M. Zirnbauer-Lectures On Advanced Quantum Mechanics
M. Zirnbauer-Lectures On Advanced Quantum Mechanics
M. Zirnbauer-Lectures On Advanced Quantum Mechanics
M. Zirnbauer
Institut für Theoretische Physik,
Universität zu Köln
WS 2010/2011
Contents
1 Scattering theory 4
1.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Types of scattering process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Observables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Lippmann-Schwinger equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4.1 Born approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5 Scattering by a centro-symmetric potential (partial waves) . . . . . . . . . . . . . 10
1.5.1 Optical theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5.2 Example: scattering from a hard ball . . . . . . . . . . . . . . . . . . . . . 13
1.5.3 Asymptotics of the spherical Bessel functions . . . . . . . . . . . . . . . . . 14
1.6 Time-dependent scattering theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.6.1 Example: potential scattering d = 1 . . . . . . . . . . . . . . . . . . . . . . 17
1.6.2 Scattering by a centro-symmetric potential in d = 3 . . . . . . . . . . . . . 20
1.6.3 On the condition range(W− ) ⊂ domain(W+† ) . . . . . . . . . . . . . . . . . 21
1.7 Time reversal and scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.7.1 Some linear algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.7.2 T -invariant scattering for spin 0 and 1/2 . . . . . . . . . . . . . . . . . . . 27
1
2.8 Spinor representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.8.1 Tensor product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.9 Transforming scalar and vector fields . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.10 Infinitesimal spin transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.11 Spin group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.12 Relativistic covariance of the Dirac equation . . . . . . . . . . . . . . . . . . . . . 46
2.13 Discrete symmetries of the Dirac equation . . . . . . . . . . . . . . . . . . . . . . 48
2.13.1 Pin group and parity transformation . . . . . . . . . . . . . . . . . . . . . 48
2.13.2 Anti-unitary symmetries of the Dirac equation . . . . . . . . . . . . . . . . 49
2.13.3 Parity violation by the weak interaction . . . . . . . . . . . . . . . . . . . 50
2.14 Dirac electron in a Coulomb potential . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.14.1 Using rotational symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.14.2 Reduction by separation of variables . . . . . . . . . . . . . . . . . . . . . 53
2.14.3 Solution of radial problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3 Second quantization 57
3.1 Bosons and fermions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.2 Weyl algebra and Clifford algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.3 Representation on Fock space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.4 Fock space scalar product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.5 Second quantization of one- and two-body operators . . . . . . . . . . . . . . . . . 66
3.6 Second quantization of Dirac’s theory . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.6.1 Motivation: problems with the Dirac equation . . . . . . . . . . . . . . . . 69
3.6.2 Stable second quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.6.3 The idea of hole theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.6.4 Mode expansion of the Dirac field . . . . . . . . . . . . . . . . . . . . . . . 73
3.7 Quantization of the electromagnetic field . . . . . . . . . . . . . . . . . . . . . . . 75
3.7.1 Review and consolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.7.2 More preparations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.7.3 Symplectic structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.7.4 Complex structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.7.5 Fock space: multi-photon states . . . . . . . . . . . . . . . . . . . . . . . . 81
3.7.6 Casimir energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.7.7 Mode expansion for the torus geometry . . . . . . . . . . . . . . . . . . . . 83
3.8 Matter-field interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.9 γ-decay of excited states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2
4.2 Basic notions of representation theory . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2.1 Borel-Weil Theorem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.3 Invariant tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.3.1 Invariant tensors of degree 2 . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3.2 Example: SO3 -invariant tensors of degree 2 . . . . . . . . . . . . . . . . . . 94
4.3.3 Example: SU2 -invariant tensors of degree 2 . . . . . . . . . . . . . . . . . . 96
4.3.4 Invariant tensors of degree 3 . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.4 The case of SU2 : Wigner 3j-symbol, Clebsch-Gordan coefficient . . . . . . . . . . 101
4.5 Integrals of products of Wigner D-functions . . . . . . . . . . . . . . . . . . . . . 106
4.6 Tensor operators, Wigner-Eckart Theorem . . . . . . . . . . . . . . . . . . . . . . 109
3
1 Scattering theory
1.1 Preliminaries
The formalism of scattering theory plays an important role in physics, as it allows to predict
experimental observations from the fundamental interactions postulated by theory. A typical
setup, say (i), for a scattering experiment looks as follows:
Thus there is a source for a beam of particles which, after preparation of the beam kinematics,
hit a target and then get observed in a detecting device. A few elementary remarks are: thick
targets causes multiple scattering events. While these might be hard to analyze in the case of an
amorphous target, they lead to characteristic interference patterns and thus structural information
in the case of a crystal (for the target).
Next, let us look at the symmetric situation, (ii), of a scattering experiment with two colliding
beams:
In this setting one wants to place the detector far from the collision zone for good angular reso-
lution. One does not place the detector in the forward direction of an incoming beam.
Remark. In the case of scattering of massive particles, the situation (i) can be reduced to the
situation (ii) by a change of inertial reference frame.
• Fermilab Chicago, Tevatron (1983; p + p; ca. 2 TeV; top quark discovered; 6 km)
4
1.2 Types of scattering process
We briefly introduce some nomenclature:
a + b −→ a + b .
a + b −→ a′ + b′ ,
internal degrees of freedom of the particles get excited (e.g., rotational or vibrational degrees
of freedom).
3. In the general category of rearrangement scattering, the identity of the scattered particles is
altered:
a + b −→ c + d + e + . . .
An example is the break-up of the deuteron into its constituents (namely, one proton and
one neutron) during a collision with another particle. A second example is nucleon-nucleon
scattering, where one has the possibilities
plus various other “channels” not shown here. (The theoretical analysis of rearrangement
scattering requires the formalism of multichannel scattering, which will not be treated in
this lecture course.)
4. By resonance scattering one means a process that involves the formation and subsequent
decay of an unstable (but possibly very long-lived) intermediate state.
5. In scattering theory one may also consider the decay of an unstable particle (a −→ b+c+. . .).
For example, a free neutron decays into a proton, an electron, and an anti-neutrino:
n −→ p + e + ν̄e .
5
1.3 Observables
Here are some quantities that are observable by scattering experiments:
We now define the differential scattering cross section, assuming the situation (i) of Section 1.1.
Let the (probability) current density of the incoming particles be homogeneous and directed along
the z-axis for a Cartesian coordinate system:
I = i dxdy .
Remark. In these lecture notes we will be concerned with the scattering of quantum mechanical
particles. We mention in passing that the notions of differential and total scattering cross section
already make sense in the setting of classical mechanics. For example, the total cross section for
hard balls (as target particles) of radius R is σtot = πR2 in the classical limit.
We now state a practical recipe by which to calculate the differential scattering cross section
for potential scattering of Schrödinger particles with energy E = ~2⃗k 2 /2m and wave vector ⃗k:
6
Find the solution of Schrödinger’s equation Hψ = Eψ with the asymptotic form
|⃗
x|=r→∞ ⃗ eikr
ψ(⃗x) −→ eik·⃗x + fE (Ω) , k = |⃗k| .
r
Then the differential cross section is
dσ
= |fE (Ω)|2 .
dΩ
[This is to enable you to do one of the problems on the first problem sheet.]
H0 ψ0 = Eψ0 .
For ε = 0+ let G0 := (E + iε − H0 )−1 (operator inverse) the free Green operator or resolvent.
Claim. If ψ satisfies
ψ0 = ψ − G0 V ψ = (1 − G0 V )ψ = G0 (G−1
0 − V )ψ.
Now let
G := (E + iε − H)−1
G−1 = G−1
0 −V and ψ0 = G0 G−1 ψ,
so that
ψ = GG−1
0 ψ0 = G (E + iε − H0 )ψ = iεGψ0 .
It follows that
E−H E − H
(E − H)ψ = iε
ε→0
ψ0 −→ 0 , since < 1.
E + iε − H E + iε − H
op
7
Now let
p2 ~2 2
H0 = =− ∇, V = V (⃗x) (“potential scattering”).
2m 2m
Then the operator G0 has the integral kernel (or “Green’s function”)
′
′ eik|⃗x−⃗x | 2m
⟨⃗x | G0 | ⃗x ⟩ = − ,
4π|⃗x − ⃗x ′ | ~2
and the abstract equation for ψ takes the explicit form of an integral equation:
∫ ′
2m eik|⃗x−⃗x |
ψ(⃗x) = ψ0 (⃗x) − 2 V (⃗x ′ ) ψ(⃗x ′ ) d3 x′
~ R3 4π|⃗x − ⃗x | ′
If ∥ G0 V ∥op < 1, one can expand the solution in a geometric series (called the Born series):
ψ = (1 − G0 V )−1 ψ0 = ψ0 + G0 V ψ0 + G0 V G0 V ψ0 + ... .
For ∥ G0 V ∥op ≪ 1 (high energy, or weak potential) one may use the lowest-order (or first) Born
approximation,
ψ ≈ ψ0 + G0 V ψ0 .
To write the first Born approximation in explicit form, let us take ψ0 (⃗x) = eikz . Then
∫ ′
m eik|⃗x−⃗x | ′
(G0 V ψ0 )(⃗x) = − V (⃗x ′ ) eikz d3 x′ .
2π~2 R3 |⃗x − ⃗x |
′
For V of finite range and ⃗x far from the scattering center of V we may expand
( )
⃗x ⃗x ′ ⃗x ′
′
k|⃗x − ⃗x | = kr − ≈ kr 1 − Ω · = kr − kΩ · ⃗x ′ .
r r r
We then see that
Rewriting this in a form which is independent of the choice of Cartesian basis, we obtain
∫
m ⃗ ⃗ ′ ′
e−i(ki −kf )·⃗x +ikz V (⃗x ′ ) d3 x′ .
(1)
fE (Ω) = − 2
2π~ R3
(1)
We thus see that the scattering amplitude in the first Born approximation, fE (Ω), is essentially
given by the Fourier transform of the potential V .
8
1.4.2 Examples
1. V (x) = V0 λ3 δ(x). Here V0 has the physical dimension of energy, and λ has the physical
dimension of length. δ is the Dirac δ-function (actually, δ-distribution), with support at
zero (the origin of the coordinate system). The Fourier transform of the δ-function is simply
a constant, so
(1) m λ V0
fE (Ω) = − V 0 λ3 = − ,
2π~ 2 4π ~ /(2mλ2 )
2
independent of the energy E and the outgoing direction Ω. Note that Ekin = ~2 /(2mλ2 ) is
a rough estimate of the kinetic energy of a Schrödinger particle with mass m confined to a
box of size λ. The ratio V0 /Ekin is a dimensionless measure of the strength of the scattering
potential. We have written the answer for fE in a form which makes it transparent that fE
has the physical dimension of length. We notice that the scattering amplitude fE is negative
in the repulsive case (V0 > 0) and positive in the attractive case (V0 < 0).
2. V (x) = V0 λ5 ∇2 δ(x). We first give some explanation of what it means to apply the Laplacian
∇2 to a δ-function. (As a side remark: this operation is mathematically well-defined if δ is
regarded as a Schwartz distribution.) For this purpose we use a Gaussian regularization of
the δ-function:
( )
2 − 2a
2
∂2 2 ∂ r2
e− 2a2
r
2 −3/2
∇ δ(x) = lim ∇ e
2
/(2πa )
2 2 3/2
= lim (2πa ) 2
+
a→0+ a→0+ ∂r r ∂r
( ) ( 2 )
2 −3/2 ∂ 2 ( r ) − r22 −3/2 r 3 2
− r2
= lim (2πa ) + − 2 e 2a = lim (2πa )
2
− e 2a .
a→0+ ∂r r 2a a→0+ a4 a2
By using the rule ∇ → ik for the Fourier transform, we obtain the following expression for
the first Born approximation to the scattering amplitude:
(1) m m
fE = + 2
V0 λ5 (ki − kf )2 = V0 λ5 k 2 (1 − cos θ) .
2π~ π~2
If λ is the range of the scattering potential (i.e., we take the Gaussian regularization parameter a
to be λ ≡ a), then from the properties of the Fourier transform we expect the following qualitative
picture (for our second example) of the scattering cross section:
9
1.5 Scattering by a centro-symmetric potential (partial waves)
In this section we consider Hamiltonians of the form
~2 2 √
H=− ∇ + V (r) , r= x2 + y 2 + z 2 ,
2m
where the potential V (r) is invariant under all rotations fixing a center (which we take to be the
origin of our Cartesian coordinate system x, y, z). We assume that V (r) decreases faster than 1/r
in the limit of r → ∞.
Our goal here is to explain the ‘method of partial waves’, which is one of the standard methods
of scattering theory. To get started, we recall a few facts known from the basic course on quantum
mechanics. Using spherical polar coordinates
If the incoming wave of the scattering wave function ψ is a plane wave eikz traveling in the z
direction, then we expect the scattering amplitude to be independent of the azimuthal angle ϕ
(by the rotational symmetry of the potential V ). In fact, the azimuthal angle ϕ will never appear
in the following discussion.
We make an ansatz for the wave function of the form
∑
∞
ψ= ψl (r) Pl (cos θ) ,
l=0
where Pl is the Legendre polynomial of degree l. We recall that Legendre polynomials are eigen-
functions of the angular part of the Laplacian:
∂ ∂ ∂
− sin θ Pl (cos θ) = l(l + 1) Pl (cos θ) . (1.1)
sin θ ∂θ ∂θ
~2 k2
We now write the energy E of the Schrödinger particle in the form E = 2m
. Our Ansatz for ψ
then leads to the following differential equation for the radial functions ψl (r) :
( )
1 d2 l(l + 1) 2m
− ◦r+ + 2 V (r) ψl (r) = k 2 ψl (r) . (1.2)
r dr2 r2 ~
For large values of r this equation and its general solution simplify to
d2 eikr e−ikr
− rψl = k 2 rψl , ψl (r) = A +B .
dr2 r r
Now it is a basic property (called conservation of probability or ‘unitarity’ for short) of quantum
~
mechanics that the divergence of the probability current density j = m
Im ψ̄∇ψ must vanish for
a solution ψ of the time-independent Schrödinger equation Hψ = Eψ. A quick computation
shows that div j ≡ 0 is possible only if the radially outgoing wave eikr /r and the radially incoming
10
wave e−ikr /r combine to a standing wave sin(kr + δ)/r . This requires that A and B have the
same magnitude |A| = |B|. Thus the large-r asympotics of any solution of the time-independent
Schrödinger equation Hψ = Eψ with azimuthal symmetry has to be
r→∞
∑
∞
Al eikr + Bl e−ikr
ψ −→ (2l + 1) Pl (cos θ) , |Al | = |Bl | . (1.3)
l=0
2ikr
The factor (2l + 1)/(2ik) has been inserted for later convenience.
Our boundary conditions for the scattering problem dictate that
r→∞ eikr
ψ −→ eikz + fE (θ) .
r
Lemma. In order for ψ to be of this asymptotic form, the amplitudes Bl in (1.3) must be
Bl = (−1)l+1 .
To prove this lemma, we need an understanding of how eikz expands in partial waves. From
the basic quantum theory of angular momentum we recall that the Legendre polynomials have
the orthogonality property
∫ π
2 δll′
Pl (cos θ) Pl′ (cos θ) sin θ dθ = . (1.4)
0 2l + 1
Moreover, the Legendre polynomials form a complete system of functions on the interval [−1, +1] ∋
cos θ. We may therefore expand eikz = eikr cos θ as
∑
∞
ikr cos θ
e = il jl (kr) (2l + 1) Pl (cos θ) , (1.5)
l=0
−l ∫ π
i
jl (kr) = eikr cos θ Pl (cos θ) sin θ dθ . (1.6)
2 0
The factor of il has been inserted in order to make the function jl (kr) coincide with the so-called
spherical Bessel functions. The lowest-order spherical Bessel functions are
sin ξ sin ξ cos ξ
j0 (ξ) = , j1 (ξ) = − .
ξ ξ2 ξ
For small values of the argument, the spherical Bessel functions behave as
jl (kr) ∼ (kr)l .
This behavior follows from the orthogonality property (1.4) of the Legendre polynomials and the
fact that any polynomial in cos θ of degree l can be expressed as a linear combination of the
Legendre polynomials Pl′ (cos θ) of degree l′ ≤ l.
For large values of the argument, the spherical Bessel functions behave as
sin(kr − lπ/2)
jl (kr) ≃ . (1.7)
kr
We will motivate this important relation at the end of the present subsection.
11
Proof of Lemma. Using the expansion of eikr cos θ and the large-r asymptotics of the spherical
Bessel functions we have
∑∞
sin(kr − lπ/2)
ikr cos θ r→∞
e −→ il (2l + 1) Pl (cos θ) .
l=0
kr
which is of the expected form fE (θ)eikr /r. We already know that by unitarity we must have
|Al | = |Bl | = 1. It is customary to put Al = e2iδl and call δl the phase shift (in the channel of
angular momentum l). We then have the following result for the scattering amplitude:
∑
∞
e2iδl (E) − 1
fE (θ) = (2l + 1)Pl (cos θ) , (1.8)
l=0
2ik(E)
where we have emphasized the energy dependence of the phase shift δl (E) and the wave number
√
k(E) = 2mE/~ .
We now use the formula (1.8) for the scattering amplitude fE (θ) to compute the total scattering
cross section. By the orthogonality property (1.4) of the Legendre polynomials we obtain
∫ ∫
4π ∑
∞
dσ
σtot = dΩ = |fE (Ω)| dΩ = 2
2
(2l + 1) sin2 δl (E) . (1.9)
S2 dΩ S2 k (E) l=0
On the other hand, since (e2iδl − 1)/(2i) = eiδl sin δl has imaginary part sin2 δl , the imaginary part
of the scattering amplitude in the forward direction is
1 ∑
∞
Im fE (θ = 0) = (2l + 1) sin2 δl (E) .
k(E) l=0
Thus the total cross section and the forward scattering amplitude are related by
4π
σtot = Im fE (0) . (1.10)
k
This relation is called the optical theorem.
12
1.5.2 Example: scattering from a hard ball
We now illustrate the method of partial waves at the example of scattering from a hard ball:
{
+∞ r>R,
V (r) =
0 r<R.
The goal is to find the scattering phase shifts δl . Having calculated these, we get the scattering
amplitude and the cross section from the formulas of the previous subsection.
The scattering wave function must vanish identically inside the ball (r < R) where the potential
is repulsive and infinite. Outside the ball (r > R) the motion is that of a free particle. The
continuity of the wave function implies Dirichlet boundary conditions at the surface of the ball:
ψ =0.
r=R
In the exterior of the ball, where the motion is free, we look for solutions ψl (r)Pl (cos θ) of the
Schrödinger equation for a free particle of angular momentum l . We recall that the equation for
the radial functions ψl (r) reads
( )
1 d2 l(l + 1)
− 2
◦r+ 2
ψl (r) = k 2 ψl (r) . (1.11)
r dr r
Solutions of this equation are the spherical Bessel functions ψl (r) = jl (kr). Indeed, we know
that the plane wave eikr cos θ is a solution of the free Schrödinger equation, and by expanding
∑ l
eikr cos θ = i jl (kr)(2l + 1)Pl (cos θ) and using the eigenfunction property (1.1) of the Legendre
polynomials, we that jl (kr) solves the radial equation (1.11).
Since the radial equation is of second order, a single solution is not enough to express the
most general solution. We need a linearly independent second solution. To find it, we recall
that jl (kr) ∼ rl for r ≪ k −1 . Using this, it is easy to see that Ψl (x) = jl (kr)Pl (cos θ) in the
limit of r → 0 contracts to a solution of the Laplace equation ∇2 ψ = 0 . Now from the chapter
on multipole expansion in electrostatics, we know that there exists a second angular momentum
l solution r−l−1 Pl (cos θ) of Laplace’s equation. (Solutions of Laplace’s equation are also called
harmonic functions.) We therefore expect that there exists a solution, say nl (kr), of the radial
equation (1.11) with the corresponding small-r asymptotics:
( )
1 d2 l(l + 1)
− 2
r+ 2
nl (kr) = k 2 nl (kr) , nl (kr) ∼ r−l−1 (r → 0) . (1.12)
r dr r
Such a solution nl (kr) does exist, and it is called the spherical Neumann function of degree l . The
spherical Neumann functions of lowest degree are
cos ξ cos ξ sin ξ
n0 (ξ) = , n1 (ξ) = − − .
ξ ξ2 ξ
We now make an ansatz for the wave function of free motion outside the hard ball:
∑∞
ψ= il (al jl (kr) + bl nl (kr)) (2l + 1)Pl (cos θ) (r > R) .
l=0
13
The spherical Neumann functions have the following large-r asymptotic behavior:
l=0
Comparison with the general expression (1.3), where Al = e2iδl and Bl = −(−1)l , then yields
al + ibl = e2iδl , al − ibl = 1, and hence al = eiδl cos δl , bl = eiδl sin δl . Thus our ansatz takes the
form
∑
∞
ψ= il eiδl (cos(δl ) jl (kr) + sin(δl ) nl (kr)) (2l + 1)Pl (cos θ) (r > R) . (1.13)
l=0
We now impose the Dirichlet boundary condition ψ r=R = 0. This gives the following condition:
jl (kR)
tan(δl ) = − ,
nl (kR)
which determines the scattering phase shifts δl . This expression will be further elaborated in the
problem class.
Here we specialize to the long wave length limit kR ≪ 1 . In this limit δl ∼ (kR)2l+1 . Thus
scattering is appreciable only in the s-wave channel (l = 0). The exact value of the s-wave
scattering phase shift is ( )
j0 (kR)
δ0 = − arctan = −kR .
n0 (kR)
From Eq. (1.9) we then have the total scattering cross section
4π
σtot ≈ sin2 δl ≈ 4πR2 .
k2
Notice that this is larger (by a factor of four) than the geometric cross section πR2 of classical
scattering from a hard ball. (There is no inconsistency in this, as the long wave length limit
kR ≪ 1 is just the opposite of the classical limit kR ≫ 1.)
Here we fill in a gap which was left in the argument above: the asymptotic behavior (1.7).
For this purpose we use the following integral representation of the Legrendre polynomials:
∫ 2π
dϕ
Pl (cos θ) = (cos θ − i sin θ cos ϕ)l , (1.14)
0 2π
which can be verified by checking that the integral on the right-hand side satisfies the differential
equation (1.1). In combination with (1.6) this gives a representation of the spherical Bessel
functions as an integral over the unit sphere S2 :
∫
i−l
jl (ξ) = eiξ cos θ+l (cos θ−i sin θ cos ϕ) sin θ dθ dϕ . (1.15)
4π S2
14
For large values of ξ , the integrand oscillates strongly everywhere on S2 with the exception of two
points: the ‘north pole’ (θ = 0) and the ‘south pole’ (θ = π). These are the points where cos θ has
vanishing derivative. The contribution to the integral from each of these points can be calculated
by using the method of stationary phase (not explained here). Consider first the north pole θ = 0.
To compute its contribution to the integral, it is helpful to change coordinates as follows:
√
dn1 dn2
cos θ = 1 − n21 + n22 , sin θ cos ϕ = n1 , sin θ sin ϕ = n2 , sin θ dθ dϕ = √ .
1 − n21 − n22
e−iξ+lπ/2
≈ .
−2iξ
By adding these two contributions, we get the claimed behavior (1.7).
15
Given a vector ψ ∈ H we look for vectors ψ± ∈ H with the approximation property
t→±∞
∥ Vt ψ − Ut ψ± ∥ −→ 0 ,
i.e., evolving ψ by the full time evolution Vt for a very long time t → ±∞ leads to the same state
as evolving ψ± by the free time evolution Ut . This idea is illustrated in the figure shown below.
It should be noted, however, that it will not always be possible to find such ψ± for every ψ .
Indeed, any free motion Ut ψ± (if it is a free motion in the usual sense) must escape to infinity. On
the other hand, if there exist bound states and the vector ψ has nonzero projection on the space
of bound states, then Vt ψ contains a component which does not escape to infinity.
To avoid this complication, we use the unitarity of Vt (for finite t) to write
∥ Vt ψ − Ut ψ± ∥ = ∥ ψ − Vt† Ut ψ± ∥ .
The operators W± are called Moller operators, or wave operators. They are isometric, i.e., they
satisfy ∥ W± ψ ∥ = ∥ ψ ∥ for all ψ ∈ H . By the remark above they are not surjective (and hence
do not have an inverse) when bound states exist.
The definitions of Ut , Vt , and W± make sense even if the Hamiltonians H0 and/or H depend
on time. In the case of time-independent H0 and H we have the explicit formulas
For the next step we require that the range of W− (i.e., the image of H under W− ) be contained
in the domain of definition of W+† :
This requirement is physically reasonable, and will be discussed in Section 1.6.3 below. If it is
met, one may form the composition W+† W− . The situation is sketched in the following diagram.
16
Definition. With the pair (H0 , H) we associate an operator
S := W+† W− , (1.17)
In the case of time-independent H0 , H the Moller operators satisfy the intertwining relations
Vt W± = W± Ut . (1.18)
As an important corollary of the relations (1.18), the scattering operator commutes with the free
time evolution:
S Ut = Ut S . (1.19)
Now by differentiating the relations (1.19) (which are valid, as we recall, in the case of time-
independent H0 , H), we deduce that the scattering operator commutes with the free Hamiltonian:
H0 S = SH0 .
From spectral theory one then knows that H0 and S can be brought to diagonal form simultane-
ously. (Note that the eigenstates of H0 usually fail to be square-integrable. Thus they do not lie
inside the Hilbert space H .)
We illustrate the above at the example of potential scattering in one dimension, with
~2 d 2
H0 = − , H = H0 + V (x) ,
2m dx2
and V (x) vanishing (or decaying sufficiently fast) outside a finite region supp V ⊂ R . The
eigenspace of H0 with energy E > 0 is two-dimensional, being spanned by the two functions
√
ϕ± (x) := e±ikx , k= 2mE/~ > 0 .
17
In order for the scattering operator S to commute with H0 , it must leave the two-dimensional
H0 -eigenspace (with fixed energy E) invariant. In other words, applying S to either one of the
functions ϕ± (x) = e±ikx we must get a linear combination of the same two functions:
Sϕ+ = ρ ϕ− + τ ϕ+ , (1.20)
Sϕ− = τ ′ ϕ− + ρ′ ϕ+ . (1.21)
The complex numbers ρ , ρ′ (resp. τ, τ ′ ) are called reflection coefficients (resp. transmission coeffi-
cients). Due to S † S = IdH they satisfy the unitarity relations
In particular, the probabilities |ρ|2 for reflection and |τ |2 for transmission sum up to unity. We
are now left with the question of how to compute the matrix elements ρ, τ, ρ′ , τ ′ of the scattering
operator. This question is answered in the sequel.
We fix an energy E = ~2 k 2 /2m > 0 . For such an energy, we know that the eigenspace of
H0 is two-dimensional, and so is the eigenspace of H. As before, let ϕ± (x) = e±ikx denote the
corresponding eigenfunctions of H0 . We now introduce two sets of basis vectors for the two-
dimensional eigenspace of H with energy E. First, consider the state vectors
in
ψ± := W− ϕ± , (1.23)
obtained by applying the Moller operator W− = limt→−∞ eiHt/~ e−iH0 t/~ to the plane waves ϕ± .
These functions x 7→ ψ±
in in
(x) are solutions of the Schrödinger equation Hψ± in
= Eψ± . Indeed, by
differentiating the intertwining relation (1.18) at t = 0 we obtain
HW± = W± H0 , (1.24)
Here we used the fact that scattering solutions of Hψ = Eψ become superpositions A eikx +B e−ikx
of plane waves in the limit of |x| → ∞.
18
in
Similarly, ψ− is a stationary scattering state which originates from a plane wave ϕ− (x) = e−ikx
moving in the negative x-direction and coming in from x = +∞. There cannot be any incoming
component at x → −∞, so ψ−
in
must be of the asymptotic form
{ ′ +ikx
asympt A ·e + 1 · e−ikx x → +∞ ,
ψ− (x) −→
in
0·e +ikx
+ D′ · e−ikx x → −∞ .
in
One refers to ψ± as stationary scattering states satisfying incoming-wave boundary conditions.
out
The second set of states, ψ± , is defined by using the other Moller operator, W+ :
out
ψ± := W+ ϕ± .
in out out
ψ+ = Dψ− + Aψ+ , (1.25)
in
ψ− = D′ ψ−
out
+ A′ ψ+
out
. (1.26)
Finally, combining all this information we can make a statement about the scattering operator
in
S. By using the definition W− ϕ+ = ψ+ and the relations (1.25) and W+† ψ±
out
= ϕ± we compute
and, similarly,
Sϕ− = D′ ϕ− + A′ ϕ+ .
D=ρ, A=τ , D′ = τ ′ , A′ = ρ′ .
19
Thus the problem of computing the matrix elements of S is reduced to finding the asymptotics of
stationary scattering states with incoming-wave boundary conditions.
in in
Summary. If the solutions of Hψ± = Eψ± with incoming-wave boundary conditions are of the
asymptotic form
{
τ · e+ikx + 0 · e−ikx x → +∞ ,
in
ψ+ −→
1 · e+ikx + ρ · e−ikx x → −∞ ,
{
ρ′ · e+ikx + 1 · e−ikx x → +∞ ,
in
ψ− −→
0 · e+ikx + τ ′ · e−ikx x → −∞ ,
~2 ∂ ∂ ~2 ∂ 2 ~ ∂
L2 = − sin θ − , Lz = .
sin θ ∂θ ∂θ sin2 θ ∂ϕ2 i ∂ϕ
If Ylm denotes the spherical harmonic of angular momentum l and magnetic quantum number m,
the functions
ϕk,l,m (r, θ, ϕ) = jl (kr)Ylm (θ, ϕ)
~2 k 2
H0 ϕk,l,m = ϕk,l,m , L2 ϕk,l,m = ~2 l(l + 1) ϕk,l,m , Lz ϕk,l,m = ~m ϕk,l,m .
2m
The joint eigenspace Ek,l,m with these eigenvalues is one-dimensional: Ek,l,m = C·ϕk,l,m . Therefore,
since S commutes with each of H0 , L2 , and Lz , the function ϕk,l,m is an eigenfunction of S :
(In some sense, the present situation is simpler than that of d = 1, as the basis of functions ϕk,l,m
completely diagonalizes S.)
We now claim that the phases δl (k) are the phase shifts of Section 1.5. To verify this, we recall
that a scattering solution ψk,l,m (r, θ, φ) = Rk,l (r)Ylm (θ, φ) of the Schrödinger equation Hψk,l,m =
Eψk,l,m has the asymptotic behavior
( )
Rk,l (r) −→ (2ikr)−1 e2iδl (k) ei(kr−lπ/2) − e−i(kr−lπ/2) .
r→∞
[Warning: we have adjusted our overall phase convention!] This solution is a solution with
incoming-wave character in the sense that its radially incoming wave component e−ikr /r is ex-
actly the same as the corresponding component of the free solution. We therefore expect that
W− ϕk,l,m = ψk,l,m . Indeed, the right factor of W− = limt→−∞ eitH/~ e−itH0 /~ sends ϕk,l,m (or, rather
20
a localized wave packet formed by superposition of k-values in a narrow range) to an incoming
wave e−ikr /r at r = ∞ in the distant past, and the left factor then produces the full scattering
state with a phase-shifted radially outgoing wave component (but unchanged radially incoming
wave component).
We still have to figure out what happens to ψk,l,m = W− ϕk,l,m when the adjoint Moller operator
W+† is applied. We know that W+† ψk,l,m = Sϕk,l,m is a unitary number times ϕk,l,m , but what is
that number? To find it, we look at the radially outgoing wave component e2iδl ei(kr−lπ/2) /(2ikr) of
ψk,l,m . The operator W+† sends this component to r = ∞ by the full time evolution and then sends
it back in by the free time evolution. In this journey to infinity and back, no scattering takes
place. Therefore, whereas W− left the radially incoming wave component unchanged, it is the
radially outgoing wave component that remains unchanged under W+† . In this way, by comparing
expressions, we see that Sϕk,l,m = W+† ψk,l,m = e2iδl (k) ϕk,l,m .
Let us make a small remark about the condition in the title of this subsection. For this we recall
from basic quantum theory that if U and V are Hilbert spaces with Hermitian scalar products
⟨·, ·⟩U resp. ⟨·, ·⟩V , then the adjoint of a linear operator A : U → V is the linear operator
A† : V → U , v 7→ A† v defined by
⟨A† v, u⟩U = ⟨v, Au⟩V
for all u ∈ U . Now if H is the Hilbert space of our problem with Hamiltonian H, let Hsc ⊂ H
denote the subspace of scattering states (i.e., the orthogonal complement of the subspace of bound
states). The adjoint of the Moller operator W+ : H → Hsc then is an operator
W+† : Hsc → H .
In other words, domain(W+† ) = range(W+ ) = Hsc . The condition above can therefore be reformu-
lated as
range(W− ) ⊂ range(W+ ) . (1.27)
If the Hamiltonian H is time-reversal invariant (see the next subsection), then one can show the
identity
range(W− ) = range(W+ ) ,
so condition (1.27) is always satisfied in that case. According to a remark made after Definition
(1.17), it follows that the scattering operator S for a time-reversal invariant system is always
unitary: S † S = IdH and SS † = IdH .
21
1.7 Time reversal and scattering
We have already mentioned the fact that unitary symmetries of the Hamiltonian (U HU −1 = H
and U H0 U −1 = H0 ) give rise to unitary symmetries of the scattering operator (U SU −1 = S).
Here we will tell a related story, describing the consequences of time-reversal symmetry (not a
unitary symmetry; see below) for scattering.
We begin with a brief discussion of time reversal in classical mechanics. In a classical phase
space with position variables q and momenta p , the operation of inverting the time (called ‘time
inversion’ or ‘time reversal’ for short) is the anti-canonical transformation Tcl defined by
It is anti-canonical because it reverses the sign of the Poisson bracket. Clearly, time reversal is an
involution, which is to say that Tcl2 is the identity transformation. A classical Hamiltonian system
is called time-reversal invariant if the Hamiltonian function satisfies H = H ◦ Tcl , i.e.,
(p − eA)2
H(q, p) = ,
2m
where A is a vector potential for B. We observe that in the presence of a magnetic field, time-
reversal symmetry in the sense of (1.28) is broken. (Here we take the viewpoint of regarding the
magnetic field B as ‘external’ or fixed. Time reversal continues to be a symmetry even for B ̸= 0
if, along with transforming q and p , we also time-reverse B 7→ −B and A 7→ −A .)
The process of quantization is known to take canonical transformations of the classical phase
space into unitary transformations of the quantum Hilbert space, H. Now since time reversal fails
to be canonical in the classical theory, we should not expect it to be represented by a unitary
operator in the quantum theory.
Rather, time reversal in quantum mechanics will turn out to be an anti-unitary operator (the
definition is spelled out below). To motivate this fact, consider the time-dependent Schrödinger
equation (with position variables x and time variable t):
∂ ~2 2
i~ ψ(x, t) = − ∇ ψ(x, t) + V (x)ψ(x, t) .
∂t 2m
By taking the complex conjugate on both sides and inverting the time argument, we see that if
(x, t) 7→ ψ(x, t) is a solution of this equation, then so is (x, t) 7→ ψ(x, −t). We therefore expect
that the operator T of time reversal in the position representation H = L2 (R3 ) (and for the present
case of Schrödinger particles) is simply complex conjugation:
22
This is in fact true.
e t) in the
Problem. Deduce from (1.29) that the time-reversal operator on wave functions ψ(p,
e t) 7→ ψ(−p,
momentum representation is given by ψ(p, e t).
We now infer two properties of the time-reversal operator T which are independent of the
representation used. The first property,
T (zψ) = z̄ T ψ (z ∈ C , ψ ∈ H) , (1.30)
is called complex anti-linearity. It says that a complex number z goes past the time-reversal
operator T as the complex conjugate, z̄ . Notice that this property distinguishes T from the usual
type of complex-linear operator, say A, which obeys the commutation rule A z = zA .
A stronger consequence of the formula (1.29) is that T preserves the Hermitian scalar product
∫
′
⟨ψ, ψ ⟩ = ψ(x, t) ψ ′ (x, t) d3 x
R3
up to complex conjugation:
T 2 A T −2 = A
for any A . This implies that T 2 = z IdH where z is some complex number. Since T 2 is unitary,
the number z = eiα must be unitary.
The possible values of the unitary number z are further constrained by associativity of the
operator product T 2 · T = T · T 2 :
z T ψ = T 2 (T ψ) = T (T 2 ψ) = T (zψ) = z̄ T ψ .
It follows that our unitary number z = eiα has to be real. This leaves but two possibilities: z = 1,
and z = −1. We have already encountered a situation where z = 1. The other case of z = −1
also occurs in physics.
23
( )
ψ↑
Fact. The operator of time reversal on a spinor ψ = (i.e., the wave function of a particle
ψ↓
with spin 1/2) is realized by
in the position representation. (This fact will be explained in the chapter on Dirac theory.)
After this brief introduction to time reversal, we turn to the consequences of time-reversal
symmetry for scattering.
Definition. A quantum Hamiltonian system is called time-reversal invariant if the Hamiltonian
stays fixed under conjugation by the time reversal operator: H = T HT −1 .
We know that the scattering operator S is obtained by taking a limit of products of time
evolution operators. Therefore, we now look at what happens to time evolution operators under
−1
conjugation by T . By using the relations T (AB)T −1 = (T AT −1 )(T BT −1 ), T eA T −1 = eT A T and
T iT −1 = −i , we get
−1 )/~
T e−itH/~ T −1 = e+it (T HT ,
Let now the Hamiltonian H0 for free motion be time-reversal invariant as well. Then by the
( )−1
same calculation we have T e−itH0 /~ T −1 = e−itH0 /~ and hence
( )−1
T eitH0 /~ e−2itH/~ eitH0 /~ T −1 = eitH0 /~ e−2itH/~ eitH0 /~ .
T ST −1 = S −1 . (1.33)
Corollary. Assume that H0 has no bound states and that H0 and H are time-reversal invariant.
Then the scattering operator is unitary on the full Hilbert space: S † S = SS † = IdH . (This
conclusion is not tied to time reversal but holds if the pair H0 , H has any anti-unitary symmetry.)
In concrete applications one usually looks at matrix elements of the scattering operator in
certain subspaces with fixed quantum numbers. One then wants to understand the consequences
of time-reversal invariance at the level of matrix elements. In this endeavor it is possible to get
confused. Indeed, you might make the following (incorrect) argument. You might say that since
T for spinless particles is just complex conjugation, the result (1.33) implies that the scattering
matrix is symmetric: S t = S̄ † = S̄ −1 = T S̄T −1 = S. Copying from one of the standard textbooks,
you would write the scattering matrix, say for potential scattering in one dimension, as
( )
τ ρ′
S= .
ρ τ′
[You would probably argue that this, not (1.22), is the ‘correct’ way of arranging the scattering
matrix elements. After all, in the limit of vanishing potential, where we have zero reflection
24
ρ = ρ′ = 0 and full transmission, τ = τ ′ = 1, the scattering matrix should turn into the identity
?
matrix.] The symmetry S = S t of the scattering matrix would then seem to imply that ρ = ρ′ .
This is false. The correct statement is that τ = τ ′ due to time-reversal invariance.
What went wrong with our argument? The answer is that we were not careful enough to
translate the result (1.33) for the operator S into a correct statement about the matrix of S.
Problem. Get the argument straightened out to show that τ = τ ′ .
In the sequel we will explain how the notion of ‘symmetry’ of the scattering operator S can be
formulated in an invariant (or basis-free) manner. For this purpose we take a time-out in order to
review some basic linear algebra.
Let V be a vector space over the number field K = C or K = R . We recall that the dual, V ∗ ,
of V is the vector space of linear functions f : V → K . Let now L : V → W be a K-linear
mapping between two K-vector spaces V and W . The canonical transpose of L is the mapping
Lt : W ∗ → V ∗ defined by
(Lt f )(v) := f (Lv) .
We call it the ‘transpose’ because, if L (resp. Lt ) is expressed with respect to bases of V and W
(resp. the dual bases of V ∗ and W ∗ ), then the matrix of Lt is the transpose of the matrix of L.
Consider now the special case of W = V ∗ . One then has W ∗ = (V ∗ )∗ = V (for this, to be
precise, we should require the vector space dimension to be finite) and the canonical transpose of
L : V → V ∗ is another mapping Lt : V → V ∗ between the same vector spaces. In this situation
we can directly compare L with Lt and give a natural meaning to the word ‘symmetric’.
Definition. A linear mapping L : V → V ∗ is called symmetric if L = Lt . It is called skew if
L = −Lt .
Remark. In the case of V = W , there is no canonical definition of ‘symmetric’ linear map
L : V → V . (The matrix of L with respect to some basis of V may be symmetric, but this
property is not preserved by a change of basis in general.) To speak of a symmetric map in this
context, one needs an identification of V with V ∗ , e.g., by a non-degenerate quadratic form on V .
m: V →V∗ , v 7→ m(v) = p .
25
2. A rigid body in motion has an angular velocity ω ∈ so3 (where so3 ≃ R3 is the Lie algebra
of the rotation group SO3 fixing some point, e.g., the center of mass of the rigid body). The
angular momentum L of the body is an element L ∈ so∗3 of the dual vector space. The
pairing L(ω) computes twice the energy of rotational motion of the body. The tensor I of
the moments of inertia of the body is a symmetric linear mapping
called the Hall conductivity. Notice that any skew (linear) mapping L : V ∗ → V for
dimV = 3 (or any other odd dimension) must have a vector e which is a null vector, i.e.,
L(e) = 0. In the case of σH this vector e coincides with the axis of the magnetic field. Since
σH (E)(E ′ ) = −σH (E ′ )(E) = 0 vanishes for E = E ′ , the Hall part of the conductivity does
not contribute to the power.
After this list of examples, we continue our review of some basic linear algebra. Let now V be
a Hermitian vector space; in other words, V is a complex vector space (K = C) equipped with a
Hermitian scalar product ⟨·, ·⟩V . We then have a canonical anti-linear bijection
In the language of Dirac, this is called the ket-bra bijection, |v⟩ 7→ ⟨v|.
Definition. Let L : V → W be a complex linear mapping between two Hermitian vector spaces
V and W . The Hermitian adjoint L† : W → V is defined as the composition
c Lt c−1
L† = cV−1 ◦ Lt ◦ cW : W −→
W
W ∗ −→ V ∗ −→
V
V . (1.35)
26
1.7.2 T -invariant scattering for spin 0 and 1/2
We are now ready to describe in what sense the scattering operator S of a time-reversal invariant
system is symmetric. Let us associate with S : H → H a complex linear operator S̃ : H → H∗ by
Se = cH ◦ T ◦ S :
c
H∗ .
S H T
H −→ H −→ H −→ (1.36)
Fact. The scattering operator S of a time-reversal invariant system of particles with spin zero
(resp. spin 1/2) is symmetric (resp. skew) in the sense that Se = +Set (resp. Se = −Set ).
Proof. We evaluate Se on a pair of vectors ψ, ψ ′ ∈ H :
e
S(ψ)(ψ ′
) = ⟨T Sψ, ψ ′ ⟩ .
By using the relation (1.33) for a system with time-reversal invariance we obtain
⟨T Sψ, ψ ′ ⟩ = ⟨S −1 T ψ, ψ ′ ⟩ = ⟨T ψ, Sψ ′ ⟩ ,
where the second equality results from the unitarity S −1 = S † of the scattering operator. In the
next step we use the anti-unitary property (1.31) of T :
e ′ )(T 2 ψ) .
⟨T ψ, Sψ ′ ⟩ = ⟨T 2 ψ, T Sψ ′ ⟩ = ⟨T Sψ ′ , T 2 ψ⟩ = S(ψ
e
S(ψ)(ψ ′ e ′ )(ψ) .
) = ϵT S(ψ
Thus Se is symmetric for spinless particles (ϵT = 1) and skew for spin-1/2 particles (ϵT = −1).
Example. A current topic of active research are so-called topological insulators with strong spin-
orbit scattering and time-reversal invariance in d = 2 or d = 3 dimensions. The one-dimensional
boundary of such an insulator in d = 2 may house a propagating spin-1/2 mode with one right-
moving and one left-moving component. Skew symmetry constrains the scattering matrix for such
a two-component mode to be of the form
( )
0 eiϑ
Se = .
−eiϑ 0
This means that the reflection coefficient vanishes identically and the absolute value of the trans-
mission coefficient τ = eiϑ always remains unity as the length of the one-dimensional boundary is
increased. (Only the phase ϑ changes). Remarkably, this ‘perfectly conducting’ property of the
boundary channel is stable with respect to the introduction of any kind of (time-reversal invariant)
disorder.
27
2 Relativistic quantum mechanics: Dirac equation
2.1 Motivation
We now turn to a very fundamental theme in relativistic quantum mechanics and quantum field
theory: the Dirac equation. Its particular importance derives from the fact that all known elemen-
tary particles of matter (the so-called leptons and quarks), are described by the Dirac equation or
its quantum field-theoretic extensions. The elementary particle we have in mind here, in particular,
is the electron.
We begin our motivation with the relativistic equation relating the positive energy E of a free
particle of mass m to its momentum p :
√
E= (mc2 )2 + (pc)2 . (2.1)
In attempting to reconcile quantum mechanics with special relativity, one looks for a relativistic
wave equation with the property that this relation is reproduced. By the quantum-theoretic
correspondences E ↔ i~ ∂/∂t and p ↔ ~∇/i , a first proposal for a relativistic wave equation (of
the electron, say) might be
∂ψ ? √
i~ = (mc2 )2 − (~c)2 ∇2 ψ .
∂t
It is, however, not clear how to make mathematical sense of the square root of an expression
involving the Laplacian ∇2 . Defining the square root by its power series, one would end up with
a ‘non-local’ differential equation (involving derivatives up to arbitrarily high order).
To avoid the problems caused by taking a square root, an alternative approach might be to
start from the energy-momentum relation (2.1) in the squared form
By using again the correspondences E ↔ i~ ∂/∂t and p ↔ ~∇/i one gets the wave equation
∂2
−~2 ψ = (mc2 )2 ψ − (~c)2 ∇2 ψ , (2.3)
∂t2
which is known as the Klein-Gordon equation. The continuity equation naturally associated to it
∫
is ρ̇ + divj = 0 (leading to a conservation law ρ d3 x = const) with
~ ( ) i~ ( )
j= ψ ∇ψ − ∇ψ ψ , ρ= 2
ψ ∂ t ψ − ∂t ψ ψ (∂t ≡ ∂/∂t) . (2.4)
2im 2mc
~ω
For solutions with time dependence ψ ∼ e−iωt and positive frequency ω one has ρ = mc2
|ψ|2 ≥0
but for ψ ∼ e+iωt the same quantity becomes negative. This means that ρ d3 x in the present case,
unlike the Schrödinger case, cannot be interpreted as a probability density. (By the way: in a
quantum-field theoretic setting ρ d3 x does have an interpretation as a charge density.)
It is not difficult to understand why the positivity of ρ is lacking for the Klein-Gordon equation:
it is because the expression for ρ contains the time derivative ∂t . This, in turn, is a consequence
of the Klein-Gordon equation being of second order in ∂t . (The Schrödinger equation, in contrast,
is of first order in ∂t .)
28
2.2 Dirac equation
The lesson learned from the Klein-Gordon equation is that, in trying to construct a relativistic
generalization of the Schrödinger equation, one should retain the first-order-in-∂t nature of the
equation. Special relativity then suggests that the new equation should also be of first order in
the spatial derivatives ∂/∂xj . In fact, Dirac (1928) pioneered the idea of looking for a first-order
differential operator
∑
3
~ ∂
2
D = βmc + αj pj c , pj = , (2.5)
j=1
i ∂xj
with algebraic objects β, α1 , α2 , α3 that remain to be specified. If these satisfy the relations
then D squares to
∑3
∂2
D = (mc ) − (~c) ∇ ,
2 2 2 2 2
∇ = 2
. (2.7)
j=1
∂x2j
∂ 2ψ
−~2 = D2 ψ . (2.9)
∂t2
By using the formula (2.7) for D2 we see that the latter is nothing but the Klein-Gordon equation
(2.3). Thus for plane wave solutions ψ of (2.8) with frequency ω = E/~ and wave vector k = p/~
one gets the desired energy-momentum relation (2.1). (At the same time, one gets expressions for
ρ and j of a more desirable form; see below.)
The question now is whether one can realize the algebraic relations (2.6), and if so, how. It
is certainly impossible to satisfy these relations while clinging to the Schrödinger viewpoint of
complex numbers β, . . . , α3 multiplying a wave function ψ with values in C.
Therefore, following Dirac we now abandon the Schrödinger viewpoint and allow that ψ may
take values in a more general vector space, say Cn with n ≥ 1. With that generalization, we can
take β, . . . , α3 to be n × n matrices multiplying the n-component vector ψ. It then turns out to
be possible to realize the relations (2.6) for n ≥ 4. Indeed, one possible choice for n = 4 is
( ) ( )
1 0 0 σj
β= , αj = , j = 1, 2, 3, (2.10)
0 −1 σj 0
Problem. Check that the choice (2.10) satisfies the relations (2.6).
29
We will elaborate on the theoretical background behind the relations (2.6) in a later subsection.
For now, we record that there exists at least one possible realization for n = 4. In our further
arguments, we will often refer to this realization for concreteness.
Definition. The Dirac equation for a free particle of mass m reads
~c ∑
3
∂ψ ∂ψ
i~ = mc2 βψ + αj , (2.11)
∂t i j=1 ∂xj
where ψ(x, t) takes values in C4 and the 4 × 4 matrices β, α1 , α2 , α3 are subject to (2.6).
With these conventions, the Dirac equation (2.11) takes the relativistic form
( ∑3 )
0 ∂ j ∂ mc
γ + γ +i ψ=0. (2.13)
∂x0 j=1 ∂xj ~
The parameter mc/~ has the physical dimension of an inverse length. Its reciprocal ~/(mc) is
called the (reduced) Compton wave length. For the electron with mass m ≈ 0.5 MeV/c2 one has
~ ~c 200 MeV · fm
= 2
≈ ≈ 400 × 10−15 m . (2.14)
mc mc 0.5 MeV
In the jargon of physics the matrices γ 0 , . . . , γ 3 are called the gamma matrices. We note that for
the choice (2.10) they have the expressions
( ) ( )
0 1 0 j 0 σj
γ = , γ = , j = 1, 2, 3. (2.15)
0 −1 −σj 0
It is customary to use Greek letters for space-time indices; i.e., µ = 0, 1, 2, 3.
Problem. Adopting this notation, show that the algebraic relations (2.6) take the concise form
γ µ γ ν + γ ν γ µ = 2g µν 1 , (2.16)
g 00 = 1 , g 11 = g 22 = g 33 = −1 , g µν = 0 for µ ̸= ν , (2.17)
30
2.4 Non-relativistic reduction
Dirac’s theory is intended to be a relativistic quantum theory of the electron. We already have
a non-relativistic quantum theory of the electron, namely the Schrödinger equation or, including
spin, the Pauli equation. By the principles of theory building, in order for a new theory to be
acceptable it must be consistent with the old theory which is already known to be true (within its
limits of validity). Therefore the logical step to be taken next is to verify that the Dirac equation
reduces to the Schrödinger/Pauli equation in the non-relativistic limit.
For this purpose we write the Dirac equation in the following block-decomposed form:
( ∑ )( ) ( )
i mc + 1 ∂
~∑ c ∂t
σj
∂
∂xj ψ+ ψ±,↑
0= , ψ± = . (2.19)
− σj ∂x∂ j i mc
~
− 1c ∂t
∂
ψ− ψ±,↓
At present, the symbols ↑ and ↓ are just some fancy notation to label the two components of
ψ± (x, t) ∈ C2 . (Later we will see that they do, in fact, reflect the spin of the electron.)
The non-relativistic limit is |v| ≪ c , or |k| ≪ mc/~ . By the correspondence k ↔ ∇/i
this means that the off-diagonal blocks of the matrix in (2.19) are to be considered as being much
smaller than the diagonal blocks. In zeroth-order approximation we neglect the off-diagonal blocks
altogether to obtain ( )
mc 1 ∂ (0) (0)
i + ψ+ = 0 , ψ− = 0 . (2.20)
~ c ∂t
(0)
We are setting ψ− = 0 by fiat because we intend to identify ψ+ with the spinor wave function of
the Pauli equation, and there is no room for additional degrees of freedom in the non-relativistic
limit. (We will learn later that ψ− describes the positron, the antiparticle of the electron. In the
(0)
present context, we envisage a situation with no positrons present. Hence our choice ψ− = 0.)
(0)
Note that the first equation in (2.20) implies that ψ+ has the time dependence
ψ+ ∼ e−imc
(0) 2 t/~
.
We now turn to a first-order (or improved) approximation. For this we write the system of
equations (2.19) in the form ( )( )
A B ψ+
=0,
C D ψ−
or equivalently,
Aψ+ + Bψ− = 0 , Cψ+ + Dψ− = 0 .
(A − BD−1 C) ψ+ = 0 ,
or explicitly,
( ) (∑ )( )−1 (∑ )
mc 1 ∂ ∂ mc 1 ∂ ∂
i + ψ+ + σj i − σj ψ+ = 0 .
~ c ∂t ∂xj ~ c ∂t ∂xj
31
Now since we know from the zeroth-order approximation that ψ+ has the leading time dependence
ψ+ ∼ e−imc ψ+ , and we may replace −D−1 Cψ+ by
(0)
, we have − 1c ∂t ψ+ ≈ i mc
2 t/~ ∂
~
( )−1 (∑ ) (∑ )
−1 mc 1 ∂ ∂ ~ ∂
−D Cψ+ = i − σj ψ+ → σj ψ+ .
~ c ∂t ∂xj 2imc ∂xj
By multiplying the equation for ψ+ by i~c and slightly rearranging the terms, we then arrive at
(1) (1)
the improved approximation ψ+ ≈ ψ+ where ψ+ satisfies
( )2
~2 ∑
(1)
∂ψ+ 2 (1) ∂ (1)
i~ = mc ψ+ − σj ψ+ .
∂t 2m ∂xj
In the final step we use the relations σj σl + σl σj = 2δjl for the Pauli matrices to find
(∑ )2 ∑ 2
∂ ∂
σj = = ∇2 .
∂xj ∂x2j
Altogether we have
∂ (1) (1) ~2 2 (1)
i~ψ+ = mc2 ψ+ − ∇ ψ+ . (2.21)
∂t 2m
This is indeed the free-particle Schrödinger equation with a constant shift of the energy by the
rest mass mc2 . Thus the Dirac equation has passed its first test.
Note added. At the same level of approximation, the equation ψ− = −D−1 Cψ+ yields
(∑ )
(1) ~ ∂ (1)
ψ− = σj ψ+ .
2imc ∂xj
(1) (1)
Thus ψ− is smaller than ψ+ by, roughly speaking, a factor of ~|k|/mc or |v|/c .
32
of the matter field ψ . It is easy to see that the expression
( )
~ ∂
− eAµ ψ
i ∂xµ
is gauge invariant in this sense. The principle of minimal substitution then tells us to enforce
gauge invariance by making in the Dirac equation (or any other charged quantum wave equation,
for that matter) the substitution
( )
∂ ∂ ie
ψ→ − Aµ ψ . (2.25)
∂xµ ∂xµ ~
Doing so, we arrive at the final form of the Dirac equation.
Definition. The Dirac equation for a particle of mass m and charge e in the presence of an
electromagnetic field (described by the gauge field Aµ ) is
( )
∂ ie mc
γ µ
− Aµ ψ + i ψ = 0 . (2.26)
∂x µ ~ ~
Problem. By following the steps of Section 2.4, show that the full Dirac equation (2.26) in the
non-relativistic limit reduces to the Pauli equation (i.e., the Schrödinger equation including the
Pauli coupling of the spin of the charged particle to the magnetic field B):
( )2
∂ (1) (1) ~2 ∑ ∂ ie (1) e~ ∑ (1)
i~ ψ+ = (mc + eΦ)ψ+ −
2
− Aj ψ+ − Bj σj ψ+ . (2.27)
∂t 2m j ∂xj ~ 2m j
Here Φ = −cA0 is the electric scalar potential, and Aj are components of the magnetic vector
potential A obeying curl A = B .
ψ has four components, as we recall, and thus takes values in C4 . We now assume that the vector
space C4 is Hermitian, i.e., is equipped with a Hermitian scalar product C4 × C4 → C . Using this
structure we define the Hermitian adjoint ψ † with values in the dual vector space (C4 )∗ , and ψ † ψ
with values in C (actually, R).
33
Now observe that the matrices β and αl in (2.10) are Hermitian: β = β † and αl = αl†
(l = 1, 2, 3). We promote this observation to an axiom of the theory, i.e., we demand that any
permissible choice of β and αl must not only obey the algebraic relations (2.6) but must also have
the property of being Hermitian. By dualizing the equation (2.28) we then obtain the following
equation for ψ † :
( ) ∑( ∂ )
∂ ie † ie mc2 †
− Φ ψ +c + Al ψ † αl − i ψ β=0. (2.29)
∂t ~ ∂xl ~ ~
Next we contract the equation (2.28) for the vector ψ with the dual vector ψ † , and similarly the
equation (2.29) for ψ † with ψ . Afterwards we add the two resulting scalar equations. The terms
√
containing i = −1 all cancel since their signs are changed by taking the Hermitian adjoint. So
we get
∂ † ∑ ∂
ψ ψ+c ψ † αl ψ = 0 . (2.30)
∂t l
∂x l
ρ := ψ † ψ , jl := c ψ † αl ψ . (2.31)
Summary. We record that if ψ is a solution of the Dirac equation (with or without electromag-
netic field), then the scalar ρ = ψ † ψ and the vector j with components jl = c ψ † αl ψ satisfy the
continuity equation
ρ̇ + div j = 0 . (2.32)
By a standard argument using the divergence theorem (a.k.a. Gauss’ theorem) it follows that the
∫
total space integral ρ d3 x is conserved.
Remark. Coming from Schrödinger quantum mechanics, it is natural to think that ρ (actually,
ρ d3 x) is the probability density for a relativistic electron, and j is the vector of the corresponding
probability current density. However, it will turn out that this (wishful) thinking is untenable.
Dirac’s theory in fact will have to be reformulated (in the framework of quantum field theory)
so as to give ρ the interpretation of charge density of the electron (actually of the quantum field
encompassing the electron as well as the positron).
34
2.7 Clifford algebra
In this and the following subsection we provide some theoretical background concerning the four-
component nature of the wave function ψ of the Dirac equation. A question which was left open
in Section 2.3 is this: how can we say a priori that the algebraic relations
γ ν γ ν + γ ν γ µ = 2g µν
are realizable for n×n matrices with n ≥ 4 , and how can such a matrix realization be constructed?
To answer this question, we shall take the liberty of going into more detail than is offered in most
physics textbooks, as the very same formalism will turn out to be relevant for the procedure of
second quantization of many-particle quantum mechanics.
In the sequel we will be concerned with a vector space V over the real number field K = R
or the complex number field K = C . We assume that V comes with a non-degenerate symmetric
K-bilinear form (also referred to as a ‘quadratic form’ for short)
Examples. We give two examples, the number field being K = R in both cases. The first example
is the Euclidean vector space V = R3 equipped with the Euclidean scalar product Q,
The second example is the example of relevance for the Dirac equation: the Lorentzian vector
space V = R4 with the Minkowski scalar product Q given by (summation convention!)
vw + wv = 2Q(v, w) . (2.35)
35
Remark. The words ‘associative algebra generated by V ⊕ K’ mean that the elements of Cl(V, Q)
are polynomial expressions in the elements of V with coefficients taken from the number field K .
The relations (2.35) imply, e.g.,
Example. Let (V = R4 , Q) be the Lorentzian vector space with Minkowski scalar product Q.
Take {e0 , e1 , e2 , e3 } to be some standard basis of V , so that Q(eµ , eν ) = gµν where g00 = −g11 =
−g22 = −g33 = 1. Then some examples of elements in Cl(V, Q) are
e0 e0 = 1 , e1 e1 = −1 , e0 e1 = −e1 e0 , e0 e1 e0 = −e1 .
All this may seem abstract and unfamiliar. It can be made more tangible as follows.
Let us think of the vectors e0 , . . . , e3 as basis elements in some four-dimensional subspace
V ≃ R4 of the linear space (or vector space) of 4 × 4 matrices with quadratic form
Because we have realized our vectors v = v µ eµ as matrices, the vector space V ≃ R4 now carries
the additional structure of an associative algebra with matrix multiplication playing the role of
the product. If Id denotes the 4 × 4 unit matrix, one sees that the relations
vw + wv = 2Q(v, w) Id (2.38)
Count the number of these elements to deduce that dim Cl(V, Q) = 2 dim V .
36
2.8 Spinor representation
While the Clifford algebra Cl(V, Q) is defined abstractly as an associative algebra with certain
binary relations, we saw that Cl(V, Q) has a concrete realization by 4×4 matrices for the Minkowski
example of V = R4 . This realization will be called the spinor representation of Cl(R4 , Q) from
now on. In Section 2.7 we simply wrote down the spinor representation without explaining where
it came from. This gap in our theoretical development is to be closed in the present subsection.
For brevity we will only consider the special case of a real vector space V ≃ R2n of even
dimension with Euclidean scalar product Q. Given this situation, let us fix some orthonormal
basis {e1 , . . . , en , f1 , . . . , fn } of V .
Although everything so far is real (K = R), we are now going to consider complex linear
combinations of the vectors v ∈ V ; mathematically speaking we pass to the complexification
V ⊗R C of V . (The symbol ⊗K denotes the tensor product of vector spaces over K ; see the end
of this subsection for a quick exposition.) In the complexification V ⊗R C we then form the linear
√
combinations (i = −1 )
Such a decomposition is called a polarization. [The notation indicates that P ∗ may be regarded
as the dual vector space of P via the non-degenerate pairing P ⊗ P ∗ → C by (c, c∗ ) 7→ Q(c, c∗ ).]
Notice that the Clifford relations (2.35) imply the following relations for our new generators:
In physics, these relations are called the canonical anti-commutation relations (CAR). They will
play a central role in the formalism of second quantization for many-fermion systems.
To formulate the algorithm for constructing the spinor representation of Cl(V, Q), we need one
more algebraic concept.
Definition. By the exterior algebra ∧(U ) of a K-vector space U one means the associative algebra
generated by U ⊕ K with relations
⊕U
dim
∧(U ) = ∧l (U ) , (2.45)
l=0
37
∑
where ∧0 (U ) ≡ K , ∧1 (U ) ≡ U , ∧2 (U ) consists of all quadratic elements ui uj , ∧3 (U ) consists
∑
of all cubic elements ui uj uk , and so on.
Remarks. i) Note that (2.44) implies u2 = 0 for any u ∈ U ⊂ ∧(U ). ii) In physics the exterior
algebra ∧(U ) is also known as the fermionic Fock space of U . (More precisely, if U is a Hilbert
space, then ∧(U ) is another Hilbert space called the fermionic Fock space associated to the single-
particle Hilbert space U .) In the Fock space setting the degree has the physical meaning of particle
number. iii) The notion of exterior algebra is basic to the calculus of differential forms.
Problem. Show that the exterior algebra ∧(U ) is isomorphic as a vector space to the Clifford
algebra Cl(U, Q). Hint: prove by construction of a basis that ∧(U ) has dimension 2 dim U .
We now recall the definition of the complex vector space P ∗ in (2.41) and make the important
observation that the exterior algebra of P ∗ (or P for that matter) is contained as a subalgebra in
Cl(V, Q) ⊗ C . By this observation the exterior algebra ∧(P ∗ ) can be turned into a representation
space for Cl(V, Q) as follows.
We are looking for an operation of Cl(V, Q) on ∧(P ∗ ),
which is a representation. In other words, we want a Clifford multiplication rule (2.46) which is
compatible with the associative product in the Clifford algebra in the sense that (i) the relation
(ab) · ξ = a · (b · ξ) (2.47)
holds for all a, b ∈ Cl(V, Q), ξ ∈ ∧(P ∗ ), and (ii) the Clifford relations (2.35) are preserved, i.e., for
all v, w ∈ V ⊂ Cl(V, Q), ξ ∈ ∧(P ∗ ), we have
v · (w · ξ) + w · (v · ξ) − 2Q(v, w) · ξ = 0 . (2.48)
v · ξ := vξ , (2.49)
i.e., the product is computed in the exterior algebra ∧(P ∗ ). Note that the action of v ∈ P ∗
increases the degree: it maps ∧l (P ∗ ) into ∧l+1 (P ∗ ).
Second, let v ∈ P . Such elements act by an operation which lowers the degree:
P ∋ v : ∧l (P ∗ ) → ∧l−1 (P ∗ ) . (2.50)
38
This operation is defined recursively. For the two lowest degrees l = 0, 1 one sets
l=0: v · 1 := 0 , (2.51)
l=1: v · v ′ := 2Q(v, v ′ ) . (2.52)
The action on higher-degree elements is then defined by the so-called anti-Leibniz rule:
v · (w · ξ) + w · (v · ξ) − 2Q(v, w) · ξ
( )( ) ( )( )
= ε(vP ∗ ) + ι(vP ) ε(wP ∗ ) + ι(wP ) ξ + ε(wP ∗ ) + ι(wP ) ε(vP ∗ ) + ι(vP ) ξ − 2Q(v, w)ξ
= ε(vP ∗ )ι(wP )ξ + ι(wP )ε(vP ∗ )ξ + ε(wP ∗ )ι(vP )ξ + ι(vP )ε(wP ∗ )ξ − 2Q(v, w)ξ
= 2Q(vP ∗ , wP )ξ + 2Q(vP , wP ∗ )ξ − 2Q(v, w)ξ = 0 .
Thus the condition (2.48) is indeed satisfied and our Claim is true. What we have learned is
summarized in the next statement — where we keep the assumption of even dimension of V but
drop the (unnecessary) condition that Q be a Euclidean structure.
Definition. Let (V, Q) be an even-dimensional vector space over the reals R with polarization
V ⊗R C = P ⊕ P ∗ . The spinor representation of the Clifford algebra Cl(V, Q) is defined to be the
representation on the exterior algebra ∧(P ∗ ) which is given by the action (2.54).
It is now straightforward to produce a matrix representation of Cl(V, Q) by fixing a basis of
∧(P ∗ ) and expanding the action of the Clifford algebra elements w.r.t. this basis.
39
Example. We illustrate the procedure at the simple example of the Euclidean plane V = R2
with orthonormal basis {e, f }. As before, let c := (e − if )/2 and c∗ := (e + if )/2. The action of
c∗ and c on the basis {1, c∗ } of ∧(P ∗ ) is computed to be
and
c · 1 ≡ ι(c)1 = 0 , c · c∗ ≡ ι(c)c∗ = 2Q(c, c∗ ) = 12 Q(e − if, e + if ) = 1 .
Thus if we make the following identifications for the basis vectors {1, c∗ } of ∧(P ∗ ):
( ) ( )
0 ∗ 1
1≡ , c ≡ ,
1 0
then the Clifford algebra elements 1, e, f , and ef are represented by the matrices
( )
1 0
matrix(1) = ,
0 1
( )
∗ 0 1
matrix(e) = matrix(c + c ) = = σ1 ,
1 0
( )
∗ 0 −i
matrix(f ) = matrix(ic − ic ) = = σ2 ,
i 0
( )
i 0
matrix(ef ) = = iσ3 .
0 −i
Problem. By following the same procedure, construct the spinor representation of the Clifford
algebra Cl(V, Q) for the Euclidean vector space V = R4 with orthonormal basis {e1 , e2 , f1 , f2 }.
We finish this subsection with a few final comments:
1. The same construction applied to the Lorentzian vector space V = R4 with Minkowski scalar
product gives the so-called Weyl form of the gamma matrices:
( ) ( )
0 0 1 j 0 σj
γ = , γ = (j = 1, 2, 3). (2.56)
1 0 −σj 0
2. The spinor representation of Cl(R2n , Q) is the one and only representation of Cl(R2n , Q)
which is both non-trivial and irreducible.
3. We recall the dimension formulas dim Cl(V, Q) = 2 dim V and dim ∧ (P ∗ ) = 2 2 dim V . For the
1
1
case of four dimensions (V = R4 ) it so happens that the dimension 2 2 dim V = 22 of the spinor
representation is also four. This coincidence of dimensions 4 = 4 is an accident.
In the construction of the spinor representation we used the notion of tensor product, which may
be unfamiliar to the students of this course. Therefore we add a few words of explanation.
We assume that the notion of direct product (of sets, groups, vector spaces, manifolds, etc.)
is understood. Now in the case of two K-vector spaces V and W one has a variant of the direct
40
product V × W which is called the tensor product (over K) and is denoted by V ⊗K W ≡ V ⊗ W .
The tensor product is distinguished by the feature that the scalars k ∈ K move between factors:
kv ⊗ w = v ⊗ kw , (2.57)
I : W ⊗ V ∗ → Hom(V, W ) ,
which is given by
I(w ⊗ f )(v) := f (v) w .
Note that I(kw ⊗ f )(v) = f (v) kw = k f (v) w = I(w ⊗ kf )(v), so the tensor product W ⊗ V ∗ (not
the direct product W × V ∗ ) in fact is the proper product to appear in this isomorphism.
A corollary of this isomorphism is the fact that every matrix of size m × n can be decomposed
uniquely (up to an ambiguity due to scalar factors moving in the tensor product) as a sum of terms
each of which is the tensor product (sometimes called a ‘dyadic’ product in physics textbooks) of
an m-component column vector with an n-component row vector:
Let now f be a mapping from M to N . Given the G-actions on M and N , there exists an induced
action of G on such mappings f : M → N by
(g • f )(x) := g · f (g −1 · x) . (2.59)
41
Problem. Show that one has (gh) • f = g • (h • f ).
Now recall from special relativity that the Lorentz group G = SO1,3 acts on space-time vectors
v ∈ R4 by linear transformations v 7→ g · v which preserve the Minkowski scalar product. A
(complex) scalar field is a function f : R4 → C . The Lorentz group acts on it by
(g • f )(v) = f (g −1 · v) . (2.60)
The wave function of the Dirac equation is neither a scalar nor a vector field but a spinor field
ψ : R4 → ∧(P ∗ ), P = C2 . In order to establish the relativistic covariance of the Dirac equation,
we will have to understand how the Lorentz group acts on spinors ξ ∈ ∧(C2 ).
Remark. Every associative algebra can be given the structure of a Lie algebra by taking the
Lie bracket to be the commutator: [X, Y ] := XY − Y X. The Jacobi identity is automatically
satisfied in this case.
Example. The Euclidean vector space R3 equipped with the standard vector product (or cross
product)
R3 × R3 → R3 , (v, v ′ ) 7→ v × v ′ = −v ′ × v ,
is a Lie algebra isomorphic to so3 , the Lie algebra of the group of proper rotations of R3 . The
Jacobi identity in this case reads
v × (v ′ × v ′′ ) = (v × v ′ ) × v ′′ + v ′ × (v × v ′′ ) .
42
Next we recall that a Lie group G is a group with the structure of a differentiable manifold
such that the product G×G → G is a differentiable mapping. Every Lie group G comes with a Lie
algebra g , which has the geometric meaning of tangent space at the neutral element of G. In the
cases of interest to us, one gets for each X ∈ g a one-parameter subgroup of G by exponentiation,
R ∋ t 7→ etX . The Lie bracket of g then results from the differentiable product of G by
∂ 2 sX tY −sX −tY
[X, Y ] := e e e e . (2.64)
∂s∂t s=t=0
which leave Q invariant. The special orthogonal group SO(V, Q) ⊂ O(V, Q) is the connected
component containing the neutral element.
Examples. i) For V = R3 with Euclidean scalar product Q one has SO(V, Q) = SO3 , the group
of proper rotations of the Euclidean vector space R3 . ii) If V = R4 is equipped with Minkowski
scalar product Q, then SO(V, Q) is the group of Lorentz transformations of V . .
We now consider the Lie algebra so(V, Q) of the Lie group SO(V, Q). By setting gt := etX and
linearizing the condition Q(gt v, gt v ′ ) = Q(v, v ′ ) at the neutral element t = 0 we get the following
characterization of so(V, Q) :
Problem. Verify from this definition that so(V, Q) is closed under the commutator, i.e., if both
X and Y are in so(V, Q), then so is [X, Y ] := XY − Y X.
The elements of the Lie algebra so(V, Q) are skew-symmetric in the following sense. Using
that Q is non-degenerate, one has an isomorphism I : V → V ∗ by I(v) = Q(v, ·). One then
associates with every element X ∈ so(V, Q) a linear mapping X̃ : V → V ∗ by X̃ := I ◦ X. By
the condition (2.66) on X this map is skew: X̃ t = −X̃.
In the case of the Euclidean vector space V = R3 with Cartesian basis {e1 , e2 , e3 } the matrix
of X ∈ so(V, Q) = so3 has the form
0 X12 X13
matrix(X) = −X12 0 X23 = matrix(X̃) .
−X13 −X23 0
In the case of the Lorentzian vector space V = R4 with Minkowski scalar product Q, the matrices
of X ∈ so(V, Q) and X̃ ∈ Hom(V, V ∗ ) with respect to a standard basis {e0 , e1 , e2 , e3 } look as
follows:
0 X01 X02 X03 0 X01 X02 X03
X01 0 X12 X13 −X01 0 −X12 −X13
matrix(X) =
X02 −X12
, matrix(X̃) = .
0 X23 −X02 X12 0 −X23
X03 −X13 −X23 0 −X03 X13 X23 0
43
We now return to considering the Clifford algebra Cl(V, Q) and in it the subspace Cl2 (V, Q)
of skew-symmetrized elements of degree two:
∑
Cl2 (V, Q) := {X ∈ Cl(V, Q) | X = (uj vj − vj uj ) ; uj , vj ∈ V } . (2.67)
j
We will now proceed by explicitly constructing the claimed Lie algebra isomorphism. For this
∑
purpose we observe that the commutator of X = (uj wj − wj uj ) ∈ Cl2 (V, Q) with v ∈ V ⊂
Cl(V, Q) is another element [X, v] ∈ V ⊂ Cl(V, Q) :
∑ ∑ ( )
[X, v] = 2 [uj wj , v] = 4 Q(wj , v) uj − Q(uj , v) wj ∈ V .
j j
τ (X)τ (Y )v − τ (Y )τ (X)v = [X, [Y, v]] − [Y, [X, v]] = [[X, Y ], v] = τ ([X, Y ])v .
Now let {u, v} := uv + vu denote the anti-commutator and do the following calculation:
In the next to last equality we used a kind of generalized Jacobi identity, which is easily checked
to be correct. Thus we see that the condition (2.66) for τ (X) to be in so(V, Q) is satisfied.
Now the representation τ : Cl2 (V, Q) → so(V, Q) is injective (we leave the verification of this
statement as a problem for the student). By the equality (2.68) of dimensions it follows that τ is
bijective and hence an isomorphism of Lie algebras.
Summary. The subspace Cl2 (V, Q) of skew-symmetrized degree-two elements of the Clifford
algebra Cl(V, Q) has the structure of a Lie algebra. We have two representations for it:
1. As a subspace of Cl(V, Q) the Lie algebra Cl2 (V, Q) acts on spinors ξ ∈ ∧(P ∗ ) by the spinor
1
representation. This representation has dimension 2 2 dim V .
2. Via the isomorphism τ the Lie algebra Cl2 (V, Q) acts on vectors v ∈ V by the fundamental
(or defining) representation of so(V, Q). The dimension of this representation is dim V .
44
2.11 Spin group
By exponentiating a Lie algebra g (which sits inside an associative algebra A or acts on some
representation space R by linear operators or matrices, so that products such as X n for X ∈ g
make sense) one gets a Lie group G :
exp X2 X3 Xn
g −→ G , X 7→ eX := 1 + X + + + ... + + ... .
2! 3! n!
(To ensure that the exponential series converges, one poses the requirement that the associative
algebra A or the representation space R be finite-dimensional, so that the Lie algebra elements
are represented by matrices of finite size.) In order to verify that exponentiation of a Lie algebra
really does yield a group, one argues on the basis of the so-called Baker-Campbell-Hausdorff series:
1 1 1
eX eY = eX+Y + 2 [X,Y ]+ 12 [X,[X,Y ]]+ 12 [[X,Y ],Y ]+... , (2.70)
which suggests that the product eX eY of two exponentiated Lie algebra elements is the exponential
of another Lie algebra element X +Y + 12 [X, Y ]+. . . . (There indeed exists a recursive construction
by which the infinite Baker-Campbell-Hausdorff series can be shown to exist and converge in the
finite-dimensional case.)
Example. By exponentiating the Lie algebra so(V, Q) in the associative algebra End(V ) we get
the Lie group SO(V, Q).
Definition. The Lie group obtained by exponentiating the Lie algebra Cl2 (V, Q) in the Clif-
ford algebra is called the spin group Spin(V, Q). The representation which results from letting
Spin(V, Q) ⊂ Cl(V, Q) act on ∧(P ∗ ) is called the spinor representation of Spin(V, Q).
Remark. Since Cl2 (V, Q) is isomorphic as a Lie algebra to so(V, Q), one might think that
Spin(V, Q) would be isomorphic as a Lie group to SO(V, Q). However, as we shall see, this is
not the case.
We now show for the case of even-dimensional V = R2n and Euclidean Q that the group
Spin(V, Q) contains an element g = −1. For this we fix any pair e, f ∈ V of orthogonal unit
vectors and consider the element X := 12 (ef − f e) ∈ Cl(V, Q). For any parameter t ∈ C we have
Thus, in particular,
eπX = −1 .
45
It follows (for Euclidean V = R2n ) that if g is an element of Spin(V, Q), then so is −g.
Next we consider the action of g = eX ∈ Spin(V, Q) ⊂ Cl(V, Q) on v ∈ V ⊂ Cl(V, Q) by
conjugation in the Clifford algebra:
1 1 ∑ τ (X)l
∞
X −X
e ve = v + [X, v] + [X, [X, v]] + [X, [X, [X, v]]] + . . . = v = eτ (X) v .
2! 3! l=0
l!
Note that since τ (X) is in so(V, Q), the exponential eτ (X) is an element of the special orthogonal
group SO(V, Q). Thus we have a mapping
This mapping is a representation of Spin(V, Q) on the vector space V . The situation is summarized
in the following commutative diagram (commutative here says that exp ◦τ = ρ ◦ exp):
by invertible transformations (GL = general linear group) of the exterior algebra ∧(P ∗ ) ≃ ∧(C2 ) ≃
C4 , and the vector representation
46
The gamma matrices γ µ appearing in the Dirac equation differ from the γµ by the position of the
index, which is raised or lowered by using the Minkowski metric: γµ = Q(eµ , eν )γ ν =: Qµν γ ν and
γ µ = Qµν γν where Qµν Qνλ = δλµ . We therefore have
σ(g)γ µ σ(g)−1 = ρ(g)γ µ = Qµν ρ(g)γν = Qµν Qλτ γ τ ρ(g)λν = ρ(g −1 )µν γ ν . (2.75)
We now introduce a notion similar to that of scalar field and vector field.
Definition. A spinor field is a mapping ψ : R4 → ∧(P ∗ ) ≃ C4 from space-time (viewed as a
vector space after fixing a space-time origin) into the spinor representation space. It transforms
under the spin group as
The next statement has the implication that the wave function ψ of the Dirac equation is to
be viewed as a spinor field.
Proposition. If ψ is a solution of the Dirac equation with gauge potential A = Aµ dxµ , then g · ψ
for g ∈ Spin1,3 is a solution of the same equation with transformed gauge potential
We multiply the equation from the left by σ(g) for g ∈ Spin1,3 and use (2.75) to obtain
( )
−1 µ ν ~ ∂
ρ(g ) ν γ − eAµ σ(g)ψ + mc σ(g)ψ = 0 . (2.79)
i ∂xµ
It remains to transform the arguments of ψ and Aµ from v to ρ(g)−1 v. For this step we use the
µ
fact that application of the operator e− ω ν xν ∂µ
(for ω µν ∈ R) to a function v 7→ f (v) yields
( µ ) ∂
e− ω ν xν ∂µ
f (v) = f (e−ω v) , ∂µ ≡ . (2.80)
∂xµ
We also use the formula
µ µ
e− ω ν xν ∂µ
∂λ e ω ν xν ∂µ
= ∂λ + ω µλ ∂µ + . . . = (eω )µλ ∂µ .
µ
Now we apply the operator e− ω ν x ∂µ for eω ≡ ρ(g) to Eq. (2.79). This gives
ν
( )
−1 µ ν ~ ∂ ′
ρ(g ) ν γ ρ(g) µ λ − eAµ (g · ψ) + mc (g · ψ) = 0 ,
λ
(2.81)
i ∂x
where A′µ (v) = Aµ (ρ(g)−1 v). Since ρ(g)λµ ρ(g −1 )µν = ρ(gg −1 )λν = δνλ we obtain
( )
~ ∂
γ µ
− e (g · A)µ (g · ψ) + mc (g · ψ) = 0 , (2.82)
i ∂xµ
which is the desired result.
47
2.13 Discrete symmetries of the Dirac equation
2.13.1 Pin group and parity transformation
Both the special orthogonal group SO(V, Q) and the spin group Spin(V, Q) are connected. (In
fact, Spin has the additional property of being simply connected, which means that every closed
path in it is contractible to a point; the latter is not the case for SO.) However the full orthogonal
group O(V, Q) has more than one connected component. In particular, the operation of space
reflection, while not contained in SO(V, Q), does belong to O(V, Q) for our case of V = R4 with
Minkowski scalar product Q. One may then ask whether there exists a group Pin(V, Q) which
is related to Spin(V, Q) in the same way that O(V, Q) is related to SO(V, Q). The answer, as it
turns out, depends on whether the quadratic form Q is definite or indefinite. In the definite (or
Euclidean) case, the answer is yes; otherwise it is only partially yes. We now briefly highlight a
few salient points in order to motivate the form of the parity operator for the Dirac equation.
The Clifford algebra comes with an automorphism α : Cl(V, Q) → Cl(V, Q) by the Z2 -degree:
α(x) = x if x is even, and α(x) = −x if x is odd. The Clifford algebra also comes with an
anti-automorphism called the transpose; this is defined by 1 = 1t , v t = v for v ∈ V , and
(xy)t := y t xt .
Definition. The Clifford group Γ(V, Q) is defined as the group of invertible elements x ∈ Cl(V, Q)
with the property that twisted conjugation
Cl(V, Q) ⊃ V ∋ v 7→ xv α(x)−1
stabilizes V , i.e., maps V into itself. The group Pin(V, Q) ⊂ Γ(V, Q) is the subgroup defined as
−u Q(u, v)
uv α(u)−1 = uv = v − 2u ∈V .
Q(u, u) Q(u, u)
48
is a space reflection, i.e., the operation of inverting each of the three Cartesian coordinates. This
operation is also called a ‘parity transformation’ in physics.
In the context of the Dirac equation one wants to work with conjugation v 7→ gvg −1 rather than
twisted conjugation gv α(g)−1 . To correct for the extra minus sign inflicted by twisted conjugation,
we multiply g = e1 e2 e3 by a pin group element x with the property that xv x−1 = −v for all v ∈ V .
If e0 with Q(e0 , ej ) = 0 (j = 1, 2, 3) is a time-like unit vector, the product x = e0 e1 e2 e3 is such an
element. Thus we arrive at xg = e0 . Since e0 is represented in the spinor representation by γ 0 ,
the next definition is well motivated.
Definition. The operator of parity transformation acts on Dirac spinors ψ as
Going beyond parity, the Dirac equation has some discrete symmetries which do not come (not
in a physically satisfactory manner anyway) from the pin group but require complex anti-linear
operations. One of these is time reversal, T . In the standard representation (2.15) of the gamma
matrices this operation is given by
49
for µ = 1, 3 and ϵ = 1 for µ = 0, 2. In this way we find that T transforms a solution ψ of the
Dirac equation (2.85) into a solution of the same equation with transformed gauge potential
( ∑ )
(T · A)(⃗x, t) = − −A0 (⃗x, −t) dx + 0
Aj (⃗x, −t) dx j
.
The Dirac equation has another complex anti-linear symmetry, which is called charge conju-
gation. This operation (again, in standard representation) is defined by
The name derives from the easily verified fact that C transforms a solution ψ of the Dirac equation
into a solution C · ψ of the same equation with conjugated charge (e → −e). Notice that C 2 = 1,
and CP = −P C, whereas CT = T C and P T = T P .
Consider now the combination of charge conjugation, parity, and time reversal:
This combined operation is more fundamental than its individual factors in that it is defined even
in the absence of a splitting R1,3 = R ⊕ R3 into time and space. Since each factor is a symmetry
of the Dirac equation, so is the product CP T . The precise statement is this: if ψ is a solution of
the Dirac equation with gauge field Aµ , then CP T · ψ is a solution with transformed gauge field
Remark. A deep result of theoretical physics is the so-called CP T theorem: any quantum field
theory which is unitary (i.e., has a Hermitian Hamiltonian for its time-evolution generator), local
(i.e., the Lagrangian contains only derivatives up to a finite order), and Lorentz-invariant, must
have CP T as a symmetry. In other words: the combined operation of running time backwards,
reflecting space, and replacing matter by anti-matter, must always be a symmetry of the funda-
mental laws of nature. Another remark is that the CP T theorem serves as important input to
the proof of the spin-statistics theorem, which says that bosons and fermions have integer and
half-integer spin, respectively.
Although CP T is always a symmetry, it turns out that there do exist fundamental interactions of
nature which break one or another of the individual factors. Here we briefly indicate how parity
is broken by the weak interaction.
In the limit of zero mass (m → 0), the Dirac equation for the 4-component spinor ψ splits into
a pair of equations for two 2-component spinors ψL and ψR , which are called Weyl or half-spinors.
This works as follows. Based on γ52 = 1 for γ5 = iγ 0 γ 1 γ 2 γ 3 , one introduces projection operators
Π± = 21 (1 ± γ5 ) . (2.90)
50
By using γ µ Π± = Π∓ γ µ , one infers from the Dirac equation (2.85) with mass m = 0 that
( ) ( )
~ ∂ ~ ∂
0 = Π± γ µ
− eAµ ψ = γ µ
− eAµ Π∓ ψ . (2.91)
i ∂xµ i ∂xµ
Thus both ψL := Π+ ψ and ψR := Π− ψ are solutions of the massless Dirac equation.
To interpret the half-spinors ψR , ψL it is best to choose the so-called Weyl representation for
the gamma matrices:
( ) ( ) ( )
0 0 1 j 0 σj 1 0
γ = , γ = , γ5 = . (2.92)
1 0 −σj 0 0 −1
In this representation, (the non-zero part of) ψL consists of the upper two components of ψ and
ψR consists of the lower two components. The decoupled pair of equations for ψL,R then reads
( ∑ ) ( ∑ )
∂t + c σj ∂j ψR = 0 , ∂t − c σj ∂j ψL = 0 (2.93)
⃗
in zero electromagnetic field. In order for ψR = ei(k· ⃗x−|k|ct) η to be a solution of the first equation,
the constant 2-component spinor η must be an eigenvector
( )
⃗ ⃗ k3 k1 − ik2
⃗σ · k η = |k| η , ⃗σ · k = ,
k1 + ik2 −k3
of ⃗σ · ⃗k/|k| with eigenvalue (also referred to as helicity) +1. In such an eigenvector the momentum
and spin angular momentum of the particle form a right-handed system (hence the notation ψR
⃗
and the name helicity). Similarly, in order for ψL = ei(k· ⃗x−|k|ct) ξ to be a solution of the second
equation, the spinor ξ must be an eigenvector of ⃗σ · ⃗k/|k| with eigenvalue −1. In this case the
momentum and the spin angular momentum form a left-handed system. Note that Π+ P = P Π− ,
so that P ψL = ψR and P ψR = ψL , in agreement with the observation that space reflection
transforms a right-handed system into a left-handed one and vice versa.
Experiments on β-decay done in the mid-1950’s showed a strong right-left asymmetry of the
radioactive decay process. This is understood nowadays (from the so-called Standard Model of
elementary particle physics) as resulting from a peculiar feature of the electro-weak interaction:
left-handed particles carry electro-weak charge, but right-handed particles do not – a clear violation
of parity symmetry! More )
( precisely, the left-handed component eL of the electron constitutes a
eL
doublet of half-spinors together with the electron neutrino, and an analogous doublet is
νL
formed by the up and down quarks. These doublets are governed by the second equation of (2.93)
with minimal substitution ∂µ → ∂µ − Aµ where Aµ , the electro-weak gauge potential, is a 2 × 2
matrix. The off-diagonal elements of Aµ quantize as the W ± boson, the diagonal elements as the
Z 0 boson. The right-handed component of the electron only couples to the electromagnetic field.
51
2.14 Dirac electron in a Coulomb potential
A major application of non-relativistic quantum mechanics is the computation of the discrete
energy spectrum of the hydrogen atom using the Schrödinger equation. In this section we carry
out a similar computation based on the Dirac equation.
The first step is to choose a gauge. We take
where Φ = Ze/(4πε0 r) is the Coulomb potential due to a positive nuclear charge Ze placed at
√
the origin of the coordinate system. As usual, r = x2 + y 2 + z 2 denotes the distance from the
origin. Adopting this gauge, the Hamiltonian for the negatively charged Dirac electron is
~c ∑ ∂ Ze2
H= αl + βmc2 − . (2.95)
i l ∂xl 4πε0 r
We are going to look for bound-state solutions of the stationary Dirac equation
Hψ = Eψ . (2.96)
Our strategy is the same as in the non-relativistic case: exploiting the spherical symmetry of the
Coulomb potential, we will reduce Hψ = Eψ to a radial differential equation. The solution of the
latter equation will then be found as a power series.
The Coulomb potential is invariant under space rotations fixing the origin (the nucleus). This
invariance will allow us to find the solution in the form of a product ansatz (separation of variables).
To begin, we recall the action of a spin group element g ∈ Spin1,3 on the spinor field ψ:
(g · ψ)(v) = σ(g)ψ(ρ(g)−1 v) .
Here we want this transformation law for the special case of g ∈ Spin3 ⊂ Spin1,3 , ρ ∈ SO3 , and
we only need it at the infinitesimal level. Hence we put gt = etX and τ (X) = ω and differentiate
with respect to the parameter t at t = 0.
To compute the contribution from the rotation v 7→ ρ(gt )−1 v of the argument of ψ , we use the
∑
formulas ρ(gt )−1 = e−tω and f (e−tω •) = e−t ωkl xl ∂k
f [see Eq. (2.80)],
∑ ∑ ∑ ~∑ ∑
e−t
i
ωkl xl ∂k
= e~t ω j Lj
, ωkl = ϵjkl ωj , Lj = ϵjkl xk ∂l = ϵjkl xk pl .
j i kl kl
Here ϵjkl is the totally anti-symmetric tensor of R3 , i.e., ϵ123 = ϵ231 = ϵ312 = 1, ϵ321 = ϵ213 = ϵ132 =
−1, and the tensor vanishes if two indices coincide.
There is a second contribution to the derivative at t = 0; this comes from multiplication
ψ 7→ σ(gt )ψ = σ(etX )ψ by the spinor representation of gt . To compute it, we use the formulas
( )
−1 1∑ i∑ σj 0
X = τ (ω) = ωkl γl γk = ωj Σ j , Σ j = ,
4 kl 2 j 0 σj
52
where γj represents the Cartesian basis vector ej ∈ R3 viewed as a generator of the Clifford algebra
(in the spinor representation). Putting everything together, we obtain
d τ −1 (tω) i ∑ ~
e · ψ = ωl Jl ψ , J = L + S , Sl = Σl . (2.97)
dt t=0 ~ l 2
The interpretation of this result is clear: the total angular momentum J is the sum of orbital
angular momentum L and spin angular momentum S.
Problem. Show by direct calculation that the Dirac Hamiltonian (2.95) has vanishing commu-
tator with each of the three components of the total angular momentum J = L + S.
It follows that [H, J 2 ] = 0. We also have [H, P ] = 0 by the parity symmetry of the Coulomb
potential. We thus have a set of four commuting operators {H, J 2 , J3 , P }, and we will be looking
for joint eigenfunctions of this set. For this purpose, let
( ) ( )
( ) 1 0
φjm := Yl ⊗ χ1/2 jm = ajm Yl, m−1/2 ⊗
l l
+ bjm Yl, m+1/2 ⊗
l
(2.98)
0 1
be a 2-component angular momentum eigenfunction:
The exact details of this construction will be spelled out in the chapter on tensor invariants of the
rotation group SO3 ; here we just mention/anticipate that the coefficients aljm and bljm are so-called
Clebsch-Gordan coefficients coupling orbital angular momentum (with quantum numbers l m±12 )
and spin angular momentum (with quantum numbers 1
2
±12 ) to total angular momentum (with
quantum numbers j m):
aljm = ⟨l m−21 , 12 1
2
| j m⟩ , bljm = ⟨l m+12 , 12 − 12 | j m⟩ . (2.101)
The main result to be used in the sequel now follows from the identity
∑
2 Li Si = (L + S)2 − L2 − S 2 = J 2 − L2 − S 2 .
i
⃗ := ∑ σi Li :
Using it we infer that φljm is an eigenfunction of the operator ⃗σ · L
( )
(⃗σ · L)
⃗ φl = ~ j(j + 1) − l(l + 1) − 3 φl = −~ (1 + κ) φl .
jm 4 jm jm (2.102)
We now proceed with the solution of Hψ = Eψ by making an ansatz for the 4-component spinor
ψ as a direct sum of two 2-component spinors ξ and η :
( )
ξ
ψ= . (2.104)
η
53
The stationary Dirac equation with Coulomb potential V (r) := −Ze2 /(4πε0 r) then decomposes
as a pair of coupled equations for ξ and η :
(E − mc2 − V ) ξ = c ⃗σ · p⃗ η , (2.105)
(E + mc2 − V ) η = c ⃗σ · p⃗ ξ . (2.106)
At this point we recall from (2.98) the angular momentum eigenfunction φljm and separate the
variables by making the product ansatz
⃗σ · ⃗x l
ξ = iF (r)φljm , η = G(r) φjm . (2.107)
r
We pause briefly to subject this ansatz to a consistency check. A short calculation shows that
⃗σ · ⃗x is invariant under infinitesimal rotations:
[∑ ] ~∑ ∑
[Jj , ⃗σ · ⃗x] = ϵjkl xk pl + 21 ~σj , ⃗σ · ⃗x = ϵjkl xk σl + i~ ϵjkl σl xk = 0 .
i
On the other hand, space reflection ⃗x → −⃗x takes ⃗σ ·⃗x into its negative. Thus ⃗σ ·⃗x is what is called
a pseudo-scalar in physics. Now a spherical harmonic Yl ml transforms under space reflection as
Yl ml → (−1)l Yl ml . Therefore our ansatz has parity P ψ = (−1)l ψ .
∑
To proceed, we use the formula σj σk = δjk + i ϵjkl σl to deduce the identities
With their help we now compute the result of applying the operator ⃗σ · p⃗ to the two-component
spinors ξ and η of (2.107):
( )
⃗σ · ⃗x ⃗σ · ⃗x ∂ 1+κ
⃗σ · p⃗ ξ = 2 ((⃗σ · ⃗x ) (⃗σ · p⃗ )) ξ = ~ 2 + F φljm , (2.110)
r r ∂r r
( )
( )G l ∂ 2 1+κ
⃗σ · p⃗ η = (⃗σ · p⃗ )(⃗σ · ⃗x ) φ = i~ − − + G φljm . (2.111)
r jm ∂r r r
Finally, by making the substitutions F = r−1 f and G = r−1 g we bring the pair of equations
(2.105, 2.106) into the form
( )
∂ κ
(E − mc − V )f = ~c − +
2
g, (2.112)
∂r r
( )
∂ κ
(E + mc − V )g = ~c + +
2
f. (2.113)
∂r r
We must now solve this pair of ordinary differential equations in the radial variable r.
We first investigate the problem in the limit r → ∞ where the Coulomb energy V (r) ∼ 1/r and
the rotational energy κ/r become negligible and our pair of equations simplifies to
∂g ∂f
(E − mc2 )f ≈ −~c , (E + mc2 )g ≈ ~c .
∂r ∂r
54
We easily see that this pair of asymptotic equations is solved by
where
√
λ = (~c)−1 (mc2 )2 − E 2 (2.114)
is a characteristic wave number of the problem. We observe that in the non-relativistic limit
0 < mc2 − E ≪ mc2 the expression for λ reduces to
√
λ = ~−1 2m△ , △ = mc2 − E ,
which is the de Broglie (imaginary) wave number for a bound state of binding energy △ .
These asymptotic considerations motivate the following ansatz:
√ √
E −λr E −λr
f = 1+ 2
e (A + B) , g = 1− e (A − B) . (2.115)
mc mc2
with unknown functions A = A(r) and B = B(r) which behave as A → 0 and B → const for
r → ∞. To process this ansatz, it is useful to introduce the following dimensionless quantities:
E mc2 e2
ρ := 2λr , kE := Zα , kM := Zα , α := , (2.116)
~cλ ~cλ 4πε0 ~c
where α ≈ 1/137 is the so-called fine structure constant. (This name is of historical origin. It is
nowadays understood that α is not a constant; being the coupling of quantum electrodynamics,
α depends logarithmically on the energy cutoff or regularization scale.)
We now insert the ansatz (2.115) into (2.112, 2.113). After some trivial (if somewhat lengthy)
algebra we arrive at the following equations for A and B :
∂A kE kM + κ ∂B kM − κ kE
=A− A− B, = A+ B. (2.117)
∂ρ ρ ρ ∂ρ ρ ρ
The final step is to find the solution to these equations in the form of power series. To prepare
this step, we investigate the behavior in the limit of ρ → 0. Let
A ≃ aργ , B ≃ bργ (ρ → 0) .
By inserting this small-ρ asymptotics into (2.117) and then sending ρ → 0, we obtain the following
system of linear equations for the numbers a and b :
(γ + kE ) a + (κ + kM ) b = 0 ,
(κ − kM ) a + (γ − kE ) b = 0 .
In order for this system to have a solution other than the trivial one (a = b = 0), the determinant
of the underlying matrix must vanish:
( )
γ + kE κ + kM
Det = γ 2 − kE2 − κ2 + kM
2
=0.
κ − kM γ − kE
55
2
From (2.116) we then see that kM − kE2 = (Zα)2 . Therefore the value of γ must be
√
γ = γ(κ) = κ2 − Z 2 α 2 , (2.118)
which makes good sense as long as the nuclear charge Z is not too large.
We finally look for the solution in the form of two power series:
∑
∞ ∑
∞
A= an ργ+n , B= bn ργ+n .
n=0 n=0
(n + γ + kE ) an + (κ + kM ) bn = an−1 ,
(κ − kM ) an + (n + γ − kE ) bn = 0 .
The equation for n = 0 (with a−1 = 0) has a non-trivial solution for a0 and b0 by our choice
(2.118) for γ. For n ≥ 1 we eliminate bn to obtain a recursion relation for an alone:
( )−1
(κ + kM )(κ − kM ) n + γ − kE
an = (n + γ + kE ) − an−1 = 2 an−1 .
n + γ − kE n + 2nγ
A sufficient condition for a bound-state solution to emerge is that the power series terminates at
some finite order n resulting in square-integrable functions f and g. The condition for termination
is that
0 = n + γ − kE
holds for some positive integer n ≥ 1. By solving this condition for E, we arrive at the following
answer for the (discrete) energy spectrum:
mc2
En,κ = √ . (2.119)
(Zα)2
1+ (n+γ(κ))2
Problem. Show that in the non-relativistic limit c → ∞ this formula reproduces the known
energy spectrum of hydrogen as computed from the Schrödinger equation. How do n , κ relate to
the quantum numbers of the non-relativistic spectrum?
56
3 Second quantization
3.1 Bosons and fermions
We are assuming that the students of this course already have some familiarity with the quantum
mechanics of a single particle and thus with quantum theory at the simplest level. As a prepara-
tion for the more advanced formalism of ‘second quantization’ introduced in this chapter, let us
summarize by means of a table what are the key elements of the correspondence between classical
mechanics and single-particle quantum mechanics.
When there is more than one particle present, new structures and phenomena related to per-
mutation symmetry appear. It turns out that the generalization to n-particle quantum mechanics
depends on whether the particles under consideration are distinguishable or indistinguishable. In
the former case the Hilbert space V of the many-particle system is simply the tensor product
of the Hilbert spaces V (1) , . . . , V (n) for the individual particles (labeled by 1, . . . , n). Thus the wave
function for n distinguishable particles in the Schrödinger representation is a linear combination
of products of single-particle wave functions,
∑
Ψ(x1 , x2 , . . . , xn ) = ai1 i2 ... in ψi1 (x1 )ψi2 (x2 ) · · · ψin (xn ) , (3.2)
with complex coefficients ai1 i2 ... in . Linear operators representing observables such as the energy
or the components of momentum, act on these wave functions in the natural way (i.e., one simply
tensors up the setting of single-particle quantum mechanics), and there isn’t a whole lot of new
material for the student to learn.
However, in the case of indistinguishable or identical particles, the many-particle Hilbert space
is spanned by wave functions which are either totally symmetric or totally anti-symmetric. Con-
sider first the latter case where we have
57
for any permutation π ∈ Sn . [Here sign(π) = +1 if the permutation π is even and sign(π) = −1
if π is odd.] Particles with this behavior under permutations are called fermions.
Gibbs, in the context of classical statistical mechanics, observed that symmetrization (more
precisely: reduction of statistical weight by identification of n-particle configurations that are
permutations of one another) is needed in order for the entropy to be additive. The next series of
statements gives a quantum-theoretic explanation of this observation.
Fact. The wave functions for (an indefinite number of) fermions with single-particle Hilbert space
V , e.g. V = L2 (R3 ), form an exterior algebra, ∧(V ). The wave function for a definite number n
of fermions is an element of the degree-n subspace ∧n (V ) ⊂ ∧(V ) of the exterior algebra. If Ψp
and Ψq are totally anti-symmetric wave functions describing two subsystems of p resp. q fermions
(all of which are identical), then the wave function for the total system of p + q identical fermions
is the exterior product Ψp ∧ Ψq defined by
1 ∑
(Ψp ∧ Ψq )(x1 , . . . , xp+q ) := sign(π)Ψp (xπ(1) , . . . , xπ(p) ) Ψq (xπ(p+1) , . . . , xπ(p+q) ) . (3.4)
p! q! π∈S
p+q
The most general n-fermion wave function ψ ∈ ∧n (V ) can be expressed as a linear combination
of exterior products of single-particle wave functions:
∑
Ψ= ai1 i2 ... in ψi1 ∧ ψi2 ∧ · · · ∧ ψin , (3.5)
with complex coefficients ai1 i2 ... in . The vacuum state is 1 ∈ ∧0 (V ) ≡ C. One also writes 1 ≡ |0⟩.
Problem. Show that the exterior product is associative: (Ψp ∧ Ψq ) ∧ Ψr = Ψp ∧ (Ψq ∧ Ψr ) .
We now turn to the alternative situation of totally symmetric wave functions:
for any permutation π ∈ Sn . Particles with this behavior under permutations are called bosons.
Definition. If V is a K-vector space, one defines the symmetric algebra S(V ) as the associative
algebra which is generated by V ⊕ K with relations vv ′ = v ′ v for all v, v ′ ∈ V . Multiplication in
the symmetric algebra with these relations is called the symmetric product. When the symmetric
product is to be emphasized, we use the expanded notation v ∨ v ′ = v ′ ∨ v. The symmetric algebra
is graded by the degree:
⊕
S(V ) = Sn (V ) , S0 (V ) ≡ K , S1 (V ) ≡ V . (3.7)
n≥0
( )
N +n−1 n
Problem. Assuming V to be of finite dimension dim V = N , show that dim S (V ) = .
n
[Thus, unlike the exterior algebra ∧(V ), the symmetric algebra S(V ) has infinite dimension even
when V is finite-dimensional.]
Fact. The wave functions for (an indefinite number of) bosons with single-particle Hilbert space
V form the symmetric algebra S(V ). Any wave function for a definite number n of bosons is
an element of the degree-n subspace Sn (V ) ⊂ S(V ) of the symmetric algebra. If Φp and Φq are
58
totally symmetric wave functions describing two subsystems of p resp. q bosons (all of which are
identical), then the wave function for the total system of p + q bosons is the symmetric product
1 ∑
(Φp ∨ Φq )(x1 , . . . , xp+q ) = Φp (xπ(1) , . . . , xπ(p) ) Φq (xπ(p+1) , . . . , xπ(p+q) ) . (3.8)
p! q! π∈S
p+q
The most general n-boson wave function Φ ∈ Sn (V ) can be expressed as a linear combination of
symmetric products of single-particle wave functions:
∑
Φ= ai1 i2 ... in ϕi1 ∨ ϕi2 ∨ · · · ∨ ϕin . (3.9)
The Weyl algebra of U is defined to be the associative algebra, W(U, A), generated by U ⊕ K with
relations
uu′ − u′ u = A(u, u′ ) . (3.11)
Remark. Without changing the content of the theory, we could enhance the similarity with the
definition of the Clifford algebra in Section 2.7 by putting 2A instead of A on the right-hand side.
Nonetheless, in the present context we prefer to use the normalization convention (3.11).
Example 1. Consider the symplectic plane (or phase plane) U = R2 with basis vectors e, f and
symplectic form
A(e, e) = A(f, f ) = 0 , A(e, f ) = −A(f, e) = −1 . (3.12)
The Weyl algebra W(R2 , A) then is the algebra of polynomial expressions in e and f with real
coefficients and the relation f e − ef = 1 being understood. A concrete realization of this algebra
is by a polynomials in a coordinate generator x ≡ e and its derivative d/dx ≡ f .
59
Example 2. We give a second example, which is closer to the setting pursued below. Let q and
p be the position and momentum operators of a one-dimensional quantum system. Let K = C ,
U = spanC {p , q}, and take A to be the Heisenberg commutator:
Then the Weyl algebra W(U, A) is the algebra of polynomials in the operators q and p . A specific
example of such a polynomial is the Hamiltonian H = p2 /2m + mω 2 q 2 /2 of the one-dimensional
harmonic oscillator. Introducing the dimensionless operators
( ) ( )
1 q ℓp † 1 q ℓp
a := √ +i , a := √ −i , (3.14)
2 ℓ ~ 2 ℓ ~
√
where ℓ = ~/(mω) is the oscillator length, one can also view the Weyl algebra W(U, A) as the
algebra of polynomials in the operators a and a† with commutation relations
a a† − a† a = 1 . (3.15)
With these definitions, the Weyl algebra W(V ⊕ V ∗ , A) is the algebraic structure underlying (the
formalism of second quantization for) many-boson quantum systems.
Definition. In physics the following notational conventions are standard. One fixes some basis
{e1 , e2 , . . .} of the single-boson Hilbert space V with dual basis {f1 , f2 , . . .} of V ∗ , and one writes
fi ≡ ai and ei ≡ a+
i . Using the duality pairing fi (ej ) = δij , one then gets the Weyl algebra
[ai , a+
j ] = δij , [ai , aj ] = [a+ +
i , aj ] = 0 (i, j = 1, 2, . . .) . (3.17)
S(V ) with its canonical Hermitian scalar product. In that setting, assuming that {e1 , e2 , . . .} is
†
i ≡ ai will turn out to be the Hermitian adjoint of ai .
an orthonormal basis, the operator a+
Example. By the canonical commutation relations one has
i aj ak al = ai ak aj al + δjk ai al .
a+ + + + +
(3.18)
60
In view of the close similarity between the Weyl algebra and the Clifford algebra, one may
ask whether a development parallel to that of Sections 2.10 and 2.11 exists on the bosonic side.
The answer is: yes, to a large extent. Indeed, the analog of SO(V, Q) is the symplectic Lie group
Sp(U, A), the analog of so(V, Q) is the symplectic Lie algebra sp(U, A), and the analog of Cl2 (V, Q)
is the subspace
∑
W2 (U, A) := {X ∈ W(U, A) | X = (ui vi + vi ui ) ; ui , vi ∈ U }
i
of symmetrized degree-2 elements of the Weyl algebra. It is still true that W2 (U, A) is a Lie
algebra which is isomorphic to sp(U, A) by letting X ∈ W2 (U, A) act on u ∈ U ⊂ W(U, A) by
the commutator, and that by exponentiating this commutator action one recovers the symplectic
Lie group. There is, however, one difference: it is not possible to exponentiate W2 (U, A) inside
the Weyl algebra to produce an analog of the spin group. The reason is that, by a theorem due
to Stone and von Neumann, any non-trivial realization of the Weyl algebra with relations (3.11)
must be on an infinite-dimensional representation space where the elements X ∈ W2 (U, A) act as
unbounded operators. For example, the Hamiltonian H of the harmonic oscillator is an element
H ∈ W2 (U, A) for our Example 2 above. The exponential e−tH exists for Re t > 0 but does not
for Re t < 0 . Nonetheless, in a suitable real framework there does exist a bosonic analog of the
spin group; it is called the metaplectic group. (For example, the quantum time evolution of the
harmonic oscillator is a one-parameter subgroup of the metaplectic group.)
This concludes our introduction of the Weyl algebra and the canonical commutation relations
for bosons. We now turn to the fermionic side: the Clifford algebra. Here we will be brief, as we
have already given a thorough discussion of the Clifford algebra in Chapter 2.
Definition. The formalism of second quantization for fermions is based on the following. Starting
from the single-fermion Hilbert space V , we take the direct sum U = V ⊕ V ∗ and we equip U with
the symmetric bilinear form B : U × U → C defined by
The Clifford algebra Cl(U, B) then is the associative algebra generated by U ⊕ C with relations
In physics it is customary to use the following notational conventions. Fixing any basis {e1 , e2 , . . .}
of the single-fermion Hilbert space V with dual basis {f1 , f2 , . . .} of V ∗ , one writes fi ≡ ci and
e i ≡ c+
i . By the duality pairing fi (ej ) = δij , one then gets the Clifford algebra relations (3.20) in
the form
j } = δij ,
{ci , c+ i , cj } = 0
{ci , cj } = {c+ +
(i, j = 1, 2, . . .) , (3.21)
where {u, u′ } := uu′ + u′ u denotes the anti-commutator. We recall from Section 2.8 that the
relations (3.21) are referred to as the canonical anti-commutation relations (CAR).
61
3.3 Representation on Fock space
In Section 2.8 we learned that the abstract Clifford algebra Cl(V ⊕V ∗ , B) has a concrete realization
by linear operators on the spinor representation space ∧(V ). [Actually, we constructed the spinor
representation for V = P ∗ . Note that V and V ∗ are on the same footing in this context, and
exactly the same construction would go through for the choice ∧(P ∗ ) = ∧(V ∗ ) of representation
space.] We will now explain that, in a very similar way, the abstractly defined Weyl algebra
W(V ⊕ V ∗ , A) has a realization by linear operators on the so-called bosonic Fock space S(V ).
Note that A(v, v ′ ) = A(φ, φ′ ) = 0 vanishes for any v, v ′ ∈ V and φ, φ′ ∈ V ∗ . One expresses this
fact by saying that V and V ∗ are Lagrangian subspaces of the complex symplectic vector space
U = V ⊕ V ∗ . It follows from the Weyl algebra relations (3.11) that vv ′ = v ′ v and φφ′ = φ′ φ.
These are the defining relations of the respective symmetric algebras, namely S(V ) and S(V ∗ ).
We thus see that both S(V ) and S(V ∗ ) are contained as subalgebras in W(V ⊕ V ∗ , A).
We can now get a representation of W(V ⊕ V ∗ , A) by letting it act on S(V ), or S(V ∗ ), or any
other subalgebra generated by a Lagrangian subspace. We focus on S(V ) for concreteness. Thus
our goal now is to define an action of W(V ⊕ V ∗ , A) on S(V ),
The notation µ(v) is used in order to distinguish the algebra element v ∈ V ⊂ W(V ⊕ V ∗ , A)
from its concrete realization as a linear operator µ(v) on S(V ). Note that this multiplication
µ(v) : Sl (V ) → Sl+1 (V ) increases the degree (in physics language: the number of bosons) by one.
Now consider φ ∈ V ∗ . Such elements act by a degree-lowering operation δ(φ) :
which is called a derivation of S(V ), as it has the algebraic properties of a derivative. We define
it recursively. For the two lowest degrees l = 0, 1 we set
The action on higher-degree elements ξ ∨ η ∈ S(V ) is then defined by the product rule (or Leibniz
rule) known from differential calculus:
62
Problem. Show that the operations of symmetric multiplication µ(v) : Sl (V ) → Sl+1 (V ) and
derivation δ(φ) : Sl (V ) → Sl−1 (V ) satisfy the canonical commutation relations (CCR):
shows that the relations (3.11) are satisfied for all v + φ = u and v ′ + φ′ = u′ . Thus we have
indeed constructed a representation of W(V ⊕ V ∗ , A).
Definition. For a complex vector space V let the direct sum V ⊕ V ∗ be equipped with the
canonical alternating form A(v + φ, v ′ + φ′ ) = φ(v ′ ) − φ′ (v). The oscillator representation of the
Weyl algebra W(V ⊕ V ∗ , A) is defined by the action
Let us now give the translation into the language and notation of physics. We recall that the
symmetric algebra S(V ) is referred to as the bosonic Fock space (of the single-particle Hilbert
space V ) in this context. The neutral element 1 ∈ S0 (V ) = C is called the vacuum state and is
denoted by 1 ≡ |0⟩. (To avoid confusion, we emphasize that |0⟩ is not the zero vector!) We also
recall that fixing an orthonormal basis {e1 , e2 , . . .} of V together with the dual basis {f1 , f2 , . . .} of
V ∗ , one writes a+
i ≡ ei and ai ≡ fi and refers to these as particle creation and particle annihilation
operators, respectively. No change of notation is made when these algebraically defined objects act
as linear operators on the bosonic Fock space. Thus ai ≡ δ(fi ) and a+
i ≡ µ(ei ). The definitions
i = µ(ei ) : S (V ) → S
of multiplication a+ (V ) and derivation aj = δ(fj ) : Sl (V ) → Sl−1 (V )
l l+1
translate to
( + + )
j · ai1 ai2 · · · ail |0⟩ := aj ai1 ai2 · · · ail |0⟩ ,
a+ + + + + +
(3.30)
( )
i1 ai2 · · · ail |0⟩ := δji1 ai2 ai3 · · · ail |0⟩
aj · a+ + + + + +
63
The action (3.31) of the particle annihilation operator aj = δ(aj ) is also defined by saying that
the canonical commutation relations (3.17) hold and
aj |0⟩ = 0 . (3.32)
This concludes, for the moment, our development of the bosonic variant of second quantization,
and we turn to a summary and adaptation of the material of Chapter 2 for the fermionic side.
Definition (Review). For a complex vector space V let the direct sum V ⊕ V ∗ be equipped
with the canonical symmetric form B(v + φ, v ′ + φ′ ) = φ(v ′ ) + φ′ (v). The spinor representation
of the Clifford algebra Cl(V ⊕ V ∗ , B) is defined by the action
on the exterior algebra ∧(V ). In physics one calls ∧(V ) the fermionic Fock space. The vacuum
is denoted by |0⟩ ≡ 1 ∈ ∧0 (V ). Introducing an orthonormal basis as usual, the actions of the
particle creation operator ε(ej ) ≡ c+
j and particle annihilation operator ι(fj ) ≡ cj are given by
( + + )
c+
j · c c
i1 i2 · · · c +
il |0⟩ j ci1 ci2 · · · cil |0⟩ ,
:= c+ + + +
(3.34)
( + )
cj · c+ i1 ci2 · · · cil |0⟩ := δji1 ci2 ci3 · · · cil |0⟩
+ + + +
− δji2 c+
i1 ci3 · · · cil |0⟩ + . . . + (−1)
+ + l−1
i1 ci2 · · · cil−1 |0⟩ .
δjil c+ + +
(3.35)
64
3.4 Fock space scalar product
We are in the process of developing a quantum-theoretic formalism for many bosons and/or
fermions. What we have learned so far is how the single-particle Hilbert space V gets promoted
to a Fock space S(V ) or ∧(V ) for many particles. In order to turn Fock space into the Hilbert
space of many-body quantum theory, we still need to introduce a Hermitian scalar product on it.
Given that the single-particle Hilbert space V already carries a Hermitian scalar product ⟨·, ·⟩V ,
there is a natural way of doing this, as follows.
For bosons and fermions alike, the Fock space scalar product is diagonal in the degree or
particle number n . For bosons one defines the scalar product ⟨·, ·⟩Sn (V ) : Sn (V ) × Sn (V ) → C by
⟨ ⟩ ∑
v1 ∨ v2 ∨ · · · ∨ vn , v1′ ∨ v2′ ∨ · · · ∨ vn′ Sn (V ) = ′
⟨v1 , vπ(1) ′
⟩V ⟨v2 , vπ(2) ′
⟩V · · · ⟨vn , vπ(n) ⟩V . (3.36)
π∈Sn
Notice that in the special case of v n = v ∨ v ∨ · · · ∨ v (n factors) one has ⟨v n , v ′ n ⟩Sn (V ) = n!⟨v, v ′ ⟩nV .
Using physics notation, this means that
It can be shown that this has all the properties required of a Hermitian scalar product; in partic-
ular, ⟨v1 ∧ · · · ∧ vn , v1 ∧ · · · ∧ vn ⟩∧n (V ) is non-negative and vanishes only if v1 ∧ · · · ∧ vn is the zero
vector. The latter happens if and only if the set of vectors v1 . . . , vn is linearly dependent. This
feature is called the Pauli exclusion principle.
i = ε(ei ) for an orthonormal basis {e1 , e2 , . . .}
Assuming again the standard physics convention c+
of V , one finds that the set of Fock space vectors
65
There exists another perspective on the Hermitian scalar product in Fock space, as follows.
We recall that the Hermitian scalar product ⟨·, ·⟩V on single-particle Hilbert space V determines
a one-to-one correspondence
cV : V → V ∗ , v 7→ ⟨v, ·⟩V ,
called the Dirac ket-bra bijection and written |v⟩ 7→ ⟨v| in Dirac’s notation. Consider now the
case of fermions for definiteness (for bosons it works the same way). The map cV induces on the
fermionic Fock space ∧(V ) a mapping
The r.h.s. ⟨0|cin · · · ci1 is a linear function on the fermionic Fock space ∧(V ) and as such an element
of ∧(V )∗ in the natural way: to evaluate this linear function on a vector c+
j1 · · · cjn |0⟩ in Fock space
+
we apply the product cin · · · ci1 and then take the vacuum component:
( + )( + + )
i1 ci2 · · · cin |0⟩
c∧(V ) c+ cj1 cj2 · · · cjn |0⟩ = ⟨0|cin · · · ci2 ci1 c+
j1 cj2 · · · cjn |0⟩ .
+ +
By using the canonical anti-commutation relations to compute the right-hand side we obtain
( + )( + + ) ∑ ∏
n
c∧(V ) c+ c
i1 i2 · · · c +
in |0⟩ c c
j1 j2 · · · c jn |0⟩ = sign(π) δil , jπ(l) ,
π∈Sn l=1
in agreement with the definition (3.38). Thus we may regard the Fock space Hermitian scalar
product (3.38) as a manifestation of the natural mapping (3.40).
66
Proof. Consider the case of fermions and let v ∈ V ⊂ ∧(V ) be any single-particle state. Appli-
b yields
cation of L
∑ ∑ ∑
b =
Lv ε(Lei )ι(fi )v = fi (v) ε(Lei )1 = fi (v) Lei = Lv .
i i i
Adopting the standard physics conventions, we now choose our basis of V to be an orthonormal
basis {|i⟩}i=1,2,... (Dirac notation), and we write ε(ei ) ≡ c+
i , ι(fi ) ≡ ci , µ(ei ) ≡ ai , δ(fi ) ≡ ai .
+
The second quantization rule (3.41) for fermions or bosons then takes the form
∑
⟨i|L|j⟩ ci cj
+
(fermions),
L 7→ Lb= ∑
i,j
(3.42)
⟨i|L|j⟩ a+ i aj (bosons).
i,j
Example. Consider Schrödinger particles confined to a cube C with side length L and periodic
boundary conditions. The single-particle Hilbert space is V = L2 (C) and an orthonormal basis of
V is given by the plane waves
ψk (x) = L−3/2 eik·x
k = (k1 , k2 , k3 ) ∈ (2πZ/L)3 =: Λ .
The one-body operator T := −~2 ∇2 /2m of the Schrödinger kinetic energy has matrix elements
~2 k 2
⟨k|T |k′ ⟩ = δkk′ , k2 = k12 + k22 + k32 ,
2m
and the second-quantized kinetic energy Tb is
∑ ~2 k 2
Tb = εk c+
k ck , εk = .
k∈Λ
2m
67
The operator c+
k ck is called the occupation number operator (for the single-particle state with
momentum p = ~k). It has two eigenvalues: 0 (empty state) and 1 (occupied state).
\
Remark. The homomorphism property [L, b M
M ] = [L, c] does not single out (3.42) but also holds
for other schemes of second quantization. For example, it holds for
∑
⟨i|L|j⟩ ci cj
+
(fermions),
e
L 7→ L = i,j ∑
(3.43)
− ⟨i|L|j⟩ ai a+j (bosons).
i,j
is an automorphism of the Weyl algebra. The fermionic version of this ‘hole quantization’ scheme
will become relevant in the reinterpretation of the Dirac equation as a quantum field theory.
We now turn briefly to the scheme of second quantization for two-body operators. By definition,
a two-body operator is a linear operator
L: V ⊗V →V ⊗V (3.44)
defined on the tensor product of two single-particle Hilbert spaces. (While these may be different
when the two particles are distinguishable, we here continue to focus on the case of identical
particles.) If we work in the Schrödinger representation and φi (x) (i = 1, 2, . . .) is a basis of
single-particle wave functions, then the two-body matrix elements of L are expressed as
∫ ∫
⟨ij|L|kl⟩ := φi (x) φj (x′ ) (L(φk ⊗ φl )) (x, x′ ) d3 x d3 x′ . (3.45)
R3 R3
Example. The Coulomb interaction energy UC between two particles with charge e is the two-
body operator given by
e2
(UC (φk ⊗ φl )) (x, x′ ) = φk (x)φl (x′ ) . (3.46)
4πε0 |x − x |
′
68
3.6 Second quantization of Dirac’s theory
3.6.1 Motivation: problems with the Dirac equation
The Dirac equation was originally conceived as a relativistic wave equation for a single charged
particle, by the analogy with the Schrödinger equation. In the non-relativistic limit with weak
electromagnetic fields, the single-particle interpretation of the Dirac equation works just fine.
However, serious problems of interpretation arise in the presence of strong electromagnetic
fields. One problematic situation has already been discussed in the problem class; we here review
it. Addressing the Dirac equation Hψ = Eψ for stationary states in the case of an electric scalar
potential with a step: {
0 z<0,
V =
V0 z>0,
let us look for a plane-wave solution ψ with incoming wave boundary conditions ψin ∼ eikz . Thus
we employ an ansatz as a sum
E + mc2 E + mc2 E − V0 + mc2
0 0 0
ψ = a+ eikz
~kc
Θ(−z) + a− e−ikz
−~kc
Θ(−z) + b eik′ z
′
Θ(z)
~k c
0 0 0
of reflected and transmitted waves. [We use the spinor representation in its standard form with
gamma matrices given by (2.10).] The stationary Dirac equation Hψ = Eψ is solved by this
√
ansatz for z < 0 if k = (~c)−1 E 2 − (mc2 )2 and for z > 0 if
√
k ′ = ±(~c)−1 (E − V0 )2 − (mc2 )2 .
The ratios a− /a+ and b/a+ are determined by requiring continuity of the solution ψ at z = 0 .
The situation of interest now occurs when V0 > E + mc2 . In that case we have a transmitted
′
wave ∼ eik z with real wave number k ′ > 0 and the stationary conserved probability current
is negative. This phenomenon implies, in particular, that the reflection probability |r|2 = |a− /a+ |2
exceeds unity! Certainly, there exists no satisfactory interpretation of such a result within the
confines of unitary single-particle quantum mechanics.
69
Mathematical analysis of the situation reveals that the paradoxical behavior has its reason in
the existence of stationary solutions with negative energy E − V0 < −mc2 .
Furthermore, the Dirac Hamiltonian, as it stands, does not have a ground state: it is bounded
neither from below nor from above. When the interaction with the quantized electromagnetic field
is switched on, the absence of a lower bound for the Dirac Hamiltonian poses the catastrophic
threat that an infinite amount of electromagnetic energy might be released by Dirac particles
making a never ending series of transitions to states with lower and lower energy.
These problems force us to abandon the single-particle interpretation of the Dirac equation.
In fact, the correct physical interpretation of the result |a− /a+ |2 > 1 above is in terms of particle
production (namely, the creation of electron-positron pairs at the potential step). The paradox
is then resolved by reinterpreting the current vector field ⃗j = ψ † α
⃗ ψ – which had originally been
intended to be a conserved probability current – as a conserved electric current.
As we have tried to indicate, there are serious problems with the original form of Dirac’s theory
due to the fact that the Dirac Hamiltonian, being unbounded from below, does not have a ground
state. Therefore, we now look for a reformulation of Dirac’s theory such that the Hamiltonian
does have a ground state. In the present section we make preparations for such a reformulation
by taking the formalism of second quantization to its most general form.
Let H ∈ End(V ) be any Hermitian Hamiltonian on a single-particle Hilbert space V ; our
interest, of course, is in Hamiltonians H which resemble the Dirac Hamiltonian in that they have
positive as well as negative spectrum neither of which is bounded. We will describe a procedure of
second quantization H 7→ H b which sends H to a non-negative operator H b ≥ 0 on the Fock space
∧(P ∗ ) for a suitably chosen subspace P ∗ ⊂ V ⊕ V ∗ . Our procedure consists of two steps.
In the first step we convert H ∈ End(V ) into an element H C ∈ Cl(V ⊕ V ∗ , B) of the Clifford
algebra of the direct sum V ⊕ V ∗ with canonical symmetric form B(v + φ, v ′ + φ′ ) = φ(v ′ ) + φ′ (v).
To describe this step, let {e1 , e2 , . . .} be any basis of V and {f1 , f2 , . . .} be the dual basis of V ∗ .
∑
Expanding Hei = j ej Hji with respect to that basis, we define
∑
H C := Hji ej fi (3.50)
ij
as a quadratic element of the Clifford algebra Cl(V ⊕ V ∗ , B). The Clifford algebra element H C is
seen to be invariantly defined, i.e., does not depend on the choice of basis. Note that our object
H C is still formal and abtract: we have not yet specified the space which H C will be acting on.
Our second step is to choose an exterior algebra ∧(P ∗ ) for Cl(V ⊕ V ∗ , B) to act on. The
choice will be engineered in precisely such a manner that the second-quantized Hamiltonian H b
representing H C comes out to be non-negative.
Assuming that H has no zero eigenvalue, let Π± be the uniquely defined orthogonal projection
70
operators for the positive and negative eigenspaces V+ = Π+ V and V− = Π− of H. In formulas:
We now recall from Section 2.8 that a spinor representation of the Clifford algebra Cl(W, B)
is constructed by choosing a polarization
W = P ⊕ P∗ (3.52)
Thus we reverse the roles of vector space and dual space (and hence the definitions of creation
and annihilation operator) when switching from the H-positive eigenspace V+ to the H-negative
eigenspace V− .
We also recall from Section 2.8 that, given a polarization W = P ⊕ P ∗ , the Clifford algebra
generators w ∈ W ⊂ Cl(W, B) act by the spinor representation on ξ ∈ ∧(P ∗ ) as a mixture
of exterior multiplication ε(wP ∗ ) (or particle creation) for the P ∗ -component and contraction ι(wP )
(or particle annihilation) for the P -component of w.
Definition. Let H = H+ +H− be the decomposition of a Hermitian Hamiltonian H ∈ End(V ) into
its positive and negative parts, neither of which is bounded. If V = V+ ⊕ V− is the corresponding
decomposition into H-positive and H-negative subspaces, let {ei }, {ẽk } be any two bases of V+ ,
V− , and {fj }, {f˜l } the dual bases of V+∗ , V−∗ , respectively. By the positive (or stable) second
quantization of H we then mean the operator H b on ∧(V+ ⊕ V ∗ ) defined by
−
∑ ∑
b :=
H (H+ )ij ε(ei ) ι(fj ) − (H− )kl ε(f˜l ) ι(ẽk ) . (3.55)
ij kl
Thus the elements of P ∗ = V+ ⊕V−∗ act as particle creation operators (ε) and those of P = V+∗ ⊕V−
as particle annihilation operators (ι).
Remark. Note that by CAR the expression (3.55) can be rewritten as
∑ ∑ ( )
b=
H (H+ )ij ε(ei ) ι(fj ) + (H− )kl ι(ẽk ) ε(f˜l ) − δkl . (3.56)
ij kl
Written in this form, stable second quantization clearly implements the standard scheme (3.54)
∑
but for the subtraction of an (infinite) constant kl (H− )kl δkl = Tr H− . This subtraction in
combination with the adjustment of the order of operators ι(ẽk ) ε(f˜l ) − δkl = −ε(f˜l ) ι(ẽk ) is called
normal ordering.
71
b defined by (3.55) is positive.
Fact. The stable second quantization H
Proof. The operator H b is a derivation of the exterior algebra ∧(P ∗ ). By this token it is positive
on ∧(P ∗ ) if it is positive on the degree-one subspace P ∗ ⊂ ∧(P ∗ ). Now H b acts on V+ ⊂ P ∗ as
b is positive on P ∗ = V+ ⊕ V ∗ as desired.
H+ > 0 and on V−∗ as −H−t > 0 . Therefore H −
In the next subsection, the mathematical idea of stable second quantization will be used to fix the
problems with the Dirac equation. Before doing so, let us give a simplified discussion.
The main physical idea is to redefine what is meant by the vacuum. To consider a very simple
example, let |0⟩ be the (naive) vacuum for a one-dimensional single-particle Hilbert space V ≃ C ,
and denote by c+ (c) the particle creation (resp. annihilation) operator as usual. The Hamiltonian
in this one-dimensional case is just a real number H = h ∈ R . If h is positive, the standard
scheme H 7→ H b = h c+ c ≥ 0 of second quantization is fine. On the other hand, for h < 0 the
scheme must be changed if H b is to be non-negative. There exist two equivalent ways of going
about this. One is to exchange the roles of c+ ↔ c and replace c+ c by cc+ − 1 = −c+ c (that’s
what we did in Section 3.6.2). Equivalently, we may leave the operators the same and, instead,
adjust the definition of the vacuum:
|vac⟩ := c+ |0⟩ .
Indeed, by the Pauli exclusion principle c+ c+ = 0 we now have c+ |vac⟩ = 0 . Thus c+ acts on the
new vacuum as an annihilation operator. At the same time we have c|vac⟩ = |0⟩ ̸= 0, so c now
plays the role of a creation operator. And by subtracting a constant from the Hamiltonian
b := hc+ c − h
H
in order to make the energy of the new vacuum vanish (which amounts to a redefinition of the
location of zero on the energy axis), we obtain
b
H(c|vac⟩) = −h (c|vac⟩)
and thus an excited state c|vac⟩ with positive energy eigenvalue −h > 0 .
Going from this example to the setting of the Dirac equation, the idea of Dirac’s hole theory is
to define the true vacuum to be the state in which each negative-energy stationary solution of the
Dirac equation is occupied by one electron (and hence Pauli blocked), while all positive-energy
solutions are unoccupied. The fully occupied negative-energy sector is often referred to as the
‘Dirac sea’. Although the true vacuum |vac⟩ has infinite energy and infinite charge relative to
the naive vacuum |0⟩, one still commands the freedom of shifting the observables of energy and
charge so that they vanish for |vac⟩. The elementary excitations of the true vacuum then are
+ |vac⟩ in positive-energy states and hole excitations c− |vac⟩ of the Dirac sea. The latter
particles c+
have positive energy as well as positive charge (as the negative energy and negative charge of one
Dirac-sea electron are missing); they are interpreted as the states of a new elementary particle
predicted by this reinterpretation of the Dirac equation: the positron.
72
.
We now implement the scheme of stable second quantization for the Hamiltonian of the Dirac
equation. As we recall, the first step of the scheme is to convert the Dirac Hamiltonian H =
∑
βmc2 + αl pl c into a quadratic Clifford algebra element H C . Using the basis offered by spinors
ψ(x) in the position representation, this is achieved by
∫ ∑
4
C
H := dx3
(Hψ)†τ (x) ψτ (x) , (3.57)
τ =1
where ψ henceforth is called the Dirac field. Its components obey the canonical anti-commutation
relations
{ψτ (x), ψτ ′ (x′ )} = 0 , {ψτ† (x), ψτ† ′ (x′ )} = 0 , {ψτ (x), ψτ† ′ (x′ )} = δτ τ ′ δ(x − x′ ) . (3.58)
The second step is to construct a good Fock space ∧(P ∗ ) for H C to act on. For this we need
the decomposition of the single-particle Hilbert space V = L2 (R3 ) ⊗ C4 into its H-positive and
H-negative subspaces V+ and V− . By the translation invariance of H the projection operators
∫ ∑
(Π± φ)τ (x) = d3 x′ (Π± )τ τ ′ (x − x′ ) φτ ′ (x′ ) (3.59)
τ′
where Π± (k) are the matrices of the projection operators at fixed wave vector k . The latter are
found by diagonalizing the Dirac Hamiltonian
( )
mc2 ~c⃗σ · ⃗k
H(k) = (3.61)
~c⃗σ · ⃗k −mc2
Its eigenvalues are ±~ω(k), each with multiplicity two. Let us (k) and vs (k) denote a corresponding
set of eigenspinors:
with normalization u†s (k)us′ (k) = vs† (k)vs′ (k) = δss′ . Because of the two-fold spin degeneracy
there remains some freedom in the definition of the quantum number s. One possible choice is
73
⃗ · ⃗k/k – cf. Eq. (2.97) – with
to take us (k) and vs (k) to be eigenvectors of the helicity operator Σ
eigenvalue s = ±1.
From the eigenspinors us (k) and vs (k) one constructs the projection matrices
∑ ∑
Π+ (k) = us (k) ⊗ u†s (k) , Π− (k) = vs (k) ⊗ vs† (k) , (3.64)
s=±1 s=±1
We are now in a position to specify the fermionic Fock space ∧(P ∗ ) for the Dirac field ψ . Let
P := Π+ V ⊕ (Π− V )∗ . Introducing
∫
c+, s (k) := d3 x e−ik·x u†s (k)ψ(x) ∈ Π+ V ⊂ P , (3.66)
we take c+, s (k) to be the annihilation operator for an electron with momentum ~k and helicity
s. On the other hand we define
∫
c†−,−s (−k) := d3 x e−ik·x vs† (k)ψ(x) ∈ Π− V ⊂ P ∗ (3.67)
to be the creation operator for a positron with momentum −~k and helicity −s. These operators
satisfy CAR with the non-vanishing anti-commutators being
∫
† ′
′
{c+, s (k), c+, s′ (k )} = δss′ d3 x ei(k−k )·x = δss′ (2π)3 δ(k − k′ ) , (3.68)
This concludes our discussion of the procedure of second quantization for the free Dirac field.
It remains to explain what happens in the presence of an electromagnetic field. This will be done
in a later section.
74
3.7 Quantization of the electromagnetic field
Quantization of the electromagnetic field proceeds by the formalism of second quantization for
bosons. Let us review that formalism and bring it into a form well suited to our present goals.
The general starting point of second quantization for bosons is a real vector space W — the phase
space of generalized positions and momenta — with symplectic form α : W × W → R . (We here
renege on our earlier practice of using the symbol A , as A in the present context stands for the
electromagnetic gauge field.) More generally, in geometric quantization one would start from a
symplectic manifold W , but since the Maxwell theory of electromagnetism is a linear theory, it
suffices for our purposes to assume the setting of a vector space. We recall from Section 3.2 that
the symplectic vector space (W, α) determines a Weyl algebra W(W, α).
To quantize the theory and construct a quantum mechanical Hilbert space, it does not suffice
just to have a symplectic structure α on W . One also needs a complex structure J, i.e., a linear
mapping
J ∈ End(W ) , J 2 = −IdW . (3.73)
(or, in the more general case of a symplectic manifold W : a Riemannian metric and thus a Kähler
structure) by the formula
Problem. Deduce from (3.75) that g(Jw, Jw′ ) = g(w, w′ ) and α(Jw, Jw′ ) = α(w, w′ ).
The choice of g isn’t a big issue as long as dim W < ∞ , but it does become essential in the
case of an infinite-dimensional phase space W where different choices of g lead to inequivalent
quantizations. In our treatment of the quantized electromagnetic field, the Euclidean structure g
will be provided by the electromagnetic energy, which is a positive function.
Example. Consider the symplectic plane
W = spanR {eq , ep } ≃ R2
with generators eq and ep . The symplectic form (dual to the Poisson bracket) is
You should think of the dual basis functions q , p as the variables of position and momentum. In
the language of differential forms, one writes α = dp ∧ dq .
75
We now equip the vector space W with a Euclidean structure g (i.e., a Euclidean scalar
product g) by picking some harmonic oscillator Hamiltonian H = p2 /2m + mω 2 q 2 /2 and adopting
the natural interpretation of H as a quadratic form H : W → R to set
2 2
g(eq , eq ) := H(eq ) = mω , g(ep , ep ) := H(ep ) = (mω)−1 ,
ω ω
and g(eq , ep ) = g(ep , eq ) = 0 . The complex structure J ∈ End(W ) is then given by
J t q = (mω)−1 p , J t p = −mωq .
Note that J is the operator of time evolution by one quarter of the oscillator period.
We continue with the general development. The complex structure J determines a polarization
W ⊗ C = V ⊕ Ve (3.76)
by complex subspaces V and Ve which are acted upon by J as J|V = −i and J|Ve = +i (hence
the name ‘complex structure’). Thus V and Ve are the eigenspaces of J corresponding to the
eigenvalues −i and +i respectively. The complex vector space Ve is called the holomorphic part of
W ⊗ C , while V is the anti-holomorphic part. Note the characterization
between V and Ve is non-degenerate. Show also that the subspaces V and Ve are Lagrangian, i.e.,
α(v, v ′ ) = α(ṽ, ṽ ′ ) = 0 for any two vectors v, v ′ ∈ V and ṽ, ṽ ′ ∈ Ve .
By the pairing between the complex vector spaces V and Ve we may identify Ve with the dual
vector space V ∗ of V . Planck’s constant ~ has not appeared in the discussion so far, but it now
does. The point is that there is some freedom in making the identification Ve ≃ V ∗ . Concretely
put, one has the freedom of inserting a multiplicative constant (−i/~) in the isomorphism
i
I : Ve → V ∗ , ṽ 7→ − α(ṽ, ·) . (3.79)
~
By this choice of quantization constant we henceforth identify Ve ∋ ṽ ≡ I(ṽ) ∈ V ∗ .
Problem. In the context of our example (the symplectic plane), show that
i
I (eq − iJeq ) = (p + iJ t p) .
~
76
~ ∂
The interpretation of this result is that p = −i~ eq = i ∂q
.
Definition/Fact. The holomorphic Lagrangian subspace Ve ⊂ W ⊗ C carries a Hermitian scalar
product h : Ve × Ve → C by
i
h(ṽ, ṽ ′ ) ≡ h(w − iJw, w′ − iJw′ ) := α(w + iJw, w′ − iJw′ ) . (3.80)
~
As before in Section 3.4, the symbol δ(wV ) stands for the operator of derivation by the V -compo-
nent of w, and µ(wV ∗ ) means multiplication by the V ∗ -component of w.
Problem. Returning to our previous example, namely the symplectic plane W = spanR {eq , ep }
with mass parameter m and oscillator frequency ω, show that
( )
e ∗ 1 q ℓp
V ≡V =C·a , a = √+ +
−i .
2 ℓ ~
Using the isomorphism V ∗ ≃ Ve and the Hermitian scalar product h prove that a+ is a unit vector:
h(a+ , a+ ) = 1. Finally, use the Fock space scalar product induced by h to compute the Hermitian
adjoint of a+ and show that ( )
+ † 1 q ℓp
(a ) = a = √ +i .
2 ℓ ~
Beginning with a quick reminder of some relevant material from classical electrodynamics, we now
apply the general formalism of (second) quantization to the electromagnetic field.
∑
The electromagnetic gauge field is a one-form A = −Φ dt + Aj dxj . From it one obtains
the electromagnetic field strength, a two-form F = B + E ∧ dt , by taking the exterior derivative
F = dA . In components one has
∂Φ ∂Aj ∂Aj ∂Ai
Ej = − − , Bij = − . (3.83)
∂xj ∂t ∂xi ∂xj
∑ ∑
The components of the excitations D = i<j Dij dxi ∧ dxj and H = Hj dxj are given by
∑ ∑
Dij = ε0 ϵijk Ek , Bij = µ0 ϵijk Hk , (3.84)
k k
77
where ε0 and µ0 are called the dielectric constant and magnetic permeability of the vacuum,
respectively. A short-hand form of these so-called constitutive laws is
D = ε0 ⋆ E , B = µ0 ⋆ H , (3.85)
To quantize the theory following the general scheme of Section 3.7.1, we need a symplectic and a
complex structure on the space W of vacuum solutions. We begin with the symplectic structure.
Because every solution is determined by a set of initial data, say B|t=0 and D|t=0 , one can
compute the symplectic structure α as an integral over the Euclidean space R3 at the initial time
t = 0 . For mathematical simplicity, let us assume that the electromagnetic field is confined to
a bounded region U ⊂ R3 with Dirichlet boundary conditions D|∂U = 0 = B|∂U at the surface
∂U . (For a precise treatment, one might also take the domain U to be connected and simply
78
connected.) The following definition of the symplectic structure is modeled after the expression
α = dp ∧ dq (or α = p ∧ q = −q ∧ p in the case of a linear system) of Section 3.7.1.
Notation. If B is the two-form of the magnetic field strength (or, for that matter, any closed
two-form) then by d−1 B we mean any one-form A such that dA = B. Equivalently, using the
language of vector calculus, if B is the magnetic field (or, for that matter, any divergenceless
vector field) then by rot−1 B we mean any magnetic vector potential A such that rotA = B.
Fact. Let D and B be closed (i.e. dD = dB = 0). Then the integral
∫ ∫
−1
D∧d B = D · rot−1 B d3 x
U U
In vector notation, the same computation for the case of a gauge transformation A = A′ + gradf
looks as follows:
∫ ∫ ∫ ∫
′
D · (A − A ) d x =
3
D · gradf d x =
3
fD·d n− 2
f div D d3 x = 0 .
U U ∂U U
Remark 1. From magnetostatics one knows that if the equation B = rotA is solved for A in the
Coulomb gauge divA = 0 (or, in the language of differential forms, B = dA with d ⋆ A = 0) then
one has ∫
rotB(x′ ) 3 ′
A(x) = d x.
U 4π|x − x′ |
Thus with this choice of gauge we obtain the expression
∫ ∫
−1 D(x) · rotB(x′ ) 3 ′ 3
D · rot B d x =
3
d x d x.
U U 4π|x − x′ |
∫
Remark 2. Under certain special circumstances the integral D ∧ d−1 B has a beautiful ge-
ometric (actually, topological) interpretation. Suppose that both D and B are localized inside
e respectively. Suppose further that these tubular
tubular neighborhoods of two curves, say γ and γ
neighborhoods do not intersect each other. Then it is possible to show that our integral coincides
with the Gauss linking number G(γ, γ ′ ) of the two curves, multiplied by the product of electric
∫ ∫
(ΦD = D) and magnetic (ΦB = B) fluxes circulating in the tubular neighborhoods of the two
curves: ∫
D ∧ d−1 B = ΦD ΦB G(γ, γ ′ ).
U
79
Fact. The symplectic structure α on the space W of vacuum solutions, as represented by the
initial data D ≡ D|t=0 and B ≡ B|t=0 , is expressed by
∫
( )
′ ′
α (D, B), (D , B ) = (D ′ ∧ d−1 B − D ∧ d−1 B ′ )
∫U
= (D′ · rot−1 B − D · rot−1 B′ ) d3 x. (3.86)
U
In particular, α is non-degenerate on W .
How do we know that α as defined in (3.86) is the proper symplectic form to use? The answer
is that α in combination with the Hamiltonian function (the electromagnetic energy) gives the
correct Hamiltonian equations of motion, Ḃ = −dE and Ḋ = dH.
To conclude this subsection, we make a dimension check. Recall that D and B have the
physical dimensions
[D] = charge , [B] = [A] = energy/current . (3.87)
The complex structure on the space W of vacuum solutions is determined by considering the total
electromagnetic energy,
∫ ∫ ∫
1 1 1
H= (D ∧ E + B ∧ H) = |D| d x +
2 3
|B|2 d3 x ≥ 0 , (3.89)
2 U 2ε0 U 2µ0 U
which is the classical Hamiltonian function of the theory. The details are as follows.
The equations of motion Ḃ = −dE and Ḋ = dH for the electromagnetic field on the bounded
domain U have a fundamental system of periodic solutions, the so-called normal modes. Explicit
expressions for these can be given if U has a simple regular shape such as a cube or a ball;
otherwise one is only assured of their existence and needs a numerical algorithm (typically involving
discretization on a grid) to compute them. To find the normal modes, one looks for solutions of
the wave equation ( )
1 ∂2
− ∆ ft = 0 , ft ∂U = 0 , (3.90)
c2 ∂t2
√
of stationary type. (∆ is the Laplacian, and c = 1/ ε0 µ0 is the speed of light.) The ansatz
ft = f e−iωt leads to the Helmholtz equation
( )
ω 2 /c2 + ∆ f = 0 , ft ∂U = 0 . (3.91)
80
For the type of domain U which we envisage, solutions of this equation exist only for a discrete
set of characteristic frequencies ωλ .
Problem. Show that if D is a closed electric-type two-form solution of the Helmholtz equation
with frequency ω, then d ⋆ D = ε0 dE is a closed magnetic-type two-form solution with the same
frequency. In vector notation, the statement is that by taking the curl of a divergenceless vector
solution D one gets a divergenceless (axial) vector solution rot D .
In the following, let (Dλ , B λ ) denote the electric and magnetic initial data of any two-form
solution of the Helmholtz equation with frequency ωλ . Such a pair (Dλ , B λ ) is called a normal
mode. The normal modes for a fixed characteristic frequency ωλ form a vector space Wλ . If there
are no degeneracies in the spectrum of the Laplacian (that’s the generic situation for a domain U
of arbitrary shape which we henceforth assume) Wλ is two-dimensional. The symplectic form α
on W restricts to a non-degenerate symplectic form αλ on Wλ . In fact, one has
∫ ∫
( )
λ λ
α (0, d ⋆ D ), (D , 0) = D ∧ ⋆D =
λ λ
|Dλ |2 d3 x ̸= 0 .
U U
The complex structure J acts diagonally w.r.t. the decomposition W = ⊕Wλ by normal modes.
To compute its action on any one of the subspaces Wλ ⊂ W , one computes the time evolution of
the initial data (Dλ , B λ ) ⊂ Wλ for one quarter of the period Tλ = 2π/ωλ .
We are now in a position to get specific.
Definition/Fact. The complex structure acts on Wλ as
( ) ( )
( λ λ) d ⋆ Bλ d ⋆ Dλ ( λ λ) rot Bλ rot Dλ
J D ,B = , − , J D ,B = , − . (3.92)
ωλ µ0 ωλ ε 0 ωλ µ 0 ωλ ε 0
Problem. Verify from the equations of motion Ḋ = d ⋆ B/µ0 and Ḃ = −d ⋆ D/ε0 that J(Dλ , B λ )
as specified is indeed the quarter-period time evolution of (Dλ , B λ ). Show also that J 2 = −Id .
(Hint: for the latter, the key step is to show that (d⋆)2 agrees with the Laplacian −∆ on closed
two-forms. In vector notation this means that rot ◦ rot = −∆ on divergenceless vector fields.)
With the symplectic and complex structures in hand, we now take a look at the induced Euclidean
scalar product g. By implementing the general prescription of (3.75) we obtain
( ) ( )
g (Dλ , B λ ), (Dλ , B λ ) = α (Dλ , B λ ), J(Dλ , B λ )
∫ ∫
1 1 2
= D ∧ ⋆D +
λ λ
B λ ∧ ⋆B λ = H(Dλ , B λ ) ≥ 0 , (3.93)
ωλ ε 0 U ωλ µ0 U ωλ
for any normal mode (Dλ , B λ ) ∈ Wλ . Thus the Euclidean length squared of a normal mode is
equal to (twice) the electromagnetic energy of that mode divided by its characteristic frequency.
Problems. (i) Show that the normal mode subspaces Wλ are orthogonal to one another with re-
spect to the Euclidean structure g. (ii) Verify the validity of the invariance properties g(Jw, Jw′ ) =
g(w, w′ ) and α(Jw, Jw′ ) = α(w, w′ ) in the present context.
81
Next we use the complex structure J to polarize the complexified phase space W ⊗ C = V ⊕ Ve
by two complex Lagrangian subspaces V = ⊕Vλ and Ve = ⊕Veλ . This can be done separately for
each normal mode λ . Hence, for every λ let Veλ be the complex one-dimensional space
( )
Veλ = C · (Dλ , B λ ) − iJ(Dλ , B λ ) (3.94)
By recalling the Hermitian scalar product h which is given by Re h = 2g/~ and Im h = 2α/~
on the holomorphic subspace Ve (cf. Section 3.7.1), we see that the normalized mode ẽλ is a unit
vector of h. The corresponding generator of the anti-holomorphic part Vλ ⊂ Wλ ⊗ C is
1 ( )
eλ := √ (D0λ , B0λ ) + iJ(D0λ , B0λ ) . (3.97)
2
We now form the orthogonal sums V := ⊕λ Vλ and Ve := ⊕λ Veλ and thus have a polarization
W ⊗ C = V ⊕ Ve (3.98)
by Lagrangian holomorphic and anti-holomorphic subspaces. The Hilbert space of the quantized
electromagnetic field is the symmetric algebra S(Ve ) completed to an L2 -space by the Fock space
Hermitian scalar product due to h. The state vectors in the degree-n subspace Sn (Ve ) are called
n-photon states. As usual, the multiplication operators a+ := µ(ẽλ ) act on Fock space S(Ve ) as
λ
are photon annihilation operators. The Fock space has a vacuum 1 ∈ C = S0 (Ve ) which is denoted
by |0⟩ in Dirac notation. The commutation relations are the canonical ones (CCR):
[aλ , a+
λ′ ] = δλλ′ , [a+ +
λ , a λ′ ] = 0 , [aλ , aλ′ ] = 0 . (3.99)
We can now write down the quantized electromagnetic field in the form of a normal mode
expansion:
∑ ( ) ∑ ( )
b=
D eD + D
λ aλ + ẽλ aλ ,
b=
B eB + B
λ aλ + ẽλ aλ , (3.100)
λ λ
82
3.7.6 Casimir energy
We are now confronted with the same situation that already existed in the case of the Dirac field:
an infinite constant must be subtracted from the Hamiltonian H b in order to make it well-defined
on the Fock space S(Ve ). Indeed, applying (3.101) to the Fock vacuum |0⟩ , which is annihilated
by each of the aλ , we get an infinite ground state energy
∑
1
2
~ωλ = ∞ . (3.102)
λ
The infinite constant to be subtracted is universal in the sense that it does not depend on the
cavity U under consideration: it is the infinite ground state energy of the electromagnetic field
for the basic situation that U covers all of space and the boundary conditions are pushed to the
edge of the universe. For the case of a finite cavity U , this universal vacuum energy differs by a
finite amount EC from the infinite ground state energy (3.102). Thus the correct formula for the
quantum Hamiltonian Hb=H bU is
∑
bU =
H ~ωλ a+
λ λ aλ + EC (U ) . (3.103)
EC is called the Casimir energy. We mention in passing that over the last decade, advanced
techniques have been developed to compute EC (U ) accurately for a cavity U of general shape.
A simple dimensional estimate of EC can be had for the case of a cavity of high symmetry, say
a ball BL of radius L. For this purpose note that the characteristic frequencies ωλ secretly carry
a factor of c (speed of light). Therefore EC must be proportional to ~c, which has the physical
dimension of energy×length. Since the ball radius L is the only length scale in the problem, there
exists only one energy that can be formed:
~c
EC (BL ) ∼ . (3.104)
L
Thus EC (BL ) = K~c/L, where K is a constant. K turns out to be negative.
The Casimir energy tends to lead to an attractive force between surfaces. This so-called
Casimir effect is predicted to be a challenge for future developments of nano-technology, as the
attractive Casimir force makes nano-particles clump together, whereas the nano-engineer would
like them to be freely movable.
Let us now consider a geometry which is simple enough to make it possible for the mode expansion
to be given in explicit form: we choose for U a 3-torus T3 , i.e., a cubic domain
with periodic boundary conditions in each direction. Aside from being a mathematical idealization
that does not correspond to the real world, this choice has a certain drawback: on T3 there exist
harmonic two-forms, i.e., solutions of the Laplace equation ∆D = 0 = ∆B. Such solutions require
a separate treatment beyond that of Section 3.7.5. For brevity, we will ignore this issue here.
83
Our main task in this section is to construct closed two-form solutions of the Helmholtz equa-
tion and thus the unit normal modes eλ and ẽλ . It is a standard fact that the Helmholtz equation
(∆ + k 2 )f = 0 for functions f on T3 has a complete orthonormal system of solutions by plane
waves
fk := (L1 L2 L3 )−1/2 ei(k1 x1 +k2 x2 +k3 x3 ) , ki ∈ 2πZ/Li .
To obtain the desired solutions for D and B, we proceed as follows. For every wave vector
(actually, wave form) k = (k1 , k2 , k3 ) we choose two so-called polarization unit vectors ϵs (k)
(s = 1, 2) with two properties: (i) they form an orthonormal basis of the plane perpendicular to
the wave vector k (that’s the null plane of the wave form k), and (ii) the 3-bein (ϵ1 (k), ϵ2 (k), k)
is a right-handed system. (Let us mention in passing that such a choice does not exist in a
globally smooth manner in the k-continuum. Indeed, on topological grounds any smooth vector
field k 7→ ϵs (k) tangent to the two-sphere must have at least two zeroes. This obstacle is of no
concern to us here, as our set of wave vectors is discrete due the periodic boundary conditions.)
For every (k, s) we then introduce a basic, spatially constant two-form
Problem. Show that the two-form bk, s := fk βk, s is closed (dbk, s = 0) and solves the Helmholtz
equation (∆ + ω 2 /c2 ) bk, s = 0 with characteristic frequency ω = |k| c .
Using the information gained from the Problem, one shows that the unit normal modes are
√ d ⋆ eB
k, s
eB
k, s :=
1
µ ~ c|k|
2 0
bk, s , ẽB B
k, s = ek, s , eD
k, s = i , ẽD D
k, s = ek, s . (3.106)
|k|cµ0
b = charge
Problems. (i) Show that the quantized fields have the correct physical dimensions [D]
b = energy/current. (ii) What special choice(s) did we make in (3.106) ?
and [B]
div A = 0 or d ⋆ A = 0 . (3.108)
84
By direct transcription from (3.107) one then has the mode expansion
∑ ( )
b=
A eA + A
k, s ak, s + ẽk, s ak, s ,
k, s
√
µ0 ~ c fk ∑
3
eA
k, s = (−1)2−s ϵ3−s (k)j dxj , ẽA A
k, s = ek, s . (3.109)
2|k| i j=1
In the presence of electric charges it is no longer possible to choose the gauge Φ = 0 . Instead,
one takes Φ to be the (instantaneous) Coulomb potential determined by Poisson’s equation
where the quantum field Db of the electric excitation is subject to the Gauss-law constraint dD
b = ρb .
∑b
The matter-field interaction here is in the term Al bjl coupling the magnetic vector potential to
the electric current density. By inserting the respective mode expansions, one obtains contributions
to the interaction Hamiltonian of the schematic form (omitting spin and polarization)
The first of these creates an electron with momentum k1 and a positron with momentum k2 while
annihilating a photon with momentum k1 + k2 . The second does the reverse (or adjoint). There
are also terms of the schematic form
and similar with c+ → c− . These cause scattering of electrons and positrons by the process of
absorption or emission of a photon.
of Fock spaces for fermions and bosons, and the Hamiltonian is a sum of three terms:
85
the fermionic Fock space ∧(VF ) and is trivial on S(VB ). (For many purposes it is of course good
enough to treat the Dirac field by non-relativistic reduction in the Schrödinger approximation.)
The radiation part (the Hamiltonian for the free electromagnetic field) operates on the bosonic
∫
Fock space S(VB ) and is trivial on ∧(VF ). The coupling Hamiltonian jl Al d3 x has factors jl
operating on ∧(VF ) and factors Al operating on S(VB ).
Consider now some excited state of matter, ψi , which decays to another state ψf (e.g., the
ground state) by the emission of a single photon with wave vector k. The transition or decay rate
( )
Γ ψf ⊗ 1γ(k) ← ψi ⊗ 0γ can be computed by using Fermi’s golden rule:
( ) 2π ⟨ ⟩
Γ ψf ⊗ 1γ(k) ← ψi ⊗ 0γ = ψf ⊗ 1γ(k) | Hc | ψi ⊗ 0γ 2 δ(Ef + ~ωk − Ei ) , (3.116)
~
∫∑
where Hc ≡ Hcoupling = jl Al d3 x is the interaction part of the Hamiltonian. The transition
matrix element
∫
⟨ ⟩ ⟨ ⟩⟨ ⟩
ψf ⊗ 1γ(k) | Hc | ψi ⊗ 0γ = d3 x ψf | jl (x) | ψi 1-photon(k) | Al (x) | 0-photon (3.117)
⟨ ⟩
is essentially the Fourier transform of the transition current density ψf | jl (x) | ψi since
⟨ ⟩
1-photon(k) | Al (x) | 0-photon ∼ eik·x . (3.118)
The typical situation is that this Fourier transform can be calculated by multipole expansion. Let
us look at the case of a heavy atomic nucleus, for example. The wave length of the emitted photon
is
2π ~c 200 MeV fm
λ= = 2π ≈ 2π . (3.119)
|k| ~ωk Ei − Ef
By inserting the typical excitation energy of a low-lying nuclear excited state, one gets a value for
λ which is very much bigger than the radius R ∼ 5 fm of a heavy nucleus such as 208 Pb . Therefore,
if the transition operator jl (x) is expanded in multipoles, the leading contributions to the decay
rate come from the multipoles of lowest order which are compatible with the angular momenta of
the initial and final states of the decay.
The multipole expansion proceeds in terms of so-called tensor operators TJM . Learning more
about these is a major motivation for the next chapter.
86
momenta separates as a product of two factors: a so-called Wigner 3j-symbol (or Clebsch-Gordan
coefficient) determined by geometry, and a reduced matrix element containing the information
about the intrinsic structure of the many-body states.
In mathematical parlance one says that ρ is a group homomorphism from G into GL(V ). The
trivial representation is the representation ρ(g) ≡ IdV mapping all group elements to the identity.
Examples.
1. The simplest group is G = {e, π} with multiplication table e2 = e, eπ = πe = π, and π 2 = e.
This group has a representation on V = R by ρ(e) = +1 and ρ(π) = −1.
2. Let G be a finite group, and take V to be the vector space of all scalar-valued functions
f : V → K . The right regular representation, R ≡ ρR , of G is defined as
( )
R : G → GL(V ) , R(g0 )f (g) := f (gg0 ) . (4.2)
3. Let S2 ⊂ R3 be the unit sphere in three dimensions. The rotation group SO3 acts on S2 in the
fundamental way by
SO3 × S2 → S2 , (g, v) 7→ g · v .
This action yields a representation of SO3 , say on the Hilbert space L2 (S2 ) of square integrable
functions f : S2 → C, by
)
(ρ(g)f (v) := f (g −1 · v) . (4.4)
87
Since the operators π(h) are linear transformations of a vector space, our space of functions
V := Func(G, W )H is still a vector space. V carries a representation ρ of G by left multiplication:
Examples.
1. If you take for H the trivial group H = {e} then IndG
{e} (trivial) = L is the left regular
representation of G.
2. Let G = SO3 , and let H = SO2 ⊂ G be the subgroup of rotations about a fixed axis, say the z-
axis. Take π to be the trivial H-representation, and let W = K = C . Then Func(SO3 , C)SO2 can
be identified with the vector space of complex-valued functions on the unit sphere S2 ≃ SO3 /SO2 .
If we require the functions to be square-integrable, the induced representation IndG
H (π) is the
where d2 v = sin θ dθ dϕ is the solid-angle two-form. The SO3 -representation given by the space
L2 (S2 ) of square-integrable functions on S2 is unitary:
∫ ∫
−1
⟨ρ(g)f1 , ρ(g)f2 ⟩ = f2 (g · v) f1 (g · v) d v =
−1 2
f2 (v ′ ) f1 (v ′ ) d2 v ′ = ⟨f1 , f2 ⟩ .
S2 S2
The second equality follows from the substitution v ′ := g −1 · v by the SO3 -invariance of d2 v.
Next we explain what is meant by saying that a representation is irreducible. For this purpose
we need the following preparation of language. Let ρ : G → GL(V ) be a representation. Then a
vector subspace U ⊂ V is called G-invariant if
∀g ∈ G : ρ(g)U ⊂ U . (4.9)
88
U is called a proper subspace of V if U is neither the zero vector space nor identical with V .
Definition. A representation ρ : G → GL(V ) is called irreducible if there exist no G-invariant
proper vector subspaces of V .
Example. The unitary representation of the previous example is not irreducible as the vector
space of, say, the constant functions is an SO3 -invariant proper subspace of L2 (S2 ) of dimension
one. However, irreducible SO3 -representations do arise from it by restriction to subspaces of
definite total angular momentum. Let
where YL,M denotes the spherical harmonic of total angular momentum L and magnetic quantum
number M . One has dim VL = 2L + 1. The SO3 -representation ρL : SO3 → GL(VL ) by
∑
L
(ρL (g)YL,M )(v) := YL,M (g −1 · v) = YL,M ′ (v) DM
L
′ ,M (g) , (4.11)
M ′ =−L
∑
L
DM
L
′ ,M ′′ (g1 g2 ) = DM
L
′ ,M (g1 ) DM,M ′′ (g2 ) ,
L
(4.12)
M =−L
89
so ρ1 (g)v ∈ ker φ . Now since V1 is irreducible, the G-invariant subspace ker φ by definition cannot
be a proper subspace. Thus there exist only two possibilities: (i) ker φ ≡ V1 , in which case φ
is the zero map, or (ii) ker φ ≡ {0}, in which case φ is injective. In the latter case φ is also
surjective. Indeed, by similar reasoning as before, im φ ⊂ V2 is a G-invariant subspace of the
irreducible G-representation space V2 , so either im φ ≡ V2 or im φ = {0}. The latter possibility is
ruled out by the known injectivity of φ for the case (ii) under consideration.
Corollary. Let (V, ρ) be a finite-dimensional irreducible G-representation over the complex num-
bers C. Then every C-linear map φ ∈ End(V ) with the property φ ◦ ρ(g) = ρ(g) ◦ φ (for all g ∈ G)
is a scalar multiple of the identity.
Proof. Over the complex number field, any linear transformation φ ∈ End(V ) has at least
one eigenvalue, say λ . By the definition of what it means for λ to be an eigenvalue, the linear
transformation φ−λ·IdV is not invertible. On the other hand, φ−λ·IdV intertwines the irreducible
representation (V, ρ) with itself by assumption. Therefore, Schur’s lemma implies that φ − λ · IdV
must be identically zero. Hence φ = λ · IdV as claimed.
Problem. If ρj : G → GL(Vj ) (j = 1, 2) are two finite-dimensional irreducible representations
over C , show that the linear space HomG (V1 , V2 ) of intertwiners φ is at most one-dimensional.
The equivalence classes [g; w] ≡ [gh−1 ; π(h)w] of this H-action constitute a so-called vector bundle,
G ×H W , over the base manifold G/H. The equivalence classes [g; w] for a fixed coset gH ∈ G/H
form a vector space which is isomorphic to W by [g; w] 7→ w; it is called the fiber over gH. A
section of the vector bundle is a mapping s : G/H → G ×H W with the property that for each
gH ∈ G/H the value s(gH) of the section is a vector [g; w] in the fiber over gH. (Depending on
the situation, one may require that s is continuous, or differentiable, or analytic, or holomorphic,
etc.). The space of sections s of the vector bundle G ×H W is denoted by Γ(G ×H W ).
Problem. Show that the mapping Func(G, W )H → Γ(G ×H W ), f 7→ s, defined by s(gH) =
[g; f (g)], is a bijection.
Thus we have an alternative way of thinking about the induced representation IndG
H (π): we
90
maximal Abelian subgroup (a so-called maximal torus). The coset space G/T can then be shown to
be a complex space. (In fact, G/T is a phase space in the sense of Section 3.7.1: it has the attributes
of symplectic and complex structure which are required for the geometric quantization procedure
described there.) If one is now given any one-dimensional representation π : T → GL(W ) on
W = C, then one has a vector bundle G ×T W ; it is called a complex line bundle because its fibres
are copies of C. The sections of such a line bundle are maps from a complex space into another
complex space. One can therefore speak of the subspace of holomorphic sections.
The following theorem holds in the mathematical setting of a connected, semisimple, compact
Lie group G as described above.
Borel-Weil Theorem. There is a one-to-one correspondence between (isomorphism classes of)
irreducible representations of G and the spaces of holomorphic sections of complex line bundles
over G/T . In particular, for every irreducible G-representation ρ : G → GL(V ) there exists a
one-dimensional T -representation π : T → GL(W ) such that ρ can be identified with the induced
representation IndG
T (π) restricted to the subspace of holomorphic sections of the complex line
bundle G ×T W .
Example. Every irreducible representation (VL , ρL ) of the rotation group G = SO3 arises in
this way. For this purpose one takes T := SO2 , the subgroup of rotations about the z-axis, and
has G/T = SO3 /SO2 = S2 , the unit sphere, which is a complex space. (The complex structure
J is rotation by π/2 = 90o in the tangent plane). To get the representation (VL , ρL ) for any
total angular momentum L ∈ N ∪ {0}, one takes π : SO2 → GL(W ) to be the one-dimensional
representation on the space of states W := C · |L, L⟩} with maximal projection of the angular
momentum:
( )
π Rz (ϕ) = eiLϕ ,
where Rz (ϕ) is the rotation about the z-axis with rotation angle ϕ . The (2L + 1)-dimensional
irreducible representation (VL , ρL ) can then be seen as the space of holomorphic sections of the
complex line bundle SO3 ×SO2 C · |L, L⟩. Details of its construction (using ladder operators L+ and
L− , but omitting the relation to holomorphic sections of a complex line bundle) were discussed
in the basic course on quantum mechanics. In particular, the 3-dimensional SO3 -representation
corresponding to angular momentum L = 1 is the same as the representation of SO3 on the space
of holomorphic vector fields or sections of the tangent bundle
An example of a holomorphic vector field is the vector field of infinitesimal rotation about the
z-axis (or any other axis for that matter).
V := V1 ⊗ V2 ⊗ · · · ⊗ Vn (4.16)
91
is a G-representation space in the natural way:
This definition makes sense because the operators ρj (g) are linear transformations.
A tensor of the form v1 ⊗ v2 ⊗ · · · ⊗ vn is called a pure tensor. The most general tensor T ∈ V
(j)
is a linear combination of pure tensors. In fact, if {ei } is a basis of Vj we can express T as
∑ (1) (2) (n)
T = Ti1 , i2 ,..., in ei1 ⊗ ei2 ⊗ · · · ein . (4.18)
∑
3
Q= fi ⊗ fi ,
i=1
and this satisfies g · Q = Q due to Q(gv, gv ′ ) = Q(v, v ′ ) for g ∈ SO3 . Another SO3 -invariant
tensor is the triple product Ω ∈ V ∗ ⊗ V ∗ ⊗ V ∗ . While the Euclidean scalar product is a symmetric
tensor, Ω is totally antisymmetric:
∑ ∑
Ω = f1 ∧ f2 ∧ f3 = sign(π) fπ(1) ⊗ fπ(2) ⊗ fπ(3) = ϵijk fi ⊗ fj ⊗ fk .
π∈S3
The situation for degree 2 is rather easy to understand by using Schur’s lemma. To that end we
recall from Section 2.8.1 the isomorphism
which is determined by
µ(v ⊗ f )(u) := f (u) v . (4.20)
If V1 and V2 are G-representation spaces, the mapping µ takes G-invariant tensors T ∈ V2 ⊗ V1∗
into G-equivariant linear transformations µ(T ) ∈ HomG (V1 , V2 ). Indeed:
Problem. Show that the invariance property g · T = T for T ∈ V2 ⊗ V1∗ translates via µ into the
intertwining property ρ2 (g)µ(T ) = µ(T )ρ1 (g).
92
In the special case of V1 = V2 = V the isomorphism µ : V ⊗ V ∗ → Hom(V, V ) = End(V )
provides us with a canonical G-invariant tensor T ∈ V ⊗ V ∗ . This is the inverse image of the
identity:
∑
µ−1 (IdV ) = ei ⊗ fi , (4.21)
i
where {ei } and {fi } are bases of V resp. V ∗ which are dual to each other by fi (ej ) = δij . This
G-invariant tensor µ−1 (IdV ) is canonical in the sense that it exists for any representation (V, ρ).
Next, consider the general case of V1 ̸= V2 and assume that both representation spaces V1
and V2 are G-irreducible. Then there exist only two possibilities for the space HomG (V1 , V2 ) of
intertwiners: if V1 and V2 belong to different isomorphism classes, then HomG (V1 , V2 ) = 0 ; and if
V1 and V2 are isomorphic, then HomG (V1 , V2 ) ≃ C by Schur’s lemma (cf. the Problem at the end
of Section 4.2). In the former case V2 ⊗ V1∗ has no G-invariant tensors; in the latter case there
exists a complex line of G-invariant tensors.
In the following we focus on the important case of V2 = V and V1 = V ∗ ; i.e., we look for
G-invariant tensors in V ∗ ⊗ V ∗ (or V ⊗ V ; it makes no big difference). This case is relevant, e.g.,
for the question whether (and if so, how) a rotationally invariant state can be formed from a pair
of states carrying angular momentum L in both cases.
We will use the fact that by the isomorphism µ : V ∗ ⊗ V ∗ → Hom(V, V ∗ ) the G-invariant
tensors T ∈ V ∗ ⊗V ∗ are in one-to-one correspondence with intertwiners µ(T ) ≡ φ ∈ HomG (V, V ∗ ).
Concerning the latter recall that HomG (V, V ∗ ) ≃ C if V and V ∗ are irreducible and belong to
the same isomorphism class. Let SymG (V, V ∗ ) and AltG (V, V ∗ ) denote the G-invariant linear
transformations from V to V ∗ which are symmetric resp. skew (or alternating).
Lemma. If an irreducible G-representation V is isomorphic to its dual V ∗ , then one has either
(i) HomG (V, V ∗ ) = SymG (V, V ∗ ) ≃ C or (ii) HomG (V, V ∗ ) = AltG (V, V ∗ ) ≃ C .
Proof. We recall from Section 1.7.1 that for a linear transformation φ : V → V ∗ there exists a
canonical adjoint φt : V → V ∗ (the transpose). If 0 ̸= φ ∈ HomG (V, V ∗ ) is an intertwiner, let
φs := 21 (φ + φt ) and φa := 12 (φ − φt ) be the symmetric and alternating parts of φ . Both φs and
φa intertwine the G-representation on V with the G-representation on V ∗ . Since V is irreducible,
we know that the space of such intertwiners is one-dimensional. Now φs = +φts and φa = −φta
cannot both be non-zero and lie on the same complex line. Therefore we must have either (i)
φa = 0 or (ii) φs = 0 .
The next statement is an immediate deduction from the isomorphism V ∗ ⊗ V ∗ ≃ Hom(V, V ∗ )
of G-representation spaces and the fact that a G-invariant tensor T ∈ V ∗ ⊗ V ∗ is the same as a
G-invariant bilinear form Q : V ⊗ V → C .
Corollary. If a complex vector space V is an irreducible representation space for a group G,
then there exists a trichotomy of possibilities for the space, say Q, of G-invariant bilinear forms
Q : V ⊗ V → C: (i) Q = 0 is trivial, or (ii) Q is generated by a symmetric bilinear form, or (iii)
Q is generated by an alternating bilinear form.
93
4.3.2 Example: SO3 -invariant tensors of degree 2
To illustrate the general statements of Section 4.3.1, we now specialize to the case of G = SO3 .
We have already met the irreducible SO3 -representations on the vector spaces
which are spanned by the spherical harmonics YL,M of total angular momentum L.
Fact. The tensor product VL ⊗ VL of irreducible representations of angular momentum L contains
an SO3 -invariant symmetric tensor,
∑
+L
T = (−1)M YL,M ⊗ YL,−M . (4.22)
M =−L
Alas, the Hermitian scalar product alone cannot deliver an invariant tensor in the sense of the
Definition given at the beginning of Section 4.3, as it involves an operation of complex conjugation
in the left argument. To repair this feature and produce a complex bilinear form
Q : V ⊗ V → C, v1 ⊗ v2 7→ ⟨τ v1 , v2 ⟩ , (4.23)
We claim that when the formula (4.24) for the invariant tensor T ∈ V ∗ ⊗ V ∗ is specialized to the
case of G = SO3 and V ∗ = VL , one arrives at the invariant tensor T ∈ VL ⊗ VL of (4.22).
94
What is still missing for this argument is the origin and nature of the SO3 -invariant anti-unitary
operator τ : VL → VL needed for the construction of T . In the remainder of the subsection, we
shall introduce τ . The point here will be that, although the SO3 -representation space VL spanned
by the degree-L spherical harmonics is complex, this representation can actually be constructed
entirely over the real numbers (without ever using C). We now indicate how this goes.
The one-dimensional representation space V0,R for L = 0 is simply given by the constant
functions on the unit sphere S 2 . The three-dimensional representation space V1,R ≃ R3 for L = 1
is spanned by the three Cartesian coordinate functions x1 , x2 , x3 restricted to S 2 :
x1 S 2 = sin θ cos ϕ , x2 S 2 = sin θ sin ϕ , x3 S 2 = cos θ .
To construct the five-dimensional irreducible representation V2,R for L = 2, one starts from the
six-dimensional space S2 (V1,R ) (degree-2 part of the symmetric algebra) of functions xi xj for i ≤ j .
This space contains a generator x21 + x22 + x23 which becomes trivial upon restriction to S 2 . One
therefore removes it by passing to a quotient of vector spaces:
/
V2,R := spanR {xi xj }1≤i≤j≤3 R · (x21 + x22 + x23 ) .
By now it should be clear how to continue this construction. For a general value L ≥ 2 of the
angular momentum, one takes the space SL (V1,R ) of degree-L monomials and quotients out the
subspace SL−2 (V1,R ) · (x21 + x22 + x23 ) ⊂ SL (V1,R ) :
/
VL,R := SL (V1,R ) SL−2 (V1,R ) · (x21 + x22 + x23 ) . (4.26)
95
Now the complex representation spaces VL are obtained by complexification:
and the action SO3 × VL → VL is defined by C-linear extension of the action SO3 × VL,R → VL,R :
For example, for L = 1 one has Y1,0 ∝ x3 , Y1,1 ∝ x1 + ix2 , Y1,−1 ∝ −x1 + ix2 .
We are now in a position to say what is the complex anti-linear operator τ needed for the
construction of the SO3 -invariant tensor (4.22): it is just the well-defined operation of complex
conjugation with respect to the real structure VL,R ⊂ VL :
τ (u + iv) := u − iv . (4.30)
Defined in this way, the operator τ commutes with the G-action on VL because the latter is
already defined on the real subspace VL,R ⊂ VL . Here ends our effort to explain the mathematical
background behind (4.22).
We have seen that all SO3 -invariant tensors of degree 2 are symmetric. For further illustration
of the Corollary at the end of Section 4.3.1, we will now describe a related situation in degree 2
where one also finds skew-symmetric tensors.
Our example will be built on the compact Lie group SU2 of special unitary 2 × 2 matrices,
The coset space G/T = SU2 /U1 then is a complex manifold isomorphic to the two-sphere S 2 .
Consider the representation πN : U1 → GL(C) by
( )
α 0
πN = αN (N ∈ N ∪ {0}) . (4.34)
0 ᾱ
96
Then according to Section 4.2 there exists an irreducible SU2 -representation on the space of
holomorphic sections of the complex line bundle SU2 ×U1 C over S 2 , where U1 acts on C by πN .
By the Borel-Weil theorem, all irreducible representations of SU2 arise in this way as N = 0, 1, 2, . . .
runs through the non-negative integers.
We give a short cut to the outcome of the Borel-Weil construction, as follows. Let SU2 act on
a fundamental pair of coordinate functions z1 , z2 for C2 by
( ) ( )( ) ( )
z1 α β z1 αz1 + βz2
g· ≡ := . (4.35)
z2 −β̄ ᾱ z2 −β̄z1 + ᾱz2
∑
N
P (z1 , z2 ) = an z1N −n z2n = z1N P (1, z2 /z1 )
n=0
ᾱz − β̄ β̄ + αz
g·z = , g −1 · z = .
βz + α ᾱ − βz
By setting P (z1 , z2 ) = z1N φ(z2 /z1 ) one then obtains an SU2 -action ρN on holomorphic polynomials
∑
φ(z) = N n
n=0 an z as ( )
( ) β̄ + αz
ρN (g)φ (z) := (ᾱ − βz) φ
N
, (4.36)
ᾱ − βz
which turns out to be none other than the induced SU2 -representation on holomorphic sections φ
of the complex line bundle SU2 ×U1 C given by the U1 -representation πN .
Let us look at the infinitesimal or Lie algebra form of this representation. The representation
of the Lie algebra su2 = Lie(SU2 ) is defined by linearization at unity,
d ( )
(ρN ∗ (X)φ)(z) := ρN (e )φ (z) .
tX
(4.37)
dt t=0
For the generators of su2 = Lie(SU2 ) we take iσ1 , iσ2 , iσ3 where σj are the standard Pauli
matrices. Actually, it is more convenient to work in the complexification sl2 = su2 ⊕ i su2 , as
this allows us to present the answer in terms of the sl2 -generator σ3 and the ladder operators
σ± = 21 (σ1 ± iσ2 ). A short computation gives
d d d
ρN ∗ (σ3 ) = 2z − N, ρN ∗ (σ+ ) = −z 2 + Nz , ρN ∗ (σ− ) = . (4.38)
dz dz dz
We observe that for N = 0 there exists only one function φ(z) = const, which spans the trivial
representation. For N = 1 we have a pair of basis functions
( ) ( )
1 ∧ 0 ∧
= {z 7→ φ↑ (z) = z} , = {z 7→ φ↓ (z) = 1} .
0 1
97
For a general value of N ≥ 0 our representation space is spanned by the functions φN, n (z) = z n
for n = 0, 1, . . . , N . We denote it by VN ≡ VNSU2 . Notice that
ρN ∗ (σ3 )φN, n = (2n − N )φN, n , ρN ∗ (σ+ )φN, n = (N − n)φN, n+1 , ρN ∗ (σ− )φN, n = n φN, n−1 .
In particular, ρN ∗ (σ+ )φN, N = 0 and ρN ∗ (σ− )φN, 0 = 0 ; these are called the states of highest resp.
lowest weight of the representation VN .
Problem. Show that the representation space VN has no SU2 -invariant proper subspaces (the
implication being that the representation on VN is irreducible).
We now turn to the subject proper of this subsection: the construction of an SU2 -invariant
tensor in the tensor product VN ⊗ VN . From the general theory of Section 4.3.1 we know that
there exists at most one such tensor (up to multiplication by scalars). Let us find one.
By linearization of the group action (4.17) on tensors, any Lie algebra element X ∈ su2 acts
on VN ⊗ VN as
(2)
ρN ∗ (X) = ρN ∗ (X) ⊗ IdVN + IdVN ⊗ ρN ∗ (X) . (4.39)
Now the elements of the tensor product VN ⊗ VN can be realized as polynomials in two complex
variables z and z ′ . Adopting this realization we get from (4.38) the expressions
d d
+ 2z ′ ′ − 2N ,
(2)
ρN ∗ (σ3 ) = 2z
dz dz
d 2 d d d
ρN ∗ (σ+ ) = −z 2 − z ′ + N (z + z ′ ) ,
(2) (2)
ρN ∗ (σ− ) = + ′ . (4.40)
dz dz ′ dz dz
We then look for an SU2 -invariant tensor T ∈ VN ⊗ VN by the ansatz
∑
N
aN ; m, n z n z ′
m
T = (4.41)
m, n=0
T = (z − z ′ )N . (4.42)
From it we immediately see that our invariant tensor T is symmetric for even N and skew for odd
N . This concludes our illustration of the Corollary (end of Section 4.3.1) at the example of SU2 .
98
We finish the subsection by relating the result (4.42) to that of Section 4.3.2. Although the Lie
groups SU2 and SO3 have identical Lie algebras su2 ≃ so3 , they are not the same group. It turns
out that the relation between them is qualitatively the same as the relation between Spin2n and
SO2n (cf. Section 2.11). As a matter of fact, SU2 = Spin3 . (In other words, SU2 = Spin3 can be
obtained by exponentiating the Lie algebra su2 = spin3 ≃ Cl2 (R3 ) of skew-symmetrized degree-2
elements in the Clifford algebra of the Euclidean vector space R3 .)
This means that there exists a 2 : 1 covering map SU2 → SO3 . It is constructed just like
the 2 : 1 covering map Spin2n → SO2n . Let us recall this construction. We conjugate the Pauli
matrices σj (taking the role of the gamma matrices γ µ ) by g ∈ SU2 . The result gσj g −1 of
conjugation is a traceless Hermitian 2 × 2 matrix. It is therefore expressible in terms of the Pauli
matrices:
∑
gσj g −1 = σi Rij (g) . (4.43)
i
The real expansion coefficients can be arranged as a real 3 × 3 matrix R(g) := {Rij (g)}i,j=1,2,3 .
By construction, the correspondence g 7→ R(g) is a group homomorphism: R(g1 g2 ) = R(g1 )R(g2 ).
Problem. Show that the linear transformation R(g) with matrix elements Rij (g) has the prop-
erties R(g)t = R(g)−1 and Det R(g) = 1 of a rotation r ≡ R(g) ∈ SO3 .
From the definition (4.43) one sees that R(−g) = R(+g). This is good evidence for the true
fact (not proved here) that R : SU2 → SO3 is a 2 : 1 covering of Lie groups. By this covering map,
every representation r 7→ ρL (r) of SO3 corresponds to a representation g 7→ ρL (R(g)) of SU2 . One
SU2
actually has V2L ≃ VLSO3 , expressing the fact that the angular momentum L translates to the
quantum number N = 2L in the present notation.
The converse is not true: the SU2 -representations for odd N do not correspond to rep-
resentations of SO3 . Morally speaking, they carry half-integer angular momentum (or spin)
S = N/2 ∈ Z + 21 . In physics texts it is sometimes said that they are ‘double-valued’ representa-
tions of SO3 .
After this rather exhaustive discussion of the situation for degree 2, we turn to degree 3. Here
our main analytical tool will be the so-called Haar measure. In the following statement, the word
‘compact Lie group’ is to be interpreted in its widest sense (which includes finite groups).
Fact. For every compact Lie group G there exists a measure dg (called Haar measure) with the
properties of invariance under right and left translation by any group element g0 ∈ G :
∫ ∫ ∫
f (g) dg = f (gg0 ) dg = f (g0 g) dg . (4.44)
G G G
99
Examples.
1. In the case of a finite group G of order ord(G), the integral with Haar measure is a sum
∫
1 ∑
f (g) dg := f (g) . (4.45)
G ord(G) g∈G
The invariance (4.44) under right and left translations here follows directly from the fact that the
mappings g 7→ gg0 and g 7→ g0 g are one-to-one.
2. Let G = U1 = {z ∈ C | z̄z = 1} be the group of unitary numbers in C . Writing z = eiφ , the
unit-mass Haar measure for U1 is given by dφ/2π:
∫ ∫ 2π ∫ 2π
1 iφ 1
f (g) dg = f (e ) dφ = f (eiφ+iφ0 ) dφ . (4.46)
U1 2π 0 2π 0
where Rj (α) means an angle-α rotation around the j-axis. The Haar measure for SO3 in these
coordinates is expressed by
∫ ∫2π∫π∫2π
1 ( )
f (g) dg = 2 f R3 (ϕ)R1 (θ)R3 (ψ) dψ sin θ dθ dϕ . (4.48)
SO3 8π
0 0 0
The proof of invariance requires some knowledge of the theory of Lie groups.
Let now G be a compact Lie group with Haar measure dg and consider the tensor product
V1 ⊗ V2 ⊗ V3 of three (finite-dimensional) representations ρl : G → GL(Vl ) (l = 1, 2, 3). In terms
of bases for the factors Vl , the most general tensor T ∈ V1 ⊗ V2 ⊗ V3 is expressed as
∑ (1) (2) (3)
T = Tijk ei ⊗ ej ⊗ ek .
i,j,k
We can produce from T a G-invariant tensor Tav by taking the Haar average of its G-translates:
∫ ∑ ∫
( (1) ) ( (2) ) ( (3) )
Tav := (g · T ) dg = Tijk ρ1 (g)ei ⊗ ρ2 (g)ej ⊗ ρ3 (g)ek dg . (4.49)
G i,j,k G
Thus µ restricts to an isomorphism between the vector space (V1 ⊗V2 ⊗V3 )G of G-invariant tensors
and the vector space HomG (V3∗ , V1 ⊗ V2 ) of G-equivariant homomorphisms.
100
4.4 The case of SU2 : Wigner 3j-symbol, Clebsch-Gordan coefficient
We now specialize to the case of G = SU2 = Spin3 and recall from Section 4.3.3 that the space
The following can be regarded as a statement about the dimension of HomSU2 (VN∗3 , VN1 ⊗ VN2 ).
Fact. The tensor product VN1 ⊗ VN2 of two irreducible SU2 -representation spaces decomposes as
To verify this, one starts from the general result (not proved here) that a tensor product of
irreducible representations for a compact Lie group G is (isomorphic to) a direct sum of finitely
many irreducible G-representations. In the present case, the irreducible summands occurring in
VN1 ⊗ VN2 can be identified by the following counting procedure. Let
be a basis of homogeneous polynomials for VN1 ⊗ VN2 , and let V k ⊂ VN1 ⊗ VN2 be the eigenspace
of eigenvalue k for the (shifted) polynomial-degree operator
∂ ∂
(ρN1 ∗ + ρN2 ∗ )(σ3 ) = 2z1 + 2z2 − N1 − N2 .
∂z1 ∂z2
The next table lists the dimension of V k as a function of k :
k dim V k
−N1 − N2 1
−N1 − N2 + 2 2
.. ..
. .
−N1 − N2 + 2 min {N1 , N2 } min{N1 , N2 } + 1
.. ..
. .
−N1 − N2 + 2 max {N1 , N2 } min{N1 , N2 } + 1
.. ..
. .
N1 + N2 − 2 2
N1 + N2 1
We see that dim V k increases linearly with k until it reaches a plateau extending from min{N1 , N2 }
to max{N1 , N2 }; after that it decreases linearly in such a way that dim V k = dim V −k .
101
By inspection of these dimensions it follows that VN1 ⊗ VN2 contains one multiplet of states
with spin Jmax = 21 (N1 + N2 ), one with spin J = 12 (N1 + N2 ) − 1, and so on, until we reach the
plateau for the spin value Jmin = 21 |N1 − N2 |. This establishes the decomposition (4.54).
Corollary. A statement equivalent to the decomposition (4.54) is
{
C if |N1 − N2 | ≤ N3 ≤ N1 + N2 , N1 + N2 + N3 ∈ 2N ,
HomSU2 (VN3 , VN1 ⊗ VN2 ) ≃ (4.55)
0 else .
So far there was never any need to specify a Hermitian scalar product on VN ; now is a good
place to fill this gap. Let φn = z n and
( )−1
N
⟨φn , φn′ ⟩VN := δn, n′ . (4.56)
n
Problem. Show that with the choice (4.56) of Hermitian scalar product one has
is called a Clebsch-Gordan coefficient (for SU2 ). Here the tensor product VN1 ⊗ VN2 is equipped
with the natural Hermitian structure
⟨v1 ⊗ v2 , ṽ1 ⊗ ṽ2 ⟩VN ⊗VN2 = ⟨v1 , ṽ1 ⟩VN1 ⟨v2 , ṽ2 ⟩VN2 (4.59)
1
|N1 − N2 | ≤ N3 ≤ N1 + N2 , N1 + N2 + N3 ∈ 2N . (4.60)
We then use the isomorphism (4.53) along with VN3 ≃ VN∗3 to look for an invariant tensor
102
by a general ansatz of the form
∑
N1 ∑
N2 ∑
N3
T = Tn1 , n2 , n3 z1n1 z2n2 z3n3 . (4.62)
n1 =0 n2 =0 n3 =0
As usual, invariance means g · T = T for all g ∈ SU2 . By linearizing this equation at g = Id and
passing to the complexified Lie algebra sl2 ≃ C3 , we obtain three linearly independent conditions:
∑3 ( ) ∑3
∂ ∂T
σ3 · T = 2zj − Nj T = 0 , σ− · T = =0,
j=1
∂z j j=1
∂z j
∑3 ( )
∂
σ+ · T = −zj2 + Nj zj T = 0 . (4.63)
j=1
∂zj
The first one is easy to implement: it just says that the coefficient Tn1 , n2 , n3 vanishes unless
n1 + n2 + n3 = 12 (N1 + N2 + N3 ) . (4.64)
(N1 − l1 ) Tl1 , l2 +1, l3 +1 + (N2 − l2 ) Tl1 +1, l2 , l3 +1 + (N3 − l3 ) Tl1 +1, l2 +1, l3 = 0 (4.66)
Inspired by Racah’s explicit formula, we can actually express the solution in closed form:
103
Proposition. Let Jl ≡ Nl /2 for l = 1, 2, 3 and
∑ (−1)s1 +s2 +s3 z1J1 −s2 +s3 z2J2 −s3 +s1 z3J3 −s1 +s2
T = (4.67)
s 1 , s2 , s3
(s1 − J1 )! (s2 − J2 )! (s3 − J3 )! (J1 + J2 − s3 )! (J3 + J1 − s2 )! (J2 + J3 − s1 )!
where the sum runs over all triples s1 , s2 , s3 such that the argument of each of the factorials is a
non-negative integer. This tensor is invariant: T = g · T for all g ∈ SU2 .
Proof. It is clear that the polynomial (4.67) is homogeneous of total degree J1 + J2 + J3 and
thus satisfies the constraint (4.64). Let us now show that it also satisfies the condition (4.65) or,
equivalently, the second equation in (4.63). Introducing
AJ (x1 , x2 , x3 ; y1 , y2 , y3 ) :=
(−1)x1 +x2 +x3
(x1 − J1 )! (x2 − J2 )! (x3 − J3 )! (J1 + J2 − y3 )! (J3 + J1 − y2 )! (J2 + J3 − y1 )!
we have
∑
T = AJ (s1 , s2 , s3 ; s1 , s2 , s3 ) z1J1 −s2 +s3 z2J2 −s3 +s1 z3J3 −s1 +s2 ,
s 1 , s2 , s3
×AJ (s1 , s2 , s3 ; s1 , s2 , s3 ) z1J1 −s2 +s3 z2J2 −s3 +s1 z3J3 −s1 +s2 . (4.68)
J1 − s2 + s3 = (J3 + J1 − s2 ) + (s3 − J3 ) ,
Two more identities of the same kind result from cyclic permutations of the index set {1, 2, 3}.
Next, in (4.68) we make a shift of summation index s3 → s3 − 1 for the second term (z2−1 ) and
s2 → s2 + 1 for the third term (z3−1 ), in order to pull out common powers of z1 , z2 , z3 and combine
the various contributions. In this way we recast the expression (4.68) as
∑(
σ− · T = AJ (s1 , s2 , s3 ; s1 , s2 + 1, s3 ) − AJ (s1 , s2 , s3 − 1; s1 , s2 , s3 )
+AJ (s1 , s2 , s3 − 1; s1 , s2 , s3 ) − AJ (s1 − 1, s2 , s3 − 1; s1 , s2 , s3 − 1)
)
+AJ (s1 , s2 + 1, s3 ; s1 + 1, s2 + 1, s3 ) − AJ (s1 , s2 , s3 ; s1 , s2 + 1, s3 )
×z1J1 −s2 +s3 −1 z2J2 −s3 +s1 z3J3 −s1 +s2 .
We see that the second term cancels the third one, and the first term cancels the last one. The
fourth term cancels the fifth one on making a common shift sj → sj + 1 (j = 1, 2, 3) which leaves
the powers of z1 , z2 , z3 unchanged. This already completes the proof that σ− · T = 0 .
104
Problem. By an adaptation of the calculation above, show that σ+ · T = 0 .
Recall that by the choice (4.56) of Hermitian scalar product for VN , the basis vectors φN, n = z n
of VN have squared length
( )−1
N n! (N − n)!
⟨φN, n , φN, n ⟩ = = .
n N!
The standard practice in physics is to work with orthonormal bases. Therefore, we now introduce
the unit vectors
( )1/2 √
2J (2J)! φ2J, J+M
eJ, M := φ2J, J+M = √ ∈ V2J (4.69)
J +M (J + M )! (J − M )!
for M = −J, −J + 1, . . . , J − 1, J.
Definition. Let Tn1 , n2 , n3 ≡ TnN11, ,N2 ,N3
n2 , n3 be the coefficients which(are determined by comparing the
)
J1 J2 J3
ansatz (4.62) with the solution (4.67). The Wigner 3j-symbol is defined as
M1 M2 M3
( )
J1 J2 J3 √
:= TJ21J+M 1 , 2 J2 , 2 J3
1 , J2 +M2 , J3 +M3
△(J1 J2 J3 )
M1 M2 M3
√
× (J1 + M1 )! (J1 − M1 )! (J2 + M2 )! (J2 − M2 )! (J3 + M3 )! (J3 − M3 )! (4.70)
where
(J1 + J2 − J3 )! (J2 + J3 − J1 )! (J3 + J1 − J2 )!
△(J1 J2 J3 ) := (4.71)
(J1 + J2 + J3 + 1)!
is called a triangle coefficient.
Remark. Note that by (4.64) the Wigner 3j-symbol vanishes unless M1 + M2 + M3 = 0 . Note
also that the Wigner 3j-symbol is just a scalar multiple of the coefficients of our tensor T given
in (4.67). The following conclusion is therefore immediate.
Corollary. The tensor
∑ ( )
J1 J2 J3
eJ1 , M1 ⊗ eJ2 , M2 ⊗ eJ3 , M3 (4.72)
M1 M2 M3
M1 , M2 , M3
is SU2 -invariant.
By the isomorphism (4.53) in conjunction with the identification VN∗3 ≃ VN3 due to the invariant
pairing VN ⊗ VN → C (cf. Section 4.3.3), we finally arrive at what is called the Clebsch-Gordan
coefficient for SU2 . It is given by the formula
√ ( )
J1 −J2 +M3 J1 J2 J3
⟨J1 J2 ; M1 M2 | J1 J2 ; J3 M3 ⟩ := (−1) 2J3 + 1 . (4.73)
M1 M2 −M3
The sign change M3 → −M3 reflects the isomorphism VN∗3 ≃ VN3 determined by the mapping
fJ3 , M3 → (−1)J3 −M3 eJ3 ,−M3 (dual basis fJ, M ). The remaining sign factor (−1)J1 −J2 −J3 is due
to what is known as the phase convention of Condon and Shortley. The normalization factor
√
2J3 + 1 is motivated by the properties (4.75) and (4.76) below. A simplified notation is
105
The Dirac-style notation for the Clebsch-Gordan coefficient indicates that it effects a change of
orthonormal basis |J1 J2 ; J3 M3 ⟩ → |J1 M1 ⟩ ⊗ |J2 M2 ⟩.
Problem. Show that the Clebsch-Gordan coefficient obeys the following relations of orthonor-
mality and completeness:
∑
⟨J1 J2 M1 M2 | J3 M3 ⟩ ⟨J1 J2 M1 M2 | J3′ M3′ ⟩ = δJ3 , J3′ δM3 , M3′ , (4.75)
M1 M2
∑
⟨J1 J2 M1 M2 | J3 M3 ⟩ ⟨J1 J2 M1′ M2′ | J3 M3 ⟩ = δM1 , M1′ δM2 , M2′ . (4.76)
J3 M3
thus it describes explicitly how the representation V2J3 occurs in the tensor product V2J1 ⊗ V2J2 .
g · (v ⊗ φ) = ρ(g)v ⊗ φ ◦ ρ(g)−1 .
Now we specialize this relation to v = el and φ = fl′ and then pass to components on each side
using fk (ρ(g)el ) = Dkl (g) and fl′ (ρ(g)−1 ek′ ) = Dl′ k′ (g −1 ). The result is
∫
Dkl (g)Dl′ k′ (g −1 ) dg = c(el , fl′ ) δkk′ .
G
106
Next we use the invariance of Haar measure dg under inversion g 7→ g −1 to exchange the roles of
the index pairs kk ′ and ll′ and infer that c(el , fl′ ) = c0 δll′ .
The constant c0 is determined by setting k = k ′ , l = l′ , and summing over l :
∫ ∫ ∑ ∑
1= dg = Dkl (g)Dlk (g −1 ) dg = c0 δll = c0 dim V,
G G l l
∫
V1 ̸= V2 in different isomorphism classes, then the integral G Dkl (g) Dl′ k′ (g −1 ) dg vanishes for all
(1) (2)
k, l, k ′ , l′ . Indeed, any non-zero value of the integral would imply a G-invariant tensor in V1 ⊗ V2∗ ,
but we know from Section 4.3.1 that no such tensor exists.
From now on we restrict the discussion to the case of G = SU2 , where we use the notation
DM
J
N (g) ≡ fJ, M (ρJ (g) eJ, N ) for the matrix elements (also known in physics as the Wigner D-
functions) of the irreducible representation of spin J. For future reference we record that
∫
J′ −1 δM M ′ δN N ′
DM
J
N (g) DN ′ M ′ (g ) dg = δJJ ′ . (4.79)
SU2 2J + 1
In the remainder of this section, we are going to derive a formula for the Haar integral of a prod-
uct of three Wigner D-functions. To that end, we consider the Haar average of all SU2 -translates
of some tensor X ∈ V2J1 ⊗ V2J2 ⊗ V2J3 in the tensor product of three irreducible representations
∫
with spins J1 , J2 , J3 . Since the Haar-averaged tensor (g · X) dg is SU2 -invariant and there exists
only one such tensor, namely that of Corollary (4.72), we have
∫ ∑ ( J1 J2 J3 )
(g · X) dg = c3 (X) eJ1 , M1 ⊗ eJ2 , M2 ⊗ eJ3 , M3 ,
SU2 M1 M2 M3
M 1 , M2 , M3
with some unknown number c3 (X) . Now let X = eJ1 , N1 ⊗ eJ2 , N2 ⊗ eJ3 , N3 . By passing to compo-
nents with respect to the chosen basis we obtain
∫ ( )
J1 J2 J3
DM1 N1 (g) DM2 N2 (g) DM3 N3 (g) dg = c3 (eJ1 , N1 ⊗ eJ2 , N2 ⊗ eJ3 , N3 )
J1 J2 J3
. (4.80)
SU2 M1 M2 M3
Our next step is to exchange the roles of vectors and co-vectors. A quick (if dirty) way of doing
this is to take the complex conjugate of both sides of the equation and exploit the unitarity of the
representation:
−1
DM
J
N (g) = DN M (g ) .
Haar measure dg does not change under the substitution of integration variable g → g −1 (this is
an immediate consequence of Haar measure being unique). Since the Wigner 3j-symbol is real it
follows that
∫ ( )
J1 J2 J3
DM1 N1 (g) DM2 N2 (g) DM3 N3 (g) dg = c3 (eJ1 , M1 ⊗ eJ2 , M2 ⊗ eJ3 , M3 )
J1 J2 J3
.
SU2 N1 N2 N3
107
Fact. For the Haar integral of a product of three Wigner D-functions one has the formula
∫ ( )( )
J1 J2 J3 J1 J2 J3
DM1 N1 (g) DM2 N2 (g) DM3 N3 (g) dg =
J1 J2 J3
. (4.81)
SU2 M1 M2 M3 N1 N2 N3
from (4.73) and the orthonormality property (4.76) for the Clebsch-Gordan coefficient. To do the
integral on the left-hand side, we use the relation
J∑
1 +J2 ∑
J
χJ1 (g)χJ2 (g) = χJ (g) , χJ (g) := DM
J
M (g) , (4.82)
J=|J1 −J2 | M =−J
Thus our integral after summation over magnetic quantum numbers becomes
∫ ∑
J3 ∫ ∫
χJ1 (g)χJ2 (g)χJ3 (g) dg = χJ (g)χJ3 (g) dg = χJ3 (g)2 dg = 1 ,
SU2 J=|J1 −J2 | SU2 SU2
Proof. Using completeness (4.76) of the Clebsch-Gordan coefficients we invert the relation (4.77):
∑
eJ1 , N1 ⊗ eJ2 , N2 = ⟨J1 J2 N1 N2 | JN ⟩ ϕJ1 J2 (eJ, N ) .
JN
On the right-hand side we insert the expression (4.77) for ϕJ1 J2 (eJ, M ) and pass to components:
∑
DM
J1
1 N1
(g) DM
J2
2 N2
(g) = ⟨J1 J2 N1 N2 | JN ⟩ ⟨J1 J2 M1 M2 | JM ⟩ DM
J
N (g) .
JM N
108
4.6 Tensor operators, Wigner-Eckart Theorem
Recall the action of g ∈ SU2 = Spin3 on the basis vectors eJ, M ∈ V2J of the representation for
spin J:
∑
g · eJ, N = eJ, M DM
J
N (g) .
M
The following definition is motivated by the quantum mechanical fact that if quantum states
transform as ψ 7→ g ψ then quantum operators transform by conjugation Op 7→ g Op g −1 .
Definition. By an irreducible tensor operator of rank J one means a set of operators {TJM }M =−J,...,J
transforming under rotations g ∈ SU2 = Spin3 as
∑
g TJN g −1 = TJM DM
J
N (g) . (4.85)
M
Example. The operation of multiplying (the angular part of the wave function) by a spherical
harmonic YLM is an irreducible tensor operator of rank J = L. Important examples of tensor
operators are furnished by the problem of expanding, say, the charge density operator ρ̂ by mul-
tipoles with respect to a distinguished point (typically the center of mass). The three (spherical)
components of the dipole part of ρ̂ form a tensor operator of rank J = 1, the five components of
the quadrupole part of ρ̂ form a tensor operator of rank J = 2, and so on.
If Jx , Jy , Jz are the operators of total angular momentum, and J± = 12 (Jx ± iJy ), then the
infinitesimal version of (4.85) reads
√
[Jz , TJM ] = ~M TJM , [J± , TJM ] = ~ (J ∓ M )(J ± M + 1) TJ, M ±1 . (4.86)
Wigner-Eckart theorem. Let TJM be (the components of) an irreducible tensor operator. Its
matrix elements between two quantum mechanical states of definite spin and spin projection are
products
⟨J2 M2 | TJM | J1 M1 ⟩ = ⟨J1 JM1 M | J2 M2 ⟩ ⟨J1 ∥ TJ ∥ J2 ⟩ (4.87)
N1 N2 N
109
This result holds true for all g ∈ SU2 and remains true if we average both sides with SU2 -Haar
measure. Hence, by using formula (4.84) we have
∑ ⟨J1 JN1 N | J2 N2 ⟩
⟨J2 M2 | TJM | J1 M1 ⟩ = ⟨J1 JM1 M | J2 M2 ⟩ ⟨J2 N2 | TJN | J1 N1 ⟩ .
N1 N2 N
2J2 + 1
This gives the formula (4.87) of the Wigner-Eckart theorem along with an expression for the
reduced matrix element:
∑
⟨J1 ∥ TJ ∥ J2 ⟩ = (2J2 + 1)−1 ⟨J2 N2 | TJN | J1 N1 ⟩ ⟨J1 JN1 N | J2 N2 ⟩ . (4.88)
N1 N2 N
qe qm ∈ 2π~ Z . (5.1)
This condition, known as the Dirac quantization condition, can be read in two directions. Given
a smallest magnetic charge µ , it quantizes the electric charge according to qe ∈ (2π~/µ) Z .
Conversely, given an electric charge quantum e , magnetic charge is quantized by qm ∈ (2π~/e) Z .
The goal of the present chapter is to offer some physical heuristics and mathematical back-
ground for the Dirac quantization condition (5.1). Please be advised that this story, albeit a corner
stone of theoretical physics, is speculative in the context of physics as an experimental science, as
magnetic monopoles have never been observed in the laboratory to this day. [However, quoting
string theorist J. Polchinski (arXiv:hep-th/0304042): “the existence of magnetic monopoles seems
like one of the safest bets that one can make about physics not yet seen”.]
To begin, here is how the equations of electrodynamics (in traditional vector notation) would
have to look for a non-zero magnetic charge density ρm and magnetic current density jm :
110
The Lorentz force on a particle of velocity v, electric charge qe , and magnetic charge qm , is
K = qe (E + v × B) + qm (H − v × D) . (5.4)
As a particular consequence of the system of equations (5.2) and (5.3), magnetic charge obeys a
continuity equation ρ̇m +div jm = 0 and hence is conserved. There is no difference between electric
and magnetic charges in that respect. However, there does exist a difference of transformation
behavior: in order for electrodynamics to be parity-invariant, ρm must transform as a pseudo-
scalar and jm as an axial vector, whereas ρe is a scalar and je is a (polar) vector. In other
words, while electric charges come with a sign (they are positive or negative), magnetic charges
carry a handedness: a magnetic monopole is neither positive nor negative, but is right-handed or
left-handed.
A quick motivation for the Dirac quantization condition (5.1) is the following. Electric and
magnetic charges are sources of electric and magnetic flux, respectively. Therefore, in the simul-
taneous presence of electric and magnetic monopoles there exist crossed electric and magnetic
fields (even in the static limit). These carry momentum (as given by the Poynting vector) as
well as angular momentum. For two monopole charges with values qe and qm the total angular
momentum is proportional to the product qe qm . The condition (5.1) then reflects the fact that
angular momentum in quantum mechanics takes quantized values only.
It should be stressed that this line of reasoning was not the route by which Dirac arrived at
(5.1); his argument will be sketched in the next section.
111
for wave functions’. It goes like this.
Fixing some reference point o ∈ E3 (the ‘origin’ of three-dimensional Euclidean space) one
works in a representation in which the wave function at any point x ∈ E3 depends on not just x
but also on a path γx connecting o with x. Thus one replaces the usual wave function x 7→ ψ(x)
by a functional on paths, γx → Ψ[γx ]. This generalization comes with a condition (in order to
remove a redundancy which would otherwise ensue): if γx and γx′ are any two paths connecting o
with the same end point x, then the values of the wave function for these paths are required to
be related by a known phase factor:
If the magnetic field is bounded, the functional γx 7→ Ψ[γx ] is continuous. In other words,
small changes of the point x and/or the path γx map to small changes in the wave function. The
same is true for the phase
( / )
arg Ψ[γx ] = (2i)−1 log Ψ[γx ] Ψ[γx ] ,
112
With these preparations made, we can state the central point of the discussion: although the
setting of quantum mechanics as we know it makes for nodal lines without any beginning or end,
there exists no fundamental principle forcing that property, and Dirac suggested the possibility
for nodal lines to have points of birth or termination. He called such points ‘nodal singularities’.
Dirac’s bold step of admitting nodal singularities does not come for free but leads to a con-
straint, as we shall now explain. When positing the relation
between phases, we must take into account the contributions from the phase singularities due to
nodal lines. Thus we need to augment the (dimensionless) magnetic flux through the surface S
by the winding numbers ni of the nodal lines γi that intersect S :
∫ ∑
e
Φ(S) = B · d2 n + ±ni . (5.7)
h S
Here the winding numbers ni (which are positive by convention) count with a sign factor which
is determined by comparing the orientation of the surface S with the sense of circulation of the
nodal line γi . If these agree, ni counts with the plus sign; otherwise the minus sign applies.
The said constraint now arises from the indeterminacy of S: there exist many oriented surfaces
S with boundary ∂S = γx′ −γx and one has no way a priori of knowing which to choose. Therefore,
in order for Dirac’s extended framework (with nodal singularities) to make sense, one has to require
that Φ(S) = Φ(S ′ ) for any pair of oriented surfaces S, S ′ with the same boundary ∂S = ∂S ′ . This
is equivalent to demanding that Φ(S) in (5.7) vanishes for every closed surface S = ∂U . Thus, for
every S = ∂U we require that the dimensionless magnetic flux through ∂U is exactly canceled by
the winding number flux due to nodal singularities in the enclosed region U .
By further analyzing the situation, Dirac concluded that his extended theoretical framework
was consistent if and only if all the wave functions shared the same set of nodal singularities. (The
nodal lines at large will of course be different for two different wave functions in general.) Dirac
took this as evidence that the nodal singularities are objects of physical reality. By the vanishing of
Φ(∂U ) in (5.7) they are sources or sinks of magnetic flux, so they are to be interpreted as magnetic
∫
monopoles. Since the integral ∂U B · d2 n computes the magnetic flux through ∂U and thus the
magnetic charge contained in U , the constraint Φ(∂U ) = 0 amounts to the Dirac quantization
condition (5.1). As a historical note: Dirac himself was so confident of his generalized formalism of
quantum mechanics in the presence of magnetic monopoles that he wrote “one would be surprised
if Nature had made no use of it” [P.A.M. Dirac, Quantized singularities of the electromagnetic
field, Proc. Roc. Soc. A 133 (1931) 60].
113
5.2 The treatment of Wu & Yang
Having come to terms with the Dirac quantization condition, we wish to learn how to do actual
calculations when magnetic monopoles are present. While Dirac’s general framework of path-
dependent wave functionals γx 7→ Ψ[γx ] is a good tool for conceptual reasoning, it is probably not
so convenient for doing concrete calculations. For that, we would like to return to the familiar
framework using wave functions x 7→ ψ(x).
However, a difficulty is now in store for us. To discuss it in very concrete terms, let us consider
the problem of a non-relativistic particle of mass m and charge e moving in the magnetic field
µ x dy ∧ dz + y dz ∧ dx + z dx ∧ dy
B= ω, ω= , (5.8)
4π (x2 + y 2 + z 2 )3/2
of a static monopole of magnetic charge µ = n h/e (with n ∈ Z) placed at the origin of a system
of Cartesian coordinates x , y , z . In vector notation one has the expression
µ er
B=
4π r2
(which is not fully accurate as er is a polar vector whereas B has to transform as an axial vector).
To treat this problem, we need a Hamiltonian H. The textbook quantum mechanics of charged
particles would tell us to fix a vector potential A ∈ rot−1 B and write
(p − eA)2
H= . (5.9)
2m
Unfortunately, in the present situation there simply does not exist any vector potential A such
that rot A = B . Indeed, by integrating B over a surface S = ∂U enclosing the magnetic monopole,
we run into a contradiction:
∫ ∫
Gauss
0 ̸= µ = B·d n =
2
div B d3 x = 0 ,
S U
since div B vanishes by div ◦ rot = 0 whenever B = rot A . Thus there isn’t any A, and it would
hence appear that there exists no Hamiltonian (5.9).
To overcome this difficulty, Dirac invented a trick which is nowadays referred to as the ‘Dirac
string’. We do not describe it here but give a more modern treatment due to Wu and Yang (1968).
The idea is to exclude the singular point o ∈ E3 and work with two overlapping domains U ± (and
corresponding coordinate charts) for the punctured space E3 \ {o} = U + ∪ U − . To be concrete,
we take U + (U − ) to be E3 \ {o} with the negative (resp. positive) z-axis removed. On U + ∩ U −
we then consider a pair of one-forms
( )
n~ n~ z x dy − y dx
A(±) = ± (1 ∓ cos θ) dϕ = ± 1∓ √ . (5.10)
2e 2e x2 + y2 + z2 x2 + y 2
Problem. Compute the exterior derivatives dA(±) to show that both A(+) and A(−) are magnetic
vector potentials for the magnetic field strength B = dA(+) = dA(−) in (5.8).
The magnetic vector potential A(+) is singular on the negative z-axis (z ≤ 0) but extends as
a smooth form to the positive z-axis (z > 0); thus A(+) is good on U + . For A(−) it is the other
114
way around (singular for z ≥ 0 but good on U − ). On the overlap U + ∩ U − of their domains of
definition, the two forms are related by a gauge transformation:
n~ n~ x dy − y dx
A(+) − A(−) = dϕ = . (5.11)
e e x2 + y 2
We now recall that by the principle of gauge invariance in quantum mechanics, a gauge transfor-
mation A → A+df of the magnetic vector potential implies a corresponding gauge transformation
ψ → e i(e/~) f ψ of the wave function. Accordingly, in the present case of (5.11) we obtain
( )n/2
(+) inϕ (−) x + iy
ψ =e ψ = ψ (−) , (5.12)
x − iy
which ensures that (d − ieA(+) /~)ψ (+) = einϕ (d − ieA(−) /~)ψ (−) on U + ∩ U − . Mathematically
speaking, the relation (5.12) means that the wave functions ψ (±) are local expressions of a globally
defined object, ψ , which is called a section of (some) complex line bundle over E3 \ {o}. Similarly,
A(±) are the local coordinate expressions for a connection, A , on the same line bundle.
In order to compute the quantum mechanics of our charged particle in the magnetic monopole
field (5.8), the prescription of Wu and Yang is to solve the Schrödinger equation on U + for ψ (+)
using A(+) , and on U − for ψ (−) using A(−) . In executing that procedure the relations (5.11) and
(5.12) are in force. Some of the computational details will be presented for a somewhat simplified
setting in the next subsection. Here we wish to conclude with the remark that the above can be
regarded as another derivation of the Dirac quantization condition; indeed, the number n is the
magnetic monopole charge (in units of the flux quantum h/e), and the relation (5.12) makes sense
(as a well-defined transition between coordinate charts) if and only if n is an integer.
We are now going to show that the wave functions for the special case of µ = 2h/e (with the
motion constrained to M ≃ S2 ) can be interpreted as vector fields on M . You might object that
as a student of quantum theory you were taught to think of the wave function of a charged particle
as a function taking values in the complex numbers, C . If so, the following observation should
115
be helpful: at every point x ∈ M of a sphere, the vectors v ∈ Tx M tangent to the sphere span a
two-dimensional vector space isomorphic to R2 , which in turn is isomorphic to C .
The argument linking the case of µ = 2h/e to vector fields goes as follows. Using stereographic
projection we introduce two sets of complex coordinate functions ζ (±) : S2 → C by
The coordinate function ζ (+) = ξ (resp. ζ (−) = η) is defined everywhere on S2 with the exception
of the south pole θ = π (resp. north pole θ = 0). Note the relation
1
ξ=− . (5.14)
η
One easily checks that our magnetic vector potentials A(±) (restricted to S2 ) have the expressions
~ ¯ − ξdξ¯
~ ξdξ ~ ~ η̄ dη − η dη̄
A(+) = (1 − cos θ) dϕ = , A(−) = − (1 + cos θ) dϕ = .
e ie 1 + |ξ|2 e ie 1 + |η|2
The metric tensor pulls back to
4 dξ¯ dξ 4 dη̄ dη
dθ2 + sin2 θ dϕ2 = = ,
(1 + |ξ|2 )2 (1 + |η|2 )2
which motivates us to introduce the following unit (or normalized) basis vector fields:
1 + |ξ|2 ∂ 1 + |η|2 ∂
eξ := , eξ̄ := eξ , eη := , eη̄ := eη . (5.15)
2 ∂ξ 2 ∂η
A short computation using (5.14) then gives
eξ = e−2iϕ eη . (5.16)
If we now interpret the complex-valued wave functions ψ (±) as the components (with respect
to the corresponding basis vector fields) of an object v defined invariantly by
then the relation (5.16) translates to the transition rule (5.12) for ψ (±) (in the case of n = 2).
Thus we learn that the wave functions ψ (±) of the Dirac monopole problem for µ = 2 h/e are in
fact local expressions for (tangent) vector fields v .
Definition. By the tangent bundle T M one means the space of tangent vectors of the manifold
M – in our case a two-sphere M ≃ S2 . Locally, i.e., in a small enough neighborhood U ⊂ M
of any base point x ∈ M , the tangent bundle T M is a direct product T M |U ≃ U × R2 . The
tangent space Tx M is also called the fiber of the tangent bundle at x. It is sometimes denoted by
Tx M ≡ π −1 (x) to indicate that Tx M can be viewed as the inverse image of the projection map
π : T M → M defined by π(Tx M ) = x for all x ∈ M . A section v ∈ Γ(M, T M ) of the tangent
bundle is a mapping which assigns to every point x ∈ M a vector v(x) ∈ Tx M . (In other words,
π ◦ v = Id.) A section v ∈ Γ(M, T M ) is also called a vector field on M .
116
Remark. The tangent bundle T S2 of the two-sphere does not factor globally as a product
S2 × R2 . (In contrast, one does have T S3 ≃ S3 × R3 .) This fact is expressed by saying that T S2
is a non-trivial vector bundle. Non-triviality is related to the fact that every smooth vector field
v ∈ Γ(M, T M ) for M = S2 has at least two zeroes. (In Dirac’s way of thinking, these correspond
to the two nodal lines emanating from a nodal singularity of charge n = 2.)
Accepting the change of mathematical description from complex scalar-valued wave functions ψ
to real tangent vector fields v, our next question is this: what mathematical object should take
√
the role of i = −1 ∈ C on the left-hand side i~ ∂ψ/∂t of the Schrödinger equation?
Problem. For the complex stereographic coordinate ζ (+) = ξ = tan(θ/2) eiϕ show that
( ) ( )
∂ 1 + |ξ|2 ∂ ∂ 1 ∂ 1 + |ξ|2
∂ ∂
= ξ + ξ¯ ¯ , = iξ − i ξ¯ ¯ . (5.18)
∂θ 2|ξ| ∂ξ ∂ξ sin θ ∂ϕ 2|ξ| ∂ξ ∂ξ
How do these relations look for ζ (−) = η = − cot(θ/2) e−iϕ ?
To answer our question, we introduce the basis vector fields
∂ 1 ∂
eθ := , eϕ := , (5.19)
∂θ sin θ ∂ϕ
which are related to those of (5.15) by
ξ ξ¯ ξ ξ¯
eθ = eξ + eξ̄ , eϕ = i eξ − i e . (5.20)
|ξ| |ξ| |ξ| |ξ| ξ̄
(We here focus on the upper hemisphere with basis eξ , eξ̄ and wave function ψ (+) ≡ ψ . In the
lower hemisphere the situation is no different.) We now change basis,
v = ψ eξ + ψ̄ eξ̄ = vθ eθ + vϕ eϕ . (5.21)
√
Using (5.20) we then see that multiplication (ψ, ψ̄) 7→ (iψ, −iψ̄) by i = −1 in the complex wave
function picture translates to a π/2 rotation
117
5.3.3 Covariant derivative
In this subsection we are going to describe a rule of differentiating vector fields. For brevity,
we shall focus on the situation of interest, namely on vector fields on the two-sphere M = S2
with SO3 -invariant geometry. We refer to differential geometry textbooks for the more general
procedure of covariantly differentiating the sections of any vector bundle.
As a warm up we recall the process of differentiating functions on a manifold M . Suppose that
we want to calculate the derivative (df )x (u) of a differentiable function f : M → C at some point
x ∈ M in the direction of the tangent vector u ∈ Tx M . To do so, we choose a differentiable curve
γ : (−ϵ, ϵ) → M with γ(0) = x, γ̇(0) = u , and compute the said derivative as
f (γ(t)) − f (γ(0))
d
(df )x (v) = f (γ(t)) := lim .
dt t=0 t→0 t
Now if we naively try to apply the same definition
v(γ(t)) − v(γ(0))
d ??
v(γ(t)) = lim ,
dt t=0 t→0 t
to a vector field v ∈ Γ(M, T M ), we face the problem that the expression v(γ(t)) − v(γ(0)) makes
no immediate sense. Indeed, we have v(γ(t)) ∈ Tγ(t) M and there is no meaning to the difference
(nor the sum) of two vectors in different vector spaces Tγ(t) M ̸= Tγ(0) M (for t ̸= 0). The point here
is that although all tangent spaces Tx M ≃ R2 are in principle the same, there exists no canonical
identification of Tx M with Tx′ M for x ̸= x′ .
In order to give a meaningful definition of the difference between v(γ(t)) and v(γ(0)), we first
have to fix some vector space isomorphism Tγ(t) M ≃ Tγ(0) M . Such an isomorphism is determined
by what is called a connection (on a vector bundle). In the case at hand (i.e., magnetic monopole
charge n = 2, tangent bundle T M ), the good choice of isomorphism turns out to be given by
parallel transport via the so-called Levi-Civita connection, as follows.
Let w ∈ Tx M = Tγ(0) M . We wish to introduce a natural scheme of parallel transporting
w along the curve γ from Tγ(0) M to Tγ(t) M . For this we need the following information. Let
M = S2 ⊂ R3 be equipped with the geometry induced by restriction of the standard Euclidean
structure of R3 . (Mathematically speaking, this is the geometry given by the Fubini-Study metric
of S2 .) Thus for each point p ∈ M we have a Euclidean scalar product gp : Tp M × Tp M → R
enabling us to measure the lengths of tangent vectors in Tp M and the angles between them.
The Fubini-Study metric is SO3 -invariant. In the present context this reflects the fact that the
magnetic field of a static monopole is invariant under SO3 rotations fixing the monopole.
The idea now is to translate w ∈ Tx M along the curve γ(t) in such a way that the length of w
and its angle with the tangent vector γ̇(t) of the curve γ stays constant. In formulas, one defines
the one-parameter family of parallel translates w(t) ∈ Tγ(t) M by the conditions w(t = 0) = w and
118
where x = γ(0), u = γ̇(0) as before. This process of parallel translation applies to any w ∈ Tx M .
Doing it for all w ∈ Tx M at once, we have a one-parameter family of isomorphisms
Definition. The Levi-Civitá connection ∇ is a rule which assigns to any vector field Y and a
differentiable vector field v a third vector field ∇Y v , called the covariant derivative of v in the
direction of Y . For any fixed point x ∈ M , the vector (∇Y v)(x) ∈ Tx M is defined by choosing a
differentiable curve γ : (−ϵ, ϵ) → M with γ(0) = x and γ̇(0) = Y (x) and taking the limit
∇(f v) = df ⊗ v + f ∇v (5.27)
119
Now we recall that eϕ = Jeθ and −eθ = Jeϕ . By switching from the vector field v to the complex
wave function ψ := vθ + ivϕ and using the correspondence J ↔ i , we then obtain the result
Notice that the wave function ψ in this equation differs from that of (5.21) by a gauge transfor-
mation ψ = e−iϕ ψ (+) , which corresponds to a change of gauge
~ ~
A = A(+) + dϕ = − cos θ dϕ (5.32)
e e
for the magnetic vector potential A . In the present choice of gauge, A has singularities at θ = 0
and θ = π (reflecting the coordinate singularities due to the choice of basis eθ , eϕ ).
We see from (5.31) that the connection ∇ is proportional to the mechanical momentum p−eA .
More precisely, ∇ = d − ieA/~ . Therefore, in view of the expression H = (p − eA)2 /2m for the
Hamiltonian, the following statement should now be plausible.
Fact. In the vector field picture the Schrödinger equation reads
~2
~ J v̇ = ∇† ∇v . (5.33)
2mR2
Here R is the radius of the two-sphere and the symbol ∇† denotes the Hermitian adjoint of ∇.
Remark. ∇† is a first-order differential operator taking sections of T ∗ M ⊗ T M into sections of
T M . (Each of the vector bundles T M and T ∗ M ⊗ T M carries a natural Hermitian structure, by
which one defines ∇† from ∇ in the usual way; see Section 1.7.1.) The second-order differential
operator ∇† ∇ on vector fields is called the Bochner Laplacian.
Problem. For v = eθ or v = eϕ verify the expression
( )
†
( ) ∂f 1 ∂g
∇ (f dθ + g sin θ dϕ) ⊗ v = − + + cot θ (f + gJ) v .
∂θ sin θ ∂ϕ
By using this expression along with (5.30), show that each of the three-parameter family (a ∈ R3 )
of vector fields
v0a = (ax cos θ cos ϕ + ay cos θ sin ϕ − az sin θ) eθ + (−ax sin ϕ + ay cos ϕ) eϕ (5.34)
~2 e|B| 2h/e
E0 = = 12 ~ω , ω= , |B| = ,
2mR2 m 4πR2
where ω is the cyclotron frequency for the magnetic field strength |B|.
120
5.4 Other Cases (µ ̸= 2h/e)
Having discussed the case of monopole charge µ = 2h/e in detail, we now wish to briefly address
the more general situation. For that purpose, we introduce another piece of information.
Problem. Let Ω = sin θ dθ ∧ dϕ be the solid-angle (or Fubini-Study) two-form on S2 . Using the
formulas for the Levi-Civitá connection ∇ given in the previous section, show that
where [X, Y ] = XY − Y X is the commutator (or Lie bracket) of the two vector fields X, Y viewed
as first-order differential operators X : f 7→ (df )(X) on functions.
Remark. Notice that there is no derivative acting on the vector field v on the right-hand side of
the formula (5.35). Thus, although
looks very much like a differential operator, it is actually just a tensor field; more precisely, a
section of the vector bundle End(T M ). One calls F ∇ ≡ R the Riemann curvature tensor of the
tangent bundle T M with Levi-Civitá connection ∇. More generally, for any vector bundle with a
connection ∇, one defines the curvature tensor by the same formula (5.36).
In all of the discussion up to now, we have been assuming the situation of a magnetic monopole
in the static limit, where the magnetic field is centro-symmetric. If our charge µ = 2h/e monopole
is in motion or if divergenceless magnetic fields are superimposed on the magnetic monopole field,
the description of the wave function as a vector field continues to apply. The only change is that
the Levi-Civitá connection ∇ is replaced by an adapted connection (still denoted by ∇) such that
the curvature tensor of the adapted ∇ reflects the variable magnetic field:
e
F ∇ (X, Y ) = − B(X, Y ) J . (5.37)
~
∫
Here B is the two-form of the magnetic field strength with total flux S2
B = 2h/e.
On the other hand, if the quantized value µ = 2h/e of the magnetic monopole charge is replaced
by a different quantized value µ = nh/e, then the vector bundle itself (not just the connection)
undergoes a change. We saw a glimpse of the vector bundles for general n ∈ Z in Section 5.2, via
the procedure of gluing together two hemispheres by means of the transition function (5.12). We
now give a global differential-geometric description of these vector bundles.
To prepare the generalization, we introduce another way of thinking about the tangent bundle
of M = S2 . For this we fix some point of S2 , say the north pole, o . The rotation group SO3 acts
transitively on S2 , which is to say that if x ∈ S2 is any other point, we can find some rotation
R ∈ SO3 such that R · o = x . Of course R is unique only up right multiplication by elements of
the SO2 subgroup fixing the north pole. Indeed, if h · o = o and R · o = x then (Rh) · o = x. Thus
M = S2 is identified as a quotient (or coset space) S2 ≃ SO3 /SO2 .
121
Next, we are going to use this setting to develop a different description of tangent vectors and
vector fields. At first, we fix some pair x ∈ S2 and R ∈ SO3 such that R · o = x . The differential
R′ : To M → Tx M then is a bijective linear mapping from the tangent plane of the north pole
to the tangent plane of x . By using its inverse, we can map any tangent vector v ∈ Tx M to a
tangent vector R′ −1 v ∈ To M . Thus we may describe any pair (x, v) with v ∈ Tx M by some pair
(R ; u) with u = R′ −1 v ∈ To M . Because R is not uniquely defined, this description is not unique.
Indeed, if h is an SO2 rotation fixing o , then the pair (Rh−1 ; h′ u) is as good a description of (x, v)
as is (R ; u). In order to eliminate this redundancy, we consider
(Rh−1 ; h′ u) ∼ (R ; u) (5.38)
as equivalent (∼) and denote the equivalence class of all such pairs by [R ; u]. Note that these
equivalence classes can be added and multiplied by scalars:
What we have described here is a special case of the following general setting.
Definition 1. Let M be a manifold and G be a group. A G-principal bundle P over M is a fiber
bundle π : P → M with typical fiber π −1 (x) ≃ G. In any local chart U ⊂ M the bundle P factors
as a direct product P |U ≃ U × G. The group G acts on P on the right, i.e., p · (g1 g2 ) = (p · g1 ) · g2 .
Definition 2. Let V be a vector space carrying a G-representation ρ : G → GL(V ). If P is a
G-principal bundle over M , the associated vector bundle P ×G V over M ≃ P/G is formed by
dividing the direct product P × V by the equivalence relation
Thus the vectors of the vector bundle P ×G V are equivalence classes [p ; v] = [p · g −1 ; ρ(g)v].
From this general scheme, we get the tangent bundle of M = S2 by setting P = SO3 , G = SO2 ,
V = R2 , and taking for ρ the fundamental representation of SO2 on R2 . Any other vector bundle
over S2 is obtained simply by changing the SO2 -representation.
Fact. Wave functions of the M = S2 Dirac monopole problem with magnetic charge µ = nh/e are
sections of the vector bundle SO3 ×SO2 R2 which is associated to SO3 by the SO2 -representation
( ) ( )
cos ϕ − sin ϕ cos(nϕ/2) − sin(nϕ/2)
ρn = . (5.41)
sin ϕ cos ϕ sin(nϕ/2) cos(nϕ/2)
122
Remark. This makes immediate sense for n ∈ 2Z . For the case of odd n ∈ 2Z + 1 it makes
sense by the observation that S2 = SO3 /SO2 = Spin3 /Spin2 = SU2 /U1 . Taking the square root
(ϕ → ϕ/2) is well-defined in Spin2 , which is a 2 : 1 covering of SO2 .
Although Dirac did not have the proper mathematical apparatus at his disposal, he recognized
the possibility for quantum mechanics to remain consistent in the presence of (quantized) magnetic
charges, and he established computational control of that challenging situation. In the present
chapter we have met several ways of arriving at the quantization condition named after him. Its
final interpretation is that of an integrality condition which lets the complex line bundle and hence
the wave function exist as globally defined objects free of singularities.
123