Lie Algebras and Lie Groups Homework II
Pádraig Sheehy
Due: 27-09-2019
Problem 1
Show that the Lie algebra of the unitary group
U (n) = {A ∈ GLn (C) | AA∗ = In }, A∗ = AT ,
is the unitary Lie algebra
u(n) = {A ∈ gln (C) | A + A∗ = 0}.
Solution:
Let G = U (n).
Consider L(G) = {A ∈ gln (C) | etA ∈ G ∀ t ∈ R}
A ∈ L(G) =⇒ etA ∈ L(G) ∀ t ∈ (R)
=⇒ (etA )(etA )∗ = In
d tA tA∗
=⇒ (e )(e ) t=0 = 0
dt
d
(I + tA + O(t2 ))(I + tA∗ + O(t2 )) t=0 = 0
=⇒
dt
d
(I + tA + tA∗ + O(t2 )) t=0 = 0
=⇒
dt
=⇒ A + A∗ = 0
=⇒ A ∈ u(n)
A ∈ u(n) =⇒ A + A∗ = 0
=⇒ A and A∗ commute
∗ ∗ ∗
=⇒ (etA )(etA ) = (etA )(etA ) = et(A+A )
= e0 = In
=⇒ etA ∈ G
=⇒ A ∈ L(G)
Thus, the Lie Algebra of the unitary group is unitary Lie Algebra.
Problem 2
1. Show that the map
1 x
(R, +) → GL2 (R), x 7→
0 1
is a group homomorphism.
2. Deduce that (R, +) is a linear group.
3. Determine the Lie algebra of the group (R, +).
1
Solution:
x 1
1. Let f : (R, +) → GL2 (R) : x 7→ .
1 0
Then we have:
1 x+y 1 x 1 y
f (x + y) = = = f (x)f (y)
0 1 0 1 0 1
Thus, f is a group homomorphism.
2. (R, +) has an embedding in GL2 (R) and is closed under the group operation (since the sum of two real
numbers is real). Group homomorphisms preserve group structure and thus (R, +) is a closed subgroup
of GL2 (R) which is, by definition, a linear group.
3. We know that the Lie algebra of G = (R, +) is the tangent space at the identity element i.e. T0 G. However,
we know T0 G = R. Furthermore since G is Abelian, the Lie bracket is:
[A, B] = 0, ∀ A, B ∈ R
Thus, the Lie algebra of G = (R, +) is R equipped with the zero bracket.
Problem 3
Consider the special unitary Lie algebra
su(2) = {A ∈ gl2 (C) | A + A∗ = 0, trA = 0},
over R and a bilinear form
1
tr(AB ∗ ),
(A, B) = A, B ∈ su(2).
2
Show that this form is symmetric. Show that su(2) is 3-dimensional and has an orthonormal basis iσ1 , iσ2 , iσ3 ,
where:
0 1 0 −i 1 0
σ1 = , σ2 = , σ3 =
1 0 i 0 0 −1
are called Pauli matrices.
Solution:
For all A, B ∈ su(2) we have:
1
(A, B) = tr(AB ∗ )
2
1
= tr((−A∗ )(−B))
2
1
= tr(A∗ B)
2
1
= tr(BA∗ )
2
= (B, A)
Thus, this bilinear form is symmetric.
a b
Consider A = ∈ su(2).
c d
tr(A) = 0 =⇒ a = −d.
A∗ = −A =⇒ b = c̄, −a = ā.
2
So, A will be determined by a choice of a ∈ iR and b ∈ C, and the dimension of su(2) over R is at most 3.
We have that for k = 1, 2, 3:
tr(iσk ) = 0
and (iσk )∗ = −iσk∗ = −(iσk )
=⇒ iσ1 , iσ2 , iσ3 ∈ su(2).
We can check orthonormality of iσ1 , iσ2 , iσ3 by directly computing the inner product for each pair.
Firstly, note that the bilinear form from the first part of this problem is an inner product since:
1
(A, A) = Σi,j |aij |2 > 0, ∀A ∈ su(2)
2
Computing the inner products:
1 1 1 0
(iσ1 , iσ1 ) = tr(iσ1 (−i)σ1∗ ) = tr =1
2 2 0 1
1 1 1 0
(iσ2 , iσ2 ) = tr(iσ2 (−i)σ2∗ ) = tr =1
2 2 0 1
1 ∗) 1 1 0
(iσ3 , iσ3 ) = tr(iσ3 (−i)σ3 = tr =1
2 2 0 1
1 1 i 0
(iσ1 , iσ2 ) = tr(iσ1 (−i)σ2∗ ) = tr =0
2 2 0 −i
1 1 0 −1
(iσ1 , iσ3 ) = tr(iσ1 (−i)σ3∗ ) = tr =0
2 2 1 0
1 1 0 i
(iσ2 , iσ3 ) = tr(iσ2 (−i)σ3∗ ) = tr =0
2 2 i 0
We get that they are indeed orthonormal. We have 3 orthogonal vectors in a vector space with dimension at
most 3, and thus conclude that su(2) has dimension 3 and iσ1 , iσ2 , iσ3 form a basis.
Problem 4
Show that:
1. If g ∈ SU (n) and A ∈ su(n), then Adg (A) := gAg −1 ∈ su(n).
2. The map
Ad : SU (n) → GL(su(n)) : g 7→ Adg
is a group homomorphism.
3. If g ∈ SU (2), then Adg : su(2) → su(2) preserves the symmetric bilinear form from the previous exercise.
4. There is a group homomorphism Ad : SU (2) → SO(su(2)) ∼
= SO(3) (special orthogonal group consisting
of orthogonal operators having determinant 1).
Solution:
1. For all g ∈ SU (n) and A ∈ su(n), we have that:
Adg (A) + (Adg (A))∗ = gAg −1 + (gAg −1 )∗
= gAg −1 + ((g −1 )∗ A∗ g ∗ )
= gAg −1 + ((gA∗ g −1 )
= g(A + A∗ )g −1
= g(0)g −1
=0
3
and that:
tr(Adg ) = tr(gAg −1 )
= tr(Ag −1 g)
= tr(A)
=0
∴ Adg (A) ∈ su(n)
2. For all g, h ∈ SU (n) and A ∈ su(n), we have:
Ad(gh)(A) = Adgh (A)
= (gh)(A)((gh)−1 )
= (gh)(A)(h−1 )(g −1 )
= ghAh−1 g −1
= g(h(A)h−1 )g −1
= Adg (Adh (A))
Thus, the map
Ad : SU (n) → GL(su(n)) : g 7→ Adg
is a group homomorphism.
3. For all g ∈ SU (2) and A, B ∈ su(2), we have:
1
(Adg (A), Adg (B)) = tr(Adg (A)(Adg (B))∗ )
2
1
= tr(gAg −1 (gBg −1 )∗ )
2
1
= tr(gAg −1 (g −1 )∗ B ∗ g ∗ )
2
1
= tr(gAg −1 gB ∗ g −1 )
2
1
= tr(gAB ∗ g −1 )
2
1
= tr(AB ∗ g −1 g)
2
1
= tr(AB ∗ )
2
= (A, B)
And so Adg : su(2) → su(2) preserves the symmetric bilinear form from the previous exercise.
4. For all g ∈ SU (2), we have from part 3. that Adg preserves the inner product and is thus orthogonal.
i.e. Adg ∈ O(su(2))
=⇒ det(Adg ) = ±1
However, SU (2) is path-connected and Ad : SU (2) → GL(su(2)) : g 7→ Adg is continuous. Thus, the
image of Ad is path-connected. However, as a set of orthogonal matrices O(su(2)) has two connected
components:
SO(su(2)) = {F ∈ O(su(2) | det(F ) = 1}
N O(su(2)) : = {F ∈ O(su(2) | det(F ) = −1}
Thus we know that either:
Im Ad ⊂ SO(su(2)) or Im Ad ⊂ N O(su(2))
and so:
det(Adg ) = +1, ∀g ∈ SU (2) or det(Adg ) = −1, ∀g ∈ SU (2)
4
But, part 2. tells us that
Ad : SU (2) → GL(su(2)) : g 7→ Adg
is a group homomorphism and thus maps the identity 1 ∈ SU (2) to the identity map Ad1 on su(2).
Therefore,
det(Ad1 ) = +1
and thus we have that:
det(Adg ) = 1 ∀g ∈ SU (2)
So the image of Ad lies in SO(su(2)) and as such there is a group homomorphism Ad : SU (2) → SO(su(2)).
Noting that su(2) is a 3-dimensional real vector space, we get that there is a group homomorphism
Ad : SU (2) → SO(su(2)) ∼
= SO(3), as required.
