Professional Documents
Culture Documents
(under construction)
Peter Woit
Department of Mathematics, Columbia University
woit@math.columbia.edu
Preface xi
iii
4.6 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 40
iv
9.4.3 Some examples . . . . . . . . . . . . . . . . . . . . . . . . 104
9.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 105
13 Quantization 141
13.1 Canonical quantization . . . . . . . . . . . . . . . . . . . . . . . . 141
13.2 The Groenewold-van Hove no-go theorem . . . . . . . . . . . . . 144
13.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 145
v
16 The Harmonic Oscillator as a Representation of the Heisenberg
Group 169
16.1 Complex structures and quantization . . . . . . . . . . . . . . . . 170
16.2 The Bargmann transform . . . . . . . . . . . . . . . . . . . . . . 173
16.3 Coherent states and the Heisenberg group action . . . . . . . . . 174
16.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 176
vi
21.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 220
vii
28 Representations of the Lorentz Group 285
28.1 Representations of the Lorentz group . . . . . . . . . . . . . . . . 285
28.2 Dirac matrices and Cliff(1,3) . . . . . . . . . . . . . . . . . . . 290
28.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 294
A Conventions 335
viii
B Problems 339
B.1 Problem Set 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
B.2 Problem Set 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
B.3 Problem Set 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
B.4 Problem Set 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
B.5 Problem Set 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
B.6 Problem Set 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
B.7 Problem Set 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
B.8 Problem Set 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
B.9 Problem Set 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
ix
x
Preface
These are the course notes prepared for a class taught at Columbia during the
2012-13 academic year. The intent was to cover the basics of quantum me-
chanics, from a point of view emphasizing the role of unitary representations of
Lie groups in the foundations of the subject. The approach to this material is
simultaneously rather advanced, using crucially some fundamental mathemati-
cal structures normally only discussed in graduate mathematics courses, while
at the same time trying to do this in as elementary terms as possible. The
Lie groups needed are relatively simple ones that can be described purely in
terms of small matrices. Much of the representation theory will also just use
standard manipulations of such matrices. The only prerequisite for the course
was linear algebra and multi-variable calculus. My hope is that this level of
presentation will simultaneously be useful to mathematics students trying to
learn something about both quantum mechanics and representation theory, as
well as to physics students who already have seen some quantum mechanics,
but would like to know more about the mathematics underlying the subject,
especially that relevant to exploiting symmetry principles.
This document will continue to be updated as time permits with errors fixed,
notation improved, minor additions of material, and whatever other improve-
ments I can think of, with the latest version available at
http://www.math.columbia.edu/~woit/QM/qmbook.pdf
Corrections, comments and suggestions about this are warmly encouraged,
with the best way to contact me email to woit@math.columbia.edu
xi
xii
Chapter 1
1.1 Introduction
A famous quote from Richard Feynman goes I think it is safe to say that no
one understands quantum mechanics.[16]. In this course well try to do the
best we can to come to grips with the subject, emphasizing its deep connections
to fundamental ideas and powerful techniques of modern mathematics. The
strangeness inherent in quantum theory that Feynman was referring to has two
rather different sources. One of them is the inherent disjunction and incom-
mensurability between the conceptual framework of the classical physics which
governs our everyday experience of the physical world, and the very different
framework which governs physical reality at the atomic scale. Familiarity with
the powerful formalisms of classical mechanics and electromagnetism provides
deep understanding of the world at the distance scales familiar to us. Supple-
menting these with the more modern (but still classical in the sense of not
quantum) subjects of special and general relativity extend our understanding
into other less accessible regimes.
Read in context though, Feynman was pointing to a second source of diffi-
culty, contrasting the mathematical formalism of quantum mechanics with that
of the theory of general relativity, a supposedly equally hard to understand
subject. General relativity can be a difficult subject to master, but its math-
ematical and conceptual structure involves a fairly straight-forward extension
of structures that characterize 19th century physics. The fundamental physical
laws (Einsteins equations for general relativity) are expressed as partial differ-
ential equations, a familiar if difficult mathematical subject. The state of the
system is determined by the set of fields satisfying these equations, and observ-
able quantities are functionals of these fields. The mathematics is just that of
the usual calculus: differential equations and their real-valued solutions.
In quantum mechanics, the state of a system is best thought of as a different
sort of mathematical object: a vector in a complex vector space, the so-called
state space. One can sometimes interpret this vector as a function, the wave-
1
function, although this comes with the non-classical feature that wave-functions
are complex-valued. Whats truly completely different is the treatment of ob-
servable quantities, which correspond to self-adjoint linear operators on the state
space. This has no parallel in classical physics, and violates our intuitions about
how physics should work.
During the earliest days of quantum mechanics, the mathematician Hermann
Weyl quickly recognized that the mathematical structures being used were ones
he was quite familiar with from his work in the field of representation theory.
From the point of view that takes representation theory as a fundamental struc-
ture, the framework of quantum mechanics looks perfectly natural. Weyl soon
wrote a book expounding such ideas[52], one which got a mixed reaction from
physicists unhappy with the penetration of unfamiliar mathematical structures
into their subject (with some of them referring to it as the Gruppenpest, the
group theory plague). One goal of this course will be to try and make some of
this mathematics as accessible as possible, boiling down Weyls exposition to its
essentials while updating it in the light of many decades of progress and better
understanding of the subject.
Weyls insight that quantum mechanics crucially involves understanding the
symmetry groups of a physical system and the unitary representations of these
groups has been vindicated by later developments which dramatically expanded
the scope of these ideas. The use of representation theory to exploit the sym-
metries of a problem has become a powerful tool that has found uses in many
areas of science, not just quantum mechanics. From this course I hope that
students whose main interest is physics will learn to appreciate the mathemat-
ical structures that lie behind the calculations of standard textbooks, helping
them understand how to effectively exploit them in other contexts. Students
whose main interest is mathematics will hopefully gain some understanding of
fundamental physics, at the same time as seeing some crucial examples of groups
and representations. These should provide a good grounding for appreciating
more abstract presentations of the subject that are part of the standard mathe-
matical curriculum. Anyone curious about the relation of fundamental physics
to mathematics, and what Eugene Wigner described as The Unreasonable Ef-
fectiveness of Mathematics in the Natural Sciences[53] should benefit from an
exposure to this remarkable story at the intersection of the two subjects.
The following sections give an overview of the fundamental ideas behind
much of the material well be covering in the class. They may initially be rather
mystifying, but should become much clearer as we work through basic examples
this semester.
2
built on the following simple axioms:
Axiom (States). The state of a quantum mechanical system is given by a vector
in a complex vector space H with Hermitian inner product < , >.
Well review in an early class some linear algebra, including the properties
of inner products on complex vector spaces. H may be finite or infinite dimen-
sional, with further restrictions required in the infinite-dimensional case (e.g.
we may want to require H to be a Hilbert space). Note two very important
differences with classical mechanical states:
The state space is always linear: a linear combination of states is also a
state.
The state space is a complex vector space: these linear combinations can
and do crucially involve complex numbers, in an inescapable way. In the
classical case only real numbers appear, with complex numbers used only
as an inessential calculational tool.
In this course we will often use the notation introduced by Dirac for vectors
in the state space H: such a vector with a label is denoted
|i
d i
|(t)i = H|(t)i
dt ~
The Hamiltonian observable H will have a physical interpretation in terms
of energy, and one may also want to specify some sort of positivity property on
H, in order to assure the existence of a stable lowest energy state.
~ is a dimensional constant, the value of which depends on what units you
use for time and for energy. It has the dimensions [energy] [time] and its
experimental values are
3
small units of 6.6 1016 eV or energies in eV and time in the very small units
of 6.6 1016 seconds. Schrodingers equation implies that if you are looking
at a system where the typical energy scale is an eV, your state-vector will be
changing on the very short time scale of 6.6 1016 seconds. When we do
computations, often we will just set ~ = 1, implicitly going to a unit system
natural for quantum mechanics. When we get our final result, we can insert
appropriate factors of ~ to allow one to get answers in more conventional unit
systems.
It is sometimes convenient however to carry along factors of ~, since this
can help make clear which terms correspond to classical physics behavior, and
which ones are purely quantum mechanical in nature. We will see that typically
classical physics comes about in the limit where
Principle (Measurements).
c1 |1 i + c2 |2 i
may not have a well-defined value for the observable O. If one attempts
to measure this observable, one will get either 1 or 2 , with probabilities
|c21 |
|c21 | + |c22 |
4
and
|c22 |
|c21 | + |c22 |
respectively.
This principle is sometimes raised to the level of an axiom of the theory,
but it is better to consider it as a phenomenological over-simplified description
of what happens in typical experimental set-ups. A full understanding of how
classical behavior emerges in experiments is a very challenging topic, with the
notion of decoherence playing an important role. See the end of this chapter
for some references useful in pursuing this question.
Note that the state c|i will have the same eigenvalues and probabilities as
the state |i, for any complex number c. So states are only really determined
up to multiplication by a complex number. It is conventional to work with
states of norm fixed to the value 1, which fixes the amplitude of c, but leaving
a remaining ambiguity which is a phase ei .
5
where GL(n, C) is the group of invertible n by n complex matrices. For the
first part of this course we will just be studying representations that are finite
dimensional and will try to make rigorous statements. Later on we will get to
representations on function spaces, which are infinite dimensional, but from then
on will be rather cavalier about the serious analytical difficulties that arise when
one tries to make mathematically precise statements in the infinite-dimensional
case.
To make matters confusing, representations (, V ) are sometimes referred to
by the map , leaving implicit the vector space V that the matrices (g) act on,
but at other times referred to by specifying the vector space V , leaving implicit
the map . One reason for this is that the map may be the identity map:
often G is a matrix group, so a subgroup of GL(n, C), acting on V ' Cn by the
standard action of matrices on vectors. One should keep in mind though that
just specifying V is generally not enough to specify the representation, since it
may not be the standard one. For example, it could very well carry the trivial
representation, where
(g) = 1n
i.e. each element of G acts on V as the identity.
In mathematics, it turns out that the most interesting classes of complex
representations are unitary, i.e. preserving the standard notion of length in
a complex vector space. Similarly in physics, the group representations we will
be interested in are those corresponding to physical symmetries, so preserving
lengths in H, since these correspond to probabilities of various observations. We
have the definition
Definition (Unitary representation). A representation (, V ) on a complex vec-
tor space V with Hermitian inner product < , > is a unitary representation if
it preserves the inner product, i.e.
U 1 = U
6
symmetries acting on it, the space of states H will carry a unitary representation
of G. For physicists working with quantum mechanics, this implies that repre-
sentation theory provides information about quantum mechanical state spaces .
For mathematicians studying representation theory, this means that physics is
a very fruitful source of unitary representations to study: any physical system
with a symmetry group G will provide one.
For a representation and group elements g that are close to the identity,
one can use exponentiation to write (g) GL(n, C) as
(g) = eA
A = A
B = B
7
Definition (Group action on a set). An action of a group G on a set M is
given by a map
(g, x) G M g x M
such that
g1 (g2 x) = (g1 g2 ) x
and
ex=x
where e is the identity element of G
((g)f )(x) = f (g 1 x)
Note the inverse that is needed to get the group homomorphism property to
work since one has
a calculation that would not work out properly for non-commutative G if one
defined ((g)f )(x) = f (g x).
One way to construct quantum mechanical state space H is as wave-functions,
meaning complex valued functions on space-time. The above shows that given
any group action on space-time, we get a representation on the state space H
of such wave-functions.
8
Note that only in the case of M a finite set of points will we get a finite-
dimensional representation this way, since only then will F un(M ) be a finite-
dimensional vector space (C# of points in M ). A good example to consider to
understand this construction is the following:
(g, xi ) g xi
This construction gives 6 three by three complex matrices, which under multipli-
cation of matrices satisfy the same relations as the elements of the group under
group multiplication. In this very special case, all the entries of the matrix will
be 0 or 1, but that is special to the permutation representation.
The discussion here has been just a quick sketch of some of the ideas behind
the material we will cover later in this course. These ideas will be covered
in much greater detail as the course unfolds, beginning with the next couple
lectures where they will appear very concretely when we discuss the simplest
possible quantum systems, those with one and two-complex dimensional state
spaces.
9
clear that quantum mechanics cannot be derived from classical mechanics, so
attempts to motivate it that way are of necessity rather unclear, although they
correspond to the very interesting historical story of how the subject evolved.
As we get further along in the course, we will come to the topic of the quantized
motion of a particle, at which point it will become much easier to follow the
standard books.
A good standard quantum mechanics textbook is [40]. Three useful text-
books on the subject aimed at mathematicians are [45], [12], and [27], as well as
the forthcoming [26]. The first few chapters of [19] provide an excellent while
very concise summary of both basic physics and quantum mechanics. One im-
portant topic we wont discuss is that of the application of the representation
theory of finite groups in quantum mechanics. For this as well as a discussion
that overlaps quite a bit with the point of view of this course while emphasizing
different areas, see [42].
The field of Quantum Information Theory gives a perspective on quantum
theory that puts the elementary two-state quantum system (the qubit) front
and center. You might want to try reading the first part of Chapter 2 of John
Preskills notes on quantum computation[34] which give similar axioms to ours,
and starts off with the same example we will soon be considering.
On the whole, in this class well be avoiding the difficult issue of how mea-
surements work and how classical physics emerges from quantum theory. Some
suggested references for these subjects are Wojciech Zureks updated version
of his 1991 Physics Today article[58] and his more recent work on quantum
darwinism[59], as well as books by Schlosshauer[38] and Haroche/Raimond[28].
10
Chapter 2
11
on a complex vector space V (with a chosen basis identifying V ' Cn ) is a
homomorphism
: G GL(n, C)
This is basically a set of n by n matrices, one for each group element, satis-
fying the multiplication rules of the group elements. n is called the dimension of
the representation. We will be mainly interested in continuous groups with an
infinite number of elements, but the subject of representations of finite groups
is a very rich one, with a great deal of interesting structure.
The groups G we are interested in will be examples of what mathematicians
call Lie groups. If you are familiar with differential geometry, such groups are
examples of smooth manifolds, and one can define derivatives of functions on G
and more generally of maps from G into spaces like GL(n, C). We will assume
that our representations are given by differentiable maps . Some difficult gen-
eral theory shows that considering the more general case of continuous maps
gives nothing new since the homomorphism property of these maps is highly
constraining. In any case, our goal in this course will be to study quite explic-
itly certain specific groups and representations which are central in quantum
mechanics, and these representations will always be differentiable.
Given two representations one can form their direct sum
Definition (Direct sum representation). Given representations 1 and 2 of
dimensions n1 and n2 , one can define another representation, of dimension
n1 + n2 called the direct sum of the two representations, denoted by 1 2 .
This representation is given by the homomorphism
1 (g) 0
(1 2 ) : g G
0 2 (g)
In other words, one just takes as representation matrices block-diagonal
matrices with 1 and 2 giving the blocks.
To understand the representations of a group G, one proceeds by first iden-
tifying the irreducible ones, those that cannot be decomposed into two repre-
sentations of lower dimension.
Definition (Irreducible representation). A representation is called irreducible
if it cannot be put in the form 1 2 , for 1 and 2 of dimension greater than
zero.
This criterion is not so easy to check, and the decomposition of an arbitrary
reducible representation into irreducible components can be a very non-trivial
problem. Recall that one gets explicit matrices for the (g) of a representation
(, V ) only when a basis for V is chosen. To see if the representation is reducible,
one needs to see not whether the (g) are all in block-diagonal form, but whether
there is some basis for V with respect to which they are all in such form,
something very non-obvious from just looking at the matrices themselves.
Digression. Another approach to this would be to check to see if the repre-
sentation has no proper non-trivial sub-representations (such a representation
12
is often called indecomposable). This is not necessarily equivalent to irre-
ducibility, since a sub-representation may have no complement that is also a
sub-representation. A simple example of this occurs for the action of upper
triangular matrices on column vectors. Such representations are however non-
unitary. In the unitary case indecomposability and irreducibility are equivalent.
In these notes unless otherwise specified, one should assume that all represen-
tations are unitary, so, in particular, this issue will not arise.
Proof. Since we are working over the algebraically closed field C, we can solve
the eigenvalue equation
det(M 1) = 0
to find the eigenvalues of M . The eigenspaces
V = {v V : M v = v}
are vector subspaces of V and can also be described as ker(M 1), the kernel
of the operator M 1. Since and this operator and all the (g) commute, we
have
v ker(M 1) = (g)v ker(M 1)
so ker(M 1) V is a representation of G. If V is irreducible, we must
have either ker(M 1) = V or ker(M 1) = 0. Since is an eigenvalue,
ker(M 1) 6= 0, so M = 1 as a linear operator on V .
(g)(h) = (h)(g)
13
for all h G. If is irreducible, Schurs lemma implies that, since they commute
with all the (g), the matrices (h) are all scalar matrices, i.e. (h) = h 1 for
some h C. is then irreducible exactly when it is the one-dimensional
representation given by (h) = h .
Definition (The group U (1)). The elements of the group U (1) are points on
the unit circle, which can be labeled by the unit complex number ei , for
R. Note that and + N 2 label the same group element. Multiplication of
group elements is just complex multiplication, which by the properties of the
exponential satisfies
ei1 ei2 = ei(1 +2 )
By our theorem from the last section, since U (1) is a commutative group,
all irreducible representations will be one-dimensional. Such an irreducible rep-
resentation will be given by a map
: U (1) GL(1, C)
Theorem. All irreducible representations of the group U (1) are unitary, and
given by
for k Z.
k (1 + 2 ) = k (1 )k (2 )
f : U (1) C
14
satisfying the homomorphism and periodicity properties is of this form. Com-
df
puting the derivative f 0 () = d we find
f ( + ) f ()
f 0 () = lim
0
(f () 1)
= f () lim (using the homomorphism property)
0
= f ()f 0 (0)
Denoting the constant f 0 (0) by C, the only solutions to this differential equation
satisfying f (0) = 1 are
f () = eC
Requiring periodicity we find
These representations are restrictions to the unit circle U (1) of the irre-
ducible representations of the group C , which are given by
k : z C k (z) = z k C
Such representations are not unitary, but they have an extremely simple form,
so it sometimes is convenient to work with them, later restricting to the unit
circle, where the representation is unitary.
Digression (Fourier analysis of periodic functions). Well later on discuss
Fourier analysis more seriously when we come to the case of the translation
groups and of state-spaces that are spaces of wave-functions on space-time.
For now though, it might be worth pointing out an important example of a rep-
resentation of U (1), the space F un(S 1 ) of complex-valued functions on the circle
S 1 . We will evade completely for now the very non-trivial analysis involved here,
by not specifying what class of functions we are talking about (e.g. continuous,
integrable, differentiable, etc.). Periodic functions can be studied by rescaling
the period to 2, thus looking at complex-valued functions of a real variable
satisfying
f ( + N 2) = f ()
for integer N , which we can think of as functions on a circle, parametrized by
angle . We have an action of the group U (1) on the circle by rotation:
(, ) U (1) S 1 + S 1
15
Last time, we saw that given an action of a group on a space X, we can lin-
earize and get a representation (, F un(X)) of the group on the functions on
the space, by taking
((g)f )(x) = f (g 1 x)
for f F un(X), x X. Here X = S 1 , the action is the rotation action and we
find
(()f )() = f ( )
since the inverse of a rotation by is a rotation by .
This representation (, F un(S 1 )) is infinite-dimensional, but one can still
ask how it decomposes into the one-dimensional irreducible representations (k , C)
of U (1). What we learn from the subject of Fourier analysis is that each (k , C)
occurs exactly once in the decomposition of F un(S 1 ) into irreducibles, i.e.
M
(, F un(S 1 )) = (k , C)
kZ
where we have matched the sin of not specifying the class of functions in F un(S 1 )
on the left-hand side with the sin of not explaining how to handle the infinite
direct sum on the right-hand side. What can be specified precisely is how
L
the irreducible sub-representation (k , C) sits inside F un(S 1 ). It is the set of
functions f satisfying
16
G = U (1) are all one-dimensional, this means that as a U (1) representation, we
have
H = Hq1 Hq2 Hqn
0 : i iR 0 (i) = iQ
17
The right-hand side of the picture is supposed to somehow represent GL(n, C),
which is the 2n2 dimensional space of n by n complex matrices, minus the locus
of matrices with zero determinant, that cant be inverted. It has a distinguished
point, the identity. The charge Q map is the differential d of at the identity
In this very simple example, this abstract picture is over-kill and likely con-
fusing. We will see the same picture though occurring in many other examples
as the course goes on, ones where the abstract technology is increasingly useful.
Mostly though, keep in mind that these maps will be just explicit exponentia-
tion operators in the examples we care about, so have very concrete incarnations
as matrices we can do explicit calculations with.
18
the group U (1)). For now though, note that for time-independent Hamiltonian
operators H, the solution to this equation is given by exponentiating H, with
|(s)i = U (s)|(0)i
for
i
U (s) = e ~ sH
If the Hamiltonian operator H and the charge operator Q commute
[H, Q] = HQ QH = 0
[H k , Q] = 0
[U (s), Q] = 0
This condition
U (s)Q = QU (s)
implies that if a state has a well-defined value qi for the observable Q at time
t = 0, it will continue to have the same value at time t = s, since
19
20
Chapter 3
The simplest truly non-trivial quantum systems have state spaces that are in-
herently two-complex dimensional, and have significantly more structure than
the charge structure of the previous section, which could be analyzed by break-
ing up the space of states into one-dimensional subspaces of given charge. Well
study these in this section, encountering for the first time the implications of
working with representations of non-commutative groups. Since they give the
simplest non-trivial realization of quantum phenomena, such systems are the
fundamental objects of quantum information theory (the qubit) and the fo-
cus of attempts to build a quantum computer (which would be built out of
copies of this sort of fundamental object). Many different possible two-state
quantum systems could potentially be used as the physical implementation of a
qubit.
One of the simplest possibilities to take would be the idealized situation
of a single electron, somehow fixed so that its spatial motion could be ignored,
leaving its quantum state described just by its so-called spin degree of freedom,
which takes values in H = C2 . The term spin is supposed to call to mind
the angular momentum of an object spinning about about some axis, but such
classical physics has nothing to do with the qubit, it is a purely quantum system.
We will see later that quantum systems can be defined with spin quantum
number N/2, for
N = 0, 1, 2, 3, . . .
with dimension of the state space given by N + 1. The N = 0 case is the trivial
case, and the case at hand is N = 1, the spin 1/2 case. In the limit N
one can make contact with classical notions of spinning objects and angular
momentum, but the spin 1/2 case is at the other limit, where the behavior is
purely quantum.
21
3.1 The two-state quantum system
3.1.1 The Pauli matrices: observables of the two-state
quantum system
For a quantum system with two-dimensional state space, observables are self-
adjoint linear operators on C2 . With respect to a chosen basis of C2 , these are 2
by 2 complex matrices M satisfying the condition M = M (M is the conjugate
transpose matrix). Any such matrix will be a (real) linear combination of four
matrices:
M = c0 1 + c1 1 + c2 2 + c3 3
with ca R and the standard choice of the basis elements given by
1 0 0 1 0 i 1 0
1= , 1 = , 2 = , 3 =
0 1 1 0 i 0 0 1
The a are called the Pauli matrices and are a pretty universal choice of basis
in this subject. They reflect an arbitary convention, in particular the convention
of diagonalizing the basis element in the 3 or z-direction of R3 . One could just
as easily have decided to choose a basis diagonal in some other direction.
Recall that the basic principle of how measurements are supposed to work
in quantum theory says that the only states that have well-defined values for
these four observables are the eigenvectors for these matrices. The first matrix
gives a trivial observable, whereas the last one, 3 , has the two eigenvectors
1 1
3 =
0 0
and
0 0
3 =
1 1
with eigenvalues +1 and 1. In quantum information theory, where this is
the qubit system, these two eigenstates are labeled |0i and |1i because of the
analogy with a classical bit of information. Later on when we get to the theory of
spin, we will see that 21 3 is the observable corresponding to the SO(2) = U (1)
symmetry group of rotations about the third spatial axis, and the eigenvalues
12 , + 12 of this operator will be used to label the two eigenstates
1 1 1 0
|+ i= and | i =
2 0 2 1
A crucial and easy to check fact is that |+ 21 i and | 12 i are NOT eigenvectors
for the operators 1 and 2 . An exercise in the first problem set is to compute
the eigenvectors and eigenvalues for these operators. It can be checked that no
pair of the three a commute, so one cannot find vectors that are simultaneous
eigenvectors for more than one a . This non-commutativity of the operators
is responsible for the characteristic classically paradoxical property of quantum
22
observables: one can find states with a well defined number for the measured
value of one observable a , but such states will not have a well-defined num-
ber for the measured value of the other two non-commuting observables. The
physics language for this phenomenon is that if we prepare states with a well-
defined spin component in the a-direction, the two other components of the
spin cant be assigned a numerical value in such a state. Any attempt to
prepare states that simultaneously have specific chosen numerical values for the
3 observables corresponding to the a is doomed. So is any attempt to si-
multaneously measure such values: if you thought you knew the value for one
observable, measuring another would ensure that this first measurement was no
longer valid.
This non-classical behavior of observables takes some getting used to, and
there are many subtleties in the theory of measurement, but the basic phe-
nomenon is captured accurately by the very simple mathematics of the eigenvec-
tors and eigenvalues of the three 2 by 2 matrices a . Many quantum mechanics
textbooks (a good example is Volume III of the Feynman lectures[15]) contain
an extensive discussion of the physical implications of this specific example.
The choice we have made for the a corresponds to a choice of basis such
that the basis vectors are eigenvectors of 3 . 1 and 2 take these basis vectors
to non-trivial linear combinations of basis vectors. It turns out that there are
two specific linear combinations of 1 and 2 that do something very simple to
the basis vectors, since
0 2 0 0
(1 + i2 ) = and (1 i2 ) =
0 0 2 0
we have
0 1 1 0
(1 + i2 ) =2 (1 + i2 ) =
1 0 0 0
and
1 0 0 0
(1 i2 ) =2 (1 i2 ) =
0 1 1 0
(1 + i2 ) is called a raising operator: on eigenvectors of 3 it either
increases the eigenvalue by 2, or annihilates the vector. (1 i2 ) is called
a lowering operator: on eigenvectors of 3 it either decreases the eigenvalue
by 2, or annihilates the vector. Note that these linear combinations are not
self-adjoint, (1 + i2 ) is the adjoint of (1 i2 ) and vice-versa.
Well recall soon in our review of linear algebra that an arbitrary self-adjoint
2 by 2 complex matrix can be diagonalized by a unitary change of basis, i.e.
there is a unitary matrix U such that
1 0
U M U 1 =
0 2
where 1 , 2 R are the eigenvalues of the matrix. If c0 = 0, so we just have a
linear combination of the three Pauli matrices, we must have 2 = 1 , since
the Pauli matrices all have trace zero.
23
3.1.2 Exponentials of Pauli matrices: unitary transforma-
tions of the two-state system
Exponentiating the identity matrix gives the diagonal unitary matrix
i
e 0
ei1 =
0 ei
This matrix commutes with any other 2 by 2 matrix and just acts on our state
space C2 by multiplication by a phase, as studied in the previous section.
If we exponentiate Pauli matrices, it turns out that since all the j satisfy
j2 = 1, their exponentials also take a simple form:
1 1
eij = 1 + ij + (i)2 j2 + (i)3 j3 +
2 3!
1 1
= 1 + ij 2 1 i 3 j +
2 3!
1 1
= (1 2 + )1 + i( 3 + )j
2 3!
= (cos )1 + ij (sin )
{j , k } = j k + k j = 2jk 1
24
We can use this to compute the exponential of the matrix iv ( R, v
a unit vector) as follows, using the Taylor series formula for the exponential:
(iv )2 (iv )3
eiv = 1 + iv + + +
2! 3!
2 3
= (1 + )1 + i( + )v
2 3!
= (cos )1 + i(sin )v
Notice that you can easily compute the inverse of this matrix:
since
Well review linear algebra and the notion of a unitary matrix in one of the next
classes, but one form of the condition for a matrix M to be unitary is
M = M 1
25
To get some more insight into the structure of the group SU (2), consider an
arbitrary 2 by 2 complex matrix
Unitarity implies that the rows are orthonormal. One can see this explicitly
from the condition that the matrix times its conjugate-transpose is the identity
1 0
=
0 1
+ = 0 = =
The condition that the first row has length one gives
+ = ||2 + ||2 = 1
Using these two relations and computing the determinant (which has to be 1)
gives
= = ( + ) = = 1
so one must have
= , =
and an SU (2) matrix will have the form
where (, ) C2 and
||2 + ||2 = 1
So, the elements of SU (2) are parametrized by two complex numbers, with
the sum of their length-squareds equal to one. Identifying C2 = R4 , these are
just vectors of length one in R4 . Just as U (1) could be identified as a space with
the unit circle S 1 in C = R2 , SU (2) can be identified with the three-sphere S 3
in R4 .
26
in particular a U (1) subgroup of SU (2). Another important set of relations
satisfied by Pauli matrices are their commutation relations:
3
X
[j , k ] = j k k j = 2i jkl l
l=1
One can easily check these relations by explicitly computing with the matrices.
Putting together the anticommutation and commutation relations, one gets a
formula for the product of two Pauli matrices:
3
X
j k = jk 1 + i jkl l
l=1
While physicists prefer to work with self-adjoint Pauli matrices and their
real eigenvalues, one can work instead with the following skew-adjoint matrices
j
Xj = i
2
which satisfy the slightly simpler commutation relations
3
X
[Xj , Xk ] = jkl Xl
l=1
or more explicitly
If these commutators were zero, the SU (2) elements one gets by exponentiat-
ing linear combinations of the Xj would be commuting group elements. The
non-triviality of the commutators reflects the non-commutativity of the group.
Group elements U SU (2) near the identity satisfy
U ' 1 + 1 X1 + 2 X2 + 3 X2
for a small and real, just as group elements z U (1) near the identity satisfy
z ' 1 + i
27
space S 3 . The Xj give a basis of the tangent space R3 to the identity of
SU (2), just as i gives a basis of the tangent space to the identity of U (1).
H = h0 1 + h1 1 + h2 2 + h3 3
|(t)i = U (t)|(0)i
where
itH
U (t) = e ~
ih0 t
The h0 1 term in H just contributes an overall phase factor e ~ , with the
remaining factor of U (t) an element of the group SU (2) rather than the larger
group U (2) of all 2 by 2 unitaries.
Using our earlier equation
28
h
valid for a unit vector v, our U (t) is given by taking h = (h1 , h2 , h3 ), v = |h|
and = t|h|
~ , so we find
ge~
H= (B1 1 + B2 2 + B3 3 )
4mc
where the Ba are the components of the magnetic field, and the physical con-
stants are the gyromagnetic ratio (g), the electric charge (e), the mass (m) and
the speed of light (c), so we have solved the problem of the time evolution of
ge~
such a system, setting ha = 4mc Ba . For magnetic fields of size |B| in the 3-
direction, we see that the two different states with well-defined energy (| + 12 i
and | 12 i) will have an energy difference between them of
ge~
|B|
mc
This is known as the Zeeman effect and is readily visible in the spectra of atoms
subjected to a magnetic field.
29
30
Chapter 4
A course in linear algebra is a prerequisite for this course, and well need a
range of specific facts from that subject. We will also need some aspects of
linear algebra that might not have made an appearance (at least not explicitly)
in whatever course you may have taken. These include the use of complex
vector spaces, as opposed to the real vector spaces that a typical linear algebra
course concentrates on. We also will consider matrices in their role as elements
of matrix groups, a topic often not emphasized in typical courses. For now our
vector spaces will be finite-dimensional. Later in the course we will come to
state spaces that are infinite dimensional, and will address the various issues
that this raises at that time.
Definition. Given a vector space V over a field k, the dual vector space V is
the set of all linear maps V k, i.e.
for , k, v, w V .
Physicists have a useful notation for elements of vector space and their duals.
31
An element of a vector space V is written as a ket vector
|vi
where v is a label for a vector. An element of the dual vector space V is written
as a bra vector
hl|
Evaluating l V on v V gives an element of k, written
hl|vi
If : V V is a linear map
h, i : V V C
that is linear in the second variable, antilinear in the first variable, i.e. for
, C and u, v, w V
hu + v, wi = hu, wi + hv, wi
||v||2 = hv, vi
32
An inner product provides an isomorphism V ' V by the map
v V lv V
where lv is defined by
lv (w) = hv, wi
In the bra-ket notation, one denotes the dual vector lv by hv|. Note that in
the inner product the angle bracket notation means something different than in
the bra-ket notation. The similarity is intentional though, since in the bra-ket
notation one has
hv|wi = hv, wi
Note that our convention of linearity in the second variable of the inner product,
anti-linearity in the first, implies
for C.
hei , ej i = ij
|ii = ei
v = v1 e1 + v2 e2 + + vn en
Digression. In this course, all our indices will be lower indices. One way to
keep straight the difference between vectors and dual vectors is to use upper
indices for components of vectors, lower indices for components of dual vectors.
This is quite useful in Riemannian geometry and general relativity, where the
inner product is given by a metric that can vary from point to point, causing
the isomorphism between vectors and dual vectors to also vary. For quantum
mechanical state spaces, we will be using a single, standard, fixed inner product,
so there will be a single isomorphism between vectors and dual vectors. The
bra-ket notation will take care of the notational distinction between vectors and
dual vectors as necessary.
33
Note that one can also think of the vi as coordinate functions on V . In this
interpretation of the symbol vi , it is a linear function on V , thus an element of
the dual V . Specifically, it is the linear function on V which takes value the
coefficient vi of the basis vector ei on a vector v V .
Because of orthonormality, these coefficients of vectors can be calculated as
vi = hei , vi
With respect to the chosen orthonormal basis ei , one can explicitly represent
vectors v as column vectors
v1
v2
..
.
vn
with
0
..
.
ei = |ii =
1
.
..
0
(i.e. a 1 in the ith position, 0s elsewhere).
The operation of taking a vector |vi to a dual vector hv| corresponds to
taking a column vector to the row vector that is its conjugate-transpose.
hv| = v1 v2 vn
Then one has
w1
w2
hv|wi = v1 v2 vn . = v1 w1 + v2 w2 + + vn wn
..
wn
34
If is a linear operator : V V , then with respect to the chosen basis it
becomes a matrix, with matrix elements
ji = hj|ii
Pi = |iihi|
is called the projection operator onto the ith basis vector, it corresponds to the
matrix that has 0s everywhere except in the ii component.
hv, wi = hv, wi
35
Generalizing the fact that
hv| = hv|
for C, one can write
hv| = hv|
Note that mathematicians tend to favor the notation over the physicists
notation . On this one Im siding with the physicists.
In terms of explicit matrices, < v| is the conjugate-transpose of |v >, so
the matrix with respect to a basis for will be given by the conjugate transpose
T of the matrix for :
( )ij = ji
In the real case, the matrix for the adjoint is just the transpose matrix. We
will say that a linear transformation is self-adjoint if = , skew-adjoint if
= .
which says that the column vectors of the matrix for are orthonormal vectors.
Using instead the equivalent condition
= 1
one finds that the row vectors of the matrix for are also orthornormal.
Since such linear transformations can be composed and are invertible, they
form a group, and some of the basic examples of Lie groups are given by these
groups for the cases of real and complex vector spaces.
36
4.4.1 Orthogonal groups
Well begin with the real case, where these groups are called orthogonal groups:
Note that since the determinant of the transpose of a matrix is the same as
the determinant of the matrix, we have
so
det() = 1
O(n) is a continuous Lie group, with two components: SO(n), the subgroup of
orientation-preserving transformations, which include the identity, and a com-
ponent of orientation-changing transformations.
The simplest non-trivial example is for n = 2, where all elements of SO(2)
are given by matrices of the form
cos sin
sin cos
so the representation theory of SO(2) is just as for U (1), with irreducible com-
plex representations one-dimensional and classified by an integer.
We will go on in the next section of the class to examine in detail the case
of SO(3), which is crucial for physical applications because it is the group of
rotations in the physical number of three space dimensions.
37
4.4.2 Unitary groups
In the complex case, groups of invertible transformations preserving the Hermi-
tian inner product are called unitary groups:
Definition (Unitary group). The unitary group U (n) in n-dimensions is the
group of invertible transformations preserving an Hermitian inner product on a
complex n-dimensional vector space V . This is also the group of n by n complex
invertible matrices satisfying
(1 )ij = ji = ( )ij
is a group homomorphism.
We have already seen the examples U (1), U (2) and SU (2). For general
values of n, the case of U (n) can be split into the study of its determinant,
which lies in U (1) so is easy to deal with, and the subgroup SU (n), which is
a much more complicated story. As a manifold, we saw that SU (2) can be
identified with the three-sphere S 3 , since an arbitrary group element can be
constructed by specifying one row (or one column), which must be a vector of
length one in C2 . For the case n = 3, the same sort of construction starts by
picking a row of length one in C3 , which will be a point in S 5 . The second
row must be orthornormal, and one can show that the possibilities lie in a
three-sphere S 3 . Once the first two rows are specified, the third row is uniquely
determined. So as a manifold, SU (3) is eight-dimensional, and one might think
it could be identified with S 5 S 3 . It turns out that this is not the case, since
the S 3 varies in a topologically non-trivial way as one varies the point in S 5 .
As spaces, the SU (n) are topologically twisted products of odd-dimensional
spheres, providing some of the basic examples of quite non-trivial topological
manifolds.
38
Complex matrices behave in a much simpler fashion than real matrices, since in
the complex case the eigenvalue equation
det(1 M ) = 0
can always be factored into linear factors, and solved for the eigenvalues . For
an arbitrary n by n complex matrix there will be n solutions, counting repeated
eigenvalues with multiplicity. Changing basis and conjugating such a matrix M ,
one can always find a basis for which the matrix is in upper triangular form.
For the case of self-adjoint matrices, things are much more constrained, and
one has the following theorem, proved in most linear algebra textbooks:
Theorem (Spectral theorem for self-adjoint matrices). Given a self-adjoint
complex n by n matrix M , one can always find a unitary matrix U such that
U M U 1 = D
where D is a diagonal matrix with entries Dii = i , i R.
Given M , one finds the eigenvalues i by solving the eigenvalue equation.
One can then go on to solve for the eigenvectors and use these to find U . For
distinct eigenvalues one finds that the corresponding eigenvectors are orthogo-
nal.
This theorem is of crucial importance in quantum mechanics, where for M
an observable, the eigenvectors are the states in the state space with well-defined
numerical values that can be observed experimentally for the observable, and
these numerical values are the eigenvalues. The theorem also tells us that given
an observable, we can use it to choose distinguished orthonormal bases for the
state space: a basis of eigenvectors, normalized to length one. This is a theorem
about finite-dimensional vector spaces, but later on in the course we will see
that something similar will be true even in the case of infinite-dimensional state
spaces.
One can also diagonalize unitary matrices themselves by conjugation by
another unitary. The diagonal entries will all be complex numbers of unit length,
so of the form eij , j R.
For the simplest examples, consider the cases of the groups SU (2) and U (2).
Any matrix in U (2) can be conjugated by a unitary matrix to the diagonal
matrix i
e 1 0
0 ei2
which is the exponential of a corresponding diagonalized skew-adjoint matrix
i1 0
0 i2
For matrices in the subgroup SU (2), one has 2 = 1 = so in diagonal form
an SU (2) matrix will be i
e 0
0 ei
39
which is the exponential of a corresponding diagonalized skew-adjoint matrix
that has trace zero
i 0
0 i
40
Chapter 5
: G U (n)
Recall that in the case of G = U (1), we could use the homomorphism property
of to determine in terms of its derivative at the identity. This turns out to
be a general phenomenon for Lie groups G: we can study their representations
by considering the derivative of at the identity, which we will call 0 . Because
of the homomorphism property, knowing 0 is often sufficient to characterize
the representation it comes from. 0 is a linear map from the tangent space
to G at the identity to the tangent space of U (n) at the identity. The tangent
space to G at the identity will carry some extra structure coming from the group
multiplication, and this vector space with this structure will be called the Lie
algebra of G.
The subject of differential geometry gives many equivalent ways of defining
the tangent space at a point of manifolds like G, but we do not want to enter
here into the large subject of differential geometry in general. One of these ways
though is to define the tangent space as the space of tangent vectors, and define
tangent vectors as the possible velocity vectors of parametrized curves g(t) in
the group G.
This can be done, and more advanced treatments of Lie group theory de-
velop this point of view (see for example [49]). In our case though, since we
are interested in specific groups that are explicitly given as groups of matrices,
we can give a more concrete definition, just using the exponential map on ma-
trices. For a more detailed exposition of this subject, using the same concrete
definition of the Lie algebra in terms of matrices, see Brian Halls book[24] or
the abbreviated on-line version[25].
Note that the material of this chapter is quite general, and may be hard
41
to until one has some experience with basic examples. The next chapter will
discuss in detail the groups SU (2) and SO(3) and their Lie algebras, as well
as giving some examples of their representations, and this may be helpful in
making sense of the general theory of this chapter.
42
Definition (Adjoint representation). The adjoint representation (Ad, g) is given
by the homomorphism
Ad : g G {X gXg 1 } GL(g)
meaning
(Ad(g))(X) = gXg 1
To show that this is well-defined, one needs to check that gXg 1 g when
X g, but this can be shown using the identity
1
etgXg = getX g 1
1
which implies that etgXg G if etX G. To check this, just expand the
exponential and use
One should notice that there is something tricky going on here: in general
a Lie algebra g is a real vector space, not a complex vector space (for instance,
u(1) = R and su(2) = R3 as vector spaces). Even if G is a group of complex
matrices, when it is not GL(n, C) itself but some subgroup, its tangent space
at the identity will not necessarily be a complex vector space. Consider for
examples the cases G = U (1) and G = SU (2), where u(1) = R and su(2) = R3 .
While the tangent space to the group of all invertible complex matrices is a
complex vector space, imposing some condition such as unitarity picks out a
subspace which generally is just a real vector space, not a complex one. So the
adjoint representation (Ad, g) is in general not a complex representation, but a
real representation, with
[, ] : (X, Y ) g g [X, Y ] = XY Y X g
43
We need to check that this is well-defined, i.e. that it takes values in g
Theorem. If X, Y g, [X, Y ] = XY Y X g
d tX tX d d
(e Y e ) = ( (etX Y ))etX + etX Y ( etX )
dt dt dt
= XetX Y etX etX Y XetX
where cjkl is a set of constants known as the structure constants of the Lie
algebra. For example, in the case of su(2), the Lie algebra of SU (2) one has a
basis X1 , X2 , X3 satisfying
3
X
[Xj , Xk ] = jkl Xl
l=1
so the structure constants of su(2) are just the totally anti-symmetric jkl .
44
In order to see what this condition becomes on the Lie algebra, write = etX ,
for some parameter t, and X a matrix in the Lie algebra. Since the transpose of
a product of matrices is the product (order-reversed) of the transposed matrices,
i.e.
(XY )T = Y T X T
and the complex conjugate of a product of matrices is the product of the complex
conjugates of the matrices, one has
(etX ) = etX
The condition
= 1
thus becomes
etX (etX ) = etX etX = 1
Since X and X commute, this becomes
et(X+X )
=1
or
X + X = 0
so the matrices we want to exponentiate are skew-adjoint, satisfying
X = X
Note that physicists often choose to define the Lie algebra in these cases as
self-adjoint matrices, then multiplying by i before exponentiating to get a group
element. We will not use this definition, one reason being that the situation of
real vs. complex vector spaces here is already tricky enough without introducing
another source of complex numbers into the story.
= etX
These give a path connecting to the identity (taking esX , s [0, t]). We
saw above that the condition T = 1 corresponds to skew-symmetry of the
matrix X
X T = X
45
So in the case of G = SO(n), we see that the Lie algebra so(n) is the space of
skew-symmetric (X t = X) n by n real matrices, together with the bilinear,
antisymmetric product given by the commutator:
n2 n
1 + 2 + + (n 1) =
2
and a basis will be given by the matrices Eij , with i, j = 1, . . . n, i < j defined
as
+1 if i = k, j = l
(Eij )kl = 1 if j = k, i = l
0 otherwise
In the next section of the class we will examine detail the n = 3 case, where
the Lie algebra so(3) is R3 , realized as the space of anti-symmetric real 3 by 3
matrices.
Note that this is a real vector space, and its definition just involves real
matrices. Thought of as the adjoint representation, it is a real representation of
SO(n). One aspect of this is that the matrices Ad(g) of the adjoint representa-
tion are not diagonalizable as real matrices. One may want to complexify to get
a complex representation in which the Ad(g) can be diagonalized. Doing that
takes so(n), the antisymmetric n by n real matrices to so(n)C , the Lie algebra
of antisymmetric n by n complex matrices.
X = X
46
subgroup SO(n) U (n)). If we take ALL complex linear combinations of skew-
Hermitian matrices, than we get the full algebra of n by n complex matrices,
i.e.
u(n)C = M (n, C) = gl(n, C)
There is an identity relating the determinant and the trace of a matrix
det(eX ) = etrace(X)
: G GL(n, C)
satisfying
([X, Y ]) = [(X), (Y )]
Such a representation is called unitary if its image is in u(n), i.e it satisfies
(X) = (X)
47
More concretely, given a basis X1 , X2 , . . . , Xd of a Lie algebra g of dimension
d with structure constants cjk , a representation is given by a choice of d n by n
complex matrices (Xj ) satisfying the commutation relations
d
X
[(Xj ), (Xk )] = cjkl (Xl )
l=1
48
satisfies
1. 0
(etX ) = et (X)
2. For g G
0 (gXg 1 ) = (g) 0 (X)((g))1
3. 0 is a Lie algebra homomorphism:
0 ([X, Y ]) = [ 0 (X), 0 (Y )]
Proof. 1. We have
d d
(etX ) = (e(t+s)X )|s=0
dt ds
d
= (etX esX )|s=0
ds
d
= (etX ) (esX )|s=0
ds
= (etX ) 0 (X)
d
So f (t) = (etX ) satisfies the differential equation dt f = f 0 (X) with
0
initial condition f (0) = 1. This has the unique solution f (t) = et (X)
2. We have
0 1 1
et (gXg )
= (etgXg )
tX 1
= (ge g )
= (g)(e tX
)(g)1
0
= (g)et (X) (g)1
Differentiating with respect to t at t = 0 gives
0 (gXg 1 ) = (g) 0 (X)((g))1
3. Recall that
d tX tX
[X, Y ] = (e Y e )|t=0
dt
so
d tX tX
0 ([X, Y ]) = 0 ( (e Y e )|t=0 )
dt
d 0 tX tX
= (e Y e )|t=0 (by linearity)
dt
d
= ((etX ) 0 (Y )(etX ))|t=0 (by 2.)
dt
d 0 0
= (et (X) 0 (Y )et (X) )|t=0 (by 1.)
dt
= [ 0 (X), 0 (Y )]
49
This theorem shows that we can study Lie group representations (, V )
by studying the corresponding Lie algebra representation ( 0 , V ). This will
generally be much easier since the 0 (X) are just linear maps. We will need this
a bit later in the course when we construct and classify all SU (2) and SO(3)
representations, finding that the corresponding Lie algebra representations are
much simpler to analyze.
For any Lie group G, we have seen that there is a distingished representation,
the adjoint representation (Ad, g). The corresponding Lie algebra representation
is also called the adjoint representation, but written as (Ad0 , g) = (ad, g). From
the fact that
Ad(etX )(Y ) = etX Y etX
we can differentiate with respect to t to get the Lie algebra representation
d tX tX
ad(X)(Y ) = (e Y e )|t=0 = [X, Y ]
dt
From this we see that ad(X) is the linear map from g to itself given by
Y [X, Y ]
where these are linear maps on g, with composition of linear maps, so operating
on Z g we have
This is called the Jacobi identity. It could have been more simply derived as
an identity about matrix multiplication, but here we see that it is true for a
more abstract reason, reflecting the existence of the adjoint representation. It
can be written in other forms, rearranging terms using antisymmetry of the
commutator, with one example the sum of cyclic permutations
[, ] : (X, Y ) A A [X, Y ] A
satisfying
50
1. Antisymmetry:
[X, Y ] = [Y, X]
2. Jacobi identity:
Such Lie algebras do not need to be defined as matrices, and their Lie bracket
operation does not need to be defined in terms of a matrix commutator (al-
though the same notation continues to be used). Later on in this course we
will encounter important examples of Lie algebras that are defined in this more
abstract way.
51
52
Chapter 6
Among the basic symmetry groups of the physical world is the orthogonal group
SO(3) of rotations about a point in three-dimensional space. The observables
one gets from this group are the components of angular momentum, and under-
standing how the state space of a quantum system behaves as a representation
of this group is a crucial part of the analysis of atomic physics examples and
many others. Unlike some of the material covered in this course, this is a topic
you will find in some version or other in every quantum mechanics textbook.
Remarkably, it turns out that the quantum systems in nature are often
representations not of SO(3), but of a larger group called Spin(3), one that
has two elements corresponding to every element of SO(3). This is called the
Spin group, and in any dimension n it exists, always as a doubled version
of the orthogonal group SO(n), one that is needed to understand some of the
more subtle aspects of geometry in n dimensions. In the n = 3 case it turns
out that Spin(3) ' SU (2) and we will study in detail the relationship of SO(3)
and SU (2). This appearance of the unitary group SU (2) is special to 3 (and to
some extent 4) dimensions.
53
This can be written as an exponential, R() = eL = cos + L sin for
0 1
L=
1 0
Here SO(2) is a commutative Lie group with Lie algebra R (the Lie bracket is
trivial, all elements of the Lie algebra commute). Note that we have a represen-
tation on V = R2 here, but it is a real representation, not one of the complex
ones we have when we have a representation on a quantum mechanical state
space.
In three dimensions the group SO(3) is 3-dimensional and non-commutative.
One now has three independent directions one can rotate about, which one can
take to be the x, y and z-axes, with rotations about these axes given by
1 0 0 cos 0 sin
Rx () = 0 cos sin Ry () = 0 1 0
0 sin cos sin 0 cos
cos sin 0
Rz () = sin cos 0
0 0 1
A standard parametrization for elements of SO(3) is in terms of 3 Euler angles
, , with a general rotation given by
Note that these are exactly the same commutation relations satisfied by
the basis vectors X1 , X2 , X3 of the Lie algebra su(2), so so(3) and su(2) are
isomorphic Lie algebras. They both are the vector space R3 with the same Lie
54
bracket operation on pairs of vectors. This operation is familiar in yet another
context, that of the cross-product of standard basis vectors ej in R3 :
e1 e2 = e3 , e2 e3 = e1 , e3 e1 = e2
(X, Y ) R3 R3 [X, Y ] R3
that makes R3 a Lie algebra so(3) = su(2) is just the cross-product on vectors
in R3 .
So far we have three different isomorphic ways of putting a Lie bracket on
R3 , making it into a Lie algebra:
[v, w] = v w
Something very special that happens for orthogonal groups only in 3-dimensions
is that the vector representation is isomorphic to the adjoint representation. Re-
call that any Lie group G has a representation (Ad, g) on its Lie algebra g. so(n)
can be identified with the anti-symmetric n by n matrices, so is of (real) dimen-
2
sion n 2n . Only for n = 3 is this equal to n, the dimension of the representation
on vectors in Rn . This corresponds to the geometrical fact that only in 3 di-
mensions is a plane (in all dimensions rotations are built out of rotations in
various planes) determined uniquely by a vector (the vector perpendicular to
the plane). Equivalently, only in 3 dimensions is there a cross-product v w
which takes two vectors determining a plane to a unique vector perpendicular
to the plane.
To see explicitly what is going on, lets look at the vector and adjoint Lie
algebra representations. The commutation relations for the li determine the Lie
algebra representation (ad, so(n)) using the definition of the adjoint representa-
tion, (ad(X))(Y ) = [X, Y ]. For instance, for infinitesimal rotations about the
x-axis, one has for the adjoint representation
l1 e1 = 0, l1 e2 = e3 , l1 e3 = e2
55
from which one can see that one has the same Lie algebra representation, with
the isomorphism identifying li = ei . At the level of the group, rotations about
the x-axis by an angle correspond to matrices
1 0 0
el1 = 0 cos sin
0 sin cos
which act by conjugation on anti-symmetric real matrices and in the usual way
on column vectors.
Our two isomorphic representations are on column vectors (vector repre-
sentation on R3 ) and on antisymmetric real matrices (adjoint representation
on so(3)), with the isomorphism given by
v1 0 v3 v2
v2 v3 0 v1
v3 v2 v1 0
On the column vectors both the Lie algebra and the Lie group representation
are given just by matrix multiplication. On the antisymmetric matrices the Lie
group representation is given by
0 v3 v2 0 v3 v2
Ad(g) v3 0 v1 = g v3 0 v1 g 1
v2 v1 0 v2 v1 0
56
cases requiring more sophisticated techniques. For n = 3 it turns out that
Spin(n) = SU (2), and for n = 4, Spin(4) = SU (2) SU (2). For perhaps the
simplest way to see how this works, it is best to not use just real and complex
numbers, but also bring in a third number system, the quaternions.
6.2.1 Quaternions
The quaternions are a number system (denoted by H) generalizing the complex
number system, with elements q H that can be written as
q = q0 + q1 i + q2 j + q3 k, qi R
with i, j, k H satisfying
i2 = j2 = k2 = 1, ij = k, ki = j, jk = i
q q = q0 q1 i q2 j q3 k
Since
q q
=1
|q|2
one has a formula for the inverse of a quaternion
q
q 1 =
|q|2
From this one can also see that the conjugation operation satisfies (for u, v H)
uv = vu
|uv|2 = uvuv = uv
vu = |u|2 |v|2
The length one quaternions thus form a group under multiplication, called Sp(1).
There are also Lie groups called Sp(n), consisting of invertible matrices with
quaternionic entries that act on quaternionic vectors preserving the quaternionic
length-squared, but these play no role in quantum mechanics so we wont study
them further. Sp(1) can be identified with the three-dimensional sphere since
the length one condition on q is
57
6.2.2 Rotations and spin groups in four dimensions
Pairs (u, v) of unit quaternions give the product group Sp(1) Sp(1). An
element of this group acts on H = R4 by
q uqv
and this action preserves lengths of vectors and is linear in q, so it must corre-
spond to an element of the group SO(4). One can easily see that pairs (u, v) and
(u, v) give the same orthogonal transformation of R4 , so the same element
of SO(4). One can show that SO(4) is the group Sp(1) Sp(1), with elements
(u, v) and (u, v) identified. The name Spin(4) is given to the Lie group that
double covers SO(4), so here Spin(4) = Sp(1) Sp(1).
Later on in the course well encounter Spin(4) and SO(4) again, but for now
were interested in its subgroup Spin(3) that only acts on 3 of the dimensions,
and double-covers not SO(4) but SO(3). To find this, consider the subgroup of
Spin(4) consisting of pairs (u, v) of the form (u, u1 ) (a subgroup isomorphic
to Sp(1), since elements correspond to a single unit length quaternion u). This
subgroup acts on quaternions by conjugation
q uqu1
an action which is trivial on the real quaternions, but nontrivial on the pure
imaginary quaternions of the form
q = ~v = v1 i + v2 j + v3 k
~v u~v u1
58
Both u and u act in the same way on ~v , so we have two elements in
Sp(1) corresponding to the same element in SO(3). One can show that is a
surjective map (one can get any element of SO(3) this way), so it is what is called
a covering map, specifically a two-fold cover. It makes Sp(1) a double-cover
of SO(3), and we give this the name Spin(3). This also tells us exactly which
3-dimensional space SO(3) is. It is S 3 = Sp(1) = Spin(3) with opposite points
on the three-sphere identified. This space is known as RP(3), real projective
3-space, which can also be thought of as the space of lines through the origin in
R4 (each such line intersects S 3 in two opposite points).
59
For those who have seen some topology, note that the covering map is an
example of a topological non-trivial cover. It is just not true that topologically
S 3 ' RP3 (+1, 1). S 3 is a connected space, not two disconnected pieces.
This topological non-triviality implies that globally there is no possible homo-
morphism going in the opposite direction from (i.e. SO(3) Spin(3)). One
can do this locally, picking a local patch in SO(3) and taking the inverse of
to a local patch in Spin(3), but this wont work if we try and extend it globally
to all of SO(3).
The relationship between rotations of R3 and unit quaternions is quite sim-
ple: for
w
~ = w1 i + w2 j + w3 k
a unit vector in R3 H, conjugation by the unit quaternion
u(, w)
~ = cos + w
~ sin
gives a rotation about the w~ axis by an angle 2. The factor of 2 here reflects
the fact that unit quaternions double-cover the rotation group SO(3): taking
+ gives a different unit quaternion that implements the same rotation
in 3-space. To see how this works, one can for example take as axis of rotation
the z-axis by choosing w~ = k. The unit quaternion
u = cos + k sin
60
has inverse
u1
= cos k sin
and acts on ~v = v1 i + v2 j + v3 k as
~v u ~v u1
=(cos + k sin )(v1 i + v2 j + v3 k)(cos k sin )
=(v1 (cos2 sin2 ) v2 (2 sin cos ))i
+ (2v1 sin cos + v2 (cos2 sin2 ))j + v3 k
=(v1 cos 2 v2 sin 2)i + (v1 sin 2 + v2 cos 2)j + v3 k
cos 2 sin 2 0 v1
= sin 2 cos 2 0 v2
0 0 1 v3
Here one sees explicitly for rotations about the z-axis the double-covering map
cos 2 sin 2 0
: u = cos + k sin Sp(1) = Spin(3) sin 2 cos 2 0 SO(3)
0 0 1
d
u(, w)
~ |=0 = (sin + w
~ cos )|=0 = w
~ = w1 i + w2 j + w3 k
d
The derivative of the map will be a linear map
0 : sp(1) so(3)
61
Using the definition as a derivative with respect to the parameter of a path,
evaluated at the identity, one finds
d
0 (k) = (cos + k sin )|=0
d
2 sin 2 2 cos 2 0
= 2 cos 2 2 sin 2 0
0 0 0 |=0
0 2 0
= 2 0 0 = 2l3
0 0 0
i j k
, , and l1 , l2 , l3
2 2 2
i j k j k i k i j
[ , ]= , [ , ]= , [ , ]=
2 2 2 2 2 2 2 2 2
62
we see that the length-squared function on quaternions corresponds to the de-
terminant function on 2 by 2 complex matrices. Taking q Sp(1), so of length
one, corresponds both to the matrix having determinant one, and its rows and
columns having length one, so the complex matrix is in SU (2).
Recall that any SU (2) matrix can be written in the form
= q0 iq3 , = q2 iq1
We see that Sp(1), Spin(3) and SU (2) are all names for the same group, geo-
metrically S 3 , the unit sphere in R4 .
Under our identification of H with 2 by 2 complex matrices, we have an
identification of Lie algebras sp(1) = su(2) between pure imaginary quaternions
and skew-Hermitian trace-zero 2 by 2 complex matrices
iw3 w2 iw1
~ = w1 i + w2 j + w3 k
w =w
w2 iw1 iw3
The basis 2i , 2j , k2 gets identified with a basis for the Lie algebra su(2) which
written in terms of the Pauli matrices is
j
Xj = i
2
with the Xj satisfying the commutation relations
We now have no less than three isomorphic Lie algebras sp(1) = su(2) =
so(3) on which we have the adjoint representation, with bases that get identified
under the following identifications
0 w3 w2
i w 3 w1 iw 2
w1 2i + w2 2j + w3 k2
w3 0 w1
2 w1 + iw2 w3
w2 w1 0
i 1
i l1
2 2
63
etc. The first of these identifications comes from the way we identify H with
2 by 2 complex matrices, these are Lie algebras of isomorphic groups. The
second identification is 0 , the derivative at the identity of the covering map :
Sp(1) = SU (2) SO(3) that is a homomorphism between two non-isomorphic
groups.
On each of these we have adjoint Lie group (Ad) and Lie algebra (ad) repre-
sentations, with Ad given by conjugation with the corresponding group elements
in Sp(1), SU (2) and SO(3), and ad given by taking commutators in the respec-
tive Lie algebras of pure imaginary quaternions, skew-Hermitian trace-zero 2 by
2 complex matrices and 3 by 3 real antisymmetric matrices.
Note that these three Lie algebras are all three-dimensional real vector
spaces, so these are real representations. If one wants a complex representa-
tion, one can complexify and take complex linear combinations of elements.
This is less confusing in the case of su(2) than for sp(1) since taking complex
linear combinations of skew-Hermitian trace-zero 2 by 2 complex matrices just
gives all trace-zero 2 by 2 matrices (the Lie algebra sl(2, C)).
In addition, recall from earlier that there is a fourth isomorphic version of
this representation, the representation of SO(3) on column vectors. This is also
a real representation, but can straightforwardly be complexified.
At the level of Lie groups we have seen that our identification of H and 2
by 2 matrices identifies Sp(1) with SU (2), taking
The relation to SO(3) rotations is that this is an SU (2) element such that,
if one identifies vectors (v1 , v2 , v3 ) R3 with complex matrices
v3 v1 iv2
v1 + iv2 v3
64
In particular, rotation about the z-axis by an angle is given by conjugation
by
0
!
ei 2 0
R(, 0) =
1 0 ei 2
In terms of the group SU (2), the double covering map thus acts on diag-
onalized matrices as
i cos 2 sin 2 0
e 0
: SU (2) sin 2 cos 2 0
ei
0
0 0 1
One can write down a somewhat unenlightening formula for the map :
SU (2) SO(3) in general, getting
65
66
Chapter 7
67
than quaternions. This irreducible representation is known as the spinor or
spin representation of Spin(3) The homomorphism spinor defining the rep-
resentation is just the identity map from SU (2) to itself.
The spin representation of Spin(3) is not a representation of SO(3). The
double cover map : Spin(3) SO(3) is a homomorphism, so given a rep-
resentation (, V ) of SO(3) one gets a representation ( , V ) of Spin(3) by
composition. But there is no homomorphism SO(3) SU (2) that would al-
low us to make the standard representation of SU (2) on C2 into an SO(3)
representation. We could try and define a representation of SO(3) by
M = c0 1 + c1 1 + c2 2 + c3 3
and
ei(c1 1 +c2 2 +c3 3 ) SU (2) U (2)
68
For an arbitrary two-state quantum system, neither the operators M nor the
group SU (2) have any particular geometric significance. In some cases though,
the group SU (2) does have a geometric interpretation, reflecting its role as the
double-cover Spin(3) and the fact that the group SO(3) acts on the physical
system by rotations of three-dimensional space. In these cases, the quantum
system is said to carry spin, in particular spin one-half (we will later on
encounter state spaces of higher spin values). We will from now on assume that
we are dealing with a spin one-half state space.
A standard basis for the observables (besides the unit operator that generates
overall U (1) transformations) is taken to be the operators which give a basis for
the Lie algebra su(2) in its defining representation. We have seen these already
several times, they are the Xj , j = 1, 2, 3, where
j
Xj = i
2
and they satisfy the commutation relations
To make contact with the physics formalism, well define self-adjoint opera-
tors
j
Sj = iXj =
2
1
which will have real eigenvalues 2 .
Note that the conventional definition of these operators in physics texts
includes a factor of ~
~j
Sj = i~Xj =
2
This is because rotations of vectors are defined in physics texts using conjugation
by the matrix
R(, w) = ei ~ wS (compare to our R(, w) = eiwS = ewX )
with the convention of dividing by a factor of ~ appearing here for reasons that
have to do with the action of rotations on functions on R3 that we will encounter
later on. For now, one can either keep factors of ~ out of the definitions of Sj
and the action of rotations, or just assume that we are working in units where
~ = 1.
States in H = C2 that have a well-defined value of the observable Sj will be
the eigenvectors of Sj , with value for the observable the corresponding eigen-
value, which will be 12 . Measurement theory postulates that if we perform the
measurement corresponding to Sj on an arbitrary state |i, then
we will with probability c+ get a value of + 12 and leave the state in an
eigenvector |j, + 21 i of Sj with eigenvalue + 12
we will with probability c get a value of 12 and leave the state in an
eigenvector |j, 21 i of Sj with eigenvalue 21
69
where if
1 1
|i = |j, + i + |j, i
2 2
we have
||2 ||2
c+ = , c =
||2 + ||2 || + ||2
2
R(, w) = ewX
= cos( )1 i(w ) sin( )
2 2
In the case of a spin one-half particle, the group Spin(3) = SU (2) acts on
states by the spinor representation with the element R(, w) SU (2) acting as
|i R(, w)|i
Taking adjoints and using unitarity, one has (thinking of vectors as column
vectors, elements of the dual space as row vectors) the following action on the
dual state space
h| h|R(, w)1
The operators Sj transform under this same group according to the vector
representation of SU (2). Recall that an SO(3) rotation on vectors is given by
conjugating by an SU (2) = Spin(3) group element according to
70
We see that if our observables transform as vectors, and states as spinors,
the expectation values remain invariant:
h|Sj |i h|R(, w)1 R(, w)Sj R(, w)1 R(, w)|i = h|Sj |i
and all eigenvalues of observables remain invariant. One can also interpret
these joint transformations on states and observables as simply a change of
coordinates, a rotation from the standard basis to a different one.
Next semester in this course we may get to the physics of electromagnetic
fields and how particles interact with them in quantum mechanics, but for now
all we need to know is that for a spin one-half particle, the spin degree of freedom
that we are describing by H = C2 has a dynamics described by the Hamiltonian
H = B
Here B is the vector describing the magnetic field, and
e
=g S
2mc
is an operator called the magnetic moment operator. The constants that appear
are: e the electric charge, c the speed of light, m the mass of the particle, and g,
a dimensionless number called the gyromagnetic ratio, which is approximately
2 for an electron, about 5.6 for a proton.
The Schrodinger equation is
d i
|(t)i = ( )( B)|(t)i
dt ~
with solution
|(t)i = U (t)|(0)i
where ge|B|
it ge ge B
U (t) = e ~ B = eit 2mc SB = et 2mc XB = et 2mc X |B|
The time evolution of a state is thus given at time t by a rotation about the
B
axis w = |B| by an angle
ge|B|t
2mc
a rotation taking place with angular velocity ge|B|
2mc .
The amount of non-trivial physics that is described by this simple system is
impressive, including:
The Zeeman effect: this is the splitting of atomic energy levels that occurs
when an atom is put in a constant magnetic field. With respect to the
energy levels for no magnetic field, where both states in H = C2 have the
same energy, the term in the Hamiltonian given above adds
ge|B|
4mc
to the two energy levels, giving a splitting between them proportional to
the size of the magnetic field.
71
The Stern-Gerlach experiment: here one passes a beam of spin one-half
quantum systems through an inhomogeneous magnetic field. One can
arrange this in such a way as to pick out a specific direction w, and split
the beam into two components, of eigenvalue + 12 and 21 for the operator
w S.
Nuclear magnetic resonance spectroscopy: one can subject a spin one-
half system to a time-varying magnetic field B(t), which will be described
by the same Schrodinger equation, although now the solution cannot be
found just by exponentiating a matrix. Nuclei of atoms provide spin one-
half systems that can be probed with time and space-varying magnetic
fields, allowing imaging of the material that they make up.
Quantum computing: attempts to build a quantum computer involve try-
ing to put together multiple systems of this kind (qubits), keeping them
isolated from perturbations by the environment, but still allowing inter-
action with the system in a way that preserves its quantum behavior.
The 2012 Physics Nobel prize was awarded for experimental work making
progress in this direction.
72
Using this with
H
Y = A, X = i
~
we find
d H i
AH (t) = [i , AH (t)] = [H, AH (t)]
dt ~ ~
and this equation determines the time evolution of the observables in the Heisen-
berg picture.
Applying this to the case of the spin one-half system in a magnetic field, and
taking for our observable S we find
d i eg~
SH (t) = [H, SH (t)] = i [SH (t) B, SH (t)]
dt ~ ~2mc
We know from earlier that the solution will be
for ge|B| B
U (t) = eit 2mc S |B|
and thus the spin vector observable evolves in the Heisenberg picture by rotating
about the magnetic field vector with angular velocity ge|B|
2mc .
73
elements forms a new geometrical space, called the coset space, often written
S 3 /U (1). This structure is called a fibering of S 3 by circles, and is known
as the Hopf fibration. Try an internet search for various visualizations of the
geometrical structure involved, a surprising decomposition of three-dimensional
space into non-intersecting curves.
The same space can be represented in a different way, as C2 /C , by taking
all elements of C2 and identifying those related by muliplication by a non-zero
complex number. If we were just using real numbers, R2 /R can be thought of
as the space of all lines in the plane going through the origin.
One sees that each such line hits the unit circle in two opposite points, so
this set could be parametrized by a semi-circle, identifying the points at the
two ends. This space is given the name RP 1 , the real projective line, and
the analog space of lines through the origin in Rn is called RP n1 . What we
are interested in is the complex analog CP 1 , which is often called the complex
projective line.
To better understand CP 1 , one would like to put coordinates on it. A
standard way to choose such a coordinate is to associate to the vector
z1
C2
z2
74
points on the complex plane. There is however one problem with this coordinate:
the coset of
1
0
does not have a well-defined value: as one approaches this point one moves off
to infinity in the complex plane. In some sense the space CP 1 is the complex
plane, but with a point at infinity added.
and the equations relating coordinates (X1 , X2 , X3 ) on the sphere and the
complex coordinate z1 /z2 = z = x + iy on the plane are given by
X1 X2
x= , y=
1 X3 1 X3
75
and
2x 2y x2 + y 2 1
X1 = , X2 = 2 , X3 = 2
x2 2
+y +1 2
x +y +1 x + y2 + 1
takes
z1 z +
z=
z2 z +
76
Notice that the inner product on vectors in H does not correspond at all
to the inner product of unit vectors in R3 . The North and South poles of the
Bloch sphere correspond to orthogonal vectors in H, but they are not at all
orthogonal thinking of the corresponding points on the Bloch sphere as vectors
in R3 . Similarly, eigenvectors for S1 and S2 are orthogonal on the Bloch sphere,
but not at all orthogonal in H.
Many of the properties of the Bloch sphere parametrization of states in H are
special to the fact that H = C2 . In the next class we will study systems of spin
n n
2 , where H = C . In these cases there is still a two-dimensional Bloch sphere,
but only certain states in H are parametrized by it. We will see other examples
of systems with coherent states analogous to the states parametrized by the
Bloch sphere, but the case H has the special property that all states (up to
scalar multiplication) are such coherent states.
77
78
Chapter 8
Representations of SU (2)
and SO(3)
1
8.0.2 The spin 2
representation
This is the representation (1 , C2 ) that comes automatically from the definition
of SU (2) as 2 by 2 complex matrices, since these act on C2 . In other words, 1
79
is just the identity map on SU (2) matrices.
The same construction gives an SU (2) representation for any SO(3) representa-
tion. As we have seen though with the case of the spinor representation, there
are SU (2) representations that are NOT SO(3) representations.
80
weights of the SU (2) representation V . This is the subgroup of elements of
SU (2) of the form i
e 0
0 ei
Our decomposition of an SU (2) representation (, V ) into irreducible repre-
sentations of this U (1) subgroup equivalently means that we can choose a basis
of V so that
iq
i e 1 0 0
eiq2
e 0 0 0
i =
0 e
0 0 eiqm
(0 , C) = C0
1
and the spinor or spin 2 representation decomposes into U (1) irreducibles of
weights 1, +1:
(1 , C2 ) = C1 C+1
.
81
For the vector or spin 1 representation, recall that our double-cover homo-
morphism takes
i cos 2 sin 2 0
e 0
SU (2) sin 2 cos 2 0 SO(3) U (3)
0 ei
0 0 1
From problem 3 on our first problem set, you can see that the upper left diag-
onal 2 by 2 block acts on C2 with weights 2, +2, whereas the bottom right
element acts trivially on the remaining part of C3 , which is a one-dimensional
representation of weight 0. So, the spin 1 representation decomposes as
(2 , C3 ) = C2 C0 C+2
1
Xj = i j
2
which satisfy the commutation relations
82
We have seen that restricting the representation (, V ) to the diagonal U (1)
subgroup of SU (2) and decomposing into irreducibles tells us that we can choose
a basis of V so that
with going around U (1) once as goes from 0 to 2, this means we can choose
a basis of V so that
iq
e 1 0 0
0 eiq2 0
(ei2S3 ) =
0 0 eiqm
Taking the derivative of this representation to get a Lie algebra representation,
using
d
0 (X) = (eX )|=0
d
we find for X = i2S3
iq
e 1 0 0 iq1 0 0
d iq2
0 e 0 0 iq2 0
0 (i2S3 ) =
=
d
0 0 eiqm |=0 0 0 iqm
We see that 0 (S3 ) will have half-integral values, and make the following
definitions
83
Definition (Weights and Weight Spaces). If 0 (S3 ) has an eigenvalue k
2, we
say that k is a weight of the representation (, V ).
The subspace Vk V of the representation V satisfying
k
v Vk = 0 (S3 )v = v
2
is called the kth weight space of the representation. All vectors in it are eigen-
vectors of S3 with eigenvalue k2 .
The dimension dim Vk is called the multiplicity of the weight k in the rep-
resentation (, V ).
One cant diagonalize any combination of S1 and S2 on Vk , but two specific
linear combinations of them do something interesting:
Definition (Raising and lowering operators). Let
0 1 0 0
S+ = S1 + iS2 = , S = S1 iS2 =
0 0 1 0
We have S+ , S sl(2, C). These are neither self-adjoint nor skew-adjoint, but
satisfy
(S ) = S
and similarly we have
0 (S ) = 0 (S )
We call 0 (S+ ) a raising operator for the representation (, V ), and 0 (S )
a lowering operator.
The reason for this terminology is the following calculation:
[S3 , S+ ] = [S3 , S1 + iS2 ] = iS2 + i(iS1 ) = S1 + iS2 = S+
which implies (since 0 is a Lie algebra homomorphism)
0 (S3 ) 0 (S+ ) 0 (S+ ) 0 (S3 ) = 0 ([S3 , S+ ]) = 0 (S+ )
For any v Vk , we have
k
0 (S3 ) 0 (S+ )v = 0 (S+ ) 0 (S3 )v + 0 (S+ )v = ( + 1) 0 (S+ )v
2
so
v Vk = 0 (S+ )v Vk+2
The linear operator 0 (S+ ) takes vectors with a well-defined weight to vectors
with the same weight, plus 2 (thus the terminology raising operator).
A similar calculation shows that 0 (S ) takes Vk to Vk2 , lowering the weight
by 2.
Were now ready to classify all finite dimensional irreducible unitary rep-
resentations (, V ) of SU (2). Finite dimensionality of V implies that if one
repeatedly applies 0 (S+ ) to a vector v V , sooner or later one must get 0
(otherwise one would have an infinite dimensional representation). We define
84
Definition (Highest weights and highest weight vectors). A non-zero vector
v Vn V such that
0 (S+ )v = 0
is called a highest weight vector, with highest weight n.
n, n + 2, , n 2, n
(, V ) = Cn Cn+2 Cn2 Cn
where all the vector spaces are copies of C, and all the maps are isomor-
phisms (multiplications by various numbers, calculable in principle using the
commutation relations).
85
In summary, we see that all irreducible finite dimensional unitary SU (2)
representations are classified up to equivalence by a non-negative integer, the
highest weight n. These representations have dimension n + 1 and we will
denote them (n , V n = Cn+1 ). Note that Vn is the nth weight space, V n is the
representation with highest weight n. The physicists terminology for this uses
not n, but n2 and calls this number the spinof the representation. We have so
far seen the lowest three examples n = 0, 1, 2, or spin s = n2 = 0, 12 , 1, but there
is an infinite class of larger irreducibles, with dim V = n + 1 = 2s + 1.
Using the action of SU (2) on C2 , we will see that this space of functions is
exactly the representation space V n that we need Explicitly, for
g= , g 1 =
86
we can construct the representation as follows:
z
(n (g)f )(z1 , z2 ) =f (g 1 1 )
z2
=f (z1 z2 , z1 + z2 )
Xn
= ak (z1 z2 )nk (z1 + z2 )k
k=0
((g)f )(x) = f (g 1 x)
87
here, since polynomial functions are essentially an algebraic construction. Later
on in the course we will need to work with function spaces that require much
more serious consideration of issues in analysis.
Restricting the differential operators 0 (X) to a finite dimensional irreducible
subspace V n , the homogeneous polynomials of degree n, if one chooses a basis
of V n , then the linear operator 0 (X) will be given by a n + 1 by n + 1 matrix.
Clearly though, the expression as a simple first-order differential operator is
much easier to work with. In the examples we will be studying in much of the
rest of the course, the representations under consideration will also be on func-
tion spaces, with Lie algebra representations appearing as differential operators.
Instead of using linear algebra techniques to find eigenvalues and eigenvectors,
the eigenvector equation will be a partial differential equation, and much of the
rest of the course will focus on using Lie groups and their representation theory
to solve such equations.
One issue we havent addressed yet is that of unitarity of the representation.
We need Hermitian inner products on the spaces V n , inner products that will be
preserved by the action of SU (2) that we have defined on these spaces. Up to an
overall scale, there is only one SU (2) invariant inner product on V n (making it
a unitary representation). A standard way to define a Hermitian inner product
on functions on a space M is to define them using an integral: for f , g functions
on M , take their inner product to be
Z
< f, g >= fg
M
88
1
For n = 1, s = 2
z1 , z2
For n = 2, s = 1
1 1
z12 , z1 z2 , z22
2 2
3
For n = 3, s = 2
1 1 1 1
z13 , z12 z2 , z1 z22 , z23
6 2 2 6
89
Taking the derivative, the Lie algebra representation is given by
x1
d d
0 (X)f = (etX )f|t=0 = f (etX x2 )|t=0
dt dt
x3
0 0 0 0 0 1 0 1 0
l1 = 0 0 1 l2 = 0 0 0 l3 = 1 0 0
0 1 0 1 0 0 0 0 0
Well also use elements l = l1 il2 of the complexified Lie algebra to create
raising and lowering operators L = i0 (l ).
As with the SU (2) case, I wont include a factor of ~ as is usual in physics
(e.g. the usual convention is La = i~0 (la )), since for considerations of action
of the rotation group it would just cancel out (physicists define rotations using
i
e ~ La ). The factor of ~ is only of significance when one wants to relate things
to the momentum operator, something that will come up later in the course.
In the SU (2) case, the 0 (Sj ) had half-integral eigenvalues, with the eigen-
values of 0 (2S3 ) the integral weights of the representation. Here the La will
have integer eigenvalues, the weights will be the eigenvalues of 2L3 , which will
be even integers.
90
0 0
0
1
0 0
t x1
d 0 1
0 x )
0 (l1 )f = f (e 2 |t=0
dt
x3
0 0 0 x1
d
= f (0 cos t sin t x2 )|t=0
dt
0 sin t cos t x3
0
d
= f ( x2 cos t + x3 sin t )|t=0
dt
x2 sin t + x3 cos t
0
f f f
=( , , ) x3
x1 x2 x3
x2
f f
=x3 x2
x2 x3
so
0 (l1 ) = x3 x2
x2 x3
and similar calculations give
0 (l2 ) = x1 x3 , 0 (l3 ) = x2 x1
x3 x1 x1 x2
The space of all functions on R3 is much too big: it will give us an infinity of
copies of each finite dimensional representation that we want. Notice that when
SO(3) acts on R3 , it leaves the distance to the origin invariant. If we work in
spherical coordinates (r, , ) (see picture)
91
we will have
x1 =r sin cos
x2 =r sin sin
x3 =r cos
so
r sin cos sin sin cos x1
1 = cos cos cos sin
sin x
r 2
1
r sin sin cos 0 x
3
This is an orthogonal matrix, so one can invert it by taking its transpose, to get
x1 sin cos cos cos sin r
= sin sin cos sin cos 1
x2 r
1
x
cos sin 0 r sin
3
So we finally have
L1 = i0 (l1 ) = i(x3 x2 ) = i(sin + cot cos )
x2 x3
L2 = i0 (l2 ) = i(x1 x3 ) = i( cos + cot sin )
x3 x1
92
L3 = i0 (l3 ) = i(x1 x3 ) = i
x3 x1
and
L+ = i0 (l+ ) = ei ( + i cot ), L = i0 (l ) = ei ( + i cot )
Now that we have expressions for the action of the Lie algebra on functions in
spherical coordinates, our two differential equations saying our function Yll (, )
is of weight l and in the highest-weight space are
l
L3 Yll (, ) = i Y (, ) = lYll (, )
l
and
L+ Yll (, ) = ei ( + i cot )Yll (, ) = 0
The first of these tells us that
Yll (, ) = eil Fl ()
( l cot )Fl ()
with solution
Fl () = Cll sinl
for arbitrary constant Cll . Finally
for m = l, l 1, l 2 , l + 1, l
The functions Ylm (, ) are called spherical harmonics, and they span the
space of complex functions on the sphere in much the same way that the ein
span the space of complex valued functions on the circle (or, equivalently, with
period 2). Unlike the case of polynomials on C2 , for functions on the sphere,
one gets finite numbers by integrating such functions of the sphere, so can define
93
an inner product on these representations for which they are unitary by simply
setting
Z Z 2 Z
< f, g >= f g sin dd = f (, )g(, ) sin dd
S2 =0 =0
The Ylm (, ) are all orthogonal with respect to this inner product. One can de-
rive various general formulas for the Ylm (, ) in terms of Legendre polynomials,
but here well just compute the first few examples, with the proper constants
that give them norm 1 with respect to the chosen inner product.
For the l = 0 representation
r
0 1
Y0 (, ) =
4
For the l = 1 representation
r r r
3 3 1 3
1
Y1 = i 0
sin e , Y1 = cos , Y1 = sin ei
8 4 8
(one can easily see that these have the correct eigenvalues for 0 (L3 ) = i
).
For the l = 2 representation one has
r r
2 15 2 i2 1 15
Y2 = sin e , Y2 = sin cos ei
32 8
r
0 5
Y2 = (3 cos2 1)
16
r r
1 15 i 2 15
Y2 = sin cos e , Y2 = sin2 ei2
8 32
We will see later that these functions of the angular variables in spherical
coordinates are exactly the functions that give the angular dependence of wave
functions for the physical system of a particle in a spherically symmetric po-
tential. In such a case the SO(3) symmetry of the system implies that state
space (the wave functions) will provide a unitary representation of SO(3), and
the action of the Hamiltonian operator H will commute with the action of the
operators L3 , L . As a result all of the states in an irreducible representation
component of will have the same energy. States are thus organized into or-
bitals, single states called s orbitals (l = 0), triplet states called p orbitals
(l = 1), multiplicity 5 states called d orbitals (l = 2), etc.
94
Definition (Casimir operator for SO(3)). The Casimir operator for the repre-
sentation of SO(3) on functions on S 2 is the second-order differential operator
[L2 , 0 (X)] = 0
for any X so(3). Knowing this, a version of Schurs lemma says that L2 will act
on an irreducible representation as a scalar (i.e. all vectors in the representation
are eigenvectors of L2 , with the same eigenvalue). This eigenvalue can be used
to characterize the irreducible representation.
The easiest way to compute this eigenvalue turns out to be to act with L2 on
a highest weight vector. First one rewrites L2 in terms of raising and lowering
operators (note that the following calculation is like the first expression for L2
above and needs reinterpretation as an equation for operators in an arbitrary
representation, but this is not important for the calculation we are doing).
so
L2 = L21 + L22 + L23 = L L+ + L3 + L23
For the representation of SO(3) on functions on S 2 constructed above,
we know that on a highest weight vector of the irreducible representation l
(restriction of to the 2l + 1 dimensional irreducible subspace of functions that
are linear combinations of the Ylm (, )), we have the two eigenvalue equations
L+ f = 0, L3 f = lf
with solution the functions proportional to Yll (, ). Just from these conditions
and our expression for L2 we can immediately find the scalar eigenvalue of L2
since
L2 f = L L+ f + (L3 + L23 )f = 0 + l + l2 = l(l + 1)
We have thus shown that our irreducible representation l can be characterized
as the representation in which L2 acts by the scalar l(l + 1).
In summary, we have two different sets of partial differential equations whose
solutions give the irreducible representation l :
L+ f = 0, L3 f = lf
which are first order equations, with the first using complexification and some-
thing like a Cauchy-Riemann equation, and
L2 f = l(l + 1), L3 f = lf
95
where the first equation is a second order equation, something like a Laplace
equation.
One can compute the explicit second order differential operator L2 in the
representation on functions, it is
in this representation and check that it behaves in the same way as in the SO(3)
case.
96
Chapter 9
Tensor Products,
Entanglement, and
Addition of Spin
If one has two independent quantum systems, with state spaces H1 and H2 ,
the combined quantum system has a description that exploits the mathematical
notion of a tensor product, with the combined state space the tensor product
H1 H2 . Because of the ability to take linear combinations of states, this
combined state space will contain much more than just products of independent
states, including states that are described as entangled, and responsible for
some of the most counter-intuitive behavior of quantum physical system.
This same tensor product construction is a basic one in representation the-
ory, allowing one to construct a new representation (1 2 , W1 W2 ) out of
representations (1 , W1 ) and (2 , W2 ). When we take the tensor product of
states corresponding to two irreducible representations of SU (2) of spins s1 , s2 ,
we will get a new representation (2s1 2s2 , V 2s1 V 2s2 ). It will be reducible,
a direct sum of representations of various spins, a situation we will analyze in
detail.
Starting with a quantum system with state space H that describes a single
particle, one can describe a system of N particles by taking an N -fold tensor
product HN = H H H. A deep fact about the physical world
is that for identical particles, we dont get the full tensor product space, but
only the subspaces either symmetric or antisymmetric under the action of the
permutation group by permutations of the factors, depending on whether our
particles are bosons or fermions. An even deeper fact is that elementary
particles of half-integral spin s must behave as fermions, those of integral spin,
bosons.
Digression. When physicists refer to tensors, they generally mean the ten-
sor fields used in general relativity or other geometry-based parts of physics.
97
The relation to what is being discussed here is only tangential (a pun I realize
after writing that...) A tensor field is a function on a manifold, taking values
in some tensor product of copies of the tangent space and and its dual space.
The simplest tensor fields are just vector fields, functions taking values in the
tangent space. A more non-trivial example is the metric tensor, which takes
values in the dual of the tensor product of two copies of the tangent space.
98
There are natural isomorphisms
C V ' V, V W ' W V
and
U (V W ) ' (U V ) W
for vector spaces U, V, W
Given a linear operator A on V and another linear operator B on W , we
can define a linear operator A B on V W by
(A B)(v w) = Av Bw
for v V, w W .
With respect to the bases ei , fj of V and W , A will be a (dim V ) by (dim V )
matrix, B will be a (dim W ) by (dim W ) and A B will be a (dim V )(dim W )
by (dim V )(dim W ) matrix (which one can think of as a (dim V ) by (dim V )
matrix of blocks of size (dim W )).
One often wants to consider tensor products of vector spaces and dual vector
spaces. An important fact is that there is an isomorphism between the tensor
product V W and linear maps from V to W given by identifying lw (l V )
with the linear map
v V l(v)w W
HT = H1 H 2
A Id + Id B
HN = H H H H
| {z }
N times
99
The symmetric group SN acts on this state space, and one has a repre-
sentation (, HN ) of SN as follows. For SN a permutation of the set
{1, 2, . . . , N } of N elements, with || the minimal number of transpositions that
combine to give , on a tensor product of vectors one has
()v = v
anti-symmetric, or fermionic if SN
()v = (1)|| v
Note that in the fermionic case, one cannot have non-zero states in HN
describing two identical particles in the same state w H, a fact that is known
as the Pauli Principle. This has to be the case, since for a transposition
interchanging the two particles, the antisymetric representation acts on the
factor H H by interchanging vectors, taking
ww HH
to itself, while the antisymmetry requires that this state go to its negative, so
the state cannot be non-zero.
Next semester we will see that when one uses quantum fields to describe
systems of identical particles, then this simple behavior under the symmetric
group occurs naturally, but within the purely quantum mechanical formalism it
is a separate physical axiom.
100
Definition (Decomposable and indecomposable vectors). A vector in V W
is called decomposable if it is of the form v w for some v V, w W . If it
cannot be put in this form it is called indecomposable.
Note that our basis vectors of V W are all decomposable since they are
products of basis vectors of V and W . Linear combinations of these basis vectors
however are in general indecomposable. If we think of an element of V W
as a dim V by dim W matrix, with entries the coordinates with respect to our
basis vectors for V W , then for decomposable vectors we get a special class
of matrices, those of rank one.
In the physics context, the language used is:
d
V0 W (X)(v w) = V W (etX )(v w)t=0
dt
d
= (V (etX )v W (etX )w)t=0
dt
d d
=(( V (etX )v) W (etX )w)t=0 + (V (etX )v ( W (etX )w))t=0
dt dt
=(V0 (X)v) w + v (W 0
(X)w)
101
which could also be written
Proof. Outline
One way to prove this result is to use highest-weight theory, raising and
lowering operators, and the formula for the Casimir operator. We will not try
and show the details of how this works out, but a little later use a very different
and quicker argument using characters. However, in outline (for more details,
see for instance section 5.2 of [36]), heres how one could proceed:
One starts by noting that if vn1 Vn1 , vn2 Vn2 are highest weight vectors
for the two representations, vn1 vn2 will be a highest weight vector in the tensor
product representation (i.e. annihilated by n0 1 +n2 (S+ )), of weight n1 + n2 .
So (n1 +n2 , V n1 +n2 ) will occur in the decomposition. Applying n0 1 +n2 (S ) to
vn1 vn2 one gets a basis of the rest of the vectors in (n1 +n2 , V n1 +n2 ). However,
at weight n1 +n2 2 one can find another kind of vector, a highest-weight vector
orthogonal to the vectors in (n1 +n2 , V n1 +n2 ). Applying the lowering operator
to this gives (n1 +n2 2 , V n1 +n2 2 ). As before, at weight n1 + n2 4 one find
another, orthogonal highest weight vector, and gets another representation, with
this process only terminating at weight |n1 n2 |.
A much more straightforward proof using character theory is given in the
next section.
102
characters of irreducible representations are for the case of G = SU (2), and use
this to get a simple argument for the Clebsch-Gordan decomposition of the ten-
sor product of SU (2) representations. For this we dont need general theorems
about the relations of characters and representations, but can directly check
that the irreducible representations of SU (2) correspond to distinct character
functions which are easily evaluated.
Definition (Character). The character of a representation (, V ) of a group G
is the function on G given by
V (g) = T r((g))
V W = V + W , V W = V W
ei(n+1) ei(n+1)
V n =
ei ei
Proof.
V n1 V n2 =V n1 V n2
ei(n2 +1) ei(n2 +1)
=(ein1 + ei(n1 2) + + ei(n1 2) + ein1 )
ei ei
i(n1 +n2 +1) i(n2 n1 +1)
(e i(n1 +n2 +1)
e ) + + (e ei(n2 n1 +1) )
=
ei ei
=V n1 +n2 + V n1 +n2 2 + + V n2 n1
103
9.4.3 Some examples
Some simple examples of how this works are:
V1V1 =V2V0
This says that the four complex dimensional tensor product of two spinor
representations (which are each two complex dimensional) decomposes
into irreducibles as the sum of a three dimensional vector representation
and a one dimensional trivial (scalar) representation.
1 0
Using the basis , for V 1 , the tensor product V 1 V 1 has a basis
0 1
1 1 1 0 0 1 0 0
, , ,
0 0 0 1 1 0 1 1
The vector
1 1 0 0 1
( )V1V1
2 0 1 1 0
These three terms are of weights 2, 0, 2 under the U (1) SU (2) gener-
ated by s3 = i3 su(2). They are symmetric under permutation of the
two factors of V 1 V 1 .
We see that if we take two identical quantum systems with H = V 1 = C2
and make a composite system out of them, if they were bosons we would
get a three dimensional state space, transforming as a vector (spin one)
under SU (2). If they were fermions, we would get a one-dimensional
system of spin zero (invariant under SU (2)). Note that in this second
case we automatically get an entangled state, one that cannot be written
as a decomposable product.
The spin-statistics relation says that if these really correspond to identical
spin-half particles, we must be in the fermionic case. This however only
104
applies to the much more complicated situation of special relativity and
particles with space-time degrees of freedom, which we will come to later
in this course. The bosonic case also occurs in particle physics, when the
SU (2) symmetry group is not the double cover of the group of rotations
in space, but an internal symmetry group acting on elementary particle
labels (in which case it is called isospin). Pions, which in some sense are
composites of two isospin one-half quarks, come in an isospin one triplet,
with names + , 0 , .
Tensor product of three or more spinors:
V 1 V 1 V 1 = (V 2 V 0 ) V 1 = (V 2 V 1 ) (V 0 V 1 ) = V 3 V 1 V 1
This says that the tensor product of three spinor representations decom-
poses as a four dimensional (spin 3/2) representation plus two copies of
the spinor representation.
One can clearly generalize this and consider N -fold tensor products (V 1 )N
of the spinor representation. Taking N high enough one can get any ir-
reducible representation of SU (2) that one wants this way, giving an al-
ternative to our construction using homogeneous polynomials. Doing this
however gives the irreducible as just one component of something larger,
and one needs a method to project out the component one wants. One
can do this using the action of the symmetric group SN on (V 1 )N and
an understanding of the irreducible representations of SN . This relation-
ship between irreducible representations of SU (2) and those of SN coming
from looking at how both groups act on (V 1 )N is known as Schur-Weyl
duality, and generalizes to the case of SU (n), where one looks at N -fold
tensor products of the defining representation of SU (n) matrices on Cn .
For SU (n) this provides perhaps the most straight-forward construction
of all irreducible representations of the group.
105
106
Chapter 10
Well now turn to the problem that conventional quantum mechanics courses
generally begin with: that of the quantum system describing a free particle
moving in physical space R3 . The state space H will be a space of complex-
valued functions on R3 , called wave-functions. There is one crucial sort of
observable to understand: the momentum operator. This operator will have
the same relationship to spatial translations as the Hamiltonian does to time
translations. In both cases, the operators are given by the Lie algebra represen-
tation corresponding to the unitary representation on the quantum state space
of groups of translations (translation in the three space and one time directions
respectively).
One way to motivate the quantum theory of a particle is that, whatever it
is, it should have the same sort of behavior as in the classical case under the
translational and rotational symmetries of space-time. Invoking the classical
relationship between energy and momentum used in non-relativistic mechan-
ics relates the Hamiltonian and momentum operators, giving the conventional
Schr odinger differential equation for the wave-function of a free particle. We
will examine the solutions to this equation, beginning with the case of periodic
boundary conditions, where spatial translations in each direction are given by
the compact group U (1) (whose representations we have studied in detail).
107
operator H, a fact that is expressed as
d
i~ |i = H|i
dt
When H is time-independent, one can understand this equation as reflecting the
existence of a unitary representation (U (t), H) of the group R of time transla-
tions on the state space H. U (t) is a homomorphism from the group R to the
group of unitary transformations of H. It is of the form
i
U (t) = e ~ tH
xax=x+a
df a2 d2 f d
f (x + a) = f (x) + a + 2
+ = ea dx f (x)
dx 2! dx
108
d
In this sense dx is the infinitesimal generator of spatial translations and the Lie
algebra representation of R on functions on R is given by the linear operator
d
dx .
Since we now want to describe quantum systems that depend not just on
time, but on space variables q = (q1 , q2 , q3 ), we will have an action by unitary
transformations of not just the group R of time translations, but also the group
R3 of spatial translations. We will define the corresponding Lie algebra rep-
resentations using self-adjoint operators P1 , P2 , P3 that play the same role for
spatial translations that the Hamiltonian plays for time translations:
Definition (Momentum operators). For a quantum system with state space
H given by complex valued functions of position variables q1 , q2 , q3 , momentum
operators P1 , P2 , P3 are defined by
P1 = i~ , P2 = i~ , P3 = i~
q1 q2 q3
These are given the name momentum operators since we will see that their
eigenvalues have an intepretation as the components of the momentum vector
for the system, just as the eigenvalues of the Hamiltonian have an interpretation
as the energy. Note that while in the case of the Hamiltonian the factor of ~ kept
track of the relative normalization of energy and time units, here it plays the
same role for momentum and length units. It can be set to one if appropriate
choices of units are made.
The differentiation operator is skew-adjoint since, using integration by parts
one has for H
Z + Z + Z +
d d d d
( )dq = ( () ( ))dq = ( )dq
dq dq dq dq
The factor of i in the definition of the Pj is just the same convention used earlier
to multiply the skew-adjoint operator of a unitary Lie algebra representation by
i to get the self-adjoint operator favored by physicists.
Note that the convention for the sign choice here is the opposite from the case
of the Hamiltonian. This means that the sign choice for the Hamiltonian makes
it minus the generator of translations in the time direction. The reason for this
comes from considerations of special relativity, where a relative sign appears in
the treatment of space and time dimensions. We will review this subject in a
later chapter but for now we just need the relationship special relativity gives
between energy and momentum. Space and time are put together in Minkowski
space, which is R4 with indefinite inner product
Energy and momentum are the components of a Minkowski space vector (p0 =
E, p1 , p2 , p3 ) with norm-squared given by the mass-squared:
109
This is the formula for a choice of space and time units such that the speed of
light is 1. Putting in factors of the speed of light c to get the units right one
has
E 2 |p|2 c2 = m2 c4
Two special cases of this are:
For photons, m = 0, and one has the energy momentum relation E = |p|c
For velocities v small compared to c (and thus momenta |p| small com-
pared to mc), one has
p p c|p|2 |p|2
E= |p|2 c2 + m2 c4 = c |p|2 + m2 c2 =
2mc 2m
This is the non-relativistic limit, and we use this energy-momentum rela-
tion to describe particles with velocities small compared to c.
1 1 ~2 2 2 2
H= (P12 + P22 + P32 ) = |P|2 = ( 2 + 2 + 2)
2m 2m 2m q1 q2 q3
The Schr
odinger equation then becomes:
~2 2 2 2 ~2 2
i~ (q, t) = ( 2 + 2 + 2 )(x, t) = (q, t)
t 2m q1 q2 q3 2m
~2 2
HE (q) = E (q) = EE (q)
2m
with eigenvalue E for the Hamiltonian operator and then using the fact that
i
(q, t) = E (q)e ~ tE
i~ (q, t) = H(q, t)
t
110
The solutions E (q) to the time-independent equation are just complex ex-
ponentials proportional to
satisfying
~2 ~2 |k|2
(i)2 |k|2 = =E
2m 2m
We see that the solutions to the Schrodinger equation are all linear combinations
of states |ki labeled by a vector k which are eigenstates of the momentum and
Hamiltonian operators with
~2
Pa |ki = ~ka |ki, H|ki = |k|2 |ki
2m
These are states with well-defined momentum and energy
|p|2
pa = ~ka , E =
2m
so satisfy exactly the same energy momentum relations as those for a classical
non-relativistic particle.
While the quantum mechanical state space H contains states with the clas-
sical energy-momentum relation, it also contains much, much more. A general
state will be a linear combination of such states. At t = 0 one has
X
|i = ck eikq
k
where ck are complex numbers, and the general time-dependent state will be
X |k|2
|(t)i = ck eikq eit~ 2m
k
111
for some constant C. But if we try and compute the norm-squared of one of
our basis states |ki we find
Z Z
hk|ki = C (eikq )(eikq )dq = C 1 dq =
R3 R3
As a result, while our basis states |ki are orthogonal for different values of k,
there is no value of C which will make them of norm-squared one.
In the finite dimensional case, a linear algebra theorem assure us that given a
self-adjoint operator, we can find an orthonormal basis of its eigenvectors. In this
infinite dimensional case this is no longer true, and a much more sophisticated
formalism (the spectral theorem for self-adjoint operators) is needed to replace
the linear algebra theorem. This is a standard topic in treatments of quantum
mechanics aimed at mathematicians emphasizing analysis, but we will not try
and enter into this here. One place to find such a discussion is section 2.1 of
[45].
One way to deal with the normalization problem is to replace the non-
compact space by one of finite volume. Well show how this works for the
simplified case of a single spatial dimension, since once one sees how this works
for one dimension, treating the others the same way is straight-forward. In this
one dimensional case, one replaces R by the circle S 1 . This is equivalent to
the physicists method of imposing periodic boundary conditions, meaning
to define the theory on an interval, and then identify the ends of the interval.
One can then think of the position variable q as an angle and define the inner
product as Z 2
1
< 1 , 2 >= 1 ()2 ()d
2 0
and take as state space
H = L2 (S 1 )
the space of complex-valued square-integrable functions on the circle.
Instead of the translation group R, we have the standard action of the
group SO(2) on the circle. Elements g() of the group are rotations of the circle
counterclockwise by an angle , or if we parametrize the circle by an angle ,
just shifts
+
Recall that in general we can construct a representation on functions from a
group action on a space by
(g)f (x) = f (g 1 x)
(g())() = ( )
112
Lie algebra representation is given by taking the derivative
d
0 (X)f () = f ( )|=0 = f 0 ()
d
so we have
d
0 (X) =
d
This operator is defined on a dense subspace of H = L2 (S 1 ) and is skew-adjoint,
since (using integration by parts)
Z 2
1 d
< 1 , 20 >= 1 2 d
2 0 d
Z 2
1 d d
= ( (1 2 ) ( 1 )2 )d
2 0 d d
= < 10 , 2 >
The eigenfunctions of 0 (X) are just the ein , for n Z, which we will also
write as state vectors |ni. These are also a basis for the space L2 (S 1 ), a basis
that corresponds to the decomposition into irreducibles of
L2 (S 1 )
(, L2 (S 1 )) = nZ (n , C)
n (g()) = ein
The theory of Fourier series for functions on S 1 says that one can expand any
function L2 (S 1 ) in terms of this basis, i.e.
+
X +
X
|i = () = cn ein = cn |ni
n= n=
113
The Lie algebra of the group S 1 is the same as that of the group (R, +),
and the 0 (X) we have found for the S 1 action on functions is related to the
momentum operator in the same way as in the R case (up to a sign, which is a
convention anyway). So, we can use the same momentum operator
d
P = i~
d
which satisfies
P |ni = ~n|ni
By changing space to the compact S 1 we now have momenta that instead of
taking on any real value, can only be integral numbers times ~. Solving the
Schr
odinger equation
P2 ~2 2
i~ (, t) = (, t) = (, t)
t 2m 2m 2
as before, we find
~2 d2
EE () = E ()
2m d2
an eigenvector equation, which has solutions |ni, with
~2 n 2
E=
2m
The general solution to the Schrodinger equaton is
+
~n2
X
(, t) = cn ein ei 2m t
n=
with the cn determined from the initial condition of knowing the wave-function
at time t = 0, according to the Fourier coefficient formula
Z 2
1
cn = ein (, 0)d
2 0
To get something more realistic, we need to take our circle to have an ar-
bitrary circumference L, and we can study our original problem by considering
the limit L . To do this, we just need to change variables from to L ,
where
L
L =
2
The momentum operator will now be
d
P = i~
dL
2~
and its eigenvalues will be quantized in units of L . The energy eigenvalues
will be
2 2 ~2 n2
E=
mL2
114
10.3 The group R and the Fourier transform
In the previous section, we used the trick of replacing the group R of translations
by a compact group S 1 , and then used the fact that unitary representations of
this group are labeled by integers. This made the analysis rather easy, with H =
L2 (S 1 ) and the self-adjoint operator P = i~ behaving much the same as in
the finite-dimensional case: the eigenvectors of P give a countable orthonormal
basis of H. If one wants to, one can think of P as an infinite-dimensional matrix.
Unfortunately, in order to understand many aspects of quantum mechanics,
we cant get away with this trick, but need to work with R itself. One reason
for this is that the unitary representations of R are labeled by the same group,
R, and we will find it very important to exploit this and treat positions and
momenta on the same footing (see the discussion of the Heisenberg group in
the next chapter). What plays the role here of |ni = ein , n Z will be the
|ki = eikq , k R. These are functions on R that are irreducible representations
under the translation action ((a) acts by translation by a)
We can try and mimic the Fourier series decomposition, with the coefficients
cn that depend on the labels of the irreducibles replaced by a function fe(k)
depending on the label k of the irreducible representation of R.
Definition (Fourier transform). The Fourier transform of a function is given
by Z
1
F = (k) =
e eikq (q)dq
2
The definition makes sense for L1 (R), Lebesgue integrable functions on
R. For the following, it is convenient to instead restrict to the Schwartz space
S(R) of functions such that the function and its derivatives fall off faster than
any power at infinity (which is a dense subspace of L2 (R)). For more details
about the analysis and proofs of the theorems quoted here, one can refer to a
standard textbook such as [44].
Given the Fourier transform of , one can recover itself:
Theorem (Fourier Inversion). For e S(R) the Fourier transform of a func-
tion S(R), one has
Z +
1
(q) = Fee = eikq (k)dk
e
2
Note that Fe is the same linear operator as F, with a change in sign of the
argument of the function it is applied to. Note also that we are choosing one
of various popular ways of normalizing the definition of the Fourier transform.
In others, the factor of 2 may appear instead in the exponent of the complex
exponential, or just in one of F or Fe and not the other.
The operators F and Fe are thus inverses of each other on S(R). One has
115
Theorem (Plancherel). F and Fe extend to unitary isomorphisms of L2 (R)
with itself. In other words
Z Z
2 2
|(q)| dq = |(k)|
e dk
116
Since p = ~k, we can easily change variables and work with p instead of k,
and often will do this from now on. As with the factors of 2, theres a choice
of where to put the factors of ~ in the normalization of the Fourier transform.
Well make the following choices, to preserve symmetry between the formulas
for Fourier transform and inverse Fourier transform:
Z +
1 pq
(p)
e = ei ~ dq
2~
Z +
1 pq
(q) = ei ~ dp
2~
Note that in this case we have lost an important property that we had for
finite dimensional H and had managed to preserve by using S 1 rather than
R as our space. If we take H = L2 (R), the eigenvectors for the operator P
(the functions eikq ) are not square-integrable, so not in H. The operator P
is an unbounded operator and we no longer have a theorem saying that its
eigenvectors give an orthornormal basis of H. As mentioned earlier, one way
to deal with this uses a general spectral theorem for self-adjoint operators on a
Hilbert space, for more details see Chapter 2 of [45].
(we wont try and specify which condition on functions f or g is chosen to make
sense of this). There are however some other very obvious linear functionals on
such a function space, for instance the one given by evaluating the function at
q = c:
f f (c)
Such linear functionals correspond to generalized functions, objects which when
fed into the formula for integration over R give the desired linear functional.
The most well-known of these is the one that gives this evaluation at q = c, it
is known as the delta function and written as (q c). It is the object which,
if it were a function, would satisfy
Z +
(q c)f (q)dq = f (c)
117
To make sense of such an object, one can take it to be a limit of actual functions.
For the -function, consider the limit as 0 of
1 (qc)2
g = e 2
2
which satisfy Z +
g (q)dq = 1
for all > 0 (one way to see this is to use the formula given earlier for the
Fourier transform of a Gaussian).
Heuristically (ignoring obvious problems of interchange of integrals that
dont make sense), one can write the Fourier inversion formula as follows
Z +
1
(x) = eikq (k)dk
e
2
Z + Z +
1 1 0
= eikq ( eikq (q 0 )dq 0 )dk
2 2
Z + Z +
1 0
= ( eik(qq ) (q 0 )dk)dq 0
2
Z +
= (q 0 q)(q 0 )dq 0
118
10.4 For further reading
Every book about quantum mechanics covers this example of the free quantum
particle somewhere very early on, in detail. Our discussion here is unusual
just in emphasizing the role of the spatial translation groups and its unitary
representations. Discussions of quantum mechanics for mathematicians (such
as [45]) typically emphasize the development of the functional analysis needed
for a proper description of the Hilbert space H and of the properties of general
self-adjoint operators on this state space. In this class were restricting attention
to a quite limited set of such operators coming from Lie algebra representations,
so will avoid the general theory.
119
120
Chapter 11
In our discussion of the free particle, we used just the actions of the groups R3 of
spatial translation and the group R of time translation, finding corresponding
observables, the self-adjoint momentum (P ) and Hamiltonian (H) operators.
Weve seen though that the Fourier transform involves a perfectly symmetrical
treatment of position and momentum variables. This allows us to introduce a
position operator Q acting on our state space H. We will analyze in detail in
this chapter the implications of extending the algebra of observable operators in
this way, most of the time restricting to the case of a single spatial dimension,
since the physical case of three dimensions is an easy generalization.
The P and Q operators generate an algebra named the Heisenberg algebra,
since Werner Heisenberg and collaborators used it in the earliest work on a full
quantum-mechanical formalism in 1925. It was quickly recognized by Hermann
Weyl that this algebra comes from a Lie algebra representation, with a cor-
responding group (called the Heisenberg group by mathematicians, the Weyl
group by physicists). The state space of a quantum particle, either free or mov-
ing in a potential, will be a unitary representation of this group, with the group
of spatial translations a subgroup. Note that this particular use of a group
and its representation theory in quantum mechanics is both at the core of the
standard axioms and much more general than the usual characterization of the
significance of groups as symmetry groups. While the group of spatial trans-
lations is a symmetry of the theory of a free particle, the action of the larger
Heisenberg group on the state space H does not commute with any non-zero
Hamiltonian operator. The Heisenberg group does not in any sense correspond
to a group of invariances of the physical situation, but rather plays a much
deeper role.
121
11.1 The position operator and the Heisenberg
Lie algebra
In the description of the state space H as functions of a position variable q, the
momentum operator is
d
P = i~
dq
The Fourier transform F provides a unitary transformation to a description of
H as functions of a momentum variable p in which the momentum operator P
is just multiplication by p. Exchanging the role of p and q, one gets a position
operator Q that acts as
d
Q = i~
dp
when states are functions of p (the sign difference comes from the sign change
in Fe vs. F), or as multiplication by q when states are functions of q.
and in particular
1 pq
hq|pi = ei ~
2~
122
11.1.2 Momentum space representation
In the momentum space description of H as functions of p0 , the state is the
Fourier transform of the state in the position space representation, so one has
Z + Z +
1 i pq
0
1 0 0
i p ~q i pq
0
0 1 pp0 0
|pi = F( e ~ )= e e ~ dq = ei ~ q dq 0 = (pp0 )
2~ 2~ 2~
d 1 p0 q 1 p0 q
Q|qi = i~ 0
( ei ~ ) = q( ei ~ ) = q|qi
dp 2~ 2~
Another way to see that this is the correct operator is to use the unitary trans-
formation F and its inverse Fe that relate the position and momentum space
representations. Going from position space to momentum space one has
Q FQFe
and one can check that this transformed Q operator will act as i~ dpd 0 on functions
of p0 .
Using the representation independent notation, one has
Z +
1 pq 0
hp|i = ( ei ~ )(q 0 )dq 0 = (p)
e
2~
123
momentum relationship. Note that for the free particle P commutes with the
P2
Hamiltonian H = 2m so there is a conservation law: states with a well-defined
momentum at one time always have the same momentum. This corresponds to
an obvious physical symmetry, the symmetry under spatial translations.
The operator Q on the other hand does not correspond to a physical sym-
metry, since it does not commute with the Hamiltonian. We will see that it
does generate a group action, and from the momentum space picture we can
see that this is a shift in the momentum, but such shifts are not symmetries of
the physics and there is no conservation law for Q. The states in which Q has a
well-defined numerical value are the ones such that the position wave-function
is a delta-function. If one prepares such a state at a given time, it will not
remain a delta-function, but quickly evolve into a wave-function that spreads
out in space.
Since the eigenfunctions of P and Q are non-normalizable, one needs a
slightly different formulation of the measurement theory principle used for finite
dimensional H. In this case, the probability of observing a position of a particle
with wave function (q) in the interval [q1 , q2 ] will be
R q2
q1
(q)(q)dq
R +
(q)(q)dq
This will make sense for states |i L2 (R), which we will normalize to have
norm-squared one when discussing their physical interpretation. Then the sta-
tistical expectation value for the measured position variable will be
h|Q|i
which can be computed with the same result in either the position or momentum
space representation.
Similarly, the probability of observing a momentum of a particle with momentum-
space wave function (q)
e in the interval [p1 , p2 ] will be
R p2
p1
(p)
e (p)dp
e
R +
(p)
e (p)dp
e
and for normalized states the statistical expectation value of the measured mo-
mentum is
h|P |i
Note that states with a well-defined position (the delta-function states in
the position-space representation) are equally likely to have any momentum
whatsoever. Physically this is why such states quickly spread out. States with
a well-defined momentum are equally likely to have any possible position. The
properties of the Fourier transform imply the so-called Heisenberg uncertainty
principle that gives a lower bound on the product of a measure of uncertainty
in position times the same measure of uncertainty in momentum. Examples
124
of this that take on the lower bound are the Gaussian shaped functions whose
Fourier transforms were computed earlier.
For much more about these questions, again most quantum mechanics text-
books will contain an extensive discussion.
Soon after this commutation relation appeared in early work on quantum me-
chanics, Weyl realized that it can be interpreted as the relation between oper-
ators one would get from a representation of a three-dimensional Lie algebra,
now called the Heisenberg Lie algebra.
Definition (Heisenberg Lie algebra). The Heisenberg Lie algebra h3 is the vec-
tor space R3 with the Lie bracket defined by its values on a basis (X, Y, Z) by
Note that this is a non-abelian Lie algebra, but only minimally so. All Lie
brackets of Z are zero. All Lie brackets of Lie brackets are also zero (as a result,
this is an example of what is known as a nilpotent Lie algebra).
The Heisenberg Lie algebra is isomorphic to the Lie algebra of 3 by 3 strictly
upper triangular real matrices, with Lie bracket the matrix commutator, by the
following isomorphism:
0 1 0 0 0 0 0 0 1
X 0 0 0 , X 0 0 1 , Z 0 0 0
0 0 0 0 0 0 0 0 0
0 x z
xX + yY + zZ 0 0 y
0 0 0
since one has
x0 z0 xy 0 x0 y
0 x z 0 0 0
[0 0 y , 0 0 y 0 ] = 0 0 0
0 0 0 0 0 0 0 0 0
125
For a higher-dimensional generalization of this, one just replaces x and y by
n-dimensional vectors x and y, giving a Lie algebra h2n+1 . The physical case is
n = 3, where elements of the Heisenberg Lie algebra can be written
0 x1 x2 x3 z
0 0 0 0 y3
0 0 0 0 y2
0 0 0 0 y1
0 0 0 0 0
1 x z + 12 xy
0 x z
exp 0 0 y = 0 1 y
0 0 0 0 0 1
so the group with Lie algebra h3 will be the group of upper triangular 3 by 3 real
matrices with ones on the diagonal, and this group will be the Heisenberg group
H3 . For our purposes though, it is better to work in exponential coordinates (i.e.
identifying a group element with the Lie algebra element that exponentiates
to it). Matrix exponentials in general satisfy the Baker-Campbell-Hausdorff
formula, which says
1
eA eB = eA+B+ 2 [A,B]+
where the higher terms all involve commutators of commutators (one can check
this term by term by expanding the exponentials, for a proof, see chapter 3
of [24]). For the case of the Heisenberg Lie algebra, since commutators of
commutators vanish, we get the simpler formula for exponentials of elements of
h3
1
eA eB = eA+B+ 2 [A,B]
and we can use this to explicitly write the group law in exponential coordinates.
1
(x, y, z) (x0 , y 0 , z 0 ) = (x + x0 , y + y 0 , z + z 0 + (xy 0 yx0 ))
2
Note that the Lie algebra basis elements X, Y, Z each generate subgroups
of H3 isomorphic to R. Elements of the first two of these subgroups generate
the full group, and elements of the third subgroup are central, meaning they
commute with all group elements. Also notice that the non-commutativity of
the Lie algebra or group depends purely on the factor xy 0 yx0 , which for the
generalization to h2n+1 of H2n+1 can be written as x y0 y x0 .
126
11.4 The Schr
odinger representation
Since it can be defined in terms of 3 by 3 matrices, the Heisenberg group H3
has an obvious representation on C3 , but this representation is not unitary and
not of physical interest. What is of great interest is the representation of the
Lie algebra h3 on functions of q given by the Q and P operators:
Definition (Schrodinger representation). The Schr odinger representation of the
Heisenberg Lie algebra h3 is the representation (~ , L2 (R)) satisfying
d
0~ (X)(q) = iQ(q) = iq(q), 0~ (Y )(q) = iP (q) = ~ (q)
dq
0~ (Z)(q) = i~(q)
Factors of i have been chosen to make these operators skew-Hermitian. They
can be exponentiated, giving the unitary representation (~ , L2 (R)) of the group
H3 . Elements of H3 of the form g = exX give one subgroup R H3 and act by
multiplication by a phase depending on q
Those of the form g = eyY give another subset R H3 and act by translation
in q
d
~ (eyY )(q) = eyiP (q) = ey~ dq (q) = (q y~)
while those of the form g = ezZ act by multiplication by a phase
The group analog of the Heisenberg commutation relations (often called the
Weyl form of the commutation relations) is the relation
as well as the same product in the opposite order, and then comparing the
results.
We have seen that the Fourier transform F takes the Schrodinger represen-
tation to a unitarily equivalent representation of H3 , in terms of functions of p
(the momentum space representation). The operators change as
~ (g) F ~ (g)Fe
when one makes the unitary transformation to the momentum space represen-
tation.
127
In typical physics quantum mechanics textbooks, one often sees calculations
made just using the Heisenberg commutation relations, without picking a spe-
cific representation of the operators that satisfy these relations. This turns
out to be justified by the remarkable fact that, for the Heisenberg group, once
one picks the value of ~ with which Z acts, all irreducible representations are
unitarily equivalent. In a sense, the representation theory of the Heisenberg
group is very simple: theres just one irreducible representation. This is very
different than the theory for even the simplest compact Lie groups (U (1) and
SU (2)) which have an infinity of inequivalent irreducibles labeled by weight or
by spin. Representations of a Heisenberg group will appear in different guises
(weve seen two, will see another in the discussion of the harmonic oscillator,
and there are yet others that appear in the theory of theta-functions), but they
are all unitarily equivalent. This statement is known as the Stone-von Neumann
theorem.
So far weve been modestly cavalier about the rigorous analysis needed to
make precise statements about the Schrodinger representation. In order to prove
a theorem like the Stone-von Neumann theorem, which tries to say something
about all possible representations of a group, one needs to invoke a great deal
of analysis. Much of this part of analysis was developed precisely to be able to
deal with general quantum mechanical systems and prove theorems about them.
The Heisenberg group, Lie algebra and its representations are treated in detail
in many expositions of quantum mechanics for mathematicians. Some excellent
references for this material are [45], and [26]. In depth discussions devoted to
the mathematics of the Heisenberg group and its representations can be found
in [30] and [18].
In these references can be found a proof of (not difficult)
Theorem. The Schr
odinger representation ~ described above is irreducible.
and the much more difficult
Theorem (Stone-von Neumann). Any irreducible representation of the group
H3 on a Hilbert space, satisfying
0 (Z) = i~1
odinger representation (~ , L2 (R))
is unitarily equivalent to the Schr
Because of the unitary equivalence ensured by Stone-von Neumann, we will
sometimes be cavalier and use the symbol ~ to refer to any representation
in the unitary equivalence class, without specifying which one we mean (e.g.
Schrodinger representation as a function of q or as a function of p).
Note that all of this can easily be generalized to the case of n spatial di-
mensions, with the Heisenberg group now H2n+1 and the Stone-von Neumann
theorem still true. If one tries however to take n , which is what happens
in quantum field theory, the Stone-von Neumann theorem breaks down. In the
case of quantum field theory one has an infinity of inequivalent irreducible repre-
sentations of the commutation relations to consider, one source of the difficulties
of the subject.
128
It is also important to note that the Stone-von Neumann theorem is for-
mulated for Heisenberg group representations, not for Heisenberg Lie algebra
representations. For infinite-dimensional representations in cases like this, there
are representations of the Lie algebra that are non-integrable: they arent the
derivatives of Lie group representations. For representations of the Heisen-
berg Lie algebra, i.e. the Heisenberg commutator relations, there are counter-
examples to the Stone von-Neumann theorem. It is only for integrable rep-
resentations that the theorem holds and one has a unique sort of irreducible
representation.
129
130
Chapter 12
We have seen that the quantum theory of a free particle corresponds to the con-
struction of a representation of the Heisenberg Lie algebra in terms of operators
Q and P . One would like to use this to produce quantum systems with a similar
relation to more non-trivial classical mechanical systems than the free particle.
From the earliest days of quantum mechanics it was recognized that the commu-
tation relations of the Q and P operators somehow corresponded to the Poisson
bracket relations between the position and momentum variables in classical me-
chanics in the Hamiltonian formalism. In this chapter well begin by reviewing
the topic of Hamiltonian mechanics and the Poisson bracket, and go on to study
the Lie algebra structure determined by the Poisson bracket. The Lie algebra
here is infinite dimensional, so not the sort of finite dimensional Lie algebra
given by matrices that we have studied so far (although, historically, it is the
sort of Lie algebra that motivated the discovery of the theory of Lie groups and
Lie algebras by Sophus Lie during the 1870s). It does have certain important
finite-dimensional subalgebras which we will study, one of which is isomorphic
to the Heisenberg Lie algebra h2d+1 seen in the last chapter, the other of which
is something new: the Lie algebra of the symplectic group Sp(2d, R).
131
Definition (Poisson bracket). There is a bilinear operation on functions on the
phase space R2 (with coordinates (p, q)) called the Poisson bracket, given by
f g f g
(f, g) {f, g} =
q p p q
df
= {f, h}
dt
Taking f to be just the functions p and q, one gets Hamiltons equations for
classical trajectories in phase space
h
p = {p, h} =
q
h
q = {q, h} =
p
p2
For a non-relativistic free particle, h = 2m and these equations become
p
p = 0, q =
m
which just says that the momentum is the mass times the velocity, and is con-
served. For a particle subject to a potential V (q) one has
p2
h= + V (q)
2m
and the trajectories are the solutions to
V p
p = , q =
q m
Anti-symmetry
{f, g} = {g, f }
132
Jacobi identity
These two properties, together with the bilinearity, show that the Poisson
bracket fits the definition of a Lie bracket, making the space of functions on
phase space into an infinite dimensional Lie algebra. There is a corresponding
group, known to physicists as the group of canonical transformations, but it
is also infinite dimensional, so not a matrix group of the sort we have been
considering so far. This Lie algebra is responsible for much of the structure of
the subject of Hamiltonian mechanics, and it was historically the first sort of
Lie algebra to be studied.
The role of symmetries in classical mechanics is best seen using this Lie
algebra. From the fundamental dynamical equation
df
= {f, h}
dt
we see that
df
{f, h} = 0 = =0
dt
and in this case the function f is called a conserved quantity, since it does
not change under time evolution. Note that if we have two functions f and g
on phase space such that
{f, h} = 0, {g, h} = 0
This property says that taking Poisson bracket with a function h acts on a
product of functions in a way that satisfies the Leibniz rule for what happens
when you take the derivative of a product. This is exactly what one expects,
since one can think of h as a Hamiltonian function and then the dynamical
equation
df
{f, h} =
dt
says that the operation of Poisson bracketing a function by h is a derivative: the
time derivative along the classical trajectory corresponding to the Hamiltonian
h.
133
Taking Poisson brackets with the linear functions p and q gives the deriva-
tives
f f
{f, p} = , {f, q} =
q p
and the Leibniz rule allows one to compute recursively what will happen when
one takes Poisson bracket by any polynomial in p and q. This will always give
some first-order partial differerential operator. A simple example would be the
action of pq which would be
f f
{f, pq} = p{f, q} + {f, p}q = p +q
p q
Notice that taking Poisson bracket with p acts on functions of q as an infinites-
imal translation, so corresponds to the symmetry of spatial translation. Other
symmetries of a Hamiltonian system will be infinitesimally generated by func-
tions in a way similar to the case of the function p , with the Poisson bracket
between such functions giving the Lie bracket relation in the Lie algebra of the
symmetry group.
Digression. If you are familiar with differential manifolds, vector fields and
differential forms, a more general notion of a phase space than just M = R2d
is the following sort of manifold:
Definition (Symplectic manifold). A symplectic manifold M is a manifold with
a two-form (, ) (called a symplectic two-form) satisfying the conditions
is non-degenerate, i.e. for a nowhere zero vector field X, (X, ) is a
nowhere zero one-form.
d = 0, in which case is said to be closed.
The cotangent bundle T N of a manifold N (i.e. the space of pairs of a
point on N together with a linear function on the tangent space at that point)
provides one class of symplectic manifolds, generalizing the linear case N = Rd ,
and corresponding physically to a particle moving on N . A simple example that
is neither linear nor a cotangent bundle is the sphere M = S 2 , with the area
two-form.
Note that there is no assumption here that M has a metric (i.e. it may
not be a Riemannian manifold). A symplectic two-form is a structure on a
manifold analogous to a metric but with opposite symmetry properties. Whereas
a metric is a symmetric non-degenerate bilinear form on the tangent space at
each point, a symplectic form is an anti-symmetric non-degenerate bilinear form
on the tangent space.
Using the symplectic form , for any symplectic manifold M , one has a
generalization of Hamiltons equations. This is the following equality of one-
forms, giving the relation between a Hamiltonian function h and the vector field
Xh that determines time evolution on M
iXh = (Xh , ) = dh
134
The Poisson bracket in this context can be defined as
{f, g} = (Xf , Xg )
and the fact that the Poisson bracket gives a Lie bracket structure on functions
corresponds to the existence of a Lie bracket on vector fields, satisfying
[Xf , Xg ] = X{f,g}
(the minus sign is due to the fact that vector fields are a Lie algebra repre-
sentation on the space of functions on M , with the standard inverse appearing
when going from a group action on M to the corresponding representation on
functions on M ).
The condition d = 0 is then equivalent to the condition that the Poisson
bracket satisfies the Jacobi identity. The infinite dimensional Lie algebra of
functions on M with Lie bracket the Poisson bracket is the Lie algebra for an
infinite dimensional group, the group of all diffeomorphisms of M that preserve
. For much more about symplectic geometry and Hamiltonian mechanics, a
good reference is [7].
From the point of view of symplectic geometry, at each point the fundamental
structure is the anti-symmetric bilinear form on tangent vectors given by .
In discussions of quantization however, what is most important is the Poisson
bracket which, in the linear approximation near each point, is an anti-symmetric
bilinear form on linear combinations of the local coordinate functions (qi , pi ).
These lie not in the tangent space but in its dual space, and it is important to
keep that in mind as we go on to exploit symmetries based on the Poisson bracket,
rather than the closely related symplectic form. Our discussion of quantization
will rely crucially on having a linear structure on phase space, so will not apply
to general symplectic manifolds.
so it is these simple relations (together with the Leibniz rule) which fundamen-
tally determine the entire infinite dimensional Lie algebra structure. The vector
space of linear polynomials on M is three dimensional, with basis elements q, p, 1
and this space is closed under Poisson bracket, so gives a three dimensional Lie
algebra. This Lie algebra is isomorphic to the Heisenberg Lie algebra h3 with
the isomorphism given by
X q, Y p, Z 1
135
This isomorphism preserves the Lie bracket relations since
[X, Y ] = Z {q, p} = 1
The homogeneous quadratic polynomials in p and q form another three-
dimensional sub-Lie algebra of the Lie algebra of functions on phase space,
since the non-zero Poisson bracket relations between them are
q 2 p2
{ , } = qp {qp, p2 } = 2p2 {qp, q 2 } = 2q 2
2 2
One can identify this Lie algebra with a Lie algebra of 2 by 2 matrices as
follows
p2 q2
0 1 0 0 1 0
E= F = qp G =
2 0 0 2 1 0 0 1
The commutation relations amongst these matrices are
[E, F ] = G [G, E] = 2E [G, F ] = 2F
which are the same as the Poisson bracket relations between the corresponding
quadratic polynomials. Mathematicians should note that what we have called
G here is more usually denoted H, but in this context using that symbol would
just be too confusing due to our use of it to denote the Hamiltonian.
The matrices E, F, G give a basis of the real 2 by 2 matrices of trace zero,
which span the Lie algebra sl(2, R). So our Lie algebra of quadratic polynomials
is the Lie algebra of the group SL(2, R) of 2 by 2 real matrices of determinant
1. This three-dimensional Lie group is non-compact and quite unlike SU (2). In
particular, all of its non-trivial irreducible unitary representations are infinite-
dimensional, forming an important topic in mathematics, but one that is beyond
the scope of this course.
Two important subgroups of SL(2, R) are
The subgroup of elements one gets by exponentiating g, which is isomor-
phic to the multiplicative group of positive real numbers
t
e 0
etG =
0 et
Here one can explicitly see that this group has elements going off to infinity.
Exponentiating the Lie algebra element E F gives rotations of the plane
cos sin
e(EF ) =
sin cos
Note that the Lie algebra element being exponentiated here is
1 2
EF (p + q 2 )
2
which we will later reencounter as the Hamiltonian function for the har-
monic oscillator.
136
We have found two three-dimensional Lie algebras (h3 and sl(2, R)) as subal-
gebras of the infinite dimensional Lie algebra of functions on phase space. They
do not commute with each other, since one has the following non-zero Poisson
bracket relations between them
{pq, p} = p, {pq, q} = q
p2 q2
{ , q} = p, { , p} = q
2 2
Taking all quadratic polynomials together, we get a six-dimensional Lie algebra
with basis elements 1, q, p, qp, q 2 , p2 . This is not the product of h3 and sl(2, R),
but an example of a structure that we will study in detail later on, called a
semi-direct product and written
h3 o sl(2, R)
137
so
a b
det = ad bc = 1
c d
which say that the change of basis matrix must be an element of SL(2, R). The
infinitesimal version of this SL(2, R) action on M is given by considering the
action of elements of sl(2, R) on the basis elements q, p. Identifying elements
of sl(2, R) with linear combinations of q 2 , p2 , pq, this action is given by taking
Poisson brackets of q, p elements of sl(2, R) using the Poisson bracket relations
{pq, p} = p, {pq, q} = q
p2 q2
{ , q} = p, { , p} = q
2 2
In higher dimensions, choosing a basis for the phase space M = R2d gives
a basis {q1 , p1 , qd , pd } for the dual space M = M consisting of the linear
coefficient functions of vectors in M . Taking as an additional basis element the
constant function 1, we have a 2d + 1 dimensional space
MR
with basis {q1 , p1 , qd , pd , 1}, and the Poisson bracket turns this space into a
Lie algebra, the Heisenberg Lie algebra h2d+1 .
This Lie bracket on h2d+1 is explicitly given on linear combinations of the
basis elements by an antisymmetric, non-degenerate bilinear form as follows.
If (u, c) and (u0 , c0 ) are elements of h2d+1 = M R, with
u = cq1 q1 + cp1 p1 + + cqd qd + cpd pd M, c R
u0 = c0q1 q1 + c0p1 p1 + + c0qd qd + c0pd pd M, c0 R
then
[(u, c), (u0 , c0 )] = (u, u0 )1
where
(u, u0 ) ={cq1 q1 + cp1 p1 + + cqd qd + cpd pd , c0q1 q1 + c0p1 p1 + + c0qd qd + c0pd pd }
=cq1 c0p1 cp1 c0q1 + cqd c0pd cpd c0qd
0
0 1 ... 0 0 cq1
1 0 ... 0 0 c0p
1
= cq1 cp1 . . . cqd cpd ... .. .. .. ..
. . . .
0 0 ... 0 1 c0qd
0 0 ... 1 0 c0pd
The group of linear transformations of M that leave invariant (i.e. linear
transformations g such that
(gu, gu0 ) = (u, u0 )
for all u M) is called the symplectic group, so we have seen that it can be
defined explicitly as
138
Definition (Symplectic group Sp(2d, R)). The group Sp(2d, R) is the group of
real 2d by 2d matrices g that satisfy
0 1 ... 0 0 0 1 ... 0 0
1 0 . . . 0 0 1 0 . . . 0 0
T .. .
.. .
.. .
.. g = . .. .. ..
g . ..
. . .
0 0 . . . 0 1 0 0 . . . 0 1
0 0 . . . 1 0 0 0 . . . 1 0
For d = 1 we have Sp(2, R) = SL(2, R), but for higher d the symplectic
groups have no such simple characterization. The Lie algebra of Sp(2d, R) is
denoted sp(2d, R) and it is isomorphic to the Lie algebra of order two homoge-
neous polynomials on M = R2d , with Poisson bracket as Lie bracket.
Our definition of the Poisson bracket determined a specific bilinear form
, but we can define more generally the class of bilinear forms that have the
properties we need:
Definition (Symplectic form). A symplectic form on a vector space V is a
bilinear map
:V V R
such that
is antisymmetric, i.e. (v1 , v2 ) = (v2 , v1 )
is nondegenerate, i.e. if v 6= 0, then (v, ) V is non-zero.
A vector space V with a symplectic form is called a symplectic vector
space. One can show that such a vector space must be even dimensional, and
a basis can always be found such that has exactly the matrix expression we
used earlier. Our interest will be in symplectic vector spaces V that are the
dual vector spaces M for some phase space M = R2d . It is for these that
Poisson brackets correspond to symplectic forms, and for these that we have an
identification with the Lie algebra h2d+1 .
To summarize more abstractly this section, we have shown explicitly that
the group Sp(2d, R) acts on the Lie algebra h2d+1 by automorphisms (maps
from the Lie algebra to itself, preserving the Lie bracket). Using the exponen-
tial map, this gives an action of Sp(2d, R) on the Heisenberg group H2d+1 by
automorphisms (maps from H2d+1 to itself, preserving the group law). When
we study semi-direct products we will see that this is exactly the information
needed to define a semi-direct product group H2d+1 o Sp(2d, R).
139
of a Lie algebra (in this case h2d+1 ) by the Poisson bracket, and it is this Lie
algebra we want to use in the next chapter when we define a quantization of
the classical system. From a purely classical point of view, one can generalize
the notion of phase space and Poisson bracket, defining them purely in terms
of Lie algebras, by starting not with h2d+1 , but with a general Lie algebra g.
Since the dual of a dual is the original vector space, one can think of this g as
the dual of g , i.e.
g = (g )
and then the Lie bracket on g has an interpretation as a Poisson bracket on
linear functions on g . So, for any Lie algebra g, its dual g is a space with a
Poisson bracket, and one can think of this as a generalized classical phase space.
Such a generalized classical phase space is best thought of as a family of
conventional phase spaces. For g = h2d+1 one has
h2d+1 = M R
su(2) = R3
140
Chapter 13
Quantization
Given any Hamiltonian classical mechanical system with phase space R2d , physics
textbooks have a standard recipe for producing a quantum system, by a method
known as canonical quantization. We will see that this is just the construc-
tion we have already seen of a unitary representation 0 of the Heisenberg Lie
algebra, the Schr odinger representation, and the Stone-von Neumann theorem
assures us that this is the unique such construction, up to unitary equivalence.
We will also see that this recipe can only ever be partially successful, with the
Schrodinger representation giving us a representation of a sub-algebra of the al-
gebra of all functions on phase space (the polynomials of degree two and below),
and a no-go theorem showing that this cannot be extended to a representation
of the full infinite dimensional Lie algebra. Recipes for quantizing higher-order
polynomials will alway suffer from a lack of uniqueness, a phenomenon known
to physicists as the existence of operator ordering ambiguities.
In later chapters we will see that this quantization prescription does give
unique quantum systems corresponding to some Hamiltonian systems (in par-
ticular the harmonic oscillator and the hydrogen atom), and does so in a manner
that allows a description of the quantum system purely in terms of representa-
tion theory.
141
was that given any classical Hamiltonian system, one could quantize it by
finding a rule that associates to a function f on phase space a self-adjoint
operator Of (in particular Oh = H) acting on a state space H such that
i
O{f,g} = [Of , Og ]
~
This is completely equivalent to asking for a unitary representation ( 0 , H)
of the infinite dimensional Lie algebra of functions on phase space (with the
Poisson bracket as Lie bracket). To see this, note that one can choose units
for momentum p and position q such that ~ = 1. Then, as usual getting a
skew-adjoint Lie algebra representation operator by multiplying a self-adjoint
operator by i, setting
0 (f ) = iOf
the Lie algebra homomorphism property
corresponds to
iO{f,g} = [iOf , iOg ] = [Of , Og ]
so one has Diracs suggested relation.
Recall that the Heisenberg Lie algebra is isomorphic to the three-dimensional
sub-algebra of functions on phase space given by linear combinations of the con-
stant function, the function q and the function p. The Schrodinger representa-
tion ~=1 = provides a unitary representation not of the Lie algebra of all
functions on phase space, but of these polynomials of degree at most one, as
follows
O1 = 1, Oq = Q, Op = P
so
d
0 (1) = i1, 0 (q) = iQ = iq, 0 (p) = iP =
dq
Moving on to quadratic polynomials, these can also be quantized using the
Schr
odinger representation, as follows
P2 p2 P2 i d2
O p2 = , 0 ( ) = i =
2 2 2 2 2 dq 2
Q2 q2 Q2 i
O q2 = , 0 ( ) = i = q2
2 2 2 2 2
For the function pq one can no longer just replace p by P and q by Q since
the operators P and Q dont commute, and P Q or QP is not self-adjoint. What
does work, satisfying all the conditions to give a Lie algebra homomorphism is
1 i
Opq = (P Q + QP ), 0 (pq) = (P Q + QP )
2 2
142
This shows that the Schr odinger representation 0 that was defined as a
representation of the Heisenberg Lie algebra h3 extends to a unitary Lie algebra
representation of a larger Lie algebra, that of all quadratic polynomials on phase
space. Restricting to just the homogeneous order two polynomials, we get a
representation of sl(2, R), called the metaplectic representation.
As a Heisenberg Lie algebra representation, the Schrodinger representation
0 exponentiates to give a representation of the corresponding Heisenberg Lie
group. As an sl(2, R) representation however, one can show that 0 has the same
sort of problem as the spinor representation of su(2) = so(3), which was not a
representation of SO(3), but only of its double cover SU (2) = Spin(3). To get
a group representation, one must go to a double cover of the group SL(2, R),
which will be called the metaplectic group and denoted M p(2, R). We will
return to this point later, since the calculation necessary to see what is going on
is best performed not in the Schr odinger representation but in a different but
unitarily equivalent one that we will find when studying the harmonic oscillator.
One should keep in mind though that 0 is not a representation of the product
of h3 and sl(2, R), but of something more non-trivial called the semi-direct
product h3 o sl(2, R). In a later chapter we will study the general theory
of such semi-direct products for Lie groups in more detail. Using that general
theory, the group we want is the semi-direct product of H3 and SL(2, R), written
H3 o SL(2, R). To construct that semi-direct product, one uses an action of
SL(2, R) on H3 , and it is this action that is at issue here.
The Schr odinger representation will be an infinite dimensional unitary
representation not of this larger group, but of a double cover of it. We will see
this more clearly later on when we work with a different but unitarily equivalent
representation based on the quantization of the harmonic oscillator. On general
grounds though, one can argue that acting by g Sp(2d, R) on elements X
h2d+1 takes 0 to a different representation 0g defined by
0g (X) = 0 (g X)
0g = U (g)0 U (g)1
These U (g) should then give a unitary representation of the group Sp(2d, R)
on our state space H, and we should have what we expect: a symmetry of the
classical phase space under quantization becomes a unitary representation on
the quantum state space. The problem with this argument is that the U (g) are
not uniquely defined. Schurs lemma tells us that since the representation on H
is irreducible, the operators commuting with the representation operators will
just be the complex scalars, but these complex scalars give a phase ambiguity
in how we define the U (g). So it turns out that the U (g) only have to give
a representation of Sp(2d, R) on H up to a phase. The more detailed explicit
information we have about the representation turns out to show that this phase
ambiguity is really just a sign ambiguity, and this is what is responsible for the
143
necessity of introducing a double cover of Sp(2d, R) called M p(2d, R) if we want
a true representation. For infinite dimensional phase spaces, the ambiguity may
not just be a sign ambiguity, a situation that is described as an anomaly, where
a classical symmetry does not survive intact as a symmetry of the quantum state
space.
i
O{f,g} = [Of , Og ]
~
and
Op = P, Oq = Q
for any Lie subalgebra of the functions on R2 larger than the subalgebra of
polynomials of degree less than or equal to two.
Proof. For a detailed proof, see section 5.4 of [7], section 4.4 of [18], or chapter
16 of [23]. In outline, the proof begins by showing that taking Poisson brack-
ets of polynomials of degree three leads to higher order polynomials, and that
furthermore for degree three and above there will be no finite-dimensional sub-
algebras of polynomials of bounded degree. The assumptions of the theorem
force certain specific operator ordering choices in degree three. These are then
144
used to get a contradiction in degree four, using the fact that the same degree
four polynomial has two different expressions as a Poisson bracket:
1 2 1
q 2 p2 = {q p, p2 q} = {q 3 , p3 }
3 9
145
146
Chapter 14
We will now turn to the physical case of 3-dimensional space, and will be working
odinger representation, so we have H = L2 (R3 ). Using the action of
in the Schr
SO(3) on R3 we get an induced representation on H by
q1
(g)(q) = (g 1 q) = (g 1 q2 )
q3
This representation was studied in detail in chapter 8, where we found that the
Lie algebra representation
q1
d d
0 (X)(q) = (etX )t=0 (q) = (etX q2 )t=0
dt dt
q3
0 (l1 ) = q3 q2
q2 q3
0 (l2 ) = q1 q3
q3 q1
0 (l3 ) = q2 q1
q1 q2
In our general study of the action of the Heisenberg group H2n+1 and the
symplectic group Sp(2n, R) on the state space H, we saw that, at least at the
147
Lie algebra level, this came from just using the operators
Pj = i~ , Qj = qj , j = 1, 2, 3
qj
One can easilly check that these satisfy the Poisson bracket relations
148
Quantization gives us corresponding self-adjoint operators on L2 (R3 )
L1 = i0~ (l1 ) = Q2 P3 Q3 P2 = i~(q2 q3 )
q3 q2
L2 = i0~ (l2 ) = Q3 P1 Q1 P3 = i~(q3 q1 )
q1 q3
L3 = i0~ (l3 ) = Q1 P2 Q2 P1 = i~(q1 q2 )
q2 q1
which satisfy
One can check that these operators Lj on H = L2 (R3 ) are (up to a factor
of ~) precisely the same ones
Lj = i0 (lj )
derived earlier using nothing about quantization and the representation , but
just the infinitesimal action of rotations on functions on R3 (the representation
). The factor of ~ appears in the quantization version since now eigenvalues of
the Lj have a correspondence to the classical angular momentum and so depend
on the relative units use for momentum and length.
It is the operators
0 i
e (lj ) = e ~ Lj
that give the operator corresponding to a rotation by about the j-axis. From
the point of view of quantum mechanics, this story is independent of ~, with
the eigenvalues of observables just the integer weights one finds in the action of
SO(3) on its representations. It is only if one wants to interpret these eigenvalues
in terms of the classical notion of angular monentum that one needs to introduce
~ and then one can characterize the eigenvalues as integer multiples of ~.
Note that the spatial geometry of R3 provides us with two symmetry groups.
There is a group R3 of translations which well call T3 , with Lie algebra t3 ,
and a group SO(3) of rotations, with Lie algebra so(3). Since rotations act
on translations, what we have is not the product of these two groups, but a
semi-direct product. This is the so-called Euclidean group
E(3) = T3 o SO(3)
and we will study the group and its representations in more detail in a later
chapter. This group is six dimensional, and its Lie algebra structure can be seen
from the Poisson brackets
149
14.2 Quantum particle in a central potential
In classical physics, to describe not free particles, but particles experiencing
some sort of force, one just needs to add a potential energy term to the
kinetic energy term in the expression for the energy (the Hamiltonian function).
In one dimension, for potential energies that just depend on position, one has
p2
h= + V (q)
2m
for some function V (q). In the physical case of three dimensions, this will be
1 2
h= (p + p22 + p23 ) + V (q1 , q2 , q3 )
2m 1
Quantizing and using the Schrodinger representation, the Hamiltonian op-
erator for a particle moving in a potential V (q1 , q2 , q3 ) will be
1
H= (P 2 + P22 + P33 ) + V (Q1 , Q2 , Q3 )
2m 1
~2 2 2 2
= ( 2 + 2 + 2 ) + V (q1 , q2 , q3 )
2m q1 q2 q3
2
~
= + V (q1 , q2 , q3 )
2m
We will be interested in so-called central potentials, potential functions that
are functions only of q12 + q22 + q32 , and thus only depend upon r, the radial
distance to the origin. For such V , both terms in the Hamiltonian will be
SO(3) invariant, and eigenfunctions of H will be representations of SO(3).
Using the expressions for the angular momentum operators in spherical coor-
dinates derived in chapter 8, one can show that the Laplacian has the following
expression in spherical coordinates
2 2 1
= 2
+ 2 L2
r r r r
where L2 is the Casimir operator in the representation , which we have shown
has eigenvalues l(l+1) on irreducible representations of dimension 2l+1 (integral
spin l). So, restricted to such an irreducible representation, we have
d2 2 d l(l + 1)
= 2
+
dr r dr r2
To solve the Schr odinger equation, we need to find the eigenfunctions of H.
These will be irreducible representations of SO(3) acting on H, which we have
seen can be explicitly expressed in terms of the spherical harmonic functions
Ylm (, ) in the angular variables. So, to find eigenfunctions of the Hamiltonian
~2
H= + V (r)
2m
150
we need to find functions glE (r) depending on l = 0, 1, 2, . . . and the energy
eigenvalue E satisfying
~2 d2 2 d l(l + 1)
( ( 2+ ) + V (r))glE (r) = EglE (r)
2m dr r dr r2
Given such a glE (r) we will have
and the
(r, , ) = glE (r)Ylm (, )
will be a 2l + 1 dimensional (since m = l, l + 1, . . . , l 1, l) space of energy
eigenfunctions for H of eigenvalue E.
For a general potential function V (r), exact solutions for the eigenvalues E
and corresponding functions glE (r) cannot be found in closed form. One special
case where we can find such solutions is for the three-dimensional harmonic
oscillator, where V (r) = 21 m 2 r2 . These are much more easily found though
using the creation and annihilation operator techniques discussed in the last
chapter.
The other well-known and physically very important case is the case of a 1r
potential, called the Coulomb potential. This describes a light charged particle
moving in the potential due to the electric field of a much heavier charged
particle, a situation that corresponds closely to that of a hydrogen atom. In
this case we have
e2
V =
r
where e is the charge of the electron, so we are looking for solutions to
~2 d2 2 d l(l + 1) e2
( ( 2+ 2
) )glE (r) = EglE (r)
2m dr r dr r r
Since having
d2
(rg) = Erg
dr2
is equivalent to
d2 2
( + )g = Eg
dr2 r
for any function g, glE (r) will satisfy
~2 d2 l(l + 1) e2
( ( 2 2
) )rglE (r) = ErglE (r)
2m dr r r
The solutions to this equation can be found through a rather elaborate pro-
cess. For E 0 there are non-normalizable solutions that describe scattering
phenomena that we wont study here. For E < 0 solutions correspond to an
151
integer n = 1, 2, 3, . . ., with n l + 1. So, for each n we get n solutions, with
l = 0, 1, 2, . . . , n 1, all with the same energy
me4
En =
2~2 n2
The unexpected degeneracy in the energy values leads one to suspect that
there is some extra group action in the problem commuting with the Hamilto-
nian, so energy eigenfunctions will come in irreducible representations of some
larger group than SO(3), one such that when one restricts to the SO(3) sub-
group, the representation is reducible, giving n copies of our SO(3) representa-
tion of spin l. In the next section we will see that this is the case, and there use
representation theory to derive the above formula for En .
We wont go through the process of showing how to explicitly find the func-
152
tions glEn (r) but just quote the result. Setting
~2
a0 =
me2
(this has dimensions of length and is known as the Bohr radius), and defining
gnl (r) = glEn (r) the solutions are of the form
r 2r l 2l+1 2r
gnl (r) e na0 ( ) Ln+l ( )
na0 na0
for
n = 1, 2, . . .
l = 0, 1, . . . , n 1
m = l, l + 1, . . . , l 1, l
1 r
100 = p 3 e a0
a0
1 r r
200 = p (1 )e 2a0
3
8a0 2a0
1 r r
211 = p 3 e 2a0 sin ei
8 a0 a0
1 r 2ar
211 = p e 0 cos
4 2a0 a0
3
1 r r
211 = p 3 e 2a0 sin ei
a
8 a0 0
153
14.3 so(4) symmetry and the Coulomb potential
The Coulomb potential problem is very special in that it has an additional
symmetry, of a non-obvious kind. This symmetry appears even in the classi-
cal problem, where it is responsible for the relatively simple solution one can
find to the essentially identical Kepler problem. This is the problem of finding
the classical trajectories for bodies orbiting around a central object exerting a
gravitational force, which also has a 1r potential. Keplers second law for such
motion comes from conservation of angular momentum, which corresponds to
the Poisson bracket relation
{lj , h} = 0
Here well take the Coulomb version of the Hamiltonian that we need for the
hydrogen atom problem
1 e2
h= |p|2
2m r
One can read the relation {lj , h} = 0 in two ways: it says both that the hamil-
tonian h is invariant under the action of the group (SO(3)) whose infinitesimal
generators are lj , and that the components of the angular momentum (lj ) are
invariant under the action of the group (R of time translations) whose infinites-
imal generator is h.
Keplers first and third laws have a different origin, coming from the existence
of a new conserved quantity for this special choice of Hamiltonian. This quantity
is, like the angular momentum, a vector, often called the Lenz (or sometimes
Runge-Lenz, or even Laplace-Runge-Lenz) vector.
Definition (Lenz vector). The Lenz vector is the vector-valued function on the
phase space R6 given by
1 q
w= (p l) e2 2
m |q|
lw =0
We wont here explicitly calculate the various Poisson brackets involving the
components wa of w, since this is a long and unilluminating calculation, but
will just quote the results, which are
{wj , h} = 0
This says that, like the angular momentum, the vector wa is a conserved
quantity under time evolution of the system, and its components generate
symmetries of the classical system.
154
{wj , lk } = jkl wl
These relations say that the generators of the SO(3) symmetry act on wj
the way you would expect, since wj is a vector.
2h
{wj , wk } = jkl ll ( )
m
This is the most surprising relation, and it has no simple geometrical
explanation (although one can change variables in the problem to try and
give it one). It expresses a highly non-trivial relationship between the two
sets of symmetries generated by the vectors l, w and the Hamiltonian h.
The wj are cubic in the q and p variables, so one would expect that the
Groenewold-van Hove no-go theorem would tell one that there is no consistent
way to quantize this system by finding operators Wj corresponding to the wj
that would satisfy the commutation relations corresponding to these Poisson
brackets. It turns out though that this can be done, although not over the
entire phase-space. One gets around the no-go theorem by doing something
that only works where the Hamiltonian h is positive (well be taking a square
root of h).
The choice of operator Wa that works is
1 Q
W= (P L L P) e2
2m |Q|2
which (by elaborate and unenlightening computation) one can show satisfy
[Wj , H] = 0
2
[Wj , Lk ] = i~jkl Ll ( H)
m
[Wj , Lk ] = i~jkl Wl
as well as
LW =WL=0
Most importantly, one has the following relation between W 2 , H and the Casimir
operator L2
2
W 2 = e4 + H(L2 + ~2 )
m
and it is this which will allow us to find the eigenvalues of H, since we know
those for L2 , and can find those of W 2 by changing variables to identify a second
so(3) Lie algebra.
To do this, first change normalization by defining
r
m
K= W
2E
155
where E is the eigenvalue of the Hamiltonian that we are trying to solve for.
Note that it is at this point that we violate the conditions of the no-go theorem,
since we must have E < 0 to get a K with the right properties, and this restricts
the validity of our calculations to a subset of phase space. For E > 0 one can
proceed in a similar way, but the Lie algebra one gets is different (so(3, 1) instead
of so(4)).
One then has the following relation between operators
2H(K 2 + L2 + ~2 ) = me4
[Lj , Lk ] = i~jkl Ll
[Kj , Lk ] = i~jkl Kl
[Kj , Kk ] = i~jkl Ll
Defining
1 1
M= (L + K), N = (L K)
2 2
one has
[Mj , Mk ] = i~jkl Ml
[Nj , Nk ] = i~jkl Nl
[Mj , Nk ] = 0
which is just two commuting copies of so(3).
Recall from our discussion of rotations in three dimensions that representa-
tions of so(3) = su(2) correspond to representations of Spin(3) = SU (2), the
double cover of SO(3) and the irreducible ones have dimension 2l + 1, with l
half-integral. Only for l integral does one get representations of SO(3), and it
is these that occur in the SO(3) representation on functions on R3 . For four di-
mensions, we found that Spin(4), the double cover of SO(4), is SU (2) SU (2),
and one thus has spin(4) = so(4) = su(2)su(2) = so(3)so(3). This is exactly
the Lie algebra we have found here, so one can think of the Coulomb problem
as having an so(4) symmetry. The representations that will occur will include
the half-integral ones, since for this larger symmetry one of our so(3) factors no
longer corresponds to the action of physical rotations on functions on R3 .
Using the fact that
LK=KL=0
one finds that
M2 = N2
so on our energy eigenstates the values of the Casimirs for the two so(3) factors
must match. The relation between the Hamiltonian and these Casimirs is
156
On irreducible representations, we will have
M 2 = ( + 1)
for some half-integral , so we get the following equation for the energy eigen-
values
me4 me4
E= 2 = 2
2~ (4( + 1) + 1) 2~ (2 + 1)2
Letting n = 2 + 1, for = 0, 21 , 1, . . . we get n = 1, 2, 3, . . . and precisely the
same equation for the eigenvalues described earlier
me4
En =
2~2 n2
The irreducible representations of a product like so(3) so(3) are just tensor
products of irreducibles, and in this case the two factors of the product are
identical due to the equality of the Casimirs M 2 = N 2 . The dimension of the
so(3) so(3) irreducibles is thus (2 + 1)2 = n2 , explaining the multiplicity of
states one finds at energy eigenvalue En .
H = L2 (R3 ) C2
The action of the rotation group is now an SU (2) = Spin(3) action on a tensor
product. It is given by the homomorphism to SO(3) and the rotation group
action of SO(3) on the first factor of the tensor product. On the second factor
it is the spinor representation. Physicists generally write the tensor product Lie
algebra representation of the SU (2) operators on H as
J=L+S
where the first term acts just on L2 (R3 ) by the rotation action, the second on
C2 as the spinor representation, so should perhaps be written as
J=L1+1S
where the + (q) is thought of as the wave-function for the electron to be in the
spin up state (+ 21 eigenvalue of S3 ), (q) the wave-function for the spin
157
down state ( 12 eigenvalue of S3 ). The Hamiltonian operator is just the identity
on the spinor factor, so the only effect of this factor is to double the number of
energy eigenstates at each energy. Electrons are fermions, so anti-symmetry of
multi-particle wave functions implies the Pauli principle that states can only be
occupied by a single particle. As a result, one finds that when adding electrons
to an atom described by the Coulomb potential problem, the first two fill up the
lowest Coulomb energy eigenstate (the 100 or 1S state at n = 1), the next eight
fill up the n = 2 states ( two each for 200 , 211 , 210 , 211 ), etc. This goes a
long ways towards explaining the structure of the periodic table of elements.
When one puts a hydrogen atom in a constant magnetic field B, the Hamil-
tonian then acts non-trivially on the spinor factor of H, through a term
e
BS
mc
This is exactly the sort of Hamiltonian we began our study of quantum mechan-
ics with for a simple two-state system. It causes a shift in energy eigenvalues
proportional to |B| depending on the eigenvalue of S3 (the Zeeman effect), giving
different energies to the states in + (q) and (q).
158
Chapter 15
In this chapter well begin the study of the most important exactly solvable
physical system, the harmonic oscillator. Later chapters will discuss extensions
of the methods developed here to the case of fermionic oscillators, as well as free
quantum field theories, which are harmonic oscillator systems with an infinite
number of degrees of freedom.
159
15.1 The harmonic oscillator with one degree of
freedom
An even simpler case of a particle in a potential than the Coulomb potential of
the last chapter is the case of V (q) quadratic in q. This is also the lowest-order
approximation when one studies motion near a local minimum of an arbitrary
V (q), expanding V (q) in a power series around this point. Well write this as
p2 1
h= + m 2 q 2
2m 2
with coefficients chosen so as to make the angular frequency of periodic motion
of the classical trajectories. These satisfy Hamiltons equations
V p
p = = m 2 q, q =
q m
so
q = 2 q
which will have solutions with periodic motion of angular frequency . These
solutions can be written as
we have
so
1 1
c+ = q(0) + i p(0)
2 2m
The classical phase space trajectories are
1 1 1 1
q(t) = ( q(0) + i p(0))eit + ( q(0) i p(0))eit
2 2m 2 2m
im 1 im 1
p(t) = ( q(0) p(0))eit + ( q(0) + p(0))eit
2 2 2 2
160
Instead of using two real coordinates to describe points in the phase space
(and having to introduce a reality condition when using complex exponentials),
one can instead use a single complex coordinate
1 i
z(t) = (q(t) + p(t))
2 m
Then the equation of motion is a first-order rather than second-order differential
equation
z = iz
with solutions
z(t) = z(0)ei(t)
The classical trajectories are then realized as complex functions of t, and paramet-
rized by the complex number
1 i
z(0) = (q(0) + p(0))
2 m
Since the Hamiltonian is just quadratic in the p and q, we have seen that we
can construct the corresponding quantum operator uniquely using the Schroding-
er representation. For H = L2 (R) we have a Hamiltonian operator
P2 1 ~2 d2 1
H= + m 2 Q2 = + m 2 q 2
2m 2 2m dq 2 2
To find solutions of the Schr odinger equation, as with the free particle, one
proceeds by first solving for eigenvectors of H with eigenvalue E, which means
finding solutions to
~2 d2 1
HE = ( 2
+ m 2 q 2 )E = EE
2m dq 2
Solutions to the Schr
odinger equation will then be linear combinations of the
functions
i
E (q)e ~ Et
Standard but somewhat intricate methods for solving differential equations
like this show that one gets solutions for E = En = (n + 21 )~, n a non-negative
integer, and the normalized solution for a given n (which well denote n ) will
be r
m 1 m m q2
n (q) = ( 2n 2
) 4 Hn ( q)e 2~
~2 (n!) ~
where Hn is a family of polynomials called the Hermite polynomials. The
n provide an orthonormal basis for H (one does not need to consider non-
normalizable wave-functions as in the free particle case), so any initial wave-
function (q, 0) can be written in the form
X
(q, 0) = cn n (q)
n=0
161
with Z +
cn = n (q)(q, 0)dq
(note that the n are real-valued). At later times, the wavefunction will be
i 1
X X
(q, t) = cn n (q)e ~ En t = cn n (q)ei(n+ 2 )t
n=0 n=0
[Q, P ] = i~1
we define
r r r r
m 1 m 1
a= Q+i P, a = Qi P
2~ 2m~ 2~ 2m~
which satisfy the commutation relation
[a, a ] = 1
162
Theorem. The number operator N = a a has eigenvalues n = 0, 1, 2, . . . on
H = L2 (R).
a|0i = 0
a|0i = 0
for |0i (the lowest energy or vacuum state) which will have energy eigenvalue
1 1
2 , then acting by a n-times on |0i to get states with energy eigenvalue n + 2 .
The equation for |0i is thus
1 1 d
a|0i = (Q + iP )0 (q) = (q + )0 (q) = 0
2 2 dq
One can check that solutions to this are all of the form
q2
0 (q) = Ce 2
163
Choosing C to make this of norm 1 gives
1 q2
0 (q) = 1 e 2
a a a 1 d q2
|ni = |0i = 1 n (q )n e 2
n 2 1 4 2 2 n! dq
In the physical interpretation of this quantum system, the state |ni, with
energy ~(n + 21 ) is thought of as a state describing n quanta. The state
|0i is the vacuum state with zero quanta, but still carrying a zero-point
energy of 12 ~. The operators a and a have somewhat similar properties to
the raising and lowering operators we used for SU (2) but their commutator
is different (just the identity operator), leading to simpler behavior. In this
case they are called creation and annihilation operators respectively, due
to the way they change the number of quanta. The relation of such quanta to
physical particles like the photon is that quantization of the electromagnetic field
involves quantization of an infinite collection of oscillators, with the quantum
of an oscillator corresponding physically to a photon with a specific momentum
and polarization. This leads to a well-known problem of how to handle the
infinite vacuum energy corresponding to adding up 21 ~ for each oscillator.
The first few eigenfunctions are plotted below. The lowest energy eigenstate
is a Gaussian centered at q = 0, with a Fourier transform that is also a Gaussian
centered at p = 0. Classically the lowest energy solution is an oscillator at rest at
its equilibrium point (q = p = 0), but for a quantum oscillator one cannot have
such a state with a well-defined position and momentum. Note that the plot
gives the wave functions, which in this case are real and can be negative. The
square of this function is what has an intepretation as the probability density
for measuring a given position.
164
15.3 The Bargmann-Fock representation
Working with the operators a and a and their commutation relation
[a, a ] = 1
makes it clear that there is a simpler way to represent these operators than
the Schr
odinger representation as operators on position space functions that we
have been using, while the Stone-von Neumann theorem assures us that this will
be unitarily equivalent to the Schrodinger representation. This representation
appears in the literature under a large number of different names, depending on
the context, all of which refer to the same representation:
Definition (Bargmann-Fock or oscillator or holomorphic or Segal-Shale-Weil
representation). The Bargmann-Fock (etc.) representation is given by taking
H = C[w]
(this is the vector space of polynomials of a complex variable z), and operators
acting on this space
d
a= , a = w
dw
165
The inner product on H is
Z
1 2
h1 |2 i = 1 (w)2 (w)e|w| dudv
C
where w = u + iv.
One has
d d n
[a, a ]z n = (wwn ) w w = (n + 1 n)wn = wn
dw dw
so this commutator is the identity operator on polynomials
[a, a ] = 1
and
The elements
wn
n!
of H for n = 0, 1, 2, . . . are orthornormal.
The operators a and a are adjoints with respect to the given inner product
on H.
The basis
wn
n!
of H for n = 0, 1, 2, . . . is complete.
Proof. The proofs of the above statements are not difficult, in outline they are
in polar coordinates.
d
To show that w and dw are adjoint operators, use integration by parts.
For completeness, assume hn| = 0 for all n. The expression for the |ni
as Hermite polynomials times a Gaussian implies that
Z
q2
F (q)e 2 (q)dq = 0
166
q2
for all polynomials F (q). Computing the Fourier transform of (q)e 2
gives
(ikq)j q2
Z
q2 X
eikq e 2 (q)dq = e 2 (q)dq = 0
j=0
j!
q2
So (q)e 2 has Fourier transform 0 and must be 0 itself. Alternatively,
one can invoke the spectral theorem for the self-adjoint operator H, which
guarantees that its eigenvectors form a complete and orthonormal set.
The operators a and a+ take one from any energy eigenstate |ni to any
other, so linear combinations in H form an irreducible representation. By
completeness there are no states |i orthogonal to all the |ni. Alterna-
tively, one can argue that if there was such a |i, by the argument that
showed that the eigenvalues of N are natural numbers, the same argu-
ment would imply that |i would be in a chain of states terminating with
a state |0i0 such that a|0i0 = 0, with |0i0 6= |0i. But we saw that there
was only one solution for a|0i = 0. In general, given a representation of
the Heisenberg algebra, one can check for irreducibility by seeing if there
is a unique solution to a|0i = 0.
167
satisfying
[Qj , Pk ] = ijk 1, [Qj , Qk ] = [Pj , Pk ] = 0
where Qj and Pj just act on the jth term of the tensor product in the usual
way. Similarly, we will have
1 1
aj = (Qj + iPj ), aj = (Qj iPj ), j = 1, . . . d
2 2
satisfying
[aj , ak ] = jk 1, [aj , ak ] = [aj , ak ] = 0
These sets of commutation relations are known as the canonical commutation
relations (often abbreviated CCR).
Using the fact that tensor products of function spaces correspond to func-
tions on the product space, in the Schrodinger representation we have
H = L2 (Rd )
H = C[w1 , , wd ]
where one should keep in mind that one can rescale each degree of freedom
separately, allowing different parameters j for the different degrees of freedom.
The energy and number operator eigenstates will be written
|n1 , , nd i
where
aj aj |n1 , , nd i = Nj |n1 , , nd i = nj |n1 , , nd i
Either the Pj , Qk or the aj , ak together with the identity operator will give
a representation of the Heisenberg Lie algebra h2d+1 on H, and by exponentia-
tion a representation of the Heisenberg group H2d+1 . Quadratic combinations
of these operators will give a representation of sp(2d, R), the Lie algebra of
Sp(2d, R). In the next two chapters we will study these and other aspects of
the quantum harmonic oscillator as a unitary representation.
168
Chapter 16
169
those with a well-defined value of position and momentum variables).
The point of this is that, abstractly, to make a real vector space a complex
vector space, we need to know how to multiply vectors by complex numbers. We
already know how to multiply them by real numbers, so just need to specify what
happens when we multiply by the imaginary unit i. The choice of a complex
structure J provides exactly that piece of information.
Given such a pair (V = R2d , J), we can now break up complex linear com-
binations of vectors in V into those on which J acts as i and those on which it
acts as i (since J 2 = 1, its eigenvalues must be i). We have
V C = VJ+ VJ
h2d+1 = M R
170
with Lie bracket determined by the anti-symmetric bilinear form
Taking complex linear combinations of h2d+1 elements to get a new Lie algebra,
the complexification h2d+1 C, we just use the same formula for the Lie bracket,
with u, u0 now complex vectors in M C, and c, c0 complex numbers.
The choice of a complex structure J on M allows one to write elements of
h2d+1 C as
(u = u+ + u , c) h2d+1 C = (M C) C = M+
J MJ C
where
u+ M+
J , u MJ
we have
171
The standard Bargmann-Fock representation 0BF corresponds to a standard
choice of complex structure J on M, one such that on basis elements
Jqj = pj , Jpj = qj
J(cq q + cp p) = cp q cq p
so J in matrix form is
c 0 1 cq cp
J q = =
cp 1 0 cp cq
It is compatible with since
0
0 1 cq T 0 1 0 1 cq
(J(cq q + cp p), J(c0q q + c0p p)) =( ) ( )
1 0 cp 1 0 1 0 c0p
0
0 1 0 1 0 1 cq
= cq cp
1 0 1 0 1 0 c0p
0
0 1 cq
= cq cp
1 0 c0p
=(cq q + cp p, c0q q + c0p p)
Basis elements of M+
J are then
1
zj = (qj + ipj )
2
since one has
1
Jzj = (pj + iqj ) = izj
2
Basis elements of M
J are the complex conjugates
1
z j = (qj ipj )
2
The definition of Poisson brackets extends to functions of complex coordi-
nates with the conjugate variable to zj becoming iz j since
1 1
{zj , z k } = { (qj + ipj ), (qj ipj )} = ijk
2 2
or
{zj , iz k } = jk
These Poisson bracket relations are just the Lie bracket relations on the Lie
algebra h2d+1 C, since as we have seen, a choice of complex structure J has
been made allowing us to write
h2d+1 C = M+
J MJ C
172
and the only non-zero Lie brackets in this Lie algebra are the ones between
elements of M+J and MJ given by the Poisson bracket relation above.
Quantization takes
zj aj , z j aj , 1 1
and these operators can be realized on the state space C[w1 , . . . , wd ]. The
aj , aj are neither self-adjoint nor skew-adjoint, but if we follow our standard
convention of multiplying by i to get the Lie algebra representation, we have
0BF (zj , 0) = iaj = i , 0BF (z j , 0) = iaj = iwj , 0BF (0, 1) = i1
wj
B : HBF HS
In practice, knowing B explicitly is often not needed, since one can use the
representation independent relation
1
aj = (Qj + iPj )
2
173
in the Bargmann-Fock representation, or purely in terms of Qj and Pj which
have a simple expression
Qj = qj , Pj = i
qj
in the Schrodinger representation.
To give an idea of what the Bargmann transform looks like explicitly, well
just give the formula for the d = 1 case here, without proof. If (q) is a state
in HS = L2 (R), then
Z +
1 2 q2
(B)(z) = 1 ez 2 +2qz (q)dq
4
One can check this equation for the case of the lowest energy state in the
Schrodinger representation, where
1 q2
|0i = (q) = 1 e 2
and
Z +
1 2 2
q2 +2qz 1 q2
B(z) = 1 ez 1 e 2 dq
4 4
Z +
1 2
q 2 +2qz
= 1 ez dq
2
Z +
1 2
= 1 e(qz) dq
2
Z +
1 2
= 1 eq dq
2
=1
174
of the state, while the second annihilates |0i, so generates group transformations
that leave the state invariant. It is only the third operator a , that takes |0i to
distinct states, and one could consider the family of states
ea |0i
for C. The transformations ea are not unitary since a is not skew ad-
joint. It is better to fix this by replacing a with the skew-adjoint combination
a a, defining
Definition (Coherent states). The coherent states in H are the states
a
|i = ea |0i
where C.
Since ea a is unitary, the |i will be a family of distinct normalized states
in H, with = 0 corresponding to the lowest energy state |0i. These are, up
to phase transformation, precisely the states one gets by acting on |0i with
arbitrary elements of the Heisenberg group H3 .
Using the Baker-Campbell-Hausdorff formula one has
||2
a
|i = ea |0i = ea ea e 2 |0i
175
One thing coherent states do provide is an alternate complete set of norm
one vectors in H, so any state can be written in terms of them. However, these
states are not orthogonal (they are eigenvectors of a non-self-adjoint operator so
the spectral theorem for self-adjoint operators does not apply). One can easily
compute that
2
|h|i|2 = e||
Digression (Spin coherent states). One can perform a similar construction
replacing the group H3 by the group SU (2), and the state |0i by a highest weight
vector of an irreducible representation (n , V n = Cn+1 ) of spin n2 . Writing | n2 i
for a highest weight vector, we have
n n n
n0 (S3 )| i = , n0 (S+ )| i = 0
2 2 2
and we can create a family of spin coherent states by acting on | n2 i by elements
of SU (2). If we identify states in this family that differ just by a phase, the
states are parametrized by a sphere.
By analogy with the Heisenberg group coherent states, with n0 (S+ ) playing
the role of the annihilation operator a and n0 (S ) playing the role of the creation
operator a , we can define a skew-adjoint transformation
1 i 0 1
e n (S ) ei n0 (S+ )
2 2
and exponentiate to get a family of unitary transformations parametrized by
(, ). Acting on the highest weight state we get a definition of the family of
spin coherent states as
1 i 0
(S ) 21 ei n
0 n
|, i = e 2 e n (S+ )
| i
2
One can show that the SU (2) group element used here corresponds, in terms of
its action on vectors, to a rotation by an angle about the axis (sin , cos , 0),
so one can associate the state |, i to the unit vector along the z-axis, rotated
by this transformation.
176
Chapter 17
In the last chapter we examined those aspects of the harmonic oscillator quan-
tum system and the Bargmann-Fock representation that correspond to quan-
tization of phase space functions of order less than or equal to one, giving a
unitary representation BF of the Heisenberg group H2d+1 . Well now turn
to what happens when one also includes quadratic functions, extending BF
to a representation of a larger group, the semi-direct product of H2d+1 and
M p(2d, R) (a double cover of Sp(2d, R)).
As a representation of M p(2d, R), BF is unitarily equivalent (by the Barg-
mann transform) to S , the Schr odinger representation version studied earlier.
The Bargmann-Fock version makes some aspects of the representation easier to
study; in particular it makes clear the nature of the double-cover that appears.
This construction involves making a specific choice of complex structure J, a
choice which corresponds to a a choice of Hamiltonian operator H. This choice
also picks out a subgroup U (d) Sp(2d, R) of transformations which commute
with J. The Hamiltonian gives the action of the U (1) diagonal subgroup of
U (d), which commutes with the U (d) action, so energy eigenstates are sets of
irreducible representations of U (d).
The representation BF thus carries an action of several different sorts of
operators: linear combinations of the Qj , Pj studied in the last chapter, one
quadratic combination that gives the Hamiltonian H, a set of quadratic com-
binations that give an SU (d) action on the energy eigenstates, and yet other
quadratic combinations that dont commute with the Hamiltonian, and change
the lowest energy state |0i (these are called Bogoliubov transformations).
177
17.1 The metaplectic representation for d = 1
In the last chapter we saw that introducing annihilation and creation operators
corresponded to choosing a complex structure on the dual phase space M =
R2d , allowing one to use complex coordinates zj , z j instead of real coordinates
qj , pj . After quantization, instead of operators Qj , Pj , we could work with
operators aj , aj , and restricting to first order operators these give the Bargmann-
Fock representation 0BF of h2d+1 on H = C[w1 , . . . , wd ], which corresponds to
the quantum harmonic oscillator system in d degrees of freedom.
Recall from our discussion of the Schrodinger representation 0S that we can
extend this quantization to include quadratic combinations of the qj , pj , getting
a unitary representation of the semi-direct product h2d+1 o sp(2d, R). Restrict-
ing attention to the sp(2d, R) factor, we get the metaplectic representation, and
it is this that we will construct explicitly using 0BF in this chapter, beginning
in this section with the case d = 1, where sp(2, R) = sl(2, R).
One can readily compute the Poisson brackets of quadratic combinations of
z and z using the basic relation {z, z} = i and the Leibniz rule, finding the
following for the non-zero cases
{zz, z 2 } = 2iz 2 , {zz, z 2 } = 2iz 2 , {z 2 , z 2 } = 4izz
In the case of the Schrodinger representation, our quadratic combinations of p
and q were real, and we could identify the Lie algebra they generated with the
Lie algebra sl(2, R) of traceless 2 by 2 real matrices with basis
0 1 0 0 1 0
E= , F = , G=
0 0 1 0 0 1
Since we have complexified, our quadratic combinations of z and z are in the
complexification of sl(2, R), the Lie algebra sl(2, C) of traceless 2 by 2 complex
matrices. We can take as a basis of sl(2, C) over the complex numbers
1
Z = E F, X = (G i(E + F ))
2
which satisfy
[Z, X+ ] = 2iX+ , [Z, X ] = 2iX , [X+ , X ] = iZ
and then use as our isomorphism between quadratics in z, z and sl(2, C)
z2 z2
X+ , X , zz Z
2 2
The element
1 0 1
zz = (p2 + q 2 ) Z =
2 1 0
exponentiates to give a SO(2) = U (1) subgroup of SL(2, R) with elements of
the form
cos sin
eZ =
sin cos
178
Note that h = 12 (p2 + q 2 ) = zz is the classical Hamiltonian function for the
harmonic oscillator.
We can now quantize quadratics in z and z using annihilation and creation
operators acting on H = C[w]. There is no operator ordering ambiguity for
d2
z 2 a2 = , z 2 (a )2 = w2
dw2
For the case of zz (which is real), in order to get the sl(2, R) commutation
relations to come out right, we must take the symmetric combination
1 1 d 1
zz (aa + a a) = a a + = w +
2 2 dw 2
(which of course is just the standard harmonic oscillator quantization of the
Hamiltonian).
Multiplying as usual by i to represent real elements of the Lie algebra as
the skew-Hermitian operator of a unitary group representation, one can now
define
Definition (Metaplectic representation of sl(2, R)). The representation 0BF
on H = C[w] given by
i i 1
0BF (X+ ) = a2 , 0BF (X ) = (a )2 , 0BF (Z) = i (a a + aa )
2 2 2
is called the metaplectic representation of sl(2, R).
One can check that we have made the right choice of 0BF (Z) to get a sl(2, R)
representation by computing
i i
[0BF (X+ ), 0BF (X )] =[ a2 , (a )2 ]
2 2
1
= (aa + a a)
2
= i0BF (Z)
Note that we have actually defined the Lie algebra homomorphism on a basis of
sl(2, C), so we have a Lie algebra representation of this larger Lie algebra, but
it is only on the real sl(2, R) subalgebra that it is a unitary representation.
This representation 0BF will be unitarily equivalent to the Schrodinger ver-
sion 0S found earlier when quantizing q 2 , p2 , pq as operators on H = L2 (R).
It is however much easier to work with since it can be studied as the state
space of the quantum harmonic oscillator, with the Lie algebra acting simply
by quadratics in the annihilation and creation operators.
One thing that can easily now be seen is that this representation 0BF does
not integrate to give a representation of the group SL(2, R). If the Lie algebra
representation 0BF comes from a Lie group representation BF of SL(2, R), we
have 0
BF (eZ ) = eBF (Z)
179
where
1 1
0BF (Z) = i(a a + ) = i(N + )
2 2
so
1
(eZ )|ni = ei(n+ 2 ) |ni
Taking = 2, this gives an inconsistency
BF (1)|ni = |ni
which has its origin in the physical phenomenon that the energy of the lowest
energy eigenstate |0i is 21 rather than 0, so not an integer.
This is precisely the same sort of problem we found when studying the
spinor representation of the Lie algebra so(3). Just as in that case, the problem
indicates that we need to consider not the group SL(2, R), but a double cover,
often called the metaplectic group and denoted M p(2, R). The behavior here
is quite a bit more subtle than in the Spin(3) double cover case, where Spin(3)
was just the group SU (2), and topologically the only non-trivial cover of SO(3)
was the Spin(3) one. Here one has 1 (SL(2, Z)) = Z, and each time one goes
around the U (1) subgroup we are looking at one gets a topologically different
non-contractible loop in the group. As a result, SL(2, R) has lots of non-trivial
covering groups, only one of which is of interest here.
Digression. This group M p(2, R) is quite unusual in that it is a finite-dim-
ensional Lie group, but does not have any sort of description as a group of
finite-dimensional matrices. This is related to the fact that its only interesting
irreducible representation is the infinite-dimensional one we are studying. The
lack of any significant irreducible finite-dimensional representations corresponds
to its not having a matrix description, which would give such a representation.
Note that the lack of a matrix description means that this is a case where the
definition we gave of a Lie algebra in terms of the matrix exponential does not
apply. The more general geometric definition of the Lie algebra of a group in
terms of the tangent space at the identity of the group does apply, although to do
this one really needs a construction of the double cover M p(2, R), which is quite
non-trivial. This is not actually a problem for purely Lie algebra calculations,
since the Lie algebras of M p(2, R) and SL(2, R) can be identified.
Another aspect of the metaplectic representation that is relatively easy to
see in the Bargmann-Fock construction is that the state space H = C[w] is not
an irreducible representation, but is the sum of two irreducible representations
H = Heven Hodd
where Heven consists of the even-degree polynomials, Hodd the odd degree poly-
nomials. Since the generators of the Lie algebra representation are degree two
combinations of annihilation and creation operators, they will not change the
parity of a polynomial, taking even to even and odd to odd. The separate irre-
ducibility of these two pieces is due to the fact that (when n = m(2)), one can
get from state |ni to any another |mi by repeated application of the Lie algebra
representation operators.
180
17.2 The metaplectic representation and the choice
of complex structure
Our construction of the metaplectic representation depends on using a choice
of a basis q, p of M = R2 to get an irreducible unitary representation of the
Heisenberg algebra, and the group SL(2, R) acts on M in a way that takes one
such construction to another unitarily equivalent one. Elements of this group
are given by matrices
a b
c d
satisfying ac bd = 1. The Bargmann-Fock construction depends on a choice
of complex structure J to provide the splitting M C = M+ M into i
eigenspaces of J (and thus into annihilation and creation operators). We have
so far used the standard complex structure
0 1
J=
1 0
and one can ask how the SL(2, R) action interacts with this choice. Only a
subgroup of SL(2, R) will respect the splitting provided by J, it will be the
subgroup of matrices commuting with J. Acting with such matrices on M (and
extending by complex linearity to an action on M C), vectors that are J
eigenvectors with eigenvalue i with transform into other J eigenvectors with
eigenvalue i, and similarly for the i eigenspace. This commutativity condition
is explicitly
a b 0 1 0 1 a b
=
c d 1 0 1 0 c d
so
b a c d
=
d c a b
which implies b = c and a = d. The elements of SL(2, R) that we want will
be of the form
a b
b a
with unit determinant, so a2 + b2 = 1. This subgroup is the SO(2) = U (1)
subgroup of matrices of the form
cos sin
sin cos
To see how this works more explicitly, recall that what is happening here
is that SL(2, R) is acting by automorphisms on the Lie algebra h3 . The in-
finitesimal version of this action is given by the Poisson brackets between the
order-two polynomials of sl(2, R) and the order-one or zero polynomials of h3 .
These are
{pq, p} = p, {pq, q} = q
181
p2 q2
{
, q} = p, { , p} = q
2 2
Once we intoduce a complex structure and variables z, z, these Poisson bracket
relations break up into two sorts, the first of which is
This expresses the fact that the Hamiltonian function h = zz is the generator
of the U (1) action on M C that commutes with J, and it acts with weight 1
on the M+ which has basis z, and with weight 1 on M which has basis z.
Quantizing, one has the following operator relations
1 1
[ (aa + a a), a] = a, [ (aa + a a), a ] = a
2 2
which are the infinitesimal versions of
+a a)
+a a)
ei 2 (aa aei 2 (aa = ei a
and
+a a) i 2 (aa +a a)
ei 2 (aa a e = ei a
We see that, on operators, conjugation by the action of the U (1) subgroup of
SL(2, R) does not mix creation and annihilation operators. On states the U (1)
subgroup acts by
1
ei 2 (aa +a a) |ni = ei(n+ 2 ) |ni
and in particular
+a a)
ei 2 (aa |0i = ei 2 |0i
leaving the lowest energy state |0i invariant (up to a phase).
The second sort of Poisson bracket relation one has are these
{z 2 , z} = 2iz, {z 2 , z} = 0, {z 2 , z} = 2iz, {z 2 , z} = 0
i(a2 (a )2 )
will give the action of the metaplectic representation in the other two direc-
tions than the U (1) direction corresponding to zz. Such transformations act
non-trivially on the chosen complex structure J, taking it to a different one
J 0 , and giving a new set of annihilation a0 and creation operators (a )0 that
mixes the original ones. This sort of transformation is known to physicists as
a Bogoliubov transformation. The lowest energy state will be the state |0iJ 0
satisfying
a0 |0iJ 0 = 0
182
(this determines the state up to phase). The different Js (or equivalently, differ-
ent possible |0iJ up to phase) will be parametrized by the space SL(2, R)/U (1).
This is a space well-known to mathematicians as having a two-dimensional hy-
perbolic geometry, with two specific models the Poincare upper half plane model
and the Poincare disk model. Acting with both SL(2, R) and the Heisenberg
group on |0i we get a generalization of our construction of coherent states in
the previous chapter, with these new coherent states parametrized not just by
an element of M, but also by an element of SL(2, R)/U (1).
Whenever one has a product involving both z and z that one would like to
quantize, the non-trivial commutation relation of a and a means that one has
different inequivalent possibilities, depending on the order one chooses for the a
and a . In this case, we chose to quantize zz as 21 (aa + a a) = a a + 21 because
it then satisfied the commutation relations for sl(2, R). We could instead have
chosen to use a a, which is an example of a normal-ordered product.
Definition. Normal-ordered product
Given any product P of a and a , the normal ordered product of P , written
: P : is given by re-ordering the product so that all factors a are on the left, all
factors a on the right, for example
: a2 a a(a )3 := (a )4 a3
The advantage of working with the normal-ordered a a instead of 21 (aa +
a a) is that it acts trivially on |0i and has integer eigenvalues on all the states.
There is then no need to invoke a double-covering. The disadvantage is that one
loses the relation to sl(2, R). One also needs to keep in mind that the definition
of normal-ordering depends upon the choice of J, or equivalently, the choice
of lowest energy state |0iJ . A better notation would be something like : P :J
rather than just : P :.
183
on the state space
H = C[w1 , , wd ]
of polynomials in d complex variables. One can easily generalize the Bargmann-
Fock inner product defined using the Gaussian from the d = 1 case to this one.
One can think of this as the quantization of the dual phase space M = R2d
given by choosing the standard complex structure J on R2d and using it to
decompose
M C = M+ M
where zj , j = 1, d are basis elements in M+ while z j , j = 1, d are
basis elements in M . As in the d = 1 case, one can ask what subgroup of
Sp(2d, R) preserves the decomposition. We wont show this here, but it turns
out to be a subgroup isomorphic to the unitary group U (d). The analog of
the d = 1 parametrization of possible |0iJ by SL(2, R)/U (1) here would be a
parametrization of such states (or, equivalently, of possible choices of J) by the
space Sp(2d, R)/U (d), known as the Siegel upper half-space.
The situation at the Lie algebra level is a bit easier to work out explicitly.
The choice of J gives a decomposition of the complexified Lie algebra sp(2d, C)
into three sub-algebras as follows:
A Lie subalgebra with basis elements zj zk (as usual, the Lie bracket is the
Poisson bracket). There are 12 (d2 + d) distinct such basis elements. This is
a commutative Lie subalgebra, since the Poisson bracket of any two basis
elements is zero.
A Lie subalgebra with basis elements z j z k . Again, it is has dimension
1 2
2 (d + d) and is a commutative Lie subalgebra.
184
lies in this sub-algebra, and its Poisson brackets with the rest of the sub-
algebra are zero since
Quantization takes
zj zk aj ak , z j z k aj ak
zj z k aj ak = ak aj
[A, B] = C
[aj , ak ] = jk
[A, B] = C
185
Proof.
is a Lie algebra representation of u(d) on the space H that the aj , aj act on.
This is exactly the Lie algebra representation 0BF , restricted to the subalgebra
u(d) sp(2d, R) and normal-ordered. It exponentiates to give a representation
of U (d) on H = C[w1 , . . . , wd ], one that commutes with the Hamiltonian. From
the physics point of view, this is useful, as it provides a decomposition of energy
eigenstates into irreducible representations of U (d). From the mathematics
point of view, the quantum harmonic oscillator provides a construction of a
large class of representations of U (d) (the energy eigenstates). With no normal-
ordering, the representation 0BF itself also provides a representation of the Lie
algebra u(d), but this is one that does not exponentiate to U (d), requiring a
double-cover of U (d) to get a true representation.
H = C[w1 , w2 ]
the polynomials in two complex variables w1 , w2 . Recall from our SU (2) dis-
cussion that it was useful to organize these polynomials into finite dimensional
sets of homogeneous polynomials of degree n for n = 0, 1, 2, . . .
H = H0 H1 H2
186
There are four annihilation or creation operators
a1 = , a = w1 , a2 = , a = w2
w1 1 w2 2
acting on H. These are the quantizations of complex phase space coordinates
z1 , z2 , z 1 , z 2 with quantization just the Bargmann-Fock Lie algebra representa-
tion of the Lie algebra these span, with Lie bracket the Poisson bracket, so
Our original dual phase space was M = R4 , with a group Sp(4, R) acting
on it, preserving the Poisson bracket. When picking the coordinates z1 , z2 , we
have made a standard choice of complex structure J, which breaks up
M C = M+ M = C2 C2
z1 z 1 , z2 z 2 , z1 z 2 , z 1 z2
and quantization
i i
0BF (z1 z 1 ) = (a1 a1 + a1 a1 ), 0BF (z2 z 2 ) = (a2 a2 + a2 a2 )
2 2
0BF (z 1 z2 ) = ia1 a2 , 0BF (z1 z 2 ) = ia2 a1
The real elements (the ones that are in sp(4, R)) are a basis for u(2).
Inside this last subalgebra, there is a distinguished element h = z1 z 1 +
z2 z 2 that commutes withe the rest of the subalgebra, which give the quantum
Hamiltonian
1 1 1
H= (a1 a1 + a1 a1 + a2 a2 + a2 a2 ) = N1 + + N2 + = w1 + w2 +1
2 2 2 w1 w2
187
This operator will just multiply a homogeneous polynomial by its degree plus
one, so it acts just by multiplication by n + 1 on Hn . Exponentiating this
operator (multiplied by i) one gets a representation of a U (1) subgroup of the
metaplectic cover M p(4, R). Taking instead the normal-ordered version
: H := a1 a1 + a2 a2 = N1 + N2 = w1 + w2
w1 w2
one gets a representation of the U (1) Sp(4, R) that acts with opposite weight
on M+ and M . In either case, H does not commute with operators coming
from quantization of the first two subalgebras. These change the eigenvalue of
H or : H : by 2 so take
Hn Hn2
in particular taking |0i to 0 or a state in H2 .
Besides h, which gives the u(1) in u(2) = u(1) su(2), a basis for the su(2)
part, with its correspondence to our basis Xj = i 2j in terms of 2 by 2 matrices
is
1 i 1
X1 (z 1 z2 + z 2 z1 ), X2 (z 2 z1 z 1 z2 ), X3 (z 1 z1 z2 z2 )
2 2 2
This relates two different but isomorphic ways of describing su(2): as 2 by 2
matrices with Lie bracket the commutator, or as quadratic polynomials, with
Lie bracket the Poisson bracket.
Quantization will give a representation of su(2) on H
i 1
0BF (X1 ) = (a2 a1 + a1 a2 ), 0BF (X2 ) = (a2 a1 a1 a2 )
2 2
i
0BF (X3 ) = (a1 a1 a2 a2 )
2
Note that another way to get this result is by the theorem of the last section,
computing the quadratic operators as
a1
a1 a2 Xj
a2
188
on homogeneous polynomials are the same. The Bargmann-Fock representation
does extend as a unitary representation to a larger group (H5 o M p(4, R)).
The fact that we have an SU (2) group acting on the state space of the d = 2
harmonic oscillator and commuting with the action of the Hamiltonian H means
that energy eigenstates will can be organized as irreducible representations of
SU (2). In particular, one sees that the space Hn of energy eigenstates of energy
n+1 will be a single irreducible SU (2) representation, the spin n2 representation
of dimension n + 1 (so n + 1 will be the multiplicity of energy eigenstates of that
energy).
L3 =Q1 P2 Q2 P1
i
= ((a1 + a1 )(a2 a2 ) (a2 + a2 )(a1 a1 ))
2
i
= (a1 a2 a1 a2 a2 a1 + a2 a1 )
2
=i(a2 a1 a1 a2 )
189
One could also use the fact that in the matrix representation of the Lie alge-
bra so(3) acting on vectors, infinitesimal rotations are given by antisymmetric
matrices, and use our theorem from earlier in this section to find the quadratic
operator corresponding to a given antisymmetric matrix. For instance, for in-
finitesimal rotations about the 3 axis,
0 1 0
l3 = 1 0 0
0 0 0
190
Chapter 18
The quantum state spaces describing fundamental physical systems that occur
in the real world often are fermionic: they have a property of anti-symmetry
under the exchange of two states. It turns out that the entire quantum me-
chanical formalism we have developed so far has a parallel version in which such
anti-symmetry is built in from the beginning. Besides providing further exam-
ples of the connections between representation theory and quantum mechanics,
this new formalism encompasses in a surprising way some mathematics of great
importance in geometry.
In this chapter well consider in detail the basic example of how this works,
as a simple variation on techniques we used to study the harmonic oscillator,
leaving for later chapters a discussion of the new mathematics and new general
formalism embodied in the example.
191
This is done by noting that the operators
Nj = aj aj
nj = 0, 1, 2, . . .
and that
d d
1X 1
(aj aj + aj aj ) =
X
H= (Nj + )
2 j=1 j=1
2
We will see later in this course that collections of identical particles can be
described using this formalism, when states are symmetric under interchange
of two particle (and are then called bosons). The state |n1 , n2 , , nN i is
interpreted as describing n1 bosons of one momentum, n2 of another momentum,
etc. The lowest energy state |0, 0, , 0i describes the vacuum state with no
particles, but it carries an energy
d
X 1
~j
j=1
2
Because of this ability to describe bosonic particles, well often call the harmonic
oscillator the bosonic oscillator.
192
(which well now denote aB and aB ) by fermionic annihilation and creation
operators called aF and aF , which satisfy a variant of the bosonic commutation
relations
[aF , aF ]+ = 1, [aF , aF ]+ = 0, [aF , aF ]+ = 0
The change from the bosonic case is just the replacement of the commutator
[A, B] AB BA
[A, B]+ AB + BA
NF = aF aF
now satisfies
2
NF2 = aF aF aF aF = aF (1 aF aF )aF = NF aF a2F = NF
2
(using the fact that a2F = aF = 0). So one has
NF2 NF = NF (NF 1) = 0
which implies that the eigenvalues of NF are just 0 and 1. Well denote eigen-
vectors with such eigenvalues by |0i and |1i. The simplest representation of the
operators aF and aF on a complex vector space HF will be on C2 , and choosing
the basis
0 1
|0i = , |1i =
1 0
the operators are represented as
0 0 0 1 1 0
aF = , aF = , NF =
1 0 0 0 0 0
Since
1
H= (a aF + aF aF )
2 F
is just 12 the identity operator, to get a non-trivial quantum system, instead we
make a sign change and set
1
1 1 0
H = (aF aF aF aF ) = NF 1 = 2
2 2 0 21
The energies of the energy eigenstates |0i and |1i will then be 12 since
1 1
H|0i = |0i, H|1i = |1i
2 2
193
Note that the quantum system we have constructed here is nothing but our
old friend the qubit. Taking complex linear combinations of the operators
aF , aF , NF , 1
In this case the state space is the tensor product of N copies of the single
fermionic oscillator state space
HF = (C2 )d = C2 C2 C2
| {z }
d times
0 1
aF j = 3 3 3 1 1
| {z } 0 0
j1 times
are satisfied for j 6= k since in these cases one will get a factor in the tensor
product of
0 0 0 1
[3 , ]+ = 0 or [3 , ] =0
1 0 0 0 +
194
While this sort of tensor product construction is useful for discussing the
physics of multiple qubits, in general it is easier to not work with large tensor
products, and the Clifford algebra formalism we will describe in the next chapter
avoids this.
The number operators will be
NF j = aF j aF j
As an example, for the case d = 3 the pattern of states and their energy
levels for the bosonic and fermionic cases looks like this
In the bosonic case the lowest energy state is at positive energy and there are
an infinite number of states of ever increasing energy. In the fermionic case the
195
lowest energy state is at negative energy, with the pattern of energy eigenvalues
of the finite number of states symmetric about the zero energy level.
No longer setting ~ = = 1 and allowing different for different degrees of
freedom, the Hamiltonian will be
d
1X
H= ~j (aF j aF j aF j aF j )
2 j=1
|n1 , n2 , , nd i
196
Chapter 19
197
any element of this algebra can be written as a sum of elements in normal order,
of the form
cl,m (aB )l am
B
with all annihilation operators aB on the right, for some complex constants cl,m .
As a vector space over C, Weyl(2, C) is infinite-dimensional, with a basis
aB = w, aB =
w
one sees that Weyl(2, C) can be identified with the algebra of polynomial coeffi-
cient differential operators on functions of a complex variable w. As a complex
vector space, the algebra is infinite dimensional, with a basis of elements
m
wl
wm
In our study of quantization and the harmonic oscillator we saw that the
subset of such operators consisting of linear combinations of
2
1, w, , w2 , , w
w w2 w
is closed under commutators, so it forms a Lie algebra of dimension 6. This Lie
algebra includes as subalgebras the Heisenberg Lie algebra (first three elements)
and the Lie algebra of SL(2, R) (last three elements). Note that here we are
allowing complex linear combinations, so we are getting the complexification of
the real six-dimensional Lie algebra that appeared in our study of quantization.
Since the aB and aB are defined in terms of P and Q, one could of course
also define the Weyl algebra as the one generated by 1, P, Q, with the Heisenberg
commutation relations, taking complex linear combinations of all products of
these operators.
198
This algebra is a four dimensional algebra over C, with basis
1, aF , aF , aF aF
since higher powers of the operators vanish, and one can use the anti-commutation
relation betwee aF and aF to normal order and put factors of aF on the right.
We saw in the last chapter that this algebra is isomorphic with the algebra
M (2, C) of 2 by 2 complex matrices, using
1 0 0 0 0 1 1 0
1 , aF , aF , aF aF
0 1 1 0 0 0 0 0
We will see later on that there is also a way of identifying this algebra with
differential operators in fermionic variables, analogous to what happens in
the bosonic (Weyl algebra) case.
Recall that the bosonic annihilation and creation operators were originally
defined in term of the P and Q operators by
1 1
aB = (Q + iP ), aB = (Q iP )
2 2
Looking for the fermionic analogs of the operators Q and P , we use a slightly
different normalization, and set
1 1
aF = (1 + i2 ), aF = (1 i2 )
2 2
so
1
1 = aF + aF , 2 = (aF aF )
i
and the CAR imply that the operators j satisfy the anti-commutation relations
[1 , 1 ]+ = [aF + aF , aF + aF ]+ = 2
[2 , 2 ]+ = [aF aF , aF aF ]+ = 2
1
[1 , 2 ]+ = [aF + aF , aF aF ]+ = 0
i
From this we see that
One could alternatively have defined Cliff(2, C) as the algebra generated
by 1, 1 , 2 , subject to the relations
[j , k ]+ = 2jk
One can just drop mention of 2 , and get an algebra Cliff(1, C), generated
by 1, , with the relation
2 = 1
This is a two-dimensional complex algebra, isomorphic to C C.
199
19.1.3 Multiple degrees of freedom
For a larger number of degrees of freedom, one can generalize the above and
define Weyl and Clifford algebras as follows:
Definition (Complex Weyl algebras). The complex Weyl algebra for d degrees
of freedom is the algebra Weyl(2d, C) generated by the elements 1, aB j , aB j ,
j = 1, . . . , d satisfying the CCR
2
1, wj , , wj wk , , wj
wj wj wk wk
or, alternatively, one has the following more general definition that also works
in the odd-dimensional case
We wont try and prove this here, but one can show that, abstractly as
algebras, the complex Clifford algebras are something well-known. Generalizing
the case d = 1 where we saw that Cliff(2, C) was isomorphic to the algebra of 2
by 2 complex matrices, one has isomorphisms
Cliff(2d, C) M (2d , C)
200
in the even-dimensional case, and in the odd-dimensional case
1, j , j k , j k l , . . . , 1 2 3 n1 n
(1 + 1 )(1 + 2 ) (1 + n )
which will have 2n terms that are exactly those of the basis listed above.
For reasons that will be explained in the next chapter, it turns out that a
more general definition is useful. We write the number of variables as n = r + s,
for r, s non-negative integers, and now vary not just r + s, but also r s, the
so-called signature.
Definition (Real Clifford algebras, arbitrary signature). The real Clifford al-
gebra in n = r + s variables is the algebra Cliff(r, s, R) over the real numbers
generated by 1, j for j = 1, 2, . . . , n satisfying the relations
[j , k ]+ = 2jk 1
201
In other words, as in the complex case different j anti-commute, but only
the first r of them satisfy j2 = 1, with the other s of them satisfying j2 = 1.
Working out some of the low-dimensional examples, one finds:
Cliff(0, 1, R). This has generators 1 and 1 , satisfying
12 = 1
Taking real linear combinations of these two generators, the algebra one
getsis just the algebra C of complex numbers, with 1 playing the role of
i = 1.
Cliff(0, 2, R). This has generators 1, 1 , 2 and a basis
1, 1 , 2 , 1 2
with
12 = 1, 22 = 1, (1 2 )2 = 1 2 1 2 = 12 22 = 1
This four-dimensional algebra over the real numbers can be identified with
the algebra H of quaternions by taking
1 i, 2 j, 1 2 k
Note that one can construct this using the aF , aF for the complex case
Cliff(2, C)
1 = aF + aF , 2 = aF aF
since these are represented as real matrices.
It turns out that Cliff(r, s, R) is always one or two copies of matrices of real,
complex or quaternionic elements, of dimension a power of 2, but this requires
a rather intricate algebraic arguments that we will not enter into here. For the
details of this and the resulting pattern of algebras one gets, see for instance
[31]. One special case where the pattern is relatively simple is when one has
r = s. Then n = 2r is even-dimensional and one finds
Cliff(r, r, R) = M (2r , R)
We will see in the next chapter that just as quadratic elements of the Weyl
algebra give a basis of the Lie algebra of the symplectic group, quadratic ele-
ments of the Clifford algebra give a basis of the Lie algebra of the orthogonal
group.
202
19.3 For further reading
A good source for more details about Clifford algebras and spinors is Chapter
12 of the representation theory textbook [46]. For the details of what what
happens for all Cliff(r, s, R), another good source is Chapter 1 of [31].
203
204
Chapter 20
The definitions given in last chapter of Weyl and Clifford algebras were purely
algebraic, based on a choice of generators. These definitions do though have a
more geometrical formulation, with the definition in terms of generators corre-
sponding to a specific choice of coordinates. For the Weyl algebra, the geometry
involved is known as symplectic geometry, and we have already seen that in the
bosonic case quantization of a phase space R2d depends on the choice of a
non-degenerate anti-symmetric bilinear form which determines the Poisson
brackets and thus the Heisenberg commutation relations. Such a also deter-
mines a group Sp(2d, R), which is the group of linear transformations of R2d
preserving . The Clifford algebra also has a coordinate invariant definition,
based on a more well-known structure on a vector space Rn , that of a non-
degenerate symmetric bilinear form, i.e. an inner product. In this case the
group that preserves the inner product is an orthogonal group. In the symplec-
tic case anti-symmetric forms require an even number of dimensions, but this is
not true for symmetric forms, which also exist in odd dimensions.
205
(u, u0 ) =cq1 c0p1 cp1 c0q1 + cqd c0pd cpd c0qd
0
0 1 ... 0 0 cq1
1 0 ... 0 0 c0p
1
= cq1 cp1 . . . cqd cpd ... .. .. .. ..
. . . .
0 0 ... 0 1 c0qd
0 0 ... 1 0 c0pd
0 1 ... 0 0 0 1 ... 0 0
1 0 ... 0 0 1 0 ... 0 0
g T ... .. .. .. g = .. .. .. ..
. . . .
. . .
0 0 ... 0 1 0 0 ... 0 1
0 0 ... 1 0 0 0 ... 1 0
This choice of is much less arbitrary than it looks. One can show that
given any non-degenerate anti-symmetric bilinear form on R2d a basis can be
found with respect to which it will be the given here (for a proof, see [7]).
This is also true if one complexifies, taking (q, p) C2d and using the same
formula for , which is now a bilinear form on C2d . In the real case the group
that preserves is called Sp(2d, R), in the complex case Sp(2d, C).
To get a fermionic analog of this, it turns out that all we need to do is replace
non-degenerate anti-symmetric bilinear form (, ) with non-degenerate sym-
metric bilinear form < , >. Such a symmetric bilinear form is actually some-
thing much more familiar from geometry than the anti-symmetric case: it is
just a notion of inner product. Two things are different in the symmetric case
The underlying vector space does not have to be even dimensional, one
can take M = Rn for any n, including n odd. To get a detailed analog of
the bosonic case though, we will mostly consider the even case n = 2d.
For a given dimension n, there is not just one possible choice of < , >
up to change of basis, but one possible choice for each pair of integers r, s
such that r + s = n. Given r, s, any choice of < , > can be put in the
206
form
We can thus extend our definition of the orthogonal group as the group of
transformations g preserving an inner product
Definition (Orthogonal group O(r, s, R)). The group O(r, s, R) is the group of
real r + s by r + s matrices g that satisfy
1 0 ... 0 0 1 0 ... 0 0
0 1 . . . 0 0 0 1 ... 0 0
g T ... ... .. .. g = .. .. .. ..
. .
.
. . .
0 0 . . . 1 0 0 0 . . . 1 0
0 0 . . . 0 1 0 0 ... 0 1
| {z } | {z }
r + signs, s - signs r + signs, s - signs
207
i.e. rotations. To see the relation between Clifford algebras and geometry,
consider first the positive definite case Cliff(n, R). To an arbitrary vector
v = (v1 , v2 , . . . , vn ) Rn
v = v 1 1 + v2 2 + + vn n
Using the Clifford algebra relations for the j , given two vectors v, w the
product of their associated Clifford algebra elements satisfies
vw + wv = [v1 1 + v2 2 + + vn n , w1 1 + w2 2 + + wn n ]+
= 2(v1 w1 + v2 w2 + + vn wn )
= 2 < v, w >
where < , > is the symmetric bilinear form on Rn corresponding to the stan-
dard inner product of vectors. Note that taking v = w one has
vw + wv = 2 < v, w >
vw + wv = 2 < v, w >
another common choice. One also sees variants without the factor of 2.
For n dimensional vector spaces over C, we have seen that for any non-
degenerate symmetric bilinear form a basis can be found such that < , > has
the standard form
< z, w >= z1 w1 + z2 w2 + + zn wn
As a result, there is just one complex Clifford algebra in dimension n, the one
we defined as Cliff(n, C).
208
For n dimensional vector spaces over R with a non-degenerate symmetric
bilinear forms of type r, s such that r+s = n, the corresponding Clifford algebras
Cliff(r, s, R) are the ones defined in terms of generators in the last chapter.
In special relativity, space-time is a real 4-dimensional vector space with an
indefinite inner product corresponding to (depending on ones choice of conven-
tion) either the case r = 1, s = 3 or the case s = 1, r = 3. The group of linear
transformations preserving this inner product is called the Lorentz group, and
its orientation preserving component is written as SO(3, 1) or SO(1, 3) depend-
ing on the choice of convention. Later on in the class we will consider what
happens to quantum mechanics in the relativistic case, and there encounter the
corresponding Clifford algebras Cliff(3, 1, R) or Cliff(1, 3, R). The generators
j of such a Clifford algebra are well-known in the subject as the Dirac -
matrices.
For now though, we will restrict attention to the positive definite case, so
just will be considering Cliff(n, R)and seeing how it is used to study the group
O(n) of n-dimensional rotations in Rn .
< v, w >
v0 = v 2 w
< w, w >
209
the reflection in w transformation is
< v, w >
v v 0 =v 2 w
< w, w >
w
=v (vw + wv)
< w, w >
but since
w < w, w >
w = =1
< w, w > < w, w >
we have for w 6= 0
w
w1 =
< w, w >
and the reflection transformation is just conjugation by w times a minus sign
v v 0 = v v wvw1 = wvw1
So, thinking of vectors as lying in the Clifford algebra, the orthogonal trans-
formation that is the result of one reflection is just a conjugation (with a minus
sign). These lie in the group O(n) since they change orientation. The result of
two reflections in different hyperplanes orthogonal to w1 , w2 will be
210
rotation matrix in the plane
vj cos sin vj
vk sin cos vk
In the SO(3) case we saw that there were three of these matrices
providing a basis of the Lie algebra so(3), in n dimensions there will be 21 (n2 n)
of them, providing a basis of the Lie algebra so(n).
Just as in the case of SO(3) where unit length quaternions were used, we can
use elements of the Clifford algebra to get these same rotation transformations,
but as conjugations in the Clifford algebra. To see how this works, consider the
quadratic Clifford algebra element j k and notice that
1 1 1 1
( j k )2 = (j k j k ) = j2 k2 =
2 4 4 4
so one has
(/2)2
e 2 j k =(1 + ) + j k (/2 )
2!
= cos( ) + j k sin( )
2 2
Conjugating a vector vj j +vk k in the j k plane by this, one finds (working
out the details here will be a problem on the problem set)
e 2 j k (vj j + vk k )e 2 j k = (vj cos vk sin )j + (vj sin + vk cos )k
211
the phase space R2n . The fermionic case is precisely analogous, with the role
of the anti-symmetric bilinear form (, ) replaced by the symmetric bilinear
form < , > and the Lie algebra sp(2n, R) replaced by so(n) = spin(n).
In the bosonic case the linear functions of the Qj , Pj satisfied the commuta-
tion relations of another Lie algebra, the Heisenberg algebra, but in the fermionic
case this is not true for the j . In the next chapter we will see that one can
define a notion of a Lie superalgebra that restores the parallelism.
212
Chapter 21
Anti-commuting Variables
and Pseudo-classical
Mechanics
The analogy between the algebras of operators in the bosonic (Weyl algebra) and
fermionic (Clifford algebra) cases can be extended by introducing a fermionic
analog of phase space and the Poisson bracket. This gives a fermionic ana-
log of classical mechanics, sometimes called pseudo-classical mechanics, the
quantization of which gives the Clifford algebra as operators, and spinors as
state spaces. In this chapter well intoduce anti-commuting variables j that
will be the fermionic analogs of the variables qj , pj . These objects will become
generators of the Clifford algebra under quantization, and will later be used in
the construction of fermionic state spaces, by analogy with the Schrodinger and
Bargmann-Fock constructions in the bosonic case.
213
is called the Grassmann algebra, and denoted (Rn ).
Note that these relations imply that generators satisfy j2 = 0. Also note that
sometimes the product in the Grassmann algebra is called the wedge product
and the product of j and k is denoted j k . We will not use a different
symbol for the product in the Grassmann algebra, relying on the notation for
generators to keep straight what is a generator of a conventional polynomial
algebra (e.g. qj or pj ) and what is a generator of a Grassman algebra (e.g. j ).
The Grassmann algebra behaves in many ways like the polynomial algebra
on Rn , but it is finite dimensional, with basis
1, j , j k , j k l , , 1 2 n
F () = c0 + c1
214
and one can take
F = c1
For larger values of n, an arbitrary function can be written as
F (1 , 2 , . . . , n ) = FA + j FB
where FA , FB are functions that do not depend on the chosen j (one gets FB by
using the anti-commutation relations to move j all the way to the left). Then
one can define
F = FB
j
This derivative operator has many of the same properties as the conventional
derivative, although there are unconventional signs one must keep track of. An
unusual property of this derivative that is easy to see is that one has
=0
j j
Taking the derivative of a product one finds this version of the Leibniz rule
for monomials F and G
(F G) = ( F )G + (1)|F | F ( G)
j j j
215
21.2 Pseudo-classical mechanics
The basic structure of Hamiltonian classical mechanics depends on an even
dimensional space phase space R2d with a Poisson bracket {, } on functions on
this space. Time evolution of a function f on phase space is determined by
d
f = {f, h}
dt
for some Hamiltonian function h. This says that taking the derivative of any
function in the direction of the velocity vector of a classical trajectory is the
linear map
f {f, h}
on functions. As we saw in chapter 12, since this linear map is a derivative, the
Poisson bracket will have the derivation property, satisfying the Leibniz rule
for arbitrary functions f1 , f2 , f3 on phase space. Using the Leibniz rule and
anti-symmetry, one can calculate Poisson brackets for any polynomials, just
from knowing the Poisson bracket on generators qj , pj , which we chose to be
where |F1 | and |F2 | are the degrees of F1 and F2 . It will also have the symmetry
property
{F1 , F2 }+ = (1)|F1 ||F2 | {F2 , F1 }+
and one can use these properties to compute the fermionic Poisson bracket for
arbitrary functions in terms of the relation for generators.
216
One can think of the j as the coordinate functions with respect to a
basis ei of Rn , and we have seen that the symmetric bilinear forms on Rn are
classified by a choice of positive signs for some basis vectors, negative signs for
the others. So, on generators j one can choose
{j , k }+ = jk
{j , k l }+ = {j , k }+ l k {j , l }+ = jk l jl k
and
{j k , l m }+ ={j k , l }+ m + l {j k , m }+
= {l , j k }+ m l {m , j k }+
= jl k m + lk j m mj l k + mk l j
The second of these equations shows that the quadratic combinations of the
generators j satisfy the relations of the Lie algebra of the group of rotations in
n dimensions (so(n) = spin(n)). The first shows that the k l acts on the j as
infinitesimal rotations in the k l plane.
Notice that this is again exactly parallel to what happens in the bosonic
case, where the quadratic combinations of the qj , pj satisfy the relations of the
Lie algebra sp(2d, R) and act on the generators qj , pj as infinitesimal Sp(2d, R)
transformations of the phase space R2d .
217
and a super Jacobi identity
where |X| takes value 0 for a usual coordinate, 1 for a fermionic coordinate.
As in the bosonic case, on functions of generators of order less than or equal
to two, the Fermionic Poisson bracket {, , }+ is a Lie superbracket, giving a
Lie superalgebra of dimension 1 + n + 12 (n2 n) (since there is one constant, n
linear terms j and 21 (n2 n) quadratic terms j k ).
We will see in the next chapter that quantization will give a representation
of this Lie superalgebra, with a parallel construction to that of the bosonic case.
{j , k }+ = jk
1, 1 , 2 , 3 , 1 2 , 1 3 , 2 3 , 1 2 3
218
an infinitesimal rotation. So the solution to the classical equations of motion
will be a time-dependent rotation of the j in the plane perpendicular to
B = (B23 , B13 , B12 )
at a constant speed proportional to |B|.
As in the bosonic case, we can make the standard choice of complex structure
J on R2d and get a decomposition
R2d C = Cd Cd
into eigenspaces of J of eigenvalue i. This is done by defining
1 1
j = (2j1 + i2j ), j = (2j1 i2j )
2 2
for j = 1, . . . , d. These satisfy the fermionic Poisson bracket relations
{j , k }+ = {j , k }+ = 0, {j , k }+ = jk
In the bosonic case we could use any complex structure J that satisfied
(Ju, Ju0 ) = (u, u0 ). For the Fermionic case we need instead < Ju, Ju0 >=<
u, u0 >. The subgroup of the SO(2d) transformations that preserve < , > that
also commute with J will be a U (d), just like in the bosonic case where a U (d)
Sp(2d, R) commuted with J, and possible choices of J will be parametrized by
the space SO(2d)/U (d).
In terms of the j , the Hamiltonian is
d
X
h= j j
j=1
219
Using the derivation property of {, }+ one finds
d
X d
X
{j , h}+ = {j , k k }+ = ({j , k }+ k k {j , k }+ ) = j
k=1 k=1
and, similarly,
{j , h}+ = j
so one sees that h is just the generator of U (1) U (d) phase rotations on the
variables j .
220
Chapter 22
In this chapter well begin by investigating the fermionic analog of the notion
of quantization, which takes functions of anti-commuting variables on a phase
space with symmetric bilinear form < , > and gives an algebra of operators
satisfying the relations of the corresponding Clifford algebra. We will then
consider analogs of the constructions used in the bosonic case which there gave
us an essentially unique representation 0 of the corresponding Weyl algebra on
a space of states.
We know that for a fermionic oscillator with d degrees of freedom, the alge-
bra of operators will be Cliff(2d, C), the algebra generated by annihilation and
creation operators aF j , aF j . These operators will act on HF , a complex vec-
tor space of dimension 2d , and this will be our fermionic analog of the bosonic
0 . Since the spin group consists of invertible elements of the Clifford algebra,
it has a representation on HF . This is known as the spinor representation,
and it can be constructed by analogy with the Bargmann-Fock construction in
the bosonic case. Well also consider the analog in the fermionic case of the
Schrodinger representation, which turns out to have a problem with unitarity,
but finds a use in physics as ghost degrees of freedom.
221
Poisson bracket, which on basis elements is
[+ (j ), + (k )]+ = jk 1
+ (j k ) = + (j )+ (k )
222
algebra Cliff(p, q, R), with the quadratic elements providing a representation of
so(p, q, R). The explicit construction of these representations will show that,
as in the bosonic case, the irreducible representations are representations of a
double-cover, in this case the double cover Spin(p, q, R) of SO(p, q, R).
223
Quantizing, we get analogs of the Q, P operators
1 1
= (+ (1 ) + + (2 )),
= (+ (1 ) + (2 ))
2 2
which satisfy anti-commutation relations
2 =
2 = 0,
+
= 1
f () = c1 1 + c2
with cj complex numbers. The inner product on this space given by the
fermionic integral Z
(f1 (), f2 ()) = f1 ()f2 ()d
with
f () = c1 1 + c2
With respect to this inner product, one has
(1, 1) = (, ) = 0, (1, ) = (, 1) = 1
This inner product is indefinite and can take on negative values, since
(1 , 1 ) = 2
224
The corresponding pseudo-classical system will be the classical fermionic
oscillator studied in the last chapter. Recall that this uses a choice of complex
structure J on the fermionic phase space R2d , with the standard choice given
by defining
1 1
j = (2j1 + i2j ), j = (2j1 i2j )
2 2
for j = 1, . . . , d. Here < , > is positive-definite, and the j are coordinates with
respect to an orthonormal basis, so we have the standard relation {j , k }+ = jk
and the j , j satisfy
{j , k }+ = {j , k }+ = 0, {j , k }+ = jk
+ (j ) = aF j , + (j ) = aF j
aF j = , aF j = multiplication by j
j
where f1 and f2 are complex linear combinations of the powers of the anticom-
muting variables j . For the details of the construction of this inner product,
see chapter 7.2 of [45].
225
We saw in chapter 20 that quadratic combinations of the generators j of the
Clifford algebra satisfy the commutation relations of the Lie algebra so(2d) =
spin(2d). The fermionic quantization given here provides an explicit realization
of a representation of the Clifford algebra
Cliff(2d, R) on the complex vector
space HF since we know that the 2+ (j ) will satisfy the Clifford algebra
relations, and we have
1 1
2j1 = (j + j ), 2j = (j j )
2 i 2
As a result, the following linear combinations of annihilation and creation op-
erators
aj + aj , i(aj aj )
will satisfy the same Clifford algebra relations as 2j1 , 2j and taking quadratic
combinations of these operators on HF provides a representation of the Lie
algebra spin(2d).
This representation exponentiates to a representation not of the group SO(2d),
but of its double-cover Spin(2d). The representation that we have constructed
here on the fermionic oscillator state space HF is called the spinor representa-
tion of Spin(2d), and we will sometimes denote HF with this group action as
S.
In the bosonic case, HB is an irreducible representation of the Heisenberg
group, but as a representation of M p(2d, R), it has two irreducible components,
corresponding to even and odd polynomials. The fermionic analog is that HF
is irreducible under the action of the Clifford algebra Cliff(2d, C). To show
this, note that Cliff(2d, C) is isomorphic to the matrix algebra M (2d , C) and its
d
action on HF = C2 is isomorphic to the action of matrices on column vectors.
While HF is irreducible as a representation of the Clifford algebra, it is the
sum of two irreducible representations of Spin(2d), the so-called half-spinor
representations . Spin(2d) is generated by quadratic combinations of the Clif-
ford algebra generators, so these will preserve the subspaces
S+ = span{|0iF , aF j aF k |0iF , } S = HF
and
S = span{aF j |0iF , aF j aF k aF l |0iF , } S = HF
226
in the spinor space, |0iF HF which (up to a scalar factor), is determined by
the condition that the quantized j annihilate the state
+ (j )|0iF = aF j |0iF = 0
which has half-integral eigenvalues, corresponding to the fact that the spinor
representation HF = S will be a true representation only of the double-cover
Spin(2d) of SO(2d). In parallel with the bosonic case, one can work with
normal-ordered products of the annihilation and creation operators, getting a
true representation of U (d) on HF , but when does this one works with a u(1)
that is different than the u(1) that lies in the so(2d) = spin(2d) determined by
the fermionic Poisson bracket on quadratic combinations of the j .
Bosonic Fermionic
227
Sp(2d, R) preserves (, ) SO(2d, R) preserves < , >
Quadratics in Pj , Qj Quadratics in j
Representation of sp(2d, R) Representation of so(2d)
J : J 2 = 1, (Ju, Jv) = (u, v) J : J 2 = 1, < Ju, Jv >=< u, v >
M C = M+ M M C = M+ M
U (d) Sp(2d, R) commutes with J U (d) SO(2d, R) commutes with J
M p(2d, R) double-cover of Sp(2d, R) Spin(2d) double-cover of SO(2d)
Metaplectic representation Spinor representation
aj , aj satisfying CCR aj F , aj F satisfying CAR
aj |0i = 0, |0i depends on J aj F |0i = 0, |0i depends on J
228
Chapter 23
Supersymmetry, Some
Simple Examples
If one considers fermionic and bosonic quantum system that each separately
have operators coming from Lie algebra or superalgebra representations on their
state spaces, when one combines the systems by taking the tensor product, these
operators will continue to act on the combined system. In certain special cases
new operators with remarkable properties will appear that mix the fermionic and
bosonic systems. These are generically known as supersymmetries. In such
supersymmetrical systems, important conventional mathematical objects often
appear in a new light. This has been one of the most fruitful areas of interaction
between mathematics and physics in recent decades, beginning with Edward
Wittens highly influential 1982 paper Supersymmetry and Morse theory [54]. A
huge array of different supersymmetric quantum field theories have been studied
over the years, although the goal of using these to develop a successful unified
theory remains still out of reach. In this chapter well examine in detail some
of the simplest such quantum systems, examples of superymmetric quantum
mechanics. These have the characteristic feature that the Hamiltonian operator
is a square.
N N
1 X 1
(aB j aB j + aB j aB j ) =
X
H= ~ (NB j + )~
2 j=1 j=1
2
229
where NB j is the number operator for the jth degree of freedom, with
eigenvalues nB j = 0, 1, 2, .
N N
1 X 1
(aF j aF j aF j aF j ) =
X
H= ~ (NF j )~
2 j=1 j=1
2
where NF j is the number operator for the jth degree of freedom, with
eigenvalues nF j = 0, 1.
Putting these two systems together we get a new quantum system with state
space
H = HB HF
and Hamiltonian
N
X
H= (NB j + NF j )~
j=1
Notice that the lowest energy state |0i for the combined system has energy 0,
due to cancellation between the bosonic and fermionic degrees of freedom.
For now, taking for simplicity the case N = 1 of one degree of freedom, the
Hamiltonian is
H = (NB + NF )~
with eigenvectors |nB , nF i satisfying
H|nB , nF i = (nB + nF )~
Notice that while there is a unique lowest energy state |0, 0i of zero energy, all
non-zero energy states come in pairs, with two states
|n, 0i and |n 1, 1i
Q+ = aB aF , Q = aB aF
which are not self adjoint, but are each others adjoints ((Q ) = Q+ ).
The pattern of energy eigenstates looks like
230
Computing anti-commutators using the CCR and CAR for the bosonic and
fermionic operators (and the fact that the bosonic operators commute with the
fermionic ones since they act on different factors of the tensor product), one
finds that
Q2+ = Q2 = 0
and
(Q+ + Q )2 = [Q+ , Q ]+ = H
1
Q1 = Q+ + Q , Q2 = (Q+ Q )
i
which satisfy
[Q1 , Q2 ]+ = 0, Q21 = Q22 = H
231
with the same energy. To find states of zero energy, instead of trying to solve
the equation H|0i = 0 for |0i, one can look for solutions to
Q1 |0i = 0 or Q2 |0i = 0
These operators dont correspond to a Lie algebra representation as H does,
but do come from a Lie superalgebra representation, so are described as gen-
erators of a supersymmetry transformation. In more general theories with
operators like this with the same relation to the Hamiltonian, one may or may
not have solutions to
Q1 |0i = 0 or Q2 |0i = 0
If such solutions exist, the lowest energy state has zero energy and is described
as invariant under the supersymmetry. If no such solutions exist, the lowest
energy state will have a non-zero, positive energy, and satisfy
Q1 |0i =
6 0 or Q2 |0i =
6 0
In this case one says that the supersymmetry is spontaneously broken, since
the lowest energy state is not invariant under supersymmetry.
There is an example of a physical quantum mechanical system that has ex-
actly the behavior of this supersymmetric oscillator. A charged particle confined
to a plane, coupled to a magnetic field perpendicular to the plane, can be de-
scribed by a Hamiltonian that can be put in the bosonic oscillator form (to
show this, we need to know how to couple quantum systems to electromagnetic
fields, which we will come to later in the course). The equally spaced energy
levels are known as Landau levels. If the particle has spin one-half, there will
be an additional term in the Hamiltonian coupling the spin and the magnetic
field, exactly the one we have seen in our study of the two-state system. This
additional term is precisely the Hamiltonian of a fermionic oscillator. For the
case of gyromagnetic ratio g = 2, the coefficients match up so that we have ex-
actly the supersymmetric oscillator described above , with exactly the pattern
of energy levels seen there.
232
We will change this definition to introduce an arbitrary superpotential W (q):
1 1
aB = (W 0 (Q) + iP ), aB = (W 0 (Q) iP )
2 2
but keep our definition of the operators
Q+ = aB aF , Q = aB aF
These satisfy
Q2+ = Q2 = 0
for the same reason as in the oscillator case: repeated factors of aF or aF vanish.
Taking as the Hamiltonian the same square as before, we find
H =(Q+ + Q )2
1 1
= (W 0 (Q) + iP )(W 0 (Q) iP )aF aF + (W 0 (Q) iP )(W 0 (Q) + iP )aF aF
2 2
1 1
= (W (Q) + P )(aF aF + aF aF ) + (i[P, W 0 (Q)])(aF aF aF aF )
0 2 2
2 2
1 0 2 2 1 0
= (W (Q) + P ) + (i[P, W (Q)])3
2 2
But iP is the operator corresponding to infinitesimal translations in Q, so we
have
i[P, W 0 (Q)] = W 00 (Q)
and
1 1 00
H= (W 0 (Q)2 + P 2 ) + W (Q)3
2 2
which gives a large class of quantum systems, all with state space
H = HB HF = L2 (R) C2
.
There may or may not be a state with zero energy, depending on whether
or not one can find a solution to the equation
If such a solution does exist, thinking in terms of super Lie algebras, one calls
Q1 the generator of the action of a supersymmetry on the state space, and
describes the ground state |0i as invariant under supersymmetry. If no such
233
solution exists, one has a theory with a Hamiltonian that is invariant under su-
persymmetry, but with a ground state that isnt. In this situation one describes
the supersymmetry as spontaneously broken. The question of whether a given
supersymmetric theory has its supersymmetry spontaneously broken or not is
one that has become of great interest in the case of much more sophisticated su-
persymmetric quantum field theories. There, hopes (so far unrealized) of making
contact with the real world rely on finding theories where the supersymmetry
is spontaneously broken.
In this simple quantum mechanical system, one can try and explicitly solve
the equation Q1 |i = 0. States can be written as two-component complex
functions
+ (q)
|i =
(q)
and the equation to be solved is
1 + (q)
(Q+ + Q )|i = ((W 0 (Q) + iP )aF + (W 0 (Q) iP )aF )
2 (q)
1 d 0 1 d 0 0 + (q)
= ((W 0 (Q) + ) + (W 0 (Q) ) )
2 dq 0 0 dq 1 0 (q)
1 0 1 d 0 1 + (q)
= (W 0 (Q) + )
2 1 0 dq 1 0 (q)
1 0 1 d + (q)
= ( W 0 (Q)3 ) =0
2 1 0 dq (q)
which has general solution
c+ eW (q)
+ (q) c
= eW (q)3 + =
(q) c c eW (q
for complex constants c+ , c . Such solutions will only be normalizable if
c+ = 0, lim W (q) = +
q
or
c = 0, lim W (q) =
q
If, for example, W (q) is an odd polynomial, one will not be able to satisfy either
of these conditions, so there will be no solution, and the supersymmetry will be
spontaneously broken.
234
the tensor product of complex-valued functions on Rd (acted on by the Weyl
algebra Weyl(2d, C)) and anti-commuting functions on Rd (acted on by the
Clifford algebra Cliff(2d, C)). There are two operators Q+ and Q , adjoints
of each other and of square zero. If one has studied differential forms, this
should look familiar. This space H is well-known to mathematicians, as the
complex valued differential forms on Rd , often written (Rd ), where here the
= (d + )2
235
algebra in 2d variables and the Clifford algebra in d variables. The state space
will be
H = L2 (Rd ) S
where S is the spinor space that the Clifford algebra in d variables acts on
irreducibly.
If our Hamiltonian operator is the Laplacian on Rd , we have
d
X
H= Pj2
j=1
and one can make this act not just on functions L2 (Rd ) but on H, the spinor-
valued functions on Rd . Doing this, we find again that there is a supersymmetry
Q, just one this time (so called N = 1, or alternately N = 12 supersymmetry),
and H is a square. The operator Q is given by
d
X
j P j
j=1
where j are the generators of the Clifford algebra. Using the defining relations
for the j , one can easily see that
Q2 = H
This operator Q is called the Dirac operator, and it plays a critical role in
physics, which we will study later in the course when we study relativistic quan-
tum systems. It also plays an important role in mathematics, since like the de
Rham differential it can be constructed for not just Rn , but arbitrary manifolds
with metrics, where it can be used to study their geometry and topology.
236
Chapter 24
In this chapter well give a rapid survey of a different starting point for devel-
oping quantum mechanics, based on the Lagrangian rather than Hamiltonian
classical formalism. Lagrangian methods have quite different strengths and
weaknesses than those of the Hamiltonian formalism, and well try and point
these out, while referring to standard physics texts for more detail about these
methods.
The Lagrangian formalism leads naturally to an apparently very different
notion of quantization, one based upon formulating quantum theory in terms of
infinite-dimensional integrals known as path integrals. A serious investigation
of these would require another and very different volume, so again well have to
restrict ourselves to outlining how path integrals work, describing their strengths
and weaknesses, and giving references to standard texts for the details.
Definition. Lagrangian
The Lagrangian L is a function
L : (q, v) Rd Rd L(q, v) R
: t [t1 , t2 ] Rd
237
which we will write in terms of their position and velocity vectors as
(t) = (q(t), q(t))
Definition. Action
The action S for a path is
Z t2
S[] =
L(q(t), q(t))dt
t1
d
S() S(s )|s=0 = 0
ds
when
s : [t1 , t2 ] Rd
is a smooth family of paths parametrized by an interval s (, ), with 0 = .
Well now ignore analytical details and adopt the physicists interpretation
of this as the first-order change in S due to an infinitesimal change =
(q(t), q(t)) in the path.
When (q(t), q(t)) satisfy a certain differential equation, the path will be a
critical point and thus a classical trajectory:
L d L
(q(t), q(t)) (
(q(t), q(t))) =0
qj dt qj
Proof. Ignoring analytical details, the Euler-Lagrange equations follow from the
following calculations, which well just do for d = 1, with the generalization to
238
higher d straightfoward. We are calculating the first-order change in S due to
an infinitesimal change = (q(t), q(t))
Z t2
S[] = L(q(t), q(t))dt
t1
Z t2
L L
= ( (q(t), q(t))q(t)
+ (q(t), q(t))
q(t))dt
t1 q q
But
d
q(t)
= q(t)
dt
and, using integration by parts
L d L d L
q(t)
= ( q) ( )q
q dt q dt q
so
Z t2
L d L d L
S[] = (( )q ( q))dt
t1 q dt q dt q
Z t2
L d L L L
= ( )qdt ( q)(t2 ) + ( q)(t1 )
t1 q dt q q q
This is just Newtons second law which says that the force coming from the
potential is equal to the mass times the acceleration of the particle.
The derivation of the Euler-Lagrange equations can also be used to study
the implications of Lie group symmetries of a Lagrangian system. When a Lie
group G acts on the space of paths, preserving the action S, it will take classical
trajectories to classical trajectories, so we have a Lie group action on the space
of solutions to the equations of motion (the Euler-Lagrange equations). In
239
good cases, this space of solutions is just the phase space of the Hamiltonian
formalism. On this space of solutions, we have, from the calculation above
L L
S[] = ( q(X))(t1 ) ( q(X))(t2 )
q q
where now q(X) is the infinitesimal change in a classical trajectory coming
from the infinitesimal group action by an element X in the Lie algebra of G.
From invariance of the action S under G we must have S=0, so
L L
( q(X))(t2 ) = ( q(X))(t1 )
q q
This is an example of a more general result known as Noethers theorem.
It says that given a Lie group action on a Lagrangian system that leaves the
action invariant, for each element X of the Lie algebra we will have a conserved
quantity
L
q(X)
q
which is independent of time along the trajectory. A basic example is when
the Lagrangian is independent of the position variables qj , depending only on
the velocities qj , for example in the case of a free particle, when V (q) = 0. In
such a case one has invariance under the Lie group Rd of space-translations.
An infinitesimal transformation in the j-direction is qiven by
qj (t) =
(q(t), q(t)) R2d
240
where pj = Lqj and this R
2d
is the Hamiltonian phase space with the conven-
tional Poisson bracket.
This transformation between position-velocity and phase space is known as
the Legendre transform, and in good cases (for instance when L is quadratic
in all the velocities) it is an isomorphism. In general though, this is not an
isomorphism, with the Legendre transform often taking position-velocity space
to a lower-dimensional subspace of phase space. Such cases require a much more
elaborate version of Hamiltonian formalism, known as constrained Hamiltonian
dynamics and are not unusual: one example we will see later is that of the
equations of motion of a free electromagnetic field (Maxwells equations).
Besides a phase space, for a Hamiltonian system one needs a Hamiltonian
function. Choosing
Xd
h= pj qj L(q, q)
j=1
to solve for the velocities qj and express them in terms of the momentum vari-
ables. In that case, computing the differential of h one finds (for d = 1, the
generalization to higher d is straightforward)
L L
dh =pdq + qdp dq dq
q q
L
=qdp dq
q
So one has
h h L
= q,
=
p q q
but these are precisely Hamiltons equations since the Euler-Lagrange equations
imply
L d L
= = p
q dt q
The Lagrangian formalism has the advantage of depending concisely just on
the choice of action functional, which does not distinguish time in the same
way that the Hamiltonian formalism does by its dependence on a choice of
Hamiltonian function h. This makes the Lagrangian formalism quite useful in
the case of relativistic quantum field theories, where one would like to exploit the
full set of space-time symmetries, which can mix space and time directions. On
the other hand, one loses the infinite dimensional group of symmetries of phase
space for which the Poisson bracket is the Lie bracket (the so-call canonical
transformations). In the Hamiltonian formalism we saw that the harmonic
oscillator could be best understood using such symmetries, in particular the
241
U (1) symmetry generated by the Hamiltonian function. The harmonic oscillator
is a more difficult problem in the Lagrangian formalism, where this symmetry
is not manifest.
If K(P ) can be chosen to depend only on the momentum operator P and V (Q)
depends only on the operator Q then one can insert alternate copies of the
identity operator in the forms
Z Z
|qihq|dq = 1, |pihp|dp = 1
where the index j goes from 1 to N and the pj , qj variable will be integrated
over. Such a term can be evaluated as
i i
hqj |pj ihpj |qj1 ie N ~ K(pj )T e N ~ V (qj1 )T
242
1 i 1 i i i
= e ~ q j pj e ~ qj1 jpj e N ~ K(pj )T e N ~ V (qj1 )T
2~ 2~
1 i pj (qj qj1 ) i (K(pj )+V (qj1 ))T
= e~ e N~
2~
1 N
The N factors of this kind give an overall factor of ( 2~ ) times something
which is a discretized approximation to
i
RT
(pqh(q(t),p(t)))dt
e~ 0
where the phase in the exponential is just the action. Taking into account the
integrations over qj and pj one has
N Z
1 NY
Z RT
i i
hqT |e ~ HT |q0 i = lim ( ) dpj dqj e ~ 0 (pqh(q(t),p(t)))dt
N 2~ j=1
and one can interpret the integration as an integral over the space of paths in
phase space, thus a phase space path integral.
This is an extremely simple and seductive expression, apparently saying that,
once the action S is specified, a quantum system is defined just by considering
integrals Z
i
D e ~ S[]
over paths in phase space, where D is some sort of measure on this space of
paths. Since the integration just involves factors of dpdq and the exponential
just pdq and h, this formalism seems to share the same sort of invariance under
the infinite-dimensional group of canonical transformations (transformations of
the phase space preserving the Poisson bracket) as the classical Hamiltonian
formalism. It also appears to solve our problem with operator ordering am-
biguities, since introducing products of P s and Qs at various times will just
give a phase space path integral with the corresponding p and q factors in the
integrand, but these commute.
Unfortunately, we know from the Groenewold-van Hove theorem that this is
too good to be true. This expression cannot give a unitary representation of the
full group of canonical transformations, at least not one that is irreducible and
restricts to what we want on transformations generated by linear functions q
and p. Another way to see the problem is that a simple argument shows that by
canonical transformations one can transform any Hamiltonian into free-particle
Hamiltonian, so all quantum systems would just be free particles in some choice
of variables. For the details of these arguments and a careful examination of
what goes wrong, see chapter 31 of [39]. One aspect of the problem is that, as a
measure on the discrete sets of points qj , pj , points in phase space for successive
values of j are not likely to be close together, so thinking of the integral as an
integral over paths is not justified.
When the Hamiltonian h is quadratic in the momentum p, the pj integrals
will be Gaussian integrals that can be performed exactly. Equivalently, the
243
kinetic energy part K of the Hamiltonian operator will have a kernel in position
space that can be computed exactly. Using one of these, the pj integrals can be
eliminated, leaving just integrals over the qj that one might hope to interpret as
a path integral over paths not in phase space, but in position space. One finds,
P2
if K = 2m
i
hqT |e ~ HT |q0 i =
N Z
m Y m(qj qj1 )2
r
i2~T N i
PN T
lim ( ) 2 dqj e ~ j=1 ( 2T /N V (qj ) N )
N N m i2~T j=1
where now the paths (t) are paths in the position space.
An especially attractive aspect of this expression is that it provides a simple
understanding of how classical behavior emerges in the classical limit as ~ 0.
The stationary phase approximation method for oscillatory integrals says that,
for a function f with a single critical point at x = xc (i.e. f 0 (xc ) = 0) and for a
small parameter , one has
Z +
1 1
dx eif / = p eif (xc )/ (1 + O())
i2 f 00 (c)
Using the same principle for the infinite-dimensional path integral, with f = S
the action functional on paths, and = ~, one finds that for ~ 0 the path
integral will simplify to something that just depends on the classical trajectory,
since by the principle of least action, this is the critical point of S.
Such position-space path integrals do not have the problems of principle
of phase space path integrals coming from the Groenewold-van Hove theorem,
but they still have serious analytical problems since they involve an attempt to
integrate a wildly oscillating phase over an infinite-dimensional space. One does
not naturally get a unitary result for the time evolution operator, and it is not
clear that whatever results one gets will be independent of the details of how
one takes the limit to define the infinite-dimensional integral.
Such path integrals though are closely related to integrals that are known
to make sense, ones that occur in the theory of random walks. There, a well-
defined measure on paths does exist, Wiener measure. In some sense Wiener
measure is what one gets in the case of the path integral for a free particle, but
taking the time variable t to be complex and analytically continuing
t it
244
So, one can use Wiener measure techniques to define the path integral, getting
results that need to be analytically continued back to the physical time variable.
In summary, the path integral method has the following advantages:
Study of the classical limit and semi-classical effects (quantum effects
at small ~) is straightforward.
Calculations for free particles and for series expansions about the free
particle limit can be done just using Gaussian integrals, and these are rela-
tively easy to evaluate and make sense of, despite the infinite-dimensionality
of the space of paths.
245
246
Chapter 25
Non-relativistic Quantum
Fields and Field
Quantization
The quantum mechanical systems we have studied so far describe a finite num-
ber of degrees of freedom, which may be of a bosonic or fermionic nature. In
particular we have seen how to describe a quantized free particle moving in
three-dimensional space. By use of the notion of tensor product, we can then
describe any particular fixed number of such particles. We would however, like
a formalism capable of conveniently describing an arbitrary number of parti-
cles. From very early on in the history of quantum mechanics, it was clear
that at least certain kinds of particles, photons, were most naturally described
not one by one, but by thinking of them as quantized excitations of a classical
system with an infinite number of degrees of freedom: the electromagnetic field.
In our modern understanding of fundamental physics not just photons, but all
elementary particles are best described in this way.
Conventional textbooks on quantum field theory often begin with relativis-
tic systems, but well start instead with the non-relativistic case. Well study a
simple quantum field theory that extends the conventional single-particle quan-
tum systems we have dealt with so far to deal with multi-particle systems. This
version of quantum field theory is what gets used in condensed matter physics,
and is in many ways simpler than the relativistic case, which well take up in a
later chapter.
Quantum field theory is a large and complicated subject, suitable for a full-
year course at an advanced level. Well be giving only a very basic introduc-
tion, mostly just considering free fields, which correspond to systems of non-
interacting particles. Most of the complexity of the subject only appears when
one tries to construct quantum field theories of interacting particles. A remark-
able aspect of the theory of free quantum fields is that in many ways it is little
more than something we have already discussed in great detail, the quantum
247
harmonic oscillator problem. However, the classical harmonic oscillator phase
space that is getting quantized in this case is an infinite dimensional one, the
space of solutions to the free particle Schrodinger equation. To describe multiple
non-interacting fermions, we just need to use fermionic oscillators.
For simplicity well set ~ = 1 and start with the case of a single spatial
dimension. Well also begin using x to denote a spatial variable instead of the q
conventional when this is the coordinate variable in a finite-dimensional phase
space.
where the superscript S means we take elements of the tensor product invariant
under the action of the group SN by permutation of the factors. We want to
consider state spaces containing an arbitrary number of particles, so we define
Definition (Bosonic Fock space, the symmetric algebra). Given a complex vec-
tor space V , the symmetric Fock space is defined as
F S (V ) = C V (V V )S (V V V )S
S N (V ) = (V V )S
| {z }
N times
248
The part of the state space with N quanta has dimension
N +d1
N
which grows with N . This is just the binomial coefficient giving the number
of d-variable monomials of degree N The quanta of a harmonic oscillator are
indistinguishable, which corresponds to the fact that the space of states with
N quanta can be identified with the symmetric part of the tensor product of N
copies of Cd .
More precisely, what one has is
We wont try and give a detailed proof of this here, but one can exhibit the
isomorphism explicitly on generators. If zj V are the coordinate functions
with respect to a basis ej of V (i.e. zj (ek ) = jk ), they give a basis of V which
is also a basis of the linear polynomial functions on V . For higher degrees, one
makes the identification
zj zj S N (V ) zjN
| {z }
N times
As polynomial functions on H1 .
249
25.1.2 Fermions and the fermionic oscillator
For the case of fermionic particles, theres an analogous Fock space construction
using tensor products:
F A (V ) = C V (V V )A (V V V )A
where the superscript A means the subspace of the tensor product that just
changes sign under interchange of two factors. This is known to mathemati-
cians as the exterior algebra (V ), with
N (V ) = (V V )A
| {z }
N times
For each of these descriptions, one can define d annihilation or creation oper-
ators aj , aj satisfying the canonical anti-commutation relations, and these will
generate an algebra of operators on the Fock space.
250
25.2.1 Box normalization
Recall that for a free particle in one dimension the state space H consists of
complex-valued functions on R, with observables the self-adjoint operators for
momentum
d
P = i
dx
and energy (the Hamiltonian)
P2 1 d2
H= =
2m 2m dx2
Eigenfunctions for both P and H are the functions of the form
p (x) eipx
2
p
for p R, with eigenvalues p for P and 2m for H.
Note that these eigenfunctions are not normalizable, and thus not in the
conventional choice of state space as L2 (R). One way to deal with this issue is
to do what physicists sometimes refer to as putting the system in a box, by
imposing periodic boundary conditions
(x + L) = (x)
This box normalization is one form of what physicists call an infrared cutoff,
a way of removing degrees of freedom that correspond to arbitrarily large sizes,
in order to make a problem well-defined. To get a well-defined problem one
starts with a fixed value of L, then one studies the limit L .
251
The number of degrees of freedom is now countable, but still infinite. In order
to get a completely well-defined problem, one typically needs to first make the
number of degrees of freedom finite. This can be done with an additional cutoff,
an ultraviolet cutoff, which means restricting attention to |p| for some
finite , or equivalently |l| < L
2 . This makes the state space finite dimensional
and one then studies the limit.
For finite L and our single-particle state space H1 is finite dimensional,
with orthonormal basis elements l (x). An arbitrary solution to the Schrodinger
equation is then given by
+ L
2
X 2l 4 2 l2
(x, t) = l ei L x ei 2mL2 t
l= L
2
where the subscript j indexes the possible values of the momentum p (which are
discretized in units of 2
L , and in the interval [, ]). The occupation number
npj is the number of particles in the state with momentum pj . In the bosonic
case it takes values 0, 1, 2, , , in the fermionic case it takes values 0 or 1.
The state with all occupation numbers equal to zero is denoted
| , 0, 0, 0, i = |0i
[apj , apk ] = jk
252
The number operator
X
N
b= apk apk
k
The Hamiltonian
X p2
k
H
b = a ap
2m pk k
k
b , np , np , np , i = (
X p2k
H| j1 j j+1
n k )| , npj1 , npj , npj+1 , i
2m
k
With ultraviolet and infrared cutoffs in place, the possible values of pj are
finite in number, H1 is finite dimensional and this is nothing but the standard
quantized harmonic oscillator (with a Hamiltonian that has different frequencies
p2j
(pj ) =
2m
for different values of j). In the limit as one or both cutoffs are removed, H1
becomes infinite dimensional and the Stone-von Neumann theorem no longer
applies. This implies that not all choices of state space are unitarily equivalent.
In particular, state spaces with different choices of vacuum state |0i can be
unitarily inequivalent, with not just the dynamics of states in the state space
dependent on the Hamiltonian, but the state space itself depending on the
Hamiltonian (through the characterization of |0i as lowest energy state). Even
for the free particle, we have here defined the Hamiltonian as the normal-ordered
version, which for finite dimensional H1 differs from the non-normal-ordered one
just by a constant, but as cut-offs are removed this constant becomes infinite,
requiring careful treatment of the limit.
253
25.2.2 Continuum normalization
A significant problem introduced by using cutoffs such as the box normalization
is that these ruin some of the space-time symmetries of the system. The one-
particle space with an infrared cutoff is a space of functions on a discrete set of
points, and this set of points will not have the same symmetries as the usual
continuous momentum space (for instance in three dimensions it will not carry
an action of the rotation group SO(3)). In our study of quantum field theory
we would like to exploit the action of space-time symmetry groups on the state
space of the theory, so need a formalism that preserves such symmetries.
In our earlier discussion of the free particle, we saw that physicists often
work with a continuum normalization such that
1
|pi = p (x) = eipx , hp0 |pi = (p p0 )
2
where formulas such as the second one need to be interpreted in terms of dis-
tributions. In quantum field theory we want to be able to think of each value
of p as corresponding to a classical degree of freedom that gets quantized, and
this continuum normalization will then correspond to an uncountable number
of degrees of freedom, requiring great care when working in such a formalism.
This will however allow us to readily see the action of space-time symmetries
on the states of the quantum field theory and to exploit the duality between
position and momentum space embodied in the Fourier transform.
In the continuum normalization, an arbitrary solution to the free-particle
Schrodinger equation is given by
Z
1 p2
(x, t) = (p)eipx ei 2m t dp
2
254
but Z +
f (p)a(p)dp
Definition (Quantum field operator). The quantum field operators for the free
particle system are
X X 1
(x)
b = p (x)ap = eipx ap
p p L
Note that these are not self-adjoint operators, and thus not themselves ob-
servables. To get some idea of their behavior, one can calculate what they do
to the vacuum state. One has
(x)|0i
b =0
255
1 X ipx
b (x)|0i = e | , 0, np = 1, 0, i
L p
While this sum makes sense as long as it is finite, when cutoffs are removed it is
clear that (x) will have a rather singular limit as an infinite sum of operators.
It can be interpreted as the operator that creates a particle localized precisely
at x.
The field operators allow one to recover conventional wave-functions, for
single and multiple-particle states. One sees by orthonormality of the occupation
number basis states that
1
h , 0, np = 1, 0, |b (x)|0i = eipx = p (x)
L
the complex conjugate wave function of the single-particle state of momentum
p. An arbitrary one particle state |1 i with wave function (x) is a linear
combination of such states, and taking complex conjugates one finds
h0|(x)|
b 1 i = (x)
Similarly, for a two-particle state of identical particles with momenta pj1 and
pj2 one finds
h0|(x b 2 )| , 0, np = 1, 0, , 0, np = 1, 0, i = p ,p (x1 , x2 )
b 1 )(x
j1 j2 j1 j2
where
pj1 ,pj2 (x1 , x2 )
is the wavefunction (symmetric under interchange of x1 and x2 for bosons) for
this two particle state. For a general two-particle state |2 i with wavefunction
(x1 , x2 ) one has
h0|(x b 2 )|2 i = (x1 , x2 )
b 1 )(x
and one can easily generalize this to see how field operators are related to wave-
functions for an arbitrary number of particles.
Cutoffs ruin translational invariance and calculations with them quickly be-
come difficult. Well now adopt the physicists convention of working directly
in the continuous case with no cutoff, at the price of having formulas that only
make sense as distributions. One needs to be aware that the correct interpreta-
tion of such formulas may require going back to the cutoff version.
In the continuum normalization we take as normalized eigenfunctions for the
free particle
1
|pi = p (x) = eipx
2
with Z
1 0
hp0 |pi = ei(pp )x dx = (p p0 )
2
The annihilation and creation operators satisfy
[a(p), a (p0 )] = (p p0 )
256
The field operators are then
Z
1
(x)
b = eipx a(p)dx
2
Z
1
b (x) = eipx a (p)dx
2
[(x),
b b 0 )] = [b (x), b (x0 )] = 0
(x
Z Z
1 0 0
[(x),
b b (x0 )] = eipx eip x [a(p), a (p0 )]dpdp0
2
Z Z
1 0 0
= eipx eip x (p p0 )dpdp0
2
Z
1 0
= eip(xx ) dp
2
=(x x0 )
There are observables that one can define simply using field operators. These
include:
. One can define a number density operator
The number operator N
b(x) = b (x)(x)
n b
257
and integrate it to get an operator with eigenvalues the total number of
particles in a state
Z
N=
b n
b(x)dx
Z Z Z
1 0 1
= eip x a (p0 ) eipx a(p)dpdp0 dx
2 2
Z Z
= (p p0 )a (p0 )a(p)dpdp0
Z
= a (p)a(p)dp
The total momentum operator Pb. This can be defined in terms of field
operators as
Z
d b
P =
b b (x)(i )(x)dx
dx
Z Z Z
1 0 1
= eip x a (p0 )(i)(ip) eipx a(p)dpdp0 dx
2 2
Z Z
= (p p0 )pa (p0 )a(p)dpdp0
Z
= pa (p)a(p)dp
The Hamiltonian H.
b This can be defined much like the momentum, just
changing
d 1 d2
i
dx 2m dx2
to find
Z Z 2
1 d2 b p
H=
b (x)(
b
2
)(x)dx = a (p)a(p)dp
2m dx 2m
258
If one can solve the one-particle Schrodinger equation for a complete set of
orthonormal wave functions n (x), one can describe this quantum system using
the same techniques as for the free particle. A creation-annihilation operator
pair an , an is associated to each eigenfunction, and quantum fields are defined
by X X
(x)
b = n (x)an , b (x) = n (x)an
n n
For Hamiltonians just quadratic in the quantum fields, quantum field the-
ories are quite tractable objects. They are in some sense just free quantum
oscillator systems, with all of their symmetry structure intact, but taking the
number of degrees of freedom to infinity. Higher order terms though make
quantum field theory a difficult and complicated subject, one that requires a
year-long graduate level course to master basic computational techniques, and
one that to this day resists mathematicians attempts to prove that many ex-
amples of such theories have even the basic expected properties. In the theory
of charged particles interacting with an electromagnetic field, when the electro-
magnetic field is treated classically one still has a Hamiltonian quadratic in the
field operators for the particles. But if the electromagnetic field is treated as a
quantum system, it acquires its own field operators, and the Hamiltonian is no
longer quadratic in the fields, a vastly more complicated situation described as
aninteracting quantum field theory.
Even if one restricts attention to the quantum fields describing one kind of
particles, there may be interactions between particles that add terms to the
Hamiltonian, and these will be higher order than quadratic. For instance, if
there is an interaction between such particles described by an interaction energy
v(y x), this can be described by adding the following quartic term to the
Hamiltonian
1 b
Z Z
(x)b (y)v(y x)(y)
b (x)dxdy
b
2
259
fermionic instead of bosonic oscillators. One gets fermionic fields that satisfy
anti-commutation relations
[(x),
b b (x0 )]+ = (x x0 )
260
that, choosing a complex structure and complex coordinates zj on phase space,
the non-zero Poisson bracket relations were
{zj , iz l } = jl
If we use the |pi basis for H1 , our complex coordinates will be (p), (p)
with Poisson bracket
{(p), i(p0 )} = (p p0 )
and under quantization we have
The fact that one gets a unitary Heisenberg Lie algebra representation 0 when
multiplying these operators by i implies the commutator relations
[a(p), a (p0 )] = (p p0 )1
Using instead the |xi basis for H1 , our complex coordinates will be (x), (x)
with Poisson brackets
{(x), i(x0 )} = (x x0 )
and quantization takes
(x) (x),
b (x) b (x), 1 1
261
262
Chapter 26
In our study of the harmonic oscillator we found that the symmetries of the sys-
tem could be studied using quadratic functions on the phase space. Classically
these gave a Lie algebra under the Poisson bracket, and quantization provided a
unitary representation 0 of the Lie algebra, with quadratic functions becoming
quadratic operators. In the case of quantum fields, the same pattern holds, with
the phase space now an infinite dimensional space, the single particle Hilbert
space. Certain specific quadratic functions of the fields will provide a Lie al-
gebra under the Poisson bracket, with quantization then providing a unitary
representation of the Lie algebra in terms of quadratic field operators.
Well begin with the most fundamental symmetry in quantum mechanics,
time translation symmetry, which determines the dynamics of the theory. For
the case of a free particle, the field theory Hamiltonian will be a quadratic
function of the fields, providing a basic example of how such functions generate
a unitary representation on the states of the quantum theory.
d
f = {f, h}
dt
263
assume here H time-independent) to the unitary operator
U (t) = eiHt
that determines how states (in the Schrodinger picture) evolve under time trans-
lation. In the Heisenberg picture states stay the same and operators evolve, with
their time evolution given by
d b
f = [fb, iH]
dt
which is the quantization of the classical dynamical equation. This has the
solution
{(x), i(x0 )} = (x x0 )
Time translation on this space is given by the Schrodinger equation, which says
that wavefunctions will evolve with time dependence given by
i 2
(x, t) = (x, t)
t 2m x2
+
1 2
Z
h= (x) (x)dx
2m x2
then we will get the single-particle Schrodinger equation from the Hamiltonian
264
dynamics, since
Note that there are other forms of the same Hamiltonian function, related
to the one we chose by integration by parts. One has
d2 d d d
(x) 2
(x) = ((x) (x)) | (x)|2
dx dx dx dx
d d d d2
= ((x) (x) ( (x))(x)) + ( 2 (x))(x)
dx dx dx dx
so neglecting integrals of derivatives (assuming boundary terms go to zero at
infinity), one could have used
Z +
1 + d2
Z
1 d
h= | (x)|2 dx or h = ( (x))(x)dx
2m dx 2m dx2
Instead of working with position space fields (x, t) we could work with
their momentum space components. Recall that we can write solutions to the
Schrodinger equation as
Z
1 p2
(x, t) = (p, t)eipx ei 2m t dp
2
where
p2
(p, t) = (p)ei 2m t
Using these as our coordinates on the H1 , dynamics is given by
p2
(p, t) = {(p, t), h} = i (p, t)
t 2m
and one can easily see that one can choose
Z 2
p
h= |(p)|2 dp
2m
265
as Hamiltonian function in momentum space coordinates. This is the same
expression one would get by substituting the expression for in terms of and
calculating h from its formula as a quadratic polynomial in the fields.
In momentum space, quantization is simply given by
Z 2
p
(p) a(p), h H b = a (p)a(p)dp
2m
266
Definition (Greens function or propagator). The Greens function or propa-
gator for a quantum field theory is the amplitude, for t > t0
G(x, t, x0 , t0 ) = h0|(x,
b t)b (x0 , t0 )|0i
The physical interpretation of these functions is that they describe the am-
plitude for a process in which a one-particle state localized at x is created at time
t0 , propagates for a time t t0 , and then its wave-function is compared to that
of a one-particle state localized at x. Using the solution for the time-dependent
field operator given earlier we find
Z + Z +
1 p2 0 p0 2 0
G(x, t, x0 , t0 ) = h0|eipx ei 2m t a(p)eip x ei 2m t a (p0 )|0idpdp0
2
Z + Z +
p2 0 0 p0 2 0
= eipx ei 2m t eip x ei 2m t (p p0 )dpdp0
Z+
0 p2 0
= eip(xx ) ei 2m (tt ) dp
One can evaluate this integral, finding
im 3 im
(x0 x)2
G(x0 , t0 , x, t) = ( 0
) 2 e 2(t0 t)
2(t t)
and that
lim G(x0 , t0 , x, t) = (x0 x)
tt0
While we have worked purely in the Hamiltonian formalism, one might won-
der what the Lagrangian for this system is. A Lagrangian that will give the
Schr
odinger equation as an Euler-Lagrange equation is
1 2
h = i +
L = i
t t 2m x2
or, using integration by parts to get an alternate form of h mentioned earlier
1 2
L = i | |
t 2m x
If one tries to define a canonical momentum for as L
on just gets i.
This justifies the Poisson bracket relation
{(x), i(x0 )} = (x x0 )
but, as expected for a case where the equation of motion is first-order in time,
such a canonical momentum is not independent of and the space of the wave-
functions is already a phase space. One could try and quantize this system
by path integral methods, for instance computing the propagator by doing the
integral Z
i
D e ~ S[]
over paths from (x, t) to (x0 , t0 ). However one needs to keep in mind the
warnings given earlier about path integrals over phase space, since that is what
one has here.
267
26.2 Spatial symmetries
The specification of the Hamiltonian gives the dynamics of the theory and de-
termines how states and operators transform under time translation. The action
of translations and rotations of the physical space R3 on a quantum field theory
is of a different nature. It is determined from the fact that the phase space H1
is a space of functions on this R3 , so any group action
xgx
(x) (g 1 x)
When this group action preserves the Poisson bracket on H1 , we expect to get
a unitary representation of the group on the quantum field theory state space,
as well as an action on the field operators by conjugation.
This will be the case for the group R3 of spatial translations and the group
SO(3) of spatial rotations, which (since rotations act non-trivially on transla-
tions) fit together into a group E(3) which is a semi-direct product R3 o SO(3),
something that we will study in detail in a later chapter. Since the action of
E(3) preserves Poisson brackets, it will have a unitary representation on the
quantum field theory, which we will study in this section.
If the action of E(3) also commutes with the group action on H1 generated
by the Hamiltonian h, then we will not only have a representation of E(3) on
the state space, but energy eigenstates of each energy will give representations
of E(3). This will be the case for the free particle, since it has a Hamiltonian h
that is invariant under rotations and translations. If one takes not free particles,
but particles in an external potential V , then one will still have a representation
of E(3) on the state space, but only the subgroup of E(3) that preserves the
potential (for instance, SO(3) for a central potential) will preserve the energy
eigenspaces.
For reasons that we will see later when we discuss special relativity, in the
single-particle theory the momentum operators P that generate space transla-
tions are taken to act oppositely to the Hamiltonian operator H (iH = t , and
iP = x ). For time translations we found that the action on field operators
was infinitesimally
d b b t), iH]
(x, t) = [(x, b
dt
which exponentiates to
b t)eia0 Hb
b t + a0 ) = eia0 Hb (x,
(x,
268
The momentum operator P b in the quantum field theory can be constructed in
terms of quadratic operators in the fields in the same way as the Hamiltonian,
just replacing the single-particle Hamiltonian operator by the single-particle
momentum operator P = i. So one has
Z
P=
b b (x)(i)(x)d
b 3
x
R3
In the last chapter we saw that, in terms of annihilation and creation operators,
this operator is just Z
P=
b p a (p)a(p)d3 p
R3
which is just the integral over momentum space of the momentum times the
number-density operator in momentum space.
For spatial rotations, we found in chapter 14 that these had generators the
angular momentum operators
L = X P = X (i)
These will generate the action of rotations on the field operators. For instance,
if R() is a rotation about the x3 axis by angle , we will have
(check sign).
269
characterized by integers, with these integers counting the number of quanta.
In the non-relativistic quantum field theory, the integer will just be the total
particle number.
On fields the U (1) group action is given by
z ei z, z ei z
and since we now want to treat fields, adding together an infinite number of
such 2d phase spaces, we will need the normal-ordered version in order to not
get 21 as the particle number eigenvalue for the vacuum state.
270
In momentum space, we simply do the above for each value of p and sum,
getting Z +
N=
b a (p)a(p)dp
where one needs to keep in mind that a (p)a(p) is really an operator valued
distribution, which must be integrated against something (here the identity
function) to get an operator.
Instead of working with a(p), the quantization of the Fourier transform of
(x), one could work with (x) directly, do the above for each value of x and
sum, getting Z +
N
b= b (x)(x)dx
b
b . b (x)(x)
with the Fourier transform relating the two formulas for N b is also an
operator valued distribution, with the interpretation of measuring the number
density at x.
On field operators, Nb satisfies
[N b = ,
b , ] b , b ] = b
b [N
which are the quantum version of the U (1) symmetry on the phase space we
began our discussion with.
An important property of N b that can be straightforwardly checked is that
Z +
1 2 b
[N , H] = [N ,
b b b b (x) (x)dx] = 0
2m x2
This implies that particle number is a conserved quantity: if we start out with
a state with a definite particle number, this will remain constant. Note that the
origin of this conservation law comes from the fact that N b is the quantized gen-
erator of the U (1) symmetry of phase transformations on complex-valued fields
. If we start with any hamiltonian function h on H1 that is invariant under the
U (1) (i.e. built out of terms with an equal number of s and s), then for such
a theory N b will commute with H b and particle number will be conserved. Note
though that one needs to take some care with arguments like this, which assume
that symmetries of the classical phase space give rise to unitary representations
in the quantum theory. The need to normal-order operator products, working
with operators that differ from the most straightforward quantization by an in-
finite constant, can cause a failure of symmetries to be realized as expected in
the quantum theory, a phenomenon known as an anomaly in the symmetry.
In quantum field theories, due to the infinite number of degrees of freedom,
the Stone-von Neumann theorem does not apply, and one can have unitarily
271
inequivalent representations of the algebra generated by the field operators,
leading to new kinds of behavior not seen in finite dimensional quantum systems.
In particular, one can have a space of states where the lowest energy state |0i
does not have the property
b |0i = 0, eiNb |0i = |0i
N
N 6 0, eiN |0i |i =
b |0i = 6 |0i (for 6= 0)
b
In this case, the vacuum state is not an eigenstate of N b so does not have a
well-defined particle number. If [N , H] = 0, the states |i will all have the same
b b
energy as |0i and there will be a multiplicity of different vacuum states, labeled
by . In such a case the U (1) symmetry is said to be spontaneously broken.
This phenomenon occurs when non-relativistic quantum field theory is used to
describe a superconductor. There the ground state will be a state without a
definite particle number, with electrons pairing up in a way that allows them to
lower their energy, condensing in the lowest energy state.
one can see that this will be invariant not just under U (1) phase transformations,
but also under transformations
1 1
2 2
.. U ..
. .
m m
272
After quantization, these quadratic combinations become quadratic combi-
nations in annihilation and creation operators ab , ab satisfying
[ab , ac ] = bc
So, for each X u(m) quantization will give us a representation of u(m) where
X acts as the operator
m
ab Xbc ac
X
b,c=1
As in the U (1) case, one gets an operator in the quantum field theory just by
summing over either the a(k) in momentum space, or the fields in configuration
space, finding for each X u(m) an operator
Z + m
bb (x)Xbc bc (x)dx
X
X
b=
b,c=1
that provides a representaion of u(m) on the quantum field theory state space.
When, as for the free-particle h we chose, the Hamiltonian is invariant under
U (m) transformations of the fields b , then we will have
[X,
b H]
b
In this case, if |0i is invariant under the U (m) symmetry, then energy eigenstates
of the quantum field theory will break up into irreducible representations of
U (m) and can be labeled accordingly. As in the U (1) case, the U (m) symmetry
may be spontaneously broken, with
X|0i
b 6 0
=
for some directions X in u(m). When this happens, just as in the U (1) case
states did not have well-defined particle number, now they will not carry well-
defined irreducible U (m) representation labels.
26.4 Fermions
Everything that was done in this chapter carries over straightforwardly to the
fermionic case. Our field operators now generate an infinite-dimensional Clifford
algebra and the quantum state space is now an infinite-dimensional version of
spinors. All the symmetries considered here have Lie algebra representations
constructed using quadratic combinations of the field operators in just the same
273
way as in the bosonic case. The method of annihilation and creation operators
used to take matrices with given commutation relations and produce quadratic
field operators with the same commutation relations works in the same way,
since if one checks the proof of this in chapter 17 one finds it continues to work
when one uses the CAR relations instead of the CCR relations.
274
Chapter 27
For the case of non-relativistic quantum mechanics, we saw that systems with
an arbitrary number of particles, bosons or fermions, could be described by
taking as the Hamiltonian phase space the state space H1 of the single-particle
quantum theory (e.g. the space of complex-valued wave-functions on R3 in the
bosonic case). This phase space is infinite-dimensional, but it is linear and it
can be quantized using the same techniques that work for the finite-dimensional
harmonic oscillator. This is an example of a quantum field theory since it is a
space of functions (fields, to physicists) that is being quantized.
We would like to find some similar way to proceed for the case of rela-
tivistic systems, finding relativistic quantum field theories capable of describ-
ing arbitrary numbers of particles, with the energy-momentum relationship
E 2 = |p|2 c2 + m2 c4 characteristic of special relativity, not the non-relativistic
2
limit |p| mc where E = |p| 2m . In general, a phase space can be thought of as
the space of initial conditions for an equation of motion, or equivalently, as the
space of solutions of the equation of motion. In the non-relativistic field theory,
the equation of motion is the first-order in time Schrodinger equation, and the
phase space is the space of fields (wave-functions) at a specified initial time,
say t = 0. This space carries a representation of the time-translation group R,
the space-translation group R3 and the rotation group SO(3). To construct a
relativistic quantum field theory, we want to find an analog of this space. It will
be some sort of linear space of functions satisfying an equation of motion, and
we will then quantize by applying harmonic oscillator methods.
Just as in the non-relativistic case, the space of solutions to the equation
of motion provides a representation of the group of space-time symmetries of
the theory. This group will now be the Poincare group, a ten-dimensional
group which includes a four-dimensional subgroup of translations in space-time,
and a six-dimensional subgroup (the Lorentz group), which combines spatial
rotations and boosts (transformations mixing spatial and time coordinates).
275
The representation of the Poincare group on the solutions to the relativistic
wave equation will in general be reducible. Irreducible such representations will
be the objects corresponding to elementary particles. Our first goal will be to
understand the Lorentz group, in later sections we will find representations of
this group, then move on to the Poincare group and its representations.
(x, y) x y = x0 y0 x1 y1 x2 y2 x3 y3
0 0 0 1
as
3
X
xy = x x
,=0
The choice of overall sign of is a convention, with the choice we are making
the most common one in discussions of relativistic quantum field theories.
Digression (Upper and lower indices). In many physics texts it is conventional
in discussions of special relativity to write formulas using both upper and lower
indices, related by
X3
x = x = x
=0
with the last form of this using the Einstein summation convention.
One motivation for introducing both upper and lower indices is that special
relativity is a limiting case of general relativity, which is a fully geometrical
theory based on taking space-time to be a manifold M with a metric g that
varies from point to point. In such a theory it is important to distinguish between
elements of the tangent space Tx (M ) at a point x M and elements of its dual,
the co-tangent space Tx (M ), while using the fact that the metric g provides an
inner product on Tx (M ) and thus an isomorphism Tx (M ) ' Tx (M ). In the
276
special relativity case, this distinction between Tx (M ) and Tx (M ) just comes
down to an issue of signs, but the upper and lower index notation is useful for
keeping track of those.
A second motivation is that position and momenta naturally live in dual
vector spaces, so one would like to distinguish between the vector space M 4 of
positions and the dual vector space of momenta. In the case though of a vector
space like M 4 which comes with a fixed inner product , this inner product
gives a fixed identification of M 4 and its dual, an identification that is also an
identification as representations of the Lorentz group. For simplicity, we will
not here try and distinguish by notation whether a vector is in M 4 or its dual,
so will just use lower indices, not both upper and lower indices.
We can take Fourier transforms with respect to the four space-time variables,
which will take functions of x0 , x1 , x2 , x3 to functions of the Fourier transform
variables p0 , p1 , p2 , p3 . The definition we will use for this Fourier transform will
277
be
Z
1
fe(p) = eipx f (x)d4 x
(2)2 M 4
Z
1
= ei(p0 x0 p1 x1 p2 x2 p3 x3 ) f (x)dx0 d3 x
(2)2 M 4
and the Fourier inversion formula is
Z
1
f (x) = eipx fe(p)d4 p
(2)2 M4
Note that our definition puts one factor of 12 with each Fourier (or inverse
Fourier) transform with respect to a single variable. A common alternate con-
vention among physicists is to put all factors of 2 with the p integrals (and
thus in the inverse Fourier transform), none in the definition of fe(p), the Fourier
transform itself.
The reason why one conventionally defines the Hamiltonian operator as i t
but the momentum operator with components i x j
is due to the sign change
between the time and space variables that occurs in this Fourier transform in
the exponent of the exponential.
0 0 0 1 0 0 0 1
The Lorentz group has four components, with the component of the iden-
tity a subgroup called SO(1, 3) (which some call SO+ (1, 3)). The other three
components arise by multiplication of elements in SO(1, 3) by P, T, P T where
1 0 0 0
0 1 0 0
P = 0 0 1 0
0 0 0 1
278
is called the parity transformation, reversing the orientation of the spatial
variables, and
1 0 0 0
0 1 0 0
T =0
0 1 0
0 0 0 1
reverses the time orientation.
The Lorentz group has a subgroup SO(3) of transformations that just act
on the spatial components, given by matrices of the form
1 0 0 0
0
=0
0
where is in SO(3). For each pair j, k of spatial directions one has the usual
SO(2) subgroup of rotations in the j k plane, but now in addition for each pair
0, j of the time direction with a spatial direction, one has SO(1, 1) subgroups
of matrices of transformations called boosts in the j direction. For example,
for j = 1, one has the subgroup of SO(1, 3) of matrices of the form
cosh sinh 0 0
sinh cosh 0 0
=
0
0 1 0
0 0 0 1
for R.
The Lorentz group is six-dimensional. For a basis of its Lie algebra one can
take six matrices J for , 0, 1, 2, 3 and j < k. For the spatial indices, these
are
0 0 0 0 0 0 0 0 0 0 0 0
0 0 1 0 0 0 0 1 0 0 0 0
J12 = 0 1 0 0 , J13 = 0 0 0 0 , J23 = 0 0 0 1
0 0 0 0 0 1 0 0 0 0 1 0
which correspond to the basis elements of the Lie algebra of SO(3) that we saw
in an earlier chapter. One can rename these as
279
Taking the first index 0, one gets three elements corresponding to infinitesi-
mal boosts in the three spatial directions
0 1 0 0 0 0 1 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0
J01 = , J = , J =
0 02 1 0 03 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 1 0 0 0
One can easily calculate the commutation relations between the Kj and Jj ,
which show that the Kj transform as a vector under infinitesimal rotations. For
instance, for infinitesimal rotations about the x1 axis, one finds
1 1
Aj = (Jj + iKj ), Bj = (Jj iKj )
2 2
one finds
[A1 , A2 ] = A3 , [A3 , A1 ] = A2 , [A2 , A3 ] = A1
and
[B1 , B2 ] = B3 , [B3 , B1 ] = B2 , [B2 , B3 ] = B1
we have
so(1, 3) C = sl(2, C) sl(2, C)
In the next section well see the origin of this phenomenon at the group level.
280
27.3 Spin and the Lorentz group
Just as the groups SO(n) have double covers Spin(n), the group SO(1, 3) has a
double cover, which we will show can be identified with the group SL(2, C) of
2 by 2 complex matrices with unit determinant. This group will have the same
Lie algebra as the SO(1, 3), and we will sometimes refer to either group as the
Lorentz group.
Recall that for SO(3) the spin double cover Spin(3) can be identified with
either Sp(1) (the unit quaternions) or SU (2), and then the action of Spin(3)
as SO(3) rotations of R3 was given by conjugation of imaginary quaternions or
certain 2 by 2 complex matrices respectively. In the SU (2) case this was done
explicitly by identifying
x3 x1 ix2
(x1 , x2 , x3 )
x1 + ix2 x3
and then showing that conjugating this matrix by an element of SU (2) was a
linear map leaving invariant
x3 x1 ix2
det = x21 + x22 + x23
x1 + ix2 x3
This provides a very useful way to think of Minkowski space: as complex self-
adjoint 2 by 2 matrices, with norm-squared the determinant of the matrix.
The linear transformation
x0 + x3 x1 ix2 x0 + x3 x1 ix2
x1 + ix2 x0 x3 x1 + ix2 x0 x3
281
It also takes self-adjoint matrices to self-adjoints, and thus R4 to R4 , since
x0 + x3 x1 ix2 x0 + x3 x1 ix2
( ) =( )
x1 + ix2 x0 x3 x1 + ix2 x0 x3
= x0 + x3 x1 ix2
x1 + ix2 x0 x3
Note that both and give the same linear transformation when they act by
conjugation like this. One can show that all elements of SO(3, 1) arise as such
conjugation maps, by finding appropriate that give rotations or boosts in the
planes, since these generate the group.
Recall that the double covering map
: SU (2) SO(3)
Digression (The complex group Spin(4, C) and its real forms). Recall that
we found that Spin(4) = Sp(1) Sp(1), with the corresponding SO(4) trans-
formation given by identifying R4 with the quaternions H and taking not just
conjugations by unit quaternions, but both left and right multiplication by dis-
tinct unit quaternions. Rewriting this in terms of complex matrices instead of
quaternions, we have Spin(4) = SU (2) SU (2), and a pair 1 , 2 of SU (2)
matrices acts as an SO(4) rotation by
x0 ix3 x2 ix1 x0 ix3 x2 ix1
1 2
x2 ix1 x0 + ix3 x2 ix1 x0 + ix3
282
so gives an element of SO(2, 2) and we see that Spin(2, 2) = SL(2, R)
SL(2, R).
The three different examples
and
Spin(2, 2) = SL(2, R) SL(2, R)
that we have seen are all so-called real forms of a fact about complex groups
that one can get by complexifying any of the examples, i.e. considering elements
(x0 , x1 , x2 , x3 ) C4 , not just in R4 . For instance, in the Spin(4) case, taking
the x0 , x1 , x2 , x3 in the matrix
x0 ix3 x2 ix1
x2 ix1 x0 + ix3
preserves this space as well as the determinant (z02 + z12 + z22 + z32 ) for 1 and 2
not just in SU (2), but in the larger group SL(2, C). So we find that the group
SO(4, C) of complex orthogonal transformations of C4 has spin double cover
and one gets from this the three different real forms described above by picking
various choices of real subspaces R4 C4 .
Since spin(4, C) = so(1, 3) C, this relation between complex Lie groups
corresponds to the Lie algebra relation
283
often not treated. A typical example is Peskin-Schroeder[33], see the beginning
of Chapter 3. Another example is Quantum Field Theory in a Nutshell by
Tony Zee, see Chapter II.3 [56] (and test your understanding by interpreting
properly some of the statements included there such as The mathematically
sophisticated say the algebra SO(3, 1) is isomorphic to SU (2) SU (2)).
284
Chapter 28
Representations of the
Lorentz Group
285
1
2 , s2= 0 is one of the half-spinor representations of Spin(4) on C2 , s1 = 0, s2 =
1
2 is the other, and s1 = s2 = 21 is the representation on four-dimensional
(complexified) vectors.
Turning now to Spin(1, 3) = SL(2, C), can take the SU (2) matrices acting
on homogeneous polynomials to instead be SL(2, C) matrices and still have
irreducible representations of dimension 2s + 1 for s = 0, 12 , 1, . . .. These will
now be representations (s , V s ) of SL(2, C). There are several things that are
different though about these representations:
They are not unitary (except in the case of the trivial representation).
1
For example, for the defining representation V 2 on C2 and the Hermitian
inner product < , >
0
10
1 1
< , >= 1 2 = 1 10 + 2 20
2 20 20
286
occurs on a Hamiltonian phase space. For the simplest case of the phase
space R2 for one degree of freedom, such a form is given with respect to
a basis as the same matrix
0 1
=
1 0
287
Theorem (Classification of finite dimensional sl(2, C) representations). The
irreducible representations of sl(2, C) are labeled by (s1 , s2 ) for sj = 0, 21 , 1, . . ..
These representations are built out of the representations (s , V s ) with the ir-
reducible (s1 , s2 ) given by
(s1 s2 , V s1 V s2 )
and having dimension (2s1 + 1)(2s2 + 1).
All these representations are also representations of the group SL(2, C) and
one has the same classification theorem for the group, although we will not try
and prove this. We will also not try and study these representations in general,
but will restrict attention to the four cases of most physical interest.
(0, 0): The trivial representation on C, also called the spin 0 or scalar
representation.
( 12 , 0): These are called left-handed (for reasons we will see later on) Weyl
spinors. We will often denote the representation space C2 in this case as
SL , and write an element of it as L .
(0, 12 ): These are called right-handed Weyl spinors. We will often denote
the representation space C2 in this case as SR , and write an element of it
as R .
( 21 , 12 ): This is called the vector representation since it is the complexifi-
cation of the action of SL(2, C) as SO(1, 3) transformations of space-time
vectors that we saw earlier. Recall that for SL(2, C) this action was
x0 + x3 x1 ix2 x0 + x3 x1 ix2
x1 + ix2 x0 x3 x1 + ix2 x0 x3
288
A standard index notation for such things is called the van der Waer-
den notation. It uses a lower index taking values 1, 2 to label the
components
1
L = =
2
and in this notation acts by
Writing elements of the dual as row vectors, our example above of a par-
ticular acts by
1 2 1 2 ei 2 3
=
where
0 1
=
1 0
289
Another common notation among physicists puts a bar over the to
denote that the vector is in this representation, but well reserve that
notation for complex conjugation. The corresponding to a rotation
about the z-axis acts as
1 1
ei 2 3
2 2
SR : This is the dual representation to SR , with SL(2, C) acting on
R SR by
1
R ( )T R
and the index notation uses raised dotted indices
1 T
(( ) )
Another copy of
0 1
=
1 0
gives the isomorphism of SR and SR as representations, by
=
290
We will represent elements of Cliff(1, 3) as such 4 by 4 matrices, but should
keep in mind that we are working in the complexification of the Clifford algebra
that corresponds to the Lorentz group, so there is some sort of condition on
the matrices that should be kept track of. There are several different choices of
how to explicitly represent these matrices, and for different purposes, different
ones are most convenient. The one we will begin with and mostly use is some-
times called the chiral or Weyl representation, and is the most convenient for
discussing massless charged particles. We will try and follow the conventions
used for this representation in the standard quantum field theory textbook of
Peskin and Schroeder[33].
Writing 4 by 4 matrices in 2 by 2 block form and using the Pauli matrices
j we assign the following matrices to Clifford algebra generators
0 1 0 1 0 2 0 3
0 = , 1 = , 2 = , 3 =
1 0 1 0 2 0 3 0
One can easily check that these satisfy the Clifford algebra relations for gener-
ators of Cliff(1, 3): they anti-commute with each other and
02 = 1, 12 = 1, 22 = 1, 32 = 1
The quadratic Clifford algebra elements 12 j k for k < j satisfy the commu-
tation relations of so(1, 3). These are explicitly
1 i 3 0 1 i 2 0 1 i 1 0
1 2 = , 1 3 = , 2 3 =
2 2 0 3 2 2 0 2 2 2 0 1
and
1 1 1 0 1 1 2 0 1 1 3 0
0 1 = , 0 2 = , 0 3 =
2 2 0 1 2 2 0 2 2 2 0 3
1 1 1
0 (J1 ) = 2 3 , 0 (J2 ) = 1 3 , 0 (J3 ) = 1 2
2 2 2
and
1 1 1
0 (K1 ) =
0 1 , 0 (K2 ) = 0 2 , 0 (K3 ) = 0 3
2 2 2
On the two commuting SL(2, C) subalgebras of so(1, 3) C with bases
1 1
Aj = (Jj + iKj ), Bj = (Jj iKj )
2 2
this representation is
i 1 0 i 2 0 i 3 0
0 (A1 ) = , 0 (A2 ) = , 0 (A3 ) =
2 0 0 2 0 0 2 0 0
291
and
0 i 0 0 0 i 0 0 0 i 0 0
(B1 ) = , (B2 ) = , (B3 ) =
2 0 1 2 0 2 2 0 3
We see explicitly that the action of the quadratic elements of the Clifford
algebra on the spinor representation C4 is reducible, decomposing as the direct
sum SL SR of two inequivalent representations on C2
L
=
R
where the index on the left takes values 1, 2, 3, 4 and the indices , on the
right each take values 1, 2.
An important element of the Clifford algebra is the one you get by multi-
plying all of the basis elements together. Physicists traditionally multiply this
by i to make it self-adjoint and define
1 0
5 = i0 1 2 3 =
0 1
This can be used to produce projection operators from the Dirac spinors onto
the left and right-handed Weyl spinors
1 1
(1 5 ) = L , (1 + 5 ) = R
2 2
There are two other commonly used representations of the Clifford algebra
relations, related to the one above by a change of basis. The Dirac representation
is useful to describe massive charged particles, especially in the non-relativistic
limit. Generators are given by
1 0 0 1 0 2 0 3
0D = , 1D = , 2D = , 3D =
0 1 1 0 2 0 3 0
and the projection operators for Weyl spinors are no longer diagonal, since
0 1
5D =
1 0
292
A third representation, the Majorana representation, is given by
1 0 0 0
0 0 0 1
|p|2
0 0 1 0 ipx i 2m t
1 0
0M , 1M = i e e 0 0
= i
0 1 0 0 0 0 1 0
1 0 0 0 0 0 0 1
0 0 0 1 0 1 0 0
0 0 1 0 1 0 0 0
2M , 3M = i
= i
0 1 0 0 0 0 0 1
1 0 0 0 0 0 1 0
with
0 1 0 0
1 0 0 0
5M = i
0
0 0 1
0 0 1 0
and
(0 1 2 3 )2 = 1
SM C = SL SR
the fact that complexifying Majorana spinors gives the two kinds of Weyl
spinors.
293
28.3 For further reading
Most quantum field theory textbook have extensive discussions of spinor rep-
resentations of the Lorentz group and gamma matrices. Typical examples are
Peskin-Schroeder[33] and Quantum Field Theory in a Nutshell by Tony Zee, see
Chapter II.3 and Appendix E[56].
294
Chapter 29
The Lorentz group and the group of space-time translations fit together into a
larger group, the Poincare group. However, before moving on to discuss that
group and its representations, well first consider what happens in general for
constructions like this, and then examine specifically two simpler such cases, the
groups E(2) and E(3) of Euclidean motions in 2 and 3 dimensions. These raise
many of the same issues as the Poincare group, and the case of E(2) appears as
an important part of the Poincare group story.
295
This is an example of the following general construction
Definition (Semi-direct product). A semi-direct product
G=N oK
of two groups N and K is a group such that
As a set
G=N K
and any element g G can uniquely be written as g = nk, with n N
and k K.
N is a normal subgroup of G, a condition that is written as
N G
There is a homomorphism
: k K k Aut(N )
such that the group law in N o K is given by
(n1 k1 )(n2 k2 ) = n1 k1 n2 k2 = n1 k1 (n2 )k1 k2
296
For any vector space V , the group V o GL(V ) described at the beginning
of this section.
Recall that we defined the Heisenberg group H2n+1 as the space R2n+1
with the group law
0
1 p p
(p, q, c) (p0 , q0 , c0 ) = (p + p0 , q + q0 , c + c0 + (( , ))
2 q q0
where p Rn , q Rn , c R and is the symplectic (non-degenerate
antisymmetric bilinear) form
0
p p
( , ) = qp0 p q0
q q0
Taking (p, 0, c) N = Rn+1 H2n+1 and (0, q, 0) K = Rn H2n+1
one has
H2n+1 = Rn+1 o Rn
In this case, since we know the group law, we find that is given by
(0,q,0) (p, 0, c) =(0, q, 0) (p, 0, c) (0, q, 0)1
=(0, q, 0) (p, 0, c) (0, q, 0)
1
=(0, q, 0) (p, q, c + p q)
2
=(p, 0, c + p q)
297
Here our map just associates to an element of SO(n) the corresponding
rotation of Rn (an automorphism of the translation group)
R (v) = Rv
One can similarly construct a double cover Rn oSpin(n) of the Euclidean group,
and defining the multiplication using
taking pairs (v, R)
R (v) = (R)v
298
be the set of
Definition (Character group). For N a commutative group, let N
characters of N , i.e. functions
:N C
We only will actually need the case N = Rn , where we have already seen
that the irreducible representations are one-dimensional and given by
p (x) = eipx
g f (x) = f (g 1 x)
V = V
, i.e.
where V is the subspace on which N acts by the character N
v V = (n)v = (n)v
If V and N were finite dimensional this would be a finite sum. We are interested
in cases where this is not true, but do not want to try and enter into the
necessary analytical machinery to properly deal with this question, so this should
be considered just a motivation for the construction that follows.
If one N occurs in this sum, so will all the other N that one
gets by acting on with elements of K. The set of these is called the K-orbit
O of in N under the action of K. Distinct orbits will correspond to distinct
299
irreducible representations, which suggests that we try and construct irreducible
representations of N o K by taking functions on K-orbits in N .
One example of this phenomenon is the Schrodinger representation of the
Heisenberg group H2n+1 = Rn+1 o R2n on functions on Rn . We will not work
out the details here, but in this case N is the commutative group Rn+1 of
elements of H2n+1 of the form (p, 0, c), and N is just the dual space, also Rn+1 .
n
The group K = R of elements of the form (0, q, 0) acts on this, and the K-
orbit O of any element of N that is non-zero on (0, 0, c) will be a copy of Rn .
Functions on this orbit will give a version of the Schrodinger representation.
p = eipx
To better understand the second case, where the representations are infinite-
dimensional function spaces, one can consider the case n = 3, and observe that
this is just the space of eigenfunctions of the free-particle Hamiltonian of a
fixed energy, and thus also the space of solutions of the Schrodinger equation
with this energy. This space of solution is an irreducible representation of the
Euclidean group E(3), with the action of the translation subgroup generated
by the momentum operators P, and the rotation subgroup SO(3) generated
by the angular momentum operators (both momentum and angular momentum
operators commute with the Hamiltonian).
300
More explicitly, consider the quantum mechanical state space H of functions
on R3 , describing a free particle. Recall that we would like to work with the
functions
k (x) = eipx = |pi
as a sort of continuous basis of H, even though these are not integrable functions.
The formalism for doing this uses distributions and the orthonormality relation
hp|p0 i = (p p0 )
The Euclidean group E(3) will act on H, with the action generated by mo-
mentum and angular momentum operators, but this representation will not be
irreducible. To get an irreducible representation, we need to restrict to an ap-
propriate subspace of H, taking only those with Fourier transforms (p)
e that
are non-zero just when |p|2 has a fixed value (these are just the SO(n)-orbits
Op of the rotation group on momentum space). One can also think of this
in terms of the action of the operator |P|2 , which plays the role of a Casimir
operator in this case. Recall that irreducible representations of SO(3) could
be characterized by the eigenvalue with which the Casimir operator L2 acted
on them. Here we want to consider states with a fixed eigenvalue of a Casimir
operator for E(3), the operator |P|2 (the operator L2 is no longer a Casimir in
this case, since it does not commute with the generators P of the translation
part of E(3)).
Instead of the Casimir operator, we can use something proportional to it,
the Hamiltonian H, and think of the subspace we are looking for as a space of
eigenfunctions of H with energy E. We have
1 1 d2 d2 d2 1
H= (P12 + P22 + P32 ) = ( 2 + 2 + 2) =
2m 2m dx1 dx2 dx3 2m
where is the Laplacian. The energy eigenfunctions of energy E form a sub-
space HE of the functions in H, those satisfying
1 d2 d2 d2
H = ( 2 + 2 + 2 ) = E
2m dx1 dx2 dx3
This has solutions
p (x) = eipx = |pi
where p satisfies |p|2 = 2mE. So we get an orbit Op in momentum space
R3 given by a sphere of radius r = 2mE and this corresponds to a single
irreducible representation of E(3) on HE . This representation is infinite di-
mensional and unitary, although we need an appropriate version of the inner
301
product, which we will not write down here, although we will later do this for
the case of the Poincare group.
One could of course do the same thing in two dimensions, getting infinite
dimensional unitary irreducible representations of E(2) corresponding to circles
in R2 of radius r > 0, as well as one dimensional representations labeled by
integers (the representations of SO(2)), corresponding to the orbit of radius 0.
In the case of E(2), this is all of the irreducible representations, but for
E(n), n > 2 there will be other irreducibles, and the outline of their construction
goes as follows. For each orbit Op in N = Rn , it turns out that we can get the
group to act not just on scalar functions on Op , but on certain vector-valued
functions. For each p Rn there is a subgroup of SO(n) of rotations that leave
p invariant. One can see by going to a coordinate system where p points in
the nth direction that rotations that just act on the other n 1 directions will
leave this invariant, and these will form a group SO(n 1). For different p
in an orbit one gets different specific SO(n 1) subgroups of SO(n), but they
are all isomorphic. SO(n 1) is called the little group of the element p of
N , or of the corresponding orbit Op . Taking not scalar functions on the orbit,
but functions that take values in a representation of SO(n 1), it is not hard
to show that one can get a consistent action of the full group E(n) on such
functions, and that this representation is irreducible and unitary. For the case
of E(3), to each orbit of radius r, we will get an irreducible representation for
each choice of a representation of the little group SO(2). These are labeled by
integers m, so our irreducible representations are labeled by pairs (r, m). In the
case of E(2), the little group of each non-zero p is trivial, so irreducibles are
labeled just by the radius r.
302
Chapter 30
The Poincar
e Group and its
Representations
In the previous chapter we saw that one can take the semi-direct product of
spatial translations and rotations and that the resulting group has infinite-
dimensional unitary representations on the state space of a quantum free parti-
cle. The free particle Hamiltonian plays the role of a Casimir operator: to get
irreducible representations one fixes the eigenvalue of the Hamiltonian (the en-
ergy), and then the representation is on the space of solutions to the Schrodinger
equation with this energy. This is a non-relativistic procedure, treating time and
space (and correspondingly the Hamiltonian and the momenta) differently. For
a relativistic analog, we will use instead the semi-direct product of space-time
translations and Lorentz transformations. Irreducible representations of this
group will be labeled by a continuous parameter (the mass) and a discrete pa-
rameter (the spin or helicity), and these will correspond to possible relativistic
elementary particles.
In the non-relativistic case, the representation occurred as a space of solu-
tions to a differential equation, the Schrodinger equation. There is an analogous
description of the irreducible Poincare group representations as spaces of solu-
tions of relativistic wave equations, but we will put off that story until succeeding
chapters.
303
We will refer to both of these groups as the Poincare group, with the
double-cover the meaning only when we need it because spinor representations
of the Lorentz group are involved. The two groups have the same Lie algebra,
so the distinction is not needed in discussions that only need the Lie algebra.
Elements of the group P will be written as pairs (a, ), with a R4 and
SO(1, 3) and the group law is
The Lie algebra LieP = LieP has dimension 10, with basis
P0 , P1 , P2 , P3 , J1 , J2 , J3 , K1 , K2 , K3
where the first four elements are a basis of the Lie algebra of the translation
group (note that we are following physicists in using the same notation for Lie
algebra basis elements and the corresponding operators in a representation of
the Lie algebra), and the next six are a basis of so(1, 3), with the Jj giving
the subgroup of spatial rotations, the Kj the boosts. We already know the
commutation relations for the translation subgroup, which is commutative so
[Pj , Pk ] = 0
To find the Lie algebra relations between translation and so(1, 3) basis ele-
ments, recall that the Lie algebra commutators are related to the group multi-
plication by
d
[X, Y ] = (etX Y etX )|t=0
dt
Using the semi-direct product relation
304
Taking = etJ1 one finds
1 0 0 0 a0 0
d d
0 1 0 0 a1
0
a|t=0 = =
dt dt 0 0 cos t sin t a2 a3
0 0 sin t cos t a3 |t=0 a2
This is a relation on finite translations (a, 1), but it also hold for infinitesimal
translations, so we have
P0 1 0 0 0 P0 0
P1 d 0 1 0 0 P1
0
P2 ] = dt 0 0 cos t sin t P2
[J1 , =P3
which shows that the Pj transform as vectors under infinitesimal rotations about
the 1-axis, so
P3 0
and in general one will have
305
= R4 is the space of characters of the translation
For the Poincare group, N
group of Minkowski space, so functions
p p
and, restricting attention to the p0 p3 plane, the picture of the orbits looks
like this
Unlike the Euclidean group case, here there are several different kinds of
orbits O . Well examine them and the corresponding stabilizer groups K
each in turn, and see what can be said about the associated representations.
One way to understand the equations describing these orbits is to note that
the different orbits correspond to different eigenvalues of the Poincare group
Casimir operator
P 2 = P02 P12 P22 P32
This operator commutes with all the generators of the Lie algebra of the Poincare
group, so by Schurs lemma it must act as a scalar times the identity on an
306
irreducible representation (recall that the same phenomenon occurs for SU (2)
representations, which can be characterized by the eigenvalue j(j +1) of the Cas-
mir operator J2 for SU (2)). At a point p = (p0 , p1 , p2 , p3 ) in energy-momentum
space, the Pj operators are diagonalized and P 2 will act by the scalar
Note that in this chapter we are just classifying Poincare group representa-
tions, not actually constructing them. It is possible to construct these represen-
tations using the data we will find that classifies them, but this would require
introducing some techniques (for so-called induced representations) that go
beyond the scope of this course. In later chapters we will explicitly construct
these representations in certain specific cases as solutions to certain relativistic
wave equations.
so q
p0 = p21 + p22 + p23 + m2
307
We will later study the case of spin 12 , where one must use the double cover
SU (2) of SO(3). The Poincare group representation will be on functions on
the orbit that take values in two copies of the spinor representation of SU (2).
These will correspond to solutions of a wave equation called the massive Dirac
equation.
For choices of higher spin representations of the stabilizer group, one can
again find appropriate wave equations and construct Poincare group represen-
tations on their space of solutions, but we will not enter into this topic.
Again, one has the same stabilizer group Km,0,0,0 = SO(3) and the same con-
stuctions of wave equations of various spins and Poincare group representations
on their solution spaces as in the positive energy case. Since negative energies
lead to unstable, unphysical theories, we will see that these representations are
treated differently under quantization, corresponding physically not to particles,
but to anti-particles.
It is not too difficult to see that the stabilizer group of the orbit is K0,0,0,m =
SO(2, 1). This is isomorphic to the group SL(2, R), and it has no finite-
dimensional unitary representations. These orbits correspond physically to
tachyons, particles that move faster than the speed of light, and there is
no known way to consistently incorporate them in a conventional theory.
308
translations acting trivially. These representations are not unitary, so not usable
for our purposes.
One such vector is p = (1, 0, 0, 1) and the orbit of the vector under the action
of the Lorentz group will be the upper half of the full null-cone, the half with
energy p0 > 0, satisfying q
p0 = p21 + p22 + p23
Representations such that the two translations act trivially. These are
irreducible representations of SO(2), so one-dimensional and characterized
by an integer n (half-integers when one uses the Poincare group double
cover).
The first of these two gives irreducible representations of the Poincare group
on certain functions on the positive energy null-cone, labeled by the integer n,
which is called the helicity of the representation. We will in later chapters
consider the cases n = 0 (massless scalars, wave-equation the Klein-Gordon
equation), n = 21 (Weyl spinors, wave equation the Weyl equation), and n =
1 (photons, wave equation the Maxwell equations).
The second sort of representation of E(2) gives representations of the Poincare
group known as continuous spin representations, but these seem not to cor-
respond to any known physical phenomena.
309
30.3 For further reading
The Poincare group and its Lie algebra is discussed in pretty much any quantum
field theory textbook. Weinberg [51] (Chapter 2) has some discussion of the
representations of the Poincare group on single particle state spaces that we have
classified here. Folland [19] (Chapter 4.4) and Berndt [7] (Chapter 7.5) discuss
the actual construction of these representations using the induced representation
methods that we have chosen not to try and explain here.
310
Chapter 31
311
31.1 The Klein-Gordon equation and its solu-
tions
Recall that a condition characterizing the orbit in momentum space that we
want to study was that the Casimir operator P 2 of the Poincare group acts on
the representation corresponding to the orbit as the scalar m2 . So, we have the
operator equation
P 2 = P02 P12 P22 P32 = m2
characterizing the Poincare group representation we are interested in. Interpret-
ing the Pj as the standard differentiation operators i x j
on a space of wave-
functions, generating the infinitesimal action of the translation group on such
wave-functions, we get the following differential equation for wave-functions:
Definition (Klein-Gordon equation). The Klein-Gordon equation is the second-
order partial differential equation
2 2 2 2
( + 2 + 2 + ) = m2
t2 x1 x2 x23
or
2
( + m2 ) = 0
t2
for functions (x) on Minkowski space (which may be real or complex valued).
This equation is the obvious simplest Lorentz-invariant wave equation to try,
and historically was the one Schrodinger first tried (he then realized it could not
account for atomic spectra and instead used the non-relativistic equation that
bears his name). Taking Fourier transforms
Z
1
(p) =
e d4 xeipx (x)
(2)2
312
In the non-relativistic case, a continuous basis of solutions of the Schrodinger
equation labeled by p R3 was given by the functions
|p|2
eipx ei 2m t
313
with the integral over the 3d hyperboloid expressed as a 4d integral over R4
with a delta-function on the hyperboloid in the argument.
The Poincare group action on the coordinates (p)
e on H1 will be given by
e = eipa (
U (a, )(p) e 1 p)
(Fix this and make it more explicit after deal with symmetries later).
The delta function distribution with argument a function f (x) depends only
on the zeros of f , and if f 0 = 0 at such these zeros, one has
X 1
(f (x)) = (f 0 (xj )(x xj )) = 0 (x xj )
|f (xj )|
xj :f (xj )=0
For each p, one can applying this to the case of the function of p0 given by
f = p20 p2
on R4 , and using
d 2
(p p2 ) = 2p0 = 2p
dp0 0
one finds
Z
1 1 i(pxp0 t)
(x, t) = ((p0 p ) + (p0 + p ))(p)e
e dp0 d3 p
(2)3/2 M 4 2p
Z 3
1 e+ (p)eip t + e (p)eip t )eipx d p
= (
(2)3/2 R3 2p
Here
e+ (p) = (
e p , p), e (p) = (
e p , p)
are the values of e on the positive and negative energy hyperboloids. We see
that instead of thinking of the Fourier transforms of solutions as taking values
on energy-momentum hyperboloids, we can think of them as taking values just
on the space R3 of momenta (just as in the non-relativistic case), but we do
have to use both positive and negative energy Fourier components, and to get
a Lorentz invariant measure need to use
d3 p
2p
instead of d3 p.
A general complex-valued solution to the Klein-Gordon equation will be
given by the two complex-valued functions e+ , e , but we could impose the
condition that the solution be real-valued, in which case one can check that the
pair of functions must satisfy the condition
e (p) = e+ (p)
Real-valued solutions of the Klein-Gordon equation thus correspond to arbitrary
complex-valued functions e+ defined on the positive energy hyperboloid, which
fixes the value of the other function e .
314
31.2 Classical relativistic scalar field theory
We would like to set up the Hamiltonian formalism, finding a phase space H1 and
a Hamiltonian function h on it such that Hamiltons equations will give us Klein-
Gordon equation as equation of motion. Such a phase space will be an infinite-
dimensional function space and the Hamiltonian will be a functional. We will
here blithely ignore the analytic difficulties of working with such spaces, and use
physicists methods, with formulas that can be given a legitimate interpretation
by being more careful and using distributions. Note that now we will take the
fields to be real-valued, this is the so-called real scalar field.
Since the Klein-Gordon equation is second order in time, solutions will be
parametrized by initial data which, unlike the non-relativistic case now requires
the specification at t = 0 of not one, but two functions,
(x) = (x, 0), (x) = (x, t)|t=0
t
the values of the field and its first time derivative.
We will take as our phase space H1 the space of pairs of functions (, ),
with coordinates (x), (x) and Poisson brackets
{(x), (x0 )} = (x x0 ), {(x), (x0 )} = {(x), (x0 )} = 0
We want to get the Klein-Gordon equation for (x, t) as the following pair of
first order equations
= , = ( m2 )
t t
which together imply
2
= ( m2 )
t2
To get these as equations of motion, we just need to find a Hamiltonian
function h on the phase space H1 such that
={, h} =
t
={, h} = ( m2 )
t
One can check that two choices of Hamiltonian function that will have this
property are Z
h= H(x)d3 x
R3
where
1 2 1
H= ( + m2 2 ) or H = ( 2 + ()2 + m2 2 )
2 2
where the two different integrands H(x) are related (as in the non-relativistic
case) by integration by parts so these just differ by boundary terms that are
assumed to vanish.
315
(Should I put in here an explicit calculation, following Das).
One could instead have taken as starting point the Lagrangian formalism,
with an action Z
S= L d4 x
M4
where
1 2
L= (( ) ()2 m2 2 )
2 t
This action is a functional now of fields on Minkowski space M 4 and is Lorentz
invariant. The Euler-Lagrange equations give as equation of motion the Klein-
Gordon equation
(2 m2 ) = 0
One recovers the Hamiltonian formalism by seeing that the canonical momentum
for is
L
= =
and the Hamiltonian density is
1
H = L = ( 2 + ()2 + m2 2 )
2
Besides the position-space Hamiltonian formalism, we would like to have one
for the momentum space components of the field, since for a free field it is these
that will decouple into an infinite collection of harmonic oscillators. For a real
solution to the Klein-Gordon equation we have
Z 3
1 e+ (p)eip t + e (p)eip t )eipx d p
(x, t) = ( +
(2)3/2 R3 2p
d3 p
Z
1
= 3/2
(e+ (p)eip t eipx + e+ (p)eip t eipx )
(2) R3 2p
where we have used the symmetry of the integration over p to integrate over
p instead of p.
We can choose a new way of normalizing Fourier coefficients, one that reflects
the fact that the Lorentz-invariant notion is that of integrating over the energy-
momentum hyperboloid rather than momentum space
e+ (p)
(p) = p
2p
d3 p
Z
1
(x, t) = ((p)eip t eipx + (p)eip t eipx ) p
(2)3/2 R3 2p
316
The (p), (p) will have the same sort of Poisson bracket relations as the
z, z for a single harmonic oscillator, or the (p), (p) Fourier coefficients in the
case of the non-relativistic field:
To see this, one can compute the Poisson brackets for the fields as follows. We
have
d3 p
Z
1 ipx ipx
(x) = (x, t)|t=0 = (i p )((p)e (p)e )
(2)3/2 R3
p
t 2p
and
d3 p
Z
1
(x) = ((p)eipx + (p)eipx ) p
(2)3/2 R3 2p
so
Z
0 1 0 0
{(x), (x )} = ({(p), i(p0 )}ei(pxp x )
2(2)3 R3 R3
0 0
{i(p), (p0 )}ei(px+p x ) )d3 pd3 p0
Z
1 0 0 0 0
= 3
3 (p p0 )(ei(pxp x ) + ei(px+p x ) )d3 pd3 p0
2(2) R3 R3
Z
1 0 0
= 3
(eip(xx ) + eip(xx ) )d3 p
2(2) R3
= 3 (x x0 )
and
d3 p
Z
1
(x, t) = (p)eip t eipx p
(2)3/2 R3 2p
317
If, as in the non-relativistic case, we interpret as a single-particle wave-
function, Hamiltons equation of motion says
{, h} =
t
and applying this to the component of + with momentum p, we just get
multiplication by ip . The energy of such a wave-function would be p , the
eigenvalue of i t . These are called positive frequency or positive energy
wave-functions. In the case of momentum components of , the eigenvalue is
p , and one has negative frequency or negative energy wave-functions.
An expression for h in terms of momentum space field coordinates that will
have the right Poisson brackets on + , is
Z
h= p (p)(p)d3 p
R3
and this is the same expression one could have gotten by a long direct calcula-
tion.
In the non-relativistic case, the eigenvalues of the action of i t on the
2
wave-functions were non-negative ( |p| 2m ) so the single-particle states had non-
negative energy. Here we find instead eigenvalues p of both signs, so single-
particle states can have arbitrarily negative energies. This makes a physically
sensible interpretation of H1 as a space of wavefunctions describing a single
relativistic particle difficult if not impossible. We will however see in the next
section that there is a way to quantize this H1 as a phase space, getting a
sensible multi-particle theory with a stable ground state.
Starting with a vacuum state |0i, by applying creation operators one can create
arbitary positive energy multiparticle states of free relativistic particles with
318
single-particle states having the energy momentum relation
p
E(p) = p = |p|2 + m2
d3 p
Z
1 ipx ipx
b(x) = (i p )(a(p)e a (p)e )
(2)3/2 R3
p
2p
By essentially the same computation as for Poisson brackets, one can com-
pute commutators, finding
[(x),
b b(x0 )] = i 3 (x x0 ), [(x),
b b 0 )] = [b
(x b(x0 )] = 0
(x),
319
b = ( m2 )b
b = [b , iH]
t
which have as solution the following equation for the time-dependent field op-
erator:
d3 p
Z
1 ip t ipx ip t ipx
(x,
b t) = (a(p)e e + a (p)e e )
(2)3/2 R3
p
2p
where M+ J is the +i eigenspace of J, MJ the i eigenspace. The quantum
state space will be the symmetric algebra S (M+ J ) or equivalently the space of
polynomials on the dual of M+ J . The choice of J is reflected in the fact that
the Bargmann-Fock construction requires a choice of distinguished state |0iJ ,
the state corresponding to the constant polynomial function 1.
In non-relativistic quantum field theory our classical phase space was just
the space of complex wave-functions (x) at a constant time, and we could
take as J just multiplication by i, so our M+ J was just the space of complex
wave-functions itself, with M J the space of complex conjugated wave-functions.
Elements of M+ J got quantized as operators , and annihilation operators on the
b
Bargmann-Fock state space, while elements of M J were quantized as adjoint
operators b , acting as creation operators on the state space.
320
In relativistic quantum field theory, we must do something quite different.
The real scalar fields satisfying the Klein-Gordon equation form a real phase
space M to be quantized. When we complexify and look at the space M C
of complex-valued solutions, it naturally decomposes as a representation of the
Poincare group into two pieces: M+ , the complex functions on the positive
energy hyperboloid and M , the complex functions on the negative energy
hyperboloid. What we do to quantize is to take the complex structure to be
the operator J that is +i on positive energy wave-functions and i on negative
energy wavefunctions. Complexified classical fields in M+ get quantized as
annihilation operators, those in M as creation operators. Since conjugation
interchanges M+ and M , non-zero real-valued classical fields have components
in both M+ and M since they are their own conjugates. As we have seen,
they end up being a superposition of annihilation and creation operators.
For t > 0, this gives the amplitude for propagation of a particle in time t
from the origin to the point x.
321
31.5 For further reading
Pretty much every quantum field theory textbook has a treatment of the rela-
tivistic scalar field with more details than here, and significantly more physical
motivation. A good example with some detailed versions of the calculations
done here is [9]. See Folland [19], chapter 5 for a mathematically more careful
treatment of the distributional nature of the scalar field operators.
322
Chapter 32
323
324
Chapter 33
325
326
Chapter 34
Quantization of the
Electromagnetic Field: the
Photon
327
328
Chapter 35
329
330
Chapter 36
An Introduction to the
Standard Model
331
332
Chapter 37
Further Topics
333
334
Appendix A
Conventions
Ive attempted to stay close to the conventions used in the physics literature,
leading to the choices listed here. Units have been chosen so that ~ = 1.
To get from the self-adjoint operators used by physicists as generators of
symmetries, multiply by i to get a skew-adjoint operator in a unitary repre-
sentation of the Lie algebra, for example
The Lie bracket on the space of functions on phase space M is given by
the Poisson bracket, determined by
{q, p} = 1
L1 = Q2 P3 Q3 P2
335
For the spin 12 representation, the self-adjoint operators are Sj = 2j ,
the Xj = i 2j give the Lie algebra representation. Unlike the integer
spin representations, this representation does not come from the bosonic
quantization map .
Given a unitary Lie algebra representation 0 (X), the unitary group action
on states is given by
0
|i (eX )|i = e (X) |i
Instead of considering the action on states, one can consider the action on
operators by conjugation
0 0
O O() = e (X) Oe (X)
|i eiaP |i
336
Operators in the Heisenberg picture satisfy
or infinitesimally
d
O(t) = [O, iH]
dt
which is the quantization of the Poisson bracket relation in Hamiltonian
mechanics
d
f = {f, h}
dt
Conventions for special relativity.
Conventions for representations on field operators.
Conventions for anti-commuting variables. for unitary and odd super Lie
algebra actions.
337
338
Appendix B
Problems
Problem 2:
Use a similar argument to the one given in class for G = U (1) to classify
the irreducible representations of the group R under the group law of addition.
Which of these are unitary?
Problem 3:
Problem 4:
339
Consider a quantum mechanical system with state space H = C3 and Hamil-
tonian operator
0 1 0
H = 1 0 0
0 0 2
odinger equation for this system to find its state vector |(t)i
Solve the Schr
at any time t > 0, given that the state vector at t = 0 was
1
2
3
for i C.
H = Bx 1
(this is the sort of thing that occurs for a spin-1/2 system subjected to a mag-
netic field in the x-direction).
Find the eigenvectors and eigenvalues of H. What are the possible energies
that can occur in this quantum system?
If the system starts out at time t = 0 in the state
1
|(0)i =
0
340
Problem 3: By using the fact that any unitary matrix can be diagonalized
by conjugation by a unitary matrix, show that all unitary matrices can be
written as eX , for X a skew-adjoint matrix in u(n).
By contrast, show that
1 1
A=
0 1
is in the group SL(2, C), but is not of the form eX for any X sl(2, C) (this
Lie algebra is all 2 by 2 matrices with trace zero.
Hint: For 2 by 2 matrices X, one can show (this is the Cayley-Hamilton
theorem: matrices X satisfy their own characteristic equation det(1 X) = 0,
and for 2 by 2 matrices, this equation is 2 tr(X) + det(X) = 0)
X 2 tr(X)X + det(X)1 = 0
For X sl(2, C), tr(X) = 0, so here X 2 = det(X)1. Use this to show that
p
p sin( det(X))
eX = cos( det(X))1 + p X
det(X)
Problem 4: Show that the Lie algebra u(n) is not a complex vector space.
1. For both Lie algebras, show that this gives a bilinear, symmetric form,
negative definite, with the basis vectors sa in one case and La in the other
providing an orthogonal basis if one uses K(, ) as an inner product.
Here the Lie algebra adjoint representation (ad, g) gives for each X g a
linear map
ad(X) : R3 R3
and thus a 3 by 3 real matrix. This K 0 is determined by taking the trace
of the product of two such matrices. How are K and K 0 related?
341
Problem 2: Under the homomorphism given in class, what elements of
SO(3) do the quaternions i, j, k (unit length, so elements of Sp(1)) correspond
to?
and using the conjugation action of SL(2, C) on these matrices. (Hint: use
determinants).
Note that the Lorentz group has a spinor representation, but it is not unitary.
deriving a correct one. The notes refer to a place this is done with another
convention. To do this calculation, note that your result will be a 3 by 3
real matrix, with the first column given by the result of applying this linear
transformation to
1
0
0
then doing the same for the next two columns. Each of these three calculations
is done by conjugating the corresponding basis element in su(2) and expanding
the result in terms of these basis elements.
342
Show that the representation on such polynomials given in class (induced
from the SU (2) representation on C2 ) is a unitary representation with
respect to this inner product.
are orthonormal with respect to this inner product (break up the integrals
into integrals over the two complex planes, use polar coordinates).
Problem 2: Using the formulas for the Y1m (, ) and the inner product
Show that this operators commutes with the 0 (X) for all X so(3). Use
this to show that L2 has the same eigenvalue on all vectors in an irreducible
representation of so(3).
343
B.5 Problem Set 5
Problem 1: Consider the action of SU (2) on the tensor product V 1 V 1 of two
spin one-half representations. According to the Clebsch-Gordan decomposition,
this breaks up into irreducibles as V 0 V 2 .
1. Show that
1 1 0 0 1
( )
2 0 1 1 0
is a basis of the V 0 component of the tensor product, by computing first
the action of SU (2) on this vector, and then the action of su(2) on the
vector (i.e. compute the action of 0 (X) on this vector, for the tensor
produ ct representation, and X basis elements of su(2)).
2. Show that
1 1 1 1 0 0 1 0 0
, ( + ),
0 0 2 0 1 1 0 1 1
give a basis for the irreducible representation V 2 , by showing that they are
eigenvectors of 0 (s3 ) with the right eigenvalues (weights), and computing
the action of the raising and lowering operators for su(2) on these vectors.
344
Problem 2: If a quantum harmonic oscillator is in a state
1
(|0i + |1i)
2
at time t = 0, find its position-space wavefunction (q, t) for all t.
Q, P, Q2 , P 2
and
Q4
Use these to find the standard deviations in the statistical distributions of ob-
served values of q and p in these states. These are
p p
Q = hn|Q2 |ni hn|Q|ni2 , P = hn|P 2 |ni hn|P |ni2
345
For a rotation by an angle in the j k plane in 4 or 6 dimensions, what
are the elements of the Spin group that correspond to this, and how do
they act on the fermionic oscillator states? Do this by expressing things
in terms of annihilation and creation operators and their action on the
spinors, thought of as a fermionic oscillator state space.
0 , t0 ) (x, t)|0i
G(x0 , t0 , x, t) = h0|(x
for the free non-relativistic particle of mass m (for t0 > t). First do this in
momentum space, showing that
2
k
e 0 , t0 , k, t) = ei 2m (t0 t)
G(k (k 0 k)
346
Problem 4: Show that the Lie algebra so(4, C) is sl(2, C)sl(2, C). Within
this Lie algebra, identify the sub-Lie algebras of the groups Spin(4), Spin(1, 3)
and Spin(2, 2).
Problem 2:
For the real scalar field theory, find the momentum operator acting on
the quantum field theory state space, in terms of the scalar quantum
field. Show that this gives the expected expression in terms of a sum of
the number operators for each momentum mode and the corresponding
momentum.
Repeat the same calculation for the complex scalar field theory.
Problem 3: In class we showed that taking two real free scalar fields, one
could make a theory with SO(2) symmetry, and we found the charge operator Q
that gives the action of the Lie algebra of SO(2) on the state space of this theory.
Instead, consider two complex free scalar fields, and show that this theory has
a U (2) symmetry. Find the four operators that give the Lie algebra action for
this symmetry on the state space, in terms of a basis for the Lie algebra of U (2).
Note that this is the field content and symmetry of the Higgs sector of
the standard model (where the difference is that the theory is not free, but
interacting, and has a lowest energy state not invariant under the symmetry).
347
348
Bibliography
[1] Alvarez, O., Lectures on quantum mechanics and the index theorem, in
Geometry and Quantum Field Theory, Freed, D, and Uhlenbeck, K., eds.,
American Mathematical Society, 1995.
[2] Arnold, V., Mathematical Methods of Classical Mechanics, Springer-Verlag,
1978.
[3] Artin, M., Algebra, Prentice-Hall, 1991.
[4] Baym, G., Lectures on Quantum Mechanics, Benjamin, 1969.
[5] Berezin, F., and Marinov, M., Particle Spin Dynamics as the Grassmann
Variant of Classical Mechanics, Annals of Physics 104 (1972) 336-362.
[6] Berg, M., The Fourier-Analytic Proof of Quadratic Reciprocity, Wiley-
Interscience, 2000.
[7] Berndt, R., An Introduction to Symplectic Geometry, AMS, 2001.
[8] Berndt, R., Representations of Linear Groups, Vieweg, 2007.
[9] Das, A., Field Theory, a Path Integral Approach, World Scientific, 1993.
[10] De Faria, E. and De Melo, W., Mathematical Aspects of Quantum Field
Theory, Cambridge University Press, 2010.
[11] Dolgachev, I., Introduction to Physics, Lecture Notes
http://www.math.lsa.umich.edu/~idolga/lecturenotes.html
[12] Fadeev, L.D., Yakubovskii, O.A., Lectures on Quantum Mechanics for
Mathematics Students, AMS, 2009.
[13] Feynman, R., Space-time approach to non-relativistic quantum mechanics,
Reviews of Modern Physics, 20(1948) 367-387.
[14] Feynman, R. and Hibbs, A., Quantum Mechanics and Path Integrals,
McGraw-Hill, 1965.
[15] Feynman, R., Feynman Lectures on Physics,Volume 3, Addison-Wesley,
1965.
349
[16] Feynman, R., The Character of Physical Law, page 129, MIT Press, 1967.
[17] Feynman, R., Statistical Mechanics: A set of lectures, Benjamin, 1972.
[18] Folland, G., Harmonic Analysis in Phase Space, Princeton, 1989.
[19] Folland, G., Quantum Field Theory: A tourist guide for mathematicians,
AMS, 2008.
[20] Gendenshtein, L., and Krive, I., Supersymmetry in quantum mechanics,
Sov. Phys. Usp. 28 (1985) 645-666.
[21] Greiner, W. and Reinhardt, J., Field Quantization Springer, 1996.
[22] Guillemin, V., and Sternberg, S., Variations on a Theme of Kepler, AMS,
1990.
[23] Guillemin, V,., and Sternberg, S. Symplectic Techniques in Physics, Cam-
bridge University Press, 1984.
[24] Hall, B., Lie Groups, Lie Algebras, and Representations: An Elementary
Introduction, Springer-Verlag, 2003.
[25] Hall, B., An Elementary Introduction to Groups and Representations
http://arxiv.org/abs/math-ph/0005032
[26] Hall, B., An Introduction to Quantum Theory for Mathematicians, (to be
published).
[27] Hannabuss, K., An Introduction to Quantum Theory, Oxford University
Press, 1997.
[28] Haroche, S. and Ramond, J-M., Exploring the Quantum: Atoms, Cavities
and Photons, Oxford University Press, 2006.
[29] Hatfield, B., Quantum Field Theory of Point Particles and Strings,
Addison-Wesley, 1992.
[30] Kirillov, A., Lectures on the Orbit Method, AMS, 2004.
[31] Lawson, H. B., and Michelsohn, M-L., Spin Geometry, Princeton University
Press, 1989.
[32] Mumford, D., Tata Lectures on Theta III, Birkhauser, 1991.
[33] Peskin, M., and Schroeder, D., An Introduction to Quantum Field Theory,
Westview Press, 1995.
[34] Presskill, J., Quantum Computing Course Notes,
http://www.theory.caltech.edu/people/preskill/ph229/#lecture
[35] Porteous, I., Clifford Algebras and the Classical Groups Cambridge Univer-
sity Press, 1995.
350
[36] Ramond, P., Group Theory: A physicists survey, Cambridge University
Press, 2010.
[37] Rosenberg, J., A Selective History of the Stone-von Neumann Theorem,
in Operator Algebras, Quantization and Noncommutative Geometry, AMS,
2004.
[38] Schlosshauer, M., Decoherence and the Quantum-to-Classical Transition,
Springer, 2007.
[39] Schulman, L., Techniques and Applications of Path Integration, John Wiley
and Sons, 1981.
[40] Shankar, R., Principles of Quantum Mechanics, 2nd Ed., Springer, 1994.
[41] Singer, S., Linearity, Symmetry, and Prediction in the Hydrogen Atom,
Springer-Verlag, 2005.
[42] Sternberg, S., Group Theory and Physics, Cambridge University Press,
1994.
[43] Stillwell, J., Naive Lie Theory Springer-Verlag, 2010.
http://www.springerlink.com/content/978-0-387-78214-0
[44] Strichartz, R., A Guide to Distribution Theory and Fourier Transforms,
World Scientific, 2003.
[45] Takhtajan, L., Quantum Mechanics for Mathematicians, AMS, 2008.
[46] Taylor, M., Noncommutative Harmonic Analysis, AMS, 1986. Benjamin,
1969.
[47] Teleman, C., Representation Theory Course Notes,
math.berkeley.edu/~teleman/math/RepThry.pdf
[48] Tung, W-K., Group Theory in Physics, World Scientific Publishing 1985.
[49] Warner, F., Foundations of Differentiable Manifolds and Lie Groups
Springer-Verlag, 1983.
[50] Weil, A., Sur certain groupes doperateurs unitaires, Acta. Math. 111(1964)
143-211.
[51] Weinberg, S., The Quantum Theory of Fields I, Cambridge University
Press, 1995.
[52] Weyl, H., The Theory of Groups and Quantum Mechanics, Dover, 1950.
(First edition in German, 1929)
[53] Wigner, E., The Unreasonable Effectiveness of Mathematics in the Natural
Sciences, Communications in Pure and Applied Mathematics, vol. 13, No.
I (February 1960).
351
[54] Witten, E., Supersymmetry and Morse Theory, J. Differential Geometry,
17(1982) 661-692.
[55] Woodhouse, N. M. J., Special Relativity, Springer-Verlag, 2002.
[56] Zee, A., Quantum Field Theory in a Nutshell, Princeton University Press,
2003.
[57] Zinn-Justin, J., Path Integrals in Quantum Mechanics, Oxford University
Press, 2005.
[58] Zurek, W., Decoherence and the Transition from Quantum to Classical
Revisited, http://arxiv.org/abs/quant-ph/0306072
352