PHYS-C0252 - Quantum Mechanics 2021

Lecture notes

Lecturers: Mikko Möttönen (first half), Tapio Ala-Nissilä (second half),

Assistants: András (Marci) Gunyhó, Niko Savola

March 26, 2022

Contents

Contents ii

1 Lecture 1 1
1.1 Intended learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Hilbert space and kets vectors . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Bra vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.5 Linear operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.6 Outer product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Lecture 2 4
2.1 Intended learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Bases of H . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 States vs. vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Operators vs. matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.5 Adjugate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.6 Properties of adjugate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.7 Eigenvalues and eigenstates . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.8 Hermitian operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.9 Postulates of quantum mechanics . . . . . . . . . . . . . . . . . . . . . . . . 8
Postulate I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Postulate II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Postulate III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Postulate IV: Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Postulate V: Effect of measurement on the state . . . . . . . . . . . . . . . . . 9
Postulate VI: Temporal evolution . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Lecture 3 11
3.1 Intended learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Expectation values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3 Variance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4 Continuous bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.5 Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.6 Classical pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.7 Quantizing a classical system . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

4 Lecture 4 17
4.1 Intended learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
 4.2 One-dimensional quantum harmonic oscillator . . . . . . . . . . . . . . . . 17
4.3 Symbolic operator differential . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.4 Solving the ground state in the position representation . . . . . . . . . . . . 20
4.5 Uncertainty relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5 Lecture 5 23
5.1 Intended learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.2 Unitary temporal evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.3 Case of temporally dependent Hamiltonian . . . . . . . . . . . . . . . . . . . 25
5.4 Properties of unitary operators . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.5 Qubit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.6 How to set up a qubit from a physical system . . . . . . . . . . . . . . . . . 26
5.7 Pauli operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.8 Bloch sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

6 Lecture 6 29
6.1 Intended learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.2 Tunable Hamiltonian for quantum gates . . . . . . . . . . . . . . . . . . . . 29
6.3 Single-qubit gates: examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.4 Qubit measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.5 2-qubit system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.6 Examples of two-qubit gates . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.7 𝑛 -qubit system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.8 Quantum algorithms for 𝑛 qubits . . . . . . . . . . . . . . . . . . . . . . . . 32
6.9 Entanglement for two qubits . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.10 Commuting operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

References 34
 Lecture 1 1
1.1 Intended learning outcomes
I Identify how the course is technically implemented
I Identify Hilbert space and subspace of physical states
I Operate to state vectors by linear operators

1.2 Preface

Quantum mechanics is a mathematical framework to model nature.

It is said to be the most successful theory in all of physics, owing
to its success in describing a wide range of phenomena, including
atomic orbitals, quantum tunneling, and superconductivity.

The aim of this course is to formulate quantum mechanics based

on a solid mathematical foundation. We start by introducing the
basic mathematical objects required to describe physical systems.
Then, we introduce the postulates of quantum mechanics, and
discuss how to quantize a classical system. We also consider
some example systems, including qubits and a brief discussion on
quantum computing. Overall, we aim to be fairly rigorous with
the mathematical details, especially in the beginning, but certain
subtleties are left out in order to fit more useful tools into the
course. You may encounter such advanced mathematics in the
Master’s courses to fill in the gaps in your knowledge. To keep you
on track, we intent to tell you when we do not prove of handle
something rigorously.

Enjoy!
 1 Lecture 1 2

Math recap on vector spaces

1.3 Hilbert space and kets vectors
A vector space 𝑉 over a scalar
field 𝐹 is a set where addition,
The fundamental mathematical structure in quantum mechanics denoted by +, is defined such that
is the Hilbert space, which is a generalization of Euclidean space for 𝑢, 𝑣, 𝑤 ∈ 𝑉 , + : 𝑉 × 𝑉 → 𝑉 ,
that may be infinite-dimensional, and whose coordinates may be 𝑎, 𝑏 ∈ 𝐹 , the following properties
hold:
complex numbers. The elements of a Hilbert space are referred to
as ket vectors, denoted by |𝜓 i . A Hilbert space H is a complete inner 1. 𝑢 + (𝑣 + 𝑤) = (𝑢 + 𝑣) + 𝑤
2. 𝑢 +𝑣 = 𝑣 +𝑢
product space, which means that it has the following properties:
3. ∃ 0 ∈ 𝑉 s.t. 𝑉 + 0 = 𝑉
1. H = {|𝜓 i} is a vector space over the scalar field C (see math 4. ∀𝑣 ∈ 𝑉 ∃ − 𝑣 ∈ 𝑉 s.t. 𝑣 +
(−𝑣) = 0
recap on the right)
5. 𝑎(𝑏𝑣) = (𝑎𝑏)𝑣
2. For any pair of elements |𝜓 i , |𝜙i ∈ H , there is a scalar (inner) 6. 1𝑣 = 𝑣 , when 1 ∈ 𝐹
product h𝜓 |𝜙i B (|𝜓 i , |𝜙i) ∈ C that satisfies 7. 𝑎(𝑢 + 𝑣) = 𝑎𝑢 + 𝑎𝑣
8. (𝑎 + 𝑏)𝑣 = 𝑎𝑣 + 𝑏𝑣
(a) h𝜓 |𝜙i = (h𝜙 |𝜓 i) ∗ = h𝜙 |𝜓 i ∗ (conjugate symmetry)
(b) h𝜓 |𝑎𝜙 1 + 𝑏𝜙 2 i = h𝜓 | (𝑎 |𝜙 1 i + 𝑏 |𝜙 2 i) = 𝑎 h𝜓 |𝜙 1 i +𝑏 h𝜓 |𝜙 2 i For our purposes, the scalar field
𝐹 is always either R or C.
(linearity)
(c) h𝜓 |𝜓 i ≥ 0; h𝜓 |𝜓 i = 0 ⇐⇒ |𝜓 i = 0 (positive definite-
ness)
3. All Cauchy sequences converge into H . That is, if ∃{|𝜓𝑖 i}
s.t. k |𝜓𝑛 i − |𝜓𝑚 i k → 0 for 𝑛, 𝑚 → ∞ then ∃ |Ψi ∈ H s.t.
|𝜓𝑚 i → |Ψi for 𝑚 → ∞.1 1: Not going to ask about Cauchy se-
p quences in the exam, but this condi-
Consequently, we can define a norm k |𝜓 i k = k𝜓 k B h𝜓 |𝜓 i ≥ 0. tion is what makes a Hilbert space
For example, the following inequalities apply: complete. Thus, if you are unfamiliar
with the term Cauchy sequence, no
I | h𝜓 |𝜙i | ≤ k𝜓 k k𝜙 k , i.e., Cauchy–Schwarz inequality need to study it for this course.
I k |𝜓 i + |𝜙i k ≤ k𝜓 k + k𝜙 k , i.e., triangle inequality

Physical states, i.e., objects that can be used to model the states
of physical systems on the quantum-mechanical level, are those
elements of H which have a norm of unity, i.e., h𝜓 |𝜓 i = 1. In this
course, we use the terms ket vector and state somewhat interchange-
ably, since we are concerned mostly with physical states. But many
of the results also hold for ket vectors with any finite norm.

1.4 Bra vectors

As discussed above, the ket vectors are the elements of H , i.e.,

H = {|𝜓 i}. For each given ket vector |𝜙i we symbolically define
an object h𝜙 | , through the inner product such that for all |𝜓 i ∈ H ,
we have h𝜙 |𝜓 i := (|𝜙i , |𝜓 i) ∈ C. Below, we justify why h𝜙 | can be
referred to as a bra vector, i.e., the set of all bra vectors {h𝜙 |} forms
2: A functional 𝑓 acting on a vector
a vector space. space 𝑉 is a mapping 𝑓 : 𝑉 → C.
𝑓 is said to be bounded, if ∀𝑣 ∈ 𝑉 ,
For a fixed |𝜙i , the inner product (|𝜙i , |𝜓 i) can be identified as there exists a constant 𝑀 such that
a mapping that takes any ket vector |𝜓 i to a complex number. |𝑓 (𝑣)| ≤ 𝑀 k𝑣 k .
Thus h𝜙 | is a linear and bounded functional2 acting on H . The 3: Not going to ask about this in the
so-called Riesz representation theorem3 states that for any linear exam.
 1 Lecture 1 3

bounded functional h𝜙 | , there exists a corresponding unique ket

vector |𝜙i ∈ H such that h𝜙 |𝜓 i = (|𝜙i , |𝜓 i) for all |𝜓 i ∈ H . This is
the definition of the bra vector h𝜙 | ; it is the functional whose action
on any |𝜓 i corresponds to taking the inner product with the ket
|𝜙i .
From the above-noted uniqueness of the ket vector |𝜙i correspond-
ing to its bra h𝜙 | , it folows that there is a one-to-one correspondence
between H = {|𝜓 i} and the set of all bra vectors H ∗ B {h𝜙 |}. The
set H ∗ is referred to as the dual space of H .

1.5 Linear operators

Definition 1.5.1 A mapping 𝐴ˆ is a linear operator on H ⇐⇒

𝐴ˆ : H → H s.t. ∀ |𝜓 i , |𝜙i ∈ H , 𝑎, 𝑏 ∈ C:

𝐴ˆ (𝑎 |𝜓 i + 𝑏 |𝜙i) = 𝑎𝐴ˆ |𝜓 i + 𝑏 𝐴ˆ |𝜙i . (1.1)

We denote the set of linear operators on H by L (H ) . We also

define the notation

ˆ i B 𝐴ˆ |𝜓 i B 𝐴ˆ (|𝜓 i) .
|𝐴𝜓 (1.2)

It follows that ∀𝐴,

ˆ 𝐵ˆ ∈ L (H ) , we have



𝐴ˆ 𝐵ˆ |𝜓 i = 𝐴ˆ𝐵ˆ |𝜓 i . (1.3)

1.6 Outer product

An important example of a linear operator is the outer product of

two vectors.

Definition 1.6.1 We define the outer product |𝜓 ih𝜙 | : H → H ,

where |𝜓 i , |𝜙i ∈ H s.t. ∀ |𝜒i ∈ H we have:

(|𝜓 ih𝜙 |) | 𝜒i = |𝜓 ih𝜙 | 𝜒i = h𝜙 |𝜒i|𝜓 i (1.4)

Note that in the second equality above, we moved the term h𝜙 | 𝜒i

to the front, since it is simply a scalar.
 Lecture 2 2
2.1 Intended learning outcomes
I Use bases to represent operators
I Identify the minimal mathematical structure to describe a
physical system quantum mechanically

2.2 Bases of H

Above, we have discussed bra and ket vectors in a very abstract

way, without a way to visualize these vectors. To make them more
tangible, we will introduce them coordinates using a basis.

Definition 2.2.1 A set of ket vectors {|𝜙𝑖 i}𝑖=

𝑁 ∈ H, 𝑁 ∈ Z , is
1 +
Í𝑁
referred to as linearly independent if 𝑖=1 𝑐𝑖 |𝜙𝑖 i = 0 implies 𝑐𝑖 =
0 ∀𝑐𝑖 ∈ C.

The dimension of H , Dim{H }, is the largest 𝑁 for which such a

linearly independent set of vectors exists.

The set {|𝜙𝑖 i}𝑖=

𝑁 is referred to as complete if ∀ |𝜓 i ∈ H , ∃ {𝑐 } 𝑁 , 𝑐 ∈
𝑘 𝑘=1 𝑘
Í𝑁1
C s.t. |𝜓 i = 𝑘=1 𝑐𝑘 |𝜙𝑘 i .1 That is, any ket vector in H may be ex- 1: Defined similarly for infinite-
pressed as a linear combination of the vectors |𝜙𝑘 i . The coefficients dimensional spaces.

𝑐𝑘 are the coordinates of |𝜓 i . We have thus arrived at the definition

of a basis:

Definition 2.2.2 A complete set of linearly independent vectors {|𝜙𝑘 i}

is referred to as a basis for H .

A basis {|𝜙𝑘 i} is referred to as orthonormal if


0, for 𝑙 ≠ 𝑚,
h𝜙𝑙 |𝜙𝑚 i = 𝛿𝑙𝑚 = (2.1)
1, for 𝑙 = 𝑚.

The symbol 𝛿𝑙𝑚 is referred to as the Kronecker delta.

An observation for the orthonormal basis {|𝜙𝑘 i}: for an arbitrary

|𝜓 i ∈ H , we have
Õ
|𝜓 i = 𝑐𝑘 |𝜙𝑘 i (2.2)
𝑘
Õ
=⇒ h𝜙𝑚 |𝜓 i = 𝑐𝑘 h𝜙𝑚 |𝜙𝑘 i = 𝑐𝑚 . (2.3)
𝑘
 2 Lecture 2 5

Thus,
Õ
|𝜓 i = 𝑐𝑚 |𝜙𝑚 i
𝑚
Õ
= h𝜙𝑚 |𝜓 i|𝜙𝑚 i
𝑚
Õ
= |𝜙𝑚 ih𝜙𝑚 |𝜓 i
𝑚
!
Õ
= |𝜙𝑚 ih𝜙𝑚 | |𝜓 i ,
𝑚

where in the last step, we have used the fact that the outer product
Í
is linear. Based on this, we conclude that 𝑚 |𝜙𝑚 ih𝜙𝑚 | = 𝐼ˆ, the
identity operator. This holds for any orthonormal basis. It is a
useful trick to insert the identity operator in strategic places, and
expand it in terms of an orthonormal basis like this.

2.3 States vs. vectors

For a given basis {|𝜙𝑘 i} and a ket vector |𝜓 i ∈ H , we may write

Õ
|𝜓 i = 𝑐𝑘 |𝜙𝑘 i (2.4)
𝑘
𝑐 1 
 
𝑐 2 
=ˆ 𝑐  ,
 
 3
.
 .. 
 

where =ˆ stands for represented by. Consequently, a basis ket vector

|𝜙𝑚 i is represented by a column vector where 𝑐𝑚 = 1 and 𝑐𝑘 = 0
for 𝑘 ≠ 𝑚 . Note that the vector representation of a state may be
infinite-dimensional.

Given a column vector representation of |𝜓 i with the coefficients

{𝑐𝑘 }, the corresponding bra vector h𝜓 | may be represented by the
conjugate transpose of the column vector representing |𝜓 i :

h𝜓 | =ˆ 𝑐 1∗ 𝑐 2∗ 𝑐 3∗ . . . .
 
(2.5)

This can be shown using the inner product.

2.4 Operators vs. matrices

Analogously to representing kets as column vectors, it is possible

to represent operators as matrices. Let 𝐴ˆ ∈ L (H ) and {|𝜙𝑚 i} be
 2 Lecture 2 6

an orthonormal basis of H . Then,

! !
Õ Õ
𝐴ˆ = 𝐼ˆ𝐴ˆ𝐼ˆ = |𝜙𝑚 ih𝜙𝑚 | 𝐴ˆ |𝜙𝑘 ih𝜙𝑘 | (2.6)
𝑚 𝑘
Õ
= |𝜙𝑚 ih𝜙𝑚 |𝐴|𝜙
ˆ 𝑘 i h𝜙𝑘 |
𝑚,𝑘
| {z }
B𝐴𝑚𝑘 ∈C
Õ
= 𝐴𝑚𝑘 |𝜙𝑚 ih𝜙𝑘 |
𝑚,𝑘
𝐴11 𝐴12 . . .
 
=ˆ 𝐴21 𝐴22 . . . .
 
 . .. . . 
 .. . .


This is the matrix representation of the operator 𝐴ˆ. Note that the
matrix might be infinite-dimensional.

Using the matrix representation, the operation of 𝐴ˆ on a ket vector

|𝜓 i may be written explicitly:
Õ
𝐴ˆ |𝜓 i = 𝐴𝑚𝑘 |𝜙𝑚 ih𝜙𝑘 |𝜙𝑙 i (2.7)
𝑚,𝑘,𝑙
Õ
= 𝐴𝑚𝑘 𝑐𝑘 |𝜙𝑚 i .
𝑚,𝑘
Í
We observe that the expression 𝑚,𝑘 𝐴𝑚𝑘 𝑐𝑘 corresponds to matrix-
vector multiplication, and conclude that

𝐴11 𝐴12 . . . 𝑐 1 
  
𝐴ˆ |𝜓 i =ˆ 𝐴21 𝐴22 . . . 𝑐 2  .
  
 . .. . .   .. 
 .. . .  . 

In other words, we obtain the column vector representation of
|𝜓 0i = 𝐴ˆ |𝜓 i by calculating the matrix-vector product between the
matrix representation of 𝐴ˆ and the column vector representation Math on complex conjugation

of |𝜓 i . ∀𝑧 ∈ C we have 𝑥, 𝑦 ∈ R and i is
the imaginary unit, then:

𝑧 = 𝑥 + i𝑦
𝑧 ∗ = 𝑥 − i𝑦
2.5 Adjugate

Let 𝐴ˆ ∈ L (H ) , and furthermore, let 𝐴ˆ be bounded.2 We define 2: An operator 𝐴ˆ is said to be

the action of 𝐴ˆ on a left-lying bra vector (i.e. an element in the dual bounded if ∀ |𝜓 i ∈ H there exists
a constant 𝑀 such that k𝐴ˆ |𝜓 i k ≤
space H ∗ ), 𝐴ˆ : H ∗ → H ∗ , ∀ |𝜙i , |𝜓 i ∈ H through the relation
𝑀 k |𝜓 i k .



h𝜙 | 𝐴ˆ |𝜓 i = h𝜙 | 𝐴ˆ |𝜓 i . (2.8)
About notation
We observe that the operation h𝜙 |𝐴ˆ is a linear functional on H and These are equivalent:
is bounded since 𝐴ˆ is bounded.

h𝜙 |𝐴|𝜓
ˆ i = |𝜙i , 𝐴ˆ |𝜓 i )
 
= 𝐴ˆ† |𝜙i , |𝜓 i

= h𝐴ˆ†𝜙 |𝜓 i
 2 Lecture 2 7

Thus it follows from the Riesz representation theorem that ∃ |𝜙 0i ∈

H s.t. h𝜙 0 | = h𝜙 |𝐴ˆ. This also defines a linear operator 𝐴ˆ† as

𝐴ˆ† |𝜙i = |𝜙 0i , (2.9)

which is referred to as the adjugate of 𝐴ˆ.

For example, we have for 𝑐 ∈ C

h𝜙 |𝑐 |𝜓 i = (|𝜙i , 𝑐 |𝜓 i) (2.10)
= 𝑐 (|𝜙i , |𝜓 i)
= (𝑐 ∗ |𝜙i , |𝜓 i) ,

Thus, 𝑐 † = 𝑐 ∗ ∈ C, i.e., the adjugate of a complex number is just the

complex conjugate.

2.6 Properties of adjugate

The following equalities are not proven here, but proofs can be
constructed based on the above definitions.
 †
𝐴ˆ† = 𝐴ˆ (2.11)
 †
𝑎𝐴ˆ† = 𝑎 ∗𝐴ˆ (𝑎 ∈ C) (2.12)

†
𝐴ˆ𝐵ˆ = 𝐵ˆ †𝐴ˆ† (2.13)
(|𝜓 ih𝜙 |) † = |𝜙ih𝜓 | (2.14)

Furthermore, given a matrix representation 𝐴 of 𝐴ˆ, the matrix

representation of 𝐴ˆ† is given by the conjugate transpose of 𝐴,
which is obtained by taking the transpose and then the complex
conjugate of each element of 𝐴.

2.7 Eigenvalues and eigenstates

For an operator 𝐴ˆ ∈ L (H ) , if a ket vector |𝜓𝑘 i ∈ H satisfies the

eigenvalue equation

𝐴ˆ |𝜓𝑘 i = 𝜆𝑘 |𝜓𝑘 i (2.15)

for some scalar 𝜆𝑘 ∈ C, we define |𝜓𝑘 i to be an eigenvector, or

eigenstate, of 𝐴ˆ with an eigenvalue 𝜆𝑘 . The subscript 𝑘 signifies
that there may be (infinitely) many eigenstates and corresponding
eigenvalues.

The set of eigenvalues {𝜆𝑘 } is referred to as the spectrum of 𝐴ˆ.

 2 Lecture 2 8

It is possible that for a given eigenvalue 𝜆𝑘 , there are multiple

eigenvectors |𝜓𝑘,𝑖 i that satisfy Eq. (2.15), with 𝑖 = 1, . . . , 𝑔𝑘 . The
number of eigenvectors 𝑔𝑘 corresponding to 𝜆𝑘 is referred to as the
degeneracy of 𝜆𝑘 .

2.8 Hermitian operators

Definition 2.8.1 The operator 𝐻ˆ ∈ L (H ) is defined to be Hermitian

3: If and only if
iff3 𝐻ˆ † = 𝐻ˆ .

Hermitian operators are very important in quantum mechanics as

discovered below.

The key feature of Hermitian operators comes from the so-called

generalized spectral theorem, which states that for a Hermitian
operator 𝐻ˆ ∈ L (H ) , there exists a complete orthonormal basis of
H , {|𝜓𝑘 i}, which satisfies

𝐻ˆ |𝜓𝑘 i = 𝜆𝑘 |𝜓𝑘 i . (2.16)

Importantly, it also follows from the spectral theorem that the

eigenvalues 𝜆𝑘 are real numbers.

The above result implies that for any Hermitian operator 𝐻ˆ , it is

always possible to find a basis such that
Õ
𝐻ˆ = 𝜆𝑘 |𝜓𝑘 ih𝜓𝑘 | . (2.17)
𝑘

In the matrix representation, this is a matrix with just the eigenval-

ues 𝜆𝑘 on the diagonal. This is useful for many reasons. For one, it
is very easy to operate on any vector with such an operator. Fur-
thermore, it turns out that many problems in quantum mechanics
boil down to finding the eigenvalues of Hermitian operators. The
process of finding such a basis in which the eigenvalues are on the
diagonal is referred to as diagonalization, and much of the effort
in theoretical physics is spent on trying to diagonalize operators
related to different physical systems.

2.9 Postulates of quantum mechanics

We finally have introduced all the necessary mathematics to start

discussing physical systems. The theory of quantum mechanics is
in essence built upon the six postulates discussed below. They are the
fundamental assumptions, or axioms, of quantum mechanics, i.e.,
they are not proven, but rather they are based on empirical evidence.
Predictions derived from the postulates have been experimentally
 2 Lecture 2 9

verified to extremely high precision. In this course, we consider

quantum mechanics simply as a model for such experimental
observations.

Postulate I
4: We discuss this later
For each physical system there exists a corresponding (rigged)4
Hilbert space.

Postulate II

Each physical state of this system can be represented by a quantum

state |𝜓 i ∈ H , where h𝜓 |𝜓 i = 1.

Postulate III

For each measurable quantity 𝐴 of the system we have a corre-

sponding operator 𝐴ˆ ∈ L (H ) s.t. 𝐴ˆ† = 𝐴ˆ. Such an operator (and
often also the corresponding quantity) is referred to as an observable
of the system.

In an ideal measurement of the quantity 𝐴, any measurement

outcome equals to an eigenvalue of 𝐴ˆ.5 5: Recall that since 𝐴ˆ is Hermitian,
this implies that all measurement out-
comes are real numbers, as one would
Postulate IV: Measurement expect.

Let |𝜓 i ∈ H and 𝐴ˆ† = 𝐴ˆ ∈ L (H ) with a discrete spectrum {𝑎𝑛 }.

As discussed in Sec. 2.8, there always exists an orthonormal basis
for H , {|𝜙𝑛,𝑖 i}𝑛,𝑖=∈ {1,...,𝑔𝑛 } , where 𝑔𝑛 is the amount of degeneracy,
such that the basis vectors are eigenstates of 𝐴ˆ.

The probability of obtaining a specific measurement result 𝑎𝑛 is

given by
𝑔𝑛
Õ 2
𝑃 (𝑎𝑛 ) B h𝜙𝑛,𝑖 |𝜓 i . (2.18)
𝑖=1

Postulate V: Effect of measurement on the state

Suppose that a system is in the state |𝜓 i ∈ H . If we measure the

quantity corresponding to 𝐴ˆ and obtain the measurement result
 2 Lecture 2 10

𝑎𝑛 (an eigenvalue of 𝐴ˆ), the state of the system collapses into the
state
𝑔𝑛
𝑃ˆ𝑛 |𝜓 i Õ
|𝜓 0i = , 𝑃ˆ𝑛 = |𝜙𝑛,𝑖 ih𝜙𝑛,𝑖 | . (2.19)
k 𝑃ˆ𝑛 |𝜓 i k 𝑖=1

𝑃ˆ𝑛 is referred to as a projector onto the subspace corresponding to

𝑔
the subspace spanned by the eigenstates {|𝜙𝑛,𝑖 i}𝑖=𝑛1 .

Definition 2.9.1 𝑃ˆ𝑛 ∈ L (H ) is a projector iff 𝑃ˆ𝑛2 = 𝑃ˆ𝑛 .

Postulate VI: Temporal evolution

If a system is in the state |𝜓 i , the temporal evolution of the state

|𝜓 (𝑡)i is determined by the Schrödinger equation:

iℏ𝜕𝑡 |𝜓 (𝑡)i = 𝐻ˆ |𝜓 (𝑡)i , (2.20)

where 𝜕𝑡 B 𝜕𝑡 𝜕
, ℏ = 1.0545718 × 10−34 Js is the reduced Planck
constant, and 𝐻ˆ = 𝐻ˆ † , 𝐻ˆ ∈ L (H ) is the Hamiltonian, the observable
6: The measurable quantity corre-
corresponding to the total energy of the system.6 sponding to the observable 𝐻ˆ is the
Hamiltonian 𝐻 from classical Hamil-
Note that 𝐻ˆ may also depend on time through temporally de-
tonian mechanics.
pendent parameters {𝛼𝑖 (𝑡)}, i.e., 𝐻ˆ = 𝐻ˆ [𝛼 1 (𝑡), 𝛼 2 (𝑡), . . . ] . This is
discussed later.
 Lecture 3 3
3.1 Intended learning outcomes
I Differentiate between a measurement outcome and its expec-
tation value
I Identify continuous bases for Hilbert spaces
I Apply Lagrangian formalism to quantize physical systems

3.2 Expectation values

Definition 3.2.1 Let 𝐴ˆ ∈ L (H ) and |𝜓 i ∈ H . The expectation value

of 𝐴ˆ when the system is in the state |𝜓 i is defined by

h𝐴i B h𝜓 |𝐴|𝜓
ˆ i. (3.1)

In particular, if we have an observable quantity 𝐴 with a corre-

sponding Hermitian operator 𝐴ˆ, the expectation value is equal
to the classical expectation value h𝐴i of 𝐴. That is, repeatedly
preparing the system in the state |𝜓 i and measuring 𝐴, one obtains
on average the result h𝐴i , even though individual measurements
only yield discrete values 𝑎𝑘 , the eigenvalues of 𝐴ˆ.

Mathematically, the above discussion maybe considered as follows:

Recall that 𝐴ˆ = 𝐴ˆ† implies that there exists an orthonormal basis
{|𝜙𝑘 i} such that 𝐴ˆ |𝜙𝑘 i = 𝑎𝑘 |𝜙𝑘 i , 𝑎𝑘 ∈ R. Subsequently, we write
Í
the state as |𝜓 i = 𝑘 𝑐𝑘 |𝜙𝑘 i , from which we obtain
!†
Õ Õ
h𝜓 |𝐴|𝜓
ˆ i= 𝑐𝑘 |𝜙𝑘 i 𝐴ˆ 𝑐𝑘 |𝜙𝑘 i (3.2)
𝑘 𝑘
!†
Õ Õ
= 𝑐𝑘 |𝜙𝑘 i 𝑐𝑘 𝑎𝑘 |𝜙𝑘 i
𝑘 𝑘
Õ
= 𝑐𝑛∗ 𝑎𝑘 𝑐𝑘 h𝜙𝑛 |𝜙𝑘 i
𝑛,𝑘
Õ Õ
= 𝑎𝑘 |𝑐𝑘 | 2 = 𝑎𝑘 𝑃 (𝑎𝑘 ).
𝑘
|{z} 𝑘
𝑃 (𝑎𝑘 )

The sum on the right side of the last equality above is the clas-
sical definition of the expectation value for the measurement
outcomes.
 3 Lecture 3 12

3.3 Variance

Definition 3.3.1 We define the variance of 𝐴ˆ when the system is in

the state |𝜓 i as

Δ𝐴2 = h𝜓 | 𝐴ˆ − h𝜓 |𝐴|𝜓

2
ˆ i |𝜓 i (3.3)
ˆ i 2


= h𝜓 |𝐴ˆ2 |𝜓 i − h𝜓 |𝐴|𝜓
!2
Õ Õ
= 𝑎𝑘2 𝑃𝑘 − 𝑎 𝑘 𝑃𝑘 .
𝑘 𝑘

As above in the case of the expectation value, if 𝐴ˆ is a Hermitian

operator corresponding to the observable 𝐴, the above definition
coincides with that of the classical variance of 𝐴.

3.4 Continuous bases

Continuous bases are sometimes required, for example, if continu-

ous variables are used.

Let {|𝜓𝛼 i} ∈ H , where 𝛼, 𝛼 0 ∈ R s.t.

Math on Dirac delta function
h𝜓𝛼 |𝜓𝛼 0 i = 𝛿 (𝛼 − 𝛼 0), (3.4)
For a smooth function, i.e., 𝑓 ∈
where 𝛿 (𝑥) is the Dirac delta function. Such a set of vectors is a 𝐶∞:
∫
continuous base for H . d𝑥 {𝛿 (𝑥)𝑓 (𝑥)} = 𝑓 ( 0)

Note that h𝜓𝛼 |𝜓𝛼 i = 𝛿 ( 0) = ∞. Thus |𝜓𝛼 i is not possible to normal-

ize. This is why we consider rigged Hilbert spaces, which allow
such states.1 1: It is possible to verify that all the
properties of a Hilbert space hold
Now, similarly as for discrete bases, we may write any |𝜓 i ∈ H even with such non-normalizable
states, but it is fairly laborious and
using this basis, but with an integral instead of a sum:
therefore we do not discuss it further.
∫
|𝜓 i = 𝑐 𝛼 |𝜓𝛼 i d𝛼 (3.5)
∫
= h𝜓𝛼 |𝜓 i|𝜓𝛼 i d𝛼
∫ ∫ 
= |𝜓𝛼 ih𝜓𝛼 |𝜓 i d𝛼 = |𝜓𝛼 ih𝜓𝛼 | d𝛼 |𝜓 i
| {z }
𝐼ˆ

Note that the index 𝛼 ∈ R of the coefficients 𝑐 𝛼 ∈ C is continuous.

Often, instead of 𝑐 𝛼 we write h𝜓𝛼 |𝜓 i B 𝜓 (𝛼) , where 𝜓 (𝛼) : R → C
is generally referred to as the wave function.

Typically, the wave function is expressed in the position basis {|𝑥i},

i.e., 𝜓 (𝑥) := h𝑥 |𝜓 i . From the above equation it follows that the
probability density for the particle to reside at position 𝑥 is given
 3 Lecture 3 13

by |𝜓 (𝑥)| 2 . Note that the wave function cannot fully describe all
quantum systems, just those where such continuous variables exist
and are sufficient.

Definition 3.4.1 In a continuous basis, the measurement probability

of the measurement outcome to reside in [𝛼, 𝛼 + d𝛼] is defined by

d𝑃 (𝛼) = |h𝜓𝛼 |𝜓 i| 2 d𝛼 . (3.6)

3.5 Commutators
ˆ 𝐵ˆ ∈ L (H ) is given by
Definition 3.5.1 The commutator of 𝐴,

[𝐴, ˆ = 𝐴ˆ𝐵ˆ − 𝐵ˆ 𝐴.
ˆ 𝐵] ˆ

If two operators satisfy [𝐴, ˆ = 0, i.e., 𝐴ˆ𝐵ˆ = 𝐵ˆ 𝐴ˆ, it is defined that 𝐴ˆ

ˆ 𝐵]
and 𝐵ˆ commute.

The commutator is an important operation between two operators

and appears in numerous places in quantum mechanics.

3.6 Classical pendulum

Above, we introduced how to connect mathematics to physics

through the postulates. The Hamiltonian of the system is the key
here. Once one has it, also the relevant Hilbert space arises from
its eigenstates in addition to the temporal evolution of any state.
However, how do we obtain the Hamiltonian for the system?

To answer this question, we take a slight detour to classical mechan-

ics. As an illustrative example, we discuss a classical pendulum,
and construct the classical Hamiltonian for it. Subsequently, in the
next section, we provide a general procedure, or recipe, for convert-
ing the classical Hamiltonian of any system to the corresponding
quantum Hamiltonian. In the next lecture, we use the obtained
Hamiltonian for the pendulum to describe the quantum harmonic
oscillator.

Recall from classical mechanics that a system with 𝑁 degrees of

freedom can be described by a set of 𝑁 generalized coordinates
{𝑞𝑖 }𝑖=
𝑁 . The coordinates may for example be just the position of
1
Figure 3.1: Ideal classical pendulum,
a particle, but often the description of the system is drastically where a mass 𝑚 is attached to a mass-
simplified if one chooses the generalized coordinates wisely. less rigid rod of length 𝑙 . The rod may
rotate without friction about a single
We consider the pendulum shown in Fig. 3.1. Even though the mass axis as described by the angle 𝜃 . We
at the end of the pendulum moves in a 2D plane, we recognize assume a uniform gravitational field
described by 𝑔.
that there is only one degree of freedom in the system; the position
 3 Lecture 3 14

is fully determined by the angle 𝜃 . We thus choose the generalized

position

𝑞 = 𝑙𝜃 . (3.7)

We drop the subscript 𝑖 because we have only one degree of freedom,

but all the definitions below apply in general for multidimensional
systems as well.

The potential energy 𝑉 depends only on 𝑞 , and not on the time

derivative 𝑞¤. For small 𝜃 , it assumes the form

𝑉 = 𝑔𝑚ℎ (3.8) Math on dot notation

1 𝑚𝑔 2
= 𝑚𝑔𝑙 ( 1 − cos 𝜃 ) ≈ 𝑚𝑔𝑙𝜃 2 = 𝑞 .
2 2𝑙 d𝑦 𝜕𝑦
𝑦¤ = ≠
d𝑡 𝜕𝑡
It is straightfowrard to write the kinetic energy 𝑇 in terms of 𝑞¤: |{z}
generally

1 1 1
𝑇 = 𝑚𝑣 2 = 𝑚𝑙 2𝜃¤2 = 𝑚𝑞¤2 . (3.9)
2 2 2 Math on Taylor series

𝑥2 𝑥4
Definition 3.6.1 The Lagrangian is defined as cos 𝑥 = 1 − + −···
2 24

𝐿 B 𝑇 −𝑉. (3.10)

Thus, the Lagrangian for our case is

1 1
𝐿 = 𝑇 − 𝑉 = 𝑚𝑙 2𝜃¤2 − 𝑚𝑔𝑙𝜃 2 (3.11)
2 2
1 2 𝑚𝑔 2
= 𝑚𝑞¤ − 𝑞 .
2 2𝑙

Definition 3.6.2 The generalized momentum corresponding to the

coordinate 𝑞𝑖 is defined as

𝜕𝐿
𝑝𝑖 B .
𝜕𝑞¤𝑖

Note that when computing the momentum from the Lagrangian,

𝑞𝑖 and 𝑞¤𝑖 should be considered independent variables. Using the
above definition, the generalized momentum is (again, dropping
the subscript)
 
𝜕 1 2 𝑚𝑔 2
𝑝= 𝑚𝑞¤ − ¤
𝑞 = 𝑚𝑞. (3.12)
𝜕𝑞¤ 2 2𝑙
 3 Lecture 3 15

Definition 3.6.3 The classical Hamiltonian is defined as

Õ
𝐻 B 𝑞¤𝑖 𝑝𝑖 − 𝐿.
𝑖

Using this, we obtain the Hamiltonian of the pendulum (i.e. the

1D harmonic oscillator):
 
2 1 2 𝑚𝑔 2
¤ − 𝐿 = 𝑚𝑞¤ − 𝑚𝑞¤ −
𝐻 = 𝑞𝑝 𝑞 (3.13)
2 2𝑙
𝑝 2 𝑚𝑔 2
= + 𝑞 =𝑇 +𝑉.
2𝑚 2𝑙

3.7 Quantizing a classical system

The term quatization refers to the process of building a quantum-

mechanical model from the classical description of the system in
question. In general, given the classical Hamiltonian of a system, it
can be quantized using the following procedure:

1. Operator substitution: Replace all generalized positions and

momenta with corresponding Hermitian operators, simply
by writing hats on the classical quantities:

𝑞𝑖 −→ 𝑞ˆ𝑖 , 𝑞ˆ𝑖 : H → H, 𝑞ˆ𝑖 = 𝑞ˆ𝑖†,

𝑝𝑖 −→ 𝑝ˆ𝑖 , 𝑝ˆ𝑖 : H → H, 𝑝ˆ𝑖 = 𝑝ˆ𝑖† .

2. Quantized Hamiltonian: Using step 1, convert the classical

Hamiltonian 𝐻 to the operator 𝐻ˆ , i.e., replace all classical
generalized positions and momenta in 𝐻 by their quantum
mechanical counterparts.
3. Canonical commutation relation: The positions and mo-
menta must satisfy [𝑝ˆ𝑖 , 𝑞ˆ𝑖 ] = 𝑝ˆ𝑖 𝑞ˆ𝑖 − 𝑞ˆ𝑖 𝑝ˆ𝑖 = −iℏ, which is
referred to as the canonical commutation relation (CCR).
4. Temporal evolution: With the above operators and the con-
straint imposed by the CCR, the temporal evolution of the
system is given by the Schrödinger equation iℏ𝜕𝑡 |𝜓 i = 𝐻ˆ |𝜓 i .

For the pendulum discussed above, the quantization procedure

simply yields

𝑝ˆ 2 𝑚𝑔 2
𝐻ˆ = + 𝑞ˆ . (3.14)
2𝑚 2𝑙
The significance of the CCR and the temporal evolution for the
harmonic oscillator will be discussed on the following lectures.
 3 Lecture 3 16

The above procedure maybe used for many different systems. For
example, it is possible to quantize electric circuits by choosing
charge as the generalized position and magnetic flux as the mo-
mentum, or vice versa.2 Another important application is the 2: This is discussed in the Quantum
quantization of the electromagnetic field, which follows a similar Circuits course.

procedure but with a continuous set of generalized coordinates.

 Lecture 4 4
4.1 Intended learning outcomes
I Apply creation and annihilation operators for a harmonic
oscillator
I Apply canonical commutation relations
I Identify Heisenberg’s uncertainty relation

4.2 One-dimensional quantum harmonic

oscillator

As we derived above in Eq. (3.14), the operator corresponding to

the classical Hamiltonian of the harmonic oscillator is given by

𝑝ˆ 2 𝑚 𝑔
𝐻ˆ = + 𝑞ˆ2 (4.1)
2𝑚 2 𝑙
|{z}
≕𝜔 2
𝑝ˆ 2 1
= + 𝑚𝜔 2𝑞ˆ2, (4.2)
2𝑚 2

where [𝑞, ˆ = iℏ, 𝑞ˆ = 𝑞ˆ† , 𝑝ˆ = 𝑝ˆ † .

ˆ 𝑝]
Math on C
Next, we wish to solve the eigenstates of the oscillator. To this
end, we try to rewrite the Hamiltonian in the following form, with For 𝑥, 𝑦 ∈ R,

𝐴, 𝐵, 𝐶 ∈ R: (𝑥 + 𝑦)(𝑥 − 𝑦) = 𝑥 2 − 𝑦 2
(𝑥 + i𝑦) (𝑥 − i𝑦) = 𝑥 2 + 𝑦 2 = |𝑧| 2
𝐻ˆ = (𝐴𝑞ˆ − i𝐵𝑝)
ˆ (𝐴𝑞ˆ + i𝐵𝑝)
ˆ + 𝐶. (4.3) | {z } | {z }
≕𝑧 =𝑧 ∗

With some algebraic manipulation and the help of the CCR, we

find

𝐻ˆ = 𝐴2𝑞ˆ2 + i𝐴𝐵𝑞ˆ𝑝ˆ − i𝐵𝑝𝐴

ˆ 𝑞ˆ + 𝐵 2𝑝ˆ 2 + 𝐶 (4.4)
2 2 2 2
= 𝐴 𝑞ˆ + 𝐵 𝑝ˆ + i𝐴𝐵 [𝑞, ˆ +𝐶.
ˆ 𝑝]
|{z}
=iℏ
| {z }
=−ℏ𝐵𝐴
 4 Lecture 4 18

Comparing this to Eq. (4.2), we choose

r
1
𝐴= 𝑚𝜔 2, (4.5)
2
r
1
𝐵= , (4.6)
2𝑚
ℏ𝜔
𝐶 = ℏ𝐴𝐵 = . (4.7)
2
With these, we may write

𝑝ˆ 2 1
𝐻ˆ = + 𝑚𝜔 2𝑞ˆ2 (4.8)
2𝑚 2
r r !† r r !
𝑚𝜔 2 1 𝑚𝜔 2 1 1
= 𝑞ˆ + i𝑝ˆ 𝑞ˆ + i𝑝ˆ + ℏ𝜔
2 2𝑚 2 2𝑚 2
"r  † r   #
𝑚𝜔 i 𝑚𝜔 i 1
= ℏ𝜔 𝑞ˆ + 𝑝ˆ 𝑞ˆ + 𝑝ˆ +
2ℏ 𝑚𝜔 2ℏ 𝑚𝜔 2
| {z }| {z }
=𝑎ˆ † ≕𝑎ˆ
 
† 1
= ℏ𝜔 𝑎ˆ 𝑎ˆ + .
2

Definition 4.2.1 For the one-dimensional quantum harmonic oscilla-

tor, we define r  
𝑚𝜔 i
𝑎ˆ B 𝑞ˆ + ˆ
𝑝 ,
2ℏ 𝑚𝜔
from which is follows that
r  
† 𝑚𝜔 i
𝑎ˆ = 𝑞ˆ − 𝑝ˆ ,
2ℏ 𝑚𝜔

and
 
† 1
𝐻ˆ = ℏ𝜔 𝑎ˆ 𝑎ˆ + .
2
.

The operator 𝑎ˆ is referred to as the lowering or annihilation operator

and 𝑎ˆ† is referred to as the raising or creation operator. Sometimes, 𝑎ˆ
and 𝑎ˆ† together are referred to as ladder operators.

Note that 𝑎ˆ ≠ 𝑎ˆ† , i.e. 𝑎ˆ is not Hermitian, which means that it does
not correspond to an observable. However, the product 𝑎ˆ†𝑎ˆ is
Hermitian. Thus it is enough to find its eigenvalues and eigen-
states to solve the quantum-mechanical problem of the harmonic
oscillator.
 4 Lecture 4 19

Let us calculate the commutator of 𝑎ˆ and 𝑎ˆ† as

 
† 𝑚𝜔 i i
[𝑎,
ˆ 𝑎ˆ ] = 𝑞ˆ + ˆ 𝑞ˆ −
𝑝, 𝑝ˆ (4.9)
2ℏ 𝑚𝜔 𝑚𝜔
   
𝑚𝜔 i i
= ˆ−
𝑞, 𝑝ˆ + ˆ 𝑞ˆ
𝑝,
2ℏ 𝑚𝜔 𝑚𝜔
i

= − [𝑞, ˆ + [𝑝,
ˆ 𝑝] ˆ = 1.
ˆ 𝑞]
2ℏ |{z} |{z}
=iℏ =−iℏ

Some observations about the quantum harmonic oscillator:

1. h𝜓 |𝐻ˆ |𝜓 i ≥ 0 ∀ |𝜓 i , since
 
1
h𝐻ˆ i = h𝜓 |ℏ𝜔 𝑎ˆ†𝑎ˆ + |𝜓 i (4.10)
2
ℏ𝜔
= + h𝜓 |ℏ𝜔 𝑎ˆ†𝑎|𝜓
ˆ i
2  
1 2
= ℏ𝜔 + k𝑎ˆ |𝜓 i k ≥ 0.
2

Thus all eigenenergies are positive.

2. Let |𝜓 i be an eigenstate of 𝐻ˆ s.t. 𝐻ˆ |𝜓 i = 𝜀 |𝜓 i . Then,
 
† 1
𝐻ˆ 𝑎ˆ |𝜓 i = ℏ𝜔 𝑎ˆ 𝑎ˆ + 𝑎ˆ |𝜓 i (4.11)
|{z} 2
=𝑎ˆ𝑎ˆ † −1
 
† 1
= 𝑎ℏ𝜔
ˆ 𝑎ˆ 𝑎ˆ + − 1 |𝜓 i
2


= 𝑎ˆ 𝐻 − ℏ𝜔 |𝜓 i
ˆ
= 𝑎ˆ (𝜀 − ℏ𝜔) |𝜓 i = (𝜀 − ℏ𝜔)𝑎ˆ |𝜓 i .

In other words, |𝜓 0i = 𝑎ˆ |𝜓 i is also an eigenstate of 𝐻ˆ , with

energy 𝜀 − ℏ𝜔 . Similarly, we have 𝐻ˆ 𝑎ˆ† |𝜓 i = (𝜀 + ℏ𝜔)𝑎 † |𝜓 i .
Thus, 𝑎ˆ lowers and 𝑎ˆ† raises the energy of the state |𝜓 i by
one quantum of energy ℏ𝜔 . This is where their names come
from.

From points 1. and 2., it follows that there exists a state | 0i ∈ H

s.t. 𝑎ˆ | 0i = 0. Thus, | 0i is referred to as the ground state, i.e., the
state with the lowest possible energy. Let us find the energy of the
oscillator in the state | 0i :
 
† 1
𝐻 | 0i = ℏ𝜔 𝑎ˆ 𝑎ˆ +
ˆ | 0i (4.12)
2
ℏ𝜔
= | 0i .
2
n  o
Thus the spectrum of 𝐻ˆ is {𝜀𝑛 } = ℏ𝜔 𝑛 + 21 , and the corre-
sponding eigenstates are simply written as {|𝑛i}. That is, 𝐻ˆ |𝑛i =
 4 Lecture 4 20

 
ℏ𝜔 𝑛 + 21 |𝑛i .

4.3 Symbolic operator differential

Let 𝑞ˆ and 𝑝ˆ be a conjugate pair and 𝑞ˆ be such that it has a continuous

spectrum.

Such a conjugate pair satisfies the commutation relation [𝑞, ˆ =

ˆ 𝑝]
iℏ. Some calculations are simplified if we symbolically define
𝑝ˆ = −iℏ𝜕𝑞ˆ , where 𝜕𝑞ˆ means we take symbolically the derivative
w.r.t. 𝑞ˆ. We will check below that this symbolical differentiation is
consistent with the commutation relation.

For example, ∀ |𝜓 i ∈ H

ˆ |𝜓 i = 𝑓 0 (𝑞)


𝜕𝑞ˆ 𝑓 (𝑞) ˆ + 𝑓 (𝑞)𝜕
ˆ 𝑞ˆ |𝜓 i , (4.13)

where 𝑓 is a continuously differentiable function and 𝑓 0 denotes

its derivative.

Let us check the above claim that [𝑞, ˆ = iℏ when 𝑝ˆ = −iℏ𝜕𝑞ˆ :

ˆ 𝑝]

[𝑞, ˆ = 𝑞ˆ𝑝ˆ − 𝑝ˆ𝑞ˆ

ˆ 𝑝] (4.14)



= 𝑞ˆ −iℏ𝜕𝑞ˆ − −iℏ𝜕𝑞ˆ 𝑞ˆ
= −iℏ𝑞𝜕
ˆ 𝑞ˆ + iℏ𝜕𝑞ˆ 𝑞ˆ


= −iℏ𝑞𝜕
ˆ 𝑞ˆ + iℏ 𝜕𝑞ˆ 𝑞ˆ +iℏ𝑞𝜕
ˆ 𝑞ˆ
|{z}
=𝐼ˆ
= iℏ.

4.4 Solving the ground state in the position

representation

Using the fact that 𝑎ˆ | 0i = 0 and the above definition of the symbolic
differential, we have

0 = h𝑥 0 |𝑎|
ˆ 0i (4.15)
r   ∫ 
0 𝑚𝜔 i
= h𝑥 | 𝑥ˆ + (−iℏ𝜕𝑥ˆ ) ˜ 𝑥˜ | | 0i
d𝑥˜ |𝑥ih
2ℏ 𝑚𝜔
| {z }| {z }
= 𝑎ˆ = 𝐼ˆ
r ∫  
𝑚𝜔 0 ℏ
= d𝑥˜ h𝑥 |𝑥i
˜ 𝑥˜ + 𝜕𝑥˜ 𝜓 0 (𝑥)
˜
2ℏ |{z} 𝑚𝜔 |{z}
˜ 0)
𝛿 (𝑥−𝑥 ≕ h𝑥˜ | 0 i
 4 Lecture 4 21

 
ℏ
⇒ 𝑥+ 𝜕𝑥 𝜓 0 (𝑥) = 0 (4.16)
𝑚𝜔
 
𝑥 2𝑚𝜔 2
⇒ 𝜓 0 (𝑥) = 𝐶 exp − , (4.17)
2ℏ

𝑚𝜔 1/4


where 𝐶 = 𝜋ℏ is a normalization coefficient.

We may further derive the wave function of the first excited state
from 𝜓 1 (𝑥) = 𝐶˜ h𝑥 | 𝑎ˆ† | 1i where we do not even need to solve a
differential equation since we know 𝜓 (𝑥)0 and we may simply
multiply it and take the first derivative. Similarly, we may proceed
to derive the wave function of any state |𝑛i . However, we do not do
this here, but come back to the harmonic oscillator on the second
half of the course where we study the wave functions of the excited
states further.

4.5 Uncertainty relations

Definition 4.5.1 The Heisenberg uncertainty relation is defined as

ℏ
Δ𝑞Δ𝑝 ≥ , (4.18)
2

where Δ𝐴2 = h𝐴ˆ2 i − h𝐴i

ˆ 2 and [𝑞, ˆ = iℏ since 𝑞ˆ and 𝑝ˆ are a canonical
ˆ 𝑝]
1: Warning: does not strictly speaking
conjugate pair .
1
apply if an operator is not bounded

Definition 4.5.2 The Robertson uncertainty relation is defined as

1
Δ𝐴Δ𝐵 ≥ h[𝐴,
ˆ 𝐵]i
ˆ , (4.19)
2
ˆ 𝐵ˆ ∈ L (H ) may be unbounded, 𝐴ˆ = 𝐴ˆ† , 𝐵ˆ = 𝐵ˆ † , and
where 𝐴,
h·i B h𝜓 | · |𝜓 i .

Let us prove the above relations. To this end, we define |𝑓 i =

𝐴ˆ − h𝐴i
ˆ |𝜓 i and |𝑔i = 𝐵ˆ − h𝐵i
ˆ |𝜓 i . Then, Math on norm of C
For 𝑧 ∈ C,
Δ𝐴2 = h𝜓 | 𝐴ˆ − h𝐴i ˆ 2 |𝜓 i


(4.20)


|𝑧| 2 = ( Re 𝑧) 2 + ( Im 𝑧) 2
= h𝜓 | 𝐴ˆ − h𝐴i ˆ 𝐼ˆ 𝐴ˆ − h𝐴i
ˆ 𝐼ˆ |𝜓 i
𝑧 − 𝑧∗ 2
 
≥ ( Im 𝑧) 2 =
| {z }
((𝐴− ˆ 𝐼ˆ) |𝜓 i ) †
ˆ h𝐴i 2i

= h𝑓 |𝑓 i = k|𝑓 ik 2 ,

and similarly,

Δ𝐵 2 = h𝑔|𝑔i = k|𝑔ik 2 . (4.21)

 4 Lecture 4 22

Then, the Cauchy–Schwarz inequality implies

|h𝑓 |𝑔i| ≤ k|𝑓 ik k|𝑔ik (4.22)

⇒ Δ𝐴 Δ𝐵 ≥ | h𝑓 |𝑔i |
2 2 2
(4.23)
|{z}
∈C
ˆ 𝐼ˆ |𝜓 i 2



= h𝜓 | 𝐴ˆ − h𝐴i
ˆ 𝐼ˆ 𝐵ˆ − h𝐵i
ˆ 𝐼ˆ |𝜓 i 2





h𝜓 | 𝐴ˆ − h𝐴i
ˆ 𝐼ˆ 𝐵ˆ − h𝐵i ˆ 𝐼ˆ 𝐴ˆ − h𝐴i
ˆ 𝐼ˆ |𝜓 i − h𝜓 | 𝐵ˆ − h𝐵i
≥
4
2
h𝜓 | [𝐴 − h𝐴i𝐼, 𝐵 − h𝐵i𝐼 ] |𝜓 i
ˆ ˆ ˆ ˆ ˆ ˆ
=
4
2
h[𝐴, 𝐵]i
ˆ ˆ
= 
4
 Lecture 5 5
5.1 Intended learning outcomes
I Apply the operator exponential to symbolically solve the
Schrödinger equation
I Differentiate between a qubit and a general quantum system
I Represent a qubit state on the Bloch sphere

5.2 Unitary temporal evolution

Let |𝜓 (𝑡)i ∈ H and 𝐻ˆ ∈ L (H ) be the Hamiltonian of a system. Let

|𝜓 (𝑡 = 0)i = |𝜓 ( 0)i be the initial state of the system, the state at
𝑡 = 0. As discussed before, the temporal evolution is then given by
the Schrödinger equation:

iℏ𝜕𝑡 |𝜓 (𝑡)i = 𝐻ˆ |𝜓 (𝑡)i (5.1)

i𝐻ˆ
⇐⇒ 𝜕𝑡 |𝜓 (𝑡)i = − |𝜓 (𝑡)i .
ℏ

Note that we have assumed that 𝐻ˆ is independent of time.

Definition 5.2.1 For 𝐴ˆ ∈ L (H ) , let

∞ ˆ𝑛
𝐴ˆ
Õ 𝐴
e B . (5.2)
𝑛=0
𝑛!

ˆ ˆ ˆ ˆ
Note that in general e𝐴 e𝐵 ≠ e𝐴+𝐵 . The equality holds if 𝐴ˆ and 𝐵ˆ
commute. 1 1: The general expression for 𝐶ˆ in
ˆ ˆ ˆ
e𝐴 e𝐵 = e𝐶 is given by the Baker–
With this definition, Campbell–Hausdorff formula.

∞ ˆ𝑛 𝑛
!
ˆ
𝐴𝑡
Õ 𝐴 𝑡
𝜕𝑡 e = 𝜕𝑡 (5.3)
𝑛=0
𝑛!
∞ ˆ𝑛 𝑛−1
Õ 𝐴 𝑛𝑡
=
𝑛=1
𝑛!
∞ ˆ 𝑛−1


Õ 𝐴𝑡
= 𝐴ˆ Math on a diff.eq.
𝑛=1
(𝑛 − 1) !
∞ ˆ 𝑛


Õ 𝐴𝑡
= 𝐴ˆ 𝜕𝑥 𝑓 (𝑥) = 𝜆𝑓 (𝑥) =⇒ 𝑓 (𝑥) = 𝐴e𝜆𝑥
𝑛=0
𝑛!
ˆ
= 𝐴ˆe𝐴𝑡 .
 5 Lecture 5 24

The temporal evolution can then be written as

|𝜓 (𝑡)i = e−i𝐻𝑡 /ℏ |𝜓 ( 0)i

ˆ
(5.4)
C 𝑈ˆ (𝑡) |𝜓 ( 0)i ,


where 𝑈ˆ (𝑡) = exp −i𝐻𝑡/ℏ
ˆ is the time-evolution operator, or some-
times referred to as the propagator of the system.

Recalling that 𝐻ˆ † = 𝐻ˆ , we observe that

 †
𝑈ˆ (𝑡) † = e−i𝐻𝑡 /ℏ
ˆ
(5.5)

= ei𝐻𝑡 /ℏ
ˆ

= 𝑈ˆ (−𝑡),

from which it follows that

𝑈ˆ (𝑡) †𝑈ˆ (𝑡) |𝜓 ( 0)i = 𝑈ˆ (𝑡) † |𝜓 (𝑡)i (5.6)

= 𝑈ˆ (−𝑡) |𝜓 (𝑡)i
= |𝜓 ( 0)i ,

or in other words, 𝑈ˆ †𝑈ˆ = 𝐼ˆ, or 𝑈ˆ † = 𝑈ˆ −1 . Such an operator 𝐴ˆ that

satisfies 𝐴ˆ†𝐴ˆ = 𝐼ˆ is said to be unitary.

Let {|𝜓𝑛 i} ∈ H be an eigenbasis of the Hamiltonian 𝐻ˆ , i.e., 𝐻ˆ |𝜓𝑛 i =

∞ ∈ R. We can expand the initial state in this
𝜆𝑛 |𝜓𝑛 i , where {𝜆𝑛 }𝑛=
Í∞0
basis as |𝜓 ( 0)i = 𝑛=0 𝑐𝑛 |𝜓𝑛 i , where 𝑐𝑛 = h𝜓𝑛 |𝜓 i ∈ C, and thus
write the state at time 𝑡 as

|𝜓 (𝑡)i = e−𝑖 𝐻𝑡 /ℏ |𝜓 ( 0)i

ˆ
(5.7)
∞
𝑛 ! Õ ∞
!
Õ −i𝐻𝑡/ℏ
ˆ
= 𝑐𝑚 |𝜓𝑚 i
𝑛=0
𝑛! 𝑚=0
∞ Õ ∞
𝑛 !
Õ −i𝐻𝑡/ℏ
ˆ
= 𝑐𝑚 |𝜓𝑚 i
𝑚=0 𝑛=0
𝑛!
∞
Õ
= e−i𝜆𝑚 𝑡 /ℏ𝑐𝑚 |𝜓𝑚 i
𝑚=0
Õ∞
= 𝑐𝑚 e−i𝜆𝑚 𝑡 /ℏ |𝜓𝑚 i .
𝑚=0
 5 Lecture 5 25

5.3 Case of temporally dependent

Hamiltonian

Let now the Hamiltonian 𝐻ˆ = 𝐻ˆ (𝑡) be time-dependent. The

Schrödinger equation still holds:

iℏ𝜕𝑡 |𝜓 (𝑡)i = 𝐻ˆ (𝑡) |𝜓 (𝑡)i , (5.8)

and the evolution is unitary. Thus ∃{𝑈ˆ (𝑡)} ∈ L (H ) s.t.

𝑈ˆ (𝑡) |𝜓 ( 0)i = |𝜓 (𝑡)i , ∀ |𝜓 (𝑡)i ∈ H (5.9)

⇒ iℏ𝜕𝑡 𝑈ˆ (𝑡) |𝜓 ( 0)i = 𝐻ˆ (𝑡) 𝑈ˆ (𝑡) |𝜓 ( 0)i (5.10)
⇒ iℏ𝜕𝑡 𝑈ˆ (𝑡) = 𝐻ˆ (𝑡)𝑈ˆ (𝑡). (5.11)

This is equivalent to the Schrödinger equation.

Exercise
Build 𝑈ˆ (𝑡) for 𝐻ˆ (𝑡) .

5.4 Properties of unitary operators

For any two unitary operators 𝑈ˆ 1 and 𝑈ˆ 2 , we have

†
−1
𝑈ˆ 1𝑈ˆ 2 = 𝑈ˆ 2†𝑈ˆ 1† = 𝑈ˆ 2−1𝑈ˆ 1−1 = 𝑈ˆ 1𝑈ˆ 2 . (5.12)

That is, 𝑈ˆ 1𝑈ˆ 2 is also unitary.

Let |𝜓 i , |𝜙i ∈ H and 𝑈ˆ † = 𝑈ˆ −1 ∈ L (H ) . We define

|𝜓 0i = 𝑈ˆ |𝜓 i and |𝜙 0i = 𝑈ˆ |𝜙i , (5.13)

for which we have

h𝜓 |𝜙i = h𝜓 |𝐼ˆ|𝜙i = h𝜓 |𝑈ˆ −1𝑈ˆ |𝜙i (5.14)

= h𝜓 |𝑈ˆ †𝑈ˆ |𝜙i = 𝑈ˆ |𝜓 i , 𝑈ˆ |𝜙i

= h𝜓 0 |𝜙 0i .
2: sometimes reflections as well
Unitary operators can be considered as rotations.2

5.5 Qubit

A qubit can refer either to a physical system or to a mathematical

construction. In either case, it is modeled by a a two-level quantum
system as follows:
 5 Lecture 5 26

Let H2 = span{| 0˜ i , | 1˜ i}, where h0˜ | 0˜ i = 1 = h1˜ | 1˜ i .

H2 fully describes all possible states of the qubit where

|𝜓 i ∈ H2 and k |𝜓 i k = 1. (5.15)

Thus the qubit Hamiltonian 𝐻ˆ 𝑞 has just two eigenvalues 𝜀 1 ≤ 𝜀 2 ∈ R

and the corresponding eigenvectors are | gi and | ei , respectively.3
3: Ground and excited

Thus,

𝐻ˆ 𝑞 = 𝜀 1 | gihg | + 𝜀 2 | eihe | (5.16)

𝜀
(𝜀 1 + 𝜀 2 ) (𝜀 1 + 𝜀 2 )
= − | gihg | + | eihe | + | gihg | + | eihe |
2 2 2
𝜀
𝜀1 + 𝜀2
= − | gihg | + | eihe | + 𝐼ˆ
2 2
| {z }
We can disregard this since it just equally changes the phase of all |𝜓 i ∈ H2

where 𝜀 = 𝜀 2 − 𝜀 1 .


Thus 𝐻ˆ 𝑞 = − 𝜀2 | gihg | − | eihe | .

We can define the qubit states | 0i B | gi and | 1i B | ei .

Thus,
𝜀
𝐻ˆ 𝑞 = − 𝜎ˆ z, 𝜎ˆ z = | 0ih0 | − | 1ih1 | (5.17)
2
 𝜀 
−2 0
=ˆ .
0 + 𝜀2

The temporal evolution is given by

|𝜓 (𝑡)i = e−i𝐻𝑞 𝑡 /ℏ |𝜓 ( 0)i ,

ˆ
|𝜓 ( 0)i = 𝑐 0 | 0i + 𝑐 1 | 1i (5.18)
+i 𝜀2 𝜎ˆ z𝑡 /ℏ
=e |𝜓 ( 0)i
i 𝜀2 𝑡 /ℏ 𝜀
=e 𝑐 0 | 0i + e−i 2 𝑡 /ℏ𝑐 1 | 1i .
5
QHO JJ
4

5.6 How to set up a qubit from a physical 3

system
Energy

2
subspace
Comp.

1
Very few physical systems are qubits. However, it is possible to take
some physical systems and confine the dynamics to a subspace 0
- - 0
of two states. For example, a spin is a natural two-level system,
Superconducting phase
but confined (for example in atoms). Another example is a non-
Figure 5.1: Non-linear harmonic os-
linear system where 𝜀 0 < 𝜀 1 < 𝜀 2 · · · are eigenvalues of 𝐻ˆ s.t. cillator with a Josephson junction (JJ).
𝜀 1 − 𝜀 0 ≠ 𝜀 2 − 𝜀 1 . See Fig. 5.1. Notice that the gap ℏ𝜔 01 ≠ ℏ𝜔 12 , i.e.,
the energy states are non-equidistant.
Figure from Ref. [1].
 5 Lecture 5 27

5.7 Pauli operators

Definition 5.7.1 The Pauli operators are

𝜎ˆ z = | 0ih0 | − | 1ih1 | (5.19)

𝜎ˆ x = | 0ih1 | + | 1ih0 | (5.20)
𝜎ˆ y = −i | 0ih1 | + i | 1ih0 | (5.21)

Properties

The Pauli operators have a number of interesting properties:

𝜎ˆ𝛼2 = 𝐼ˆ ∀𝛼 ∈ {x, y, z} (5.22)

𝜎ˆ𝛼† = 𝜎ˆ𝛼 ∀𝛼 (5.23)
Õ
[𝜎ˆ𝑖 , 𝜎ˆ 𝑗 ] = 2i𝜎ˆ𝑘 𝜀𝑖 𝑗𝑘 , ∀𝑖, 𝑗 ∈ {x, y, z}, (5.24)
𝑘 ∈ {x,y,z }

where

 +1 if (𝑖, 𝑗, 𝑘) ∈ {( x, y, z), ( z, x, y), ( y, z, x)},




𝜀𝑖 𝑗𝑘 = −1 if (𝑖, 𝑗, 𝑘) ∈ {( y, x, z), ( z, y, x), ( x, z, y)}, (5.25)

 0
 otherwise,

is the Levi-Civita symbol.

Definition 5.7.2

𝜎ˆ − = | 0ih1 | (5.26)
+ − †
𝜎ˆ = (𝜎ˆ ) = | 1ih0 | . (5.27)

Exercise
Show that
ˆ
ei𝜑 𝑎® ·𝜎® = 𝐼ˆ cos 𝜑 + i 𝑎® · 𝜎®ˆ sin 𝜑,

® = 1 and 𝑎® · 𝜎®ˆ = 𝑎 x𝜎ˆ x + 𝑎 y𝜎ˆ y + 𝑎 y𝜎ˆ z .

where 𝑎® ∈ R3 , k𝑎k

5.8 Bloch sphere

Definition 5.8.1 A qubit state can always be expressed as

𝜃 𝜃
|𝜓 i = cos | 0i + ei𝜑 sin | 1i , (5.28)
2 2
where 𝜑 is the azimuthal angle and 𝜃 is the polar angle.
Figure 5.2: Bloch sphere representa-
tion [2].
 5 Lecture 5 28

Note that since a global phase of the state ei𝛼 does not affect any
measurement outcome, i.e.,

ˆ i = h𝜓 |𝐴ˆe−i𝛼 ei𝛼 |𝜓 i = h𝜓 | e−i𝛼 𝐴ˆei𝛼 |𝜓 i = hei𝛼 𝜓 |𝐴ˆei𝛼 |𝜓 i ,

h𝜓 |𝐴|𝜓
(5.29)

we can always choose 𝑐 0 ∈ R in |𝜓 i = 𝑐 0 | 0i + 𝑐 1 | 1i .

Thus, for each state there are unique 𝜃 ∈ [ 0, 𝜋) and 𝜑 ∈ [ 0, 2𝜋)

which correspond to a point on a unit sphere as shown in Fig. 5.2.

Exercise
Show that 𝑈ˆ (𝑡) are rotations of the Bloch vectors.
 Lecture 6 6
Last lecture from Mikko.

6.1 Intended learning outcomes

I Apply tensor product to construct a quantum register of 𝑁
qubits
I Identify the constituents of a quantum algorithm
I Apply the commutator to identify conserved quantities

6.2 Tunable Hamiltonian for quantum

gates

Let span {| 0i , | 1i} = H2 and assume that control over the Hamilto-
® · 𝜎®ˆ , where 𝑎® ∈ R3, k𝑎k
nian s.t. 𝐻ˆ = 𝜀 0𝑎(𝑡) ® = 1, and 𝜀 0 ∈ R has units
of energy.
1: An unitary operation on a qubit is
Thus any unitary evolution1 𝑈ˆ = 𝐼ˆ cos 𝜃 + i 𝑏® · 𝜎®ˆ sin 𝜃 can be referred to as a single-qubit gate
implemented, for example, by a control sequence

 𝑡 <0
 0,



𝑎(𝑡) ®
® = −𝜀 0𝑏, 0 ≤ 𝑡 ≤ 𝜃 ℏ/𝜀 0 . (6.1)


 0,
 𝜃 ℏ/𝜀 0 < 𝑡

There are many other ways of course. Note that there is also a way
to use
𝜀
𝐻ˆ = − 𝜎ˆ z (6.2)
2

and apply a field 𝐻®ex (𝑡) = Ω2 𝜎ˆ x sin (𝜔𝑡 + 𝜙) , where 𝜔 = 𝜀ℏ . That will
result in so-called Rabi oscillations to be discussed later.

6.3 Single-qubit gates: examples

 
0 1
I The NOT gate corresponds to 𝜎ˆ x = | 0ih1 | + | 1ih0 | = ˆ .
1 0
 
1 1 1
I Hadamard gate corresponds to 𝐻 g =
ˆ √ 1
(𝜎ˆ x + 𝜎ˆ z ) =ˆ √ .
2 2 1 −1
 
1 0
I Phase flip corresponds to 𝜎ˆ z = | 0ih0 | − | 1ih1 | = ˆ .
0 −1
 6 Lecture 6 30

Exercise
Find 𝑎(𝑡)
ˆ implementing:

𝐻ˆ g𝜎ˆ x𝐻ˆ g = 𝜎ˆ z
𝐻ˆ g† = 𝐻ˆ g = 𝐻ˆ g−1 .

6.4 Qubit measurement

Let |𝜓 i ∈ H2 be a qubit state. Thus we may write |𝜓 i = 𝑐 0 | 0i +

𝑐 1 | 1i , where 𝑐 0, 𝑐 1 ∈ C s.t. |𝑐 0 | 2 + |𝑐 1 | 2 = 1. Thus the measurement
probabilities are given by

𝑃 0 = |h0 |𝜓 i| 2 = |𝑐 0 | 2 (6.3)
2 2 2
𝑃 1 = |h1 |𝜓 i| = |𝑐 1 | = 1 − |𝑐 0 | (6.4)

After applying a quantum gate 𝑈ˆ on |𝜓 i the probabilities are given

by

𝑃 0 = h0 |𝑈ˆ |𝜓 i = h𝜓 |𝑈ˆ † | 0ih0 |𝑈ˆ |𝜓 i = h0˜ |𝜓 i ,

2 2
(6.5)

2 2
where | 0˜ i = 𝑈ˆ † | 0i . Similarly for 𝑃 1 = h1 |𝑈ˆ |𝜓 i = h1˜ |𝜓 i .

6.5 2-qubit system

( 1) ( 2) On the tensor product
The Hilbert space H4 = H2 ⊗ H2 of a system composed of two
qubits is 4-dimensional. The symbol ⊗ denotes the tensor product The tensor product (or Kronecker
of Hilbert spaces or vectors. Single-qubit operators are of the form product) is a bilinear composition

𝐴ˆ1 ⊗ 𝐼ˆ and 𝐼ˆ ⊗ 𝐴ˆ2 , where 𝐴ˆ1 ∈ L (H2( 1) ) and 𝐴ˆ2 ∈ L (H2( 2) ) .

of the two vector spaces (with min-
imal constraints).
Let 𝐴ˆ ⊗ 𝐵ˆ = 𝐶ˆ ∈ L (H4 ) and 𝐷ˆ ⊗ 𝐸ˆ = 𝐹ˆ ∈ L (H4 ) . From the properties
of the tensor product, it follows that





𝐶ˆ 𝐹ˆ = 𝐴ˆ ⊗ 𝐵ˆ 𝐷ˆ ⊗ 𝐸ˆ = 𝐴ˆ𝐷ˆ ⊗ 𝐵ˆ 𝐸ˆ . (6.6)

Definition 6.5.1 We construct the basis for the two-qubit Hilbert

space H4 as

| 00i B | 0i ⊗ | 0i (6.7)
| 01i B | 0i ⊗ | 1i (6.8)
| 10i B | 1i ⊗ | 0i (6.9)
| 11i B | 1i ⊗ | 1i , (6.10)
 6 Lecture 6 31

( 1) ( 2)
where {| 0i , | 1i} is an orthonormal basis for H2 and H2 , respec-
tively.
Thus for |𝜓 i ∈ H4 , we may write

3
Õ
|𝜓 i = 𝑐𝑘 |𝑘i (6.11)
𝑘=0
= 𝑐 0 | 00i + 𝑐 1 | 01i + 𝑐 2 | 10i + 𝑐 3 | 11i
= 𝑐 0 | 0i + 𝑐 1 | 1i + 𝑐 2 | 2i + 𝑐 3 | 3i

where |𝑘i B |𝑘 1𝑘 2 i , where 𝑘 1𝑘 2 is a binary representation of 𝑘 .

Additionally, for 𝐶ˆ = 𝐴ˆ ⊗ 𝐵ˆ ∈ L (H4 ) , we have

3
Õ 3
Õ
𝐶ˆ |𝜓 i = 𝐶ˆ 𝑐𝑘 |𝑘i = 𝑐𝑘 𝐶ˆ |𝑘i (6.12)
𝑘=0 𝑘=0
3
Õ
= 𝑐𝑘 𝐴ˆ ⊗ 𝐵ˆ |𝑘i
𝑘=0
|{z}
∈H4
3
Õ
= 𝑐𝑘 𝐴ˆ |𝑘 1 i ⊗ 𝐵ˆ |𝑘 2 i .
𝑘=0
|{z} |{z}
( 1) ( 2)
∈H2 ∈H2

6.6 Examples of two-qubit gates

Controlled NOT (CNOT) gate where qubit 1 is the control qubit

and qubit 2 is the target qubit corresponds to
( 1,2)
𝐶ˆNOT = | 0ih0 | ⊗ 𝐼ˆ + | 1ih1 | ⊗ 𝜎ˆ x (6.13)
1 0 0 0
   
 
0 1 0 0 𝐼 0
=ˆ  =
 
.
0 0 0 1 0 𝜎x
 
0 0 1 0
 

Exercise
I Construct the above matrix representations
I Express CNOT that has qubit 1 as the target qubit

6.7 𝑛-qubit system

For a system with 𝑛 qubits, we usually define 𝑁 B 2𝑛 = dim {H2𝑛 },

and we have the following properties:
 6 Lecture 6 32

( 1) ( 2) (𝑛)
I H2𝑛 = H2 ⊗ H2 ⊗ · · · ⊗ H2
Í2𝑛 −1
I |𝜓 i = 𝑘= 0 𝑐 𝑘 |𝑘i = 𝑐 0 | 00 · · · 0i + 𝑐 1 | 0 · · · 01 i + · · · , where
𝑛 zeroes 𝑛−1 zeroes
again |𝑘i means |𝑘 1𝑘 2 . . . 𝑘𝑛 i , where 𝑘 1𝑘 2 . . . 𝑘𝑛 is 𝑘 written in
binary
I 𝐼ˆ ⊗ · · · ⊗ 𝐼ˆ ⊗𝐴⊗
ˆ 𝐼ˆ ⊗ · · · ⊗ 𝐼ˆ is a single-qubit operator for qubit 𝑚 .
𝑚−1 𝑛−𝑚

6.8 Quantum algorithms for 𝑛 qubits

In general, a quantum algorithm is a procedure consisting of the

following steps:

1. Initialize qubits to | 0i .2 2: Not necessarily all qubits

2. Apply a desired 𝑛 -qubit gate U.3 3: Can be constructed from single
3. Measure qubits.4 and two-qubit gates
4. Use measurement data and go to 1, unless algorithm fin- 4: Not necessarily all qubits
ished.5 5: In the simplest case one goes only
once through 1. → 4. and initializes
and measures all qubits in 1. and 3.,
Exercise respectively.
Deustch algorithm

6.9 Entanglement for two qubits

Definition 6.9.1 A quantum state of two qubits is defined to be

entangled iff it cannot be represented as a product of two single-qubit
states.
( 1)
Thus ∀ |𝜓 i ∈ H4 that are not entangled ∃ |𝜓 1 i ∈ H2 and |𝜓 2 i ∈
H2( 2) s.t.

|𝜓 i = |𝜓 1 i ⊗ |𝜓 2 i (6.14)

Examples of so-called maximally entangled states are Bell states

1
|Φ± i = √ (| 00i ± | 11i) (6.15)
2
1
|Ψ± i = √ (| 01i ± | 10i) (6.16)
2

6.10 Commuting operators

ˆ 𝐵ˆ ∈ L (H ) be Hermitian operators with [𝐴,

Let 𝐴, ˆ = 0. In this
ˆ 𝐵]
case, it can be shown that there exists a complete eigenbasis of 𝐴ˆ
that is also an eigenbasis of 𝐵ˆ .
 6 Lecture 6 33

Especially if [𝐴,
ˆ 𝐻ˆ (𝑡)] = 0, ∀𝑡 , the eigenvalues of 𝐴ˆ are referred to
as conserved quantities since we have

𝐴ˆ |𝜓 (𝑡)i = 𝐴ˆ𝑈ˆ (𝑡) |𝜓 ( 0)i = 𝑈ˆ (𝑡)𝐴ˆ |𝜓 ( 0)i (6.17)

= 𝜆𝑈ˆ (𝑡) |𝜓 ( 0)i = 𝜆 |𝜓 (𝑡)i ,

where we have assumed that 𝐴ˆ |𝜓 ( 0)i = 𝜆 |𝜓 ( 0)i , i.e., we start from

an eigenstate of 𝐴ˆ.
 References

[1] Niko Savola. Electromagnetic simulation of superconducting

qubit–resonator coupling. English. Bachelor’s Thesis. 2020. url:
http://urn.fi/URN:NBN:fi:aalto- 202009155417 (cited
on page 26).
[2] Wikimedia Commons. File:Bloch Sphere.svg — Wikimedia Com-
mons, the free media repository. [Online; accessed 5th August
2021]. 2020. url: %5Curl%7Bhttps://commons.wikimedia.
org/w/index.php?title=File:Bloch_Sphere.svg&oldid=
514662857%7D (cited on page 27).