QM2020

Quantum Mechanics I
Oscar Loaiza-Brito
August 13, 2020
Abstract
This is a serie of lectures on advanced quantum mechanics for one semester

on the graduate program at Universidad de Guanajuato. The lectures are
written according to the timetable of the oficial calendar at Universidad de
Guanajuato consisting on 16 weeks.
i
Contents
Horarios y calificación v
Bibliografía . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
Topics you should know before taking this course! . . . . . . . . . . . . . . . . . viii
1 Introduction 1
1.1 A brief review on wave mechanics . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Bounded states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Into a mathematical framework for quantum mechanics . . . . . . . . . . 4
1.3.1 Physical States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Exercises, due to T2-day . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Fundamental concepts I 5
2.1 Physical states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Dual vector space H∗ . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Clousure relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.3 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.4 Adjoint operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Fundamental Concepts II 11
3.1 Unitary transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.1 Example: U (1)-operators . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Change of basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 Observables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3.1 Eigenvalues and eigenvectors . . . . . . . . . . . . . . . . . . . . . 14
3.4 Hermitian observables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.5 Observables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 Observables 16
4.1 Compatible observables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Measurement and Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3 Space and Momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.3.1 Measuring space and momentum . . . . . . . . . . . . . . . . . . . 20
ii
Contents iii
4.3.2 Uncertainty on measuring position and momentum . . . . . . . . . 21

4.4 Suggested exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5 Quantum Dynamics 23
5.1 Postulates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.2 Energy states and time evolution. . . . . . . . . . . . . . . . . . . . . . . . 24
5.3 Momentum as generator of translations . . . . . . . . . . . . . . . . . . . 26
5.4 Heisenberg Picture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
6 Harmonic Oscillator 29
6.1 Motivitaion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.2 Energy states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.3 Observables and uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . 31
7 Harmonic Oscillator II 32
7.1 Time evolution of X̂ and P̂ . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7.2 Energy wave functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
7.3 Hermite Polynomials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8 Gauge Transformations in Quantum Mechanics 36

8.1 Classical description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
8.2 Gauge invariance in quantum mechanics . . . . . . . . . . . . . . . . . . . 37
9 Gauge quantum effects 39

9.1 Bohm-Aharonov effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
9.2 Tool Box 1. Complex Analysis and Ampere’s law . . . . . . . . . . . . . . 42
9.3 Magnetic Monopole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
9.4 Landau levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
10 Angular Momentum 49
10.1 General Angular Momentum . . . . . . . . . . . . . . . . . . . . . . . . . 49
10.2 Eigenvectors and eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . 52
10.3 Matrix elements of Angular Momentum . . . . . . . . . . . . . . . . . . . 53
10.4 Construction of a finite rotation operator . . . . . . . . . . . . . . . . . . 54
11 Spinors 55
11.1 A warming up example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
11.2 Rotations generated by SU (2) . . . . . . . . . . . . . . . . . . . . . . . . . 57
11.3 Finally...spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
11.4 Spinors and Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
11.5 Energy states and spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
11.6 Spin precession . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
11.7 Pauli’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
11.8 Tool Box 2. Classical description of Angular Momentum and its interaction
with an external magnetic field. . . . . . . . . . . . . . . . . . . . . . . . . 67
iv Contents
12 Addition of Angular Momentum 69

12.1 Irreducible representations . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
12.2 Addition of Angular Momentum . . . . . . . . . . . . . . . . . . . . . . . 70
12.2.1 Addition of two 1/2 spin particles . . . . . . . . . . . . . . . . . . 72
12.3 Clebsch-Gordon Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . 73
12.3.1 Properties of the Clebsch-Gordon coefficients . . . . . . . . . . . . 74
13 Hydrogen atom 76
13.1 Reducing the system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.2 Quantum mechanical description . . . . . . . . . . . . . . . . . . . . . . . 77
13.2.1 Projecting into the configuration space . . . . . . . . . . . . . . . . 78
13.3 Radial equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
13.3.1 Limiting our ignorance . . . . . . . . . . . . . . . . . . . . . . . . 79
13.3.2 Solution to Radial Equation . . . . . . . . . . . . . . . . . . . . . . 80
13.3.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
13.4 Angular equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
A Some notes on Lie Groups 85
B Stern-Gerlacht experiment 87
C Transformations and unitary generators 88
D Probability currents 90
Agradecimientos
Estas notas son el resultado de haber impartido el curso de Mecánica Cuántica 1, en

la maestría en Física de la Universidad de Guanajuato. Agradezco a todos los alumnos
del curso Agosto-Diciembre de 2019, con quienes empecé a estructurar estas notas. De
igual manera, agradezco a los estudiantes del curso Enero-Junio 2020, quienes no solo
me permitieron continuar con este proyecto, sino que colaboraron en la corrección de
los múltiples errores de escritura y cálculo: a Marco Antonio García Sánchez, ,Karen
González Flores, Roberto Hinojosa Dominguez, Leonardo Morales Padilla, Jorge Alberto
Oropeza Muñoz, Marco Antonio Ortiz Villacana, José de Jesús Pérez, Gemma Elizabeth
Pérez Cuellar, Ana Laura Ramírez Díaz, Kevin Alexander Rodríguez, Aprecio el interés y
esfuerzo realizado por estos estudiantes durante la pandemia de COVID-19.
v
Horarios y calificación
Los temas que se verán son los siguientes:
Bibliografía
Los exámenes así como los ejercicios, estarán basados en estas notas. Para su elaboración
se consultó las siguientes referencias:
• Modern Quantum Mechanics, J.J. Sakurai.
• Quantum Mechanics Vol. 1, Cohen-Tannoudji
• The Theory of Spinors, Cartan.
• Lectures on Quantum Mechanics, S. Weinberg.
• Quantum Mechanics, Susskind and Friedman
• Curso Abreviado de Física Teórica, Libro 2, L. Landau y E. Lifshitz.
vi
Topics you should know before taking this course! vii
Topics you should know before taking this course!
This is what you should know in detail before starting this advanced course. That means
that all the following topics shall not be covered in these lectures.
• Experiments indicating that classical mechanics is not a suitable framework to

explain the associated phenomena, as Black Body Radiation, Photoelectric effect,
Compton scattering, etc.
• Borh’s model of hydrogen atom
• Wave functions and Scŕodinger equation’s solutions for different types of wells.
• Physical meaning of the uncertainty principle.
• Physical motivation for the wave-particle dual behaviour of matter.
• De Broglie’s hypothesis.
In case you think your kwnoledge on these topics are weak, please go directly to your
text books and review them carefully before starting this course!
Chapter 1
Introduction
In this lecture we are going to develop the mathematical formalism required to translate
all the quantum ideas we have developed in basic courses on quantum mechanics into
precise mathematical concepts. This will gives a powerful tool to compute physical
relevant quantities as well as to conclude some fascinating ideas concerning the quantum
behaviour of matter.
As in all physical studies, we need to understand the limits of a description and realize
the size and energy ranges in which our theory is valid. Hence, in this course we are
not considering particles with high velocities compared to light’s nor the presence of
strong gravitational fields. In that sense, we are going to study a non-relativistic quantum
behaviour of matter in absence of gravitational fields. Notice the implications of these
assumptions: a relativistic version of quantum mechanics is expected as well as a quantum
version of gravity. The former is known as Quantum Field Theory (QFT) while the latter
is still an open question in Theoretical Physics.
1.1 A brief review on wave mechanics
As previously studied, we already know that matter behaves as waves under certain
circumstances. Following De Broglie’s ideas it is possible to associate a wave function for a
matter particle, as a particle behaviour was associated to wave functions as electromagnetic
fields. In pedestrian words, there are relationships among physical important concepts as
energy and momentum to wave’s quantities as frequency and wave length. The relation
among these quantities is given through the appareance of a universal constant known as
Planck’s constant
~ =, (1.1)
establishing the following relations:
E = ~ω, p = ~κ, (1.2)
where ω is the frequency and κ is the vector number of a wave function φ. Notice that
the wave equation
∂2φ 2 ∂φ
2
= v , (1.3)
∂t2 ∂x2
1
2 Chapter 1 Introduction
has a a generic particular solution φ(x, t) = Aei(±kx+ωt) , with k the wave vector indicating
the direction of propagation. The solution gives us the expression v 2 = ω 2 /k 2 which is
the squared magnitude of the phase group velocity.
By substituting 1.2 into the above expressions, we get that E 2 = p2 v 2 implying that
the energy of a particle associated to a wave depends on its velocity. For electromagnetic
fields with v = c one gets that E 2 = p2 c2 which indeed is the energy for relativistic
massless particles. This is somehow expected since time and space are equally inserted in
the wave equation.
Let’s repeat the same experiment in the other way around and start with the non-
relativistic expression for energy E = p2 /2m this time referring to the kinetic energy
associated to a massive particle. Again by substituting relations (1.2) one gets
~2 k 2
~ω = . (1.4)
2m
Let us then assume that the wave function φ related to a massive particle goes also like
eikx (why?), hence we get that
~2 ∂ 2 φ
− = Eφ, (1.5)
2m ∂x2
which is the Schrödinger equation for a stationary free (i.e. non-time dependent and no
forces acting on it) wave function.
Different situations and examples establish a relation between the momentum of the
particle and a spatial derivative of the wave function. According to the above mentioned
one can say that
ψ(x) = Aei~xp (1.6)
for which given a fixed momentum p one gets that
∂ψ
− i~ = pψ. (1.7)
∂x
Although momentum implies movement on spatial directions and a relation between
momentum and a spatial partial derivative could be expected, it is not clear which
assumption is more fundamental> a wave function behaiving like ei~px or computing
momentum through a spatial derivative of the wave function. We shall formally derive
this result.
Notice as well that a free particle with an associated wave function ψ(x) = ei~px has a
magnitude equal to unity for all space. That means that it is equally probable, according
to the interpretation of the modulus of ψ as the probability to find the particle described
by such wave function, to find the particle in every point in space. This is non-sense, at
least classically. In other words, what we are saying is that if we know the precise value
of p, we cannot say where the particle is and viceversa. One way to face the problem is
1.2 Bounded states 3
to consider a bunch of particles with different momentum associated. In the language of

wave particles, that means having a linear combination or superposition of different wave
functions with different associated momentum, i.e., consider the wave packet
Z
ψ(x) = dpg(p)ei~px , (1.8)
which can be verified it is still a solution for the wave equation and satisfies the De
Broglie’s relations. The function g(p) establishes a different profile for each component of
the wave packet, or in other words, it fixed that a bunch of particles has a wide values of
momentum fixed by the weight function g(p). The momentum associated to the wave
packet can be easily computed in terms of superposition of different waves, this is, the
momentum of the wave particle represented by P [Ψ] is given by
Z
P [Ψ] = dp p g(p)ei~px , (1.9)
meaning thet the wave packet has not a definite momentum value but a wide range
according to the function g(p). This implies that we can also have a finite range of
possible values for locating the wave-packet around some point in space since
Z
2
|Ψ| = dp |g(p)|2 , (1.10)
which clearly is different than 1. However we can normalize the wave function Ψ(x) such
that Z
dx|Ψ(x)|2 = 1, (1.11)
indicating that the probability to describe the particle by the wave function Ψ is one, but
depends on every point in space in which the probability density varies from point to
point. It seems that we should to restrict to quadratic integrable wave functions. This is
a key concept and we shall come back to it later on.
1.2 Bounded states
Another interesting issue in quantum mechanics, comes by discretization of certain

physical quantities. We have some iconically examples as the discrete atomic spectrum,
the photoelectric effect, black body radiation and Compton effect. In all of these cases,
physical quantities as the energy and momentum are transferred into different particles
by discrete quantities. However, this does not happen for the free particle as we have
seen, since its energy or momentum have continuos values. What does determine the
discrete nature of certain physical quantities that we can actually measure? It seems that
when a particle is under the influence of a bounded potential, its energy and momentum
will be discretized, as in the well potentials, harmonic oscillators and the hydrogen atom.
See exercises I for some representative examples.
4 Chapter 1 Introduction
1.3 Into a mathematical framework for quantum mechanics
The above facts lead us to realize that many physical assumptions are in fact consequences
of a deeper mathematical structure which seems to be fundamental at the quantum level.
Surprisingly we shall see that linear algebra of finite and infinite dimensional spaces are
the correct mathematical structure to represent all physical states. For that one must be
very careful to understand wether we are dealing with an abstract object and wether we
can actually deduce something we can measure in an experiment.
Until now we have some experience relating wave functions to physical systems as
free particles or particles under the action of some force in terms of its potential function
and we infer some properties the wave function must have. Based on these experience we
are ready now to properly establish a formal reference frame in which all physical states
can be represented in a quantum theory.
1.3.1 Physical States

Let us then think on a physical state as a vector state in an abstract vectorial linear
space (of course linearity follows from physical superposition). Every physical state is
represented by this vector denoted |ψi. Let us consider for the moment a finite dimensional
vector space for which the vector state |ψi has n components, this means that in terms
of some basis, we can describe the state |ψi as a vector column with n components, as

ψ1
 ψ2 
|ψi −→  .  . (1.12)
 
 .. 
ψn
Notice that we are calling vectors as in the case of elements in a vector space, i.e., in
a mathematical sense. For a vector state to be actually a vector in the physical space,
we require that under rotations, the state transforms accordingly. In the same sense we
relate a dual state vector hψ| to the row vector
hψ| −→ (ψ1∗ , ψ2∗ , . . . ψn∗ ) (1.13)
such that
hψ| · |ψi = hψ|ψi = |ψ|2 = ψ12 + ψ22 + · · · + ψn2 , (1.14)
following the matrix order multiplication. Notice that the components of |ψi are in
general complex functions and that they correspond to components on a certain given
basis in the corresponding vector space. We should be able to know which basis we must
select according to the physical system under study.
1.4 Exercises, due to T2-day

Chapter 2
Fundamental concepts I
2.1 Physical states
A quantum particle or object is described in terms of a wave function Ψ(x) or equivalently

by a vector state in an abstract vectorial space we shall denote H. This is
quantum particle −→ Ψ(x) −→ |Ψi . (2.1)
As previously mentioned, the physical state |Ψi can be represented by a vector column in
an appropriate basis of H. Since |Ψi ∈ H, we have that
1. There exist the summation operation + such that for |i ∈ H
• (c + d) |ai + (c + d) |bi = (c + d)(|ai + |bi) ∈ H, for c, d ∈ C.

• There exists a unique element |∗i = 0 such that |ai + 0 = |ai for all |ai ∈ H.
• For each |ai ∈ H there exists a unique − |ai such that |ai + (− |ai) = 0.
This defines a vector space. Our departing point is to identify each physical state
of a given particle or system with some vector state in this space. As it is known, each
vector state can be also described in terms of a basis. There is an infinity number of
different basis for a vector space of finite dimension. Observe we are using the notions of
dimension in a physical meaning. Therefore,
Definition 1. A basis {|an i} of H is a set of vectors, with n = 1, . . . , N such that all

vectors in H are written as
XN
|ai = λn |an i ∈ H. (2.2)
n=1
If the above vector vanishes, then λn ∈ C vanishes for all n.
Notice this defines the basis for a vector space of dimension N .
5
6 Chapter 2 Fundamental concepts I
2.1.1 Dual vector space H∗

Each vector space H has an associated dual space denoted H∗ . It is constructed as follows.
Consider a mapping ha| : H −→ C such that
ha| (|bi) = λ ∈ C. (2.3)
We call these mappings functionals of H. If these functionals have the following

properties:
1. (ha| + hb|)(|vi) = ha| (|vi) + hb| (|vi),
2. hca| (|dv + wi) = c∗ d ha| (|vi) + c∗ ha| (|wi),
with c, d ∈ C, we define the set of linear functionals
{ha| | ha| : H → C}, (2.4)
as the dual space H∗ of H.
The dual space H∗ has the same dimension as H and there is a duality between the
two spaces given by an orthonormal basis on H. Orthonormality is a special condition
on the internal product defined on a vector space. In our case it is assumed that the
vector space containing physical states has an internal product, this is H is a vector space
with an internal product. The internal product is defined through the set of functionals,
and here becomes clear the usefulness of Dirac’s equations by kets and bra’s, since the
internal product is defined by the bilinear mapping
· : H∗ × H → C, (2.5)
this is
|ai · |bi = ha|bi = ha| (|bi). (2.6)
Two vectors are said to be orthogonal if ha|bi = 0. Let {|aj i} be an orthonormal basis of
H, i.e., hai |aj i = δij . Since H∗ is also a vector space, then there is also a basis for H∗ .
Let {haj |} be such basis. Since (H∗ )∗ = H we then have that
ha| · hb| = hb| (|ai) = hb|ai . (2.7)
Notice that the order is important since linearity on bra-states are related to the ket-states
by the complex conjugate. Specifically
ha|bi = hb|ai∗ . (2.8)
Given a basis {|aj i} in H, we define the dual basis by the set {haj |} such that hai |aj i = δij .
Notice that the dual basis is defined by the internal product.
2.1 Physical states 7
P
Let us now consider the vector state |ai = j λj |aj i. In terms of this basis we can
see that it is possible to represent the vector state by the column vector
 
λ1
 λ2 
|ai −→  .  . (2.9)
 
 .. 
λN
P
Similarly if hb| = j σj hbj |, we can represent the bra-state as a row-vector
hb| = (σ1 , σ2 , . . . , σN ) (2.10)
such that
N
X
hb|ai = σj λj . (2.11)
j=1
Notice that for a given |ai there is a unique ha| with ha| = (λ∗1 , . . . , λ∗N ) in the same basis,
such that
N
X
| |ai |2 = ha|ai = |λj |2 . (2.12)
j=1
It is important to notice that linearity on bra-states implies that if hb| = hλa| i.e.,
hb| = (λσ1 , . . . , λσN ) in a given basis, then hb| = λ∗ ha| with ha| = (σ1 , . . . , σN ).
Consider now the state |ai = N

P
j=1 λj |aj i. Given that |aj i is an orthonormal basis,
we can apply on both sides an internal product with the related bra haj | such that
N
X N
X
haj |ai = λi haj |ai i = λi δij = λj , (2.13)
i=i i=1
Pn
from which one gets λj = haj |ai. Similarly if hb| = j=1 σj haj | then σj = hb|aj i.
2.1.2 Clousure relation

The use of Dirac notation pays off in the following identity. Take a general state |ai and
write it in terms of a basis {|aj i} then,
N
X
|ai = λj |aj i ,
j=1
N
X
= haj |ai |aj i ,
i=1
XN
= ( |aj i haj |) |ai . (2.14)
i=1
We notice that the identity operator Iˆ defined as Iˆ |ai = |ai can be written in terms of
the basis {|aj i} as
XN
Iˆ = |aj i haj | . (2.15)
i=1
It is very important to notice that the internal product hb|ai is a complex number, while
the expression |ai hb| represents an object acting on ket states and giving as a results a,
in general, different state:
|ai hb| (|ci) = hb|ci |ai = λ |ai , (2.16)
with λ = hb|ci ∈ C, this is |ai hb| : H → H, which is also a linear map.
2.1.3 Operators
Objects acting on ket states as |ai hb| represent a very important set of entities called
operators. We define an operator on the vector space H as a map
Â : H −→ H, (2.17)
such that
Â |ai = |bi ∈ H. (2.18)
Operators satisfy the following properties:
1. (Â + B̂) |ai = Â |ai + B̂ |ai,
2. Â(c |ai + |bi) = cÂ |ai + Â |bi,
Using these properties we can obtain a representation of an operator Â in terms of a given

basis. Consider the orthonormal basis {|aj i} and the action of Â on a single element of
the basis,
Â |aj i = IÂ̂ |aj i ,

N
!
X
= |ai i hai | Â |aj i ,
i=1
N
X
= hai | Â |aj i |ai i ,
i=1
N
X
= Aij |ai i , (2.19)
i=1
where we have defined the matrix element of Â with respect to (wrt) the basis {|aj i}
as Aij = hai | Â |aj i. Notice that in terms of column vectors, the operator has a matrix
2.1 Physical states 9
representation:
  
A11 A12 · · · A1N 0
 A21 A22 · · · A2N  0
 
 
 ..   .. 
 .  .
  
  0
Â |ai i −→   . (2.20)

 Aij  1
 
  0
  
 .. ..   .. 
 . .  .
AN 1 AN 2 · · · AN N 0
As we have mentioned before, an operator can also be written in terms of a ket-bra state.
This follows since
Â = IÂ̂Iˆ
!  
X X
= |ai i hai | Â  |aj i haj | ,
i j
X
= |ai i hai | Â |aj i haj | ,
i,j
X
= Aij |ai i haj | . (2.21)
i,j
2.1.4 Adjoint operators
As we have seen, given a ket state |ai we can relate a bra state ha|. The question here is
if such ket state is the result of an operator acting on a different state, which operator is
acting on the bra state to get the same final state? This is, if |ai = Â |bi, is it possible
to write the corresponding bra state as ha| = hb| B̂? in such a case, is there a relation
between Â and B̂?
P
Consider an orthonormal basis {|aj i}, then we can write |ai = i λi |ai i, from which
it follows
X
ha|ai = λ∗i λi ,
i
hai | Â |bi∗ hai | Â |bi ,

X
=
i
X
= A∗ij Aij | haj |bi |2 . (2.22)
i,j
On the other hand, by assuming that ha| = hb| B̂, one gets that
ha|ai = hb| B̂ Â |bi ,

X
= hb| B̂ |ai i hai | Â |bi , (2.23)
i
X
= Bji Aij | haj |bi |2 . (2.24)
i,j
By comparing the above two expressions we get that B̂ is actually Â† (complex conjugate
transpose).
Chapter 3
Fundamental Concepts II
We continue with a formal description of operators on the vector space H and their
identification with physical quantities.
3.1 Unitary transformations
There are some special type of operators which are going to be very useful in quantum
mechanics, since they shall be related to symmetries.
Let us assume that {|ai i} and {|bi i} be 2 different orthonormal basis1 such that for
an operator Û ,
Û |ai i = |bi i . (3.1)
Hence, we can say that hbi | = hai | Û † , from which we can write an expression for any
unitary operator in terms of the vectors |ai i and |bi i as follows
Û = Û Iˆ !
X
= Û hai |ai i
i
X
= |bi i hai | . (3.2)
i
Similarly,
n
X
Û † = |ai i hbi | . (3.3)
i=1
1
In Homework 2, you should show that if {|ai i} is an orthornormal basis, then Û |ai i = |bi i is also
an orthonormal basis.
11
12 Chapter 3 Fundamental Concepts II
It follows that
! 
X X
Û Û † = |bi i hai |  |aj i hbj |
i j
X
= |bi i δij hbj | ,
i,j
X
= |bi i hbi | ,
i
ˆ
= I. (3.4)
Operators such that Û Û † = Iˆ are called unitary operators. Similarly, as seen, Uni-
tary operators are those which transform an orthonormal basis into another
orthonormal basis.
3.1.1 Example: U (1)-operators

†
Consider the operator Û = eiÂ and Û † = e−iÂ . In general we have that2
† 1 † ]+···
Û Û † = ei(Â−Â ) e 2 [Â,Â . (3.5)
ˆ with
Let us take the simplest case, i.e., Â = Â† which is accomplished by taking Â = αI,
iα
α ∈ C, for which one has that Û is unitary. Unitary operators of the form Û = e , can
be written for small values of α as
Û ∼ Iˆ + iαI,
ˆ (3.6)
and
Û |ai ∼ |ai + iα |ai , (3.7)
meaning that the state |ai changes under the action of Û to the state (1 + iα) |ai. In other
words, the state |ai changes under the action of an infinitesimal transformation generated
by α by δ |ai = iα |ai. Under finite transformations it is clear that |ai → eiα |ai from
which we see that the state |ai changes by a phase. Unitary transformation produces a
transformation by a phase, from which | |ai |2 keeps invariant. This is a symmetry. Any
transformation which keeps invariant the modulus of |ai is called a symmetry. When α
does not depend on space or time coordinates, we call it a global transformation. If α is
a function on spatial or time coordinates, the transformation is referead as local.
3.2 Change of basis
Consider two different orthonormal basis {|ai i} and {|bi i} such that a state |ψi
X X
|ψi = ci |ai i = di |bi i , (3.8)
i j
2
Homework 2.
3.3 Observables 13
By establishing the change of basis we can obtain a relation between coefficients ci and
di by the use of an unitary operator (since such operators relate orthonormal basis)
X
|ψi = di |bi i ,
i
X
= di IÛ̂ |ai i ,
i
X
= Uij di |aj i ,
i,j
X
= cj |aj i , (3.9)
j
where3 X
cj = Uij di , (3.10)
i
this is c = Ud where c and d are the vector representations of states |ci and |di and U
is the matrix representation of Û in the {|ai i}-basis.
A similar analysis can be done for a matrix representation of Â. Let Aij = hai | Â |aj i
eij = hbi | Â |bj i. Then,
and A
X
Â = |ai i Aij haj | ,
i,j
X
= eij haj | Û † ,
Û |ai i A
ij
 
eij U †  ham | ,
X X
= |ak i  Uki A jm (3.11)
k,m i,j
from which we find that

eij U †
X
Akm = Uki A jm (3.12)
i,j
e †.
or in terms of matrix representations, we have that A = UAU
3.3 Observables
The notion of an observable is one of the most important concepts in the formulation of
quantum mechanics. It establishes a link between operators and those quantities we can
actually measure. So, at this point we have to think about the meaning of measuring in
the context of the mathematical language we have developed till this point. Hence, by
measuring some physical quantity a of a particle, we mean applying an operator Â acting
on the physical state |ψi representing the particle in the vector space H such that
Â |ψi = a |ψi . (3.13)
3
Clearly justify each step in the above calculation.
14 Chapter 3 Fundamental Concepts II
Notice that the operator does not change the sates is acting on. This property is very
important and those states which are not changed under the action of an operator are
called proper vectors or eigenvectors of Â with a proper value or eigenvalue a.
The number a is in general a complex number, although we shall see that for an observable
we shall restrict it to be real.
3.3.1 Eigenvalues and eigenvectors

For a given operator Â with an orthonormal basis {|ai i} we have that if |ψi is an eigenstate
of Â,
hai | Â |ψi = a hai |ψi ,

X
= Aij haj |ψi , (3.14)
j
we see that since hai |ψi = ci are the coefficients on the expansion of |ψi in the corre-
sponding basis, then
X
aci = Aij cj , (3.15)
j
or X
(Aij − aδij ) cj = 0. (3.16)
j
from which the eigenvalues a can be obtained by solving the characteristic equation
det (A − aI) = 0, (3.17)
and then, one can find the corresponding eigenvectors.
3.4 Hermitian observables
Consider the special case in which Â = Â† . We call these operators Hermitian opera-
tors. Two important results follows from the very definition of an Hermitian operator:
1. The eigenvalues of a Hermitian operator are real. The proof is left to the
student.
2. Two eigenvectors of a Hermitian operator with two different eigenvalues,

are orthogonal. Let us consider a Hermitian operator Â such that Â |ai = a |ai
and Â |bi = b |bi. Since
hb| Â |ai = a hb|ai = b hb|ai , (3.18)
it follows that (a − b) hb|ai = 0. For a 6= b we have that hb|ai = 0.

3.5 Observables 15
In the case in which two different eigenvectors have the same eigenvalue wrt
aHermitian
operator, we say there is degeneracy, i.e., more than one eigenvectors aj (observe
the index i is upstairs) with the same eigenvalue a. In order to measure the degeneracy,
E
let us consider the following. Let us say that we have a set of eigenvectors {aji } (not

necessarily a basis) of Â such that for a given i there are gi different eigenvectors with
the same eigenvalue ai , this is E E
Â aji = ai aji , (3.19)

D E
with j = 1, . . . , gi . Since Â is hermitian, it follows that aji aki = δ jk , and in consequence

ham n
i |ak i = δ
mn δ . The closure relation is now extended to
ij
gi E D
N X
j
aji .
X
Iˆ = ai (3.20)

i=1 j=1
Hence,
E any eigenstate of a Hermitian operator can be expressed in a basis formed by
{aji }. Such operator is then called an observable.

3.5 Observables
There are 3 important properties that observables satisfy. Let Â and B̂ be two commutable
operators, i.e., [Â, B̂] = 0. It follows that

1. If Â |ψi = a |ψi, then Â B̂ |ψi = aB̂ |ψi. Notice that this means that every
subspace of Â is globally invariant under the action of B̂.
2. If Â |ai = a |ai and B̂ |bi = b |bi with a 6= b, then ha| B̂ |bi = 0.
3. There exists an orthonormal basis of H whose elements are common eigenvectors of

Â and B̂.
The proof of these statements are left to the student. Finally, we say that a set of
observables is complete if none of the eigenvalues is degenerate. In practice that opens
up the possibility to measure new degrees of freedom.
Chapter 4
Observables
4.1 Compatible observables
Think about having two different kets. Before going on, let us think about the meaning of
the word "different". How can we distinguish between two kets? Certainly by measuring
"something" which make them be different. Yes, the argument is very weak so far. This
forces us to concretely define the meaning of measuring. So, we are going to relate a
physical state of a particle to a vector ket in the vector space H. Consider then an
eigenket |ai of an observable Â such that Â |ai = a |ai. Consider also another ket |bi
which is also an eigenket of Â with the same eigenvalue a, this is Â |bi = a |bi. How can
be sure that that these two states are really different? We cannot if we only characterize
the ket by its proper value. We need another observable which can make the difference,
i.e., that still is an eigenvector of Â and B̂ but with different eigenvalues wrt B̂. This is
B̂ |ai = b |ai ,
B̂ |bi = b0 |bi . (4.1)
Therefore, we require two observables to differentiate our kets, which also must be
eigenkets of both observables, i.e., that [Â, B̂] = 0. Let us generalize this idea and take g
kets which are degenerate wrt to an observable Â1 , with a degree of degeneracy equal to
g, i.e., Â1 |ai i = a1 |ai i for all i = 1, . . . g. Let us say that there is another observable Â2
which can differentiate only one of those g kets by different proper values, i.e.,
Â2 |ai i = a02 |ai i ,
Â2 |aj i = a2 |aj i , (4.2)
for i = 6 j and for some value j. Let us assume that there is another observable Â3 , with
[Â1 , Â3 ] = [Â2 , Â3 ] = 0, such that only one of the g − 1 degenerated states has a different
eigenvalue wrt Â3 , this is
Â3 |ai i = a03 |ai i ,
Â3 |ak i = a3 |ak i , (4.3)
with i 6= k but not necessarily different than j. So far we have broken 2 degeneracies by
considering two extra commutable observables. Continuing under this approach, we can
16
4.2 Measurement and Uncertainty 17
consider g different observables, commuting to each other such that all the eigenkets have
different set of eigenvalues wrt to all g observables. The set of observables which breaks
the degeneracy is called a complete set of compatible observables.
Exercise 1. Show that if a basis {|ai i} are eigenvectors of two compatible observables
Â and B̂, then their matrix representation in such a base are diagonal.
4.2 Measurement and Uncertainty
Consider a physical state P|Ψi. In terms of a basis of eigenvectors {|ai i} of an observable Â,
we can write that |Ψi = i λi |ai i, with Â |ai i = ai |ai i. Clearly the matrix elements of
Â are given by Aij = ai δij . Each of these elements are the values obtained by measuring
the quantity A related to the observable Â if the state |Ψi was given only by the eigenket
|ai i. Eigenvalues are those quantities we can measure. In other words, if |Ψi = |ai i the
measurement of the quantity A is given only by the value ai . The measurement would
be precise and non uncertainty is expected. The situation of course is a little bit more
complicated once we add the rest of possible states, represented by the rest of eigenkets.
Following this idea, the value of the physical quantity A related to the state Ψ is
given then by hΨ| Â |Ψi, which is
X
< Â >= hΨ| Â |Ψi = |λi |2 ai . (4.4)
i
Notice this is an average value weighted by the numbers |λi |2 . were we have taken a
normalized state, this is hΨ|Ψi = 1. Since there are many different values ai we can
measure a natural uncertainty is expected. We define this uncertainty ∆A as
(∆A)2 =< Â2 > − < Â >2 , (4.5)
for a given state |Ψi. In our case we have then that

n
! n
!2
X X
2 2 2 2
(∆A) = |λi | ai − |λi | ai , (4.6)
i=1 i=1
which clearly vanishes for n = 1 recovering the above mentioned result. Observe than
once we have performed a measurement and found that the obtained values is given
by a specific ai , it means that the phsyical state |Ψi is described only by the eigenket
|ai i (we shall call this step as the collapse of the wave function Ψ(x)). Therefore, after
one measurement the uncertainty of getting the value ai in subsequents measurements
vanishes. The uncertainty of getting some precise value on a physical state given by just
one eigenket of the corresponding observable is always zero.
Let us now move to the interesting question of performing two measurements related
to non-commutable observables Â and B̂ and let us start by expressing the |Ψi in terms
18 Chapter 4 Observables
P
of eigenvectors of the observable Â, i.e., |Ψi = λi |ai i. Define the operator
δ Â = Â− < Â >, (4.7)
with a similar expression for B̂. Notice that
< (δ Â)2 >=< Â2 > − < Â >2 = (∆A)2 . (4.8)
Consider now the action of an operator δ Â and δ B̂ on a ket state |Ψi. By Schwarz
inequality,
| hΨ| δ Âδ B̂ |Ψi |2 ≤ hΨ| (δ Â)2 |Ψi hΨ| (δ B̂)2 |Ψi . (4.9)
The left hand side of this equation can be rewritten by using that
1
δ Âδ B̂ = [δ Â, δ B̂] + {δ Â, δ B̂} , (4.10)
2
Since [Â, B̂]† = −[Â, B̂] the commutator is a anti-Hermitian operator, which expectation
value is always a pure imaginary number, while the anti-commutator is always a real
number.
Therefore, we can write

1
< (δ Â)2 >< (δ B̂)2 > ≥ | < [δ Â, δ B̂] > |2 + | < {δ Â, δ B̂} > |2 , (4.11)
4
and clearly conclude that
1
∆A∆B ≥ | < [Â, B̂] > |. (4.12)
2
Exercise 2. Show that if Ĉ is an anti-Hermitian operator, then its expectation value wrt
any state is a pure imaginary number. Prove also that for a Hermitian operator, its
expectation value is a real number.
Exercise 3. Prove that [δ Â, δ B̂] = [Â, B̂] and explain in detail why we have not consider
the other possible inequality
1
(δA)2 (δB)2 ≥ | < {δ Â, δ B̂} > |2 . (4.13)
4
4.3 Space and Momentum
As we have seen, an observable is the abstraction of performing a measurment on some

physical state. The actual measured value correspond to the eigenvalue of the physical
state wrt to the observable. Up to now we have studied discrete eigenvalues corresponding
to some finite set of discrete eigenkets which in turn generates a finite dimensional vector
space. A question arises about the existence of continuous values for some physical
quantity we could be interested on measuring. Position and momentum are some of
4.3 Space and Momentum 19
them, since the possible values, at least for a free particle, seem to vary continuously.
We deine such basis as the eigen states of the observables of position X̂ and momentum P̂ as
X̂ |xi = x |xi , P̂ |pi = |pi . (4.14)

Therefore, a physical state given by the ketstate |Ψi can be written in terms of this
basis as Z
|Ψi = dx Ψ(x) |xi , (4.15)
where Ψ(x) are the function representing the continuous set of coefficients in the linear
infinitesimal combination. From this assertion it follows that Ψ(x) = hx|Ψi and since
Z
|Ψi = dx hx|Ψi |xi ,
Z
= dx |xi hx| |Ψi (4.16)
it follows that Z
Iˆ = dx |xi hx| . (4.17)
Inserting this relation into the expression hx|Ψi one gets that
hx|Ψi = hx| Iˆ |Ψi ,

Z
dx0 xx0 x0 Ψ ,

= (4.18)
from which we obtain the orthonormal condition for a continuous basis,

0
xx = δ(x − x0 ). (4.19)
A similar analysis can be done in terms of another continuous parameter as momentum.

A ket state can also be written in terms of a basis on momentum-space with similar
expressions as for those in configuration space, this is:
Z
|Ψi = dp φ(p) |pi , (4.20)
hp|Ψi = φ(p), (4.21)
Z
Iˆ = dp |pi hp| , (4.22)

0
pp = δ(p − p0 ). (4.23)
Hence, the projection of ket states on the space or momentum basis, represents the wave
function we have refered to at the beginning of these lectures. Notice that by considering
a continuous value for the space and momentum coordinates, we are implicitly studying a
free particle. We shall see that our intuiton is indeed correct.
We now turn into a much more interesting question. What is the projection of the
space basis |xi into the momentum basis |pi? For that, let us do the following,
δ(x − x0 ) = xx0 ,

Z
dp hx|pi px0 .

= (4.24)
Let us use a common expression for Dirac’s delta function given by

Z
0 1 0
δ(x − x ) = eiω(x−x ) dω. (4.25)
2π
Comparing both expressions, and by taking ω = p/~, one gets that
1
hx|pi = √ eipx/~ , (4.26)
2π~
which is actually the expression for a plane wave representing a free particle wave function!
Observe that
1
hp|xi = √ e−ipx/~ , (4.27)
2π~
according to hx|pi = hp|xi∗ . Therefore,
Z
Ψ(x) = dp hx|pi Φ(p),
Z
1
= √ dp eipx/~ Φ(p). (4.28)
2π~
which is actually the wave packet formed by plane waves with different momentum
weighted by the function Φ(p). Similarly
Z
1
Φ(p) = √ dx e−ipx/~ Ψ(x), (4.29)
2π~
establishing that Ψ(x) and Φ(p) are Fourier transforms to each other.
4.3.1 Measuring space and momentum

We are interesting in the ket vector X̂ |Ψi and its projection either on the state-space or
the momentum-space. Notice that |Ψi is not an eigenvector of X̂. Hence,
hx| X̂ |Ψi = xΨ(x), and hp| P̂ |Ψi = pΦ(p). (4.30)
The question now is how the action of the momentum operator acts on a ket state in the
ocnfiguration space. This is hx| P̂ |Ψi. Then,
Z
dp0 xp0 p0 P̂ |Ψi ,

hx| P̂ |Ψi =
Z
1 0
dp0 p0 eip x/~ p0 Ψ ,

= √
2π~ Z
1 ~ d ip0 x/~
0
= √ dp0 e p Ψ ,
2π~ i dx
d
= −i~ hx|Ψi . (4.31)
dx
4.3 Space and Momentum 21
This means that the action of momentum operator on a ket state |Ψi is represented in the
space of functions as a derivative on space coordinates. Notice that this also implies that
d
− i~ Ψ(x) = pΨ(x), (4.32)
dx
which precisely has as a solution the plane wave Ψ(x) = Aeipx/~ . In terms of energy
E = p2 /2m + V (X), one could naively say that
~2 d2
EΨ(x) = − Ψ(x) + V (x)Ψ(x), (4.33)
2m dx2
which as we shall see and you know, is the famous Schrödinger equation.
Exercise 4. Show that hp| X̂ |Ψi = i~ hp|Ψi.
4.3.2 Uncertainty on measuring position and momentum

Our next step is to compute the projection on coordinate space of the action of the
operator [X̂, P̂ ] on a state Ψ. Therefore,
hx| [X̂, P̂ ] |Ψi = hx| X̂ P̂ |Ψi − hx| P̂ X̂ |Ψi ,

= x hx| P̂ |Ψi − hx| P̂ X̂ |Ψi ,
d d
= −i~x hx|Ψi + i~ hx| X̂ |Ψi ,
dx dx
= i~ hx|Ψi . (4.34)
Hence we conclude that

ˆ
[X̂, P̂ ] = ihI. (4.35)
Position and momentum operators do not commute. What are the physical implications
of that? Notice that it is not possible to have a common basis of eigenvectors for both
operators, implying that an uncertainty for measuring both observables would be different
from zero, as expected from the wave packets constructed at the beginning. Therefore,
let us compute such uncertainty.
According to the uncertainty relation (4.12), we have
~
∆X ∆P ≥ . (4.36)
2
For a free particle, defining a momentum for the plane wave implies a complete uncertainty
about the localization of the particle in the entire space and viceversa. For a wave packet
the uncertainty in position is now limited to some finite region in space but the uncertainty
in momentum has also become greater than zero. Notice that momentum cannot
have a discrete spectrum.
Exercise 5. Compute [X̂, Pˆ2 ]. What does it mean?
Exercise 6. Take the kinetic energy operator Êk defined as P̂ 2 /2m. Compute ∆Ek ∆X.
What does it imply?
~ˆ × P~ˆ . Show that

Exercise 7. Take the angular momentum operator as L̂ = X
~ˆ 2 ] = 0.
1. [Ĥ, L
2. [Ĥ, L̂i ] = 0, ∀i,
3. [L̂i , L̂j ] = iijk L̂k ,

ˆ Why?
where we have used that [X̂i , X̂j ] = [P̂i , P̂j ] = 0 for all i, j and [X̂i , P̂j ] = i~δij I.
Also use that V (X̂) is a function solely of ||X̂|| (consider that X̂ is a vector operator).
4.4 Suggested exercises
1. Consider a Gaussian wave packet. Show the uncertainty in position and momentum
is minimal.
2. How is the matrix element of Lz in terms of eigenvectors of Lz . Give a general

expression in terms of its proper values.
3. Compute the matrix representation of L2 , Lx and Ly wrt the basis of eigenvectors

of Lz .
Chapter 5
Quantum Dynamics
In this lecture we shall see how a state or an operator evolves in time. This will ensure
that quantum mechanics is indeed deterministic when it comes about probabilities, since
a physical state evolution in time is determined by a differential equation known as
Schrödinger equation. Equivalentely we can describe a time evolution on operator, a
point of view known as Heisenberg picture.
5.1 Postulates
Since the fundamental origins of what we have observed are the dynamical laws of quantum
mechanics, are not known, we start by enunciating some fundamental postulates. They
are the following:
1. A physical state related to some particle or system is described by a ket state |Ψi
in a vector space H of finite or infinite dimension.
2. A physical measurable quantity A related to some physical system is represented

by a Hermitian operator Â : H → H.
3. Let |Ψi a state expressed in terms of eigenvectors of Â where degeneracy is assumed.

Then
n Xgi E
λi aji .
X
|Ψi =

i=1 j=1
E the value ai and therefore the probability that the

The probability to measure
system be in the state aji is given by

gi D E
| aji Ψ |2 .
X
P(ai ) = (5.1)

j=1
Notice that X
P(ai ) = 1.
i
23
24 Chapter 5 Quantum Dynamics
4. If before measurement the state is given by a linear combination of all possible

eigenstates of an observable Â, after measuring the physical value ai the state is said
to collapse to |ai i. The collapase of the state or wave function is mathematically
described in terms of a projector operator P̂i given by
gi E D
j
aji ,
X
P̂i = ai (5.2)

j=1
where
P̂i2 = P̂i ,
which is the proper definition of a projector operator. Therefore, the collapse of the
state is then
P̂i |Ψi
|Ψi −→ .
||P̂i |Ψi ||2
5.2 Energy states and time evolution.
OPEN QUESTION: Think about the following. If distance and momentum

cannot be simultaneously measured, how can the energy P̂ 2 /2M + V (X̂) be
measured with precision?
Consider a ket state |Ψi at a given time t0 . We are interesting in describing its
evolution through time, this is, we want to know how the system evolves to another time
t > t0 . We would like to find an operator that does the job, this is
|Ψ(t)i = Û (t, t0 ) |Ψ(t0 )i . (5.3)
Let us start by expanding the ket state in time1 ,
d
|Ψ(t)i = |Ψ(t0 )i + |Ψ(t0 )i ∆t + . . .
dt
(5.4)
On the other hand, following expression 5.3, we have that
hΨ(t)|Ψ(t)i = hΨ(t0 )|Ψ(t0 )i = 1, (5.5)
since we are not assuming an interaction with another particle. Therefore one concludes
ˆ indicating that the time evolution operator is unitary. This is expected
that Û † Û = I,
since there are not interactions with another systems, and therefore the state changes
only through a phase, which precisely represents the action of a unitary operator. Let us
describe it in more detail: since Û is unitary can be represented by an element of the
1
Notice that time and space are different objects in this description, since space coordinates can
be measured through the space operator X̂ while time is just a continuous parameter. This makes our
description non-relativistic. Notice also teh assumption about continuity and density in space and time
coordinates.
5.2 Energy states and time evolution. 25
group U (1), this is U (t,ˆ t0 ) = eiβ Ĥ∆t .
Hence we can write

|Ψ(t)i = Û (t, t0 ) |Ψ(t0 )i = eiβ∆tĤ |Ψ(t0 )i = Iˆ + iβ∆tĤ . . . |Ψ(t0 )i , (5.6)
from which we conclude that

d
iβ Ĥ |Ψ(t0 )i = |Ψ(t0 )i . (5.7)
dt
We need to elucidate the nature of operator Ĥ. For that, consider the following. We
know from the second Newton’s law that
dp d
= − V (x),
dt dx
which we can use and extend it to the momentum associated to a wave function hx|Ψi by
d d
hx| P̂ |Ψi = − hx| V (X̂) |Ψi . (5.8)
dt dx
Hence, since
d d d
hx| P̂ |Ψi = −i~ hx|Ψi ,
dt dtdx
d d
= − i~ hx|Ψi ,
dx dt
d
= − (V (X̂) + F(P̂ )) hx|Ψi , (5.9)
dx
where in the last equality we have used second Newton’s law. F(P̂ ) represents a function
on momentum operators P̂ with energy units. Therefore we can safely assume that
F(P̂ ) = P̂ 2 /2m. Then,
d
i~ |Ψi = Ĥ |Ψi , (5.10)
dt
where Ĥ is the Hamiltonian, and we have fixed β = −i/~. In this form we have find that
the Hamiltonian is the generator of time evolution from t0 to t and we can write
Û (t, t0 ) |(t0 )i = |Ψ(t)i = e−iĤ∆t/~ |Ψ(t0 )i . (5.11)
Taking Ψ(t0 ) as an energy state, i.e.,
Ĥ |Ψ(t0 )i = E |Ψ(t0 )i
we find that Ψ(x, t) = eiE∆t/~ Ψ(x).
Suggested Exercise. Explain why we can have energy states

Notice also that the time-dependent wave function for a free particle given by
Ψ(x, t) = ei(Et−px)/~) ,
has precisely the relation among Energy, time, momentum and space as the solution
of a classical wave in electromagnetic theory. But most important, see that this also
corresponds to the relativistic quantity pµ xν . Relativity is encoded in electromagnetic
and wave theory. So if we promote particles to have a wave behaviour, what the meaning
of saying that we are developing a non-relativistic version of quantum mechanics?
5.3 Momentum as generator of translations
Evolution in time generated by the action of an hermitian operator, in this case the Hamil-
tonian, suggests us the possibility to realize any change on some continuous parameter by
the action of an operator in tun generated by some observable. That’s the case of spatial
translations, which as we shall see, are generated by momentum P̂ .
Consider the wave function Ψ(x0 ) which describes the wave front around a coordinate
x0 . If our reference system is moved wrt x0 or if we, as observers mve wrt to the wave
function, we would oike to get an equivalent description given in terms of a wave function
Ψ(x0 + ∆x). Since the spatial coordinates are continuous we can think on an infinitesimal
spatial translation, although we would choose to describe a finite one. Hence, we can say
that
∞
X dn (∆x)n
Ψ(x + ∆x) = n
Ψ(x0 ) . (5.12)
dx n!
n=0
In terms of ket states we then have that
Ψ(x0 + ∆x) = hx0 + ∆x|Ψi , (5.13)
where we are expecting that
K̂(∆x) |x0 i = |x0 + ∆xi , (5.14)
with K̂(∆x) being an operator to be determined. Now, since we now that
d dΨ(x)
−i~ hx|Ψi = −i~ = hx| P̂ |Ψi ,
dx dx
it follows that
∞
∆x n 1
X
hx + ∆x|Ψi = hx| P̂ n |Ψi ,
−i~ n!
n=0
∞
!n !
X iP̂ ∆x 1
= hx| |Ψi ,
~ n!
n=0
iP̂ ∆x/~
= hx| e |Ψi . (5.15)
5.4 Heisenberg Picture 27
Therefore, we define the Translation operator T̂ (∆x) as
T̂ (∆x) |xi = e−iP̂ ∆x/~ |xi = |x + ∆xi , (5.16)

and we say that momentum generates spatial translations.
5.4 Heisenberg Picture
Up to now we have consider the case in which states evolve in time while operators are
kept invariant. However this picture is not unique. We can also consider the opposite: a
set of ket states which are invariant under time but with time dependent operators. This
is called the Heisenberg picture.
Let us fixed the state at t = t0 as |Ψ0 i. In Schrödinger’s picture, |Ψ(t)i = Û (t, t0 ) |Ψ0 i,
and then
< Â > (t) = hΨ(t)| Â |Ψ(t)i = hΨ0 | Û † ÂÛ |Ψ0 i . (5.17)
By defining the time-dependent operator ÂH (t) as
ÂH (t) = Û † ÂÛ , (5.18)
we have an alternative picture in which a physical system can be described in terms

of operator dynamics. Notice that < Â >=< ÂH (t) >. Similarly as states satisfy
Schrödinger’s equation, we can find a differential equation a time-dependent observable
must follows. This is
dÂH dÛ † dÛ

= ÂÛ + Û Â ,
dt dt dt
1
= Ĥ Û † ÂÛ − Û † ÂĤ Û ,
~
i
= Ĥ ÂH − ÂH Ĥ , (5.19)
~
where in the last line we have used that [Û , Ĥ] = 0. Then, the equation that a time-
dependent observable satisfies is
dÂH i
= [Ĥ, ÂH ]. (5.20)
dt ~
Notice that an observable represents a conserved measurable quantity if it com-
mutes with the Hamiltonian in the Heisenberg picture.
5.5 Exercises
1. Prove the Ehrenfest Theorem,
d < P̂ > d
< X̂ >= , < P̂ >= − < ∇V (X̂) >, (5.21)
dt m dt
by using the Heisenberg Picture. Show also that the momentum center of the wave
packet is represented by < P̂ >. Does the center of the wave packet obey the laws of
classical mechanics? Take V (X̂) = λX̂ n and show that < ∇V >6= ∇V (< X̂ >).
2. Show that X
ÂH (t) = Anm ei(En −Em )t/~ |ni hm| , (5.22)
n,m
where |ni is an energy state.
3. Let us say that the Hamiltonian operator Ĥ is given by Ĥ = κÂ. Assume that there
are only two different eigenstates of Â given by |1i and |2i. Estimate the time at
which a change of energy states is expected.
4. In the above system, consider that at initial time, |Ψi = C1 |1i + C2 |2i. Write the
expression for |Ψ(t)i, and compute the probability of measuring the state |1i at time
t = 0 and at t.
5. Consider also in the above system that a third observable is taken into account such
that [Â, B̂] 6= 0. How can you express the probability of measuring the state |1iB ?
where B̂ |iiB = bi |iiB .
~
6. Consider now a vector observable Â, such that the Hamiltonian only commutes
with Âz . Show that the average measurement of the physical quantity Â shows a
precession.
7. Describe in detail the collapse of the wave function by measuring two different
observables which do not commute, and show how an initial measurement of some
of the possible values of one of the observables is altered by measuring the second
one.
Chapter 6
Harmonic Oscillator
The study of the Harmonic Oscillator is crutial in our comprehension of quantum physics.
In this lecture we shall see some important features about it and some rasons about its
significnce.
6.1 Motivitaion
From a classical point of view, every scalar potential on coordinates x can be approximated
around a minimal point x0 as
dV (x0 ) 1 d2 V (x0 )
V (x) = V (x0 ) + (x − x0 ) + (x − x0 )2 + . . . , (6.1)
dx 2 dx2
from which any potential around a critical stable point can be approximated as a harmonic
oscillator,
1
V (x) = Θ2 x2 , (6.2)
2
where we have chosen the origin at x0 = 0 and Θ = d2 V /dx2 (x0 ). By promoting any
scalar potential to a function of position operators, we shall consider the hamiltonian of a
quantum particle immersed on a harmonic oscillator potential as,
P̂ 2 1
Ĥ = + mw2 X̂ 2 . (6.3)
2m 2
6.2 Energy states
In order to algebraically find the energy values, let us write the above hamiltonian operator
as the product of two complex operators:

Ĥ = αX̂ − iβ P̂ αX̂ + iβ P̂ + αβ[X̂, P̂ ], (6.4)
where α and β are real numbers. It follows that

1 1
α2 = mω 2 , β2 = , (6.5)
2 2m
29
30 Chapter 6 Harmonic Oscillator
from which it follows that

1ˆ
†
Ĥ = ~ω â â + I , (6.6)
2
with
r
mω 1
â = X̂ + i P̂ ,
2~ mω
r
† mω 1
â = X̂ − i P̂ . (6.7)
2~ mω
Notice that commutaion between X̂ and P̂ implies that
[â, â† ] = I.
ˆ (6.8)
By defining the Number Operator as
N̂ = â† â, (6.9)
we clearly see that Ĥ and N̂ share a common basis of energy states we shall denote as
{|ni}, with Ĥ |ni = En |ni and N̂ |ni = n |ni.
Exercise 6.1. Show that energy states |ni are non-degenerate.
Now, considering the operators â and â† one can ask about the physical meaning of
states upon which the above operators act as â† |ni. In order to elucidate this, let us
consider operating this state with the Number operator,
N̂ â† |ni = â(N̂ + I)

ˆ |ni = (n + 1)â† |ni ,
ˆ |ni = (n − 1)â |ni .
N̂ â |ni = â(N̂ − I) (6.10)
From this we see that the states â |ni and â† |ni are eigenvector of N̂ with eigenvalues
given by (n ± 1) respectively. Since there is no degeneracy it follows that
Cn |n + 1i = â† |ni , and

Dn |n − 1i = â |ni . (6.11)
Therefore, by taking normalized energy states, we have that

√
â |ni = n |n − 1i ,
†
√
â |ni = n + 1 |n + 1i . (6.12)
Exercise 6.2 Compute the commutators between N̂ , â and â† .
This tells us that the operators â and â† lower and rise the energy eigenvalues, a reason
by which they get their names rising and lowering operators or annihilation and
creation operators. Notice however that the vector â |0i is null and in consequence the
6.3 Observables and uncertainty 31
ground state is identified with the state |0i with a ground energy given by E0 = ~ω/2.
Exercise 6.3 Prove the above assertion, i.e., â |0i = 0.
By applying recursively the rising operator on the ground state, we can construct any
energy state |ni, since
|1i = â† |0i ,
√
2 |2i = â† |1i = (â† )2 |0i ,
√ √
3 |3i = â† |2i = 2(â† )3 |0i ,
..
. (6.13)
from which
(â† )n
|ni = √ |0i , (6.14)
n!
1
with an energy level given by En = ~ω(n + 2 ).
Exercise 6.4. By computing the norm of the state â |ni, show that n ≥ 0.
6.3 Observables and uncertainty
Given the basis of energy levels or pure states |ni we can compute the expectation
values of any compatible observable and its corresponding matrix representation. Let us
concentrate on the observables X̂ and P̂ . We have that
r r
0 1 2~ †
0 ~
hn| X̂ n = hn| â + â n = (δn,n0 −1 + δn,n0 +1 ). (6.15)
2 mω 2mω
Notice that the matrix representation of X̂ is non diagonal. A similar computation shows
that
r r
0 1 2 †
0 ~
hn| P̂ n = mω hn| â − â n = −i (δn,n0 −1 − δn,n0 +1 ). (6.16)
2i ~ 2mω
Observe that the matrix elements of P̂ are in general imaginary numbers. This is not a
problem since the expectation value with respect to a pure state gives a null result.
Exercise 6.5. Show that hn| P̂ 2 |ni = ~mω 2

2 and that hn| X̂ |ni =
~
2mω . Using this, show
that ∆X∆P = (n + 1/2)~ with respect to a pure state |ni.
P 6.6. Compute the expectation value of X̂ and P̂ with respect to a general state
Exercise
|Ψi = λn |ni.
Exercise 6.7. Show the viral Theorem, i.e.,

P̂ 2 hn| Ĥ |ni
hn| |ni = hn| V (X̂) |ni = .
2m 2
Chapter 7
Harmonic Oscillator II
7.1 Time evolution of X̂ and P̂
As mentioned, expectation values with respect to pure states vanish and the oscillation
one would expect from this system seems to be hidden in some sense. Let us then study
the evolution of the space and momentum operators in the Heisenberg picture.
In the Heisenberg picture X̂H = Û † X̂ Û , the equations for these two observables are
coupled,
dX̂H P̂H
= ,
dt m
dP̂H
= −mω 2 X̂H . (7.1)
dt
However, the corresponding equations for the rising and lowering operators are decoupled,
dâH dâ†H
= −iωâH , = iωâ†H . (7.2)
dt dt
The solutions show that operators â and â† evolve in time by a Hamiltonian given by
Ĥ = ~ω, this is
âH (t) = e−iωt â(0), â†H (t) = eiωt â† (0). (7.3)
From this we get that
P̂ (0)
X̂H (t) = X̂(0)cosωt + sinωt,
mω
P̂H (t) = −mω X̂(0)sinωt + P̂ (0)cosωt (7.4)
Although the operators oscillate in time as expected, it is their expectation values wrt to
some physical state what we can actually measure. Therefore notice that hn| X̂H |ni =
hn| P̂ |ni = 0 as in the Schrödinger picture. We see therefore that a stationary state will
have always a null expectation value. In general if the physical state Ψ consists on a
32
7.2 Energy wave functions 33
linear combination of different energy states, the result will be non-zero,

X
λ∗n λn0 hn| X̂ n0

hΨ| X̂ |Ψi =
n0 ,n
r
~ X ∗
− = λ λn0 (δn,n0 −1 + δn,n0 +1 ),
2mω 0 n
n,n
r
~
= (λ∗ λn+1 + λ∗n λn−1 ) (7.5)
2mω n
P
Exercise 6.8. Find an appropriate linear combination of energy states |Ψi = λn |ni
such that hΨ| X̂ |Ψi = Acosωt + Bsinωt. Such state is called coherent state.
7.2 Energy wave functions
Once we have a way to build high energy states from the ground sates, it is possible
to use this method to construct the corresponding wave functions. Consider the wave
function associated to the ground state, Ψ0 (x) = hx|0i. Now, since â |0i = 0 we have that
r
mω P̂
0 = hx| â |0i = hx| X̂ + i |0i
2~ mω
r
mω ~ d
= x+ Ψ0 (x). (7.6)
2~ mω dx
A solution for the above differential equation is

2
1 −1 x
Ψ0 (x) = 1/4 √ e 2 x0
, (7.7)
π x0
where r
~
x0 = . (7.8)
mω
The next wave function, i.e., the wave function related to the energy level 1 corresponding
to the ket state |1i is constructed as follows.
Ψ1 (x) = hx|1i = hx| â† |0i

r !
mω P̂
= |xi X̂ − i |0i ,
2~ mω

1 2 d
= √ x − x0 Ψ0 (x). (7.9)
2x0 dx
Applying recursively the operator â† one obtains

n 2
1 1 2 d
1 x
Ψn (x) = √ n+1/2
x − x0 e 2 x0 . (7.10)
π 1/4 n
2 n! x0 dx
34 Chapter 7 Harmonic Oscillator II
7.3 Hermite Polynomials
The wave functions related to a Harmonic Oscillator are also described in terms of specific
functions known as Hermite polynomials. In this section we want to emphasize how easy
is to obtain and derive many properties of special functions departing from the quantum
analysis we have studied.
The wave function for then-th energy state can be written as
Ψn (ξ) = hξ|ni = Λ0 Fn (ξ), (7.11)
where
x
ξ = ,
x
r0
~
x0 = ,
mω
1
Λ0 = 1/4
√ n ,
π 2 n!x0
n
d 1 2
Fn (ξ) = ξ− e− 2 ξ . (7.12)
dξ
By using the closure relation for the basis |ξi,
Iˆ
Z
dξ |ξi hξ| = , (7.13)
x0
one gets that Z
dξ Fm (ξ)Fn (ξ) = π 1/2 2n n!δn,m . (7.14)
However, since we can show that

2 /2
Ψn (ξ) = Λ0 e−ξ Hn (ξ), (7.15)
where
2 dn −ξ2
Hn (ξ) = (−1)n eξ e , (7.16)
dξ n
we can express the orthonormal relation for the functions Fn (ξ) in terms of the Hermite
polynomials Hn (ξ) as
Z
2
dξ e−ξ Hn (ξ)Hm (ξ) = π 1/2 2n n!δn,m . (7.17)
Hermite polynomials consist on a family of n-different solutions of a particular differential

equation. Here we want to show that such differential equation is in fact the Schrödinger
equation. Let us depart from the Schrödinger equation in terms of the variable x,
!
P̂ 2 1
hx| − + mω 2 X̂ 2 |Ψn i = En hx|Ψn i , (7.18)
2m 2
7.3 Hermite Polynomials 35
which in terms of the wave function Ψ(x) it reads
~2 d2 Ψn (x) 1 1
2
+ mω 2 x2 Ψn (x) = ~ω(n + )Ψn (x). (7.19)
2m dx 2 2
Using the fact that
d2 1 d2
= , (7.20)
dx2 x20 dξ 2
we get the differential equation for Hn (ξ),
d2 Hn dHn
2
− 2ξ + 2nHn = 0, (7.21)
dξ dξ
known as the Hermite equation.
Exercise 6.9. Find the orthonormal relation for the continuous basis |ξi and the action
of the momentum operator P̂ on the ξ-space.
Chapter 8
Gauge Transformations in Quantum Mechanics
Gauge transformations in classical electromagnetism play an important role in a deeper

understanding about the dynamics of electromagnetic waves, their propagation among
others. It is then desirable to study its presence in Quantum Mechanics. Up to now we
have seen that some measurable physical quantities keep unaltered by the displacement of
some parameters. For example, we have seen that if displacement on spatial coordinates
does not alter our system, there is a symmetry under the action of a U (1)-group generated
by the linear momentum. The same happens for time displacements or time evolution
which is also U (1)-symmetry generated by the Hamiltonian operator.
Any element g of the Lie Group1 U (1) can be written as g = eiaÂ with a a real constant
number and Â the corresponding observable. In this lecture we shall see that the presence
of electromagnetic fields can be understood as the presence of another U (1)-symmetry on
the physical state |Ψi not related to some change on a measurable quantity. On virtue of
that it is called an internal symmetry.
8.1 Classical description
Consider a probe particle with electric charge q in the presence of an electromagnetic

~ The corresponding Lagrangian is then
field described by the classical potential φ, A.
1 ~
L = mv 2 − q(φ − ~v · A), (8.1)
2
where ~v = ~x˙ . The canonical momentum pi is given by
∂L
pi = = πi + qAi , (8.2)
∂ ẋi
1
The Lie Group U (1) is defined as the set of all complex numbers with a unitary modulus.
36
8.2 Gauge invariance in quantum mechanics 37
with πi = mẋi the mechanical momentum. In this context the Hamiltonian is given
by
X ∂L
H = ẋi − L, (8.3)
∂ ẋi
i
~ 2
p − q A)
(~
= + qΦ. (8.4)
2m
Therefore the dynamics of a charged particle in the presence of an electromagnetic field is
equivalent to that of a free particle in an electric potential by substituting the canonical
momentum p~ → p~ − q A.~ This is call the minimal coupling.
Exercise 8.1. Consider a charged particle at rest in an inertial frame S in the presence of
~ r)). Apply a Galileo
a non-trivial (i.e., it is not made of pure gauge) vectorial potential A(~
0
Transformation to another inertial frame S which moves in direction of positive coor-
dinates along axis x with respect to a static observer in S. What does see an observer in S 0 ?
8.2 Gauge invariance in quantum mechanics
We want to study the dynamics of quantum charged particle in the presence of electro-
magnetic fields. For that we shal assume that the particle is a probe one such that we
can ignore the possible quantum nature of the electromagnetic fields and the quantum
implications in the corresponding interactions with our particle (big assumption!).
The most obvious and straightforward way to do this, is just to promete the potential
~ to vector functions on the space observables and time by using the same
fields Φ and A
classical Hamiltonian. However we want to consider a different and more useful approach.
Let us ask ourselves about the symmetries of the Schrödinger equation.
Let us start by seeing that Schrödinger equation is invariant under a phase global
mapping, this is
|Ψi → eiλ |Ψi ,
where λ is a constant. All physical measurable quantities are invariant. Our next question
is if this invariance can be preserved by promoting λ to be a function on space coordinates,
i.e., for local phase transformations.
Since
∇ eiλ Ψ(x, t) = (iΨ(x, t)∇λ + ∇Ψ(x, t)) eiλ ,
the Schrödinger’s equation becomes
~2

2
∂λ ∂
(i∇λ + ∇) Ψ(x, t) + V (x)Ψ(x, t) = i~ i + Ψ(x, t), (8.5)
2m ∂t ∂t
38 Chapter 8 Gauge Transformations in Quantum Mechanics
clearly showing that Schrödinger’s equation is not invariant under such transformations.
This is the point at which we want to look at this problem with a different approach,
one that will be essential in the way we are going to explain forces and interactions. We
have seen that Schrödinger equation is not invariant under local phase transformations.
Is there a way to make it invariant? Look carefully at the resulting equation 8.5. A way
to make it invariant is if we can say that the derivative operator ∇ satisfies in some way
that
∂λ
∇ → ∇ + i∇λ, and V (x) → V (x) + ,
∂t
as the wave function transform under a local phase mapping. This is possible if the
derivative also contains electric and magnetic potentials, this is if
e~ e
∇ → D = ∇ − i A(x, t), and V (x) → V (x) + Φ(x, t)
~ ~
since the electromagnetic potentials satisfy gauge transformations under Λ with λ = eΛ/~.
ˆ ˆ
Exercise 8.2. Show that the momentum mapping P~ → P~ − eA(
~ X̂) implies in the
space of functions that
e~
∇ → ∇ − i A(~
x).
~
This is very important since it tells us that Schödinger equation posses a U (1)
symmetry (i.e., is invariant under local phase transformations) only if electromagntic
fields are present. In other words, the existence of electromagnetic fields can be thought as
the result of having a U (1) symmetry. Hence, the U (1)-symmetric Schrödinger equation
is
~2 2

∂ e
D Ψ(x, t) + V (x)Ψ(x, t) = − Φ(x, t) Ψ(x, t). (8.6)
2m ∂t ~
~ has the same units of a derivative operator in a system in
Exercise 8.3. Show that eA
which c = 1.
Exercise 8.4. Sakurai exercises we have already done.
Notice that the U (1)-symmetric Schrödinger equation is not relativistic. Clearly space
and time play different roles, as well as momentum and energy. But also in this version,
magnetic and electric potential are treated also in different way (one appears quadratically
while the other just linearly).
Exercise 8.5 Write down the Schrödinger equations from the perspective of two equivalent
observers in different frameworks. Let us say that Ana observes a charged particle in rest
with respect to her. An electric potential is present. Another observer, say Beto, observes
the same charged particle, but Beto moves in the x-direction at a constant velocity v. To
him, the particle moves to his left producing a current J~ = −qvêx . As usual, Beto really
thinks that his observations are correct and that surely Ana is wrong. Discuss about Beto’s
stupidity and show under which conditions both are correct.
Chapter 9
Gauge quantum effects
The presence of electromagnetic fields as a consequence of the existence of an internal

U (1) symmetry has very important and non-trivial consequences in the study of quantum
mechanics and the interaction of a particle with electromagnetic fields and potentials.
The effects we are going to study in this lecture are present only at the quantum level.
9.1 Bohm-Aharonov effect
Let us take the existence of the local U (1) symmetry to the next step and explore some
direct consequence starting from the simplest case: the absence of electromagnetic fields.
Hence, let us say that a particle with electric charge q is placed in a space with null
~
electromagnetic fields. This means that ∇ × A(x) = 0 and ∇Φ(x, t) = 0. From the gauge
invariance we can select a potential fields made of pure gauge, this is
~ t) = ∇Λ(x, t), ∂Λ
A(x, Φ(x, t) = . (9.1)
∂t
Hence, for a non-constant Λ the ket state is invariant under the local phase transformation
q
|Ψi → e−i ~ Λ(x,t) |Ψi ,
according to the local U (1)-symmetry.
Consider now the following situation. Let us say there is a charged particle in x = 0
meaning that it has associated a wave function h0|Ψi = Ψ(0, t). Due to the above local
q
symmetry we can remplace the ket space |Ψi by |Ψ0 i = e−i ~ Λ(0,t) |Ψi. Notice that since
we do not know anything about quantization of electromagnetic fields, evaluating gauge
fields at points in space is taken in the same context as the classical case.
Now, let us take a displacement of dx, such that the wave function is given by
hdx|Ψi = Ψ(dx, t). In order to keep the local U (1)-symmetry we need realize how the
local phase transformation adapts to this case. This is
P̂ q
hdx|Ψi = h0| e−i ~ dx |Ψi → hdx| e−i ~ Λ(dx,t) |Ψi , (9.2)
39
40 Chapter 9 Gauge quantum effects
dΛ
where Λ(dx, t) ≈ Λ(0, t) + dx (0, t)dx. Hence we get that for a finite displacement
q Rx
~
~ dl
|Ψi → ei ~ 0 A·
|Ψi , (9.3)
~
where the trajectory from point 0 to x is given locally by dl.
Exercise 9.1. Show the above assertion.
Consider now a closed trajectory in a two-dimensional space. A well-defined wave

function implies that
q H ~ ~
|Ψi → ei ~ A·dl |Ψi , (9.4)
~ r, t), since
which is fulfilled in the absence of a magnetic field B(~
I I Z
A ~ = ∇×A
~ · dl ~ = B
~ · ds ~ = φB ,
~ · ds (9.5)
where φB is the magnetic flux through the surface limited by the arbitrarily closed
trajectory we have selected.
Now, we are going to study a very particular and interesting case in which a magnetic
field vanishes in all space except at a point inside the above closed trajectory with
a transversal direction wrt the x-y plane. Notice that this implies the presence of a
singularity since
~ r, t) = B(t)δ(0, 0)êz ,
B(~
for which we are going to change to complex coordinates. First of all, notice that we can
establish an equality as I I
~ ~
F · dl = F (z)dz, (9.6)
where F (z) = Fx − iFy with Fx and Fy the components of F~ and dz = dx + idy for a
generic vector field F~ (x, y).
Exercise 9.2 Prove the above assertion and the exact conditions under which the equality
makes sense following all steps in the Box Tool about Ampere’s law and Complex Analysis
at the end of this section..
The presence of a singular point z0 at which a complex function F (z) diverges implies
that I
F (z)dz = 2πi Res(F (z0 )). (9.7)
By taking a magnetic field B ~ which diverges at the origin (where we have located a
singular point representing an outcoming lineal current) we conclude that
~ r, t) = 2πb êφ ,
B(~ (9.8)
r
9.1 Bohm-Aharonov effect 41
~ = ∇ × A,
where −b is the imaginary part of the Residue of B(z). Now since B ~
Z I
~ · dS
A ~ = ∇×A ~ = 2πb,
~ · dφ (9.9)
implying that
~ y) = 2b eˆz ,
A(x, (9.10)
r2
from which we conclude that in the presence of a singluar B field, the local U (1)-symmetry
indicates that q H ~ ~ q
|Ψi → e−i ~ A·dl |Ψi = e−i ~ φB |Ψi . (9.11)
Although we only have a change by a phase and therefore no implications on the
probability given by the corresponding wave function, notice that the phase of the particle
is affected by the magnetic field even though the trajectory of the particle is located in a
zone where B ~ = 0.
Exercise 9.3 Prove Eq.(9.10).
To think: Why are we talking about trajectories?
The singular point in the x-y plane is a point at which the magnetic potential A ~
diverges and corresponds to a localized current transversal to the x − y-plane. Observe
~ = 0 at all points different to the one at which the current is located.
that B
Now, let us consider two particles, located at the same point and moving towards
a second common point by two different trajectories, we call I and II, surrounding the
singular point. At the initial point the total wave function is given by ΦI (0, t) + ΦII (0, t)
from which the probability is given as
|Ψ(0, t)|2 = |ΨI (0, t)|2 + |ΨII (0, t)|2 + 2Re(ΨI (0, t)ΨII (0, t)∗ ). (9.12)
After arriving at the final point ~x, the wave functions are transformed by a phase given
by the path integral from 0 to ~x, this is
q R
~
~ dl q R
~
~ dl
ΨI (~x, t) = e−i ~ I A·
ΨI (0, t), ΨII (~x, t) = e−i ~ II A·
ΨII (0, t). (9.13)
Therefore, the interference pattern given by the last term in Eq.(9.12) changes by the
factor
Z Z I
q ~ −
~ · dl ~
~ · dl q ~
~ · dl
cos −i A A = cos −i A
~ I II ~

qφB
= cos −i . (9.14)
~
This is the Bohm-Aharonov effect. Two charged particles traveling in different paths
surrounding a point at which there is a non vanishing magnetic field, eventhough the
field vanishes at the trajectories, establishes an interference patterns which depends on
the magnetic flux wrt to the surface limited by the afore mentioned trajectories. This
implies that the values of the magnetic potential indeed have a physical meaning since it
manifests in measurable quantities as the interference pattern.
9.2 Tool Box 1. Complex Analysis and Ampere’s law
In this box we want to show how to use complex analysis in electromagnetism and how to relate contour
integration with closed path integrals. Let us start by specifiying the relation between a contour integral
and a real closed integral.
~ (x, y) which in the time independent frame formed by the orthonormal

Step 1. Consider a vector field F
basis êx , êy is given by
~ (x, y) = Fx êx + Fy êy .
F
The trajectory C, along which the integration is performed, is given by the map:
~ : I → R2 ,
φ
~
such that φ(t) = φx (t)êx + φy êy . A tangential vector along C is then given by
~
dφ ~˙ = φ̇x êx + φ̇y êy ,
=φ
dt
~ given by N ¨
~
~ = φ ¨ ~ ~
~
while a normal vector to C is determined by normalizing the vector N − (φ · φ)T , where
~ ~
T̂ = φ̇/||φ̇||. Hence the trajectory integral can we be written as
I Z b
F ~=
~ (x, y) · dφ ~˙
~ (x(t), y(t)) · φ(t)
F dt,
a
~
where φ(a) ~
= φ(b). Notice that in terms of a mobil basis formed by (T̂ , N̂ ) along C, the above integral is
written as I Z b
~ · dφ
F ~˙
~ = ||φ|| FT (t)dt,
a
where F~ = FT T̂ + FN N̂ . This mobil orthonormal basis along a curve is called the Frenet-Serret basis, and
expressing a two-dimensional vector in terms of polar coordinates and basis êr and êθ is an example of that.
~ being a conservative field, the closed integral vanishes.
Notice as well that for F
Step 2. Now we want to see under which conditions we can relate a contour complex integral to the above
path integral. Let us start by considering a complex function F (z) and a curve in the complex plane given
by
φ : I → C,
H
such that φ(t) = f (x, y) + ig(x, y), with t ∈ I. The contour integral F (z)dz takes into account values of
z along a curve C, this means that z = φ(t) ∈ C. Therefore
I Z
F (z)dz = (u + iv)(df + idg),
C
where F (z) = u(x, y) + iv(x, y). Hence,
I Z b Z b
F (z)dz = (uf˙ − v ġ)dt + i (uġ + v f˙)dt.
a a
Recall that if F (z) is a holomorphic function all over the region limited by the contour C, the closed integral
vanishes.
~ equals F (z)dz when

~ · dφ
H H
Step 3. We say that the integral F
I I I
Re F (z)dz = F ~
~ · dφ, and Im F (z)dz = 0.
(continue next page)

9.1 Bohm-Aharonov effect 43
(...continue from page above)
Observe that this holds if
• uf˙ − v ġ = Fx φ˙x + Fy φ˙y ,
• uġ + v f˙ = 0.
It follows that this is true if Fx = u, Fy = −v, φ̇x = f and φ̇y = g. It is important to observe that a non-
conservative force is translated by this relationship into a non-holomorphic function. In the same context,
the source for a force to be non-conservative comes from a singularity under this relation.
Under this approach and since

I
F (z)dz = 2πi Res(F (z0 )),
where z0 is the point in the complex plane at which F (z) diverges (has a pole, a singularity) and by denoting
Res(F (z0 )) = −ib, with a, b ∈ R, we can say that
I
~ = 2πb,
~ · dφ
F
and therefore
∞
b X n
F (z) = − + an z .
z − z0 n=0
Be aware that this applies for a pole of order 1. Surely the reader can generalize such relation to poles of
order n.
~ in terms of the tangential vector T̂

~ · dφ
H
It is equally important to observe, from the expression of F
~ does not
H
that the fact that Im F (z)dz vanishes, implies that the normal component of the vector field F
contribute to the closed integral, from which it follows (since the trajectory can be deformed into a circle
surrounding the singularity) that BN = 0. Hence we can say that
~ (x, y) = FT T̂ .
F
Example: Ampere’s law. Now, we are going to use the above identification between a real and a complex
closed integral to compute the magnetic field of a line current, this is Ampere’s law is just a specific case
deduced from the Residual Theorem in complex analysis if we identify certain physical configurations as
singularities in the complex plane.
Hence, consider a current line placed on the z-axis. We are not consider a non-zero radius for the
wire carrying the electric current. In that sense, the line current is a point in the transverse plane x − y.
The magnitude of the magnetic field produced by the line current increases as we go closer to it such that
~ → ∞ as r → 0 with r 2 = x2 + y 2 . Hence we can use the above identification by promoting the vector
||B||
~ into a complex non-holomorphic function B(z), this is we complexify the vector field B.
field B ~
We take
~
B(x, y) = Bx êx + By êy → B(z) = Bx − iBy ,
H
and our purpose is to use the complex integral B(z)dz to deduce Ampere’s law. Then it is necessary to
fulfilled the constraints derived from
H our identification between the real integral and the complex one. In
particular, the imaginary part of B(z(dz must vanish, independently of the trajectory we have used to
integrate B(z) around the singularity related to the line current. This is
Bx sinθ + By cosθ = 0 = Br ,
indicating that the fact that the trajectory can be modified into a circle around the singularity, implies that
~ does not contribute to the integral. We can deduce from this relationship the
the radial part of B
direction of B~ which in real integration must be assumed. Hence
I Z 2
B(z)dz = r πBθ dθ = 2πb,
0
(continue next page...)

(...continue from page above)
where ResB(0) = −ib. Hence we find that

~ = 2πb êθ .
B
r
Comparing with Ampere’s law we see that
µo I
b= ,
4π
and then
∞
µ0 I X n
B(z) = −i + an z .
4πz n=0
Ampere’s law is under this approach, the result of promoting a point-located source of current as a singular
point in a complexified version of the fields.
9.3 Magnetic Monopole
A second very important consequence of the interaction of electromagnetic fields with

quantum particles is the quantization of electric charge by the simple assumption of the
existence of a magnetic monopole. First of all we must understand under which conditions
on the magnetic potential A ~ we say that there is no magnetic monopole. For that let
us remember that the no existence of magnetic monopoles follows from the Maxwell’s
equation ∇ · B~ = 0, which in turn implies that we can express the magnetic field in terms
of a magnetic potential as B~ = ∇ × A.~
The key point is that we are using a globally well-defined vector potential A, ~ this
is, there is no point in space at which the magnetic potential is not well behaved. We
shall see that by assuming the existence of a magnetic monopole, we must abandon the
concept of a globally well-defined vector potential. This is as follows. Consider a point-like
magnetic monopole, this is
∇·B ~ = 4πρm , (9.15)
where the magnetic charge is given by

I
Qm = dv ρm , (9.16)
with ρm (~r, t) = Qm δ(~r). Then

I
1 ~
~ · ds.
Qm = B (9.17)
4π
Under this assumption, it is not possible to have a globally well-defined vector potential A~
since ∇ · ∇ × A~ = 0. The solution is defining two different vector potentials A ~ I and A
~ II
such that they both cover all space although each of them are not globally well-defined.
Let us start by taking I Z Z
B ~ =
~ · ds B ~ +
~ · ds ~
~ · ds,
B (9.18)
I II
9.3 Magnetic Monopole 45
where regions I and II can be thought as part of the surface limiting the volume on
which integration of the divergence of ~b is performed. Hence since they are open sets,
they have a boundary, and then
Z I
B ~ = A~I · dl,
~ · ds ~ (9.19)
I
~ is the tangential differential line element on the boundary of region I and A

where dl ~ I is
well-defined over region I but no over region II. A similar situation holds for region II.
In order for the two potential to cover the whole space, we need that they intersect at
some region. By pedagogical purposes, let us think on a sphere sorounding the magnetic
monopole. Region I is then the upper half hemisphere and Region II the lower half
hemisphere and they intersect on region around the equator. Since they are gauge fields
we expect the physics they imply is the same for both of them, therefore we require that
their difference is given by a gauge transformation, this is
~ I (~r) = A
A ~ II (~r) + ∇λ(~r). (9.20)
~ r) = Qm /r2 êr for the magnetic monopole located at the origin. By
Consider then that B(~
expressing ∇ × A~ in spherical coordinates we get that
~ r) = Qm (1 − cosθ) êφ .
A(~ (9.21)
r sin θ
Exercise 9.4. Prove the above equation.
We can select
~ I (~r) = Qm 1 − cosθ êφ ,
A for 0 ≤ θ < π, (9.22)
r sinθ
and
A~ II (~r) = − Qm 1 + cosθ êφ , for 0 < θ ≤ π. (9.23)
r sinθ
~ I diverges at θ = π while A
Notice that A ~ II diverges at θ = 0.
~ I and A
Exercise 9.5. Prove that A ~ II are indeed solutions for the magnetic monopole.
It follows that in the overlapping region,
A ~ I = − 2Qm êφ ,
~ II − A (9.24)
rsinθ
from which we conclude that the gauge function λ is linear on the angle φ, this is
λ = −2Qm φ. (9.25)
Consider now the equator in the overlapping region, at which we have that θ = π/2 and
both functions are well-defined. Since the wave function is well defined up to a local
phase given precisely by the gauge function λ, we have that
qλ
ΨI (~r) = ΨII (~r)ei ~ , (9.26)
where q is the electric charge. For the wave function to be single valued after a 2π + n
rotation with n ∈ Z,
2qQm
= n ∈ Z, (9.27)
~
implying that electric charge is quantized by the presence of a single magnetic monopole.
This is Dirac quantization.
Exercise 9.6. Show that the presence of a magnetic monopole implies the existence of a
magnetic current. Write the corresponding Maxwell’s equations.
Exercise 9.7 *. Sketch a different procedure by assuming the existence of a current of

magnetic monopoles J~m . In this case A
~ is time-dependent. Could we conclude the same?
Could you do it explicitly?
9.4 Landau levels
In the previous section we left as an exercise (exercise 2.36, Sakurai) to analyze the
dynamics of a charged particle immersed in a constant magnetic field pointing along the
z-direction. So we encourage the reader to do such an exercise due to its relevance for
this section.
~ = Bêz . Its
Consider then an electric charged particle immersed in a magnetic field B
Hamiltonian operator is then given by
(p~ˆ − q A)
~ 2
Ĥ = , (9.28)
2m
which can be written as

pˆz 2 ~qB 1ˆ
Ĥ = + (N̂ + I), (9.29)
2m m 2
with the number operator N̂ is determined by the algebra of the mechanical momentum
operators Π̂x and Π̂y given by
ˆ
[Π̂x , Π̂y ] = i~qB I. (9.30)
Hence, it is possible to define the rising and lowering operators â and â† in analogy to
the Harmonic Oscillator, as
" ! !#
1 X̂ i Ŷ i
â = √ + Pˆx −i + Pˆy ,
2 2 ~ 2 ~
" ! !#
1 X̂ i Ŷ i
â† = √ − Pˆx +i − Pˆy , (9.31)
2 2 ~ 2 ~
9.4 Landau levels 47
such that N̂ = â† â. However it is also possible to define other two operators as
" ! !#
1 X̂ i ˆ Ŷ i ˆ
b̂ = √ + Px + i + Py ,
2 2 ~ 2 ~
" ! !#
1 X̂ i Ŷ i
b̂† = √ − Pˆx − i − Pˆy , (9.32)
2 2 ~ 2 ~
with [b̂† , b̂] = −Iˆ and hence, they satisfy the same algebra as for raising and lowering
operators. This tells us that the system can be understood as a Harmonic Oscillator with
two sets of raising and lowering operators.
Also notice that while the magnetic field has a direct consequence on the dynamics of
the particle in the x-y plane, it makes no restriction on the z direction along which it
moves as a free particle. This means that there must be a z-component of the angular
momentum L̂z . Since [Ĥ, L̂z ] = 0 both operators share a common base of eigenkets we
shall denote as |m, ni, so we can write that
L̂z |m, ni = −~m |m.ni , (9.33)
with m an integer1 . Moreover we can show that

L̂z = ~ â† â − b̂† b̂ . (9.34)
Define N̂b = b̂† b̂. Since

[N̂b , N̂ ] = [N̂b , H] = 0, (9.35)
we can write that N̂b |m, ni = ~nb |m, ni where nb is related to m, n by
~nb = ~(m − n). (9.36)
Exercise 9.8. Prove equations 9.34 and 9.36.
ˆ
Exercise 9.9. Prove that [b̂, b̂† ] = I.
Since nb ≥ 0 it follows that m ≥ n. This means that for each value of n we can rise
an infinite number of times the number m. The most general state is then given by
(b̂† )m+n (â† )n

|m, ni = p √ |0, 0i . (9.37)
(m + n)! n!
Notice that the energy En only depends on the number n for which we have an m-times
degenerated system.
1
See Chapter 10
The general wave function is then given by

z̄ n n+m − |z|2
Ψm,n (x, y) = ∂z − z e 4 , (9.38)
4
where z = x + iy is a complex coordinate. In general we can write that
|z|2
Ψ0,0 (x, y) = H(z)e− 4 , (9.39)
where H(z) is an analytical function.
Exercise 9.10 Prove Eqs. 9.37, 9.38 and 9.39.
Project Landau Levels **

The next project is optional. Please do it just in case you feel interested.
1. Consider a charged particle moving in the x − y-plane with a constant magnetic

field directed along the z-axis. Also consider the presence of a second magnetic
~ 2 = B2 δ(0, 0)êz , this is a magnetic field at a singular point as that studied in
field B
the Bohm-Aharonov effect. In mathematical terms we have included a singularity.
What is the effect on the Landau Levels?
2. Consider a magnetic monopole in the presence of an electric field directed along

z-axis. Can we have a dual effect of Landau Levels?
3. Consider electric charged particles and a magnetic monopole in an xy-plane in the

presence of a magnetic and electric field directed along z-axis. Study the system
and tell if Landau levels are still present in this much more complex system.
Chapter 10
Angular Momentum
We have seen that basic transformation on the wave functions correspond to the action of
a specific operator on the ket state in Hilbert space (see Appendix B). So far we have
studied translations, time evolution and also internal gauge transformations connected to
U (1)-symmetries. However we have not consider which kind of operators drive a rotation
on the physical state in configuration or momentum space. This is equivalent to look for
the operator which generates a rotation. Along the way we shall learn that there exists a
physical observable responsible for that, called General Angular Momentum, which under
specific circumstances reduces to the well known Angular momentum defined by the cross
product between position and linear momentum.
10.1 General Angular Momentum
Let us start by asking the following question: Is there an operator in Hilbert space which
produces a rotation on the space of quadratically integrable functions? The abstract
idea behind this question is to find an operator D̂(R) depending on some parametrizable
object R such that
D̂(R) |Ψi = |ΨR i , (10.1)
where the wave function h~r|Ψi = Ψ(~r) is rotated as
Ψ(~r) → h~r| D̂(R) |Ψi = ΨR (~r),

D E
= r~0 Ψ = Ψ(r~0 ), (10.2)

E E
where r~0 = D̂† (R) |~ri and in consequence
~0
r = |R~ri with r~0 = R~r with R ∈ SO(3)

a rotation matrix. In other words, R is the matrix representation in R3 of the operator D̂† .
In order to get more information about operator D̂ let us consider its action on a
vector operator V ~ˆ (as linear momentum). Let us say that
~ˆ =
X
V V̂i eî , (10.3)
i
49
50 Chapter 10 Angular Momentum
where V̂i are operators while êi are the corresponding orthonormal basis. Consider now
two different states |αi and |βi such that
hα| V̂i |βi → hαR | V̂i |βR i ,

= Rij hα| Vˆj |βi ,
= hα| D̂† V̂i D̂ |βi , (10.4)
from which we conclude that the matrix element transforms as
D̂† V̂i D̂ = Rij Vj . (10.5)

As usual and with the goal to elucidate more properties of this operator D̂ let us consider
an infinitesimal transformation. Since R ∈ SO(3) we can write
R ' 1 + ω, (10.6)
where 1 is the unit matrix. Since the group SO(3) is defined as
SO(3) = {M3×3 (R)|M M T = 1}, (10.7)
we obtain that ω is an antisymmetric matrix, i.e., ω T = −ω.
Exercise 10.1. Prove that ω is an antisymmetric matrix.
This implies we only have 3 degrees of freedom for ω and then we can expand the
operator D̂(R) as
X3
D̂(R) = Iˆ + β ωij Jîj + O(ω 2 ), (10.8)
i=1
with β a numerical factor to be fixed. Notice the introduction of the operators Jîj . Since
ω is antisymmetric we also conclude that Jîj = −Jˆji implying there are as well only 3
different operators Jîj . Let us fix a more suitable notation1 by defining the operator Jî by
1
Jî = ijk Jˆjk , (10.9)
2
which follows by defining Jˆ1 = Jˆ23 , Jˆ2 = Jˆ31 and Jˆ3 = Jˆ12 . In terms of these 3 operators
Jî we can rewrite Ec.(10.8) as
X
D̂(R) = Iˆ + 2β θi Jî , (10.10)
i
with θi = 21 ijk ωjk . Notice that by a dimensional analysis we can fix β = −i/2~.
1
Which also shall help us to find the algebraic structure underlying such operators and allowing us to
define what Angular Momentum is.
10.1 General Angular Momentum 51
Exercise 10.2. i) Prove Eq.(10.10). ii) Discuss fixing β as above.
This result suggests us that it may be possible to write the operator D̂(R) as
i ˆ
D̂(R) = e− ~ Jk θk ,
as the responsable to produce a rotation by an angle θk with respect to the k-axis. The
rotation is then generated by the operator Jˆk . It is then our purpose to elucidate the
nature of Jˆk .
Algebraic properties of Jˆk
From Eq. (10.5) and by substituting D̂ and R by its infinitesimal expression we obtain
that
i
Rij Vj = Vi − ijk Vj θk = Vi + θk [Jˆk , V̂i ], (10.11)
~
from which we find that h i
Jˆk , V̂i = i~kij V̂j . (10.12)
This tells us how a vector operator V̂i behaves under the action of the infinitesimal
generator of rotations Jî . In other words, we can conclude for arbitrary operators that
indeed they are part of a vectorial operator if they are related under the action of Jî by
the above algebra.
In consequence, we can see if Jî is a component of a vector operator by looking at its

algebra. Hence, consider that
D̂(R1 )D̂(R2 ) = D̂(R1 R2 ), (10.13)
and parametrizing rotations by angles θ and φ on the same direction we can write
D̂(Rφ )D̂(Rθ )D̂(Rφ−1 ) = D̂(Rφ Rθ R−φ ). (10.14)
By taking again the infinitesimal expression for D̂ and R, we get

1
θk Jˆk − θk φj [Jˆj , Jˆk ] · · · = θk Jˆk + jkl φj θk Jˆl + . . . (10.15)
~
and then h i
Jî , Jˆj = i~ijk Jˆk . (10.16)
Notice that this is the same algebra that satisfies the Angular Momentum Operator
L̂i = ijk X̂j P̂k ,
ˆ
which allows to cal the generator of infinitesimal rotations J~ = (Jˆ1 , Jˆ2 , Jˆ3 ) as the operator
of General Angular Momentum. The Angular momentum operator L ~ˆ is just an
example of this more general kind of object.
Exercise 10.3. Prove Equations (10.12) and (10.16).

10.2 Eigenvectors and eigenvalues
Our next step is to study the General Angular Momentum’ spectrum. Firs of all define
ˆ
the angular momentum vector operator J~ as
ˆ
J~ = (Jˆ1 , Jˆ2 , Jˆ3 ), (10.17)

[Jˆ2 , Jî ] = 0, for all i, (10.18)
and in consequence these two operators share a common base of eigenvectors we shall
denote as |a, bi, this is
Jˆ2 |a, bi = a |a, bi , Jˆz |a, bi = b |a, bi . (10.19)
Notice a very important selection we have made. Since the componentes of the Angular
Momentum do not commute, they don’t share a common basis of eigenvectors. So we
must select one component and we select Jˆ3 = Jˆz . We still need to find the precise values
for the eigenvalues a and b. For that define the operators
Jˆ± = Jˆx ± iJˆy , (10.20)
which satisfy the following algebra,

h i
Jˆ+ , Jˆ− = 2~Jˆz ,
h i
Jˆ2 , Jˆ± = 0,
h i
Jˆz , Jˆ± = ±~Jˆ± . (10.21)
It follows that

Jˆz Jˆ+ |a, bi = ~Jˆ+ + Jˆ+ Jˆz |a, bi = (~ + b)Jˆ+ |a, bi . (10.22)
This implies that

Jˆ+ |a, bi = C |a, b + ~i , (10.23)
where C is a constant to be determined. Notice that Jˆ+ increases by ~ the eigenvalue b
suggesting that it will be useful to express b as a multiple of ~.
Exercise 10.4. Prove Eqs. 10.18 and 10.21. Second part *: Prove that there is no
degeneracy in 10.23 and then C is indeed a normalization constant.
Exercise 10.5. By defining Jˆ2 in terms of Jˆz , Jˆ± prove that a − b2 ≥ 0.
It follows from exercise 10.5, that for a fixed value of a, |b| has a maximum value
denoted by bmax . At this value,
Jˆ+ |a, bmax i = 0, (10.24)

10.3 Matrix elements of Angular Momentum 53

Jˆ− Jˆ+ |a, bmax i = 0. (10.25)
From this relation we find that
a = bmax (bmax + ~). (10.26)
Now, since b also has a minimum value bmin such that bmin = −bmax we have that there
exists an integer number n such that
Jˆ+
n
|a, bmin i = |a, bmax i , (10.27)
implying that
bmax = −bmax + n~, (10.28)
or equivalently
bmax = j~, (10.29)
where j is an integer or half-integer. Henceforth we denote the corresponding eigenvalues
as
a = ~2 j(j + 1), and b = m~, (10.30)
with m = −j, −j + 1, . . . , 0, . . . , j − 1, j. Then, for a given integer or semi-integer value
of j there are 2j + 1 values for m. Because of this we denote the eigenstates of Jˆ2 and Jˆz
as |j, mi.
Exercise 10.6. Prove all relations from Eq.(10.24) to Eq. (10.30). Prove also that n
must be an integer.
Exercise 10.7*. Show that the angular Momentum Operator L̂2 has integer values for
j.
10.3 Matrix elements of Angular Momentum
Since Jˆ2 and Jˆz share a common basis denoted by |j, mi we can compute the matrix
elements for these operators in terms of such basis. It must be clear for the reader that
in this case the matrices are diagonal since
Jˆ2 |j, mi = ~2 j(j + 1) |j, mi , Jˆz |j, mi = ~m |j, mi . (10.31)
We are now interested in the action of non-compatible components of Angular Momentum

on the same basis. For that we can see that by normalizing the states Jˆ± |j, mi, we obtain
that
Jˆ± |j, mi = ~ j(j + 1) − m(m ± 1) |j, m ± 1i .
p
(10.32)
Since
j, mj 0 , m0 = δjj 0 δmm0 ,

we have that

0 0
j , m Jˆ± |j, mi = ~ j(j + 1) − m(m ± 1)δj 0 j δm0 ,m±1 .
p
(10.33)
Exercise 10.8. a) Take j = 1. Write explicitly the matrices for Jˆx , Jˆy and Jˆz . b)
Express the operators Jî in terms of kets and bras in the corresponding basis, this is
Jî = |j, mi Jmn hj, n|.
10.4 Construction of a finite rotation operator
So far we have seen that for an infinitesimal rotation determined by the angle δθ in the n̂
direction, the corresponding operator can be safely written as
i ˆ
D̂(Rδθ,n̂ ) ' Iˆ − (J~ · n̂)δθ. (10.34)
~
A finite rotation can be constructed by applying N -times the above infinitesimal rotation
for a very big number N (which also can be thought as a limit to infinity), this is,
N
θ
D̂(Rθ,n̂ ) = D̂( , n̂)
lim ,
N →∞ N

ˆ i ˆ θ θ 2
= lim I − (J · n̂) + O( ) , (10.35)
N →∞ ~ N N
from which we conclude that

ˆ
D̂(Rθ , n̂) = e−i(J·n̂)θ/~ . (10.36)
Rotation are indeed generated by General Angular Momentum. But, what’s

the meaning of a rotation generated by Momentum angular with a half-integer value for
j?
Chapter 11
Spinors
We have seen that the operator Jî generates a rotation on the spatial dependence of the
wave function Ψ. However, this opens up the posibility to study a much more wider
phenomena in Quantum Mechanics. Let us start by enumerating some precise statements
about what we already know.
11.1 A warming up example
First of all, notice that all operators Â have an associated matrix which we shall call its
matrix representation. This matrix element of course depends on the basis we choose
and there is not a unique matrix representation. Based on this, let us think the following.
Consider the angular momentum component L̂z acting on the state |l, mi (notice we have
replaced j by l), we know that
L̂z |l, mi = ~m |l, mi ,
Since Lˆ2 has integer eigenvalues, we can consider l = 1 and realize that the matrix
element for L̂z is a 3 × 3 matrix. Therefore the operator D̂ = exp(−i/~L̂z φ) has a matrix
representation with respect to the basis |1, mi given by
 iφ 
e 0 0
[D̂z (φ)]|1,mi =  0 1 0  (11.1)
0 0 e −iφ
once again, with respect to the basis formed by eigenvectors of L̂z .
However, this is not the matrix representation we are looking for. We want the matrix
representation action on R3 , given by R. Such representation must be given in terms
of the unitary vectors in R3 in a given coordinate system. In order to clarify this, let
us compute the matrix representation of the same operator D̂z (φ) and L̂z in R3 with
respect to the coordinate basis êi . Notice that we expect that
 
cosφ sinφ 0
[D̂z† (φ)]êi = −sinφ cosφ 0 (11.2)
0 0 1
55
56 Chapter 11 Spinors
since this corresponds to a Rotation matrix acting on a 3-dimensional vector around the
z-axis, i.e, a rotation around z-axis is generated by angular momentum L̂z . To prove this,
consider the infinitesimal transformation
i
[D̂z† (φ)]êI = [Iˆ + φL̂z + . . . , ]êi
 ~ 
1 φ 0
= −φ 1 0 + O(φ2 ), (11.3)
0 0 1
and therefore
 
0 −1 0
[L̂z ]êi = i~ 1 0 0 (11.4)
0 0 0
is the matrix representation of L̂z in the coordinates basis êi in R3 . Notice the both matrix
representations satisfies the general momentum algebra, for which we conclude that the
algebra is a fundamental property of the generators independent of their representation.
The group of generators L̂i satisfying the angular momentum algebra receives the name
of so(3) and they are defined as
so(3) = {X3×3 (R) ∈ SO(3)|X † = X, Tr(X) = 0}. (11.5)
In our example it must be clear that the generators ~1 [L̂z ]êi belong to so(3). One way
to see that generators must satisfy these properties is the following. Take an element of
SO(3) let us say A = e−iφX . Since A ∈ SO(3), then det(A) = 1. However
det(A) = e−iφTr(X) , (11.6)
and in consequence Tr(X) = 0. Observe also that there is another map relating both
representations
h : so(3) → so(3), (11.7)
such that
h([L̂z ]|1,mi ) = [L̂z ]êi , (11.8)
consisting on a finite number of row operations. Formally h is called an homomorphism

between algebras so(3).
Exercise 11.1. Find the matriz representation wrt the coordinate basis in R3 of L̂x and
Lˆy for l = 1.
Exercise 11.2.* Prove that det(A) = e−iφTr(X) .

11.2 Rotations generated by SU (2) 57
11.2 Rotations generated by SU (2)
Our next step consist on doing the same for generators related to a half-integer eigenvalue
j. Explicitly let us take j = 1/2. First of all, let study some properties of the operator
D̂1/2 (φ) = e−i/~φŜz , where the general angular momentum Jî has been renamed Ŝi for
the case in which j = 1/2. For consistency1 we change our nomenclature and replace all
j’s by s’s. We shall also use the following notation for the eigenvalues of Ŝ 2 and Ŝi ,

1 1 1 1
,− |−i
= , 2 2 = |+i .
and , (11.9)
2 2
Matrix representations of Ŝi wrt to the above basis are given by

~ 1 0 ~ 0 −i ~ 0 1
[Ŝz ]|±i = , [Ŝy ]|±i = , [Ŝx ]|±i = (11.10)
2 0 −1 2 i 0 2 1 0
which can be expressed as well in terms of the Pauli matrices as

~
[Ŝi ]|±i = σi . (11.11)
2
According to our initial statements, these matrices must satisfy similar properties as
the angular momentum components L̂i we have reviewed in the last subsection. The
properties are
[σi , σj ] = 2iijk σk ,
{σi , σj } = 2Iδij , (11.12)
i.,e, Pauli matrices are h2 × 2 matrices,

i matrix representations of operators Ŝi in the |±i
basis. Because of that D̂z,1/2 (φ) is a matrix in SU (2), and since
|±i
i
D̂z,1/2 (φ) = e− ~ Ŝz φ , (11.13)
we conclude that Ŝz is an element of the SU (2) algebra, called su(2), where
su(2) = {iX ∈ SU (2)|X † = X, T rX = 0}. (11.14)
from which one concludes that Pauli matrices {σi } are matrix representations of elements
in su(2).
Exercise 11.3. Prove the following properties:
1. Any element U in SU (2) can be written as U = x4 I + ixk σk .
2. For any X ∈ su(2), X = xi σi .
3. eiσi φ = Icosφ + iσi sinφ.

1
As the reader suspects, the s’s are after the term spin.
Now we want to understand how a 2 × 2 complex matrices as Pauli matrices can

generate a rotation in R3 . This means we have to clearly define a map between SU (2)
and SO(3) and similarly on their algebras, this is from su(2) to so(3).
Define the map

hU : su(2) → so(3), (11.15)
for all U ∈ SU (2), such that
hU (X) = U XU † , (11.16)
with X ∈ su(2). In order to determine the matrix representation of the mapping hU we
evaluate the above mapping on the elements of the basis in su(2) such that we can then
have a matrix representation in terms of those basis for hU .
In general if we evaluate on an element of the basis, we have that

X
hU (iσk ) = Λkj (iσj ) (11.17)
j
for which the matrix representation of hU in terms of the basis of su(2) is

[hU ]su(2) = Λ, (11.18)
where λ is actually a rotation matrix, this is Λ ∈ SO(3). This defines a map
h : SU (2) → SO(3), (11.19)
by associating a matrix representation
h(U ) = [hU ]su(2) . (11.20)
In order to have a concrete example and to realize some important consequences of this
mapping, let us take for simplicity that U = e−iφσ3 and evaluate the map hU on each
element of the basis on su(2), this is on iσk for each k = 1, 2, 3. We have then
he−iφσ3 (iσ1 ) = (e−iφσ3 )(iσ1 )(eiφσ3 )
= (cos2φ)iσ1 + (sin2φ)iσ2 , (11.21)
and similarly for hU (iσ2 ) and hU (iσ3 ), from which in terms of the basis iσi we find that
 
cos 2φ sin 2φ 0
h(U ) = − sin 2φ cos 2φ 0 . (11.22)
0 0 1
This is an amazing result. It tells us that we have constructed a map between SU (2)
and SO(3) matrices and even more, that a full period in SU (2) its mapped into a doble
rotation in SO(3). This is the reason why SU (2) is called the double cover of SO(3).
Notice that this also means that ±U is mapped to same element in SO(3) by h, a reason
from which one can write that
SO(3) ' SU (2)/Z2 . (11.23)
Exercise 11.4. Compute explicitly the matrix h(U ) in the above example.
Exercise 11.5. Compute the matrix representation for D̂z,1/2 (φ) in the |±i basis.
11.3 Finally...spinors 59
11.3 Finally...spinors
We have seen that for j = 1/2 the corresponding operators D̂1/2,i (φ) belong to SU (2)
group. In this case a straightforward question is: What is the meaning of h~r|±i?
Let us start by considering a ket state as a linear combination of the eigenvectors of

Ŝ 2 and Ŝi , this is
|χi = C+ |+i + C− |−i . (11.24)
In terms of the basis of eigenvectors |±i, the projection of the ket state into the configu-
ration space has only two possible components. Let us see what is the effect of applying
an operator D̂ on the above state (notice we have denoted the ket state as |χi):
D̂1/2,z (φ) |χi = e−iφ/2 C+ |+i + eiφ/2 C− |−i , (11.25)
which can be expressed in terms of the corresponding matrix representations as

−iφ/2
e 0
h~r| D̂1/2,z (φ) |χi = h~r|χi . (11.26)
0 eiφ/2
Notice that a rotation generated by Ŝz implies half a rotation that the one generated by
Jˆz with j an integer. The objects related to these generators ande denoted by
χ(~r) = h~r|χi , (11.27)
are called spinors. Notice that by associating

1 0
|+i → , |−i → , (11.28)
0 1
we can also express that

[D̂z,1/2 (φ)]|±i = e−iσz φ , (11.29)
and that
h~r| D̂z (r) |χi = e−iσz φ χ(~r),

−iφ/2
e 0
= χ(~r) (11.30)
0 eiφ/2
where
χ(~r) = C+ χ+ (~r) + C− χ+ (~r), (11.31)
and χ± (~r) = h~r|±i.
’
Exercise 11.6. Using the definitions of Jˆ± adapted for spin operators, compute the
matrix elements for Ŝx and Ŝy .
Exercise 11.7 Show that

~X †
< Ŝk >= χi (σk )ij χj (11.32)
2
ij
where < Ŝk >= hχ| Ŝk |χi.
Exercise 11.8. Show that
(~σ · ~a)(~σ · ~b) = ~a · ~b + i~σ · (~a × ~b). (11.33)
Exercise 11.9. Find Pauli matrices (we have already constructed σz )

0 1 0 i 1 0
σx = σy = σz = . (11.34)
1 0 −i 0 0 −1
Exercise 11.10. Consider an arbitrary unitary vector n̂. Show that
cos φ2 − i(sin φ2 )nz −i(sin φ2 )n+

~
−iS·n̂φ/~ −i~
σ ·n̂φ/2
[e ]|±i = e = , (11.35)
−i(sin φ2 )n− cos φ2 + i(sin φ2 )nz
where n̂± = n̂x ± in̂y .
With the last set of exercises it is easy to see that after a rotation a spinor transforms
as
χ(~r) → e−i~σ·n̂φ/2 χ(~r). (11.36)
11.4 Spinors and Uncertainty
The experimental proof of the existence of spin was established by the famous Stern-
Gerlacht experiment. We shall describe it once we have defined the spin states. If you are
interested in a more physical description of the actual experiment2 , please see Appendix B.
Let us then continue with our description of the experiment. Consider a system in
which we are going to consider only the spin eigenkets. This is, the ket state is just a
linear combination of spin states for s = 1/2, then
|Ψi = C+ |+i + C− |−i (11.37)
Hence the probability P(±) of measuring the state |±i is given by
P(±) = |C± |2 = | hψ|±i |2 . (11.38)
Since our system is only characterized by the spin degrees of freedom, we can say that
hψ|ψi = 1. It follows then that
|C+ |2 + |C− |2 = 1 (11.39)
defining a circle in the |±i-space. Giving equal probabilities for the state to be described
by |+i or |−i (why?) we have that
1
|C± |2 = . (11.40)
2
2
Watch this video → https://www.youtube.com/watch?v=PH1FbkLVJU4.
11.4 Spinors and Uncertainty 61
In general we can parametrize the coefficients such that
C+ = sin θ/2, C− = eiφ cos θ/2, (11.41)
with 0 ≤ θ ≤ π and 0 ≤ φ ≤ 2π. We have now use the equation for a circle and
parametrize it such that it now describes the equation of a 3D-sphere. In this space of
parameters, e.g., the state |+i corresponds to C+ = 1 and C− = 0. However now we can
describe as well states |±ix and |±iy . For instance, the state |+i is now described by
φ = 0 and θ = π/2, and
π π 1
|+ix = sin |+i + ei·0 cos |−i = √ (|+i + |−i), (11.42)
4 4 2
as expected. Since the basis |±i is orthonormal, we have that the closure relation in this
case reads
Iˆ = |+i h+| + |−i h−| , (11.43)
from which we can write
Ŝz = IˆŜz I,
ˆ
= |+i h+| Ŝz |+i h+| + |−i h−| Ŝz |−i h−| ,
~
= (|+i h+| − |−i h−|) , (11.44)
2
and deduce the matrix representation of Ŝz wrt |±i,

~ 1 0
[Ŝz ]|±i = , (11.45)
2 0 −1
in concordance with all we have mentioned.
After a measure of the system, the quantum state collapse to one of the two possible
states |±i and the probability to obtain such a result is given precisely by 1/2. Let us
think that we measure the state |+i, this is |ψi = |+i and now the probability to measure
(by a second measurement) the state |+i or |−i is 1 or 0 respectively. However, since the
components of the spin operator Ŝi do not commute among them there is an uncertanty
by measuring two spin components at once. Let us say we want to measure the spin
x-component. For that we need to express the state |+i in terms of eigenstates of Ŝx .
This is illustrated as follows. Consider the eigenstates of the spin operator in components
x, |±ix , such that
~
Ŝx |±ix = ± |±ix , (11.46)
2
and we want to write the state |+i as
|+i = A+ |+ix + A− |−ix , (11.47)

and to know the values of A± . Using matrix representations for the spin z-projections, it
is easy to show that
Ŝx = |+i h−| + |−i h+| ,

1
|±ix = √ (|+i ± |−i), (11.48)
2
√
from which A± = 1/ 2. We are now in conditions to know how the state |+i collapses
by measuring the spin projection on the axis x. The state |+i can be expressed as
1
|+i = √ (|+ix + |−ix ), (11.49)
2
then, by measuring the observable Ŝx there is an equal probability to obtain the states
|±ix . Let us say that the state |+i collapses to |+ix after this second measurement. What
is the probability for the state to be described by |−i after this second measurement?
Once again we need to write the state |+ix in terms of |±i, which is given by Eq.(11.48).
Hence the probability for the state to be described by |−i after being collapsed into
|−i is non-zero, since it involves a second measurement of another spin projection. The
fact that these two observables do not commute implies that there is an uncertainty by
simultaneously measure them.
Exercise 11.11. Express states |±iy in terms of |±i and |±ix . Perform a series of
measurements with different options. Interpret your results.
Exercise 11.12. Construct the operator Ŝy in terms of the basis |±i.

~
∆Si ∆Sj ≥ . (11.50)
4

~
Ŝx |+i = |−i . (11.51)
2
Notice this nice result: Ŝx changes the state |+i into |−i !
11.5 Energy states and spinors
As we have seen, spin states |±i are just the corresponding eigenvectors for a special
Angular Momentum, As for the case of angular momentum L̂i we expect to have an
energy contribution from it, as in classical mechanics. Let us then review very fast this
classical result.
11.5 Energy states and spinors 63
Let us consider a particle with kinetic energy with only kinetic energy and with a
Lagrangian given by
1
L = m~r˙ (t) · ~r˙ (t) (11.52)
2
where ~r˙ (t) = r(t)êr + rθ̇êθ . The fact that θ is not a constant implies that the trajectory
is not following a straight line and then velocity is not constant. This means that we can
associate angular momentum to the particle’s motion. The interesting part is that it is
possible to decompose the kinetic energy into two parts: kinetic energy by displacement
and kinetic energy by angular momentum. This last component sometimes is absorbed
by a scalar potential defining an effective potential. In this context the above Lagrangian
gives rise to a Hamiltonian given by
p2r L2
H= + , (11.53)
2m 2mr2
where L ˙
~ is the angular momentum ~r × p~ and pr is the radial component of m~r(t). There
is also another terms which contributes to the Hamiltonian in which Angular Momentum
plays a key role. In the presence of an external non-trivial magnetic field, an electrical
charged particle with angular momentum defines a potential energy. This is as follows.
~ r(t)). As we know, we can construct a magnetic
Consider an external magnetic field B(~
potential energy such that we can compute the force through the gradient, this is
F~B = −∇U
~ (~r), (11.54)
where U = −m ~ The quantity m

~ · B. ~ is known as the magnetic dipole moment and it
corresponds to the first approximation object given rise to a general magnetic field. It is
represented by a circular current by a particle with charge q. Since the particle circulates,
it also has associated an angular momentum, such that it is possible to write the dipole
magntic momentum in terms of Angular momentum. This is known as the giromagnetic
relation and it is given by
~
qL
m~ = . (11.55)
2m
with a potential energy given by
~ ·B
qL ~
U =− . (11.56)
2m
Hence by promoting to operators the above relation, we see that angular momentum,
coming from a particle describing a local or global rotation, can be expressed as
P̂ 2 L̂2 q ~ ~
Ĥ = + − L · B, (11.57)
2m 2mr2 2m
Let us go back to Quantum Mechanics. By promoting the above Hamiltonian to an
operator, we have that angular momentum L ~ˆ contributes to energy and since [Ĥ, Lˆ2 ] = 0
they share a common base of eigenvectors. This is, eigenvectors of L̂2 are also energy
states. Since [L̂2 , L̂z ] = 0 as well, we have a set of commutable observables given by Ĥ, L̂2
and Lˆz with a common basis denoted by |n, l, mi where n denotes the energy levels for
a confined potential. Hence, for the basis |n, l, mi to be an eigenvector for the above
Hamiltonian, we can choose an external magnetic field to be aligned along the z-axis,
although a more general situation can be considered.
Finally, we are going to take a big step. We shall generalize the interaction between
angular momentum, L ~ˆ and magnetic field B
~ to our general angular momentum J. ~ˆ This
is supported from the experimental evidence about the existence of spin states through
the Stern-Gerlacht experiment. Then in the following sections we are going to consider
an energy term given by
gq ~ˆ ~ ˆ
− J · B(~r), (11.58)
2m
where g is called the gyromagnetic ratio for the particle we are considering. In particular
ˆ ~ˆ
we are going to consider the electron spin, for which we are taking J~ = S and3 g = 2.
Exercise 11.15. Show that indeed, for a spin 1/2 particle, g = 2.
11.6 Spin precession
Let us consider an electron immersed in an external magnetic field and let us take
~ = Bêz . The interaction given by an electron in such a field is given by
B
q
Ĥ = − B Ŝz . (11.59)
m
Notice that energy states denoted |ni are also spin states, i.e., a linear combination of
states |±i. Our goal here is to study these states. For that, we apply Ĥ to |ni,
eB
Ĥ |±i = − Ŝz |±i ,
m
~qB
= ∓ |±i , (11.60)
2m
from which there are 2 different energy states given by
~
E± = ∓ ω, (11.61)
2
with the frequency ω = qB/2m and Ĥ = ω Ŝz . Notice that so far, we have considered a
constant magnetic field and therefore the Hamiltonian is not time dependent. The energy
states are stationary states, while for |Ψi = C+ |+i + C− |−i
Û (t, 0) |Ψ(0)i = e−iĤt/~ |Ψi ,

= e−iωtŜz /~ |Ψi ,
= |Ψ(t)i , (11.62)
3 α
Actually ge = 2(1 + 2π
+ . . . ) where α is the fine-structure constant, a quantity that comes out from
relativistic corrections to our current framework.
11.7 Pauli’s equation 65
with
|Ψ(t)i = e−iωt/2 C+ |+i + eiωt/2 C− |−i . (11.63)
where |C± |2 is the probability for the state |Ψ(t)i to be |±i.
Exercise 11.16 . Show that
• If the state at t = 0 is given by |+i, the probability for |Ψ(t)i to be given by the
state |+i is one and < Ŝz >= ~/2,
• If |Ψ(0)i = |+ix , show that the probability for the state |Ψ(t)i to be in the state
|+i is cos2 (ωt/2).
• Show that < Ŝx >= ~2 cos ωt where the average value is taken with respect to a
general state |Ψ(t)i. . Compute also < Ŝy >.
Under this context it is possible to show that that if the state |Ψ(t0)i is prepared in
the state |+ix , the probability for the state |Ψ(t)i to be in the state |±ix is given by
ωt ωt
|x < +|Ψ(t) > |2 = cos2 ( ), |x < −|Ψ(t) > |2 = sin2 ( ). (11.64)
2 2
From this we see that the average measure of Ŝx of a spin 1/2 particle interacting with an
external magnetic field aligned along the z-axis shows a precession around the direction
of the magnetic field. The same happens to the observable Ŝy .
~ˆ · n̂ is the projection
Exercise 11.17. Compute the average value for Ŝn where Ŝn = S
of the spin vector operator along a general unitary vector n̂.
11.7 Pauli’s equation
We have seen that for an electrical charged particle with spin 1/2 immersed in an external
magnetic field, the most general Hamiltonian is given by
P̂ 2 L̂2 q ~ˆ + L)
~ˆ · B,
~
Ĥ = + 2
− (2S (11.65)
2m 2mr 2m
where L~ˆ is the angular momentum derived from the presence of a force acting on the
charged particle and r is the distance of the particle position to the center of the force.
For simplification, we shall ignore this term along this section.
The question we want to ask at this point is if we can deduce such an expression
without assuming the interaction between the external magnetic field and the spin particle.
We want to obtain this interaction term from some more basic assumptions in the same
context as we can deduce the way in which a charged particle interacts with an external
~ · P~ˆ where P~ˆ is the canonical momentum promoted to

magnetic field through the term A
an operator. Hence, recalling:
Free particle Hamiltonian −→ q-charged particle in a magnetic field Hamiltonian,

P̂ 2 ~ 2
(P̂ − q A) Π̂2 2q
Ĥ = −→ Ĥ = = − Ai Π̂i . (11.66)
2m 2m 2m m
ˆ ~ˆ
where P~ = Π ~ It is probable that at this point you ask yourself about all different
+ q A.
ways to write a Hamiltonian containing angular momentum and its interaction with an
external magnetic field. Consult Tool Box 2 for details.
Pauli realized that the spin-magnetic field interaction can be deduced if the momentum
of a free particle is now mapped into
2
2 2 2 ~ˆ ~ˆ ~
P̂ −→ P̂ = S · (P − q A) , (11.67)
~
such that the Hamiltonian is given by

1 2
Ĥ = P̂ . (11.68)
2m
~ = 1B
Exercise 11.18. Prove the above assertion by considering that A ~ × ~r.
2
Therefore the quantum equation describing the dynamics of a non-relativistic one-half

spin particle is given by
∂
Ĥ |χ(t)i = −~ |χ(t)i , (11.69)
∂t
which after projecting into the configuration space gives us the so called Pauil’s equa-
tion:
1 ~ + q A)
~
2 ∂χ(t)
~σ · (i~∇ χ(t) = i~ (11.70)
2m ∂t
~ s as
Notice this allow us to define a new derivative operator we denote as ∇
~ s = √1 ~σ · (i~∇

− i~∇ ~ + q A)
~ eˆj , (11.71)
3 j
although a much more concrete definition of the spin-derivative operator comes from a
relativistic version of quantum mechanics.
11.7 Pauli’s equation 67
11.8 Tool Box 2. Classical description of Angular Momentum

and its interaction with an external magnetic field.
Here we want to describe in detail some different, but related way to write a Hamiltonian for a classical
particle with angular momentum and electric charge. For that, let us start by saying that a Lagrangian for
a free particle Lfree and its corresponding Hamiltonian Hfree can be written in different ways according to
the coordinates you are using:
Lfree = 1
r˙ (t)||2
2m ||~ Lfree = 1
2 m(ṙ
2
+ ω2 )
1
π ||2 1 2 1 ~ · L.
~
Hfree = 2m ||~ Hfree = 2m πr + 2mr 2
L
r (t) = r(t)êr (t), ω = ||r(t)ê˙ r || the angular velocity, L

where ~ ~ the angular momentum vector and ~
π the
mechanical momentum given by
∂Lfree
πi = , (11.72)
∂ ẋi
where xi are generalized coordinates. For a 2D example, and by using polar coordinates we get that
∂Lfree ∂Lfree 2
πr = = mṙ, πθ = = mr θ̇, (11.73)
∂r ∂θ
with πθ being the effective angular momentum in 2D.
Exercise TB2a. Write the Hamiltonian in 3D in spherical and cylindrical coordinates. Then, give ex-
plicit expressions for the Angular Momentum vector.
Now, let us turn on an external magnetic potential A(~ ~ r ). An electric charged particle with a non-
zero velocity will interact with the external magnetic field B~ = ∇×A ~ deviating its geodesic straight-line
trajectory to a curved one due to the presence of the Lorentz force. Hence the Lagrangian contains a
new term describing the position-dependent potential the particle is immerse into. Similarly to the above
diagram, we have that in general,
LA = 1
r˙ (t)||2
2m ||~ r˙ · A
+ q~ ~ LA = 1
2 m(ṙ
2
+ ω2 ) + q
m~π ~
·A
L2 q2 ~ 2

2 ~ 2 q 1 ~ ~
HA = 1
2m π = 1
2m (~
p − q A) HA = 1
2m p2r + r2
+ 2m A − m pr A r + rA ·L
where the canonical momentum is

∂LA
pi = = πi + qAi . (11.74)
∂ ẋi
There are some notes we must emphasize before going forward:
1. The Hamiltonian in terms of the mechanical momentum is exactly the same Hfree = HA , however
π 6= p
since ~ ~ all interactions with the external field arise from expressing the Hamiltonian in terms
of the canonical momentum.
2. By using angular coordinates (spherical, cylindrical, polar, etc) we see that kinetic energy is essen-
tially given by a displacement on the radial coordinate and the presence of an angular momentum
~ due to a change in angular coordinates. In the presence of an external magnetic field B
L ~ this parts
continues describing the kinetic energy of the probe particle but now new terms arise showing the
interaction with the magnetic field. These terms can be split in the interaction between momentum
(linear and angular) with the corresponding component of A ~ and a term which goes as A ~ 2 . This last
one represents the energy contained in the external magnetic field affecting the particle’s dynamics.
With the purpose of clarify the above concepts, we shall concentrate on the case with a constant
~ This allows us to write
magnetic field B.
A~ = 1B~ ×~r. (11.75)
2
In this case
1 2 2 ~2
q
~ ,
HA = ~ +q A
p − ~·A
p
2m m
1 2 2 2 2
q ~ ~
= p
~ +q r B − B · L. (11.76)
2m 2m
~ 2 /r 2 and that we have used that

~2 = p2r + L
Also notice that p
p ~ ×~
~ · (B ~ · (~
r) = B r×p
~)
Let us now consider a simpler model of a particle moving on a 2D plane and immersed in an external
~ along z-axis. For that we are using polar coordinates. We observe that in this case
magnetic field B
1 2 2 q2 r2 B 2 q
H = (p + pz + )− BLz ,
2m r 4 2m
and the last quantity can be interpreted as the potential energy U for the probe particle, which is exactly
the same potential energy related to a magnetic dipole m,
~ this is
~ = −~
m ~
r · ∇U.
Notice also that the angular part of the momentum is only related to the polar angle θ since the potential
energy does not depend on z for which we do not expect a force acting on z-axis. This follows from Hamilton
equations which read
∂H q2 B 2
ṗr = − =− r,
∂r m
∂H
p˙z = = 0, (11.77)
∂z
where ṗr is the force along the radial direction, this is the centrifugal (centripetal) force. Observe that
the q/m rate establishes the radius of the trajectory depicted by the charged particle. Therefore the
quantity q 2 B 2 r 2 in the Hamiltonian refers to the kinetic energy due to the particle moving on a non-
straight trajectory. In general, we have a particle moving on a spiral along the z-axis. Its quantum version
is described by Landau levels.
Chapter 12
Addition of Angular Momentum
In the last chapter we have seen some particular properties for quantum states carrying a
spin s = 1/2. We are now going back to some general properties we can deduce from the
Angular Momentum’s formalism we have developed so far.
Our first question is to understand a basic difference between angular momentum of

a particle as a consequence of its dynamics and the intrinsic angular momentum they
carry. This last is called spin. Hence spin can be integer or half-integer and it refers to
the intrinsic angular momentum of an object, which is not related to the spinning around
its symmetry axis since we are considering a particle as a point-like object, but as the
generators of rotations in the Hilbert space.
The second point we want to study is the addition of angular momentum. In general
this means that we can have two different particles with spin s1 and s2 . All the available
states for each particle can be mixed up and formed new states as a pair of particles. In
other words, we are interested in the available states for the whole system.
12.1 Irreducible representations
Let us consider the matrix representation of a generic rotation operator D̂(R) in the basis
|j, mi. This is a representation of the group SO(3). Each matrix element is given by
(j) ˆ
Dm0 ,m (R) = j, m0 e−iθJ·n̂/~ |j, mi ,

(12.1)
for a given value of j. Matrix elements for different values of j vanish since [D̂(R), Jˆ2 ] = 0.
Hence by considering all possible values for j, the matrix representation of a rotation
69
70 Chapter 12 Addition of Angular Momentum
generator is given by
 
(0)
 Dm0 ,m 1×1 0 0 0 0 
 
(1/2)
0 Dm0 ,m 0 0 0
 
 
 2×2 
 
(1) , (12.2)

 0 0 Dm0 ,m 0 0 
 3×3 
 .. .. .. .. 

 . . . . 

 (j) 
Dm0 ,m
2j+1×2j+1
which clearly shows a block-diagonal form. It is precisely this result which shows that the
SO(3) representation is reducible. Each of the block represents an irreducible representa-
tion of D̂ since for a given value of j there are no invariant subspaces under the operator
Dˆ(j) . Irreducible representations of SO(3) are quantum labels for objects which carry
them. Such objects are called particles. Hence,
A fundamental particle is a physical quantum state corresponding to an irreducible

representation (irrep) of SO(3). There are particles with spin 0, 1/2, 1, 3/2, 2, ....
Exercise 12.1 Compute the matrix representations of D̂1/2 (θ, êz ) and D̂2 (θ, êz ).
12.2 Addition of Angular Momentum
Let us consider a system formed by two particles with angular momentum j1 and j2 . As
a whole system, can we associate some value for the total angular momentum? By "total
angular momentum" we mean the possible association of a single angular momentum to
this pair of particles with their corresponding eigenvalues.
We start by thinking about the meaning of a pair of particles at the quantum level.
Both particles have many different states related to their angular momentum. Particle 1
2
has 2j1 + 1 different states, all of them eigenstates of Jˆ1 and Jˆ1z , denoted by |j1 , m1 i,
with mi ∈ {−j1 , −j1 + 1, . . . , 0, . . . , j1 − 1, j1 }. Similarly, for Particle 2, we have the
2
states |j2 , m2 i, eigenstates of Jˆ2 and Jˆ2z , with m2 ∈ {−j2 , −j2 + 1, . . . , 0, . . . , j2 − 1, j2 }.
Hence, by consider both particles as a whole system, we need to consider all possible pairs
of eigenstates as shown in Table 12.1, where we have denoted the state |j1 , m1 ; j2 , m2 i as
the pair (|j1 , m1 i , |j2 , m2 i), this is
2
Jˆ1 |j1 , m1 ; j2 , m2 i = ~2 j1 (j1 + 1) |j1 , m1 ; j2 , m2 i ,
Jˆ1z |j1 , m1 ; j2 , m2 i = ~m1 |j1 , m1 ; j2 , m2 i ,
2
Jˆ2 |j1 , m1 ; j2 , m2 i = ~2 j2 (j2 + 1) |j1 , m1 ; j2 , m2 i ,
Jˆ2z |j1 , m1 ; j2 , m2 i = ~m2 |j1 , m1 ; j2 , m2 i . (12.3)
12.2 Addition of Angular Momentum 71
⊗ |j1 , −j1 i |j1 , −j1 + 1i ... |j1 , j1 − 1i |j1 , j1 i

|j2 , −j2 i |j1 , −j1 ; j2 , −j2 i |j1 , −j1 + 1; j2 , −j2 i ... |j1 , j1 − 1; j2 , −j2 i |j1 , j1 ; j2 , −j2 i
|j2 , −j2 + 1i |j1 , −j1 ; j2 , −j2 + 1i |j1 , −j1 + 1; j2 , −j2 + 1i . . . |j1 , j1 − 1; j2 , −j2 + 1i |j1 , j1 ; j2 , −j2 + 1i
.. ..
. .
|j2 , j2 − 1i |j1 , −j1 ; j2 , j2 − 1i |j1 , −j1 + 1; j2 , j2 − 1i . . . |j1 , j1 − 1; j2 , j2 − 1i |j1 , j1 ; j2 , j2 − 1i
|j2 , j2 i |j1 , −j1 ; j2 , j2 i |j1 , −j1 + 1; j2 , j2 i ... |j1 , j1 − 1; j2 , j2 i |j1 , −j1 ; j2 , j2 i
Table 12.1 All possible states related to two different particles with angular momentum j1 and
j2 . This defines the Tensorial Product ⊗.
Then, all possible states for the pair of particles are denoted through the Tensor Product
as defined in Table 12.1, i.e.
|j1 , m1 ; j2 , m2 i = |j1 , m1 i ⊗ |j2 , m2 i , (12.4)
and the corresponding Hilbert space is also given by the tensorial product of the corre-
sponding spaces: H = H1 ⊗ H2 . Notice that the space H contains (2j1 + 1) × (2j2 + 1)
states1 .
Once we have defined the Hilbert space we are dealing with and the specific quantum
states, it is time to define the general operators of Angular Momentum. This means that
all states in H ⊗ H must be also eigenvectors of another angular momentum operator
ˆ ˆ
defined through J~1 and J~2 . However, since we have constructed the states by applying a
tensorial product on the eigenstates of each particle, it is easy to see that it is convenient
ˆ
to define the angular momentum J~ acting on each state of H ⊗ H as
ˆ ˆ ˆ
J~ = J~1 ⊗ 1̂2 + 1̂1 ⊗ J~2 , (12.5)
where 1̂ is the identity operator in each Hilbert space, meaning that angular momentum
(or spin) operators, act trivially on the other subsystems. In consequence we have that
[Jˆ1i , Jˆ2j ] = 0, ∀i, j. (12.6)
Then since
ˆ ˆ
Jˆ2 = Jˆ12 + Jˆ22 + 2J~1 · J~2 , (12.7)
we find that
[Jˆ2 , Jâ,z ] = 0, for a = 1, 2. (12.8)
This is quite important since we can add an extra compatible observable to the four
observables we had. However the states |j1 , m1 ; j2 , m2 i are not eigenvectors of Jˆ2 , for
which if we want to consider it as an observable, we must change our eigenvector basis.
1
Also be aware that this space is not H ⊕ H. Why?
12.2.1 Addition of two 1/2 spin particles

Let us consider two particles with spin 1/2. The whole set of eigenvectors with respect to
2 and Ŝ
the spin operators Ŝ1,2 (1,2)z are
|1/2, m1 ; 1/2, m2 i = {|++i , |+−i , |−+i , |−−i}. (12.9)
Now, define
Jˆz = Jˆ1z + Jˆ2z , (12.10)
where the sum must be interpreted as the formal tensor product previously defined. In
terms of this operator,
~
Jˆz |++i = Jˆ1z |++i + Jˆ2z |++i = (|++i + |++i) = ~ |++i ,
2
~
Jˆz |−−i = Jˆ1z |−−i + Jˆ2z |−−i = − (|−−i + |−−i) = −~ |−−i ,
2
~
Jˆz |+−i = Jˆ1z |+−i + Jˆ2z |+−i = (|+−i − |+−i) = 0,
2
~
Jˆz |−+i = Jˆ1z |−+i + Jˆ2z |−+i = (− |−+i + |−+i) = 0. (12.11)
2
Addition of angular momentum means the possibility to re-arrange these states into spin
irrep. As we see, there are 4 states characterized by the spin projection eigenvalue of J, ~ˆ
this is 2 states with m = 0, one state with m = 1 and one state with m = −1. These
states can be associated to two different spin irrep. For j = 0, m = 0 there is a single
states, while 3 states with values for m equal to −1, 0, 1 correspond to irrep for j = 1. In
order to be sure of that, we have to construct explicitly the states that are eigenvector of
the operator Jˆ2 exhibiting the above properties.
Exercise 12.2. It is easy to check that the state

1
√ (|+−i − |−+i),
2
has an eigenvalue j = 0 and m = 0, while
1
|−−i , √ (|−+i + |+−i) , |++i ,
2
have eigenvalues j = 1 and m = −1, 0, 1, respectively.
From the above one can notice that by taking all possible spin states for two particles
with s = 1/2 the set can be arrange into irrep for different spins. This means that it is
possible to add spin states saying that2
1 1
⊗ = 0 ⊕ 1. (12.12)
2 2
2
There is an alternative notation which counts states in each representation. Each spin 1/2 particle
has two-states and it is representated by the symbol 2. Hence, we can also say that 2 ⊗ 2 = 1 ⊕ 3.
12.3 Clebsch-Gordon Coefficients 73
ˆ
Another way to see this is by appliying the operator rotation generated by J~ to the
generic state 2! , m1 ; 12 , m2 . This is

ˆ
D̂zj (2π) |m1 , m2 i = e−2iπJz /~ |m1 , m2 i ,
= ei(±1±1)π |m1 , m2 i . (12.13)
Hence we can see that a 2π rotation always keep invariant the state |m1 , m2 i showing
that acts as a state with integer spin.
1
Exercise 12.3. Compute 1 ⊗ 2 and 1 ⊗ 1.
Exercise 12.4. Prove that Jˆ2 , Jˆz , Jˆ12 and Jˆ22 form a set of commutable observables.
12.3 Clebsch-Gordon Coefficients
So far, a ket state representing a quantum state for a two-particle system is generically
given by |j1 , m1 ; j1 , m2 i, which explicitly tells us that there are 4 different quantum
labels (eigenvalues) j1 , m1 , j2 and m2 related to 4 different and commutable observables
Jˆ12 , Jˆ22 , Jˆ1z and Jˆ2z . We also have seen that we can construct two extra observables for
this system given by Jˆ2 and Jˆz2 . However the ket |j1 , m1 ; j2 , m2 i is not a proper vector of
the last two observables. Instead we can prove (see Excercise 11. ) that Jˆ2 , Jˆz , Jˆ12 and
Jˆ22 are also a set of commutable operators, for which we can select a set of eigenvectors.
We shall denote such states as |j, m; j1 , j2 i and
Jˆ2 |j, m; j1 , j2 i = ~2 j(j + 1) |j, m; j1 , j2 i ,

Jˆz |j, m; j1 , j2 i = ~m |j, m; j1 , j2 i ,
Jˆ2 |j, m; j1 , j2 i = ~2 j1 (j1 + 1) |j, m; j1 , j2 i ,
1
Jˆ22 |j, m; j1 , j2 i = ~2 j2 (j2 + 1) |j, m; j1 , j2 i . (12.14)
This new basis of eigenvectors can be transformed into the previous basis through
a matrix C. For given values of angular momentum, matrix elements are called the
Clebsch-Gordon coefficient (CG), and are explicitly given by
m
Cm 1 m2
= hj1 , m1 ; j2 , m2 |j, m; j1 , j2 i , (12.15)
such that X
m
|j, m; j1 , j2 i = Cm 1 m2
|j1 , m2 ; j2 , m2 i . (12.16)
m1 ,m2
Lema. The matrix C satisfies the following properties:

m
1. Cm = 0 unless m = m1 + m2 .
1 m2
m
2. Cm = 0 unless |j1 − j2 | ≤ j ≤ j1 + j2 .
1 m2
Proof. Since
Jˆz = Jˆ1z + Jˆ2z ,
it follows that
(Jˆz − Jˆ1z − Jˆ2z ) |j, m; j1 , j2 i = 0.
Then (m − m1 − m2 )C = 0 from which it follows that C vanishes unless m = m1 + m2 .
Exercise 12.5. To prove assertion number 2, notice at first that our goal is to compare
the angular momentum value j with j1 and j2 . Consider the angular momentum operators
as vectors and use the triangular inequality to prove the statement.
Notice that if j1 and j2 are both half-integer spins, the resulting angular momentum
addition has integer values.
12.3.1 Properties of the Clebsch-Gordon coefficients

In order to deduce some important recursion relations among CG coefficients, let us define
the following operators:
Jˆ± = Jˆ1± ⊗ 12 + 11 ⊗ Jˆ2± , (12.17)
where
Jâ± = Jâx ± iJây , a = 1, 2. (12.18)
It is straightforward to prove that
Jˆ1± |j1 , m2 ; j2 , m2 i = ~ ((j1 ∓ m1 )(j1 ± m1 + 1))1/2 |j1 , m1 ; j2 , m2 i . (12.19)
For that, and since [Jˆz , Jˆ± ] = ±~Jˆ± , we have
Jˆ1z Jˆ1+ |j1 , m1 ; j2 , m2 i = (~Jˆ1+ + Jˆ1+ Jˆ1z ) |j1 , m1 ; j2 , m2 i ,

= ~(m1 + 1)Jˆ1+ |j1 , m1 ; j2 , m2 i , (12.20)
therefore, and by assuming there is no degeneracy on m eigenvalues, we get that
Jˆ1+ |j1 , m1 ; j2 , m2 i = κ |j1 , m1 + 1; j2 , m2 i . (12.21)
for a κ ∈ C. Normalizing the state and using that
Jˆ− Jˆ+ = Jˆ2 − Jˆz2 − ~Jˆz , (12.22)
one obtains that
κ2 = ~2 j1 (j1 + 1) − m21 − m1 ,

= ~2 (j12 + j1 − m21 − m1 + j1 m1 − j1 m1 ),
= ~2 (j1 − m1 )(j1 + m1 + 1), (12.23)
from which Ec. (12.19) follows. Similar expressions hold for Jˆ1− and Jˆ2± .
12.3 Clebsch-Gordon Coefficients 75
Jˆ± |j1 , m1 ; j2 , m2 i = ((j ∓ m)(j ± m + 1))1/2 |j1 , m1 ; j2 , m2 i .
Now, since we can change the basis on which we are acting the operators, it is possible
to write
X
Jˆ± |j, m : j1 , j2 i = Cm1 m2 Jˆ± |j1 , m1 ; j2 , m2 i
m1 ,m2
X
= Cm1 m2 (Jˆ1± + Jˆ2± ) |j1 , m1 ; j2 , m2 i
m1 ,m2
X hp
= (j1 ∓ m1 )(j1 ± m1 + 1)Cm1 m2 |j1 , m1 ± 1; j2 , m2 i +
m1 ,m2
p i
+ (j2 ∓ m2 )(j2 ± m2 + 1)Cm1 m2 |j1 , m1 ; j2 m2 ± 1i , (12.24)
from which it is possible to show the recursion relation

p m±1
p m
(j ∓ m)(j ± m + 1)Cm 1 m2
= (j1 ∓ m1 + 1)(j1 ± m1 )Cm 1 ∓1,m2
+
p m
+ (j2 ∓ m2 + 1)(j2 ± m2 )Cm 1 ,m2 ∓1
, (12.25)
with m1 + m2 = m ± 1. Notice that the operator Jˆ± has shifted m to m ± 1. In this

sense Jˆ± relates three different Clebsch-Gordon coefficients as in the next diagram:
J+
Cm1 −1,m2 Cm1 ,m2 Cm1 +1,m2
J+ J−
Jˆ+
Cm1 ,m2 −1 Cm1 ,m2 +1 Cm1 +1,m2
Exercise 12.7. Prove Equation 12.25.
Exercise 12.8. Compute Clebsch Gordon coefficients and show that the above relations
hold for j1 = j2 = 1/2.
Chapter 13
Hydrogen atom
One of the most intriguing questions at the beginning of the XX century was the proper
existence of the atom. According to classical mechanics, a positive and a negative elec-
tric charged particles feel an attraction due to the electromagnetic force. If the less
charged particle also has a relative angular momentum, then it would be possible to
explain the physics behind the atom, exactly as a solar system by using gravitational force.
However there is, at least one extra factor in the interaction through the electromag-
netic force which makes it different from the classical description of gravitation. This
is radiation. An electrical charged particle loses energy if it is accelerated. The energy
flows in form of electromagnetic fields propagating with some definite patterns, away
from the source. Besides those important patterns1 , the fact that a particle has angular
momentum such that can be confined to move around the positive charge, implies that it
loses energy. Therefore it must gradually collide to the other particle making the atom to
be an unstable entity.
In this chapter we shall apply all our acquired knowledge on quantum phjysics to
the simplest version of the atom: the hydrogen atom, consisting on one single negative
charged particle and a single one withy positive charge.
13.1 Reducing the system
Consider the Lagrangian of a couple of particles under electrical force, and angular
momentum:
1 1 1 ~ × P~ )2 + 1 (~r × p~)2 − V (R,
~ ~r),
L = M VR2 + mvr2 + (R (13.1)
2 2 2M R2 2mr2
where
~
R(t) = R(t)êr , ~r(t) = r(t)êr , (13.2)
1
Some fields behave as E ∼ 1/r such that the energy goes like E 2 ∼ 1/r2 , which once integrated over
a sphere of radius r becomes independent of r itseld, since d3 x = r2 sin θ. Fields of this type can be
detected at large distances r. They are called radiation.
76
13.2 Quantum mechanical description 77
where R ~ and ~r denote the position vectors of the two particles with velocities V ~ =R ~˙
and ~v = ~r˙ . Now consider a central potential, this is a potential that only depends on the
distance between the two particles. In that case the system can be reduced to a single
particle moving with respect to the center of force acting on it2 ,
1 L2 qQK
L = mṙ2 + 2
+ , (13.3)
2 2mr r
where µ is the reduced mass µ = mM/m + M , r is the relative distance |R ~ − ~r| and
K is the Coulomb’s constant on the potential energy indicating the electrical potential
between the particles with charge q and Q. Notice that this is expected by assuming that
m is much shorter than M such that it is a probe particle. In that case the center of
force coincides with the position of the particle of mass M which remains almost at rest,
implying that the dynamics of the particle is only limited to describe the movement of
the particle of mass m.
13.2 Quantum mechanical description
Let us now write the Hamiltonian operator constructed by promoting the classical
Hamiltonian function derived from the above Lagrangian:
P̂r2 L̂2
Ĥ = + + V (r̂), (13.4)
2µ 2µr2
where in spherical coordinates (see Tool Box 2),
2
p̂2φ
L̂ = p̂2θ + , (13.5)
sin2 θ
from which we can define the Hamltonian in terms of an effective potential as
P̂r2
Ĥ = + Veff (r̂), (13.6)
2µ
with Veff = L̂2 /2µr2 + V (r). Hence we want to solve the eigenvalue equation
Ĥ |Ψi = E |Ψi . (13.7)
For that it is necessary to identify the complete set of commutable observables which let
us uniquely identify each eigenstate in this Hilbert space. By the expression of the Hamil-
tonian it is easy to see hat such a set is conformed by the operators Ĥ, L̂2 and Lˆz . Since
we are not considering an external magnetic field, observe that interaction of angular mo-
mentum or spin with such a field is not playing a role in this description. Also notice that
we are ignoring the term Ŝ 2 . This is because all particles, actually electrons, have spin 1/2.
2
Notice that the force diverges at this point making difficult to describe the interaction even at
classical level. We shall make some assumption to evade a concrete description of the force at that point.
78 Chapter 13 Hydrogen atom
Let us then denote the corresponding eigen state as
|Ψi = |n, l, mi , (13.8)
such that
Ĥ |n, l, mi = En |n, m, li ,
L̂2 |n, l, mi = ~2 l(l + 1) |n, l, mi ,
L̂z |n, l, mi = ~m |n, l, mi . (13.9)
with, as we know, m = −l, . . . , l.
13.2.1 Projecting into the configuration space

In contrast to the Harmonic Oscillator, the equation to be solved seems to be more
tractable in the configuration space than in the Hilbert space. Hence, let us project the
Schrödinger equation into the configuration space, where
Ψ(r, θ, φ) = h~r|Ψi , (13.10)
such that the equation reads

!
P̂r2 L̂2
h~r| Ĥ |Ψi = h~r| + + V (r̂) |Ψi ,
2µ 2µr2
2
~ 1 ∂2 L2

= − r+ + V (r) Ψ(r, θ, φ),
2µ r ∂r2 2µr2
= En Ψ(r, θ, φ), (13.11)
where we have used that

1 ∂2 ∂2 1 ∂2

2 1 1 ∂
∇ = r+ 2 + + . (13.12)
r ∂r 2 r ∂θ2 tanθ ∂θ sin2 θ ∂φ2
Observe this allows us to separate the wave function Ψ(r, θ, φ) as
Ψ(~r) = Rn (r)Θm,l (θ)Φm,l (φ), (13.13)
implying we must solve three different differential equations. It is also important to

notice that since we have a bounded potential, due mainly by the presence of an angular
momentum, we expect to have a discrete energy spectrum. Because of this, we can assign
a label n to the energy eigenvalue.
13.3 Radial equation
The differential equation we must solve is

2
~ 1 d2 l(l + 1)~2

− r+ + V (r) R(r) = En R(r) (13.14)
2µ r dr2 2µr2
13.3 Radial equation 79
Although we have used an expected index n, we can see that solution for R(r) must
depend at least on two indexes. One is clearly l and by considering the possibility to
have degeneracy, we shall use another one, labeled k. Latero on we shall realize a relation
among these two indexes to n.
13.3.1 Limiting our ignorance

In these spherical coordinates, the center of force is located at r = 0. What do you expect
from a wave function to behave at this point? How much force it "feels"? As we approach
to r = 0 the force diverges making our quantum tools to be insufficient to describe the
atom’s nature. Quantum mechanics, at the non-relativistic limit, does not allows to
study the corresponding phenomena. We then have two ways to proceed: stop trying to
understand the atom till we are able to have a quantum description of the electromagnetic
field which implies to have a relativistic quantum mechanics as well, or to describe the
phenomena in an effective way, meaning that we can encapsule our ignorance around
r = 0 to some probably false, but useful condition which allows to study the rest of the
atom. We shall take the second approach since we have learned over the years that this
way to study natural phenomena has been proved to be a useful way to understand nature.
Let us then assume that R(r) goes like rs as r approaches to 0, expecting to have a
finite description of R at r = 0. After substituting into the radial equation we have that
the only consistent solution is
R(r) ∼ rl , for r → 0. (13.15)
Some steps follow:
1. Change to new functions: Rk,l (r) = 1r uk,l (r),
2. uk,l (r) ∼ rl+1 for r near to 0,
3. uk,l (0) = 0.
The differential equation now reads
2 2
l(l + 1)~2 e2

~ d
− + − uk,l (r) = Ek,l uk,l (r). (13.16)
2µ dr2 2µr2 r
Still, the equation seems to be difficult to be solved. We perform then the following
variable definition:
µe2
ρ = r,
~2
1/2
2~2 Ek,l

λk,l = − (13.17)
µe4
after which the differential equation is
2
d l(l + 1) 2 2
− + − λk,l uk,l (ρ) = 0, (13.18)
dρ2 ρ2 ρ
13.3.2 Solution to Radial Equation

To solve the above differential equation , we proceed as follows:
1. Asymptotic behavior: Equivalently to the case for ρ approaching to 0, we expect

a finite solution for large values to ρ, this is for ρ → ∞. In that case the second
and third terms in the differential equation are negligible and the equation can be
approximated to 2
d 2
− λk,l uk,l (ρ) = 0, (13.19)
dρ2
yielding to the general solution given by
uk,l (ρ) = e−ρλk,l yk,l (ρ), (13.20)
where yk,l is a function such that
lim uk,l = 0. (13.21)

ρ→∞
2. After substituting this solution into the differential equation (13.18), we can consider
a solution in the form of power series, as
∞
X
yk,l (ρ) = ρs cq ρq , (13.22)
q=0
with yk,l (0) = 0.

3. The above series gives us the condition
(−l(l + 1) + s(s + 1))c0 = 0, (13.23)
where c0 is the first non-zero coefficient in the series. This implies that s can be l + 1
or −l, however only for s = l + 1 we have a finite solution at the origin. Therefore
the solution is
1
Rk,l (r) = uk,l (r),
r
1/2
1 − − 2µE

k,l
2 r
= e ~ yk,l (ρ),
r
1/2 ∞ l+1+q
1 − − 2µE µe2

k,l
r l+1
X
= e ~2 r cq rq ,
r ~2
q=0
2µE 1/2 ∞ l+1+q
− − k,l
r
X µe2
= e ~2
cq rq+l . (13.24)
~2
q=0
4. The second derivative in the differential equation leads to the condition
q(q + 2l + 1)cq = 2 ((q + l)λk,l − 1) cq−1 , (13.25)
from which we can fix all cq once c0 is fixed.

13.3 Radial equation 81
Still we have not encounter a physical significance of the label k. This however comes
out in a very important physically way. Let us say that all terms in the series solution
are nonzero, implying that
cq 2λk,l
∼ , for large q. (13.26)
cq−1 q
However since for the function e2ρλk,l the q-th and q − 1-th terms in the Taylor’s series
have a similar result, i.e, we can write
X (2λk,l )q
e2ρλk,l = ρq , (13.27)
q
q!
we see that having all terms non-zero in the series is equivalent to have a positive expo-
nential solution, making our solution to be finite at large radius. This is not acceptable
for which one concludes that there exists a value of q at which ell subsequent terms ion
the series solution, vanish. Let us denote such value precisely by k, this is, for q ≥ k, all
cq are null.
By the recurrence relation among Cq coefficients, we have that

1
λk,l = ,
k+l
and then
µe2 1
Ek,l = − . (13.28)
2~2 (k + l)2
Since k and l are integers, we conclude that the energy levels are quantized, and by
defining
n = k + l, (13.29)
we have that
E0
En = − , (13.30)
n2
µe2
with E0 = 2~2
.
13.3.3 Examples
Using the recursion relations, it is easy to show that
−2 2 (k − 1)!

(2l + 1)!
cq = c0 . (13.31)
k+l (k − q − 1)! q!(q + 2l + 1)!
From this we can explicitly compute the radial functions R(r). For instance for k = 1,
l = 0, we have
2 3/2
µe 2 2
R1,0 = 2 2
e−µe r/~ . (13.32)
~
Notice that |R(r)|2 goes to zero as r grows keeping a spherical symmetry. The important
thing to note (mainly because commonly is misinterpreted) is that the probability to find
teh electron in any region of space is different from zero in the radial direction. This will
be important to remember at the end of this chapter.
Observe that
E1 = E1,0 = −E0 , (13.33)
and there is a single state at this energy level. However, for n = 2, we can have at least
to options with values for k and l, this is k = 2, l = 0 and k = 1, l = 1. But this is not all
possible degeneracy we have, since l has also associated different states. For l = 1 there
are 3 mor states with m = −1, 0, 1. Hence we have 4 different states related to n = 2
overall. Total degeneracy is given by
n−1
X
gn (2l + 1) = n2 . (13.34)
l=0
13.4 Angular equation
Let us go back to the differential equation (13.11). The angular component can be written
as
L̂2 ~2 l(l + 1)
h~r| |Ψi = h~r|Ψi , (13.35)
2µr2 2µr2
which can be reduce to
2
1 ∂2

1 1 ∂ 1 ∂ l(l + 1)
2 2 2
+ + 2 2
Yl,m (θ, φ) = ~2 Yl,m (θ, φ), (13.36)
2µr r ∂θ tanθ ∂θ sin θ ∂φ 2µr2
where
h~r|Ψi = hr, θ, φ|Ψi ,

= hr|Ψi hθ, φ|Ψi ,
= R(r)Ym,l (θ, φ), (13.37)
and Ym,l are called Spherical Harmonics. Notice that Yl,m = Θ(θ)Φ(φ). Since
|Ψi = |n, l, mi = |ni ⊗ |l, mi , (13.38)
it is easy to see that actually
Yl,m (θ, φ) = hθ, φ|l, mi . (13.39)
Properties of Spherical Harmonics are easily followed from the orthonormality of the
corresponding kets |m, li.
Exercise. Deduce all properties of Spherical Harmonics.

13.4 Angular equation 83
Now, since L̂z |l, mi = ~m |l.mi, after projecting into the configuration space, we find
that
hθ, φ| L̂z |l, mi = m~Yl,m (θ, φ), (13.40)
which translates into the differential equation
∂
−i Yl,m (θ, φ) = ~mYl,m (θ, φ). (13.41)
∂φ
A solution is given by
Yl,m (θ, φ) = Θ(θ)l,m eimφ , (13.42)
with the function θ being a suitable function such that Ψ is squared-integrable. The
following are some steps which allows us to construct the Spherical Functions:
1. Consider L̂± |l, mi = C± |l, m ± 1i. From this we get that L̂+ |l, li = 0 and the
corresponding differential equation leads to the conclusion that
Θ(θ)l,l = cl (sin θ)l . (13.43)
2. The coefficient C± can be fixed after normalizing the functions, such that
p
[L± ]Yl,m (θ, φ) = ~ l(l + 1) − m(m ± 1)Yl,m (θ, φ), (13.44)
where [L± ] is the differential operator representing the operator L̂± in the configu-
ration space.
3. After solving the corresponding differential equation, we have that

s
(−1)l 2l + 1 (l + m)! imφ dl−m
Yl,m (θ, φ) = l e (sin θ)−m (sin θ)2l . (13.45)
2 l! 4π l − m)! d(cos θ)l−m
This completely solves our initial equation, Ĥ |Ψi = E |Ψi. The wave function is then
given. As an example, let us consider the first energy level, this is for n = 1. In that case,
as we have seen, there is a single case with l = 0. Then, since m = 0, the wave function
is given by
2 3/2
1 µe 2 2
Ψ1,0,0 (r, θ, φ) = √ 2
e−µe r/~ . (13.46)
π ~
This is the wave function related to the first orbital. Notice the spherical symmetry, from
which one concludes that the probability to find the electron with these quantum numbers
is
1 µe2 −2µe2 r/~2
P(1, 0, 0) = e , (13.47)
π ~2
representing a sphere with radius P. It is important to notice that this sphere has not a
boundary on r but it is non-zero for all r. The orbital (1, 0, 0) is extended over all space.
However the maximum of density of probability occurs at r0 with
d(Pr2 )
(r0 ) = 0 (13.48)
dr
corresponding to
~2
r0 = . (13.49)
µe2
This is the first orbital, with n = 1, l = 0, m = 0 with a spectroscopic notation given by
1s2 (where s2 means two spin states in each level).
Appendix A
Some notes on Lie Groups
Let us start by defining a Group. A group is a set G of elements with an operation ∗

satisfying the following1 :
1. For g1 and g2 ∈ G, then g1 ∗ g2 ∈ G for all g ∈ G.
2. There exists a unique e ∈ G such that g ∗ e = e ∗ g = g, ∀g ∈ G.
3. For each g ∈ G there exists a unique g −1 such that g ∗ g −1 = e.
4. If g1 ∗ g2 = g2 ∗ g1 we say thet G is an Abelian Group.
Among the set of groups, there is a subset of particular interest in Physics called Lie
Groups. Lie groups have a very interesting property which is quite useful in particle
physicas as well as in General Relativity: the group is homeomorphic to a differentiable
manifold. If you are not familiar with this mathematical concepts, is enough to know
that you can identify a Lie Group with some soft surface of any dimension.
Example 1. The Lie group U (1) clearly can be mapped into the circle S1 , since
g = eiα = x + iy where x2 + y 2 = 1.
Example 2. Consider the Lie Group SU (2) defined as
SU (2) = {A2×2 (C)|AA† = I, detA = 1}. (A.1)
Since we are dealing with 2 × 2 complex matrix we have in principle 8 real variables
(physically speaking, we have 8 degrees of freedom). However we have two restrictions:
AA† = I and det A = 1. Altogether establish 3 complex restrictions or 6 real constraints
implying that we can write any matrix in SU (2) as

a b
(A.2)
−b∗ a∗
1
This means that the set G is a group wrt to the operation ∗. The same set can be also a group wrt
to another operation.
85
86 Appendix A Some notes on Lie Groups
with a and b complex numbers and such that aa∗ +bb∗ = 1. Writing these complex numbers
as a = x + iy and b = v + iw, the condition det A = 1 is written as x2 + y 2 + v 2 + w2 = 1
which is the equation of a 3-dimensional sphere S3 ∈ R4 . Hence we say that SU (2) is
homeomorphic to S3 or simply
SU (2) ' S3 . (A.3)
~
Exercise. Show that if g ∈ SU (2) is written as g = ei(~α·A) with Ai matrices, then
T r Ai = 0.
Appendix B
Stern-Gerlacht experiment
87
Appendix C
Transformations and unitary generators
88
Appendix C Transformations and unitary generators 89
Symmetries
Symmetry Operator Generator Group
−i P̂ x i /~
Translation e i ~x U (1)
Peugeot 508 Berlina Gasolina
Chrysler Voyager Monovolumen Gasolina
Land Rover Defender Todoterreno Gasolina
Table C.1 Coches disponibles
Appendix D
Probability currents
90

QM2020

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

QM2020

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

QM2020

Uploaded by

Copyright:

Available Formats

Quantum Mechanics I

August 13, 2020

This is a serie of lectures on advanced quantum mechanics for one semester

4.3.2 Uncertainty on measuring position and momentum . . . . . . . . . 21

8 Gauge Transformations in Quantum Mechanics 36

9 Gauge quantum effects 39

12 Addition of Angular Momentum 69

A Some notes on Lie Groups 85

C Transformations and unitary generators 88

Estas notas son el resultado de haber impartido el curso de Mecánica Cuántica 1, en

Los temas que se verán son los siguientes:

• Modern Quantum Mechanics, J.J. Sakurai.

• Quantum Mechanics Vol. 1, Cohen-Tannoudji

• The Theory of Spinors, Cartan.

• Lectures on Quantum Mechanics, S. Weinberg.

• Quantum Mechanics, Susskind and Friedman

• Curso Abreviado de Física Teórica, Libro 2, L. Landau y E. Lifshitz.

Topics you should know before taking this course!

• Experiments indicating that classical mechanics is not a suitable framework to

• Borh’s model of hydrogen atom

• Physical meaning of the uncertainty principle.

• Physical motivation for the wave-particle dual behaviour of matter.

1.1 A brief review on wave mechanics

to consider a bunch of particles with different momentum associated. In the language of

1.2 Bounded states

Another interesting issue in quantum mechanics, comes by discretization of certain

1.3 Into a mathematical framework for quantum mechanics

1.3.1 Physical States

hψ| −→ (ψ1∗ , ψ2∗ , . . . ψn∗ ) (1.13)

1.4 Exercises, due to T2-day

2.1 Physical states

A quantum particle or object is described in terms of a wave function Ψ(x) or equivalently

quantum particle −→ Ψ(x) −→ |Ψi . (2.1)

1. There exist the summation operation + such that for |i ∈ H

• (c + d) |ai + (c + d) |bi = (c + d)(|ai + |bi) ∈ H, for c, d ∈ C.

Definition 1. A basis {|an i} of H is a set of vectors, with n = 1, . . . , N such that all

If the above vector vanishes, then λn ∈ C vanishes for all n.

Notice this defines the basis for a vector space of dimension N .

2.1.1 Dual vector space H∗

ha| (|bi) = λ ∈ C. (2.3)

We call these mappings functionals of H. If these functionals have the following

1. (ha| + hb|)(|vi) = ha| (|vi) + hb| (|vi),

2. hca| (|dv + wi) = c∗ d ha| (|vi) + c∗ ha| (|wi),

with c, d ∈ C, we define the set of linear functionals

{ha| | ha| : H → C}, (2.4)

as the dual space H∗ of H.

ha| · hb| = hb| (|ai) = hb|ai . (2.7)

ha|bi = hb|ai∗ . (2.8)

hb| = (σ1 , σ2 , . . . , σN ) (2.10)

Consider now the state |ai = N

2.1.2 Clousure relation

|ai hb| (|ci) = hb|ci |ai = λ |ai , (2.16)

with λ = hb|ci ∈ C, this is |ai hb| : H → H, which is also a linear map.

Operators satisfy the following properties:

1. (Â + B̂) |ai = Â |ai + B̂ |ai,

2. Â(c |ai + |bi) = cÂ |ai + Â |bi,

Using these properties we can obtain a representation of an operator Â in terms of a given

Â |aj i = IˆÂ |aj i ,

3. [L̂i , L̂j ] = iijk L̂k ,