. 10
( 78 .)



Equations of Motion. The principle of least action

d4 xL
δS = 0, S=

yields when varying the gauge ¬elds
δSY M = ’ d4 x tr {Fµν δF µν } = ’ d4 x tr Fµν [Dµ , δAν ]

d4 x tr {δAν [Dµ , Fµν ]}
40 F. Lenz

the inhomogeneous ¬eld equations

[Dµ , F µν ] = j ν , (84)

with j ν the color current associated with the matter ¬elds
j aν = . (85)

For QCD and the Georgi“Glashow model, these currents are given respectively
¯ ν „ ψ , j aν = gf abc φb (Dν φ)c .

j = g ψγ (86)
As in electrodynamics, the homogeneous ¬eld equations for the Yang“Mills ¬eld
Dµ , F µν = 0 ,

with the dual ¬eld strength tensor
F µν = µνσρ
Fσρ , (87)
are obtained as the Jacobi identities of the covariant derivative

[Dµ , [Dν , Dρ ]] + [Dν , [Dρ , Dµ ]] + [Dρ , [Dν , Dµ ]] = 0 ,

i.e. they follow from the mere fact that the ¬eld strength is represented in terms
of gauge potentials.

Gauge Transformations. Gauge transformations change the color orientation
of the matter ¬elds locally, i.e. in a space-time dependent manner, and are de¬ned
U (x) = exp {ig± (x)} = exp ig± (x) a
, (88)
with the arbitrary gauge function ±a (x). Matter ¬elds transform covariantly
with U
ψ ’ U ψ , φ ’ U φU † . (89)
The transformation property of A is chosen such that the covariant derivatives
of the matter ¬elds Dµ ψ and Dµ φ transform as the matter ¬elds ψ and φ
respectively. As in electrodynamics, this requirement makes the gauge ¬elds
transform inhomogeneously

‚µ U † (x) = AµU ] (x)
Aµ (x) ’ U (x) Aµ (x) + [
resulting in a covariant transformation law for the ¬eld strength

Fµν ’ U Fµν U † . (91)
Topological Concepts in Gauge Theories 41

Under in¬nitesimal gauge transformations (|g±a (x) | 1)

Aa (x) ’ Aa (x) ’ ‚µ ±a (x) ’ gf abc ±b (x) Ac (x) . (92)
µ µ µ

As in electrodynamics, gauge ¬elds which are gauge transforms of Aµ = 0 are
called pure gauges (cf. (8)) and are, according to (90), given by

‚µ U † (x) .
Apg (x) = U (x) (93)

Physical observables must be independent of the choice of gauge (coordinate
system in color space). Local quantities such as the Yang“Mills action density
tr F µν (x)Fµν (x) or matter ¬eld bilinears like ψ(x)ψ(x), φa (x)φa (x) are gauge
invariant, i.e. their value does not change under local gauge transformations.
One also introduces non-local quantities which, in generalization of the trans-
formation law (91) for the ¬eld strength, change homogeneously under gauge
transformations. In this construction a basic building block is the path ordered
„¦ (x, y, C) = P exp ’ig = P exp ’ig dxµ Aµ .
dσ Aµ x(σ)
dσ C

It describes a gauge string between the space-time points x = x(s0 ) and y = x(s).
„¦ satis¬es the di¬erential equation

= ’ig Aµ „¦. (95)
ds ds
Gauge transforming this di¬erential equation yields the transformation property
of „¦
„¦ (x, y, C) ’ U (x) „¦ (x, y, C) U † (y) . (96)
With the help of „¦, non-local, gauge invariant quantities like

tr„¦ † (x.y, C)F µν (x)„¦ (x, y, C) Fµν (y) , ¯
ψ(x)„¦ (x, y, C) ψ(y),

or closed gauge strings “ SU(N) Wilson loops

tr „¦ (x, x, C)
WC = (97)
can be constructed. For pure gauges (93), the di¬erential equation (95) is solved
„¦ pg (x, y, C) = U (x) U † (y). (98)
While ψ(x)„¦ (x, y, C) ψ(y) is an operator which connects the vacuum with meson
states for SU (2) and SU (3), fermionic baryons appear only in SU (3) in which
gauge invariant states containing an odd number of fermions can be constructed.
42 F. Lenz

In SU(3) a point-like gauge invariant baryonic state is obtained by creating three
quarks in a color antisymmetric state at the same space-time point

ψ(x) ∼ abc
ψ a (x)ψ b (x)ψ c (x).

Under gauge transformations,

ψ(x) ’ abc
Ua± (x)ψ ± (x)Ubβ (x)ψ β (x)Ucγ (x)ψ γ (x)
= det U (x) abc ψ a (x)ψ b (x)ψ c (x) .

Operators that create ¬nite size baryonic states must contain appropriate gauge
strings as given by the following expression

ψ(x, y, z) ∼ [„¦(u, x, C1 )ψ(x)]a [„¦(u, y, C2 )ψ(y)]b [„¦(u, z, C3 )ψ(z)]c .

The presence of these gauge strings makes ψ gauge invariant as is easily veri¬ed
with the help of the transformation property (96). Thus, gauge invariance is
enforced by color exchange processes taking place between the quarks.

Canonical Formalism. The canonical formalism is developed in the same way
as in electrodynamics. Due to the antisymmetry of Fµν , the Lagrangian (80)
does not contain the time derivative of A0 which, in the canonical formalism,
has to be treated as a constrained variable. In the Weyl gauge [33,34]

a = 1....N 2 ’ 1,
Aa = 0, (99)

these constrained variables are eliminated and the standard procedure of canon-
ical quantization can be employed. In a ¬rst step, the canonical momenta of
gauge and matter ¬elds (quarks and Higgs ¬elds) are identi¬ed
δLY M δLmq δLmH
= iψ ± † ,
a = ’E
= πa .
, ± a
‚0 Ai ‚0 ψ ‚0 φ
By Legendre transformation, one obtains the Hamiltonian density of the gauge
HY M = (E 2 + B 2 ), (100)
and of the matter ¬elds
Hm = ψ † γ γ Di + γ 0 m ψ,
QCD : (101)
12 1
Hm =
π + (Dφ)2 + V (φ) .
Georgi“Glashow model : (102)
2 2
The gauge condition (99) does not ¬x the gauge uniquely, it still allows for time-
independent gauge transformations U (x), i.e. gauge transformations which are
generated by time-independent gauge functions ±(x) (88). As a consequence the
Hamiltonian exhibits a local symmetry

H = U (x) H U (x)† (103)
Topological Concepts in Gauge Theories 43

This residual gauge symmetry is taken into account by requiring physical states
|¦ to satisfy the Gauß law, i.e. the 0-component of the equation of motion (cf.
[Di , E i ] + j 0 |¦ = 0.
In general, the non-abelian Gauß law cannot be implemented in closed form
which severely limits the applicability of the canonical formalism. A complete
canonical formulation has been given in axial gauge [35] as will be discussed
below. The connection of canonical to path-integral quantization is discussed in
detail in [36].

5 ™t Hooft“Polyakov Monopole
The t™ Hooft“Polyakov monopole [37,38] is a topological excitation in the non-
abelian Higgs or Georgi“Glashow model (SU (2) color). We start with a brief
discussion of the properties of this model with emphasis on ground state con¬g-
urations and their topological properties.

5.1 Non-Abelian Higgs Model
The Lagrangian (82) and the equations of motion (84) and (85) of the non-
abelian Higgs model have been discussed in the previous section. For the follow-
ing discussion we specify the self-interaction, which as in the abelian Higgs model
is assumed to be a fourth order polynomial in the ¬elds with the normalization
chosen such that its minimal value is zero
»(φ2 ’ a2 )2 ,
V (φ) = » > 0. (104)
Since φ is a vector in color space and gauge transformations rotate the color
direction of the Higgs ¬eld (89), V is gauge invariant

V (gφ) = V (φ) . (105)

We have used the notation

gφ = U φU † , g ∈ G = SU (2).

The analysis of this model parallels that of the abelian Higgs model. Starting
point is the energy density of static solutions, which in the Weyl gauge is given
by ((100), (102))
1 1
(x) = B2 + (Dφ)2 + V (φ). (106)
2 2
The choice
A = 0, φ = φ0 = const. , V (φ0 ) = 0 (107)
minimizes the energy density. Due to the presence of the local symmetry of
the Hamiltonian (cf. (103)), this choice is not unique. Any ¬eld con¬guration
44 F. Lenz

connected to (107) by a time-independent gauge transformation will also have
vanishing energy density. Gauge ¬xing conditions by which the Gauß law con-
straint is implemented remove these gauge ambiguities; in general a global gauge
symmetry remains (cf. [39,35]). Under a space-time independent gauge transfor-
g = exp ig± , ± = const , (108)
applied to a con¬guration (107), the gauge ¬eld is unchanged as is the modulus
of the Higgs ¬eld. The transformation rotates the spatially constant φ0 . In such
a ground-state con¬guration, the Higgs ¬eld exhibits a spontaneous orientation
analogous to the spontaneous magnetization of a ferromagnet,
|φ0 | = a .
φ = φ0 ,
This appearance of a phase with spontaneous orientation of the Higgs ¬eld is
a consequence of a vacuum degeneracy completely analogous to the vacuum
degeneracy of the abelian Higgs model with its spontaneous orientation of the


. 10
( 78 .)