2) is the map

X → gXg −1

and this takes the vector


x = (x1 , x2 , x3 )

3
to a new vector
x0 = (x01 , x02 , x03 )
where
x0 = Ax
for some matrix A ∈ SO(3).
The adjoint action of the Lie algebra on itself is given by the commutation
relations for Sk
ad(Sk ) : X → [Sk , X]
One can work out what this means explicitly in terms of matrices, for instance
 
0 0 0
ad(S1 ) = 0 0 1
0 −1 0

References
[1] Borel, A., Essays in the History of Lie Groups and Algebraic
Groups,American Mathematical Society, 2001, Page 5.

You might also like