3 Teoria de Grupos Arfken

Group Theory
Disciplined judgment, about what is neat

and symmetrical and elegant, has time and
time again proved an excellent guide to
how nature works.
Murray Gell-Mann
4.1 Introduction to 6roup Theory
In classical mechanics the aymmetry of a physical system leads to conaerva-
tion Iawa. Conservation of angular momentum (L) is a direct consequence of
rotational symmetry, which means invariance of some physical observables
(suchasL
2
) or geometrical quantities (suchas lengthof a vector or distancebe-
tween points) under spatial rotations. In the nrst third of this century, Wigner
and others realized that invariance was a key concept in understanding the
new quantum phenomena and in developing appropriate theories. For exam-
ple, Noether's theorem establishes a conserved current from an invariance
of the Lagrangian of a neld theory. Thus, in quantum mechanics the concept
of angular momentum and spin has become even more central. Its generaliza-
tions, iaoapin in nuclear physics and the Bavor aymmetry in particle physics,
are indispensable tools in building and solving theories. Generalizations of the
concept of gauge invariance of classical electrodynamics to the isospin sym-
metry lead to the electroweak gauge theory.
In each case the set of these symmetry operations forms a group, a math-
ematical concept we shall soon denne. Group theory is the mathematical tool
to treat invariants and symmetries. It brings unincation and formalization of
229
Create PDF files without this message by purchasing novaPDF printer (http://www.novapdf.com)
230 Chatcr 1 6roup Theory
principles such as spatial reections, or parity, angular momentum, and geom-
etry that are widely used by physicists.
In geometry the fundamental role of group theory was recognized more
than a century ago by mathematicians (e.g., Felix Klein's Erlanger Programm).
In Euclidean geometry the distance between two points, the scalar product of
two vectors or metric, does not change under rotations or translations. These
symmetries are characteristic of this geometry. In special relativity the metric,
or scalar product of four-vectors, differs from that of Euclidean geometry in
that it is no longer positive dennite and is invariant under Lorentz transforma-
tions.
For a crystal, the symmetry group contains only a nnite number of rotations
at discrete values of angles or reections. The theory of such diacreteor Bnite
groups, developed originally as a branch of pure mathematics, is now a useful
tool for the development of crystallography and condensed matter physics.
When the rotations depend on continuously varying angles (e.g., the Euler
angles of Section J.J) the rotation groups have an innnite number of elements.
Such continuous (or Lie)
1
groupa are the topic of this chapter.
Befinition of 6roup
A group G may be denned as a set of objects or, in physics usually, symmetry
operations (such as rotations or Lorentz transformations), called the elements
q of G, that may be combined or multiplied" to form a well-denned product
in G that satisnes the following conditions:
1. If a and b are any two elements of G, then the product ab is also an element
of G. In terms of symmetry operations, b is applied to the physical system
beforeainthe product, andtheproduct ab is equivalent toa singlesymmetry
operation in G Multiplication associates (or maps) an element ab of G with
the pair (a, b) of elements of G, this property is known as cIoaure under
multiplication.
z. This multiplication is aaaociative, (ab)c a(bc).
J. There is a unit or identity element
z
1 in G such that 1a a1 a for every
element a in G.
4. G must contain an inverae or reciprocaI of every element a of G, labeled
a
1
such that aa
1
a
1
a 1.
Note that the unit is unique, as is the inverse. The inverse of 1 is 1 because
1a a1 a for a 1 yields 1 1 1 If a second unit 1 existed we would
have 11 1 1 1 and 1 1 11 1 Comparing we see that 1 1 Similarly,
if a second inverse a
1
existed we would have a
1
a aa
1
1 aa
1
Multiplying by a
1
, we get a
1
a
1
1
After the Norwegian mathematician Sophus Lie.
z
Following E. Wigner, the unit element of a group is often labeled E, from the German Einheit
(i.e., unit) or just 1 or ! for identity.
1.l Introduction to 6roup Theory 231
EXAHPIE 4.1.1
Coordinate Rotation An example of a group is the set of counterclockwise
coordinate rotations
( )
a
cos sin
sin cos
a
(4.1)
through an angle of the a-coordinate system to a new orientation (see
Fig. z.z0). The product of two rotations is denned by rotating nrst by the angle
z
and then by
1
. According to Eqs. (J.J) and (J.J7), the product of the
orthogonal z z matrices, (
1
) (
z
), describes the product of two rotations
cos
1
sin
1
sin
1
cos
1
cos
z
sin
z
sin
z
cos
z
cos(
1 z
) sin(
1 z
)
sin(
1 z
) cos(
1 z
)
,
(4.z)
using the addition formulas for the trigonometric functions. The product is
clearly a rotation represented by the orthogonal matrix with angle
1 z
. The
product is the associative matrix multiplication. It is commutative or AbeIian
because the order in which these rotations are performed does not matter. The
inverse of the rotation with angle is that with angle . The unit corresponds
to the angle 0. The group's name is SO(z), which stands for apeciaI
orthogonaI rotationa in two dimenaiona, where special means the z z
rotation matrices have determinant 1, and the angle varies continuously
from 0 to z , so that the group has innnitely many elements. The angle is the
group parameter.
A aubgroup G of a group G is a group consisting of elements of G so that the
product of any of its elements is again in the subgroup G , that is, G is cIoaed
under the multiplication of G. For example, the unit 1 of G always forms a
subgroup of G, and the unity with angle 0 and the rotation with
about some axis form a nnite subgroup of the group of rotations about that
axis.
If qq q
1
is an element of G for any q of G and q of G , then G is called
an invariant aubgroup of G. If the group elements are matrices, then the
element qq q
1
corresponds to a similarity transformation [see Eq. (J.10S)] of
q in G by an element q of G (discussed in Chapter J). Of course, the unit 1
of G always forms an invariant subgroup of G because q1q
1
1 When an
element q of G lies outside the subgroup G , then qq q
1
may also lie outside
G Let us illustrate this by three-dimensional rotations.
EXAHPIE 4.1.2
SimiIarity TranaIormation Rotations of the coordinates through a nnite
angle counterclockwise about the :-axis in three-dimensional space are de-
scribed as
a
:
:
( )
a
:
cos sin 0
sin cos 0
0 0 1
a
:
, (4.J)
which form a group by a generalization of Eq. (4.z) to our special J J matri-
ces that keep their special form on multiplication. Moreover, the order of the
rotations in a product does not matter, just like in Eq. (4.z), so that the group
is Abelian. A general rotation about the a-axis is given by the matrix
a
( )
1 0 0
0 cos sin
0 sin cos
Now consider a rotation E
a
by 90 about the a-axis. Its matrix is
a
1 0 0
0 0 1
0 1 0
,
and its inverse is
1
a
1 0 0
0 0 1
0 1 0
,
corresponding to the angle 90 This can be checked by multiplying them:
a
1
a
1 Then
a :
( )
1
a
1 0 0
0 0 1
0 1 0
1 0 0
0 cos sin
0 sin cos
1 0 0
0 0 1
0 1 0
cos 0 sin
0 1 0
sin 0 cos
,
which is a rotation by about the -axis and no longer a rotation about the
:-axis so that this element lies outside the subgroup of rotations about the :-
axis, and this subgroup is not an invariant subgroup. The set of these elements
for all form a group called conjugate to the group of rotations about the
:-axis.
Orthogonal n n matrices form the group O(n), and they form SO(n) if their
determinants are 1 (S stands for special" and O for orthogonal"), with
elements denoted by

Because

(see Section J.J for orthogonal

J J matrices that preserve the lengths of vectors and distances between
points in three-dimensional Euclidean space), we see that the product
1 z

z

1
1
z
1
1
(
1 z
)
1
is also an orthogonal matrix in O(n) or SO(n). The inverse is the transpose
(orthogonal) matrix. The unit of the group is 1
n
. A real orthogonal n nmatrix
has n(n 1) z independent parameters. For n z, there is only one parameter:
one angle in Eq. (4.1). For n J, there are three independent parameters:
for example, the three Euler angles of Section J.J, and SO(J) is related to
rotations in three dimensions, just as SO(z) is in two dimensions. Because
O(n) contains orthogonal transformations with determinant 1, this group
includes reections of coordinates or parity inversions. Likewise, unitary n n
matrices form the group U(n), and they form SU(n) if their determinants are
1. Because

1
(see Section J.4 for unitary matrices, which preserve

the norm of vectors with complex components and distances between points
in n-dimensional complex space), we see that
(
1 z
)
z 1
1
z
1
1
(
1 z
)
1
so that the product is unitary and an element of U(n) or SU(n). Each unitary
matrix obviously has an inverse, which again is unitary. Orthogonal matrices
are unitary matrices that are real so that SO(n) forms a subgroup of SU(n), as
does O(n) of U(n).
EXAHPIE 4.1.8
SimpIe Unitary Groupa The phase factors c
, with real angle , of quan-

tummechanical wave functions forma group under multiplication of complex
numbers because the phase angles add on multiplying c
1
c
z
c
(
1 z
)
More-
over, (c
) c

and c
c

1 show unitarity and inverse, and 0 gives
the unit. This group (of unitary 1 1 matrices) has one continuous real param-
eter and is therefore called U(1). The two elements 1 form a nnite (unitary)
subgroup, and the four elements 1, form another subgroup.
A nnite unitary group of z z matrices is denned by the two-dimensional
unit matrix and one of the three Pauli matrices,

, using matrix multiplication.
Because
z
1
z
, the inverse
1

and 1
1
z
1
z
When a potential has spherical symmetry we choose polar coordinates, and
the associated group of transformations is a rotation group. For problems with
spin (or other internal properties such as isospin or avor), unitary groups play
a similar role. Therefore, in the following we discuss only the rotation groups
SO(n) and the unitary group SU(z) among the classical Lie groups.
Hioqrohicol Doto
Lie, Sophua. Lie, who was born in 1S4z in Nordfjordeid, Norway, and died
in 1S99 in Kristiana (now Oslo), started his analysis of continuous groups of
transformations in Paris and continued it throughout his life.
Wigner, Eugen PauI. Wigner, who was born in 190z in Budapest, Hungary,
and died in 199b in Princeton, New 1ersey, studied in Berlin, moved to the
United States in the 19J0s, and received the Nobel prize in 19J for his con-
tributions to nuclear theory and applications of fundamental principles of
symmetry, such as the charge independence of nuclear forces. He developed
the unitary representations of the Lorentz group.
Bomomorphism and Isomorphism
There may be a correspondence between the elements of two groups: one-to-
one, two-to-one, or many-to-one. If this correspondence preserves the group
multiplication, the two groups are homomorphic. If the correspondence is
one-to-one, still preserving the group multiplication,
J
then the groups are iao-
morphic. An example is the rotations of the coordinates through an angle
counterclockwise about the :-axis in three-dimensional space described
by Eq. (4.J). If we identify every rotation with its orthogonal z z subma-
trix, we have a one-to-one map that preserves the group multiplication ac-
cording to Eq. (4.z), an isomorphism or IaithIuI repreaentation. Of course,
we can also identify each rotation with its J J matrix in Eq. (4.J), which
is also an isomorphism. Therefore, matrix representations are by no means
unique, but each isomorphism is a group multiplication preserving faithful
map.
Hatrix Representations: Reducible and Irreducible
The isomorphisms betweenrotations and groups of matrices just discussedare
examples of matrix representations of a group (of rotations). Such represen-
tations of group elements by matrices are a very powerful technique and have
been almost universally adopted by physicists. The use of matrices imposes
no signincant restriction. It can be shown that the elements of any nnite group
and of the continuous groups of Sections 4.z4.4 may be represented by ma-
trices. In each case, matrix multiplication follows from group multiplication.
The same group can be represented by matrices of different rank. Examples
are the rotations about the :-axis described in Eqs. (4.1) and (4.J).
To illustrate how matrix representations arise from a symmetry, consider
the time-independent Schr odinger equation (or some other eigenvalue equa-
tion, such as

!

for the principal moments of inertia of a rigid body in
classical mechanics, for example)
H E (4.4)
Let us assume that Eq. (4.4) stays invariant under a group G of transformations
in G For example, for a spherically symmetric Hamiltonian H the group G
would be SO(J) Consequently, H is the same in a rotated coordinate sys-
tem, where the Hamiltonian is given by the similarity transformation H
1
according to Eq. (J.10S) of Chapter J. Hence,
H
1
H, or H H , (4.b)
that is, rotations" from G and H commute. Now take a solution of Eq. (4.4)
and rotate" it with , an element from G: . Then has the aame
J
Suppose the elements of one group are labeled q
, and the elements of a second group are labeled

h
. Then q
is a one-to-one correspondence for all values of . If q
q
]
q
k
and h
h
]
h
k
,
then q
k
and h
k
must be the corresponding group elements.
vaIue oI energy E because multiplying Eq. (4.4) by and using Eq. (4.b)
yields
H E( ) H( ) (4.)
In other words, all rotated solutions are degenerate in energy or form a
vector space that physicists call a muItipIet. For example, the spin-up and
-down states of an electron form a doublet, and the states with projection
quantumnumbers u |, | 1, , 0, 1, , | of orbital angular momentum
| form a multiplet with z| 1 basis states. (The magnetic neld in the Zeeman
effect lifts the degeneracy of these states and breaks the rotational symmetry
because the magnetic neld points in some direction.)
Let us now assume that this vector space V of transformed solutions has
a nnite dimension n. Let
1
,
z
, ,
n
be a basis. Since
]
is a member of
the multiplet, we can expand it in terms of its basis,
]
k
r
]k k
(4.7)
Thus, with each transformation in G we can associate a matrix (r
]k
),
and this map ( ) is denned as a repreaentation of G. If we can take any
element of V and by rotating with all elements of G transform it into aII
other elements of V , then the representation is irreducibIe. The spin-up and
-down states of the hydrogen ground state form an irreducible representation
of SU(z) (they can be rotated into each other), but the zs state and zj states of
principal quantum number n z of the hydrogen atom have the same energy
(i.e., they are degenerate) and form a reducible representation because the zs
state cannot be rotated into the zj states and vice versa (angular momentum
is conserved under rotations). With this case in mind, we see that if all ele-
ments of V are not reached, then V splits into a direct sum of two or more
vector subspaces (see Chapter J), V V
1
V
z
, which are mapped into
themselves by rotating their elements (e.g., zs zs, zp zp). The direct aum
of two vector spaces is spanned by the basis vectors of both vector spaces. In
this case, the representation is called reducibIe. Then we can nnd a unitary
matrix so that
(r
]k
)
1
z
(4.S)
for aII of G and aII matrices (r
]k
). Here,
1
,
z
, , are matrices of lower
dimension than (r
]k
) that are lined up along the diagonal and the are matrices
made up of zeros, that is, is bIocI-diagonaI.
For example, for the zs states of hydrogen,
1
would be a unitary z z
matrix
1
a b
c d
If the electron spin is ignored,
1
1 of the zs state would be a one-dimensional
unit matrix. We maysaythat the representation has beendecomposed into
1 z
along with V V
1
V
z
, where the smaller

are irreducible.
In any representation, each irreducible representation may be repeated a nnite
number of times, and some do not appear at all.
The irreducible representations play a role in group theory that is anal-
ogous to that of the unit vectors of vector analysis. They are the simplest
representationsall others can be built from them.
Often, groups occur in physics as sets of transformations of coordinates or

symmetry transformations that leave a partial or ordinary differential equation
unchanged, or group elements denne changes of bases of some vector space.
Therefore, matrix representations of groups are common in physics. The most
important application in quantum mechanics is based on the fact that the
degenerate states of multiplets in a spectrum reect the symmetry group of
the Hamiltonian. Multiplets correspond to representations of the symmetry
group of a Hamiltonian.
EXERCISES
4.1.1 Show that an n n orthogonal matrix has n(n 1) z independent pa-
rameters.
Hnl. The orthogonality condition, Eq. (J.79), provides constraints.
4.1.2 The special linear group SL(z,C) consists of all z z matrices (with
complex elements) having a determinant of 1. Show that such matri-
ces form a group.
Nolc. The SL(z,C) group can be related to the full Lorentz group in
Section 4.4, much as the SU(z) group is related to SO(J).
4.1.3 Show that rotations about the :-axis form a subgroup of SO(J). Is it an
invariant subgroup?
4.1.4 Show that if E, S, T are elements of a group G so that ES T, and
E (r
k
), S (s
k
) is a representation according to Eq. (4.7), then
(r
k
)(s
k
) l
k
n
r
n
s
nk
,
that is, group multiplication translates into matrix multiplication for any
group representation.
4.1.5 A subgroup H of G has elements h
. Let a be a nxed element of the

original group G and not a member of H. The transform
ah
a
1
, 1, z,
generates a conjugate aubgroup aHa
1
. Show that this conjugate sub-
group satisnes each of the four group postulates and therefore is a group.
1.? 6enerators of fontinuous 6roups 23?
4.2 6enerators of fontinuous 6roups
Acharacteristic of continuous groups known as Lie groups is that the elements
are functions of parameters having derivatives of arbitrary orders suchas cos
and sin in Eq. (4.1). This unlimited differentiability of the functions allows us
to develop the concept of generator and reduce the study of the whole group
to a study of the group elements in the neighborhood of the identity element.
Lie's essential idea was to study elements in a group G that are innnites-
imally close to the unity of G. Let us consider the SO(z) group as a simple
example. The z z rotation matrices in Eq. (4.1) can be written in exponential
form using the Euler identity [Eq. (J.1SJ)] as
( )
cos sin
sin cos
1
z
cos
z
sin exp(
z
) (4.9)
From the exponential form it is obvious that multiplication of these matrices
is equivalent to addition of the arguments
(
z
) (
1
) exp(
z z
) exp(
z 1
) exp(
z
(
1 z
)) (
1 z
)
Of course, the rotations close to 1 have small angle 0.
This suggests that we look for an exponential representation
exp( ), 0, (4.10)
for group elements in G close to the unity 1. The operators of the innnites-
imal transformations are called generators of G. Therefore,
z
in Eq. (4.9)
is the generator of rotations about the :-axis. Thus, for SO(z) as denned by
Eq. (4.1) there is only one linearly independent generator,
z
. In SO(J) there is
a generator for rotations about each axis. These generators forma linear space
because multiplication of the elements of the group translates into addition
of generators S its dimension is denned as the order of G. Therefore, the or-
der of SO(z) is 1, and it is J for SO(J). One can also show that the commutator
of two generators is again a generator
[S
]
, S
k
]
|
c
|
]k
S
|
,
where the c's are denned as the atructure conatanta of the group. The vector
space of generators can be endowed with a multiplication by denning the
commutator as the product of two generators. This way the vector space of
generators becomes an algebra, the so-called Lie aIgebra.
Because does not change the volumethat is, det( ) 1we use Eq.
(J 1S4) to see that
det( ) exp(trace(ln )) exp( trace( )) 1,
which implies that generatora are traceIeaa:
tr( ) 0 (4.11)
This is the case for the rotation groups SO(n) and unitary groups SU(n).
23S Chatcr 1 6roup Theory
If of G in Eq. (4.10) is unitary, then S S is Hermitian, which is also the
case for SO(n) and SU(n). Hence the in Eq. (4.10).
Returning to Eq. (4.b), we now emphasize the most important result from
group theory. The inverse of is just
1
exp( ). We expand H
E
accord-
ing to the BakerHausdorff formula [Eq. (J.1Sb)], taking the Hamiltonian H
and to be operators or matrices we see that
H H
E
exp( )H exp( ) H [ , H]
1
z
z
[ [ , H]] (4.1z)
We subtract H from Eq. (4.1z), divide by , and let 0. Then Eq. (4.1z)
implies that for any rotation close to 1 in G the commutator
[ , H] 0 (4.1J)
We see that ia a conatant oI the motion: A aymmetry oI the ayatem
haa Ied to a conaervation Iaw. If and H are Hermitian matrices, Eq. (4.1J)
states that and H can be simultaneously diagonalized, that is, the eigenvalues
of are constants of the motion. If and H are differential operators like the
Hamiltonian and orbital angular momentum L
z
, L
:
in quantum mechanics,
then Eq. (4.1J) states that and H have common eigenfunctions, and that the
degenerate eigenvalues of H can be distinguished by the eigenvalues of the
generators . Theseeigenfunctionsandeigenvalues, s, aresolutions of separate
differential equations,
s
s
s
, so that group theory (i.e., symmetries) leads
to a separation of variables for a partial differential equation that is invariant
under the transformations of the group. For examples, see the separation of
variables method for partial differential equations in Section S.9 and special
functions in Chapter 11. This is by far the most important application of group
theory in quantum mechanics.
In the following sections, we study orthogonal and unitary groups as ex-
amples to understand better the general concepts of this section.
Rotation 6roups S(2) and S(8)
For SO(z) as denned by Eq. (4.1) there is only one linearly independent gener-
ator,
z
, and the order of SO(z) is 1. We get
z
from Eq. (4.9) by differentiation
at the unity of SO(z) (i.e., 0),
d ( )
d
0
sin cos
cos sin
0
0 1
1 0
z
(4.14)
For the rotations
:
( ) about the :-axis described by J J matrices in
Eq. (4.J), the generator is given by
d
:
( )
d
0
:
0 0
0 0
0 0 0
(4.1b)
1.? 6enerators of fontinuous 6roups 239
Here,
z
is recognized in the upper left-hand corner of
:
. The rotation
:
( )
through an innnitesimal angle may then be expanded near the unity ( 0)
as
:
( ) 1
J

:
, (4.1)
with terms of order ( )
z
and higher omitted. A nnite rotation ( ) may be
compounded of successive innnitesimal rotations
:
(
1 z
) (1
J

1 :
)(1
J

z :
) (4.17)
Let N for N rotations, with N . Then,
:
( ) lim
N
[
:
( N)]
N
lim
N
[1
J
( N)
:
]
N
exp(
:
), (4.1S)
which is another way of getting Eq. (4.10). This form identines
:
as the gen-
erator of the group
:
, an Abelian subgroup of SO(J), the group of rotations in
three dimensions with determinant 1. Each J J matrix
:
( ) is orthogonal
(hence unitary), and trace(
:
) 0 in accordance with Eq. (4.11).
By differentiation of the coordinate rotations
a
( )
1 0 0
0 cos sin
0 sin cos
,

( )
cos 0 sin
0 1 0
sin 0 cos
, (4.19)
we get the generators
a
0 0 0
0 0
0 0
,

0 0
0 0 0
0 0
(4.z0)
of
a
(

), the subgroup of rotations about the a- (-)axis.
Rotation of Functions and rbital Angular Homentum
In the foregoing discussion the group elements are matrices that rotate the
coordinates. Any physical system being described is held nxed. Now let us
hold the coordinates nxed and rotate a function (a, , :) relative to our nxed
coordinates. With to rotate the coordinates,
, (4.z1)
we denne on by
(a, , :) (a, , :) ( ) (4.zz)
In words, operates on the function , creating a new Iunction that
is numerically equal to ( ), where are the coordinates rotated by . If
rotates the coordinates counterclockwise, the effect of is to rotate the
pattern of the function counterclockwise, as shown in Fig. z.z0.
Returning to Eqs. (4.J), (4.1b), and (4.z0) consider an innnitesimal rotation
again, . Then, using
:
[Eq. (4.J)], we obtain
:
( ) (a, , :) (a , a , :) (4.zJ)
The right side may be expanded to nrst order in to give
:
( ) (a, , :) (a, , :) a a O( )
z
(1 L
:
) (a, , :), (4.z4)
where the differential expression in curly brackets is the orbital angular mo-
mentumL
:
(Exercise 1.7.1z). This shows how the orbital angular momentum
operator arises as a generator. Since a rotation of nrst and then about the
:-axis is given by
:
( )
:
( )
:
( ) (1 L
:
)
:
( ) , (4.zb)
we have (as an operator equation)
d
:
d
lim
0
:
( )
:
( )
L
: :
( ) (4.z)
In this form, Eq. (4.z) integrates immediately to
:
( ) exp( L
:
) (4.z7)
Note that
:
( ) rotates functions (counterclockwise) relative to nxed coordi-
nates [so Eqs. (4.z7) and (4.10) are similar but not the same] and that L
:
is the
:-component of the orbital angular momentumL. The constant of integration
is nxed by the boundary condition
:
(0) 1.
If we recognize that the operator
L
:
(a, , :)
:
a
:
, (4.zS)
it becomes clear why L
a
, L
, and L
:
satisfy the same commutation relation
[L
, L
]
]
]k
L
k
(4.z9)
as
a
,

, and
:
and yield the structure constants
]k
of SO(J).
Special tnitary 6roup St(2)
Since unitary z z matrices transform complex two-dimensional vectors pre-
serving their norm, they represent the most general transformations of (a basis
in the Hilbert space of) spin
1
z
wave functions in nonrelativistic quantum me-
chanics. The basis states of this system are conventionally chosen to be
1
0
,
0
1
,
corresponding to spin
1
z
up and down states, respectively. We can showthat the
apeciaI unitary group SU(z) of such unitary z z matrices with determinant
1.? 6enerators of fontinuous 6roups 241
1 has the three Pauli matrices

as generators. Therefore, we expect SU(z)
to be of order J and to depend on three real continuous parameters , , ,
which are often called the CayIey~KIein parameters and are essentially the
SU(z) analog of Euler angles. We start with the observation that orthogonal
z z matrices [Eq. (4.1)] are real unitary matrices, so they form a subgroup of
SU(z). We also see that
c
0
0 c

is unitary for real angle with determinant 1. Therefore, these simple and
manifestly unitary matrices form another subgroup of SU(z) from which we
can obtain all elements of SU(z)that is, the general z z unitary matrix
of determinant 1 For a two-component spin
1
z
wave function of quantum
mechanics, this diagonal unitary matrix corresponds to multiplication of the
spin-up wave function with a phase factor c
and the spin-down component

with the inverse phase factor. Using the real angle instead of for the rotation
matrix and then multiplying by the diagonal unitary matrices, we construct a
z z unitary matrix that depends on three parameters and is clearly a more
general element of SU(z):
c
0
0 c

cos sin
sin cos
c
0
0 c

c
cos c
sin
c

sin c

cos
c
0
0 c

c
( )
cos c
( )
sin
c
( )
sin c
( )
cos
Denning , , we have in fact constructed the general element
of SU(z):
z
( , , )
c
cos c
sin
c

sin c

cos
a b
b a
, (4.J0)
where a
z
b
z
1 It is easy to check that the determinant det(
z
) 1 by
the product theorem of Section J.z and that
z
z
1
z
z
holds provided
, , are real numbers.
To get the generators, we differentiate
z
0, 0
1 0
0 1
J
, (4.J1a)
z
0, 0
0
0
z
(4.J1b)
To avoid a factor 1 sin for 0 upon differentiating with respect to , we
use instead the right-hand side of Eq. (4.J0) for
z
for pure imaginary b
with 0 Differentiating such a
z
, we get the third generator
1
z
1
z
0
1
z
1
z
0
0 1
1 0
1
(4.J1c)
Being generators, these Pauli matrices are all traceless and Hermitian. Acorre-
spondence with the physical world is obtained if we scale the SU(z) generators
sothat they yieldthe angular momentumcommutators. Withthe Pauli matrices
as generators, the elements
1
,
z
,
J
of SU(z) may be generated by
1
exp(
1
z),
z
exp(
z
z),
J
exp(
J
z) (4.Jz)
The three parameters are real, and we interpret them as angles. The extra
scale factor 1 z is present in the exponents because

z satisfy the same
commutation relations,
4
[

,
]
]
]k k
, (4.JJ)
as the orbital angular momentum in Eq. (4.z9).
Using the angular momentum matrix
J
, we have as the corresponding
rotation operator
:
( ) exp(
J
z) in two-dimensional (complex wave
function) space, analogous to Eq. (4.J) that gives the operator for rotating the
Cartesian coordinates in the three-space.
For rotating the two-component vector wave function (spinor) or a spin
1
z
particle relative to nxed coordinates, the rotation operator is
:
( )
exp(
J
z) according to Eq. (4.z7).
Using in Eq. (4.Jz) the Euler identity [Eq. (J.1SJ)] we obtain
]
cos( z)
]
sin( z), (4.J4)
etc. Here, the parameter appears as an angle, the coefncient of an angular
momentummatrixlike in Eq. (4.z7). With this identincation of the exponen-
tials, the general form of the SU(z) matrix (for rotating functions rather than
coordinates) may be written as
( , , ) exp(
J
z) exp(
z
z) exp(
J
z), (4.Jb)
where the SU(z) Euler angles , , differ from the , , used in the denni-
tion of the CayleyKlein parameters , , by a factor of 1 z Further discus-
sion of the relation between SO(J) and orbital angular momentum appears in
Sections 4.J and 11.7.
The orbital angular momentum operators are the generators of the rotation
groupSO(J) and(1 z) thePauli spinmatricesarethosefor SU(z), thesymmetry
group of the Schr odinger equation for a spin
1
z
particle such as the electron.
Generators obey commutation relations characteristic of the group.
4
Thestructureconstants (
]k
)leadtotheSU(z) representations of dimensionzJ 1for generators
of dimension zJ 1, J 0, 1 z, 1, . The integraI 1 cases also lead to the representations of
SO(J).
1.3 rbital Angular Homentum 243
EXERCISES
4.2.1 (i) Showthat thePauli matricesare the generators of SU(z) without using
the parameterization of the general unitary z z matrix in Eq. (4.J0).
Hnl. Exploit the general properties of generators.
4.2.2 Prove that the general form of a z z unitary, unimodular matrix is
a b
b a
,
with a a b b 1. Based on this result, derive the parameterization of
Eq. (4.J0).
4.2.3 A tranaIation operator T(a) converts (a) to (a a),
T(a) (a) (a a)
In terms of the (quantum mechanical) linear momentum operator j
a
d da, show that T(a) exp(aj
a
) (i.e., j
a
is the generator of transla-
tions).
Hnl. Expand (a a) as a Taylor series.
4.2.4 Consider the general SU(z) element Eq. (4.J0) to be built up of three
Euler rotations: (i) a rotation of a z about the :-axis, (ii) a rotation of
b z about the newa-axis, and (iii) a rotation of c z about the new:-axis.
(All rotations are counterclockwise.) Using the Pauli generators, show
that these rotation angles are determined by
a z z
b z
c z z
Nolc. The angles a and b here are not the a and b of Eq. (4.J0).
4.2.5 We know that any z z matrix can be expanded as a
0
1 a ,
where 1 is the two-dimensional unit matrix. Determine a
0
and a for the
general SU(z) matrix in Eq. (4.J0).
4.2.6 Rotate a nonrelativistic wave function

( , ) of spin
1
z
about the
:-axis by a small angle d . Find the corresponding generator.
4.8 rbital Angular Homentum
The classical concept of angular momentum L
class
r p is presented in
Section 1.J to introduce the cross product. Following the usual Schr odinger
representation of quantum mechanics, the classical linear momentum p is
replaced by the operator . The quantum mechanical orbital angular
momentum operator becomes
b
L
M
r (4.J)
This is used repeatedly in Sections 1.7, 1.S, and z.4 to illustrate vector differ-
ential operators. From Exercise 1.7.1J the angular momentum components
satisfy the commutation relations
[L
, L
]
]
]k
L
k
(4.J7)
The
]k
is the LeviCivita symbol of Section z.9. A summation over the index
k is understood.
The differential operator corresponding to the square of the angular mo-
mentum
L
z
L L L
z
a
L
z
L
z
:
(4.JS)
may be determined from
L L (r p) (r p), (4.J9)
which is the subject of Exercises 1.S. and z.b.1z(c). Since L
z
is invariant under
rotations, [L
z
, L
] 0, which can also be verined directly.

Equation (4.J7) presents the basic commutation relations of the com-
ponents of the quantum mechanical angular momentum. Indeed, within the
framework of quantum mechanics and group theory, these commutation rela-
tions denne an angular momentum operator. We shall use them now to con-
struct the angular momentum eigenstates and nnd the eigenvalues. For the
orbital angular momentum, these are the spherical harmonics of Section 11.b.
Iadder perator Approach
Let us start with a general approach, where the angular momentum J we
consider may represent an orbital angular momentum L, a spin z, a total
angular momentum L z, etc. Thus,
1. J is an Hermitian operator whose components satisfy the commutation
relations
[J
, J
]
]
]k
J
k
, [J
z
, J
] 0 (4.40)
Otherwise J is arbitrary. (See Exercise 4.J.1.)
z. u is a normalized eigenfunction (or eigenvector) of J
:
with eigenvalue
uand an eigenfunction
of J
z
,
J
:
u u u , J
z
u u (4.41)
b
For simplicity, h is set equal to 1. This means that the angular momentum is measured in units
of h.
That u is an eigenfunction of both J

:
and J
z
follows from [J
:
, J
z
] 0 in Eq. (4.40). Note also
that eigenvalues are in small letters, whereas operators are in capitals.
1.3 rbital Angular Homentum 245
We shall show that ]( ] 1), with ] either integral or half-integral,
and then nnd other properties of the u . The treatment will illustrate the
generality and power of operator techniques, particularly the use of ladder
operators.
7
The Iadder operatora are denned as
J J
a
J
, J J
a
J
(4.4z)
In terms of these operators J
z
may be rewritten as
J
z
1
z
(J J J J ) J
z
:
(4.4J)
From the commutation relations, Eq. (4.40), we nnd
[J
:
, J ] J , [J
:
, J ] J , [J , J ] zJ
:
(4.44)
Since J commutes with J
z
(Exercise 4.J.1),
J
z
(J u ) J (J
z
u ) (J u ) (4.4b)
Therefore, J u is still aneigenfunction of J
z
with eigenvalue , andsimilarly
for J u . However, from Eq. (4.44)
J
:
J J (J
:
1), (4.4)
or
J
:
(J u ) J (J
:
1) u (u 1)J u (4.47)
Therefore, J u is still an eigenfunction of J
:
but with the eigenvalue u 1.
J has raised the eigenvalue by 1 and so is called a raiaing operator. Similarly,
J lowers the eigenvalue by 1, it is a Iowering operator.
Taking the diagonal matrix element (also called expectation value) and
using J
a
J
a
, J
, we get
uJ
z
J
z
:
u u J
z
a
J
z
u J
a
u
z
J
u
z
0
and see that u
z
0, so uis bounded, and 0. Let ] be the Iargeat
u value. Then J ] 0, which implies J J ] 0. Hence, combining
Eqs. (4.4J) and (4.44) to get
J
z
J J J
:
(J
:
1), (4.4S)
we nnd from Eq. (4.4S) that
0 J J ] (J
z
J
z
:
J
:
) ] ( ]
z
]) ]
Therefore,
]( ] 1) 0 (4.49)
7
Ladder operators can be developed for other mathematical functions. Compare Section 1J.1 for
Hermite polynomials.
We now relabel the states u ]u . Similarly, let ] be the amaIIeat u
value. Then J ] ] 0. From
J
z
J J J
:
(J
:
1), (4.b0)
we see that
0 J J ] ] (J
z
J
:
J
z
:
) ] ] ( ] ]
z
) ] ] (4.b1)
Hence,
]( ] 1) ] ( ] 1) ( ])( ] 1)
Therefore, ] ], and ] 0 because ] ] Moreover, uruns in integer
atepa from ] to ],
] u ], (4.bz)
so that z ] must be a positive integer. Thus, ] is either an integer or half of an
odd integer.
Starting from ] ] and applying J repeatedly, we reach all other states
]u . Hence, the ]u form an irreducible representation, u varies and ] is
nxed.
Then using Eqs. (4.40), (4.4S), and (4.b0) we obtain
J J ]u [ ]( ] 1) u(u 1)] ]u ( ] u)( ] u 1) ]u ,
J J ]u [ ]( ] 1) u(u 1)] ]u ( ] u)( ] u 1) ]u
(4.bJ)
Because J and J are Hermitian conjugates,
S
J J , J J , (4.b4)
the eigenvalues in Eq. (4.bJ) must be positive or zero.
9
This follows from
]u J (J ]u ) (J ]u ) J ]u 0 (4.bb)
Examples of Eq. (4.bJ) are provided by the matrices of Exercises J.z.9 (spin
1
z
),
J.z.11 (spin 1), and J.z.1J (spin
J
z
). For the orbital angular momentum ladder
operators L and L , explicit forms are given in Exercise z.b.10.
Since J raises the eigenvalue uto u 1, we relabel the resultant eigen-
function ]u 1 . From Eqs. (4.47) and (4.bJ) we see that
J ]u ( ] u)( ] u 1) ]u 1 , (4.b)
taking the positive square root and not introducing any phase factor. By the
same arguments
J ]u ( ] u)( ] u 1) ]u 1 (4.b7)
S
The Hermitian conjugation or adjoint operation is denned for matrices in Section J.4 and for
operators in general in Section 9.1.
9
For an excellent discussion of adjoint operators and Hilbert space, see Messiah, A. (191). van-
lvu Mcchancs, Chapter 7. Wiley, New York.
1.3 rbital Angular Homentum 24?
As shown later, orbital angular momentum is described with integral ]. From
the spins of some of the fundamental particles and of some nuclei, we get
] 1 z, J z, b z, . Our angular momentum is quantized, essentially as a
result of the commutation relations.
EXAHPIE 4.8.1
Spin
1
2
Statea The spin raising operator is given by
S
1
z
1
z
(
a

)
1
z
0 1 ( )
1 0
0 1
0 0
,
so that
S
0 1
0 0
0
1
1
0
,
which is consistent with [
1
z
(
1
z
)](
1
z
1
z
1) 1 of Eq. (4.b).
In spherical polar coordinates , the functions , |u Y
u
|
( , ) are the
spherical harmonics of Section 11.b. Similarly, we can work out Eq. (4.b7)
for the orbital angular momentum lowering operator L L
a
L
c

( cot ) from Exercise z.b.10b acting on the spherical harmonic
, 11
J
S
sin c
Y
11
( , ) We nnd
L Y
11
c

cot ( 1)
J
S
sin c
c

J
S
(cos cos )c

4
cos z , 1, 0 zY
10
,
where (1 1)(1 1 1) z. That is, the ladder formulas (Eqs. 4.bb and
4.b) apply to the spherical harmonics and are equivalent to using the differ-
ential operators for L
Generators for the classical Lie groups can be organized into those denning
additiveeigenvaluesandladder operators that raiseor lower theseeigenvalues.
For the rotation group SO(J) these are L
:
and L Altogether, they denne the
selection rules of a symmetry group.
EXERCISES
4.3.1 Show that (a) [ J , J
z
] 0, (b) [ J , J
z
] 0.
4.3.2 Write down all matrix elements ] u ]u of the angular momentum
operator J
z
, J
:
, and J
4.3.3 Construct matrix representations for J and J
:
for angular momentum
J 1, J z, z, b z
24S Chatcr 1 6roup Theory
4.3.4 Let a, b be a complete set of common eigenfunctions of the Hermitian
operators and that is,
a, b a a, b , a, b b a, b
Show that [ , ] 0 Is the inverse conclusion valid?
4.3.5 The three z z matrices
]
for ] 1, z, J satisfy the same commutation
relations as the Pauli matrices. Show that there is one matrix so that
]
]
1
for ] 1, z, J Interpret this result in your own words.
4.3.6 Determine the eigenvalues of the orbital angular momentum operators
L
z
and L
:
for the functions c
r
z
a
z
cos and c
r
z
a
z
sin c

, where a is
some constant length.
4.3.? Derive the generators of SU(J). Determine the order of SU(J). Write
down various raising and lowering operators.
4.3.S Explain why the theorem of Exercise 4.J.b does not hold for three cor-
responding J J generator matrices of SU(J).
4.4 Bomogeneous Iorentz 6roup
Generalizing the approach to vectors of Section z., in special relativity we
demand that our physical laws be covariant
10
under
space and time translations,
rotations in real, three-dimensional space, and
Lorentz transformations.
The demand for covariance under translations is based on the homogeneity of
space and time. Covariance under rotations is an assertion of the isotropy of
space. The requirement of Lorentz covariance follows from special relativity.
Space rotations and pure Lorentz transformations together formthe homo-
geneous Lorentz group, and they form the Poincar e group when translations
are included as well.
We nrst generate a subgroupthe Lorentz transformations in which the
relative velocity v is along the a a
1
axis. The generator may be determined
by considering spacetime reference frames moving with a relative velocity ,
an innnitesimal.
11
The relations are similar to those for rotations in real space
(Sections z. and J.J), except that here the angle of rotation is pure imaginary.
Lorentz transformations are linear not only in the space coordinates a
but
also in time l. They originate from Maxwell's equations of electrodynamics,
10
To be covariant means to have the same form in different coordinate systems, often called
inertial frames, so that there is no preferred reference system (compare Section z.).
11
This derivation, with a slightly different metric, appears in an article by 1. L. Strecker, Au. J.
Phs. 35, 1z (197).
1.1 Bomogeneous Iorentz 6roup 249
which are invariant under Lorentz transformations. Lorentz transformations
leave the quadratic form
c
z
l
z
a
z
1
a
z
z
a
z
J
a
z
0
a
z
1
a
z
z
a
z
J
invariant, where a
0
cl and c is the velocity of light that is the same in all
inertial frames. We see this invariance if weswitchona light source at theorigin
of the coordinate system. At time l light has traveled the distance cl a
z
so that c
z
l
z
a
z
1
a
z
z
a
z
J
0. Special relativity requires that in any inertial
frame whose coordinates are a
that moves with velocity c in any direction

relative to the a
system and has the same origin at time l 0,

c
z
l
z
a
z
1
a
z
z
a
z
J
0
holds also.
Four-dimensional spacetime with the metric
a a a
z
a
z
0
a
z
1
a
z
z
a
z
J
is called Minkowski space, with the scalar product of two four-vectors denned
as
a b a
0
b
0
a b
Using the metric tensor
(q ) (q )
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
(4.bS)
we can raise and lower the indices of a four-vector, such as the coordinates
a (a
0
, ), so that a q a (a
0
, ), a q a a
z
0

z
, Einstein's
summation convention being understood. It is a common convention to use
Greek letters for four-vector indices and, because a
0
a
0
, we usually write a
0
for the time variable. For the gradient that is a covariant four-vector we have
a
0
,
a
,
a
0
, ,
so that
z
a
0
z
z
is a Lorentz scalar, just like the metric a
z
a
z
0

z
.
For c, in the nonrelativistic limit, a Lorentz transformation must be
Galilean. Hence, to derive the form of a Lorentz transformation along the
a
1
-axis, we start with a Galilean transformation for innnitesimal relative ve-
locity :
a
1
a
1
l a
1
a
0
(4.b9)
Here, as usual, c. By symmetry, we also write
a
0
a
0
a a
1
, (4.0)
with a a parameter that is nxed by the requirement that a
z
0
a
z
1
be invariant,
a
z
0
a
z
1
a
z
0
a
z
1
(4.1)
Remember that a (a
0
, ) is the prototype four-dimensional vector in Min-
kowski space. Thus, Eq. (4.1) is simply a statement of the invariance of the
square of the magnitude of the distance" vector under Lorentz transformation
in Minkowski space. Here is where the special relativity is brought into our
transformation. Squaring and subtracting Eqs. (4.b9) and (4.0) and discarding
terms of order ( )
z
, we nnd a 1. Equations (4.b9) and (4.0) may be
combined as a matrix equation
a
0
a
1
(1
z 1
)
a
0
a
1
, (4.z)
where
1
happens to be the Pauli matrix
1
, and the parameter represents
an innnitesimal change. Using the same techniques as in Section 4.z, we repeat
the transformation N times to develop a nnite transformation with the velocity
parameter N . Then
a
0
a
1
1
z
1
N
N
a
0
a
1
(4.J)
In the limit as N ,
lim
N
1
z
1
N
N
exp(
1
) (4.4)
As in Section 4.z, the exponential is expanded as a Maclaurin expansion
exp(
1
) 1
z 1
1
z!
(
1
)
z
1
J!
(
1
)
J
(4.b)
Noting that
z
1
1
z
,
exp(
1
) 1
z
cosh
1
sinh (4.)
Hence, our nnite Lorentz transformation is
a
0
a
1
cosh sinh
sinh cosh
a
0
a
1
(4.7)
1
has generated the representations of this pure Lorentz transformation. Cosh
and sinh may be identined by considering the origin of the primed coordi-
nate system, a
1
0, or a
1
l. Substituting into Eq. (4.7), we have
0 a
1
cosh a
0
sinh (4.S)
With a
1
l and a
0
cl,
tanh c (4.9)
Note that the rapidity c except in the limit as 0. The rapidity is the
additive parameter of a pure Lorentz transformation (boost") along some
axis that plays the same role as the angle of a rotation about some axis.
Using 1 tanh
z
(cosh
z
)
1
,
cosh (1
z
)
1 z
, sinh (4.70)
The preceding special case of the velocity parallel to one space axis is easy,
but it illustrates the innnitesimal velocity-exponentiationgenerator tech-
nique. Now this exact technique may be applied to derive the Lorentz transfor-
mation for the relative velocity v not parallel to any space axis. The matrices
given by Eq. (4.7) for the case of v `
a
form a subgroup. The matrices in
the general case do not. The product of two Lorentz transformation matrices,
(v
1
) and (v
z
), yields a third Lorentz matrix (v
J
) if the two velocities v
1
and
v
z
are parallel. The resultant velocity v
J
is related to v
1
and v
z
by the Einstein
velocity addition law (Exercise 4.4.J). If v
1
and v
z
are not parallel, no such
simple relation exists. Specincally, consider three reference frames S, S , and
S , with S and S related by (v
1
) and S and S related by (v
z
). If the veloc-
ity of S relative to the original system S is v
J
, S is not obtained from S by
(v
J
) (v
z
) (v
1
). Rather, one can show that
(v
J
) (v
z
) (v
1
), (4.71)
where is a J J space rotation matrix. With v
1
and v
z
not parallel, the nnal
system S is rotated relative to S. This rotation is the origin of the Thomas
precessioninvolved inspin-orbit coupling terms in atomic andnuclear physics.
Because of its presence, the (v) by themselves do not form a group.
\ector Analysis in Hinkowski Space--Time
We have seen that the propagation of light determines the metric
r
z
c
z
l
z
0 r
z
c
z
l
z
,
where a (cl, r) is the coordinate four-vector. For a particle moving with
velocity v the Lorentz invariant innnitesimal version
cd da da c
z
dl
z
dr
z
dl c
z
v
z
dennes the invariant proper time on its track. Because of time dilation in
moving frames, a proper time clock rides with the particle (in its rest frame)
and runs at the slowest possible rate compared to any other inertial frame (e.g.,
of an observer). The four-velocity of the particle can now be denned properly
as
da
d
v
c
c
z
v
z
,
v
c
z
v
z
so that v
z
1, and the four-momentum j cuv (
E
c
, p) yields Einstein's
famous energy relation
E
uc
z
1 v
z
c
z
uc
z
u
z
v
z
A consequence of v
z
1 and its physical signincance is that the particle is on
its mass shell j
z
u
z
c
z
Now we formulate Newton's equation for a aingIe particIe of mass uin
special relativity as
dj
d
K , with K denoting the force four-vector, so that
its vector part of the equation coincides with the usual form. For 1, z, J
we use d dl 1 v
z
c
z
and nnd
1
1 v
z
c
z
dp
dl
F
1 v
z
c
z
K
determining K in terms of the usual force F Next, we need to nnd K
0
We
proceed by analogy with the derivation of energy conservation, multiplying
the force equation into the four-velocity
uv
dv
d
u
z
dv
z
d
0,
because v
z
1 const. The other side of Newton's equation yields
0 v K
K
0
1 v
z
c
z
F v c
1 v
z
c
z
z
,
so that K
0
F v c
1 v
z
c
z
is related to the work done by the force on the particle.
Now we turn to two-body collisions in which energy-momentumconserva-
tion takes the form j
1
j
z
j
J
j
4
, where j
are the particle four-momenta.

Because the scalar product of any four-vector with itself is an invariant under
Lorentz transformations, it is convenient to denne the Lorentz invariant energy
squared s (j
1
j
z
)
z
P
z
, where P is the total four-momentum, and use
units where the velocity of light c 1 The laboratory system(lab) is denned as
the rest frame of the particle with four-momentum j
z
(u
z
, 0) and the center
of momentum frame (cms) by the total four-momentum P (E
1
E
z
, 0)
When the incident lab energy E
L
1
is given, then
s j
z
1
j
z
z
zj
1
j
z
u
z
1
u
z
z
zu
z
E
L
1
is determined. Now the cms energies of the four particles are obtained from
scalar products
j
1
P E
1
(E
1
E
z
) E
1
s
so that
E
1
j
1
(j
1
j
z
)
s
u
z
1
j
1
j
z
s
u
z
1
u
z
z
s
z s
,
E
z
j
z
(j
1
j
z
)
s
u
z
z
j
1
j
z
s
u
z
z
u
z
1
s
z s
,
E
J
j
J
(j
J
j
4
)
s
u
z
J
j
J
j
4
s
u
z
J
u
z
4
s
z s
,
E
4
j
4
(j
J
j
4
)
s
u
z
4
j
J
j
4
s
u
z
4
u
z
J
s
z s
,
by substituting
zj
1
j
z
s u
z
1
u
z
z
, zj
J
j
4
s u
z
J
u
z
4
Thus, all cmsenergies E
dependonly ontheincident energybut not onthescat-

tering angle. For elastic scattering u
J
u
1
, u
4
u
z
so that E
J
E
1
, E
4
E
z
The Lorentz invariant momentum transfer squared
l (j
1
j
J
)
z
u
z
1
u
z
J
zj
1
j
J
depends linearly on the cosine of the scattering angle.
EXAHPIE 4.4.1
Kaon Decay and Pion Photoproduction ThreahoId Find the kinetic en-
ergies of the muon of mass 10 Me\ and massless neutrino into which a K
meson of mass 494 Me\ decays in its rest frame.
Conservation of energy and momentum gives u
K
E E s Apply-
ing the relativistic kinematics described previously yields
E
j (j j )
u
K
u
z
j j
u
K
,
E
j (j j )
u
K
j j
u
K
Combining both results, we obtain u
z
K
u
z
zj j so that
E T u
u
z
K
u
z
zu
K
zbS 4 Me\,
E T
u
z
K
u
z
zu
K
zJb Me\
As another example, in the production of a neutral pion by an incident photon
according to j
0
j at threshold, the neutral pion and proton are
created at rest in the cms. Therefore,
s (j j)
z
u
z
j
zu
j
E
L
(j j )
z
(u u
j
)
z
so that E
L
u
u
z
zu
j
144 7 Me\.
The Lorentz group is the symmetry group of electrodynamics: It governs spe-

cial relativity of the electroweak gauge theory and of the strong interactions
described by quantumchromodynamics. The metric of Minkowski spacetime
is Lorentz invariant and expresses the propagation of light, that is, the veloc-
ity of light is the same in all inertial frames. Newton's equations of motion are
straightforward to extend to special relativity. The kinematics of two-body col-
lisions are important applications of vector algebra in Minkowski spacetime.
Hioqrohicol Doto
Lorentz, HendriI Antoon. Lorentz, a Dutch physicist, was born in 1SbJ
in Arnhem and died in 19zS in Haarlem. He obtained his Ph.D. in 1S7b at Lei-
den University, where he returned J years later as a professor of theoretical
physics and stayed until his death. He renned Maxwell's theory of radia-
tion, and light in particular, attributing it to oscillations of charged particles
within matter at a time when atoms were not universally recognized. These
were later identined as electrons by 1. 1. Thomson and ions by Arrhenius. As
proof, he suggested subjecting atoms to magnetic nelds, predicting effects
that were demonstrated by his student Zeeman in 1S9. He also analyzed the
physical consequences (LorentzFitzgerald contraction) of the invarianceof
Maxwell's equations under transformations that depend on the relative ve-
locity of the inertial frames (now named after him) and differ from those
(Galilean transformations) of Newton's equations of motion. Thus, he was
a forerunner of Einstein's special relativity.
EXERCISES
4.4.1 Two Lorentz transformations are carried out in succession:
1
along the
a-axis and then
z
along the -axis. Show that the resultant transfor-
mation (given by the product of these two successive transformations)
cannot be put in the form of a single Lorentz transformation.
Nolc. The discrepancy corresponds to a rotation so that pure Lorentz
transformations (boosts) do not form a group.
4.4.2 Rederive the Lorentz transformation working entirely in Minkowski
space (a
0
, a
1
, a
z
, a
J
) with a
0
a
0
cl. Show that the Lorentz transfor-
mation may be written (v) exp( ), with
0
0 0 0
0 0 0
0 0 0
,
and , , the direction cosines of the velocity v.
4.4.3 Using the matrix relation, Eq. (4.7), let the rapidity
1
relate the Lorentz
reference frames (a
0
, a
1
) and (a
0
, a
1
). Let
z
relate (a
0
, a
1
) and
(a
0
, a
1
). Finally, let relate (a
0
, a
1
) and (a
0
, a
1
). From
1 z
,
derive the Einstein velocity addition law
1 z
1
1 z
c
z
4.4.4 (a) A particle A decays into B and C. What is the energy of the particle B
in the rest frame of A? An example is a neutral pion that decays into two
photons. (b) Explain why a photon of sufncient energy can produce an
electronpositron pair only in the presence of matter.
Nolc. The particle masses obey u
A
u
E
u
C
4.4.5 Determine the minimum frequency of a -ray that disintegrates a
deuteron into a neutron of mass u
n
9J9 bb Me\ c
z
and a proton
of mass u
j
9JS z7z Me\ c
z
. The deuteron has a binding energy of
z zz Me\.
Hnl. Ignore the Fermi motion, that is, take the neutron and proton at
rest. Explain why this is a good approximation.
4.4.6 An observer moving with four-velocity v measures the energy E of a
particle with four-momentum j Show that E cj v
4.4.? Derive the relativistic generalization of the equation of motion
dp
dl
qE qv B of a charged particle in an electromagnetic neld.
4.4.S Show that the Lorentz covariant equation of motion
dj
d
q
0
uc
F j
of a charged particle in an electromagnetic neld is consistent with (has
as a solution) the on-mass-shell relation j
z
u
z
c
z
. Here, uis the rest
mass of the particle and its proper time (i.e., the Lorentz invariant
time measured in its rest frame), j ucv is its four-momentum, and
F A A is the electromagnetic neld tensor.
Additional Reading
Buerger, M. 1. (19b). E|cucnlar Crsla||oqrajh. Wiley, New York. A com-
prehensive discussion of crystal symmetries. Buerger develops all Jz point
groups and all zJ0 space groups. Related books by this author include
Conlcujorar Crsla||oqrajh. McGraw-Hill, New York (1970), Crsla|
Slrvclvrc Ana|ss. Krieger, NewYork(1979), and!nlrodvclonlo Crsla|
Gcouclr. Krieger, New York (1971, reprint 1977).
Burns, G., and Glazer, A. M. (197S). Sjacc Grovjs for So|d Slalc Sccnlsls.
AcademicPress, NewYork. Awell-organized, readable treatment of groups
and their application to the solid state.
de-Shalit, A., and Talmi, I. (19J). Nvc|car Shc|| Modc|. Academic Press, New
York. We adopt the CondonShortley phase conventions of this text.
Edmonds, A. R. (19b7). Anqv|ar Moucnlvu n vanlvu Mcchancs. Prince-
ton Univ. Press, Princeton, N1.
Falicov, L. M. (19). Grovj Thcor and !ls Phsca| Ajj|calons [Notes
compiledbyA. Luehrmann]. Univ. of ChicagoPress, Chicago. Grouptheory
with an emphasis on applications to crystal symmetries and solid-state
physics.
Greiner, W., andM uller, B. (19S9). vanlvuMcchancs Suuclrcs. Springer,
Berlin. We refer to this textbook for more details and numerous exercises
that are worked out in detail.
Hamermesh, M. (19z). Grovj Thcor and !ls Ajj|calon lo Phsca| Prob-
|cus. Addison-Wesley, Reading, MA. A detailed, rigorous account of both
nnite and continuous groups. The Jz point groups are developed. The
continuous groups are treated with Lie algebra included. A wealth of ap-
plications to atomic and nuclear physics.
Heitler, W. (1947). Thc vanlvu Thcor of Eadalon, znd ed. Oxford Univ.
Press, Oxford. Reprinted, Dover, New York (19SJ).
Higman, B. (19bb). Ajj|cd GrovjThcorclc and Malra Mclhods. Claren-
don, Oxford. A complete and unusually intelligible development of matrix
analysis and group theory.
Messiah, A. (191). vanlvu Mcchancs, \ol. z. North-Holland, Amsterdam.
Panofsky, W. K. H., and Phillips, M. (19z). C|assca| E|cclrcl and Maq-
nclsu, znd ed. Addison-Wesley, Reading, MA. The Lorentz covariance of
Maxwell's equations is developed for both vacuum and material media.
Panofsky and Phillips use contravariant and covariant tensors.
Park, D. (19S). Resource letter SP-1 on symmetry in physics. Au. J. Phs.
36, b77bS4. Includes a large selection of basic references on group the-
ory and its applications to physics: atoms, molecules, nuclei, solids, and
elementary particles.
Ram, B. (197). Physics of the SU(J) symmetry model. Au. J. Phs. 35, 1.
An excellent discussion of the applications of SU(J) to the strongly inter-
acting particles (baryons). For a sequel to this, see Young, R. D. (197J).
Physics of the quark model. Au. J. Phs. 41, 47z.
Rose, M. E. (19b7). E|cucnlar Thcor of Anqv|ar Moucnlvu. Wiley, New
York. Reprinted, Dover, New York (199b). As part of the development of
the quantum theory of angular momentum, Rose includes a detailed and
readable account of the rotation group.
Wigner, E. P. (19b9). Grovj Thcor and !ls Ajj|calon lo lhc vanlvu Mc-
chancs of Alouc Sjcclra (translated by 1. 1. Grifnn). Academic Press,
New York. This is the classic reference on group theory for the physicist.
The rotation group is treated in considerable detail. There is a wealth of
applications to atomic physics.

3 Teoria de Grupos Arfken

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

3 Teoria de Grupos Arfken

Uploaded by

Copyright:

Available Formats

Group Theory

Disciplined judgment, about what is neat

(see Section J.J for orthogonal

(see Section J.4 for unitary matrices, which preserve

, with real angle , of quan-

, and the elements of a second group are labeled

is a one-to-one correspondence for all values of . If q

Often, groups occur in physics as sets of transformations of coordinates or

. Let a be a nxed element of the

and the spin-down component

] 0, which can also be verined directly.

That u is an eigenfunction of both J

that moves with velocity c in any direction

system and has the same origin at time l 0,

are the particle four-momenta.

dependonly ontheincident energybut not onthescat-

The Lorentz group is the symmetry group of electrodynamics: It governs spe-

You might also like