Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 14
(Released 29.1.2025 – Discussed 5.2.2025)

E14.1 Let A : D(A) → H be a self-adjoint operator on a separable Hilbert space H. Let

u0 ∈ H and assume that
Z T
1
MT = |e−itA u0 ihe−itA u0 | * M∞
T 0
weakly in the Hilbert–Schmidt topology when T → ∞, namely Tr[MT B] → Tr[M∞ B] for
every Hilbert–Schmidt operator B on H.
(a) Prove that e−itA M∞ eitA = M∞ for all t ∈ R.
(b) Prove that M∞ commutes with A.
(c) Deduce that there exist λn ≥ 0 and an orthonormal family {un } of eigenfunctions
of A such that
∞
X
M∞ = λn |un ihun |.
n=1
(d) Conclude that if u0 is orthogonal to all eigenfunctions of A, then M∞ = 0.
Hint: E13.3 is helpful for (d). We used these ingredients in the proof of RAGE theorem.

E14.2 Let d ≥ 1, p ∈ [1, 2] and 1/p + 1/p0 = 1. Prove the dispersive estimate
− d2 ( p1 − p10 )
keit∆ ϕkLp0 (Rd ) ≤ |t| kϕkLp (Rd ) .

Hint: You may interpolate from two cases p = 1 and p = 2.

E14.3 Let V ∈ L2 (R3 , R) + Lp (R3 , R) for some 2 < p < 3. Prove that the wave operators

Ω± = lim e−itA e−it∆

t→±∞

are well defined on L2 (Rd ).

Hint: You can use Cook method and E14.1

E14.4 A : D(A) → L2 (Rd ) is a self-adjoint operator and ϕ ∈ L2 (Rd ) such that the limit

Ω± ϕ = lim e−itA e−it∆ ϕ

t→±∞

exists. Prove that the vectors Ω± ϕ are orthogonal to all eigenfunctions of A.

1
2

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 13
(Released 22.1.2025 – Discussed 29.1.2025)

E13.1 Let A : D(A) → H be a self-adjoint operator on a separable Hilbert space H.

(a) Prove that for every x0 ∈ H and t ∈ Rd , the vector x(t) = e−itA x0 satisfies that
d
hϕ, ix(t)i = hAϕ, x(t)i, ∀ϕ ∈ D(A).
dt
(b) Prove that kx(t)k = kx0 k for all t ∈ R if x0 ∈ H. Moreover, kAx(t)k = kAx0 k for
all t ∈ R if x0 ∈ D(A).
(c) Prove that for every x0 ∈ H, if y(t) satisfies
d
hϕ, iy(t)i = hAϕ, y(t)i, ∀ϕ ∈ D(A), ∀t ∈ R
dt
and y(0) = x0 , then y(t) = x(t) for all t ∈ R (namely the weak solution is unique).

E13.2 Consider the free Schrödinger evolution eit∆ on L2 (Rd ). Let f ∈ L1 (Rd ) ∩ L2 (Rd ).
(a) Prove that for all ε > 0, we have
|x|2
 
2 1
e−(it+ε)|2πk| fb(k) = G
\ ε ∗ f (k), with Gε (x) = exp −
(4π(it + ε))d/2 4(it + ε)
(b) Deduce that
|x − y|2
Z  
1
eit∆ f (x) = exp − f (y)dy.
(4πit)d/2 Rd 4it

E13.3 Let A be a self-adjoint operator on L2 (Rd ) and f ∈ span{eigenfunctions of A}.

Prove that for every ε ∈ (0, 1), there exists R = Rε > 0 such that for all t ∈ R,
Z Z
−itA 2
|(e f )(x)| dx ≥ (1 − ε) |f |2 .
|x|≤R Rd

d
E13.4 In H = L2 ([0, 1]), we have seen in the lecture that the momentum operator pα = i dx
is self-adjoint when defined on

D(pα ) = {f ∈ J : f absolutely continious with derivative in H, f (0) = eiα f (1)}

for a phase 0 ≤ α < 2π. Find an explicit expression for the one parameter group Uα (t) =
eitpα provided by Stone’s theorem.
Hint: Note that Uα is defined on all of H but taking the t-derivative must yield the
different pα with their respective domains!
 3

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 12
(Released 15.1.2025 – Discussed 22.1.2025)
In this homework we only consider the dimension d ≥ 3.

E12.1 Let {un }N 1 d N 2 d

n=1 ⊂ H (R ) such that {∇un }n=1 are orthonormal in L (R ).
√
(a) Prove that { −∆un }N
n=1 are orthonormal.
(b) Prove that for all E > 0 we have
N Z 2
X d−2
e2πik·x ûn (k)dk ≤ Cd E 2 .
n=1 |k|2 ≤E

(c) Prove that

N Z
X d
N= k∇un k2L2 ≥ Kd ρ d−2 .
n=1 Rd
Here Cd > 0 and Kd > 0 are constants depending only on d.

E12.2 Let V ∈ Ld/2 (Rd , R) and let A = −∆ + V (x) be the self-adjoint operator on L2 (Rd )
by Friedrichs’ method (c.f. E8.3).
(a) Use the characterization in E11.5 to prove directly that inf ess (A) ≥ 0.
(b) Assume that A has N negative eigenvalues and let W be the space spanned by
the corresponding eigenfunctions. Prove that hϕ, Aϕi ≤ 0 for all ϕ ∈ W and that
√ √
dim −∆W = N . Deduce that we can find N functions {un }N N
n=1 such that { −∆un }n=1
are orthonormal and that hun , Aun i < 0 for all n = 1, 2, ..., N .
(c) Use E12.1 to prove that the number of negative eigenvalues of A is bounded by
Cd Rd |V− |d/2 . This is the CLR inequality. Here V− (x) = min(V (x), 0).
R

E12.3 Consider A = −∆+|x|2 . We know that it is a self-adjoint operator with eigenvalues

0 < µ1 ≤ µ2 ≤ ... and µn → ∞.
(a) Use the CLR inequality to prove that for all λ > 0, the number of eigenvalues {µn }
in the interval (0, λ) is bounded from above by Cd λd .
(b) Prove that µn ≥ cd n1/d for all n = 1, 2, ...
Here Cd > 0, cd > 0 are constants depending only on d.

E12.4 (a) Let A and B be two self-adjoint operators on a separable Hilbert space H such
that D(A) ∩ D(B) is dense in H. Prove that if A ≥ B, then the number of negative
eigenvalues of A is smaller than or equal to the number of negative eigenvalues of B.
d/2
(b) Prove that V ∈ Lloc (Rd , R) and V (x) ≥ −λ|x|−2 for |x| large, with a constant
0 < λ < (d − 2)2 /4, then the Schrödinger operator −∆ + V (x) has only finitely many
negative eigenvalues.
4

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 11
(Released 8.1.2025 – Discussed 15.1.2025)

E11.1 Let A = Ma be the multiplication operator on L2 (Rd , µ), with µ a Borel locally
finite measure on Rd and a ∈ L∞ d
loc (R , R). Let λ ∈ R.
(a) Prove that λ is an eigenvalue of A if and only if µ(a−1 ({λ})) > 0.
(b) Prove that if λ ∈ σ(A) but λ is not an eigenvalue, then limε→0 µ(a−1 ((λ−ε, λ+ε)) =
0. Deduce that there exists a sequence εn → 0 such that

µ(a−1 ((λ − εn , λ + εn )\(λ − εn+1 , λ + εn+1 )) = 0, ∀n ∈ N.

(c) Prove that if λ is an isolated point of σ(A), namely here exists ε > 0 such that
σ(A) ∩ (λ − ε, λ + ε) = {λ}, then λ is an eigenvalue and

|A − λ|2 ≥ ε2 1Ker⊥ , Kλ = Ker(A − λ).

λ

Hint: We used these properties to prove Weyl’s criterion for spectrum. Here a−1 (U ) =
{x : a(x) ∈ U }. You can use σ(A) = ess-range(a) (see E3.5).

E11.2 Use Weyl’s criterion to prove that σ(−∆) = σess (−∆) = [0, ∞). Here −∆ is the
Laplacian in L2 (Rd ).

E11.3 Let A : D(A) → H be a self-adjoint operator on a separable Hilbert space and

B : D(A) → H be a symmetric operator. Assume that B is A-relatively compact, namely
B(A + i)−1 is a compact operator.
(a) Prove that
lim kB(A + in)−1 k = 0.
n→∞
(b) Prove that B is A-relatively bounded with any relative bound ε ∈ (0, 1), namely

kBuk ≤ εkAuk + Cε kuk, ∀u ∈ D(A).

Hint: You can show that An = (A + i)(A + in)−1 satisfies kAn uk → 0 for all u ∈ H.

E11.4 Let d ∈ N. Let V ∈ Lp (Rd ) with max(2, d/2) ≤ p < ∞ and additionally 2 < p if
d = 4. Prove that V is ∆-relatively compact.

E11.5 Let A : D(A) → H be a self-adjoint operator on an (infinite dimensional) separable

Hilbert space H which is bounded from below. Prove that

inf σess (A) = inf{lim inf hun , Aun i : {un } ⊂ D(A) orthonormal}
n→∞

= inf{lim inf hun , Aun i : {un } ⊂ D(A) normalized, un * 0}.

n→∞
 5

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 10
(Released 20.12.2024 – Discussed 8.1.2025)

E10.1 Show that the space of states of M at(2 × 2, C) as a convex space is the unit ball
in R3 . Hint: The identity I and the Pauli matrices σx , σy , σz for a basis of the hermitean
2 × 2 matrices. Show explicitly that the pure states are those that come from rays in C2 .

E10.2 Let ρ be a density matrix for M at(2 × 2, C). Work out explicitly the GNS con-
struction for ωρ (A) = Tr(ρA). What is its dimension? You can assume that ρ is diagonal
(why?) and you will have to distinguish two cases.

E10.3 Alice and Bob have each their separted labs each with their local Hilbert spaces
HA and Hb respectively which we assume to be finite dimesional. The whole “bipartite”
set up is described by the Hilbert space H = HA ⊗ HB and observables A = B(H ) =
B(HA ⊗ HB ). Let ω : A → C be a state.
Alice can only make measurements in her own lab, thus she has only access to observ-
ables of the form A ⊗ I for A ∈ B(HA ). Show that those form a sub-C*-algebra of A and
→ B(HA ⊗ HB ). Accordingly, we can define a state
we have thuse an inclusion B(HA ) ,−
on Alices’s subsystem B(HA ) via

ωA := ω|B(HA ) .

Express the density matrix for ωA in terms of the density matrix of ω!

6

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 9
(Released 11.12.2024 – Discussed 17.12.2024)

E9.1 Instead of two types of generators, one can also write the CCR algebra in terms of
operators W (z) for z ∈ C and

W (s + it) = e−ist/2 U (s)V (t)

with relations W (z1 )W (z2 ) = ei=(z1 z̄2 )/2 W (z1 + z2 ). (In fact, this can be defined for any
symplectic vector space, that is a real vector space with a non-degenerate anti-symmetric
bi-linear form σ. In this case only replace =(z1 z̄2 ) by the symplectic form σ). Show that
all W (z) are unitary. Thus their spectrum has to be contained in the unit circle.
Then compute W (u)W (z)W (−u) and use this to argue that the spectrum of W (z)
is either empty (this is impossible) or it is all of the unit circle. You can use that the
spectrum is invariant under unitary conjugation.
This tells you about the spectrum of W (z) − I. Finally conclude that for z 6= 0

kW (z) − Ik = 2

E9.2 Prove the Stone-von Neumann theorem. Assume that you have a regular irreducible
represenations π : CCR → B(H). Show first that
Z
1 2
P := d2 z π(W (z))e−|z| /4
2π
(defined in terms of matrix elements) is an orthogonal projection. Show also that
2 /4
P π(W (z))P = e−|z| P.

If Ω ∈ P H normalized, show that the span of all π(W (z)) for z ∈ C is invariant under
the action of all Weyl-operators W . From the irriducubility of the representation conclude
that P H is one dimensional.
Use this to construct the unitary equivalence of two regular irreducible representations
of CCR.

E9.3 Construct an irreducuble representation of CCR on a Hilbert space of functions

ψ : R → C where π(U (s)) is the multiplication operator eisx and (π(V (t))ψ)(x) = ψ(x+t),
both act as unitary operators. Different from the Schrödinger representation, let H contain
the function that is 1 everywhere (assume this to be normalized to k1k = 1). Show that
this representation is inequivalten to the Schrödinger representation.
 7

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 8
(Released 4.12.2024 – Discussed 11.12.2024)

E8.1 Let d ≥ 1, s > d/2 and f ∈ H s (Rd ). Prove that f is Hölder’s continuous, namely
|f (x) − f (y)|
sup ≤ Ckf kH s (Rd )
x6=y |x − y|ε
for some constants ε ∈ (0, 1) and C = C(d, s, ε) independent of f .
Hint: You can use the inverse Fourier and |1 − eia | ≤ min(2, |a|) for all a ∈ R.

In the following we always consider the Schrödinger operator A = −∆+V (x) on L2 (Rd )
with domain D(A) = Cc2 (Rd ). Here V : Rd → R is a real-valued potential.

E8.2 Let d = 2.
(a) Prove that for all q ∈ [2, ∞) we have the Sobolev inequality

kukLq (R2 ) ≤ Cq kukH 1 (R2 ) , ∀u ∈ H 1 (R2 ).

Hint: You can mimic the proof of the case d ≥ 3.

(b) Prove that −∆ + V is bounded from below if V ∈ Lp (R2 ) + L∞ (R2 ) for p > 1.

E8.3 Let d ≥ 3.
(a) Prove that the Sobolev inequality kukL2∗ (Rd ) ≤ Ck∇ukL2 (Rd ) , with 2∗ = 2d/(d − 2),
is equivalent to the fact that −∆ + V ≥ 0 if kV kLd/2 small enough.
(b) Prove that −∆ + V is bounded from below if V ∈ Ld/2 (Rd ) + L∞ (Rd ).

E8.4 Prove that for all dimensions d ≥ 1 and V ∈ L1+d/2 (Rd ) we have
Z
−∆ + V ≥ −C |V (x)|1+d/2 (x).
Rd

E8.5 Let {un }∞ 1 d 2 d

n=1 ⊂ H (R ) be orthonormal functions in L (R ) such that
Z
|x|2 |un (x)|2 dx ≤ C, ∀n ∈ N.
Rd
Prove that Z
lim |∇un (x)|2 dx = ∞.
n→∞ Rd
8

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 7
(Released 27.11.2024 – Discussed 4.12.2024)

E7.1 Let s, d ∈ N and f ∈ H s (Rd ). Let χ ∈ Cc∞ (Rd ) such that χ(x) = 1 if |x| ≤ 1 and
χ(x) = 0 if |x| ≥ 2. Define
Z
fn (x) = χ(x/n) nd χ̂(n(x − y))f (y)dy.
Rd

Prove that fn ∈ Cc∞ (Rd ) and fn → f in H s (Rd ).

E7.2 Prove that if f, g ∈ H 1 (Rd ), then we have the integration by part

Z Z
(∂j f )(x)g(x)dx = − f (x)(∂j g)(x)dx, j = 1, 2, ..., d.
Rd Rd
Hint: You first consider f, g smooth, and then use a density argument (thanks to E7.1).

E7.3 (a) Let d ≥ 1. Prove that the equation

 (d − 2)2 
−∆− f (x) = 0
4|x|2
has a solution f (x) = |x|−α for a suitable α > 0. Deduce that
(d − 2)2
Z Z
|∇u(x)|2 dx ≥ 2
|u(x)|2 dx
Rd Rd 4|x|
for all u ∈ Cc∞ (Rd ) such that u = 0 in a neighborhood of 0.
(b) Let d ≥ 3. Prove that the above Hardy inequality holds true for all u ∈ H 1 (Rd ).
Hint: You can use the Perron-Frobenius principle for (a) and a density argument for (b).

E7.4 (a) Use the hydrogen atom lower bound to prove the following inequality
Z Z 2
1
|∇u(x)|2 dx ≥ |u(x)|2 dx , ∀u ∈ H 1 (R3 ), kukL2 (R3 ) = 1.
R3 R3 |x|
(b) Use the above inequality to prove the Heisenberg uncertainty principle
Z  Z 
2
|∇u(x)| dx |x| |u(x)| dx ≥ 1, ∀u ∈ H 1 (R3 ), kukL2 (R3 ) = 1.
2 2
R3 R3

Note: The optimal constant in the Heisenberg uncertainty principle is d2 /4 = 9/4.

E7.5 Let d ≥ 3 and 1 ≤ p ≤ ∞. We denote by Cd,p a constant depending only on d, p.

(a) Prove that if kukLp (Rd ) ≤ Cd,p k∇ukL2 (Rd ) for all u ∈ H 1 (Rd ), then p = 2∗ = 2d
d−2 .
(b) Prove that if kukLp (Rd ) ≤ Cd,p kukH 1 (Rd ) for all u ∈ H 1 (Rd ), then 2 ≤ p ≤ 2∗ .
 9

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 6
(Released 20.11.2024 – Discussed 27.11.2024)

E6.1 Let A : D(A) → H be a symmetric operator on a Hilbert space H. Let A : D(A) be

the closure of A.
(a) Prove that A is a closed operator, namely the graph {(x, Ax) : x ∈ D(A)} is close
in H ⊕ H. Here (x, y)H⊕H = kxk + kyk with the norm k.k in H.
(b) Prove that if A is a closed operator, namely if the graph {(x, Ax) : x ∈ D(A)} is
close in H ⊕ H, then A = A.

E6.2 Let A : D(A) → H be an operator on a Hilbert space H and A ≥ 1. Prove that

hx, yiQ = hx, AyiH , ∀x, y ∈ D(A)

k.kQ
is an inner product on D(A). (We used this to defined Q(A) = D(A) .)

E6.3 Let A : D(A) → H be a symmetric operator on a Hilbert space (H, k.k) with A ≥ 1.
Let AF be the Friedrichs extension of A.
√
(a) Prove that the quadratic form domain Q(A) is the same with the domain D( AF ).
(b) Prove that

inf hx, Axi = inf kxk2Q(A) = inf hx, AF xi.

x∈D(A),kxk=1 x∈Q(A),kxk=1 x∈D(AF ),kxk=1

E6.4 Let A : D(A) → H be a symmetric operator on a Hilbert space H such that A ≥ 1.

Prove that if A is essentially self-adjoint, namely A is self-adjoint, then A coincides with
the Friedrichs extension of A.

E6.5 Let A : D(A) → H be a symmetric operator on a Hilbert space H. Prove that A is

self-adjoint if and only if C = (A + i)(A − i)−1 is well-defined as a unitary operator on H.
(This is called the Cayley transform.)
10

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 5
(Released 13.11.2024 – Discussed 20.11.2024)

E5.1 Let A : D(A) → H be a symmetric operator on a Hilbert space. Let i2 = −1.

(a) Prove that k(A ± i)xk ≥ kxk for all x ∈ D(A).
(b) Prove that if Ran(A + i) = H, then (A + i)−1 is a bounded operator.
(c) Prove that A is self-adjoint if and only if Ran(A ± i) = H.

E5.2 Let A : D(A) → H be a symmetric operator on a Hilbert space. Prove that A is

self-adjoint if and only if its spectrum is real, namely σ(A) ⊂ R.

E5.3 Let A : D(A) → H be a self-adjoint operator on a Hilbert space. Then S = (A + i)−1

is a bounded operator by Problem E5.1. Prove that S is a normal operator, namely
SS ∗ = S ∗ S. (Note: You are not allowed to use the spectral theorem here, since we used
this fact to prove the spectral theorem for self-adjoint unbounded operators.)

E5.4 Let A : D(A) → H be a self-adjoint operator on a Hilbert space H. Define the

spectral projection
An = 1(A ≤ n), n∈N
by functional calculus (i.e. via the spectral theorem). Prove that An converges to the
identity 1 strongly, namely An u → u strongly for every u ∈ H as n → ∞.

E5.5 (a) Let (X, k.kX ) and (X, k.kY ) be two Banach spaces (the same vector space X
with two different norms). Assume that kukX ≤ kukY for all u ∈ X. Prove that there
exists a constant C > 0 such that kukY ≤ CkukX for all u ∈ X.
(b) Let (Ω, µ) be a sigma-finite measure space. Let a : Ω → R be a measurable function
such that af ∈ L2 (Ω, µ) for all f ∈ L2 (Ω, µ). Prove that a ∈ L∞ (Ω, µ).
(c) Let A : D(A) → H be a self-adjoint operator on a Hilbert space H. Prove that if
D(A) = H, then A is a bounded operator. (This is called the Hellinger–Toeplitz theorem.)
 11

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 4
(Released 8.11.2024 – Discussed 13.11.2024)

E4.1 Let A be a self-adjoint compact operator on a Hilbert space with the representation
X
A= λn |un ihun |
n≥1

where {un } is an orthonormal family and λn ∈ R, λn → 0. Prove that

[
σ(A) = {0} {λn : n ≥ 1}.

E4.2 Let B be a bounded operator on a Hilbert space H such that

kuk ≤ CkBuk, kuk ≤ CkB ∗ uk, ∀u ∈ H

with a constant C > 0 independent of u. Prove that B −1 is a bounded operator.

E4.3 In the lecture, we have proved that if A is a self-adjoint bounded operator on a

Hilbert space H, then
sup |hu, Aui| ≤ sup |σ(A)| ≤ kAk.
kuk≤1
Use these inequalities to prove that sup |σ(A)| = kAk (this is what we actually need to
prove the Spectral theorem). Hint: You can apply the above inequalities to A and A2 .

E4.4 Let A be a self-adjoint bounded operator on a Hilbert space H. For any polynomial
f (t) = N
P j
PN j
j=1 αj t , αj ∈ C, we defined f (A) = j=1 αj A .
(a) Prove that for any polynomial, σ(f (A)) = f (σ(A)), and then deduce that

||f (A)|| = sup |f (t)|.

t∈σ(A)

Hint: You can consider the factorized form of λ − f (t) (fundamental theorem of algebra).
(b) Prove that we can extend the definition f (A) for any f ∈ C(σ(A), C) by Weierstrass
theorem. Moreover, we have

f (A)g(A) = (f g)(A), ∀f, g ∈ C(σ(A), C).

This means that f 7→ f (A) is a C ∗ - isomorphism from C(σ(A), C) to B(H).

E4.5 Let A be a self-adjoint bounded operator on a Hilbert space H. Let u ∈ H and let
µu be the spectral measure associated with A. Prove that

µu (σ(A)) = kuk2 .
12

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 3
(Released 30.10.2024 – Discussed 6.11.2024)

E3.1 Let A : H → H be a trace class operator on a Hilbert space. Let {xn }n≥1 be an
orthonormal basis for H (not necessary the eigenfunctions of A). Prove that
X
Tr(A) = hxn , Axn i.
n≥1

E3.2 Let A : H → H be a bounded self-adjoint operator on a Hilbert space.

(a) Assume that A ≥ 0 and that there exists an orthonormal basis {xn }n≥1 such that
X
hxn , Axn i < ∞.
n≥1

Prove that A is trace class.

(b) Can we relax the condition A ≥ 0 in (a)? Namely, if we only assume that A
is bounded self-adjoint and that there exists an orthonormal basis {xn }n≥1 satisfying
P
n |hxn , Axn i| < ∞, then can we conclude that A is trace class?

E3.3 Let A ≤ 0 be a compact operator on a Hilbert space H. Prove that its eigenvalues
λ1 ≤ λ2 ≤ ... (counting multiplicity) satisfy the min-max principle

λn = min max hu, Aui, ∀n = 1, 2, ...

M ⊂H u∈M
dimM =n kuk=1

Hint: The minimum is attained when M is the space spanned by the first n eigenfunctions.

E3.4 Prove that if A is a bounded operator on a Hilbert space and B ∈ S p , the Schatten
space with 1 ≤ p ≤ ∞, then both AB and BA belong to S p .
Hint: You can estimate the singular values of X via those of X ∗ X or XX ∗ and use E3.3.

E3.5 Let (Ω, µ) be a σ-finite measure space. Let a : Ω → C be a measurable. Define the
multiplication operator Ma on L2 (Ω, µ) by

(Ma f )(x) = f (x)a(x), D(Ma ) = {f ∈ L2 : af ∈ L2 }.

Prove the following statements:

(a) The spectrum of Ma is equal to the essential range of a.
(b) Ma is bounded if and only if a is bounded. Moreover, in this case kMa k = kakL∞ .
(c) Ma is self-adjoint if and only if a is real-valued almost everywhere.
 13

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 2
(Released 25.10.2024 – Discussed 30.10.2024)

E2.1 Let A : D(A) → H be a densely defined operator on a Hilbert space. Prove that the
following statements are equivalent:
(a) A is a symmetric operator.
(b) hx, Axi ∈ R for all x ∈ D(A).
(c) A ⊂ A∗ .

E2.2 Let A : H → H be a linear operator on a Hilbert space.

(a) Prove that A is bounded if and only if A∗ is bounded, and in this case kAk = kA∗ k.
(b) Prove that A is compact if and only if A∗ is compact.

E2.3 Let A : H → H be a linear operator on a Hilbert space. Prove that A is bounded

if and only if A maps weak convergence to weak convergence, namely if xn * x weakly
then Axn * Ax weakly.

E2.4 Let A : H → H be a bounded operator on a Hilbert space. Prove that A∗ A and AA∗
have the same non-zero eigenvalues with the same multiplicities. Moreover, A maps each
eigenspace of A∗ A with a non-zero eigenvalue to the corresponding eigenspace of AA∗ .

E2.5 Let A : H → H be a bounded operator on a Hilbert space such that for every
orthonormal sequence {xn }, we have Axn → 0 strongly. Can we conclude that A is a
compact operator?
14

Mathematisches Institut LMU Mathematical Quantum Mechanics

Robert Helling & Phan Thành Nam Winter 2024-2025

Homework Sheet 1
(Released 18.10.2024 – Discussed 23.10.2024)

E1.1 Let (H, k.k) be a Banach space. Prove that it is a Hilbert space if and only if the
norm satisfies
ku + vk2 + ku − vk2 = 2(kuk2 + kvk2 ), ∀u, v ∈ H.
How can we write the inner product hu, vi in terms of the norm of suitable vectors?

E1.2 Let V be a (complex) vector space and let Q : V × V → C be a bilinear form (in
particular Q(u, v) is linear in v and anti-linear in u). Prove that if Q(u, u) ≥ 0 for all
v ∈ H, then we have the Cauchy–Schwarz inequality
p p
|Q(u, v)| ≤ Q(u, u) Q(v, v), ∀u, v ∈ V.

E1.3 Let V be a dense subspace of a separable Hilbert space H. Prove that there exists
an orthonormal basis {un }n≥1 for H such that un ∈ V for all n ≥ 1.

E1.4 (a) Argue that for symmetric operators A, B on a Hilbert space H, we have
1
h∆Ai · h∆Bi ≥ hIm[A, B]i2
4
where h∆Ai = hψ, A ψi − hψ, Aψi , h∆Bi = hψ, B 2 ψi − hψ, Bψi2 and hIm[A, B]i =
2 2

Imhψ, [A, B]ψi with a suitable vector ψ ∈ H which belongs to the domain of all relevant
operators. When does the equality occur?
(b) For A = −i∂x (momentum operator) and B = x (multiplication operator) on
H = L2 (R), what do we get from (a)? Determine all functions Ψ ∈ L2 (R) satisfying the
corresponding equality.

E1.5 Let p, q ∈ (1, ∞) such that 1/p + 1/q = 1.

(a) Prove Young’s identity
ap bq
 
= sup ab − , ∀a ∈ [0, ∞).
p b∈[0,∞) q
(b) Deduce Hölder’s inequality
Z
kf kLp = sup fg , ∀f ∈ Lp (Rd )
kgkLq ≤1 Rd

and the triangle inequality

kf + gkLp ≤ kf kLp + kgkLp , ∀f, g ∈ Lp (Rd ).