Detailed Lecture Notes in Quantum Field Theory

Detailed lecture notes in Quantum Field
Theory
Diego Restrepo
Instituto de Fı́sica
Universidad de Antioquia
2016
ii
License
All materials are licensed under the Creative Commons Attribution-Share Alike 3.0 United
States License.
iii
iv
Contents
Introduction 1
1 Classical Field Theory 3

1.1 Lagrangian Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 Ecuaciones de Euler-Lagrange . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.2 Teorema de Noether para simetrı́as internas . . . . . . . . . . . . . . . . . . . 5
1.1.3 Teorema de Noether para simetrı́as externas . . . . . . . . . . . . . . . . . . . 7
1.2 Global gauge invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Local phase invariance in the Scrödinger’s Lagrangian . . . . . . . . . . . . . . . . . . 14
1.4 Notación relativista . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4.1 Ejemplos de cuadrivectores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.4.2 Lorentz tranformation for fields . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.5 Vector field Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.5.1 Energı́a del campo electromagnético . . . . . . . . . . . . . . . . . . . . . . . . 30
1.6 Scrödinger Equation in presence of the electromagnetic field . . . . . . . . . . . . . . 33
1.6.1 Euler-Lagrange equation for ψ ∗ . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.6.2 Euler-Lagrange equation for Aµ . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.6.3 Conserved currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
1.7 Gauge Transformation Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
1.8 Proca Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
1.9 Klein-Gordon Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.9.1 Complex scalars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.10 Lorentz transformation of the fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.11 Dirac’s Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.11.1 Lorentz transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.11.2 Corriente conservada y Lagrangiano de Dirac . . . . . . . . . . . . . . . . . . 51
1.11.3 Tensor momento-energı́a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1.11.4 Ecuaciones de Euler-Lagrange . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
v
vi CONTENTS
1.11.5 Propiedades de las matrices de Dirac . . . . . . . . . . . . . . . . . . . . . . . 57

1.11.6 Lorentz Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
1.11.7 Lorentz invariance of the Dirac Action . . . . . . . . . . . . . . . . . . . . . . 62
1.11.8 Dirac’s Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
1.12 Electrodinámica Cuántica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
1.13 Cromodinámica Cuántica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
1.13.1 Ecuaciones de Euler–Lagrange . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
1.13.2 Derivada covariante adjunta . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
1.14 Spontaneous symmetry breaking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
1.14.1 Superconductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
1.15 Fermiones quirales de cuatro componentes . . . . . . . . . . . . . . . . . . . . . . . . 90
1.16 Standard model Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
1.16.1 Spontaneous symmetry breaking in SU (3)c × SU (2)L × U (1)Y . . . . . . . . . 97
1.16.2 Yukawa Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
1.16.3 Fermion-gauge interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
1.16.4 Self-interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
1.16.5 Lagrangiano del modelo estándar para la primera generación . . . . . . . . . . 108
1.16.6 Dinámica de sabor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
1.17 Fenomenologı́a Electrodébil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
1.17.1 Decaimientos débiles mediados por corrientes cargadas . . . . . . . . . . . . . 117
2 Second quantization 121

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
2.2 Quantization of the nonrelativistic string . . . . . . . . . . . . . . . . . . . . . . . . . 122
2.2.1 The clasical string . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
2.2.2 Quantization of the string . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
2.2.3 Generalization to three dimensions . . . . . . . . . . . . . . . . . . . . . . . . 136
2.3 Quantization of the Klein-Gordon field . . . . . . . . . . . . . . . . . . . . . . . . . . 137
2.4 Fock space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
2.5 Fock space for the harmonic oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
2.6 Propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
2.6.1 Complex scalar field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
2.7 Quantization of Fermions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
2.7.1 Quantization of Weyl spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3 Quantization of the electromagnetic field 163

3.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
3.2 Quantization of the electromagnetic field . . . . . . . . . . . . . . . . . . . . . . . . . 165
CONTENTS vii
4 S–matrix 169
4.1 The S–matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.2 Relativistic and no relativistic normalizations . . . . . . . . . . . . . . . . . . . . . . 171
4.3 Process probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.4 Cross Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.4.1 2–to–2 cross section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.5 Decay Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
4.5.1 Two body decays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
4.5.2 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
4.6 Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
5 Feynman Rules 195

5.1 Interaction picture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
5.2 Useful identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
5.3 Dirac delta function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
5.4 Yukawa interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
5.5 Wick Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
5.5.1 Basics concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
5.5.2 Wick theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
5.6 Mandelstam variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
5.7 Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.7.1 Higgs mediated scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.7.2 Other two-particle states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
5.7.3 Vector interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
5.7.4 Example: Direct detection cross sections . . . . . . . . . . . . . . . . . . . . . 250
5.8 Dirac fermion case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
5.9 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
6 Neutrino masses 257

6.1 Standard model particle content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
6.2 Weinberg operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
6.3 Type-I seesaw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
6.4 Dirac operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
6.5 Dirac Type-I seesaw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
7 Leptogenesis 265
viii CONTENTS
8 One-loop neutrino masses 267

8.1 Extra particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
8.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
8.3 Radiative seesaw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
8.4 Singlet-doublet fermions with scalar singlets . . . . . . . . . . . . . . . . . . . . . . . 274
8.4.1 Zee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
8.4.2 Minimal Zee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
9 Incubator 279
A Computational QFT 281

A.1 LanHEP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
A.2 CalcHEP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
A.3 LanHEP/CalcHEP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
B Neutrinos 293
B.1 Weinberg operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
C Two body decays 295

C.1 Particle decays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
C.2 Width decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
C.3 e+ e− → µ+ µ− . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
D Three body decays 301

D.1 Muon decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
D.1.1 Amplitude estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
D.1.2 Amplitude calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
D.2 three body decays in radiative seesaw . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
D.A Sample point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
D.B Preliminary discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
D.3 Diagonalization of a symmetric 2 × 2 matrix . . . . . . . . . . . . . . . . . . . . . . . 323
Introduction
We have organized the topics in order of complexity, and, in the same spirit than in previous book
[1], we have tried to write the calculations as detailed as possible. In Chapter 2 we included the
building blocks of quantum field theory, in Chapter 4 we introduce the S–matrix in the Scrödinger
Picture separating the kinematical and normalization factors from the matrix element. Then the
expressions for the decay rates and cross sections are obtained. The explicit calculation of the
matrix element from the expansion of the S–matrix to obtain the Feynman rules, is postponed to
Chapter 5. In Chapter C we use the Feynman rules necessary to calculates the matrix element, and
develop the techniques associated to the squaring of the matrix element. In Chapter 5 we obtain the
Feynman rules used in two body decays directly from the first order expansion of the S–matrix in
the interaction picture. The subsequent chapters have applications of the techniques developed to
the calculation of tree-level, Chapter D and loop processes.
This notes are based in books [2], [3], [4]. In each Chapter or Section the main reference used
is cited. Also, we have included material developed by students Juan Alberto Yepez, José David
Ruiz Álvarez. This notes are written in English, because at this level it is expected that any physics
student be fluently in reading technical texts in this language.
Some parts are still in Spanish.
This work have been partially supported by “Dedicación Exclusiva 2008-2009” project: RR 26663
1
2 CONTENTS
Chapter 1
Classical Field Theory
This chapter is a summary of the main topics developed in the course “Hacia la teorı́a cuántica
de campos” [1]. We will introduce special relativity as the necessary ingredient to guarantee the
local conservation of electric charge in quantum mechanics. The symmetries of the electromagnetic
Lagrangian will be extended to include the electron, as one Dirac spinor. The resulting Quantum
Electrodynamics theory will be used as a paradigm to explain the other fundamental interactions.
1.1 Lagrangian Formulation

1.1.1 Ecuaciones de Euler-Lagrange
Definamos
∂
∂µ = , (1.1)
∂xµ
En tres dimensiones, la acción de la se puede escribir como:
Z
S[φ, ∂µ φ] = d4 xL(φ, ∂µ φ) (1.2)
R
donde d4 x = dt dx dy dz. Considere primero una variación sólo de los campos, tal que (x = xµ )
δφ(x) = φ0 (x) − φ(x) (1.3)
De otro lado, con δx = x0 − x, la expansión de Taylor para f (x + δx) es
∂f
f (x + δx) = f (x) + δx + · · · (1.4)
∂x
3
4 CHAPTER 1. CLASSICAL FIELD THEORY
Para L, tenemos de la ec. (1.3)

L(φ0 , ∂µ φ0 ) = L(φ + δφ, ∂µ φ + ∂µ (δφ))
∂L ∂L
=L+ δφ + ∂µ (δφ) (1.5)
∂φ ∂(∂µ φ)
Entonces, de imponer que δS = 0, tenemos
Z Z
0 4 0 0
δS = S − S = d x L(φ , ∂µ φ ) − d4 x L(φ, ∂µ φ)
Z R R
∂L ∂L
= d4 x δφ + ∂µ (δφ)
R ∂φ ∂(∂µ φ)
Z Z
4 ∂L ∂L 4 ∂L
= dx − ∂µ δφ + d x ∂µ δφ
R ∂φ ∂(∂µ φ) R ∂(∂µ φ)
Z Z
4 ∂L ∂L ∂L
δS = dx − ∂µ δφ + δφ dσµ = 0. (1.6)
R ∂φ ∂(∂µ φ) σ ∂(∂µ φ)
Donde hemos aplicado el Teorema de Gauss

Z Z
3
∇ · Ad x = A · dS (1.7)
V S
generalizado a cuatro dimensiones. Como la variación de δφ es cero sobre la hipersuperficie σ resulta
Z
4 ∂L ∂L
dx − ∂µ δφ = 0. (1.8)
R ∂φ ∂(∂µ φ)
Como δφ es cualquier posible variación entre las fronteras de la hipersuperficie, el integrando debe
anularse y resultan las ecuaciones de Euler-Lagrange:

∂L ∂L
∂µ − = 0. (1.9)
∂(∂µ φ) ∂φ
La densidad Lagrangiana
L0 = L + ∂µ (η(x)) (1.10)
donde η(x) es cualquier función de los campos de la densidad Lagrangiana original, da lugar a la
Acción
Z Z Z
0 4 0 4
S = d xL = d xL + d4 x ∂µ η
R ZR ZR
4
= d x L + ηdσµ
R σ
=S , (1.11)
1.1. LAGRANGIAN FORMULATION 5
para una hipersuperficie suficientemente grande. De modo que dos densidades lagrangianas que
difieran solo en derivadas totales dan lugar a la misma Acción.
Usando el principio de mı́nima acción en términos del campo φ, tenemos que para la densidad
Lagrangiana (??)
" 2 #
2
1 1 ∂φ ∂φ
L= 2
− , (1.12)
2 v ∂t ∂z
las ecuaciones de Euler-Lagrange (1.9)

∂L ∂L ∂L
∂0 + ∂3 − =0
∂(∂0 φ) ∂(∂3 φ) ∂φ

∂ ∂L ∂ ∂L
+ =0
∂t ∂(∂φ/∂t) ∂z ∂(∂φ/∂z)

1 ∂ ∂φ ∂ ∂φ
− =0
v 2 ∂t ∂t ∂z ∂z
1 ∂ 2φ ∂ 2φ
− 2 =0 , (1.13)
v 2 ∂t2 ∂z
que corresponde a la ecuación de onda.
Generalizando a tres dimensiones vemos que la ecuación para una onda propagandose a una
velocidad v,
1 ∂ 2φ
2 2
− ∇2 φ = 0 , (1.14)
v ∂t
proviene de una densidad Lagrangiana (hasta derivadas totales)
" #
2
1 1 ∂φ
L= − ∇φ · ∇φ
2 v 2 ∂t

1 1
= ∂0 φ ∂0 φ − ∂i φ ∂i φ . (1.15)
2 v2
1.1.2 Teorema de Noether para simetrı́as internas

Para un campo complejo la ec. (1.2) se generaliza a
Z
∗ ∗
S[φ, φ , ∂µ φ, ∂µ φ ] = d4 x L(φ, φ∗ , ∂µ φ, ∂µ φ∗ ) (1.16)
R
Usando el mismo procedimiento, se obtiene

Z Z
∂L ∂L ∂L ∂L
δS = d x 4
− ∂µ δφ + 4
dx ∗
− ∂µ ∗)
δφ∗
R ∂φ ∂(∂ µ φ) ∂φ ∂(∂µ φ
Z R
∂L ∂L
+ d4 x ∂µ δφ + δφ∗ = 0. (1.17)
R ∂(∂µ φ) ∂(∂µ φ∗ )
Usando de nuevo el Teorema de Gauss resultan las ecuaciones de Euler Lagrange para φ y φ∗

∂L ∂L ∂L ∂L
∂µ − = 0, ∂µ ∗
− ∗ = 0. (1.18)
∂(∂µ φ) ∂φ ∂(∂µ φ ) ∂φ
De otro lado, si asumimos que φ y φ∗ satisfacen las ecuaciones de Euler–Lagrange, en lugar de asumir
que δφ y δφ∗ se anulan sobre la hipersuperficie, los dos primeros términos de la ec. (1.17) se anulan
y tendremos que para que δS = 0: Z
d4 x (∂µ J µ ) = 0, (1.19)
R
donde,
µ ∂L ∗ ∂L
J = δφ + δφ (1.20)
∂(∂µ φ) ∂(∂µ φ∗ )
Entonces J µ satisface la ecuación de continuidad:
∂µ J µ = 0 (1.21)
0
∂J
+∇·J=0 (1.22)
∂t
Integrando con respecto al volumen
Z Z
∂J 0 3
d x+ ∇ · J d3 x = 0,
∂t
ZV ZV
∂J 0 3
d x + J · dS = 0, (1.23)
V ∂t S
Escogiendo una superficie suficientemente grande que abarque toda la fuente de densidad ρ = J 0 , de
la corriente J, el segundo integrando es cero y
Z
d
ρ d3 x = 0. (1.24)
dt V
Este resultado es conocido como Teorema de Noether. Éste establece que para toda transformación
continua del tipo (1.3), debe existir una cantidad conservada, dQ/dt = 0, que en este caso corresponde
a Z
Q= ρ d3 x. (1.25)
V
1.1.3 Teorema de Noether para simetrı́as externas

Para el caso de una simetrı́a externas, por ejemplo la correspondiente a una traslación espacio–
temporal
µ
xµ → x0 =xµ + δaµ
δxµ =δaµ (1.26)
tenemos
φ0 (x0 ) = φ0 (x + δa) (1.27)

∂φ0 (x) µ
≈ φ0 (x) + δa (1.28)
∂xµ
∂
= [φ(x) + δφ(x)] + [φ(x) + δφ(x)]δaµ (1.29)
∂xµ
∂φ(x) µ
≈ φ(x) + δφ(x) + δa , (1.30)
∂xµ
donde, por simplicidad, φ es de nuevo un campo real, y en el último paso hemos despreciado un
término de orden δφδaµ . Entonces,
∂φ(x) µ
∆φ(x) ≡ φ0 (x0 ) − φ(x) = δφ(x) + δa . (1.31)
∂xµ
Para una traslación, ∆φ(x) = 0, ver figura 1.1. De modo que
δφ = −(∂µ φ)δaµ , (1.32)
y la transformación del campo φ como consecuencia de la traslación es
φ(x) → φ0 (x) = φ(x) + δφ(x) = φ(x) − (∂µ φ(x))δaµ . (1.33)
Si aµ es constante (un análisis más general es hecho en [?])
d 4 x0 = d 4 x (1.34)
En este caso, asumiendo que el campo satisface las ecuaciones de Euler-Lagrange y usando la ec. (1.32)
Figure 1.1: Traslación de función y coordenadas en una dimensión: φ(x) = φ0 (x0 )
y (1.9) tenemos
Z Z
4 0 0 0µ
δS = d x L(φ , ∂µ φ , x ) − d4 x L(φ(x), ∂µ φ(x), x)
ZR R Z
4
= µ
d x L(φ + δφ, ∂µ φ + ∂µ (δφ), x + δa ) − µ
d4 x L
ZR R
Z
4 ∂L ∂L
≈ d x L+ δφ + ∂µ (δφ) + (∂µ L)δa − µ
d4 x L
∂φ ∂(∂ µ φ)
ZR R
∂L ∂L
= d4 x δφ + ∂µ (δφ) + (∂µ L)δaµ
∂φ ∂(∂µ φ)
ZR
4 ∂L ∂L µ
= dx ∂µ δφ + ∂µ (δφ) + (∂µ L)δa
R ∂(∂µ φ) ∂(∂µ φ)
Z
4 ∂L µ
= d x ∂µ δφ + (∂µ L)δa
R ∂(∂µ φ)
Z
4 ∂L µ
= d x ∂µ δφ + Lδa
R ∂(∂µ φ)
Z
4 ∂L ν µ
= d x ∂µ − ∂ν φδa + Lδa
R ∂(∂µ φ)
Z
4 ∂L ν µ ν
= d x ∂µ − ∂ν φδa + Lδν (δa )
R ∂(∂µ φ)
Z
4 ∂L µ ν
= d x ∂µ − (∂ν φ) + δν L δa
R ∂(∂µ φ)
(1.35)
Z
= d4 x ∂µ (Tνµ δaν ) = 0. (1.36)
R
Y por consiguiente
∂µ Tνµ δaν = 0, (1.37)
De modo que para cada ν, con δaν 6= 0, se satisface:
∂µ Tνµ = 0, (1.38)
donde
∂L
Tνµ = (∂ν φ) − δνµ L (1.39)
∂(∂µ φ)
El tensor Tνµ proviene de asumir la homogeneidad del espacio y el tiempo y es llamado el tensor de
momentum–energı́a.
Para una traslación temporal: ν = 0, se genera entonces la ecuación de continuidad:
∂µ T0µ = 0 (1.40)
Donde la densidad de Energı́a, o más de forma más general: la densidad Hamiltonina corresponde a
T00
∂L
H = T00 = φ̇ − L (1.41)
∂ φ̇
∂φ(x)
= π(x) − L. (1.42)
∂t
Comparando con la expresión correspondiente en la formulación Lagrangiana de la Mecánica Clásica,
tenemos que si φ(x) es la variable canónica, la variable canónica conjugada es π(x)
∂L
π(x) = . (1.43)
∂(∂φ(x)/∂t)
El teorema de Noether en este caso establece que la invarianza de la Acción bajo traslaciones tem-
porales da lugar a la ecuación de continuidad (1.38) para ν = 0
∂µ T0µ = 0 (1.44)
cuya carga conservada corresponde a la energı́a
Z Z
3 0
H= d x T0 = d3 x H. (1.45)
V V
De igual forma la invarianza bajo traslaciones espaciales de lugar a ecuaciones de continuidad para
cada componente ν = i (i = 1, 2, 3)
∂µ Tiµ = 0, (1.46)
cuyas densidad de cargas conservadas, Ti0 , que en forma vectorial escribiremos como T0 , dan lugar
a la conservación del momentum
Z
P= d3 x T0 . (1.47)
V
Generalizando a un campo complejo

∂L ∂L
Tνµ = (∂ν φ) + (∂ν φ∗ ) − δνµ L (1.48)
∂(∂µ φ) ∂(∂µ φ∗ )
1.2 Global gauge invariance

Haciendo ~ = 1, el Lagrangiano que da lugar a la ecuación de Schrödinger es

∗ ∗ 1 ∗ i ∗ ∂ψ ∂ψ ∗
L(ψ, ψ , ∂µ ψ, ∂µ ψ ) = ∇ψ · ∇ψ − ψ − ψ + ψ∗V ψ (1.49)
2m 2 ∂t ∂t
1 i
= ∂i ψ ∗ ∂i ψ − (ψ ∗ ∂0 ψ − ∂0 ψ ∗ ψ) + ψ ∗ V ψ.
2m 2
Aplicando las ecuaciones de Euler-Lagrange (1.18) para la función de onda ψ ∗ obtenemos la ecuación
de Scrödinger con ~ = 1:

∂L ∂L ∂L ∂L ∂L
0 = ∂µ ∗
− ∗
= ∂0 ∗
+ ∂i ∗
− . (1.50)
∂(∂µ ψ ) ∂ψ ∂(∂0 ψ ) ∂(∂i ψ ) ∂ψ ∗
Como
∂L i ∂L i
= − ψ∗ ∗
= ψ
∂(∂0 ψ) 2 ∂(∂0 ψ ) 2
∂L 1 ∂L 1
= ∂i ψ ∗ = ∂i ψ (1.51)
∂(∂i ψ) 2m ∂(∂i ψ ∗ ) 2m
∂L i ∂L i
= ∂0 ψ ∗ + ψ ∗ V = − ∂0 ψ + V ψ.
∂ψ 2 ∂ψ ∗ 2
Entonces, reemplazando la ec. (1.51) en la ec. (1.50), tenemos

∂L ∂L i 1 i
0 = ∂µ − = ∂0 ψ + ∂i ∂i ψ − − ∂0 ψ + V ψ
∂(∂µ ψ ∗ ) ∂ψ ∗ 2 2m 2
i 1 i
= ∂0 ψ + ∂i ∂i ψ + ∂0 ψ − V ψ. (1.52)
2 2m 2
1.2. GLOBAL GAUGE INVARIANCE 11
Que puede escribirse como

∂ 1 2
i ψ= − ∇ + V ψ. (1.53)
∂t 2m
El Lagrangiano en ec (1.49), y por consiguiente la Acción, es invariante bajo una transformación
de fase
ψ → ψ 0 = eiθ ψ. (1.54)
Por consiguiente, de acuerdo al Teorema de Noether, debe existir una cantidad conservada. La
corriente conservada se obtine de la ec. (1.20). Para los campos ψ y ψ ∗ , tenemos
δψ = ψ 0 − ψ = (eiθ − 1)ψ ≈ iθψ (1.55)

δψ ∗ ≈ −iθψ ∗ . (1.56)
Usando además la ec. (1.51) en la definición de J 0 dada por la ec. (1.20), tenemos

0 ∂L ∗ ∂L
J = δψ + δψ
∂(∂0 ψ) ∂(∂0 ψ ∗ )
i i
= − ψ ∗ (iθψ) + (−iθψ ∗ ) ψ
2 2
= θψ ∗ ψ, (1.57)
y

i ∂L ∗ ∂L
J = δψ + δψ
∂(∂i ψ) ∂(∂i ψ ∗ )
1 1
= ∂i ψ ∗ (iθψ) + (−iθψ ∗ ) ∂i ψ
2m 2m
iθ
= (∂i ψ ∗ ψ − ψ ∗ ∂i ψ) . (1.58)
2m
Entonces, normalizando apropiadamente la corriente escogiendo θ = 1, tenemos
J 0 = ψ∗ψ (1.59)
i
J= (ψ∇ψ ∗ − ψ ∗ ∇ψ) . (1.60)
2m
De acuerdo a la ec. (1.59), la cantidad conservada corresponde a la probabilidad de la función de
onda y normalizando apropiadamente la ec. (1.25)
Z
Qρ = ψ ∗ ψ d3 x = 1. (1.61)
V
En cuanto a las simetrı́as externas, tenemos de la ec. (1.39) que da lugar a las ecuaciones de
continuidad (1.44)(1.46)
∂µ T0µ = 0,
∂µ Tiµ = 0 (1.62)
Las cargas conservadas corresponden entonces a T00 y Ti0 . Usando las ecs. (1.51) en la ec. (1.48)
∂L ∂L
Ti0 = (∂i ψ) + (∂i ψ ∗ )
∂(∂0 ψ) ∂(∂0 ψ ∗ )
i i
Ti0 = − ψ ∗ (∂i ψ) + (∂i ψ ∗ )ψ (1.63)
2 2
Entonces, definiendo
i
T0 = (ψ∇ψ ∗ − ψ ∗ ∇ψ) (1.64)
2
Además
i
T0 = (∇(ψ ∗ ψ) − ψ ∗ ∇ψ − ψ ∗ ∇ψ)
2
i
= −iψ ∗ ∇ψ + ∇(ψ ∗ ψ) . (1.65)
2
Integrando en el volumen
Z Z Z
∗ i
0 3
T d x = −i 3
ψ ∇ψ d x + ∇ ψ ∗ ψ d3 x (1.66)
V V 2 V
De acuerdo a la ec. (1.61), la última integral es una constante y
Z Z
0 3
T d x = −i ψ ∗ ∇ψd3 x
V ZV
hb
pi = ψ∗pb ψd3 x (1.67)
V
De modo que hb
pi son las cargas conservadas asociadas al valor esperado el operador de momentum
b = −i∇ .
p (1.68)
De otro lado
∂L ∂L
T00 = ∂0 ψ + ∂0 ψ ∗ −L
∂(∂0 ψ) ∂(∂0 ψ ∗ )
i i 1 i
= − ψ ∗ ∂0 ψ + ∂0 ψ ∗ ψ − ∂i ψ ∗ ∂i ψ + (ψ ∗ ∂0 ψ − ∂0 ψ ∗ ψ) − ψ ∗ V ψ
2 2 2m 2
1
=− ∂i ψ ∗ ∂i ψ − ψ ∗ V ψ (1.69)
2m
1.2. GLOBAL GAUGE INVARIANCE 13
Como las corrientes solo están determinadas hasta un factor de proporcionalidad, definimos
1
H ≡ −T00 = ∇ψ ∗ · ∇ψ + ψ ∗ V ψ
2m
1 1 ∗ 2
= ∇ · (ψ ∗ ∇ψ) − ψ ∇ ψ + ψ ∗ V ψ. (1.70)
2m 2m
Integrando sobre el volumen y usando la ec. (1.67)
Z Z Z
3 1 ∗ ∗ 1 2
Hd x = ∇ · (ψ ∇ψ) + ψ − ∇ + V ψ d3 x
V 2m V V 2m
Z Z
1 ∗ ∗ 1 2
= ∇ · (ψ ∇ψ) + ψ − ∇ + V ψ d3 x
2m V 2m
Z V
i ∗ 1 2
= ∇ · hbpi + ψ − ∇ + V ψ d3 x
2m V 2m
Z
∗ 1
= ψ − ∇ + V ψ d3 x .
2
(1.71)
V 2m
Entonces
Z Z
3 ∗ 1 2
H≡ Hd x = ψ − ∇ +V ψ d3 x
V V 2m
Z
= b = hHi.
d3 x ψ ∗ Hψ b (1.72)
V
Que es un resultado bien conocido de la mecánica cuántica.

Como
b = 1 p̂2 + Vb ,
H (1.73)
2m
podemos escribir la ec. (1.53) como
∂ b .
i ψ = Hψ (1.74)
∂t
Podemos identificar entonces los operadores de energı́a y momentum.
b =i∂,
H p̂ = −i ∇. (1.75)
∂t
Retornando a la ec. (1.67), tenemos que para la solución de partı́cula libre de la ecuación de
Schrödinger
ψ = A e−ik·x , (1.76)
la condición de normalización en ec. (1.61) implica que |A|2 = 1/L3 , y

Z
T0 d3 x = k. (1.77)
V
Ejercicio: De la ec. (1.72) obtenega la densidad Hamiltoniana, y usando la ec. (1.41) encontrar la densidad
Lagrangiana (1.49).
1.3 Local phase invariance in the Scrödinger’s Lagrangian

When we discuss the wave function ψ(x), x represents the point in space at which we want to know
the value of the wave function. Since complex numbers are, well, complex, you can’t represent
them by a position on a simple number line. Instead, the have to be represented by a point in a
two–dimensional plot.
In addition the length of the arrow pointing to the complex number we also need an angle to
specify exactly how to draw the arrow pointing to the complex number. The observable is encoded
into the length of the arrow representing the value of the complex valued wave function at that point
of the space–time. Its angle is unobservable.
The complex number ψ(x) in the Scrödinger equation is just the number whose square is the
relative probability of finding the object at that point.
Now, suppose that you arbitrarily decide to make a change of phase of the wave function –to
change, at every point in space, the angle θ of the complex number ψ makes with the real axis. Here
is the critical point: Is this change phase is global, if the phase that you change the phase angle θ
is the same everywhere in space, the this change of phase will not destroy the delicate and essential
balance between the kinetic and potential energy in the Scrödinger equation.
However, in the view implemented by Einstein’s relativity, the need to require that quantum–
mechanical systems be unaltered only by global changes of phase seemed to be very unnatural. Once
you choose the phase of the wave function at one space-time point, the requirement of global phase
invariance fixes it at all other space-time points:
As usually conceived however, this arbitrariness is subject to the following limitation: once one
choose [the phase of the wave function] at one space–time point, one is then not free to make
any choices at other space–time points.
It seems that it is not consistent with the localized field concept that underlies the usual physical
theories. In the present paper we wish to explore the possibility of requiring all the interactions
to be invariant under independent [change of phases] at all space-time points.
Yang-Mills, Physical Review, 1954

1.3. LOCAL PHASE INVARIANCE IN THE SCRÖDINGER’S LAGRANGIAN 15
This is similar to what happens in electromagnetic theory expressed in terms of scalar and vector
potentials. The can be changed by arbitrary functions in a such way that the measured electric
and magnetic fields remain invariant. As we will see, this feature is deeply connected with the local
conservation of electric charge.
We start again with the Scrödinger Lagrangian as written in eq. (1.49):

∗ ∗ 1 ∗ i ∗ ∂ψ ∂ψ ∗
L(ψ, ψ , ∂µ ψ, ∂µ ψ ) = ∇ψ · ∇ψ − ψ − ψ + ψ∗V ψ (1.78)
2m 2 ∂t ∂t
1 i
= ∂i ψ ∗ ∂i ψ − (ψ ∗ ∂0 ψ − ∂0 ψ ∗ ψ) + ψ ∗ V ψ.
2m 2
This Lagrangian is not invariant under local phase changes of the wave function:

∂µ ψ → ∂µ ψ 0 =∂µ eiθ(x) ψ

= ∂µ eiθ(x) ψ + eiθ(x) ∂µ ψ
=eiθ(x) (i∂µ θ(x)) ψ + eiθ(x) ∂µ ψ
=eiθ(x) [i∂µ θ(x) + ∂µ ] ψ . (1.79)
In order to have a new Lagrangian invariant under local phase changes, or local gauge transformations,
we need to introduce a new term to compensate for the term arising from the derivate of eiθ(x) :

Dµ ψ → Dµ0 ψ 0 =(∂µ + Xµ0 ) eiθ(x) ψ

=eiθ(x) [i∂µ θ(x) + ∂µ ] ψ + Xµ0 eiθ(x) ψ

=eiθ(x) i∂µ θ(x) + ∂µ + Xµ0 ψ . (1.80)
The transformation condition of the new term Xµ , in order to compensate for the term arising from
the derivative of the local phase, i∂µ θ(x), is just that
Xµ → Xµ0 = Xµ − i∂µ θ(x) . (1.81)
Replacing back in Eq. (1.80) we have

Dµ ψ → (Dµ ψ)0 = Dµ0 ψ 0 =(∂µ + Xµ0 ) eiθ(x) ψ
=eiθ(x) [i∂µ θ(x) + ∂µ + Xµ − i∂µ θ(x)] ψ
=eiθ(x) [∂µ + Xµ ] ψ
=eiθ(x) (Dµ ψ) . (1.82)
Note that Dµ ψ transforms like the field ψ, and because of this is called the covariant derivative of
ψ. Similarly
∗ ∗
(Dµ ψ)∗ → (Dµ ψ)0 =(∂µ + Xµ0 ) ψ ∗ e−iθ(x)

= −i∂µ θ(x) + ∂µ + Xµ∗ + i∂µ θ(x) ψ ∗ e−iθ(x)

= ∂µ + Xµ∗ ψ ∗ e−iθ(x)
= (Dµ ψ)∗ e−iθ(x) . (1.83)
It is convenient to redefine Xµ in terms of Aµ :
1
Aµ ≡ Xµ , (1.84)
iq
such that the covariant derivative can be conveniently written as
Dµ = ∂µ + iqAµ . (1.85)
The transformation properties of Aµ can be obtained from the Xµ transformation in eq. (1.81):
iqAµ →iqA0µ = iqAµ − i∂µ θ(x)

1
Aµ →A0µ = Aµ − ∂µ θ(x) . (1.86)
q
We define local gauge invariance as an arbitrary way of choosing the complex phase factor of a
charged field1 at all space time points.
In this way, we can change the original Lagrangian for a new one which is invariant under local
phase transformations:
1 i
L(ψ, ψ ∗ , ∂µ ψ, ∂µ ψ ∗ , Aµ ) = (Di ψ)∗ Di ψ − [ψ ∗ D0 ψ − (D0 ψ)∗ ψ] + ψ ∗ V (x)ψ. (1.87)
2m 2
where
1
Aµ → A0µ = Aµ − ∂µ θ(x) . (1.88)
q
1
like the electron field as described by the usual Scrödinger equation.
1.4. NOTACIÓN RELATIVISTA 17
This is just the gauge transformation which left the Electromagnetic fields invariant. In fact, the
new Lagrangian is now invariant under the local phase transformations
1 ∗ i ∗ ∗ ∗
L → L0 = (Di ψ)0 (Di ψ)0 − ψ 0 (D0 ψ)0 − (D0 ψ)0 ψ 0 + ψ 0 V (x)ψ 0
2m 2
1 ∗ −iθ(x) iθ(x)
= (Di ψ) e e (Di ψ)
2m
i
− ψ ∗ e−iθ(x) eiθ(x) (D0 ψ) − (D0 ψ)∗ e−iθ(x) eiθ(x) ψ + ψ ∗ e−iθ(x) eiθ(x) V (x)ψ.
2
=L . (1.89)
To preserve invariance one notices that it is necessary to counteract the variation of θ with x, y,
z, and t by introducing the electromagnetic field Aµ . In this way, the electromagnetic interaction is
obtained as the result of impose local gauge invariance under U (1) (local phase transformations). To
fully implement the gauge principle, i.e, the paradigm to obtain the interactions as the result of the
gauge invariance, we need to introduce some concepts of special relativity to be developed below.
1.4 Notación relativista

Las transformaciones de Lorentz se definen como la transformaciones que dejan invariante al producto
escalar en el espacio de Minkowski definido como
2 2
a2 = gµν aµ aν ≡ aν aν = a0 − ai ai = a0 − a · a (1.90)
donde µ, ν = 0, 1, 2, 3, i = 1, 2, 3 y se asume suma sobre ı́ndices repetidos. Además
aν ≡ gµν aµ (1.91)
Finalmente la métrica usada se define como

 
1 0 0 0
0 −1 0 0
{gµν } =  
0 0 −1 0  (1.92)
0 0 0 −1
donde {gµν } denota la forma matricial del tensor gµν .

El producto de dos cuadrivectores se define en forma similar como
aν bν = gµν aµ bν = a0 b0 − a · b (1.93)
El inverso de la métrica es
{g µν } ≡ {gµν }−1 = {gµν } (1.94)
tal que
g µα gαν = δνµ and aµ = g µν aν (1.95)
Bajo una transformación de Lorentz.
µ
aµ → a0 =Λµ ν aν (1.96)
aµ → a0 µ =Λµ ν aν
La invarianza del producto escalar en ec. (1.93)
µ
a0 b0 µ =aµ bµ
α β
gαβ a0 b0 = gµν aµ bν
gαβ Λα µ aµ Λβ ν bν = gµν aµ bν
Λα µ gαβ Λβ ν aµ bν = gµν aµ bν , (1.97)
da lugar a
gµν = Λα µ gαβ Λβ ν or {gµν } = {Λµ α }T {gαβ } Λβ ν . (1.98)
En notación matricial
g = ΛT gΛ . (1.99)
From eq. (1.98) we also have
g ρµ gµν =g ρµ Λα µ gαβ Λβ ν
δνρ =Λβ ρ Λβ ν , (1.100)
or
Λα µ Λα ν = δνµ . (1.101)
Since
µ
Λ−1 α
Λα ν = δνµ (1.102)
the inverse of Λ is
µ
Λ−1 α
= Λα µ , (1.103)
or
µ
Λ−1 ν
= Λν µ , (1.104)
• Example: Lorentz invariance

µ
aµ bµ → a0µ b0 =Λµ ν aν Λµ ρ bp
=Λµ ν aν Λµ ρ bp
ν
= Λ−1 µ Λµ ρ aν bp
=δρν aν bp
=aν bν .
Como un ejemplo de Transformación de Lorentz considere un desplazamiento a lo largo del eje x

   0   t+vx    
t t √ cosh ξ sinh ξ 0 0 t
1−v 2
 x   x 0  √x+vt   sinh ξ cosh ξ 0 0   x 
{xµ } = 
y 
 →  0  =  1−v2  = 
y   y   0
   = {Λµ ν } {xν } , (1.105)
0 1 0 y 
z z0 z 0 0 0 1 z
donde
1
cosh ξ = γ sinh ξ = vγ, and γ=√ . (1.106)
1 − v2
y, por ejemplo:
t + vx
t cosh ξ + x sinh ξ = γ(t + vx) = √ . (1.107)
1 − v2
El Λµ ν definido en la ec. (1.105) satisface la condición en ec. (??),
   
cosh ξ sinh ξ 0 0 1 0 0 0 cosh ξ sinh ξ 0 0
 sinh ξ cosh ξ 0 0 0 −1 0 0  
ΛT gΛ =     sinh ξ cosh ξ 0 0
 0 0 1 0 0 0 −1 0   0 0 1 0
0 0 0 1 0 0 0 −1 0 0 0 1
  
cosh ξ − sinh ξ 0 0 cosh ξ sinh ξ 0 0
 sinh ξ − cosh ξ 0 0   sinh ξ cosh ξ 0 0
 
=  0

0 −1 0   0 0 1 0
0 0 0 −1 0 0 0 1
 2 2

cosh ξ − sinh ξ cosh ξ sinh ξ − cosh ξ sinh ξ 0 0
cosh ξ sinh ξ − cosh ξ sinh ξ sinh2 ξ − cosh2 ξ 0 0
= 

0 0 −1 0 
0 0 0 −1
=g (1.108)
Denotaremos los cuadrivectores con ı́ndices arriba como
aµ = (a0 , a1 , a2 , a3 ) = (a0 , a) (1.109)
Entonces el correspondiente cuadrivector con ı́ndices abajo, usando la ec. (1.91), es
aµ = (a0 , a1 , a2 , a3 ) = (a0 , −a1 , −a2 , −a3 ) = (a0 , −a). (1.110)
Con esta notación, el producto escalar de cuadrivectores puede expresarse como el producto escalar
de los dos vectores de cuatro componente aµ y aµ .
1.4.1 Ejemplos de cuadrivectores
xµ =(x0 , x1 , x2 , x3 ) = (t, x, y, z) = (t, x) (1.111)

pµ =(p0 , p1 , p2 , p3 ) = (E, px , py , pz ) = (E, p) (1.112)
De la relatividad especial tenemos que
E =γm
p =γmv . (1.113)
Por lo tanto, ya que v 2 = v2 = |v|2
E 2 − p2 = γ 2 m2 (1 − v 2 ) = m2 . (1.114)
El invariante de Lorentz asociado a pµ corresponde a la ecuación de momento energı́a una vez se

identifica la masa de una partı́cula con su cuadrimomentum
p2 = pµ pµ = m2 = E 2 − p2 (1.115)
De [?]
The intuitive understanding of this equation is that the energy of a particle is partially
due to its motion and partially due to the intrinsic energy of its mass. The application
to particle detectors is that if you know the mass of a particular particle, or if it’s going
so fast that its energy and momentum are both huge so that the mass can be roughly
ignored, then knowing the energy tells you the momentum and vice versa
Para p = 0, es decir cuando la partı́cula está en reposo se reduce a la famosa ecuación E = mc2
(c = 1)
Del electromagnetismo tenemos
J µ = (J 0 , J) = (ρ, J) (1.116)
Aµ = (A0 , A) = (φ, A) (1.117)

Del cálculo vectorial

µ ∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
∂ ≡ = , , , = ,− ,− ,−
∂xµ ∂x0 ∂x1 ∂x2 ∂x3 ∂x0 ∂x1 ∂x2 ∂x3

∂ ∂ ∂ ∂
= ,− ,− ,−
∂t ∂x ∂y ∂z
=(∂0 , −∇) = (∂ 0 , −∇) (1.118)

∂ ∂ ∂ ∂ ∂
∂µ = = , , , = (∂0 , ∇) (1.119)
∂xµ ∂t ∂x ∂y ∂z
Por consiguiente:
∂
∇= (1.120)
∂x
Producto escalar:
aµ bµ = gµν aµ bν = a0 b0 − a1 b1 − a2 b2 − a3 b3 = a0 b0 − ai bi = a0 b0 − a · b (1.121)
Entonces
∂a0
∂ µ aµ = +∇·a (1.122)
∂t
La ecuación de continuidad ∂µ J µ = 0 es un invariante bajo transformaciones de Lorentz: ∂µ0 J 0 µ =
∂µ J µ = 0 El operador cuadrático es, usando la ec. (1.90)
∂2 ∂2 ∂2 ∂2
≡ ∂µ ∂ µ = ∂ 0 ∂ 0 − ∇ 2 = − − − (1.123)
∂t2 ∂x2 ∂y 2 ∂z 2
Los operadores de energı́a y momentum de la mecánica cuántica también forman un cuadrivector
b p̂)
p̂µ = (p̂0 , p̂) = (H, (1.124)
b y p̂ dados en la ec. (1.75). Entonces
con H,
∂
p̂µ = i∂ µ = i(∂ 0 , ∂ i ) = i( , −∇) (1.125)
∂t
Las derivada covariante en la ec, (1.85) es en términos de componentes:

D0 =∂0 + iqA0
Di =∂i − iqAi . (1.126)
Definiendo D de la misma forma que el gradiente, tenemos
Di =∂i + iqAi
D =∇ + iqA . (1.127)
Podemos definir el cuadrivector
Dµ =(D0 , D)
=(∂0 , ∇) + iq(A0 , −A)
=(∂0 , ∂i ) + iq(A0 , −Ai )
=(∂0 , ∂i ) + iq(A0 , Ai )
=(∂0 + iqA0 , ∂i + iAi )
=(D0 , Di )
=(D0 , Di ) , (1.128)
donde hemos usado
Di = ∂i + iqAi . (1.129)
Además Aµ tiene la transformación gauge
∂χ
A → A0 = A + ∇χ A0 → A00 = A0 − (1.130)
∂t
En notación de cuadrivectores

µ 0µ 0 ∂χ
A →A = A − , A + ∇χ
∂t

0 ∂χ i
= A − , A + ∂i χ
∂t

= A0 − ∂ 0 χ, Ai − ∂ i χ

= A0 , Ai − ∂ 0 χ, ∂ i χ
µ
Aµ → A0 =Aµ − ∂ µ χ . (1.131)
Note that the eq. (1.131) can be written as
Aµ → A0µ = Aµ − ∂µ χ(x) (1.132)
which is just the transformation obtained in eq. (1.88).
1.4.2 Lorentz tranformation for fields

The scalar field is defined by their properties under Lorentz transformation. In section 1.1.3 we
study the behavior of one scalar field under a space–time translation. Under a general Lorentz
transformation
µ
xµ → x0 = Λµ ν xν , (1.133)
Now we will study the effect of a Lorentz tranformation on the field φ(x), for example under a boost.
By definition the scalar field does not change by the Lorentz transformation, the functional form is
unaltered the scalar field still satisfy
φ(x) → φ0 (x0 ) = φ(x) . (1.134)
By using eq. (1.133) we have
φ0 (x0 ) = φ(Λ−1 x0 ) . (1.135)
Therefore, for an arbitrary space-time point we have that the scalar field transforms under a Lorentz
transformation as
φ(x) → φ0 (x) = φ(Λ−1 x) . (1.136)
In order to check the Lorentz invariance of the scalar we need to obtain the Lorentz transformation
properties for ∂µ . It is convinient to invert eq. (1.133)
µ α µ
Λ−1 α x0 = Λ−1 α Λα ν xν
=δνµ xν
=xµ , (1.137)
1
−1 µ 1
ν = Λ , (1.138)
x0 ν xµ
or
1
−1 ν 1
= Λ , (1.139)
x0 µ µ xν
and the defintion of the Lorentz transformation itself:

µ ν
g µν = Λ−1 ρ g ρσ Λ−1 σ . (1.140)
From eq. (1.139) we can obtain the Lorentz transformation for ∂µ = ∂/∂xµ :
∂ ∂
−1 ν
= Λ
∂x0 µ µ
∂x
ν
0 −1 ν
∂µ = Λ ∂ ,
µ ν
(1.141)
The field Aµ (x) transforms simultaneously as field and as vector under Lorentz transformation
µ
Aµ (x) → A0 (x0 ) = Λµ ν Aν (Λ−1 x) . (1.142)
1.5 Vector field Lagrangian

We are now are in position to answer the following question: What is the most general Lagrangian
for a the four–components field Aµ compatible with Lorentz invariance and the gauge transformation
µ
Aµ → A0 = Aµ − ∂ µ χ(x) ? (1.143)
Definiendo
F µν = ∂ µ Aν − ∂ ν Aµ
Gµν = ∂ µ Aν + ∂ ν Aµ
El Lagrangiano que da lugar a una Acción invariante de Lorentz para el cuadrivector Aµ es, hasta
derivadas totales y potencias en los campos de hasta dimensión 4:
1 1 1
L = − F µν Fµν − Gµν Gµν − J µ Aµ + m2 Aµ Aµ + λ1 ∂ν Aν (x)Aµ (x)Aµ (x) + λ2 Aµ Aµ Aν Aν
4 4 2
µν µν
+ λ3 F (x)Aµ (x)Aν (x) + λ4 G (x)Aµ (x)Aν , . (1.144)
• Ejercicio: Show that terms like ∂ µ Aν (x)∂µ Aν (x), and hence F µν Fµν , transforms as

∂ µ Aν Λ−1 x ∂µ Aν Λ−1 x (1.145)
Hint: use the Lorentz transformation properties of ∂µ in eq. (1.141).
In the case of J µ Aµ :
µ ν
J µ (x)Aµ (x) → gµν J 0 (x)A0 (x) =gµν Λµ ρ J ρ Λ−1 x Λν σ Aσ Λ−1 x

=Λµ ρ gµν Λν σ J ρ Λ−1 x Aσ Λ−1 x

=gρσ J ρ Λ−1 x Aσ Λ−1 x , (1.146)
1.5. VECTOR FIELD LAGRANGIAN 25
in the case ∂ν Aν (x)Aµ (x)Aµ (x):

ν µ σ
∂ν Aν (x)Aµ (x)Aµ (x) → ∂ 0 ν A0 (x0 )A0 µ (x0 )A0 (x0 ) = Λ−1 Λν ρ ∂σ Aρ Λ−1 x Aµ Λ−1 x Aµ Λ−1 x
ν

=δρσ ∂σ A Λ−1 x Aµ Λ−1 x Aµ Λ−1 x
ρ

=∂ρ Aρ Λ−1 x Aµ Λ−1 x Aµ Λ−1 x ,
(1.147)
and similarly for the other terms. Under a Lorentz transformation the full Lagrangian transform as
L(x) → L0 (x) = L(Λ−1 x) (1.148)
Since the Action involves the integration over all the points, it is invariant under the Lorentz transfor-
mation. The J µ (x) does not involves the introduction a new vector field, because it will be identified
later as the 4–current.
Terms like
Kν Aν (x)Aµ (x)Aµ (x) , (1.149)
(for Kν constant) are not Lorentz invariant:
ν µ
Kν Aν (x)Aµ (x)Aµ (x) → Kν A0 (x)A0 µ (x)A0 (x) =Kν Λν ρ Aρ Λ−1 x Aµ Λ−1 x Aµ Λ−1 x . (1.150)
Kν (x)Aν (x)Aµ (x)Aµ (x) is Lorentz covariant but not gauge-invariant (see below).
Bajo la transformación gauge (1.132)
µν ν µ
F µν → F 0 =(∂ µ A0 − ∂ ν A0 )
=∂ µ Aν − ∂ µ ∂ ν χ − ∂ ν Aµ + ∂ ν ∂ µ χ
=∂ µ Aν − ∂ ν Aµ − ∂ µ ∂ ν χ + ∂ µ ∂ ν χ
=F µν (1.151)
Si queremos que la Acción refleja las simetrı́as de las ecuaciones de Maxwell debemos mantener
sólo los términos del Lagrangiano para Aµ en (1.144) que sean invariantes hasta una derivada total.
Bajo una transformación gauge, cada uno de los términos
1 1
− Gµν Gµν + m2 Aµ Aµ + λ1 ∂µ Aµ Aν Aν + λ2 Aµ Aµ Aν Aν + λ3 F µν Aµ Aν + λ4 Gµν Aµ Aν + Kν (x)Aν Aµ Aµ
4 2
dan lugar a un δL = 6 ∂µ (algo) y la Acción no es invariante bajo la transformación gauge. Para los
términos restantes
1
L = − F µν Fµν − J µ Aµ , (1.152)
4
usando la ec. (1.166), tenemos

1 µν 0 1
δL = L0 − L = − F 0 Fµν − J µ A0µ + F µν Fµν + J µ Aµ
4 4
= − J µ Aµ + J µ ∂µ θ(x) − J µ Aµ
=∂µ (J µ χ) − (∂µ J µ )θ(x) (1.153)
For the action

Z
δS = d4 x [∂µ (J µ χ) − (∂µ J µ )θ(x)]
Z
= − d4 x(∂µ J µ )θ(x)
Z Z ∞
3
=− d x dt(∂µ J µ )θ(x) . (1.154)
−∞
In order to have δS = 0 we need to assume for the while that ∂µ J µ = 0. However we will see that
this is just a self-consistent condition.
In summary, if the electromagnetic current is conserved, then the Lagrangian is invariant under
the gauge transformation (1.143). Note that the Lagrangian density is not locally gauge invariant.
However, the action (and hence the theory) is gauge invariant.
Por lo tanto, el Lagrangiano
1
L = − F µν Fµν − J µ Aµ (1.155)
4
es el más general que da lugar a una Acción invariante de Lorentz e invariante gauge local.
The definition of F µν already includes the homogeneous Maxwell equations. To see this we note
first that the only non-zero F µν components are
(
F µ0 = F i0 ν = 0
F µν = (1.156)
F µl = F ml ν = l
For ν = 0 we have
F i0 = ∂ i A0 − ∂ 0 Ai
∂A0 ∂Ai
=( − )
∂xi ∂x0
∂A0 ∂Ai
= −( i + 0 )
∂x ∂x
i
=E (1.157)
where
∂A
E = −∇φ − . (1.158)
∂t
while for ν = l we have
F ml = ∂ m Al − ∂ l Am
= (δlj δmi − δli δmj )∂ i Aj
= −(δlj δmi − δli δmj )∂i Aj
= (δli δmj − δlj δmi )∂i Aj
∂Aj
= (δli δmj − δlj δmi ) i
∂x
∂Aj
= lmk ijk i
∂x
= lmk (∇ × A)k
= lmk B k , (1.159)
where
B = ∇ × A. (1.160)
Then we have
 
0 −E 1 −E 2 −E 3
E1 0 213 B 3 312 B 2 
{F µν } = 
E 2

123 B 3
0 321 B 1 
E3 132 B 2 231 B 1 0
 
0 −E 1 −E 2 −E 3
E 1 0 −B 3 B 2 
=E 2
. (1.161)
B3 0 −B 1 
E3 −B 2 B 1 0
From eqs. (1.158), and (1.160)

∂
∇ × E = −∇ × ∇φ − ∇×A
∂t
∂B
=− ,
∂t
and
∇ · B = ∇ · (∇ × A)
=0
which are just the homogeneous Maxwell equations. Therefore the expression
F µν = ∂ µ Aν − ∂ ν Aµ . (1.162)
with the {F µν } given in (1.161), is just an equivalent form for the homogeneous Maxwell equations.
The remaining Maxwell equations can be obtained from the Euler-Lagrange equations for Aν :
Con miras a calcular las ecuaciones de Euler-Lagrange para el Lagrangiano en ec. (1.155), tenemos
F ρσ Fρσ =(∂ ρ Aσ − ∂ σ Aρ )(∂ρ Aσ − ∂σ Aρ )

=∂ ρ Aσ ∂ρ Aσ − ∂ ρ Aσ ∂σ Aρ − ∂ σ Aρ ∂ρ Aσ + ∂ σ Aρ ∂σ Aρ
=g ρα g σβ (∂α Aβ ∂ρ Aσ − ∂α Aβ ∂σ Aρ − ∂β Aα ∂ρ Aσ + ∂β Aα ∂σ Aρ ).
Entonces
∂
F ρσ Fρσ =g ρα g σβ (δαµ δβν ∂ρ Aσ + ∂α Aβ δρµ δσν − δαµ δβν ∂σ Aρ − ∂α Aβ δσµ δρν
∂(∂µ Aν )
− δβµ δαν ∂ρ Aσ − ∂β Aα δρµ δσν + δβµ δαν ∂σ Aρ + ∂β Aα δσµ δρν ).
=g ρµ g σν ∂ρ Aσ + g µα g νβ ∂α Aβ − g ρµ g σν ∂σ Aρ − g να g µβ ∂α Aβ
− g ρν g σµ ∂ρ Aσ − g µα g νβ ∂β Aα + g ρν g σµ ∂σ Aρ + g να g µβ ∂β Aα
=∂ µ Aν + ∂ µ Aν − ∂ ν Aµ − ∂ ν Aµ − ∂ ν Aµ − ∂ ν Aµ + ∂ µ Aν + ∂ µ Aν
=4(∂ µ Aν − ∂ ν Aµ )
∂
F ρσ Fρσ = 4F µν (1.163)
∂(∂µ Aν )
Usando la ec. (1.163), tenemos

∂L ∂L
∂µ − =0
∂(∂µ Aν ) ∂Aν

1 ∂ ∂Aρ
− ∂µ (F ρσ Fρσ ) + J ρ =0
4 ∂(∂µ Aν ) ∂Aν
−∂µ F + J ρ δρν
µν
=0
∂µ F µν = Jν. (1.164)
Como era de esperarse una Acción invariante de Lorentz e invariante gauge local, expresada en
términos del Lagrangiano (1.155), da lugar a la Teorı́a Electromagnética.
Tomando la derivada con respecto a ν en ambos lados tenemos
∂ν ∂µ F µν = ∂ν J ν . (1.165)
De la parte izquierda de ésta ecuación tenemos
∂ν ∂µ F µν = 12 (∂ν ∂µ F µν + ∂ν ∂µ F µν )
= 12 (∂ν ∂µ F µν + ∂µ ∂ν F νµ ) intercambiando ı́ndices mudos
= 12 (∂ν ∂µ F µν + ∂ν ∂µ F νµ ) conmutando derivadas
= 12 (∂ν ∂µ F µν − ∂ν ∂µ F µν ) usando antisimetrı́a de F µν
= 0,
Por consiguiente, la cuadricorriente J µ es conservada:
∂µ J µ = 0 . (1.166)
Again, for ν = 0, we have
∂µ F µ0 = J 0
∂i F i0 = J 0
∂ i0
F = J0
∂xi
∂E i
= J0 , (1.167)
∂xi
and therefore
∇ · E = ρ. (1.168)
while for ν = k we have

∂µ F µk = J k
∂i F ik + ∂0 F 0k = J k
−∂i F ki − ∂0 F k0 = J k
∂(ikj B j ) ∂E k
− i
− = Jk
∂x ∂t
∂B j ∂E k
ijk i − = Jk
∂x ∂t
∂E k
(∇ × B)k − = J k. . (1.169)
∂t
and therefore
∂E
∇×B− = J. (1.170)
∂t
In this way the expression
∂µ F µν = J ν where F µν = ∂ µ Aν − ∂ ν Aµ , (1.171)
is completely equivalent to the full set of Maxwell equations:

∂B
∇ · B = 0, ∇×E+ =0 (1.172)
∂t
∂E
∇ · E = ρ, ∇×B− = J. (1.173)
∂t
1.5.1 Energı́a del campo electromagnético

Necesitamos la expresión para Fµν ,
(
F0i = F0ν = g00 gij F 0j = −F 0i para µ = 0
Fµν = gµρ gνη F ρη ⇒ (1.174)
Fij = Fiν = gik gjl F kl = F ij para µ = i
De la ec. (1.39), se tiene

∂L
Tνµ = (∂ν Aλ ) − δνµ L
∂(∂µ Aλ )
= −F µλ (∂ν Aλ ) − δνµ L (1.175)
La energı́a del campo, corresponde a la componente T00 :
T00 = −F 0λ (∂0 Aλ ) − L
1
= −F 0λ (∂0 Aλ ) + F µν Fµν + J µ Aµ
4
Usando las ecuaciones (??), (??), (1.174)
1
T00 = −F 0λ (∂0 Aλ ) + F µν Fµν + J µ Aµ
4
ν=0 ν=i
0µ 1 z µ0}| { 1 z µi}| {
= −F (∂0 Aµ ) + F Fµ0 + F Fµi +J µ Aµ
4 4
0µ 1 µ0 1
µ0
= −F ∂µ A0 − F Fµ0 + F Fµ0 + F µi Fµi + J µ Aµ . (1.176)
4 4
Tenemos dos partes
µ=0 µ=j
µ0 1 µ0 1 µi i0 1 i0 1 z 0i}| { 1 z ji}| {
−F Fµ0 + F Fµ0 + F Fµi = −F Fi0 + F Fi0 + F F0i + F Fji
4 4 4 4 4
1 i0 1 i0 1
= −F Fi0 + F Fi0 + F Fi0 + F ji Fji
i0
4 4 4
1 i0 1 ji
= − F Fi0 + F Fji . (1.177)
2 4
Además
−F 0µ ∂µ A0 + J µ Aµ = − ∂µ (A0 F 0µ ) + A0 ∂µ F 0µ + J µ Aµ
= − ∂µ (A0 F 0µ ) − A0 ∂µ F µ0 + J µ Aµ
= − ∂µ (A0 F 0µ ) − A0 J 0 + J µ Aµ
= − ∂i (A0 F 0i ) − J · A . (1.178)
Entonces
1 1
T00 = −∂i (A0 F 0i ) − F i0 Fi0 + F ji Fji − J · A
2 4
1 1
= −∂i (A0 F 0i ) + F i0 F i0 + F ji F ji − J · A, suma también sobre i, j
2 4
1 i i 1
= E E + ijk B k ijl B l + ∂i (A0 E i ) − J · A, suma también sobre i, j
2 4
1 2 1
= E + δkl B k B l + ∇ · (A0 E) − J · A
2 2
1 2 1 2
= E + B + ∇ · (A0 E) − J · A (1.179)
2 2
Entonces, en ausencia de corrientes
1 1
H = E2 + B2 + ∇ · (A0 E) . (1.180)
2 2
Similarmente la densidad Lagrangiano puede escribirse como
1 1 2
L = − F µν Fµν = E − B2 (1.181)
4 2
En vista a la ec. (1.176), ya que la densidad Lagrangiana está definida hasta una derivada total,
como ∇ · (A0 E) = ∂µ (A0 F µ0 ), la densidad Hamiltoniana también estará definida hasta una derivada
total. De hecho, el Hamiltoniano es
Z Z
1 3 2 2
H= d x (E + B ) + d3 x ∇ · (A0 E)
2 V V
Z
1
= d3 x (E2 + B2 ), (1.182)
2 V
y corresponde a la expresión conocida para la energı́a del campo electromagnético. Hemos usado el
hecho que en ausencia de corrientes todo lo que entra a un volumén debe salir y por consiguiente las
integrales sobre el volumen de la divergencia de cualquier vector es cero.
Similarmente el momentum total del campo, en ausencia de corrientes, corresponde al vector de
1.6. SCRÖDINGER EQUATION IN PRESENCE OF THE ELECTROMAGNETIC FIELD 33
Pointing:
∂L
Ti0 = ∂i Aν
∂(∂0 Aν )
= − F 0ν ∂i Aν
= − F 0j (∂i Aj − ∂j Ai ) − F 0j ∂j Ai
= − F 0j Fij − F 0j ∂j Ai
= − F 0j F ij − ∂j (F 0j Ai ) + (∂j F 0j )Ai
=E j jik B k + ∂j (E j Ai ) + (J 0 )Ai
= − (E × B)i − ∇ · (Ai E) − ρAi (1.183)
En ausencia de cargas y corrientes
Z Z Z
3
i
P =− d x Ti0 = 3 i
d x (E × B) + d3 x ∇ · (Ai E)
V ZV V
P= d3 x (E × B) . (1.184)
V
1.6 Scrödinger Equation in presence of the electromagnetic

field
Once we have established the set of fields, as in this case ψ, ψ ∗ , and Aµ , we should write the most
general Lagrangian. Therefore
1 X i 1
L(ψ, ψ ∗ , ∂µ ψ, ∂µ ψ ∗ , Aµ ) = (Di ψ)∗ Di ψ − [ψ ∗ D0 ψ − (D0 ψ)∗ ψ] + ψ ∗ V (x)ψ − F µν Fµν − J ν Aν .
2m i 2 4
(1.185)
If we further assume that all interactions are obtained from the covariant derivative, then we
need only consider the free Lagrangian of each field, but with the normal derivative replaced by the
covariant one:
1 X i 1
L(ψ, ψ ∗ , ∂µ ψ, ∂µ ψ ∗ , Aµ ) = (Di ψ)∗ Di ψ − [ψ ∗ D0 ψ − (D0 ψ)∗ ψ] − F µν Fµν . (1.186)
2m i 2 4
The expansion of the Lagrangian in terms of the field ψ, ψ ∗ , and Aµ is
1 X i 1
L= (∂i ψ + iqAi ψ)∗ (∂i ψ + iqAi ψ) − [ψ ∗ (∂0 ψ + iqA0 ψ) − (∂0 ψ + iqA0 ψ)∗ ψ] − F µν Fµν
2m i 2 4
1 X i 1
= (∂i ψ ∗ − iqAi ψ ∗ ) (∂i ψ + iqAi ψ) − [ψ ∗ (∂0 ψ + iqA0 ψ) − (∂0 ψ ∗ − iqA0 ψ ∗ ) ψ] − F µν Fµν
2m i 2 4
1 X
= ∂i ψ ∗ ∂i ψ − iqψ ∗ Ai ∂i ψ + iq∂i ψ ∗ Ai ψ + q 2 Ai Ai ψ ∗ ψ
2m i
i ∗ 1
− [ψ ∂0 ψ + iqψ ∗ A0 ψ − (∂0 ψ ∗ )ψ + iqA0 ψ ∗ ψ] − F µν Fµν
2 4
1 X
= ∂i ψ ∗ ∂i ψ − iqψ ∗ Ai ∂i ψ + iq∂i ψ ∗ Ai ψ + q 2 Ai Ai ψ ∗ ψ
2m i
i ∗ 1
− [ψ ∂0 ψ − (∂0 ψ ∗ )ψ + 2iqψ ∗ A0 ψ] − F µν Fµν . (1.187)
2 4
By using the sum convention upon repeated indices we have
1
L=− ∂i ψ ∗ ∂ i ψ − iqψ ∗ Ai ∂i ψ + iq∂i ψ ∗ Ai ψ + q 2 Ai Ai ψ ∗ ψ
2m
i 1
− [ψ ∗ ∂0 ψ − (∂0 ψ ∗ )ψ + 2iqψ ∗ A0 ψ] − F µν Fµν . (1.188)
2 4
From this we can obtain the Euler-Lagrange equation for each field.
1.6.1 Euler-Lagrange equation for ψ ∗

In particular for ψ ∗ we have

∂L ∂L
∂µ ∗
− =0
∂(∂µ ψ ) ∂ψ ∗

∂L ∂L ∂L
∂0 ∗
+ ∂i ∗
− =0
∂(∂0 ψ ) ∂(∂i ψ ) ∂ψ ∗

i 1 i i
1 i 2 i
i
∂0 ψ − ∂i ∂ ψ + iqA ψ − − −iqAi ∂ ψ + q Ai A ψ − (∂0 ψ + 2iqA0 ψ) =0
2 2m 2m 2
1
i∂0 ψ − qA0 ψ − ∂i ∂ i ψ + iqAi ψ + iqAi ∂ i ψ + iqAi ψ =0
2m
1
i(∂0 + iqA0 )ψ − (∂i + iqAi )(∂ i ψ + iqAi ψ) =0
2m
1 X
iD0 ψ + Di Di ψ =0 , (1.189)
2m i
If we define
D ≡ ∇ − iqA . (1.190)
we have in components:
D i = ∂i − iqAi
D i = ∂i + iqAi . (1.191)
Then we have the new wave equation:
1
iD0 ψ = − D · Dψ
2m
1 2
iD0 ψ = − D ψ, (1.192)
2m
que corresponde a la ecuación de Scrödinger con la derivada normal reemplazada por la derivada
covariante.
Expandiendo esta ecuación tenemos

∂ 1 X
i + iqA0 ψ =− (∂i + iqAi )2 ψ
∂t 2m i

∂ 1 X
i − qA0 ψ =− (∂i − iqAi )2 ψ
∂t 2m i

∂ 1
i − qφ ψ =− (∇ − iqA)2 ψ
∂t 2m
−i2
Hb − qφ ψ =− (∇ − iqA)2 ψ
2m
1
= (i∇ + qA)2 ψ
2m
1
= (−i∇ − qA)2 ψ
2m
1
= (bp − qA)2 ψ . (1.193)
2m
In this way, the Scrödinger equation in presence of the electromagnetic field, can be obtained from
the original Scrödinger equation but with the minimum substitution:
b →H
H b − qφ b →b
p p − qA . (1.194)
De la ecuación (1.193) podemos obtener la ecuación de Schödinger en presencia de un campo

electromagnético

∂ 1 2
i ψ= (−i∇ − qA) + qA0 ψ . (1.195)
∂t 2m
Para que la mecánica cuántica sea consistente con las ecuaciones de Maxwell es necesario que las
transformaciones gauge (1.86) de los potenciales de Maxwell estén acompañados por una transfor-
mación de la función de onda, ψ → ψ 0 , donde ψ 0 satisface la ecuación
0 1 02 0
iD0 ψ 0 = − D ψ
2m
∂ 0 1
i ψ = (−i∇ − qA ) + qA 0 ψ 0 .
0 2 0
(1.196)
∂t 2m
Como la forma de la ecuación (1.196) es exactamente la misma que la forma de (1.195) entonces
ambas describen la misma fı́sica. Se dice que la ec. (1.195) es covariante gauge, lo que significa que
mantiene la misma forma bajo una transformación gauge.
• Ejemplo:
Demuestre que la ec. (1.196) es covariante:
Como
ψ → ψ 0 = eiθ(x) ψ (1.197)
Entonces
D 0 ψ 0 = [(∇ − iqA) − i∇θ] eiθ(x) ψ

= i(∇θ)eiθ(x) ψ + eiθ(x) ∇ψ − iqAeiθ(x) ψ − i(∇θ)eiθ(x) ψ
= eiθ(x) (∇ − iqA)ψ
= eiθ(x) (Dψ) (1.198)
y
2
D 0 ψ 0 = D 0 (D 0 ψ 0 )
= [(∇ − iqA) − i∇θ] eiθ(x) (Dψ)
= i(∇θ)eiθ(x) (Dψ) + eiθ(x) ∇(Dψ) − iqAeiθ(x) (Dψ) − i∇θeiθ(x) (Dψ)
= eiθ(x) (∇ − iqA)(Dψ)
= eiθ(x) (D 2 ψ) (1.199)
De la misma manera
0
D0 ψ 0 = eiθ(x) (D0 ψ) (1.200)
De modo que
µ
Dµ ψ → D0 ψ 0 = eiθ(x) (Dµ ψ) (1.201)
y la derivada covariante del campo transforma como el campo. Tenemos entonces que
0 1 02 0
iD0 ψ 0 = − D ψ
2m
1 iθ(x) 2
ieiθ(x) D0 ψ = − e D ψ
2m
1 2
iD0 ψ = − D ψ (1.202)
2m
En resumen, para
Dµ = ∂ µ + iqAµ (1.203)
y reemplazando θ → qθ tenemos
Aµ → Aµ0 = Aµ − ∂ µ θ(x)
ψ → ψ 0 = eiqθ(x) ψ
µ
Dµ ψ → D0 ψ 0 = eiqθ(x) (Dµ ψ) . (1.204)
En esta convención q corresponde al generador de la transformación y θ al parámetro de la transfor-
mación.
1.6.2 Euler-Lagrange equation for Aµ

Para el campo Aν tenemos

∂L ∂L
∂µ − = 0. (1.205)
∂(∂µ Aν ) ∂Aν
Usando el Lagrangiano en (1.187) y el resultado de (1.164) tenemos
∂L
∂µ (−F µν ) − =0
∂Aν
que da lugar a dos conjuntos de ecuaciones, una para Ai
∂L
∂µ (F µj ) + =0
∂Aj
1
∂µ (F µj ) − [iq(∂ i ψ ∗ )ψ − iqψ ∗ (∂ i ψ) + 2q 2 Ai ψ ∗ ψ] =0
2m
iq
∂µ (F µj ) − [(∂ i ψ ∗ )ψ − ψ ∗ (∂ i ψ) − 2iqψ ∗ ψAi ] =0
2m
iq
∂µ (F µj ) − [(∂ i ψ ∗ )ψ − iqψ ∗ ψAi − ψ ∗ (∂ i ψ) − iqψ ∗ ψAi ] =0
2m
iq
∂µ (F µj ) − {[(∂ i − iqAi )ψ ∗ ]ψ − ψ ∗ (∂ i + iqAi )ψ} =0
2m
iq
∂µ (F µj ) − [(Di ψ)∗ ψ − ψ ∗ Di ψ] =0 , (1.206)
2m
y otra para A0
∂L
∂µ (F µ0 ) + =0
∂A0
∂µ (F µ0 ) + qψ ∗ ψ =0
(1.207)
Entonces
∂µ F µν = j ν
con
(
−qψ ∗ ψ ν=0
j ν = iq (1.208)
2m
[(Di ψ)∗ ψ − ψ ∗ Di ψ] ν = i
Que incluye el término corriente para una partı́cula cargada y es diferente de la corriente de proba-
bilidad en ec. (1.57). En otras palabras es la carga eléctrica la que se converva localmente.
1.6.3 Conserved currents

The 4-current can be obtained directly from the Noether’s Theorem:
∂L ∂L
Jµ = δψ + δψ ∗
∂µ ψ ∂µ ψ ∗
(
∂L
∂0 ψ
δψ + δψ ∗ ∂∂L
0ψ
∗ µ=0
= ∂L ∗ ∂L
. (1.209)
∂i ψ
δψ + δψ ∂i ψ∗ µ=i
i i
J 0 = − ψ ∗ (iqθ)ψ − iqθψ ∗ ψ
2 2
∗
=qθψ ψ , (1.210)
1
Ji = [(∂i − iqAi ) ψ ∗ iqθψ − iqθψ ∗ (∂i + iqAi ) ψ]
2m
iqθ
Ji = [(Di ψ)∗ ψ − ψ ∗ (Di ψ)] . (1.211)
2m
When θ is fixed to 1 as in ec. (1.57) to define the probability, we get eq. (1.465).
It is worth to notice that for T00 , and Ti0 we should obtain
b =i ∂ − qφ
H b = − i∇ − qA .
p (1.212)
∂t
1.7 Gauge Transformation Group

• Ejemplo:
Muestre que los campos electromagnéticos son invariantes bajo las siguientes transformaciones
∂χ
A → A0 = A + ∇χ φ → φ0 = φ − (1.213)
∂t
Ya que
∂ ∂A ∂
E → E0 = −∇φ + ∇χ − − ∇χ = E (1.214)
∂t ∂t ∂t
B → B0 = ∇ × A + ∇ × ∇χ = B (1.215)
| {z }
=0
Esto implica que diferentes observadores en diferentes puntos del espacio, usando diferentes
calibraciones para sus medidas, obtienen los mismos campos. Las ecs. (1.213), corresponden a
transformaciones gauge locales
En notación de cuadrivectores
µ
Aµ → A0 =Aµ − ∂ µ χ (1.216)
Sea U un elemento del Grupo de Transformaciones U (1):
U = eiθ(x) ∈ U (1) (1.217)
El Grupo está definido por el conjunto infinito de elementos Ui = eiθ(xi ) . Entonces
• Producto de Grupo
U1 · U2 = ei[θ(x1 )+θ(x2 )] ≡ eiθ(x3 ) ∈ U (1)
• Identidad:
θ(x) = 0 tal que UI = 1
• Inverso
θ(−x) = −θ(x) tal que U −1 = e−iθ(x)
Note que si
µ i
Aµ → A0 = U Aµ U −1 + (∂ µ U )U −1 (1.218)
q
1.8. PROCA EQUATION 41
y si θ es suficientemente pequeño
U = eiθ(x) ≈ 1 + iθ(x) + O(θ2 ) U −1 = e−iθ(x) ≈ 1 − iθ(x) + O(θ2 ) (1.219)
Entonces
i
Aµ0 =[1 + iθ(x) + O(θ2 )]Aµ [1 − iθ(x) + O(θ2 )] + (i∂ µ θ(x))[1 + iθ(x) + O(θ2 )][1 − iθ(x) + O(θ2 )]
q
1
=Aµ − ∂ µ θ(x) + O(θ2 ) (1.220)
q
which is just the eq. (1.86)
1.8 Proca Equation

Consideraremos ahora el efecto de adicionar un término de masa a la teorı́a de Maxwell. Los campos
vectoriales masivos juegan un papel importante en fı́sica. Campos como W µ , Z µ que median las
interacciones débiles son ejemplos de campos de este tipo. Las implicaciones de una masa finita para
el fotón pueden inferirse de un conjunto de postulados que hacen de las ecuaciones de Proca la única
generalización posible de las ecuaciones de Maxwell [?].
Teniendo en cuenta sólo el término de masa en la ec. (1.155)
1 1
L = − F µν Fµν + m2 Aµ Aµ − J µ Aµ . (1.221)
4 2
Usando las ecuaciones de Euler-Lagrange, tenemos

1 ∂ ρη ∂ 1 2 ρ ρ
− ∂µ F Fρη − m A Aρ − J Aρ = 0
4 ∂(∂µ Aν ) ∂Aν 2
∂µ F µν + m2 Aν = J ν . (1.222)
Tomando la cuadridivergencia a ambos lados de la ecuación y usando la ec. (??), tenemos
∂ν ∂µ ∂ µ Aν − ∂ν ∂ ν ∂µ Aµ + m2 ∂ν Aν = ∂ν J ν
∂ν ∂µ ∂ µ Aν − ∂µ ∂ µ ∂ν Aν + m2 ∂ν Aν = ∂ν J ν
m2 ∂ν Aν = ∂ν J ν (1.223)
De este modo, en ausencia de corrientes, la ecuaciones de Proca dan lugar a la condición de Lorentz.
De otro lado, si asumimos que la corriente se conserva, la condición de Lorentz también aparece. Por
consiguiente, si la masa de campo vectorial es diferente de cero, la condición de Lorentz, ec. (??),
emerge como una restricción adicional que debe ser siempre tomada en cuenta. De este modo la
libertad gauge de las ecuaciones de Maxwell se pierde completamente en la ecuaciones de Proca, que
sin perdida de generalidad se pueden reescribir, usando ∂µ Aµ = 0 y las ec. (1.222), como:
∂µ F µν + m2 Aν = J ν
∂µ ∂ µ Aν − ∂µ ∂ ν Aµ + m2 Aν = J ν
( + m2 )Aν =J ν (1.224)
donde esta definido en la ec. (1.123). En ausencia de corrientes, cada una de las componentes del
campo vectorial satisface la ecuación de Klein-Gordon (??). Por consiguiente m corresponde a la
masa del campo vectorial Aµ .
Aplicando la condición de Lorentz a la ec. (1.221), obtenemos el Lagrangiano de la Ecuación de
Proca (1.224)
1 1
L = − F µν Fµν + m2 Aµ Aµ − J µ Aµ
4 2
1 µ ν 1
= − (∂ A ∂µ Aν + ∂ ν Aµ ∂ν Aµ − ∂ µ Aν ∂ν Aµ − ∂ ν Aµ ∂µ Aν ) + m2 Aµ Aµ − J µ Aµ
4 2
1 µ ν 1 2 ν
= ∂ A ∂µ Aν − m A Aν + J ν Aν , (1.225)
2 2
donde hemos reabsorbido un signo global que no afecta las ecuaciones de movimiento. El primer
término que incluye sólo derivadas de los campos es llamado término cinético y dependen sólo del
espı́n de las partı́culas. El término cuadrático en los campos corresponde al término de masa, y el
último corresponde a la interacción del campo con una corriente. Cuando un Lagrangiano contiene
sólo términos cinéticos y de masa diremos que el campo que da lugar al Lagrangiano es libre de
interacciones, o simplemente que es un campo libre. Las otras partes del Lagrangiano serán llamadas
Lagrangiano de Interacción. De este modo podemos reescribir el Lagrangiano (1.225) como
L = Lfree + Lint ,
donde,
1 1
Lfree = ∂ µ Aν ∂µ Aν − m2 Aν Aν
2 2
Lint = J ν Aν . (1.226)
Debido a que la teorı́a masiva ya no es invariante gauge, la condición de Lorentz aparece au-
tomáticamente como la única restricción apropiada sobre el campo vectorial.
Una vez se toma en cuenta la condición de Lorentz el campo masivo libre puede expandirse en
ondas planas con tres grados de libertad independientes de polarización. Dos de estos corresponden
1.9. KLEIN-GORDON EQUATION 43
a los dos estados transversos que aparecen en las ondas electromagnéticas (A1 , A2 ), y el tercero (A3 )
corresponde a un estado longitudinal en la dirección del momento de la partı́cula [?].
Aunque hemos hecho el análisis de la ecuación de Proca permitiendo un término de masa para
el fotón, las implicaciones experimentales de una teorı́a de este tipo dan lugar a restricciones muy
fuertes sobre la masa del fotón[?]. El lı́mite actual sobre la masa del fotón es m < 6 × 10−17 eV
(1.1 × 10−52 Kg) [?]. Debido al principio gauge local, desde el punto teórico se espera que la masa del
fotón sea exactamente cero. En general, los campos vectoriales puede ser generados a partir de otras
cargas no electromagnéticas y pueden ser masivos. El reto durante varias décadas fue entender como
las masa de los campos vectoriales de la interacción débil podrı́a hacerse compatible con el principio
gauge local.
1.9 Klein-Gordon Equation

De la componente escalar de la ecuación de Proca, (1.225), obtenemos la ecuación de Klein–Gordon
para un campo escalar real φ = A0 ,
1 1
L = ∂ µ φ∂µ φ − m2 φ2 + ρφ (1.227)
2 2
Donde ρ es la densidad de carga que actua como fuente del campo φ. El Lagrangiano más general
posible que cuya acción sea invariante de Lorentz, para el campo escalar real φ(x) es
1 1
L = ∂ µ φ∂µ φ − m2 φ2 − V (φ) , (1.228)
2 2
donde V (φ) es alguna función de φ con operadores de dimensión menor o igual a 4. Para demostrar
la invarianza de Lorentz.
The kinetic part of Klein-Gordon Lagrangian transforms as
∂µ φ(x)∂ µ φ(x) →g µν ∂ 0 µ φ0 (x)∂ 0 ν φ0 (x)
h i −1 σ
µν −1 ρ −1 −1
=g Λ µ
∂ρ φ Λ x Λ ν
∂σ φ Λ x
µν −1 σ
−1 ρ

= Λ µ
g Λ ∂ φ Λ x ∂σ φ Λ−1 x
ν ρ
−1

=g ρσ ∂ρ φ Λ−1 x ∂σ φ Λ−1 x

=∂ρ φ Λ−1 x ∂ ρ φ Λ−1 x . (1.229)
Since φ(x) → φ0 (x) = φ (Λ−1 x), under a Lorentz transformation the full Lagrangian transform as
L(x) → L0 (x) = L(Λ−1 x) (1.230)
Since the Action involves the integration over all the points, it is invariant under the Lorentz trans-
formation.
• Ejercicio Demuestre que un término
φ(x)aµ ∂µ φ(x) . (1.231)
con aµ constante, no es invariante de Lorentz, de modo que en efecto, el Lagrangiano en de

Klein-Gordon en (1.228) es el más general posible (hasta términos de intercción en φ(x).)
El campo φ puede pensarse como proveniente de una fuente de la misma manera como el campo
electromagnético surge de partı́culas cargadas. Como en el caso del electromagnetismo, en esta
sección podemos considerar los campos sin preocuparnos de las fuentes. En tal caso tendremos una
teorı́a en la cual el campo escalar juega el papel de partı́cula mediadora de la interacción.
Si el campo escalar se generaliza para que pueda tener otros números cuánticos, como carga
eléctrica, entonces estos pueden ser las fuentes de las respectivas cargas y corrientes en la ecuaciones
para campos vectoriales. Esto se estudiará en la sección ??. En tal caso podrı́amos tener por ejemplo
“átomos” formados de partı́culas escalares que se excitan emitiendo fotones.
La ecuaciones de Euler-Lagrange para V (φ) = −ρφ dan lugar a:
( + m2 )φ = ρ .
2
∂ 2 2
− ∇ + m φ = ρ. (1.232)
∂t2
Con el cuadrivector (1.125) podemos construir la siguiente ecuación
p̂µ p̂µ φ = m2 φ
i∂µ i∂ µ φ = m2 φ
−∂µ ∂ µ φ = m2 φ
2
∂ 2 2
− ∇ + m φ = 0. (1.233)
∂t2
Que corresponde a la ecuación de Klein-Gordon (??). Una expresión escrita en términos de productos
escalares de Lorentz se dice que esta en forma covariante.
De acuerdo a la ec. (1.226), tenemos
1 1
Lfree = ∂µ φ∂ µ φ − m2 φ2
2 2
Lint = ρφ (1.234)
∂L
Tνµ = ∂ν φ − δνµ L , (1.235)
∂(∂µ φ)
T00 =∂ 0 φ∂0 φ − L
=π(x)∂0 φ(x) − L
1 1 1
=∂ 0 φ∂0 φ − ∂0 φ∂ 0 φ − ∂i φ∂ i φ + m2 φ2
2 2 ! 2
1 X 1
= ∂0 φ∂ 0 φ + ∂i φ∂i φ + m2 φ2
2 i
2
" #
2
1 ∂φ 1
H= + (∇φ)2 + m2 φ2 . (1.236)
2 ∂t 2
donde
∂L
π(x) =
∂ (∂φ/∂t)
∂φ
= . (1.237)
∂t
La densidad de momentum es
∂L
Ti0 = ∂i φ
∂(∂0 φ)
Ti0 =∂ 0 φ∂i φ
∂φ
T0 = ∇φ . (1.238)
∂t
1.9.1 Complex scalars

En la sección anterior se trabajo con un campo escalar real que sólo podrı́a describir un pion neutro.
Para describir piones cargados debemos construir un campo escalar complejo. En mecánica cuántica
la función de onda compleja puede describir parcialmente a un electrón cargado. Sin embargo la
función de onda del electrón también debe ser generalizada para poder dar cuenta del espı́n. Esto
corresponde al función de onda de la ecuación de Dirac en la sección ??.
De hecho, algunas consecuencias fı́sicas interesantes surgen si consideramos un sistema de dos
campos escalares reales, φ1 y φ2 , que tengan la misma masa m. Entonces
1 1 1 1
L = [∂ µ φ1 ∂µ φ1 − m2 φ21 ] + [∂ µ φ2 ∂µ φ2 − m2 φ22 ] (1.239)
2 2 2 2
Si definimos
φ1 + iφ2
φ= √ then (1.240)
2
φ 1 − iφ2
φ∗ = √ , and (1.241)
2
√
2φ =(φ1 + iφ2 )
√ ∗
2φ =(φ1 − iφ2 ). Therefore
√ ∗
2(φ + φ ) =2φ1
√
2(φ − φ∗ ) =2iφ2 . Then
∗
φ+φ
φ1 = √ (1.242)
2
φ − φ∗
φ2 = √ . (1.243)
2i
Reemplazando la ecuaciones (1.242) y (1.243) en la ec. (1.239), tenemos
1 1
L = [∂ µ (φ + φ∗ )∂µ (φ + φ∗ ) − m2 (φ + φ∗ )2 ]
4 2
1 1
+ i2 [∂ µ (φ − φ∗ )∂µ (φ − φ∗ ) − m2 (φ − φ∗ )2 ]
4 2
1 µ
= [∂ φ∂µ φ + ∂ µ φ∗ ∂µ φ∗ + 2∂ µ φ∗ ∂µ φ − m2 (φ2 + φ∗ 2 ) + 2φ∗ φ]
4
1
− [∂ µ φ∂µ φ + ∂ µ φ∗ ∂µ φ∗ − 2∂ µ φ∗ ∂µ φ − m2 (φ2 + φ∗ 2 ) − 2φ∗ φ]
4
1 µ ∗
= [4∂ φ ∂µ φ − 4m2 φ∗ φ]
4
L =∂ µ φ∗ ∂µ φ − m2 φ∗ φ (1.244)
De la ec. (1.18) de la sección ??,

De las ecuaciones de Euler-Lagrange para φ∗ , usando el Lagrangiano en ec. (1.244)

∂L ∂L
∂µ ∗
− ∗ =0
∂(∂µ φ ) ∂φ
∂µ ∂ φ + m2 φ = 0
µ
( + m2 )φ = 0, (1.245)
y de la ecuaciones de Euler-Lagrange para φ,
( + m2 )φ∗ = 0. (1.246)
De este modo tanto φ, como φ∗ , satisfacen la ecuación de Klein-Gordon. Cada campo además
corresponde a una partı́cula de masa m como en el caso de φ1 y φ2
Estamos ahora interesado en las simetrı́as internas del Lagrangiano. Entonces la corriente con-
servada puede definida en la sección ??, eq. (1.20)
∂L ∂L
Jµ = δφ + δφ∗
∂(∂µ φ) ∂(∂µ φ∗ )
J µ =∂ µ φ∗ δφ + δφ∗ ∂ µ φ. (1.247)
Además de la invarianza de Lorentz, el Lagrangiano en ec, (1.244) también es invariante bajo el

grupo de transformaciones U(1) definido en las sección ??, pero con una fase constante
U = eiθ ≈ 1 + iθ.
Entonces
U
φ −→ φ0 = eiθ φ ≈ (1 + iθ)φ
= φ + iθφ. (1.248)
Entonces,
δφ = iθφ (1.249)
δφ∗ = −iθφ∗ . (1.250)
Reemplazando en ec. (1.247)

J µ ∝ −iθ(φ∂ µ φ∗ − φ∗ ∂ µ φ), (1.251)
y
∂φ∗ ∂φ
0
ρ = J ∝ −iθ(φ − φ∗ ). (1.252)
∂t ∂t
Definimos J µ como
J µ = i(φ∗ ∂ µ φ − φ∂ µ φ∗ ), (1.253)
Como ρ puede ser negativo no puede interpretarse como una probalidad, como se hizo con la función
de onda de la ecuación de Scrödinger. Esto presentó un obstaculo en la interpretación inicial de la
ecuación de Klein-Gordon. Sin embargo una vez se cuantiza el campo escalar la probabilidad de los
estados cuánticos queda bien definida [?].
1.10 Lorentz transformation of the fields

Note again, that a term like
φ∗ (x)aµ ∂µ φ(x) , (1.254)
does not left the Action invariant. To have a proper formulation of the quantum mechanics through
the general equation
∂
i ψ = Ĥψ , (1.255)
∂t
with some, to be determined, relativistic Hamiltonian operator H, b we should be able to build a

Lagrangian with temporal derivatives of order one. Therefore, the Lorentz invariant requires all the
derivatives of order one.
Consider spinor fields, which transforms as
ψa (x) → ψa0 (x) = Sab (Λ)ψb (Λ−1 x) , (1.256)
where S(Λ) is some spinorial representation of the Lorentz Group. We will check in next section if
a Action with a term like
ψa∗ (x)aµab ∂µ ψb (x) (1.257)
could be invariant under Lorentz transformations, for some internal representation of the Lorentz
Group.
In summary we have the following Lorentz’s transformation properties for the fields
φ(x) → φ0 (x0 ) =φ(x) Scalar field,

µ
Aµ (x) → A0 =Λµ ν Aν (Λ−1 x) Vector field,
ψ(x) → ψ 0 (x) =S(Λ)ψ(Λ−1 x) Spinor field. (1.258)
1.11 Dirac’s Action

The Scrodinger equation can be written as
∂
i ψ = ĤS ψ , (1.259)
∂t
1.11. DIRAC’S ACTION 49
where
ĤS = (1.260)
In order to have a well defined probabilty in relativistic quantum mechanics it is necessary that
Lagrangian be linear in the time derivative, in order to obtain the general Sccödinger equation:
∂
i ψ = Ĥψ , (1.261)
∂t
like the Scrödinger Lagrangian. However, this automatically imply that the Lagrangian will be also
linear in the spacial derivatives. A pure scalar field cannot involve a Lorentz invariant term of only
first derivatives (see eq. (1.254)). Therefore the proposed field must have some internal structure
associated with some representation of the Lorentz Group. Therefore we build the Lagrangian for a
field of several components
 
ψ1
 ψ2 
 
ψ =  ..  (1.262)
 . 
ψn
1.11.1 Lorentz transformation

If the field is to describe the electron. it must have spin and in this way it must transform under
some spin representation of the Lorentz Group

ψ(x) → ψ 0 (x) = S(Λ)ψ Λ−1 x . (1.263)
One possible invariant could be the term ψ † (x)ψ(x). However, under a Lorentz transformation we
should have ψ † S † Sψ. As we cannot assume that S(Λ) is unitary, the solution is to define the adjoint
spinor
ψ = ψ†b . (1.264)
which transforms as
0 †
ψ(x) → ψ (x) = ψ 0 (x)b = ψ † Λ−1 x S † (Λ)b , (1.265)
and,
0
ψ(x)ψ(x) → ψ (x)ψ 0 (x) = ψ † Λ−1 x S † (Λ)bS(Λ)ψ Λ−1 x (1.266)
The condition that must be fulfilled for Lorentz invariance of the Action is
S † (Λ)bS(Λ) =b , (1.267)
and therefore,
0
ψ(x)ψ(x) → ψ (x)ψ 0 (x) = ψ Λ−1 x ψ Λ−1 x , (1.268)
and:
0
ψ(x) → ψ (x) = ψ † Λ−1 x bS −1 (Λ)

= ψ Λ−1 x S −1 (Λ) . (1.269)
A Action with a Lagrangian term linear in the derivatives, could be Lorentz invariant if, taking
into account:
−1 µ
ρ
ψ(x)γ µ ∂µ ψ(x) → ψ 0 (x)γ µ ∂µ ψ 0 (x) = ψ a Λ−1 x Sab (Λ)γbc Λ−1 µ ∂ρ Scd (Λ)ψd Λ−1 x
ρ
= ψψ Λ−1 x Λ−1 µ S −1 (Λ)γ µ S(Λ) ∂ρ ψ Λ−1 x
= ψ(x)γ µ ∂µ ψ(x) , (1.270)
if the following condition is satisfied:
S −1 (Λ)γ µ S(Λ) = Λµ σ γ σ . (1.271)
the most general Lagrangian for this field is
L = iψγ µ ∂µ ψ − mψψ , (1.272)
Where the coefficients have been already fixed by convenience. Since the Action is real, it is convenient
to rewrite this as
L = iψγ µ ∂µ ψ − mψψ
1
= − ∂µ iψγ µ ψ + iψγ µ ∂µ ψ − mψψ
2
i i
= − (∂µ ψ)γ µ ψ − ψγ µ ∂µ ψ + iψγ µ ∂µ ψ − mψψ
2 2
i µ i
= ψγ ∂µ ψ − (∂µ ψ)γ µ ψ − mψψ . (1.273)
2 2
Para que este nuevo Lagrangiano sea real se requiere que,
b† = b
b2 = I
bγµ† b = γµ (1.274)
ya que

† i † † i
L = ψ γµ b∂µ ψ − ∂µ ψ γµ bψ − mψ † bψ
† †
2 2

i † 2 † i
= † 2 †
ψ b γµ b∂µ ψ − ∂µ ψ b γµ bψ − mψ † bψ
2 2

i † i †
= ψ̄bγµ b∂µ ψ − ∂µ ψ̄bγµ bψ − mψ̄ψ
2 2

i i
= ψ̄γµ ∂µ ψ − ∂µ ψ̄γµ ψ − mψ̄ψ
2 2
1.11.2 Corriente conservada y Lagrangiano de Dirac

De la ec. (??)
" #
∂L ∂L
J0 = δψ + δψ
∂ (∂0 ψ) ∂ ∂0 ψ
= iψγ 0 δψ (1.275)
El Lagrangiano es invariante bajo transformaciones de fase globales, U (1)
ψ → ψ 0 = e−iα ψ ≈ ψ − iαψ, (1.276)
de modo que
δψ = −iαψ. (1.277)
Por consiguiente
J 0 = αψγ 0 ψ (1.278)
Para que J 0 pueda interpretarse como una densidad de probabilidad, se debe cumplir
bγ 0 = I (1.279)
La densidad de corriente es
J 0 ∝ ψ†ψ . (1.280)
Que podemos interpretar como una densidad de probabilidad.
De la ec. (1.279), ya que la inversa de es única:
b = γ0 . (1.281)
ψ se define como la adjunta de ψ:

ψ = ψ†γ 0 . (1.282)
It is convenient at this point to summarize the properties for γ 0 :
† 2
γ 0 =γ 0 γ 0 =1 γ 0 γ µ† γ 0 =γ µ
S † (Λ)γ 0 S(Λ) =γ 0 . (1.283)
En general
" #
∂L ∂L
Jµ ∝ δψ + δ ψ̄
∂ (∂µ ψ) ∂ ∂µ ψ̄
∝ iψ̄γ µ (−iαψ)
∝ iψ̄γ µ (−iαψ)
= ψ̄γ µ ψ (1.284)
y
J µ = ψ † bγ µ ψ . (1.285)
1.11.3 Tensor momento-energı́a
∂L ∂L
T00 = ∂0 ψ + ∂0 ψ̄ −L
∂ (∂0 ψ) ∂ ∂0 ψ̄
= iψ̄γ 0 ∂0 ψ − L
= −iψ̄γ i ∂i ψ + mψ̄ψ,
= ψ̄(γ · p + m)ψ,
= ψ † γ 0 (γ · p + m)ψ,
= ψ † Ĥψ, (1.286)
donde
Ĥ = γ 0 (γ · p + m) (1.287)
la ecuación de Scröndinger de validez general es entonces:
∂
i ψ = Ĥψ (1.288)
∂t
y, como en mecánica clásica usual Z
hĤi = ψ † Ĥψ d3 x. (1.289)
Además
∂L ∂L
Ti0 = ∂i ψ + ∂i ψ̄
∂ (∂0 ψ) ∂ ∂0 ψ̄
= iψ̄γ 0 ∂i ψ
= −ψ † (−i∂i )ψ (1.290)
de modo que Z
hp̂i = ψ † p̂ψ d3 x (1.291)
1.11.4 Ecuaciones de Euler-Lagrange

Queremos que el Lagrangiano de lugar a la ecuación de Scröndinger de validez general
∂
i ψ = Ĥψ (1.292)
∂t
con el Hamiltoniano dado en la ec. (1.289), que corresponde a un Lagrangiano de sólo derivadas de
primer orden y covariante, en lugar del Hamiltoniano para el caso no relativista.
De hecho, aplicando las ecuaciones de Euler-Lagrange para el campo ψ̄ al Lagrangiano en ec. (??)
,tenemos
" #
∂L ∂L
∂µ − =0
∂ ∂µ ψ̄ ∂ ψ̄
∂L
=0
∂ ψ̄
iγ µ ∂µ ψ − mψ = 0. (1.293)
Expandiendo
iγ 0 ∂0 ψ + iγ i ∂i ψ − mψ = 0
iγ 0 ∂0 ψ − γ · (−i∇)ψ − mψ = 0,
iγ 0 ∂0 ψ = (γ · p̂ + m)ψ,
de donde
2 ∂
iγ 0 ψ = γ 0 (γ · p + m)ψ. (1.294)
∂t
tenemos que
2
γ0 = 1. (1.295)
De la ec. (1.287)
Ĥ = γ 0 (γ · p + m), (1.296)
A este punto, sólo nos queda por determinar los parámetros γ µ .
La ec. (1.292) puede escribirse como

∂
i − Ĥ ψ = 0. (1.297)
∂t
El campo ψ también debe satisfacer la ecuación de Klein-Gordon. Podemos derivar dicha ecuación
aplicando el operador
∂
−i − Ĥ
∂t
De modo que, teniendo en cuenta que ∂ Ĥ/∂t = 0,

∂ ∂
−i − Ĥ i − Ĥ ψ = 0
∂t ∂t

∂ ∂ψ
−i − Ĥ i − Ĥψ = 0
∂t ∂t
!
∂ 2ψ ∂ Ĥ ∂ψ ∂ψ
+i ψ + iĤ − iĤ + Ĥ 2 ψ = 0
∂t2 ∂t ∂t ∂t
2
∂ 2
+ Ĥ ψ = 0. (1.298)
∂t2
De la ec. (1.296), y usando la condición en ec. (1.295), tenemos
Ĥ 2 = (γ0 γ · p + γ0 m)(γ0 γ · p + γ0 m)
= (γ0 γ · p)(γ0 γ · p) + mγ0 γ · pγ0 + mγ02 γ · p + m2 (1.299)
Sea
β = γ0
αi = βγ i
γ i = βαi (1.300)
Ĥ 2 = (α · p)(α · p) + mα · pβ + mβα · p + m2
= (α · p)(α · p) + m(αβ + βα) · p + m2 (1.301)
Sea A una matriz y θ en un escalar. Entonces tenemos la identidad
X 2 2
X
(A · θ)2 = Ai θi + Ai , Aj θi θj (1.302)
i i<j
• Demostración
XX
[(A · θ)]αβ = Aiαγ θi Ajγβ θj
ij γ
X X
= i j
θθ Aiαγ Ajγβ
ij γ
XX
= θi θj Aiαγ Ajγβ
γ ij
!
X X 2
X X
= θi Aiαγ Aiγβ + θi θj Aiαγ Ajγβ + θi θj Aiαγ Ajγβ
γ i i<j i>j
!
X X X X
i2
= θ Aiαγ Aiγβ + θθi j
Aiαγ Ajγβ + j i
θθ Ajαγ Aiγβ
γ i i<j j>i
" #
X X X
i2
= θ Aiαγ Aiγβ + θi θj Aiαγ Ajγβ + Ajαγ Aiγβ
γ i i<j
" #
X 2 X
= θi Ai Ai αβ + θi θj Ai , Aj αβ
i i<j
" #
X X
i2 i2
= θ A + θi θ j
Ai , Aj . (1.303)
i i<j αβ
Entonces
X
Ĥ 2 =αi2 p2i + {αi , αj } pi pj + m(αi β + βαi )pi + m2 (1.304)
i<j
(suma sobre ı́ndices repetidos). Si

αi2 = 1
{αi , αj } = 0 i 6= j
αi β + βαi = 0 (1.305)
Ĥ 2 = −∇2 + m2 (1.306)
y reemplazando en la ec. (1.298) llegamos a la ecuación de Klein-Gordon para ψ
2
∂ 2 2
−∇ +m ψ =0
∂t2

+ m2 ψ = 0 (1.307)
En términos de las matrices γ µ las condiciones en ec. (1.305) son

2
γ0 =1

i 2

i 2

i 2
α = 1 → γ 0γ iγ 0γ i = − γ = 1 → γ = −1

γ iγ 0 + γ 0γ i = γ i, γ0 = 0 (1.308)
De modo que

αi , αj = γ 0 γ i γ 0 γ j + γ 0 γ j γ 0 γ i = 0 i 6= j
−γ 0 γ 0 γ i γ j − γ 0 γ 0 γ j γ i = 0 i 6= j
i j j i
γ γ +γ γ =0 i 6= j
i j
γ ,γ = 0 i 6= j (1.309)
Las ecuaciones (1.308)(1.309) pueden escribirse como
{γ µ , γ ν } ≡ γ µ γ ν + γ ν γ µ = 2g µν 1 (1.310)
donde
γ µ = (γ 0 , γ i ) (1.311)
Además, de la ec. (1.283)

γ 0 γ µ† γ 0 = γ µ . (1.312)
Cualquier conjunto de matrices que satisfagan el álgebra en ec. (1.310) y la condición en ec. (1.312),
se conocen como matrices de Dirac. A ψ se le llama espinor de Dirac.
En términos de la matrices γ µ , el Lagrangiano de Dirac y la ecuación de Dirac, son respectivamente
de las ecs. (??) y (??)
L = ψ̄ (iγ µ ∂µ − m) ψ, (1.313)
iγ µ ∂µ ψ − mψ = 0, (1.314)
donde
ψ̄ = ψ † γ 0 . (1.315)
1.11.5 Propiedades de las matrices de Dirac

De la ec. (1.312) ( †
γ0 = γ0 µ=0
γ µ† = γ 0γ µγ 0 ⇒ † 2 . (1.316)
γ i = −γ 0 γ i = −γ i µ=i
Definiendo
γ5 = iγ0 γ1 γ2 γ3 , (1.317)
entonces,
γ52 = − γ0 γ1 γ2 γ3 γ0 γ1 γ2 γ3
γ52 = + γ02 γ1 γ2 γ3 γ1 γ2 γ3
γ52 = + γ1 γ2 γ3 γ1 γ2 γ3
γ52 = − γ2 γ3 γ2 γ3
γ52 =γ2 γ2 γ3 γ3
γ52 =1 . (1.318)
γ52 = 1, (1.319)
Teniendo en cuenta que γµ2 ∝ 1 y conmuta con las demás matrices, tenemos por ejemplo
γ5 γ3 =iγ0 γ1 γ2 γ32 = γ32 iγ0 γ1 γ2 = −γ3 iγ0 γ1 γ2 γ3 = −γ3 γ5

γ5 γ2 = − iγ0 γ1 γ22 γ3 = −γ22 iγ0 γ1 γ3 = −γ2 iγ0 γ1 γ2 γ3 = −γ2 γ5
γ5 γ1 =iγ0 γ12 γ2 γ3 = γ12 iγ0 γ2 γ3 = −γ1 iγ0 γ1 γ2 γ3 = −γ1 γ5
γ5 γ0 =iγ0 γ1 γ2 γ3 γ0 = −γ02 iγ1 γ2 γ3 = −γ0 γ5 . (1.320)
De modo que
{γµ , γ5 } = 0. (1.321)
Expandiendo el anticonmutador tenemos
γµ γ5 = −γ5 γµ
γ5 γµ γ5 = −γµ
Tr (γ5 γµ γ5 ) = − Tr γµ
Tr (γ5 γ5 γµ ) = − Tr γµ
Tr γµ = − Tr γµ , (1.322)
y por consiguiente
Tr γµ = 0. (1.323)
De otro lado, si
γ˜µ ≡ U γµ U † , (1.324)
para alguna matriz unitaria U , entonces γ˜µ corresponde a otra representación de álgebra de Dirac
en ec. (1.310), ya que

{γ̃ µ , γ̃ ν } = U γ µ U † , U γ ν U †
= U {γ µ , γ ν } U †
= 2g µν U U †
= 2g µν 1. (1.325)
Claramente, la condición en ec. (1.312) se mantiene para la nueva representación. Como γ0 es

hermı́tica, siempre es posible escoger una representación tal que γ˜0 ≡ U γ0 U † sea diagonal. Como
γ02 = 1, sus entradas en la diagonal deben ser ±1, y como Tr γ̃0 = 0, debe existir igual número de
+1 que de −1. Por lo tanto la dimensión de γ0 (y de γµ ) debe ser par: 2, 4, . . .. Para un fermion sin
masa
L = iψ † γ 0 γ 0 ∂0 ψ + iψ † γ 0 γ i ∂i ψ = iψ † ∂0 ψ + iψ † αi ∂i ψ , (1.326)
solo se requieren tres matrices 2 × 2 que satisfacen

αi , αj = 2δ ij , (1.327)
y por lo tanto pueden identificarse con las tres matrices de Pauli. Como en general tenemos 4 matrices
independientes, su dimensión mı́nima debe ser 4.
†
Como γ̃ i = γ 0 γ i γ 0 = γ i = −γ i , podemos definir la representación de paridad
γ̃ 0 =γ 0 , γ̃ i = −γ i , para U =γ 0 (1.328)
1.11.6 Lorentz Group

We must build a representation of the Lorentz Group in the Dirac space of n dimensions. First, let
us consider a simpler group, corresponding to the rotation group in tree dimensions. The generators
are the angular momentum operators J i , which satisfy the commutation relations
i j
J , J = iijk J k . (1.329)
The Pauli matrices are set of matrices satisfying this commutation relations:
i j
τ τ τk
, = i ijk (1.330)
2 2 2
donde τ i
0 1 0 −i 1 0
τ1 = τ2 = τ3 = (1.331)
1 0 i 0 0 −1
dividas por dos, corresponden a los generadores del Grupo. Las constantes de estructura del Grupo
corresponden a ijk . Como los generadores no conmutan, SU (2) es un Grupo de Lie no Abeliano.
Definiendo los generadores de SU (2) como
τi
Ti = , (1.332)
2
un elemento del Grupo puede escribirse como
iθ τi
U = eiT i
≈ 1 + iT i θi = 1 + i θi . (1.333)
2
Como antes, θi es el parámetro de la transformación.
Las matrices de Pauli y por consiguiente Ti satisfacen
τi† = τi
Tr (τi ) = 0 (1.334)
Además
det (τi ) = −1
{τi , τj } = 2δij · I ⇒ τi2 = I

Tr τ i τ j = 2δ ij
τi τj = iijk τk + δij (1.335)
In [12]:
It is generally true that one can find matrix representations of a continuous group by
finding matrix representations of the generators of the group (which must satisfy the
proper commutation relations), then exponentiating these infinitesimal representations.
For our present problem, we need to know the commutation relations of the generators of
the group of Lorentz transformations. For the rotation group, one can work the commu-
tation relations by writing the generators as differential operators; from the expression
J = x × p = x × (−i∇) , (1.336)
the angular momentum commutation relations (1.329) follow straightforwardly.

The last equation can be written as (summation of repeated indices)
J k = [x × (−i∇)]k = − iijk xi ∂j = iijk xi ∂ j (1.337)
J lm ≡ lmk J k =ilmk ijk xi ∂ j

=i(δli δmj − δlj δmi )xi ∂ j
=i(xl ∂ m − xm ∂ l ) . (1.338)
Involving three generators. The generalization to four-dimensions give to arise three further genera-
tors J 0i :
J µν = i(xµ ∂ ν − xν ∂ µ ) . (1.339)
The six generators J µν satisfy the algebra
[J µν , J ρσ ] =i(g νρ J µσ − g µρ J νσ − g νσ J µρ + g µσ J νρ ) . (1.340)
From [12]:
Any matrices that are to represent this algebra must obey these same commutation rules.
The exponentiation of the generators give to arise to group elements

J µν
Λ = exp −iωµν (1.341)
2
To find a representation of the usual boosts and rotations, consider a boost
   0   t+vx    
t t √ cosh ξ sinh ξ 0 0 t
1−v 2
 x   0  x+vt     
{xµ } =  → x0  =  √1−v2  =  sinh ξ cosh ξ 0 0 x = {Λµ ν } {xν } , (1.342)
y  y   y   0 0 1 0 y 
0
z z z 0 0 0 1 z
Since
∞
X ξ 2n
cosh ξ = ≈ 1 + O(ξ 2 )
n=0
2n!
X∞
ξ 2n+1
sinh ξ = ≈ ξ + O(ξ 2 ) , (1.343)
n=0
(2n + 1)!
one infinitesimal boost along x is

 
1 ξ 0 0
ξ 1 0 0
{Λµ ν }x−boost ≈
0
. (1.344)
0 1 0
0 0 0 1
Similarly a rotation by an infinitesimal angle θ = θ3 along xy–plane (or about the z–axis)
 
1 0 0 0
0 1 θ 0
{Λµ ν }xy−rotation ≈  
0 −θ 1 0 . (1.345)
0 0 0 1
In general we define the six independent Lorentz–Group parameters:
ω0i = −ωi0 ≡ξi
ω12 = −ω21 ≡θ3 ω32 = −ω23 ≡ − θ2 ω13 = −ω31 ≡θ1 . (1.346)
The 4 × 4 matrices
(J µν )αβ = i (δ µ α δ ν β − δ ν β δ µ α ) , (1.347)
where µ and ν label which of the six matrices we want, while α and β label components of the
matrices. These matrices satisfy the commutations relations (1.340), and generate the three boosts
and three rotations of the ordinary Lorentz 4-vectors:
i
Λα β ≈ δ α β − ωµν (J µν )α β (1.348)
2
1
Λ = 1 + ξi bi + θi ijk rjk , (1.349)
2
bi = − iJ i0 rjk = −iJ jk . (1.350)
1.11.7 Lorentz invariance of the Dirac Action

We need to satisfy the following conditions
S −1 (Λ)γ µ S(Λ) =Λµ ν γ ν
S † (Λ)γ 0 S(Λ) =γ 0 or S † (Λ)γ 0 = γ 0 S −1 (Λ) . (1.351)
In order to find a representation of the Lorentz Group in terms of the Dirac matrices we propose
1
S(Λ) = 1 + ξi B i + θi ijk Rjk . (1.352)
2
Instead of show the Lorentz invariance of the Dirac Action, we use the conditions derived from the
invariance, to find a representation in terms of the Dirac matrices for B i and Rjk . As a consistency
check, the resulting representation would satisfy the Lorentz algebra. In this way, by using eq. (1.349)
and (1.352), we obtain from
S −1 (Λ)γ µ S(Λ) = Λµ ν γ ν , (1.353)
that
1
Bi = γ 0γ i
2
1
Rjk = γ j γ k , (1.354)
2
which can be written in covariant form if we define
i µ ν
S µν = [γ , γ ] . (1.355)
4
In fact, the six set of non-zero independently generators are
i 0 i i 0 i
S 0i =γ γ − γ iγ 0 = γ γ = iB i
4 2
i i i j
ij
S = γ iγ j − γ j γ i = γ γ = iRij . (1.356)
4 2
It is worth notices that in fact S µν satisfy the Lorentz algebra, and therefore are the generators of
the Lorentz group elements:

S µν
S(Λ) = exp −iωµν
2
i
≈1 − ωµν S µν . (1.357)
2
Another consistency check is
S † (Λ)γ 0 S(Λ) =γ 0 , (1.358)

or equivalently
S † (Λ)γ 0 =γ 0 S −1 (Λ)

i µν † 0 0 i µν
1 + ωµν S γ =γ 1 + ωµν S
2 2
µν † 0 0 µν
S γ =γ S . (1.359)
Taking into account that

2 2
γ µ† γ ν † γ 0 = γ 0 γ µ† γ 0 γ ν †γ 0 = γ 0γ µγ ν , (1.360)
we have
i µ ν† 0
S µν † γ 0 = − [γ , γ ] γ
4
i
= − γ ν † , γ µ† γ 0
4
i µ† ν † 0
= γ ,γ γ
4
i
= [γ µ , γ ν ] γ 0
4
=γ 0 S µν
(1.361)
1.11.8 Dirac’s Lagrangian

Para una matriz de n dimensiones existen n2 matrices hermı́ticas (o anti–hermı́ticas) independientes.
Si se sustrae la identidad quedan n2 − 1 matrices hermı́ticas (o anti–hermı́ticas) independientes de
traza nula. En el caso n = 2 corresponden a las 3 matrices de Pauli. En el caso de la ecuación de
Dirac se requieren 4 matrices independientes, por lo tanto deben ser matrices 4 × 4. En efecto para
n = 4 existen 15 matrices independientes de traza nula dentro de las cuales podemos acomodar sin
problemas las 4 γ µ .
De [19]:
All Dirac matrix elements will now be written in the form
ψ(x)Γψ(x) , (1.362)
where Γ is a 4×4 complex matrix. The most general such matrix can always be expanded
in terms of 16 independent 4 × 4 matrices multiplied by complex coefficients. In short
Matriz Γ Transformación Número Escalar en Dirac

1 Escalar (S) 1 ψ̄ψ
γ5 Pseudoescalar (P) 1 ψ̄γ5 ψ
γµ Vector (V) 4 ψ̄γµ ψ
γµ γ5 Vector axial (A) 4 ψ̄γµ γ5 ψ
σµν = 2i [γµ , γν ] Tensor antisimétrico (T) 6 ψ̄σµν ψ
16
Table 1.1: Matrices Γi .
the matrices Γ can be regarded as a 16–dimensional complex vector space spanned by 16

matrices.
It is convenient to choose the 16 matrices, Γi , so that they have well defined transformation
properties under the Lorentz Transformations. Since the γ µ ’s have such properties, we
are lead to choose the following 16 matrices for this basis:
En la Tabla 1.1 se muestran las matrices de traza nula con sus propiedades de transformación
bajo el Grupo de Lorentz. En la última se muestra el correspondiente escalar en el espacio de Dirac
ψ̄Γψ. Demostración
J µ (x) ≡ ψ̄(x)γ µ ψ(x) →ψ̄(Λ−1 x)S −1 (Λ)γ µ S(Λ)ψ(Λ−1 x)

=Λµ ν ψ̄(Λ−1 x)γ ν ψ(Λ−1 x)
=Λµ ν J ν (Λ−1 x) . (1.363)
In [19]: Problem 5.4:
ψγ5 ψ → ψS −1 (Λ)γ 5 S(Λ)ψ = (det Λ)ψγ5 ψ (1.364)
The solution is in Appendix C. of Burgess book, by using
i
γ5 = µναβ γ µ γ ν γ α γ β (1.365)
24
and
det Λ = µναβ Λµ 1 Λν 2 Λα 3 Λβ 4 . (1.366)

1.12 Electrodinámica Cuántica

Para hacer el Lagrangiano en ec. (1.313) invariante gauge local bajo U (1)Q , procedemos de la forma
usual. El campo transforma como
ψ → ψ 0 = e−iθ(x)Q ψ
ψ̄ → ψ̄ 0 = ψ̄eiθ(x)Q , (1.367)
donde Q es el generador de carga eléctrica en unidades de la carga del electrón.

La derivada covariante se define de manera que transforma de la misma forma que el campo,
introduciendo el campo gauge Aµ
∂ µ → Dµ = ∂ µ − ieQAµ , (1.368)
donde e es la carga eléctrica del electrón. De esta forma, si ψe es el campo que representa al electrón
eQψe = e(−1)ψe = −eψe . (1.369)
El Lagrangiano correspondiente a la interacción de un fermión y el campo electromagnético cor-

responde al Lagrangiano de Dirac con la derivada normal reemplzada por la derivada covariante, y el
correspondiente término cinético invariante gauge y de Lorentz asociado al nuevo campo introducido
en la derivada covariante: Aµ . Este campo es necesario para compensar los cambios en la energı́a y
momentum que sufre el electrón como consecuencia de imponer la invarianza de la Acción bajo un
cambio de fase local
L = ψ (iγ µ Dµ − m) ψ − 41 F µν Fµν , (1.370)
y es invariante bajo transformaciones locales U (1)Q . Desarrollando la expresión anterior, tenemos
L = ψ̄ [iγ µ (∂µ − ieQAµ ) − m] ψ − 14 F µν Fµν

= ψ̄ (iγ µ ∂µ − m) ψ + eQψ̄γ µ ψAµ − 14 F µν Fµν . (1.371)
Este Lagrangiano da lugar a la Acción de la teorı́a conocida como Electrodinámica Cuántica (QED
de sus siglas en inglés).
Aplicando las ecuaciones de Euler-Lagrange para ψ̄, tenemos
(iγ µ ∂µ − m)ψ + eQγ µ Aµ ψ = 0

(iγ µ ∂µ − i2 eQγ µ Aµ − m)ψ = 0
[iγ µ (∂µ − ieQAµ ) − m]ψ = 0
(iγ µ Dµ − m)ψ = 0. (1.372)
1.12. ELECTRODINÁMICA CUÁNTICA 67
Que corresponde a la ecuación de Dirac en presencia del campo electromagnético. Mientras que para
el campo Aµ , tenemos
ρη
1 F Fρη ∂Aρ
− ∂µ − eQψ̄γ ρ ψ =0
4 ∂ (∂µ Aν ) ∂Aν
∂µ F µν = −eQψ̄γ ν ψ (1.373)
Definimos entonces la corriente electromagnética generada por el fermión como
j µ = −eQψ̄γ µ ψ. (1.374)
De nuevo, la aparición de la interacción electromagnética es una consecuencia de la invarianza gauge

local.
El cálculo directo de la corriente
∂L ∂L
Jµ = δψ + δ ψ̄
∂ (∂µ ψ) ∂ ∂µ ψ̄
= iψγ µ (−iθ(x)Q)ψ = θ(x)Qψγ µ ψ , (1.375)
y para la ecuación de Dirac, a diferencia de la ecuación de Schrödinger, la corriente de probabilidad

tiene la misma forma que la corriente electromagnética.
De esta manera podemos reescribir el Lagrangiano en términos de un Lagrangiano libre y otro
de interacción
L = Lfree + Lint , (1.376)
Lf ree =ψ̄ (iγ µ ∂µ − m) ψ − 14 F µν Fµν

Lint =eQψ̄γ µ ψAµ . (1.377)
Para la QED sólo hay un término de interacción que es suficiente para explicar todos los fenoménos
electromagnéticos y su interacción con la materia. Este esta representado por el diagrama de Feynman
mostrado en la Figura 1.2
La repulsión electromagnética esta representada por la figura 1.3. En la Figura (a) el primer
electrón emite un fotón y se dispersa, mientras que el segundo absorbe el fotón y se dispersa en
la dirección opuesta. En la Figura (b) el primer electón absorve el fotón emitido por el segundo
electrón. Los dos diagrams se representa por uno único con el fotón en horizontal como se muestra
en la Figura (c).
Figure 1.2: Feynman rule for QED
Figure 1.3: Electromagnetic repulsion. The diagrams (a) and (b) are summarized in the diagram (c)
1.13. CROMODINÁMICA CUÁNTICA 69
1.13 Cromodinámica Cuántica
Los protones, neutrones, piones, kaones y demás hadrones, son partı́culas compuestas de consti-
tuyentes elementales llamados quarks. Por ejemplo los protones, neutrones y piones están constitu-
idos de quarks up y down. Los hadrones están dividos en bariones, B, constituidos de tres quarks,
y los mesones, M , de dos. Para satisfacer el principio de exclusión de Pauli, y justificar el confi-
namiento de los hadrones, se requiere que cada quark contenga Nc cargas diferentes, llamadas cargas
de color, de manera que la carga de color de un hadrón sea cero. Muchos resultados experimentales
respaldan la existencia de tres cargas de color para cada quark, Nc = 3. De este modo cada quark
q = u, d, c, s, t, b viene en tres colores
qα = q1 , q2 , q3 = qr , qb , qg , (1.378)
donde los últimos subı́ndices hacen referencia a los colores red, blue, green. De este modo los Bariones
y mesones están descritos por combinaciones singletes de color del tipo qr qb qg y qr q̄r ,
1 1
B = √ αβγ |qα qβ qγ i M = √ δ αβ |q̄α qβ i (1.379)
6 3
Estos estados son singletes de color. Una de las determinaciones de Nc proviene del observable
σ(e+ e− → hadrones)
R≈ (1.380)
σ(e+ e− → µ+ µ− )
Para f = u, d, s, c, b, t, (en orden de masa) tenemos que para una energı́a donde se pueden producir
hadrones compuestos de hasta quarks fmax
Pfmax PNc + − ¯
f =u α=1 σ(e e → fα fα )
R≈
σ(e+ e− → µ+ µ− )
Pfmax + − ¯
f =u σ(e e → f f )
R ≈Nc (1.381)
σ(e+ e− → µ+ µ− )
De este modo R esta dado por la suma de las cargas eléctricas al cuadrado
P 2
f Qf
R ≈Nc
Q2µ
fmax
X
=Nc Q2f
f =u

2 2 −1 2 2

Nc [( 3 ) + 2( 3 ) ] = 3 Nc f = u, d, s, fmax = s
= Nc [2( 32 )2 + 2( −1 )2 ] = 10 Nc fmax = c

 2 2
3
−1 2
9
11
Nc [2( 3 ) + 3( 3 ) ] = 9 Nc fmax = b


2 Nc = 3, fmax = s
10
= 3 Nc = 3, fmax = c (1.382)

 11
3
Nc = 3, fmax = b
√
En la figura, tomada de [?], se muestra el gráfico de R con respecto a s√(la energı́a de centro de masa
√ que va hasta una energı́a s ≈ 4 GeV que corresponden
de la colisión). Se observan dos escalones, uno
a f = u, d, s, con un R ≈ 2, y otro hasta s ≈ 40 GeV que corresponde a f = u, d, s, c, b, con un
R ≈ 3.7 ≈ 11/3. Los dos valores de R son compatibles con los esperados de la ec. (1.382). Como
referencia también se señalan los valores para Nc = 4 (en rojo).
Si queremos que el color sea una carga conservada como la carga eléctrica, ésta debe ser la
consecuencia de una simetrı́a gauge local. Para tener tres cargas diferentes la posibilidad más simple
es imponer la simetrı́a SU (3)c , tal que tengamos un vector compuesto de 3 espinores de Dirac en el
espacio de color:    
ψr qr
Ψ=  ψ b
 
= qb  . (1.383)
ψg qg
El Lagrangiano de Dirac con invarianza gauge global SU (3), para un quark, se puede escribir como
Lglobal = iΨ̄γ µ ∂µ Ψ − mΨ̄Ψ, (1.384)
donde
0 λa
Ψ → Ψ = exp iθa Ψ. (1.385)
2
a = 1, . . . , 8, λa /2 son los ocho generadores de SU (3) y θa son los parámetros de la transformación
global. Los generadores de SU (3)
λa
Λa ≡ , (1.386)
2
Figure 1.4: Datos para R
satisfacen el álgebra
λa λb λc
, = if abc , (1.387)
2 2 2
donde f abc son las constantes de estructura fina de SU (3).
En un análisis similar al de la sección 1.11.8 tenemos que la Acción invariante gauge local bajo
SU (3)c , se obtiene de reemplazar la derivada normal por la derivada covariante
1
Llocal = iΨ̄γ µ Dµ Ψ − mΨ̄Ψ − Tr (Gµν Gµν ) , (1.388)
2
donde
Ψ → Ψ0 = U (x)Ψ
Dµ Ψ → (Dµ Ψ)0 = U (x)Dµ Ψ, (1.389)
con la matriz 3 × 3

λa
U (x) = exp iθa (x) , (1.390)
2
y
λa a
Dµ = ∂µ − igsG ≡ ∂µ − igs Gµ (1.391)
2 µ
donde hemos definido la matriz 3 × 3 Gµ , como

λa
(Gµ )αβ = Gaµ (1.392)
2 αβ
Este Lagrangiano da lugar a la interacción fuerte y es llamado el Lagrangiano de la Cromodinámica

Cuántica, o el Lagrangiano de la QCD de sus siglas en Inglés.
De (1.389), tenemos
Dµ Ψ → (Dµ Ψ)0 =Dµ0 Ψ0 = U (x)Dµ Ψ

Dµ0 U Ψ = U (x)Dµ Ψ . (1.393)
Por consiguiente
µ
D0 U = U Dµ (1.394)
Dµ → (Dµ )0 = U Dµ U −1 (1.395)
Desarrollando a ambos lados
Dµ ψ → (Dµ ψ)0 = Dµ0 ψ 0 =Dµ0 ψ 0

µ
(∂ µ − igs G0 )U ψ =U Dµ U −1 U ψ
µ
(∂ µ − igs G0 )U ψ =U (∂ µ − igs Gµ )ψ
µ
U ∂ µ ψ + (∂ µ U )ψ − igs G0 U ψ =U ∂ µ φ − igs U Gµ ψ
µ
(∂ µ U )ψ − igs G0 U ψ = − igs U Gµ ψ
µ
−igs G0 U ψ = − (∂ µ U )ψ − igs U Gµ ψ , (1.396)
de modo que
µ 1 µ
G0 U = (∂ U ) + U Gµ
igs
µ i
G0 = − (∂ µ U )U −1 + U Gµ U −1 . (1.397)
gs
Como U es unitaria, la transformación de los campos gauge puede escribirse como
i µ
Gµ → (Gµ )0 = U Gµ U −1 − (∂ U ) U † . (1.398)
gs
Entonces
µ i
Λa G0 a ≈(1 + iθb Λb )Λc Gµc (1 − iθd Λd ) − [i(∂ µ θe )Λe (1 − iθf Λf )]
gs
i
=(Λc + iθb Λb Λc )(1 − iθd Λd )Gµc − [i(∂ µ θe )Λe (1 − iθf Λf )]
gs
1
≈[Λc − iθd Λc Λd + iθb Λb Λc ]Gµc + Λe ∂ µ θe
gs
1
=[Λc − iθb (Λc Λb − Λb Λc )]Gµc + Λe ∂ µ θe
gs
1
=Λa Gµa − i(if acb Λa )Gµc θb + Λa ∂ µ θa
gs

a µ 1 µ acb µ
=Λ Ga + ∂ θa + f Gc θb (1.399)
gs
de donde
µ 1 µ
Gµa → G0 a ≈Gµa + ∂ θa + f abc Gµb θc (1.400)
gs
que se reduce al caso Abeliano cuando las constates de estructura son cero. Como era de esperarse
cada campo gauge tiene asociado un parámetro de transformación gauge θa (x).
Similarmente, definiendo la matriz 3 × 3,
i µ ν λa
Gµν = [D , D ] ≡ Gµν , (1.401)
gs 2 a
tenemos
i µ
Gµν ψ = [∂ − igs Gµ , ∂ ν − igs Gν ]ψ
gs
i
= [(∂ µ − igs Gµ ) (∂ ν − igs Gν ) ψ − (∂ ν − igs Gν ) (∂ µ − igs Gµ ) ψ]
gs
i µ ν
= ∂ ∂ ψ − gs2 Gµ Gν ψ − igs [∂ µ (Gν ψ) + Gµ ∂ ν ψ] − ∂ ν ∂ µ ψ + gs2 Gν Gµ ψ + igs [∂ ν (Gµ ψ) + Gν ∂ µ ψ]
gs
i
= {(∂ µ ∂ ν − ∂ ν ∂ µ )ψ − gs2 (Gµ Gν − Gν Gµ )ψ − igs [(∂ µ Gν ) − (∂ ν Gµ )]ψ
gs
− igs [Gν ∂ µ ψ + Gµ ∂ ν ψ − Gµ ∂ ν ψ + Gν ∂ µ ψ]}
=[∂ µ Gν − ∂ ν Gµ − igs (Gµ Gν − Gν Gµ )]ψ
={∂ µ Gν − ∂ ν Gµ − igs [Gµ , Gν ]}ψ (1.402)
De modo que
Gµν =∂ µ Gν − ∂ ν Gµ − igs [Gµ , Gν ] , (1.403)
que se reduce al caso Abeliano cuando los bosones gauge conmutan. En términos de componentes
b µ
Λa Gµν a µ ν a ν µ c ν
a =Λ ∂ Ga − Λ ∂ Ga − igs [Λ Gb , Λ Gc ]
=Λa ∂ µ Gνa − Λa ∂ ν Gµa − igs [Λb , Λc ]Gµb Gνc
=Λa ∂ µ Gνa − Λa ∂ ν Gµa − igs (iΛa fabc )Gµb Gνc
=Λa ∂ µ Gνa − Λa ∂ ν Gµa + Λa gs fabc Gµb Gνc . (1.404)
Por consiguiente
Gµν µ ν ν µ abc µ ν eµν + gs f abc Gµ Gν ,
a = ∂ Ga − ∂ Ga + gs f Gb Gc ≡ Ga b c (1.405)
con
eµν = ∂ µ Gν − ∂ ν Gµ
G (1.406)
a a a
A diferencia del caso Abeliano Gµν ya no es invariante bajo transformaciones gauge
µν i 0µ 0ν
Gµν → G0 = D ,D
gs
i
= U Dµ U −1 , U Dν U −1
gs
= U Gµν U −1 . (1.407)
Note que con la definición (1.401), la derivada covariante de la matrix Gµν , transforma como la
matrix Gµν
Dµ Gµν → (Dµ Gµν )0 = U Dµ Gµν U −1 . (1.408)
Para poder obtener un invariante bajo transformaciones gauge a partir del producto Gµν Gµν ,
debemos utilizar la traza
µν
Tr (Gµν Gµν ) → Tr G 0 G 0 µν = Tr U Gµν U −1 U Gµν U −1

= Tr U Gµν Gµν U −1

= Tr U −1 U Gµν Gµν
= Tr (Gµν Gµν ) . (1.409)
Teniendo en cuenta la normalización de las matrices de Gell-Man

Tr λa λb =2δ ab
1
Tr Λa Λb = δ ab , (1.410)
2
tenemos (suma sobre indices repetidos de SU (3))
µν
Tr (Gµν Gµν ) → Tr G 0 G 0 µν = Tr Λa Gµν b b
a Λ Gµν

= Tr Λa Λb Gµν b
a Gµν
1
= δ ab Gµν
a Gµν
b
2
1
= Gµν Ga . (1.411)
2 a µν
Expandiendo el Lagrangiano en ec. (1.388), tenemos

λa a 1
µ
L =iΨ̄γ ∂µ − igs Gµ Ψ − mΨ̄Ψ − Tr (Gµν Gµν )
2 2

λa 1
=iΨ̄γ µ ∂µ − igs Gaµ Ψ − mΨ̄Ψ − Gµν Ga
2 4 a µν
λa 1
=iΨ̄γ µ ∂µ Ψ − mΨ̄Ψ + gs Ψ̄γ µ Gaµ Ψ − Gµν Ga
2 4 a µν
λa 1 eµν ea
=iΨ̄γ µ ∂µ Ψ − mΨ̄Ψ + gs Ψ̄γ µ ΨGaµ − G G
2 4 a µν
1 eµν eaµν + gs2 f abc fade Gµ Gνc Gdµ Geν

− gs Ga fade Gdµ Geν + gs f abc Gµb Gνc G b
4
=Lfree + Lgauge + LSI , (1.412)
donde
Lfree =iΨ̄γ µ ∂µ Ψ − mΨ̄Ψ

λa 1 eµν ea
Lgauge =gs Ψ̄γ µ ΨGaµ − G G
2 4 a µν
1
LSI = − e µν d e abc µ ν e a 2 abc µ ν d e
gs Ga fade Gµ Gν + gs f Gb Gc Gµν + gs f fade Gb Gc Gµ Gν . (1.413)
4
Hemos divido el Lagrangiano en tres partes
• El Lagrangiano libre de Dirac
• Una parte gauge que puede escribirse como un Lagrangiano electromagnético:

1 µ ν
Lgauge = − (∂ Ga − ∂ ν Gµa ) ∂µ Gaν − ∂ν Gaµ − Jaν Gaν , (1.414)
4
dende
λa
Jaµ = −gs Ψ̄γ µ Ψ, (1.415)
2
es la nueva corriente conservada de interacción fuerte que surge como consecuencia de la invar-
ianza gauge local SU (3); y
• Una parte de auto-interacciones gauge:
gs abc ea µ ν gs2 abc

LSI = − f Gµν Gb Gc + f fade Gµb Gνc Gdµ Geν
2 4
gs abc µ ν g2
= − f (∂ Ga − ∂ ν Gµa ) Gbµ Gcν − s f abc fade Gµb Gνc Gdµ Geν . (1.416)
2 4
que se desaparecen en el caso Abeliano.
El Lagrangiano de interacción es:
λa gs g2
Lint = gs Ψ̄γ µ ΨGaµ − f abc (∂ µ Gνa − ∂ ν Gµa ) Gbµ Gcν − s f abc fade Gµb Gνc Gdµ Geν . (1.417)
2 2 4
From [21] (pag 136):
The quarks have an additional type of polarization that is not related to geometry. The
idiot physicists, unable to come up with any wonderful Greek words anymore, call this
type of polarization by the unfortunate name of “color”, which has nothing to do with
color in the nornal sense. At a particular time, a quark can be in one of three conditions,
or “colors”–R, G, or B (can you guess what they stand for?). A quark’s “color” can be
changed when the quark emits or absorbs a gluon. The gluons come in eigth diffent types,
according to the “colors” they can couple with. For example, if a red quark changes to
green, it emits a red-antigreen gluon–a gluon that takes the red from quark and gives
it green (“antigreen” means the gluon is carrying green in the opposite direction). This
gluon could be absorved by a green quark, which changes to red (see Fig. 1.5). There are
eigth different possible gluons, such as red-antired, red-antiblue, red-antigreen, and so on
u d
green red
red
t
green
red green
u d
Figure 1.5: Quark–gluon interaction
(you’d think there’d be nine, but for technical reasons, onw is missing)2 . The theory is
not very complicated. The complete rule of gluons is: gluons couple with things having
“color”–it just requires a little bookkeeping to keep track of where the “colors go”. There
is, however, an interesesting possibility created by this rule: gluons can couple with other
gluons (see Fig. 1.6).
El primer término da lugar a interacciones de cambio de color de quarks como la que se ilustra
en la Figura 1.5
Mientras que el segundo y tercer término dan lugar a autointeracciones de los gluones como se
muestra en la Figura 1.6
Todas las interacciones están determinadas en términos de una única constante de acoplamiento
gs . Las autointeracciones gauge pueden explicar aspectos de la interacción fuerte como la libertada
asintótica, que consiste en que las interacciones fuertes se vuelven más débiles a distancias cortas.
En términos de ı́ndices de color la corriente, y las otras partes del Lagrangiano, pueden escribirse
2
 
rr̄ rb̄ rḡ
 br̄ bb̄ bḡ  , with rr̄ + bb̄ + gḡ = 0
gr̄ g b̄ gḡ
Figure 1.6: Triple–gluon self–interaction. The anticolors are the colors running back in time.
como

λa
Jaµ α µ β
= −gs q̄ γ q . (1.418)
2 αβ
Note que tanto para la Electrodinámica Cuántica como para la Cromodinámica Cuántica la corriente
ψ̄Γψ es vectorial. Para las interacciones débiles la estructura es más complicada y requiere un
conocimiento más profundo de la ecuación de Dirac y sus soluciones.
1.13.1 Ecuaciones de Euler–Lagrange
Sigiendo los mismos procedimientos anteriores debemos llegar a los siguientes resultados. Para el
campo Ψ
(iγ µ Dµ − m)Ψ = 0 , (1.419)


∂L ∂L
∂µ a
−
∂ (∂µ Gν ) ∂Gaν

1 ∂ λa
e µν
=∂µ −Ga − gs fdbc Gb Gc ρ σ
∂ρ Gσ − ∂σ Gρ − gs Ψγ ν Ψ
d d
2 a
∂ (∂µ Gν ) 2
gs dbc eρσ g s
+ f Gd (δρν δba Gcσ + Gbρ δσν δca ) + f ibc fide (g ρα g σβ Gbα Gcβ Gdρ Geσ )
2
4
=∂µ −G µν 1
ea − gs fdbc G Gc (δρµ δσν δda − δσµ δρν δda ) − gs Ψγ ν λa Ψ
ρ σ
b
2 2
gs dac eνσ c gs dba eρν b
+ f Gd Gσ + f Gd Gρ
2 2
gs ibc
+ f fide g g (δαν δba Gcβ Gdρ Geσ + Gbα δβν δca Gdρ Geσ + Gbα Gcβ δρν δda Geσ + Gbα Gcβ Gdρ δσν δea )
ρα σβ
4
= − ∂µ G ea − gs f G G + gs fabc G G − gs Ψγ ν λa Ψ
µν 1 abc µ ν 1 ν µ
b c b c
2 2 2
gs eνσ gs adb eνρ b
− f adc G c
d Gσ − f Gd Gρ
2 2
gs gs gs
+ f iac fide g ρν g σβ Gcβ Gdρ Geσ + f iba fide g ρα g σν Gbα Gdρ Geσ + f ibc fiae g να g σβ Gbα Gcβ Geσ
4 4 4
gs ibc ρα νβ b c d
+ f fida g g Gα Gβ Gρ . (1.420)
4
Desarrollando los cuatro últimos términos, tenemos
gs iac gs gs
f fide g ρν g σβ Gcβ Gdρ Geσ + f iba fide g ρα g σν Gbα Gdρ Geσ + f ibc fiae g να g σβ Gbα Gcβ Geσ
4 4 4
gs ibc ρα νβ b c d
+ f fida g g Gα Gβ Gρ
4
gs iac gs gs gs
= f fide Gνd Gµc Geµ + f iba fide Gνe Gµb Gdµ + f ibc fiae Gνb Gµc Geµ + f ibc fida Gνc Gµb Gdµ
4 4 4 4
gs dac gs dca gs dbc gs
= f fdje Gj Gc Gµ + f fdje Ge Gc Gµ + f fdae Gb Gc Gµ + f dbc fdea Gνc Gµb Geµ
ν µ e ν µ j ν µ e
4 4 4 4
gs dac gs dca gs gs
= f fdje Gj Gc Gµ + f fdje Ge Gc Gµ + fdac f Gj Gce Gµ + fdca f dje Gνe Gµj Gcµ
ν µ e ν µ j dje ν µ c
4 4 4 4
gs abc b gs abc b gs gs
= − f Gµ fcej Ge Gj − f Gµ fcej Ge Gj − fabc Gµ f Ge Gj − fabc Gbµ f cej Gµe Gνj
µ ν µ ν b cej µ ν
4 4 4 4
b cej µ ν
= − gs fabc Gµ f Ge Gj (1.421)
Entonces

∂L ∂L
∂µ −
∂ (∂µ Gaν ) ∂Gaν

eµν abc µ ν ν λa eµν − gs facd Gcµ f dej Gµe Gνj
=∂µ −G a − g s f Gb c − gs Ψγ
G Ψ − gs f acd Gcµ Gd
2
acd c µν λa
= − ∂µ Gµν
a − gs f Gµ Gd − gs Ψγ ν Ψ = 0 . (1.422)
2
Entonces las Ecuaciones de Euler Lagrange para Gaν , son
acd c µν λa
∂µ Gµν
a + gs f Gµ Gd = −gs Ψγ ν Ψ. (1.423)
2
Definiendo
λa
Jaµ = −gs Ψ̄γ µ Ψ, (1.424)
2
La ec.(1.423) puede reescribirse como:

ν λa
∂µ Gµν
a = −gs fabc Gbµ Gµν
c + Ψ̄γ Ψ (1.425)
2
y usando el hecho que ∂µ ∂ν = ∂ν ∂µ :
∂ν ∂µ Gµν eµν + gs ∂ν ∂µ (fabc Gµ Gν )
a =∂ν ∂µ G b c
1
=0 + [gs ∂ν ∂µ (fabc Gµb Gνb ) + gs ∂ν ∂µ (fabc Gµb Gνc )]
2
1
= [gs ∂ν ∂µ (fabc Gµb Gνb ) + gs ∂µ ∂ν (fabc Gµb Gνc )]
2
1
= [gs ∂ν ∂µ (fabc Gµb Gνb ) + gs ∂ν ∂µ (facb Gνc Gµb )]
2
1
= [gs ∂ν ∂µ (fabc Gµb Gνb ) − gs ∂µ ∂ν (fabc Gµb Gνc )]
2
=0 , (1.426)
como en el caso Abeliano, tenemos la corriente conservada
∂ν j ν = 0 , (1.427)
donde

ν λa
jaν = − gs fabc Gbµ Gµν
c + Ψ̄γ Ψ . (1.428)
2
El primer término corresponde a las autointeracciones y el segundo a la corriente de color generada
por los quarks.
1.13.2 Derivada covariante adjunta

Toda el algebra de SU (3) se puede escribir en notación vectorial en términos de vectores de 8 com-
ponentes asociados al espacio de los generadores de SU (3). Este nos permitira entender como las
autointeracciones gauge emergen también de la derivada covariante cuando se escribe en la repre-
sentación adjunta de SU (3).
Definiendo el producto vectorial de SU (3) como
(A × B)a = fabc Ab B c , (1.429)
si Gµ es un vector en el espacio SU (3) con las 8 componentes Gµa , entonces podemos escribir (1.400)
como
1
Gµ → G0µ = Gµ + ∂ µ θ + Gµ × θ . (1.430)
gs
Podemos escribir también la ec. (1.405) en términos de vectores en el espacio SU (3) como:
Gµν = ∂ µ Gν − ∂ ν Gµ + gs Gµ × Gν , (1.431)
donde Gµν es el vector en el espacio SU (3) con las 8 componentes Gµν

a .
De igual forma, podemos escribir (1.423) en forma vectorial como
∂µ Gµν + gs Gµ × Gµν = Jν (1.432)
y la corriente conservada como

ν µν νλ
j = − gs −G × Gbν + Ψ̄γ Ψ . (1.433)
2
Como Gµν es una matrix 8 × 8, su derivada covariante debe estar en la representación adjunta de
SU (3)
(Λa )bc = −i (f a )bc , (1.434)
con

Λa , Λb = ifabc Λc . (1.435)
En esta representación la derivada covariante queda
Dµ = ∂µ − igs Λa Gaµ (1.436)

En componentes
(Dµ )ab =δab ∂µ − igs (Λc )ab Gcµ

=δab ∂µ − gs fcab Gcµ
=δab ∂µ + gs facb Gcµ . (1.437)
Aplicada sobre la componente Gµν

b , queda
(Dµ )ab Gµν µν c µν

b =δab ∂µ Gb + gs facb Gµ Gb
(Dµ )ab Gµν µν µν
b =∂µ Ga + gs (Gµ × G )a
(Dµ Gµν )a =∂µ Gµν µν
a + gs (Gµ × G )a , (1.438)
podemos escribir la derivada covariante de Gµν = (Gµν µν

1 , · · · , G8 ) como
Dµ Gµν =∂µ Gµν + gs Gµ × Gµν . (1.439)
Entonces las las Ecuaciones de Euler Lagrange para Gaµν , en (1.423) se pueden escribir como
Dµ Gµν =Jν , (1.440)
donde el vector en espacio SU (3) Jν , tiene por componentes
λa
Jaµ = −gs Ψ̄γ µ Ψ. (1.441)
2
Para escribir el Lagrangiano en forma vectorial en el espacio SU (3), debemos reescribir la trans-
formación gauge de Gµν en términos de vectores de SU (3). Como
Gµν → G0µν =U Gµν U †

=(1 + iθb Λb )Λc Gcµν (1 − iθb Λb ) (1.442)
podemos realizar los mismos pasos que en (1.399). El resultado es

a
Gaµν → G0 µν ≈Gaµν + f abc Gbµν θc . (1.443)
Note que en el caso Abeliano fabc = 0, el tensor correspondiente es invariante gauge, como ocurre el
caso electromagnético. En notación de vectores de SU (3):
Gµν → G0µν ≈ Gµν + Gµν × θ . (1.444)

Utilizando la propiedad cı́clica del triple producto escalar
A · (B × C) = B · (C × A) = C · (A × B) , (1.445)
podemos construir el invariante
µν
Gµν a
a Gµν = G
µν
· Gµν → G0 · G0µν ≈Gµν · Gµν + Gµν · (Gµν × θ) + (Gµν × θ) · Gµν
=Gµν · Gµν + Gµν · (θ × Gµν ) + (Gµν × θ) · Gµν
=Gµν · Gµν − (Gµν × θ) · Gµν + (Gµν × θ) · Gµν
=Gµν · Gµν . (1.446)
El Lagrangiano de la QCD escrito en forma de vectores de SU (3) es

µ λ 1
L = iΨ̄γ ∂µ − igs · Gµ Ψ − mΨ̄Ψ − Gµν · Gµν . (1.447)
2 4
El Lagrangiano para los campos gauge, el cual puede generalizarse para cualquier teorı́a SU (N ), es
1
Lgluon = Lgauge + LSI = − Gµν · Gµν − Jν · Gν , (1.448)
4
Da lugar la ecuaciones de Maxwell pero con la derivada normal reemplzada por la derivada covariante
Dµ Gµν =Jν , (1.449)
donde
Dµ Gµν =∂µ Gµν + gs Gµ × Gµν . (1.450)
Note que en el caso Abeliano, U (1), la derivada covariante del tensor de campo se reduce a la derivada
normal de dicho tensor. El término extra en la derivada covariante da lugar a las autointeracciones
de los campos gauge.
• Ejercicio:
Muestre que la derivada covariante de Gµν , transforma como Gµν .
Figure 1.7: V (φ) = 12 µ2 φ2 con µ2 > 0
1.14 Spontaneous symmetry breaking

Escribamos el Lagrangiano para una partı́cula escalar real de masa m como
L = 21 ∂ µ φ∂µ φ − V (φ) (1.451)
con
V (φ) = 12 µ2 φ2 . (1.452)
Este Lagrangiano es simétrico bajo la transformación discreta φ → −φ.
Cuando µ2 > 0, el campo tiene excitaciones alrededor del mı́nimo del potencial que cuestan
energı́a y dicho término se interpreta como la masa de la partı́cula. Ver figura 1.7. En Teorı́a
Cuántica de Campos al estado de mı́nima energı́a se le llama el vacı́o y las excitaciones alrededor del
vació corresponden a las partı́culas.
Si µ2 < 0, no existe un mı́nimo del potencial alrededor del cual el campo pueda oscilar. Además
el alejamiento del campo del punto de simetrı́a del potencial no cuesta energı́a. Por consiguiente en
ese caso, el término de interacción
V (φ) = 12 µ2 µ2 < 0, (1.453)
no puede interpretarse como un término de masa en el Lagrangiano dado por la ec. (1.451).
Consideremos ahora el potencial
V (φ) = 21 µ2 φ2 + 14 λφ4 µ2 < 0, λ > 0 (1.454)

1.14. SPONTANEOUS SYMMETRY BREAKING 85
Figure 1.8: V (φ) = 12 µ2 φ2 + 14 λφ4 con µ2 < 0, y λ > 0. Simetrı́a exácta
que mantiene la simetrı́a bajo la transformación discreta φ → −φ. λ > 0 garantiza la aparición de los
dos mı́nimos que se muestran el la figura 1.8. Si la energı́a es suficientemente alta como se muestra
en la figura 1.8, las excitaciones son simétricas con respecto al máximo del potencial y el término en
µ2 no puede interpretarse como masa para la partı́cula escalar.
Sin embargo, si la energı́a es suficientemente baja como se muestra en la figura 1.9, las excitaciones
alrededor del mı́nimo dan lugar a la aparición de un término de masa para el campo escalar. Además,
dichas excitaciones no respetan la simetrı́as φ → −φ. En tal caso decimos que la simetrı́a ha sido
espontáneamente rota: aunque el Lagrangiano mantiene la simetrı́a original, el vacı́o la rompe.
Para analizar cuantitativamente el espectro de partı́culas es necesario expandir el campo alrededor
del mı́nimo y determinar las excitaciones. Establezcamos en primer lugar los mı́nimos del potencial.
La ∂V /∂φ = 0 da lugar a
µ2 φ + λφ3 = 0 (1.455)
φ(µ2 + λφ2 ) = 0, (1.456)
con extremos φmax = 0, y r
−µ2
φmin ≡ hφi ≡ v = ± . (1.457)
λ
De hecho
∂ 2V
2
= µ2 + 3λφ2 . (1.458)
∂φ
p
φ = 0 corresponde a un máximo, mientras que la segunda derivada para φ = ± −µ2 /λ es −2µ2 > 0
y corresponden a los mı́nimos. Expandiendo el campo alrededor del mı́nimo
φ(x) = H(x) + v (1.459)
Figure 1.9: V (φ) = 12 µ2 φ2 + 14 λφ4 con µ2 < 0, y λ > 0. Simetrı́a espontáneamente rota.
V (φ) = 12 µ2 φ2 + 41 λφ4
= 12 µ2 (H + v)2 + 41 λ(H + v)4
= 21 µ2 (H + v)2 + 41 λ(H + v)4
2
= 21 µ2 H 2 + 2vH + v 2 + 41 λ H 2 + 2vH + v 2
h 2 i
= 21 µ2 H 2 + 2vH + v 2 + 41 λ H 4 + 2H 2 2vH + v 2 + 2vH + v 2

= 21 µ2 H 2 + 2vH + v 2 + 41 λ H 4 + 4vH 3 + 2H 2 v 2 + 4v 2 H 2 + 4v 3 H + v 4

= 21 µ2 H 2 + 2vH + v 2 + 41 λ H 4 + 4vH 3 + 6H 2 v 2 + 4v 3 H + v 4

= 21 µ2 H 2 − 23 H 2 µ2 + µ2 vH − µ2 vH + 21 µ2 v 2 − 41 µ2 v 2 + 14 λ H 4 + 4vH 3

V (H) = 12 −2µ2 H 2 + λvH 3 + 41 λH 4 + 14 µ2 v 2 , (1.460)
y

LH = 21 ∂ µ H∂µ H − 1
2
−2µ2 H 2 − λvH 3 − 14 λH 4 + constant. (1.461)
Entonces H adquiere una masa −2µ2 y no es invariante bajo H → −H.

Otro método es usar las ecuaciones de mı́nimo −µ2 = λv 2 , para eliminar un parámetro del
potencial:
V (φ) = − 21 λv 2 φ2 + 41 λφ4

= − 12 λv 2 H 2 + 2vH + v 2 + 41 λ H 4 + 4vH 3 + 6H 2 v 2 + 4v 3 H + v 4
=λv 2 H 2 + λvH 3 + 41 λH 4 + constant. (1.462)
Figure 1.10: Potential for complex scalar field
Podemos escribir el potencial en términos del nuevo campo como
1 1 m2H 3 1 m2H 4
V (H) = m2H H 2 + H + H . (1.463)
2 2 v 8 v2
donde

m2H = 2 µ2 = 2λv 2 (1.464)
Consideremos ahora un campo escalar complejo sin término de masa, pero con potencial:
L = ∂ µ φ∗ ∂µ φ − V (φ) (1.465)
V (φ) = µ2 φ∗ φ + λ(φ∗ φ)2 µ2 < 0, λ > 0 (1.466)
La simetrı́a del Lagrangiano corresponde a U (1) global. Este potencial corresponde al “sombrero
mexicano”, como se ilustra en la Figura 1.10. Para una energı́a suficientemente baja de manera que
el campo deba oscilar alrededor del mı́nimo aparecen dos tipos de excitaciones. Una sobre las paredes
que cuestan energı́a y corresponden a un campo escalar masivo como en el caso anterior, y otra a lo
largo de la circunferencia de mı́nimo, que corresponde a una partı́cula escalar sin masa, y es llamada
bosón del Golstone.
El Lagrangiano escalar complejo es equivalente al Lagragiano de dos campos escalares reales con
los mismos paramétros. Para un conjunto de N campos reales tenedremos (suma sobre i) [12]3 :
1 1 1 2
L = ∂µ φi ∂µ φi − µ2 φi φi − µ2 φi φi , (1.467)
2 2 2
que es invariante bajo una simetrı́a O(N )
i
φi → φ0 = Rij φj , (1.468)
para cualquier matriz N × N ortogonal R. El análisis para N = 2 da lugar a un bosón de Goldstone.

El anális para N > 2 es el mismo y por cada campo real que se introduzca aparece un nuevo bosón
de Goldstone [12]:
[. . . ] there are not continuous symmetries for N = 1, while for N = 2 there is a single
direction of rotation. A rotation in N dimensions can be in any one of N (N −1) planes, so
the O(N )–symmetric theory has N (N − 1)/2 continuous symmetries. After spontaneous
symmetry breaking there are (N − 1)(N − 2)/2 remaining symmetries corresponding to
rotations of the (N − 1) [non massive] fields. The number of broken symmetries is the
difference, N − 1.
Entonces tenemos el siguiente teorema [12]
Goldstone’s theorem states that for every spontaneously broken continuous symmetry, the
theory must contain a massless particle.
Also from [12]4
In a global symmetry that is spontaneously broken the symmetry currents are still con-
served and interactions are similarly restricted [the Lagrangian keeps the symmetry], but
the vacuum state does not respect the symmetry and the particles do not form obvi-
ous symmetry multiplets. Instead, such a theory contains massless particles, Goldstone
bosons, one for each generator of the spontaneously broken symmetry. The third case
is that of a local, or gauge, symmetry. [. . . ] such a symmetry requires the existence of
a massless vector field for each symmetry generator, and the interactions among these
fields are highly restricted.
It is now only natural to consider a fourth possibility: What happens if we include both
local gauge invariance and spontaneous symmetry breaking in the same theory?
3
Sec. 11.1
4
Introduction to Chapter 20
En el caso de la Acción invariante gauge local bajo el Grupo U (1), tenemos el Lagrangiano (??):
L = (Dµ φ)∗ Dµ φ − µ2 φ∗ φ − λ (φ∗ φ)2 − 41 F µν Fµν µ2 < 0 and λ > 0 (1.469)
Para obtener directamente el espectro después de la ruptura espontánea de simetrı́a es conveniente

usar la transformación gauge de la ec. (??). Haciendo θ(x) = −η(x):

0 iθ(x) iη(x) H(x) + v H(x) + v
φ→φ =e e √ = √ (1.470)
2 2
∗ 2
L → L0 = (Dµ )0 φ0 (Dµ )0 φ0 − µ2 (φ∗ )0 φ0 − λ (φ∗ )0 φ0 − 41 (F µν Fµν )0

= 12 ∂ µ H + igA0 (H + v) [∂µ H − igA0 µ (H + v)] − 12 µ2 (H + v)2 − 41 λ(H + v)4 − 41 (F µν Fµν )0 .
µ
(1.471)
En adelante omitiremos las primas, aunque debe estar claro que se esta trabajando en el gauge
especı́fico de la ec. (1.470). Entonces
L = 12 ∂ µ H∂µ H − 12 µ2 (H + v)2 − 41 λ(H + v)4 + 21 g 2 Aµ Aµ (H + v)2 − 41 F µν Fµν . (1.472)
Usando la ec. (1.460)

L = LH + LAµ + 21 g 2 Aµ Aµ H 2 + g 2 vAµ Aµ H, (1.473)
donde LH esta dado por la ec. (1.461) y
LAµ = − 41 F µν Fµν + 12 g 2 v 2 Aµ Aµ . (1.474)
Teniendo en cuenta la ec. (1.221) para el Lagrangiano de Proca, vemos que como consecuencia de la
ruptura espontánea de simetrı́a el campo gauge ha adquirido una masa
mAµ = gv. (1.475)
El mecanismo completo mediante el cual, a partir de un Lagrangiano invariante gauge local, los
bosones gauge adquieren masa se llama mecanismo de Higgs [?]. La partı́cula escalar que adquiere
masa se llama Higgs, mientras que el bosón de Goldstone es absorbido por campo gauge como modo
longitudinal.
El número de grados de libertad independientes en el Lagrangiano original en la ec. (1.469) es
cuatro. Correspondientes a los dos grados de libertad del bosón gauge no masivo y los dos del campo
escalar complejo. En el Lagrangiano final en la ec. (1.473) no aparece el bosón de Goldstone. Sin
embargo esto no es un problema porque dicho Lagrangiano también tiene cuatro grados de libertad
correspondientes a los tres grados de libertad del bosón gauge masivo y al grado de libertad del bosón
de Higgs.
1.14.1 Superconductivity
A review ot the use of the Proca Equations for a massive photon in superconductivity is given in [24].
A popularization review along this lines is in the book of Frank Wilczek “The Lightness of Being”
(see Additional references).
The photon mass inside a superconductor is 10−11 GeV (or 1/1000 of the electron mass accordin
to [24]). Also from the article in Beamline λ ∼ 10 µm y Mγ = ~/λc
There are two important length scales in a superconductor. The first measures how efficientrly
the condensate expels a magnetic field. In fact, the expulsion is not
Additional references:
• The Lightness of Being: Mass, Ether, and the Unification of Forces, Frank Wilczek, http:
//www.amazon.com/The-Lightness-Being-Unification-Forces/dp/0465018955
• http://www.scholarpedia.org/article/Englert-Brout-Higgs-Guralnik-Hagen-Kibble_
mechanism_(history)
• Elementary Particle Physics: Volume 1: Quantum Field Theory and ..., Volume 1 By Yorikiyo
Nagashima, ElementaryParticlePhysics:Volume1:QuantumFieldTheoryand...,Volume1ByYorikiyoNa
• From Superconductors to Supercolliders by LANCE DIXON http://www.slac.stanford.

edu/pubs/beamline/26/1/26-1-dixon.pdf
• Electrodynamics of Superconductors http://www.physics.buffalo.edu/phy514/w11/index.

html
1.15 Fermiones quirales de cuatro componentes

Sea
1 − γ5
PL ≡
2
1 + γ5
PR ≡ . (1.476)
2
Además
ψ L ≡ PL ψ
ψR ≡ PR ψ . (1.477)
1.15. FERMIONES QUIRALES DE CUATRO COMPONENTES 91
Entonces
ψ = ψL + ψR . (1.478)
Las matrices PL,R tienen las propiedades
2
PL + P R = 1 PL,R = PL,R PL,R = PL,R
†
PL P R = 0 PL,R = PL,R . (1.479)
1 ∓ γ5 µ 1 ± γ5
PL,R γ µ = γ = γµ = γ µ PR,L (1.480)
2 2
Para escribir el Lagrangiano en término de los nuevos ψL,R debemos tener en cuenta que
ψL,R = (PL,R ψ)† γ 0 = ψ † PL,R γ 0 = ψ † γ 0 PR,L = ψPR,L (1.481)
L =iψγ µ ∂µ ψ − mψψ
=iψ(PL + PR )γ µ ∂µ ψ − mψ(PL + PR )ψ
=iψPL γ µ ∂µ ψ + iψPR γ µ ∂µ ψ − mψPL ψ − mψPR ψ
=iψPL PL γ µ ∂µ ψ + iψPR PR γ µ ∂µ ψ − mψPL PL ψ − mψPR PR ψ
=iψPL γ µ ∂µ PR ψ + iψPR γ µ ∂µ PL ψ − mψPL PL ψ − mψPR PR ψ
=iψR γ µ ∂µ ψR + iψL γ µ ∂µ ψL − m(ψR ψL + ψL ψR ) . (1.482)
En términos de espinores izquierdos y derechos de cuatro componentes la transformación de paridad
t→t x → −x ψL (t, x) →ψR (t, −x), ψR (t, x) → ψL (t, −x)
∂0 → ∂0 ∇ → −∇ ψL (t, x) →ψR (t, −x), ψR (t, x) → ψL (t, −x) . (1.483)
Además L = r × p → (−r) × (−p) = L, y como γ µ esta asociado al momento angular intrı́nsico,
entonces también γ µ → γ µ
Entonces la transformación de paridad da lugar a (sin tener en cuenta el cambio de argumento
en los campos que desaparece en la integral de la Acción)
ψR γ µ ∂µ ψR = ψR γ 0 ∂0 ψR + ψR γ · ∇ψR →ψL γ 0 ∂0 ψL − ψL γ · ∇ψL
=ψL γ 0 ∂0 ψL + ψL γ † · ∇ψL
=ψL γ 0 γ 0 γ 0 ∂0 ψL + ψL γ 0 γγ 0 · ∇ψL
=ψL γ̃ 0 ∂0 ψL + ψL γ̃ · ∇ψL
=ψR γ̃ µ ∂µ ψR . (1.484)
Entonces
L → L0 =iψR γ̃ µ ∂µ ψR + iψL γ̃ µ ∂µ ψL − m(ψR ψL + ψL ψR ) , (1.485)
donde γ̃ µ = U γ µ U † , con U = γ 0 . Como las dos representaciones dan lugar a la misma fı́sica,
podemos decir que la Acción en términos de espinores L, R de cuatro componentes es invariante bajo
la transformación de paridad.
La existencia de ambos espinores ψL,R garantizan que el Lagrangiano de Dirac es invariante bajo
la transformación de paridad.
La corriente de la electrodinámica cuántica en ec. (1.374) (o la de la cromodinámica cuántica,
ec. (1.418)) conservan paridad ya que, siguiendo los mismos pasos que en la ec. (1.482)
ψγ µ ψ = ψL γ µ ψL + ψR γ µ ψR → ψL γ̃ µ ψL + ψR γ̃ µ ψR . (1.486)
Si para alguna partı́cula, como es el caso del neutrino, no existe la componente derecha, entonces
la correspondiente interacción vectorial viola paridad y no puede tener ni interacciones electro-
magnéticas ni fuertes, es decir, no se acopla con el fotón o los gluones. Además dicha partı́cula
no puede tener masa de Dirac. En el caso del neutrino esto se entiende pues al no tener carga
eléctrica sólo requiere dos grados de libertad independientes.
De otro lado, si una determinada interacción, como es el caso de la interacción débil, solo participa
la componente izquierda de la ec. (1.486), está corresponde a una interacción del tipo
ψ L γ µ ψL = ψPR γ µ PL ψ = ψγ µ PL ψ

µ 1 − γ5
= ψγ ψ
2
= 12 ψ (γ µ − γ µ γ5 ) ψ , (1.487)
que de acuerdo a la asignación en la Tabla corresponde a una corriente V–A.
1.16 Standard model Lagrangian

The known matter is build from the elementary set of particles defined in table
we further define the color triplets of quarks as
   
u1 d1
u = u 2  d = d2  (1.488)
u3 d3
1.16. STANDARD MODEL LAGRANGIAN 93
Type Name Symbol Charge

leptons electron e −1
neutrino ν 0
quarks up quark u1 , u2 , u3 2/3
down quark d1 , d2 , d3 −1/3
Table 1.2: Elementary fermions. The symbol represent both the particle, e.g e− , as the antiparticle,
e.g, e+ . The lectric chage is given in units of the electron chage e
The free Lagrangian containing this particles can be written as
Lfree =ieγ µ ∂µ e − me ee + iνγ µ ∂µ ν + iuγ µ ∂µ u − mu uu + idγ µ ∂µ d − md dd

=ieL γ µ ∂µ eL + ieR γ µ ∂µ eR − me (eR eL + eL eR ) + iνL γ µ ∂µ νL
+ iνR γ µ ∂µ νR + iuL γ µ ∂µ uL + iuR γ µ ∂µ uR
− me (uR uL + uL uR )idL γ µ ∂µ dL + idR γ µ ∂µ dR − me (dR dL + dL dR ) . (1.489)
donde,
νL,R =PL,R ν, eL,R =PL,R e

uL,R =PL,R u, dL,R =PL,R d . (1.490)
Corrientes V–A
En las interacciones débiles sólo participan las partes izquierdas de los campos. Esto nos permite
prescindir del νR , pues no tiene carga eléctrica, fuerte, o débil
Lfree =ieL γ µ ∂µ eL + ieR γ µ ∂µ eR − me (eR eL + eL eR ) + iνL γ µ ∂µ νL

+ iuL γ µ ∂µ uL + iuR γ µ ∂µ uR
− mue(uR uL + uL uR ) + idL γ µ ∂µ dL + idR γ µ ∂µ dR − md (dR dL + dL dR ) . (1.491)
Simetrı́a global SU (3)c × SU (2)L × U (1)Y

En el contexto de las interacciones débiles un eL es completamente equivalente a un campo νL .
Es decir, el Lagrangiano debe ser invariante bajo una transformación SU (2)L de esos campos. La
diferencia entre ellos son sus respectivas cargas electricas y sus masas. Asumiendo que ambos campos
tienen una misma hipercarga, asociada a una nueva simetrı́a Abeliana U (1)Y , podrı́amos esperar que
la corriente electromagnética apropiada pueda obtenerse a partir del Grupo semisimple SU (2)L ×
U (1)Y . Además las respectivas masas se podrı́an obtener a partir del mecanismo de Higgs.
La simetrı́a SU (2)L entre las partes izquierdas del neutrino y el electrón, y entre las partes
izquierdas de los quarks up y down, se establece definiendo los dobletes:

νL u
L≡ Q= L , (1.492)
eL dL
De otro lado, La invarianza bajo U (1)Y requiere que
YL =YνL = YeL
YQ =YuL = YdL . (1.493)
El generador de carga eléctrica Q,b se va obtener a partir de una combinación lineal del generador
diagonal de SU (2)L , T3 , y del generador de hipercarga, Y .
Bajo la simetrı́a SU (2)L , los campos transforman como:
L → L0 = exp(iT i θi )L ≈ (1 + iT i θi )L
Q → Q0 = exp(iT i θi )Q ≈ (1 + iT i θi )Q
eR →e0R = eR
uR →u0R = uR
dR →d0R = dR . (1.494)
donde
τi
Ti = , (1.495)
2
y τ i son las matrices de Pauli dadas en la ec. (1.331).
Claramente el término de masa me en la ec. (??) no es invariante bajo la simetrı́a SU (2)L . El
Lagrangiano en la ec. (??), sin término de masa, puede reescribirse de manera que exhı́ba de forma
más explicita la invarianza bajo SU (2)L como
Lfree =ieL γ µ ∂µ eL + ieR γ µ ∂µ eR + iνL γ µ ∂µ νL

+ iuL γ µ ∂µ uL + iuR γ µ ∂µ uR + idL γ µ ∂µ dL + idR γ µ ∂µ dR
=iLγ µ ∂µ L + iQγ µ ∂µ Q + ieR γ µ ∂µ eR + iuR γ µ ∂µ uR + idR γ µ ∂µ dR . (1.496)
Simetrı́a gauge local SU (3)c × SU (2)L × U (1)Y

Para obtener la interacciones del modelo estándar, reemplazamos las derivadas normales por derivadas
covariantes.
Proponemos entonces el Lagrangiano

L =iQγ µ Dµ Q + iLγ µ Dµ L + ieR γ µ Dµ eR + idR γ µ Dµ dR + iuR γ µ Dµ uR
µν
− 14 Gµν a 1 i 1 µν
a Gµν − 4 Wi Wµν − 4 B Bµν , (1.497)
Donde
λa µ τi
Dµ ≡ ∂ µ − igs Ga − ig Wiµ − ig 0 Y B µ . (1.498)
2 2
donde
λa
Λa ≡ , a = 1, 2, . . . , 8 8 generadores de SU (3)c
2
τi
T i ≡ , i = 1, 2, 3 3 generadores de SU (2)L
2
Y generador de U (1)Y
All the particles in this Lagrangian are massless. It is only good for the gluons and the Abelian
gauge boson, but is no realist for the fermions of the weak gauge bosons Wµi . To solve this problem,
we postulate a new complex scalar doublet with four degree of freedom:
+
φ φ1 + iφ2
Φ= = . (1.499)
φ0 φ3 + iφ4
The “+” and 0 superindexes for just for later convenience. The full Lagrangian involving those fields
are
L =iQγ µ Dµ Q + iLγ µ Dµ L + ieR γ µ Dµ eR + idR γ µ Dµ dR + iuR γ µ Dµ uR
µν
− 14 Gµν a 1 i 1 µν
a Gµν − 4 Wi Wµν − 4 B Bµν
+ (Dµ Φ)† Dµ Φ − µ2 Φ† Φ − λ(Φ† Φ)2

− (he LΦeR + hd QΦdR + hu QΦu e R + h.c)
=Lfermion + Lgauge + LW BH − LYukawa . (1.500)
donde µ2 < 0, y λ > 0,
e = iτ2 Φ∗ ,
Φ (1.501)
Lfermion =iQγ µ Dµ Q + iLγ µ Dµ L + ieR γ µ Dµ eR + idR γ µ Dµ dR + iuR γ µ Dµ uR

µν
Lgauge = − 41 Gµν a 1 i 1 µν
LW BH =(Dµ Φ)† Dµ Φ − µ2 Φ† Φ − λ(Φ† Φ)2

LYukawa =he LΦeR + hd QΦdR + hu QΦue R + h.c (1.502)
Bajo una transformación gauge local las derivadas covariantes de los campos (y por consiguiente
los campos) transforman como:

Dµ L → (Dµ L)0 = exp −iθi T i − iβYL Dµ L

Dµ Q → (Dµ Q)0 = exp −iαa Λa − iθi T i − iβYQ Dµ Q

Dµ Φ → (Dµ Φ)0 = exp −iθi T i − iβYΦ Dµ Φ
Dµ eR → (Dµ eR )0 = exp (−iβYeR ) Dµ eR = exp (−iβQeR ) Dµ eR
Dµ dR → (Dµ dR )0 = exp (−iαa Λa − iβYdR ) Dµ dR = exp (−iαa Λa − iβQdR ) Dµ dR
Dµ uR → (Dµ uR )0 = exp (−iαa Λa − iβYuR ) Dµ uR = exp (−iαa Λa − iβQuR ) Dµ uR . (1.503)
donde QeR = −1, etc, son las cargas eléctricas asociadas a los campos.
Para los campos del Lagrangiano, debemos asegurarnos de que todos los términos invariantes
gauge locales y renormalizables sean considerados. De hecho un término de interacción entre fermiones
y el campo escalar, correspondiente a una interacción de Yukawa: LΦeR y QΦdR son invariantes bajo
transformaciones SU (3)c × SU (2)L × U (1)Y si
−YL + YΦ + QeR =0
−YQ + YΦ + QdR =0
−YQ + YΦe + YuR = −YQ − YΦ + QuR =0 ,
From this set of three equations we obtain the three doublet hypercharges:
1 1 1
YL = − , YΦ = , YQ = . (1.504)
2 2 6
En el análisis anterior hemos fijado YΦe = −YΦ . Esto es debido a que si QΦ es un invariante SU (2)L ,
el término Φ̃† Q también es un invariante de SU (2). Explı́citamente
e † Q =(iτ2 Φ∗ )† Q
Φ
0 ∗ †
φ
= Q
−φ−

0 +
uL
= φ −φ
dL
=φ0 uL − φ+ dL
=12 Q1 Φ2 + 21 Q2 Φ1
=ab Qa Φb . (1.505)
Bajo una transformación SU (2)L
e0 † Q0 = ab Q0a Φ0b =ab Uac Ubd Qc Φd

e †Q → Φ
Φ
=cd det UQc Φd
=cd Qc Φd
e †Q .
=Φ (1.506)
Sin perdida de generalidad los cuatro grados de libertad de Φ, pueden escribirse en la forma

iηj (x)T j 0
Φ =e √1 [H(x) + v]
. (1.507)
2
El potencial escalar, definido por
V (Φ) = µ2 Φ† Φ + λ(Φ† Φ)2 (1.508)
se reduce a
1 1
V (H) = µ2 (H + v)2 + λ(H + v)4 . (1.509)
2 4
1.16.1 Spontaneous symmetry breaking in SU (3)c × SU (2)L × U (1)Y

Retornando al doblete de Higgs del modelo estándar en la ec. (1.507), los cuatro grados de libertad
de Φ, pueden escribirse en forma polar con la parte real neutra desplazada para generar la ruptura
espontánea de la simetrı́a SU (2)L × U (1)Y

iηj T j 0
Φ =e √1 (H(x) + v)
(1.510)
2
√ +
1 1 + iη3 2iη 0
≈ √ −
2 2iη 1 − iη3 √1 (H(x) + v)
2
+ +

1 iη H + viη
= 1
2 √2 (H + v − iη3 H − iη3 v)

G+
= √1 .
2
(H(x) + v − iG0 )
Para SU (2)L ×U (1)Y tenemos cuatro generadores y cuatro bosones gauge. De acuerdo a la parametrización
en ec. (1.510) esperamos que aparezcan tres bosones de Goldstone y un campo de Higgs con masa,
de manera que quedará un generador no roto correspondiente a una simetrı́a remanente del vacı́o
U (1)Q
hΦi
SU (2)L × U (1)Y −→ U (1)Q . (1.511)
Se espera entonces que el espectro consista de un bosón de Higgs, tres bosones gauge masivos, y
un bosón gauge sin masa.
Podemos hacer una transformación gauge similar a la de la ec. (??) sobre el campo Φ, tal que

0 0
Φ → Φ = √1 , (1.512)
2
(H(x) + v)
que define el gauge unitario. En adelante sin embargo omitiremos las primas sobre los campos
transformados Φ0 y Wµν
0
.
Comenzaremos analizando la parte escalar del Lagrangiano del Modelo dada en la ec. (1.502)
†
1 µ 0 0
LW BH = D Dµ − V (H) , (1.513)
2 H(x) + v H(x) + v
donde V (H) dado en la ec. (1.463), incluye el término de masa para el bosón de Higgs (1.464):

m2H = 2 µ2 = 2λv 2 (1.514)
Como

10 1 1 1 0 −i 2 1 1 0
Wµ = Ti Wµi = Wµ + Wµ + Wµ3
21 0 2 i 0 2 0 −1
 
3
√ Wµ1 − i Wµ2
Wµ 2 √
1 2 
=  
2 √ Wµ + i Wµ 
1 2
√ 3
2 −Wµ
2
3
√
1
√ Wµ 2Wµ+
≡
2 2Wµ− −Wµ3
!
1 √1 W +
W3 µ
= √12 − 2
1 3 . (1.515)
2
W µ − 2
W µ
Dµ corresponde a la matrix 2 × 2
!
1
∂µ − i gWµ3 + g 0 Y Bµ − √i2 gWµ+
Dµ = 2 . (1.516)
− √i2 gWµ− ∂µ − i − 12 gWµ3 + g 0 Y Bµ
Entonces

− √i2 gWµ+ (H + v)
Dµ Φ = . (1.517)
∂µ H − i − 21 gWµ3 + g 0 YΦ Bµ (H + v)
De modo que
†
1 − √i2 gW µ + (H + v)
LW BH = ·
2 ∂ µ H − i − 21 gW3µ + g 0 YΦ B µ (H + v)

− √i2 gWµ+ (H + v)
− V (H)
∂µ H − i − 21 gWµ3 + g 0 YΦ Bµ (H + v)
1 √i µ− µ 1 µ 0 µ

= gW (H + v) ∂ H + i − 2
gW 3 + g Y Φ B (H + v) ·
2 2
− √i2 gWµ+ (H + v)
− V (H)
∂µ H − i − 21 gWµ3 + g 0 YΦ Bµ (H + v)
1
= g 2 W µ− Wµ+ (H + v)2 − V (H)
4
1
+ ∂ µ H + i − 21 gW3µ + g 0 YΦ B µ (H + v) ×
2
∂µ H − i − 21 gWµ3 + g 0 YΦ Bµ (H + v)
1
= − V (H) + g 2 W µ− Wµ+ (H + v)2 +
4
1 µ 1 2
+ ∂ H∂µ H + − 21 gW3µ + g 0 YΦ B µ (H + v)2 (1.518)
2 2
donde la última lı́nea corresponde a la magnitud del “número” complejo:

∂µ H − i − 21 gWµ3 + g 0 YΦ Bµ (H + v) (1.519)
Entonces
1
LW BH = ∂ µ H∂µ H − V (H)
2
gv 2 1 1
+ W µ− Wµ+ + g 2 W µ− Wµ+ H 2 + vg 2 W µ− Wµ+ H + LZAH , (1.520)
4 4 2
donde
1 1 2 µ 3 1 0 µ 1 0 µ 02 2 µ

LZAH = g W 3 W µ − gg Y W
Φ 3 µB − gg Y W
Φ 3 µB + g Y Φ B Bµ ×
2 4 2

2
H 2 + 2vH + v 2 (1.521)
Haciendo YΦ = 1/2 como en la ec. (1.504),
3
1 µ µ
g 2 −gg 0 Wµ 2 2

LZAH = W3 B 2 H + 2vH + v (1.522)
8 −gg 0 g 0 Bµ
Sea

cos θW sin θW 1 g g0
V = =p , (1.523)
− sin θW cos θW g2 + g02 −g 0 g
con tan θW = g 0 /g, tal que g sin θW = g 0 cos θW , como en la ec. (??). Note que V es una matrix
ortogonal que satisface V V T = V T V = 1. Si (ver ec. (??)),
3
Wµ3 Zµ Zµ T Wµ
=V ó =V (1.524)
Bµ Aµ Aµ Bµ
entonces
2 3
1 3µ µ
T g −gg 0 T Wµ 2 2

LZAH = W B VV 2 V V H + 2vH + v
8 −gg 0 g 0 Bµ
2
1 g −gg 0
Zµ
= Z µ Aµ V T 0 02 V H 2 + 2vH + v 2 (1.525)
8 −gg g Aµ
3
T g 2 −gg 0 1 g + gg 0 2 −g 2 g 0 − g 0 3 g g0
V V = 2
−gg 0 g 0 2 2
g + g 0 2 +g 2 g 0 − g 2 g 0 −gg 0 + gg 0
2
−g 0 g
3
1 g + gg 0 2 −g 2 g 0 − g 0 3 g g0
= 2
g + g02 0 0 −g 0 g
4
1 g + g 2 g 0 2 + g 2 g 0 2 + g 0 4 g 3 g 0 + gg 0 3 − g 3 g 0 − gg 0 3
= 2
g + g02 0 0
2 02

g +g 0
= (1.526)
0 0

1 g2 + g02
LZAH = Z µ Zµ H 2 + 2vH + v 2
2 4

1 g 2
= 1 + tan2 θW Z µ Zµ H 2 + 2vH + v 2
2 2
2
1 g
= Z µ Zµ H 2 + 2vH + v 2
2 2 cos θW
2 2
1 gv 1 g
= µ
Z Zµ + Z µ Zµ H 2
2 2 cos θW 2 2 cos θW
2
g
+ vZ µ Zµ H (1.527)
2 cos θW
Retornando a la ec. (1.520), tenemos tenemos
2
LW BH = (Dµ Φ)† Dµ Φ − µ2 Φ† Φ − λ Φ† Φ
1
= ∂ µ H∂µ H − V (H)
2
1 1
+ g 2 W µ− Wµ+ H 2 + vg 2 W µ− Wµ+ H
4 2
2 2
1 g µ 2 g
+ Z Zµ H + v Z µ Zµ H
2 2 cos θW 2 cos θW
1 2 1 1
+ mW W µ− Wµ+ + m2W W µ− Wµ+ + m2Z Z µ Zµ , (1.528)
2 2 2
donde:
• Masas gauge:
gv gv
mW = mZ = , (1.529)
2 2 cos θW
y
mW
mZ = . (1.530)
cos θW
•
V (H) = 21 m2H H 2 + λvH 3 + 41 λH 4
1 m2 1 m2H 4
= m2H H 2 + H H 3 + 2
H
2 2v 4 2v

1 H H2
= m2H H 2 1 + + 2 . (1.531)
2 v 4v
con
m2H = −2µ2 = 2λv 2 . (1.532)
• 3
Wµ cos θW sin θW Zµ
= , (1.533)
Bµ − sin θW cos θW Aµ
tal que
g sin θW = g 0 cos θW . (1.534)
1.16.2 Yukawa Lagrangian

In the Unitary gauge
LYukawa =he LΦeR + hd QΦdR + hu QΦu e R + h.c

1
= √ he (eL eR + eR eL ) + hd (dL dR + dR dL ) + hu (uL uR + uR uL ) ×
2
[H(x) + v]

v H(x)
= √ he ee + hd dd + hu uu +1 , (1.535)
2 v
definiendo
hf v
mf = √ (1.536)
2
tenemos
me md mu
LYukawa =me ee + md dd + mu uu + eeH + ddH + uuH . (1.537)
v v v
1.16.3 Fermion-gauge interactions

De la ec. (1.502) tenemos
Lfermion =iQγ µ Dµ Q + iLγ µ Dµ L + ieR γ µ Dµ eR + idR γ µ Dµ dR + iuR γ µ Dµ uR . (1.538)

Los términos de interacción generados por la simetrı́a gauge para el campo L son:
iLγ µ Dµ L − iLγ µ ∂µ L =iLγ µ (−igTi Wµi − ig 0 YL Bµ )L
=Lγ µ (gT1 Wµ1 + gT2 Wµ2 + gT3 Wµ3 + g 0 YL Bµ )L

µ g 0 Wµ+ 3 0
=Lγ √ − + gT3 Wµ + g YL Bµ L
2 Wµ 0

µ g 0 Wµ+ µ
3 0

=iLγ √ − L + Lγ gT3 W µ + g YL Bµ L
2 Wµ 0

µ g eL Wµ+
=Lγ √ − + LAZL
2 νL W µ
g
= √ νL γ µ eL Wµ+ + eL γ µ νL Wµ− + LAZL
2
=LW L + LAZL , (1.539)
donde
g
LW L = √ νL γ µ eL Wµ+ + eL γ µ νL Wµ−
2

LAZL =Lγ µ gT3 Wµ3 + g 0 YL Bµ L (1.540)
Generalizando para todos los campos:
g
LW L → √ νL γ µ eL Wµ+ + uL γ µ dL Wµ+ + h.c . (1.541)
2
LAZL =Lγ µ [gT3 (cW Zµ + sW Aµ ) + g 0 YL (−sW Zµ + cW Aµ )] L
=Lγ µ [gT3 cW Zµ + gT3 sW Aµ − g 0 YL sW Zµ + g 0 YL cW Aµ ] L
=Lγ µ [(gcW T3 − g 0 sW YL ) Zµ + (gsW T3 + g 0 cW YL ) Aµ ] L , (1.542)
donde cW = cos θW , sW = sin θW . Usando la relación entre g y g 0 (1.534):

µ s2W
LAZL =Lγ gcW T3 − g YL Zµ + (gsW T3 + gsW YL ) Aµ L
cW
=gsW Lγ µ [(cot θW T3 − tan θW YL ) Zµ + (T3 + YL ) Aµ ] L . (1.543)
Como el generador asociado a Aµ debe ser el generador de carga eléctrica, tenemos que
e = g sin θW (1.544)
donde e es la carga eléctrica del electrón, y el generador de carga

b = T3 + Y ,
Q (1.545)
de modo que

b = (T3 + Y )L = Qν 0 0 0
QL L= L. (1.546)
0 Qe 0 −1
La ec. (1.545), se conoce como la relación Gell-Mann–Nishijima, y establece la condición que se debe
satisfacer para obtener apropiadamente la QED a partir de la interacción electrodébil asociada al
grupo semisimple SU (2)L × U (1)Y . De esta forma
LAZL =eLγ µ (cot θW T3 − tan θW YL ) LZµ + eLγ µ Q bL LAµ
h i
=eLγ µ cot θW T3 − tan θW Q bL − T3 LZµ + eLγ µ Q bL LAµ
e h i
= bL LZµ + eLγ µ Q
Lγ µ τ3 − 2s2W Q bL LAµ . (1.547)
2cW sW
Generalizando para los otros campos, tenemos
X e h i
LAZL → bL F Zµ + eF γ µ Q
F γ µ τ3 − 2s2W Q b L F Aµ . (1.548)
F =Q,L,e ,d ,u
2c W sW
R R R
Usando los acoplamientos gauge de los quarks con los gluones (1.413), de los fermiones con el
Wµ± (1.541) y con Zµ y Aµ (1.548) para expandir Lfermion en (1.538), tenemos
Lfermion =iQγ µ Dµ Q + iLγ µ Dµ L + ieR γ µ Dµ eR + idR γ µ Dµ dR + iuR γ µ Dµ uR
=iuL γ µ ∂µ uL + iuR γ µ ∂µ uR + idL γ µ ∂µ dL + idR γ µ ∂µ dR
+ ieL γ µ ∂µ eL + ieR γ µ ∂µ eR + iνL γ µ ∂µ νL

λa λa λa λa
+ gs uL γµ uL + uR γµ uR + dL γµ dL + dR γµ dR Gµa
2 2 2 2
g
+ √ νL γ µ eL Wµ+ + uL γ µ dL Wµ+ + h.c
2
X e h i
+ F γ µ τ3 − 2s2W Q b L F Zµ
F =Q,L,eR ,dR ,uR
2cW sW

+ e eL γ µ Q be eL + eR γ µ Q
b e eR

+uL γ µ Qbu uL + uR γ µ Q b u u R + dL γ µ Q
b d dL + dR γ µ Q
bd dR Aµ . (1.549)
Para escribir este Lagrangiano en terminos de espinores de 4 componentes, tomemos algunos
casos especı́ficos:
u d νe e
2vf 1 − sin2 θW
8
3
−1 + sin2 θW
4
3
1 −1 + 4 sin2 θW
2af 1 −1 1 −1
Table 1.3: Acoplamientos de corrientes neutras
νL γ µ eL Wµ+ = νPR γ µ PL eWµ+

= νγ µ PL2 eWµ+
= νγ µ PL eWµ+
1
= νγ µ (1 − γ5 )eWµ+ , (1.550)
2
•
h i
µ 2 b 2 µb 2 µb
Qγ τ3 − 2sW QQ Q − 2sW uR γ Qu uR − 2sW dR γ Qd dR Zµ
" !
µ 1 − 2s2W Q bu 0 uL
= u L dL γ 2 b
0 −1 − 2sW Qd dL
i
bu uR − 2s2 dR γ µ Q
−2s2W uR γ µ Q bd dR Zµ
W

= uL γ uL − dL γ dL − 2s2W (uL + uR ) γ µ Qu (uL + uR ) + dL + dR γ µ Qd (dL + dR ) Zµ
µ µ

1 µ 1 µ 2 µ µ

= uγ (1 − γ5 )u − dγ (1 − γ5 )d − 2sW uγ Qu u + dγ Qd d Zµ
2 2

µ 1 2 1 µ 1 2 1
= uγ − 2sW Qu − γ5 u + dγ − − 2sW Qd + γ5 d Zµ
2 2 2 2
µ µ

= uγ (vu − au γ5 ) u + dγ (vd − ad γ5 ) d Zµ , (1.551)
donde
vf =T3f − 2 sin2 θW Qf af =T3f γ5 (1.552)
Los valores explı́citos para vf y af en el modelo estándar, están dados en la Tabla 1.3.
Usando las expresiones para pasar de fermiones L, R a los fermiones de Dirac de cuatro compo-
mentes, y las ecuaciones (1.550), (1.551) tenemos
Lfermion =iuγ µ ∂µ u + idγ µ ∂µ d + ieγ µ ∂µ e + iνL γ µ ∂µ νL

λa λa
+ gs uγµ u + dγµ d+ Gµa
2 2
g µ X e
+ √ νγ (1 − γ5 )eWµ+ + uγ µ (1 − γ5 )dWµ+ + h.c + f γ µ (vf − af γ5 ) f
2 2 f =u,d,ν,e
2c s
W W

+ e eγ µ Q be e + uγ µ Qu u + dγ µ Qd d Aµ
X X λa
= if γ µ ∂µ f + gs qγµ qGµa
f =u,d,ν,e q=u,d
2
g
+ √ νγ µ (1 − γ5 )eWµ+ + uγ µ (1 − γ5 )dWµ+ + h.c
2 2
X e
+ f γ µ (vf − af γ5 ) f
f =u,d,ν,e
2c s
W W
X
+e f γ µQbf f Aµ , (1.553)
f =u,d,ν,e
donde Qf están dadas en la Tabla 1.2 y vf , af en la Tabla 1.3.
1.16.4 Self-interactions
El Lagrangiano gauge
µν
Lgauge = − 14 Gµν a 1 i 1 µν
(1.554)
se puede escribir como
Lgauge = − 14 F µν Fµν − 14 Z µν Zµν − 21 (FW † µν

) (FW )µν − 14 G eµν ea
a Gµν
h i
† µν + µν − − + µν
− ie cot θW (FW ) Wµ Zν − (FW ) Wµ Zν + Wµ Wν Z
h i
† µν
− ie (FW ) Wµ+ Aν − (FW )µν Wµ− Aν + Wµ− Wν+ F µν
e2 h i
µ− 2 − ν−
− 2 W +
µ W − W +
µ W µ+
W ν W − e2 cot2 θW Wµ+ W µ− Zν Z ν − Wµ+ Z µ Wν− Z ν
2 sin θW

− e2 cot2 θW 2Wµ+ W µ− Aν Z ν − Wµ+ Aµ Wν− Z ν − Wµ+ Z µ Wν− Aν

− e2 Wµ+ W µ− Aν Aν − Wµ+ Aµ Wν− Aν
1 eµν
eaµν + gs2 f abc fade Gµ Gνc Gdµ Geν ,
− gs Ga fade Gdµ Geν + gs f abc Gµb Gνc G b (1.555)
4
donde
(FW )µν =∂µ Wν+ − ∂ν Wµ+

eµν =∂ µ Gν − ∂ ν Gµ .
G (1.556)
a
1.16.5 Lagrangiano del modelo estándar para la primera generación

Recopilando los resultados para LW BH (1.528), LYukawa (1.537), Lfermion (1.553), y Lgauge (1.555),
tenemos para f = νe , e, u, d; q = u, d [con f 0 = e (d) para f = νe (u) ]
X
L1 gen = if¯ (γ µ ∂µ − mf ) f
f
− 41 F µν Fµν − 14 Z µν Zµν − 21 (FW † µν

) (FW )µν − 14 G eµν G
ea
a µν
2

1 H H
+ 12 ∂ µ H∂µ H − m2H H 2 1 + + 2
2 v 4v

2 µ− + 1 2 µ H H2
+ mW W Wµ + mZ Z Zµ 1+2 + 2
2 v v
X X
λa
+ gs q̄γ µ q Gaµ + e f¯γ µ Qf f Aµ
q
2 f
e X
+ f¯γ µ (vf − af γ5 )f Zµ
2 cos θW sin θW f
" #
g X X mf
+ √ f¯γ µ (1 − γ5 )f 0 Wµ+ + h.c + f¯f H
2 2 f f
v
h i
† µν + µν − − + µν
− ie cot θW (FW ) Wµ Zν − (FW ) Wµ Zν + Wµ Wν Z
h i
† µν
− ie (FW ) Wµ+ Aν − (FW )µν Wµ− Aν + Wµ− Wν+ F µν
e2 h i
+ µ− 2 + µ+ − ν−
− W µ W − W µ W W ν W
2 sin2 θW

− e2 cot2 θW Wµ+ W µ− Zν Z ν − Wµ+ Z µ Wν− Z ν


1 eµν
ea + g 2 f abc fade Gµ Gν Gd Ge .
− gs Ga fade Gdµ Geν + gs f abc Gµb Gνc G µν s b c µ ν (1.557)
4
1.16.6 Dinámica de sabor

El Modelo Estándar esta compuesto de las siguientes tres familias de fermiones i = 1, 2, 3. A cada
familia se le asigna una carga de sabor diferente
i e µ τ
νL νL νL ν
Li = i : L1 = L2 = L3 = L eiR : eR , µR , τR
eL eL µL τL
iα α α α
α uL α uL α cL α t
Qi = iα : Q1 = α Q2 = α Q3 = Lα uiR : uR , cR , tR
dL dL sL bL
diR : dR , sR , bR . (1.558)
Con
1 1 2 1
YLi = − YQ i = YeiR = − 1 YuiR = YdiR = − . (1.559)
2 6 3 3
De los procesos entre familias, es decir de cambio de sabor, sabemos que
• No se han observado procesos de corrientes neutras que cambian sabor.
• Los bosones gauge cargados Wµ± decaen siempre a leptones de la misma generación y con la
misma intensidad.
Proponemos entonces el Lagrangiano

X 0 0 µ 0i µ

µ 0 0 0i µ 0i 0i 0i µ 0i
L =i Qi γ Dµ Qi + Li γ Dµ Li + eR γ Dµ eR + dR γ Dµ dR + uR γ Dµ uR
i
0 0 D 0 0 U 0e 0
− (hE
ij Li ΦeR j + hij Qi ΦdR j + hij Qi ΦuR j + h.c)
− 41 Wiµν Wµν
i
− 14 B µν Bµν
+ (Dµ Φ)† Dµ Φ − µ2 Φ† Φ + λ(Φ† Φ)2 . (1.560)
Para aclarar la notación, obviando de momento la definición definitiva de hij y las primas sobre los
campos, consideremos el Lagrangiano de Yukawa para el sector down
†
L ⊃hD
ij dRi Φ Qj + h.c
⊃hD ea b
ij dRi ab Φ Qj + h.c
α a b
⊃hD e
ij ab dR i Φ Qjα + h.c
α ηρ a b
⊃hD † e
ij ab (d ) γ0 Φ Qjαρ + h.c , (1.561)
Rη i
donde i, a, α, η son ı́ndices en los espacios de familia, SU (2)L , SU (3)c y de Dirac, respectivamente.
Por ejemplo el primer termino de la sumatoria
† ηρ e 1 21
L ⊃hD
11 (dRη )1 γ0 Φ Q11ρ + . . .
0∗ r r
⊃hD
11 dR φ dL + . . . (1.562)
corresponde a la interacción de Yukawa del quark down rojo (r) con un campo escalar complejo
neutro en carga eléctrica pero de isospı́n débil 1/2. En forma compacta la primera expresión en la
ec. (1.561) puede escribirse como (en el gauge unitario)
†
L ⊃dR hD Φ† Q + QL ΦhD dR
u
D H(x)+v L
⊃ dR h 0 √ + h.c
2 dL

D H(x) + v
⊃ dR h √ dL + h.c
2
hD hD v
⊃ dR √ H(x)dL + dR √ dL + h.c
2 2
D
h
⊃ dR √ H(x)dL + dR MD dL + h.c . (1.563)
2
La matrix 3 × 3 M3 es en general una matriz compleja no diagonal, la cual se debe diagonalizar con
una transformación biunitaria (de similaridad). Retornado a la ec. (1.560), tenemos que para definir
apropiadamente la masa de los quarks, rotamos de los autoestados de interacción a los autoestados
de masa con la matrices unitarias
0 †
dR,L 0j = (VR,L
D
)jk dR,L k D
dR,L j = dR,L k (VR,L )kj (1.564)
Tal que
D † D †
(VR,L )ij (VR,L )jk = δik (VLD )ki MijD (VRD )jl = mD
k δkl (1.565)

† D† † D† D† D
VLD D
M M VLD = VLD MD VR
D
V R M VL
†
=MD D
diag Mdiag
2
= MDdiag , (1.566)
donde MD D D D
diag = diag(m1 , m2 , m3 ). Similarmente
2
D† †
VR MD MD VRD
= MD
diag . (1.567)
Con definiciones similares para los campos ui y ei .

D
0 hij v
LYukawa ⊃dL i √ dR 0j
2
0
=dL i MijD dR 0j
†
=dLk (VLD )ki MijD (VRD )jl dRl
=dLk mD
k δkl dRl
=mD
k dLk dRk (1.568)
Para las diferentes combinaciones de términos de corrientes
†
uL 0i γ µ dL 0i =uLk γ µ (VLU )ki (VLD )il dLl
=Vkl uLk γ µ dLl
νL 0i γ µ eL 0i =νL 0i γ µ (VLE )ij eLj
=νL 0i (VLE )ij γ µ eLj
=νLj γ µ eLj (1.569)
Donde hemos definido la matriz de Cabibbo–Kobayashi–Maskawa (CKM) como
†
V =VLU VLD
† †
X X X X
V † V =VLD VLU VLU VLD = 1 ⇒ Vij† Vjk = δik ⇒ Vji∗ Vjk = δik ⇒ |Vji |2 = |Vij |2 = 1
j j j j
(1.570)
y los autoestados débiles de los neutrinos como
†
νL 0i = (VLE )ij νLj (1.571)
Con esta definición, las corrientes débiles cargadas para los leptones siguen siendo universales. Sim-
ilarmente
†
uL 0i γ µ uL 0i =uLk γ µ (VLU )ki (VLU )il uLl
=δkl uLk γ µ uLl
=uLk γ µ uLk (1.572)
De modo que todas las corrientes neutras permanecen universales después de la redefinición de los
campos fermiónicos. A éste resultado, basado en la unitariedad de las transformaciones biunitarias se
le llama Mecanismo GIM. En muchas extensiones del Modelo Estándar las matrices que transforman
los fermiones a sus autoestados de masa no son unitarias y dan lugar a corrientes débiles neutras que
cambian sabor (FCNC de sus siglas en inglés).
Teniendo en cuenta estos resultados podemos escribir finalmente el Lagrangiano completo del
Modelo Estándar en la Gauge Unitario, para
f =νe , νµ , ντ , e, µ, τ, u, c, t, d, s, b; q =u, c, t, d, s, b; l =e, µ, τ (1.573)
X
LSM = if¯ (γ µ ∂µ − mf ) f
f
− 41 F µν Fµν − 14 Z µν Zµν − 12 (FW † µν

) (FW )µν − 41 G eµν ea
a Gµν

1 H H2
+ 12 ∂ µ H∂µ H − m2H H 2 1 + + 2
2 v 4v

2 µ− + 1 2 µ H H2
+ mW W Wµ + mZ Z Zµ 1+2 + 2
2 v v
X X
λa
+ gs q̄γ µ q Gaµ + e f¯γ µ Qf f Aµ
q
2 f
e X
+ f¯γ µ (vf − af γ5 )f Zµ
2 cos θW sin θW f
" τ #
g X X
µ + µ +
+ √ ν̄l γ (1 − γ5 )lWµ + Vij ūi γ (1 − γ5 )dj Wµ + h.c
2 2 l=e ij
X mf
+ f¯f H
f
v
h i
† µν
− ie cot θW (FW ) Wµ+ Zν − (FW )µν Wµ− Zν + Wµ− Wν+ Z µν
h i
† µν + µν − − + µν
− ie (FW ) Wµ Aν − (FW ) Wµ Aν + Wµ Wν F
e2 h i
+ µ− 2 + µ+ − ν− 2 2 + µ− ν + µ − ν
− W µ W − W µ W W ν W − e cot θW W µ W Zν Z − Wµ Z W ν Z
2 sin2 θW


1 eµν d e abc µ ν e a 2 abc µ ν d e

− gs Ga fade Gµ Gν + gs f Gb Gc Gµν + gs f fade Gb Gc Gµ Gν . (1.574)
4
donde mνl = 0.
1.17 Fenomenologı́a Electrodébil

El Lagrangiano del Modelo contiene los parámetros gs , g, sin θW , v, mH . Alternativamente uno puede
escoger como parámetros, en lugar de g, sin θW , v [?]
GF = 1.166 371(6) × 10−5 GeV−2

α−1 = 137.035 999 679(94)
mZ = 91.1876(20) GeV
αs (mZ ) = 0.1176(20) . (1.575)
donde αi = gi2 /(4π). Esto tiene la ventaja de usar las tres cantidades experimentales mejor medidas.
Las relaciones
m2W πα
sin2 θW =1 − , m2W sin2 θW = √ (1.576)
m2Z 2GF
determinan entonces
sin2 θW =0.212
mW =80.94 GeV (1.577)
Si se usa α(MZ ) ≈ 1/128 entonces
sin2 θW =0.233
mW =79.84 GeV (1.578)
Los valores medidos son sin2 θW = 0.23149(13), mW = 80.398(25) GeV, y pueden ser reproducidos por
el modelo estándar una vez se tienen en cuenta correcciones perturbativas inducidas por partı́culas
virtuales.
El acelerador e+ e− LEP, que funcionó hasta desde 1998 hasta el 2000 [?], operó a energı́as su-
ficientes para producir millones de Z. Combinado con otros resultados experimentales, se pudo
verificar todo el Lagrangiano del Modelo Estándar hasta un nivel del 1 por mil. Con excepción de
las interacciones asociadas con el Higgs.
La universalidad de los decaimientos del Z está soportada por los resultados experimentales
siguientes donde sólo se muestran los decaimientos leptónicos del Z diferentes de cero [?]
Γ(Z → e+ e− ) = 83.92(12) MeV Γ(Z → µ+ µ− ) = 83.99(18) MeV Γ(Z → τ + τ − ) = 84.08(22) MeV

Br(Z → e+ e− ) = 3.363(4)%, Br(Z → µ+ µ− ) = 3.366(7)%, Br(Z → τ + τ − ) = 3.370(8)%
(1.579)
1.17. FENOMENOLOGÍA ELECTRODÉBIL 115
Mientras que para el W ± , en %,
Br(W − → ν̄e e− ) = 10.65(17), Br(W − → ν̄µ µ− ) = 10.59(15), Br(W − → ν̄τ τ − ) = 11.44(22)

(1.580)
La diferencia de ν̄τ τ respecto a los otros representa un efecto a 2.8σ. La universalidad de los
acoplamientos leptónicos de W puede comprobarse también indirectamente a través de los decaimien-
tos débiles mediados por corrientes cargadas. Los datos actuales verifican la universalidad de los
acoplamientos de corrientes cargadas leptónicas al nivel del 0.2% [?]. Sin necesidad de entrar en
detalles de los cálculos de las amplitudes de decaimiento, podemos usar el hecho de que ellas son
proporcionales a los acoplamientos al cuadrado correspondiente, de modo que un cociente entre am-
plitudes de decaimiento es igual, en primera aproximación, a los cocientes de los acoplamientos al
cuadrado. Tendremos en cuenta además que el Branching es la amplitud de decaimiento a un canal
especifico divido por la suma de las amplitudes de decaimiento a todos los canales posibles.
Para los decaimientos del Z el Modelo Estándar predice, además de la ausencia de eventos del
tipo Z → e+ µ− , que para un cierto l = e, µ, τ , o q = d, s, b
Br(Z → l+ l− ) (|vl |2 + |al |2 )

≈
Br(Z → q̄q) Nc (|vq |2 + |aq |2 )
h 2 1 i
1 2
− 2 + 2 sin θW + 4
= h 2 i
Nc − 21 + 23 sin2 θW + 14

0.338 Nc = 2
0.776 
≈ = 0.225 Nc = 3 (1.581)
Nc 

0.169 Nc = 4
Para ser comparado con el resultado experimental de por ejemplo
Br(Z → e+ e− ) 3.363(4)
= ≈ 0.222 (1.582)
Br(Z → b̄b) 15.12(5)
que de nuevo da lugar al Nc = 3, que seguiremos tomando en adelante.

Los Branchings de decaimiento en la ec. (1.580) y ec. (1.579) pueden ser calculados sin entrar en
detalles del cálculo de las amplitudes. Teniendo en cuenta que el canal Z → t̄t esta cerrado
Γ(Z → e+ e− )
Br(Z → e+ e− ) =
Γtotal
(|ve |2 + |ae |2 )
=P P2 P3
|2 + |a |2 ) + (|v |2 + |a |2 )] + N [ |2 + |a |2 ) + 2 2
l [(|vl l νl ν l c i=1 (|v ui ui i=1 (|vdi | + |adi | )]
(|ve |2 + |ae |2 )
=
3[(|ve |2 + |ae |2 ) + (|vνe |2 + |aνe |2 )] + 3[2(|vu |2 + |au |2 ) + 3(|vd |2 + |ad |2 )]
(|ve |2 + |ae |2 )
=
21|ae |2 + 3[|ve |2 + |vνe |2 ] + 3[2|vu |2 + 3|vd |2 ]
(−1 + 4s2 θW )2 + 1
=
21 + 3[(−1 + 4s2 θW )2 + 1] + 3[2(1 − 83 s2 θW )2 + 3(−1 + 43 s2 θW )2 ]
2 − 8s2 θW + 16s4 θW
=
42 − 80s2 θW + 320 3
s4 θ W
≈ 3.43% (1.583)
Para W ± tenemos por ejemplo
Γ(W − → ν̄e e− )
Br(W − → ν̄e e− ) = (1.584)
Γtotal
donde, teniendo en cuenta que los canales a top están cerrados, y usando la condición de unitariedad
de la matriz CKM en ec. (1.570), tenemos
X X
Γtotal = Γ(W − → ν̄l l− ) + Nc [Γ(W − → ū1 di ) + Γ(W − → ū2 di )]
l i
X
− −
=Γ(W → ν̄e e ){3 + Nc [|V1i |2 + |V1i |2 ]}
i
− −
=Γ(W → ν̄e e )(3 + 2Nc )
(1.585)
entonces
1
Br(W − → ν̄e e− ) = = 11.1% (1.586)
3 + 2Nc
Una mejor predicción de dichos resultados en el contexto del Modelo Estándar requiere tener en
cuenta las correcciones radiativas.
El ME también tiene una predicción concreta para la amplitud del Z a neutrinos, Γinv :
P
Γinv Γ(Z → ν̄l νl )
= l
Γl Γ(Z → e+ e− )
Nν Γ(Z → ν̄e νe )
=
Γ(Z → e+ e− )
Nν (|vνe |2 + |aνe |2 )
≈
|ve |2 + |ae |2
2Nν
=
(−1 + 4 sin2 θW )2 + 1
(
5.865 Nν = 3
≈ , (1.587)
7.819 Nν = 4
mientras que el valor medido experimentalmente para esta cantidad 5.942(16) [?], es una evidencia
muy fuerte de que sólo exiten tres neutrinos livianos.
1.17.1 Decaimientos débiles mediados por corrientes cargadas

De la corrientes cargadas para leptones tenemos
" #
g X
Lcc ⊃ √ ν̄l γ µ (1 − γ5 )lWµ+ + ¯lγ µ (1 − γ5 )νl Wµ− (1.588)
2 2 l
Esto da lugar a los posibles diagramas para decaimientos de leptones a bosones virtuales, y bosones
a leptontes mostrados en la figura 1.11. Las flechas representan el flujo de número leptónico. La
flecha de tiempo es de izquierda a derecha. Al lado izquierdo del vértice entran partı́culas y salen
antipartı́culas. Mientras que al lado derecho entran antip artı́culas y salen partı́culas Del primer y
cuarto diagrama obtenemos el diagrama de Feynman para el decaimiento µ− → νµ e− ν̄e , mostrado
en la figura 1.12 El propagador para el bosón W de momentum q resulta ser

e µν = 1 q µ q ν
D gµν − 2 . (1.589)
q 2 − m2W mW
Para los propósitos actuales la obtención de este resultado no es necesaria, el punto importante es
que cuando los momentum de las partı́culas iniciales y finales son mucho más pequeñas que mW ,
esto se reduce a
De µν = − gµν . (1.590)
m2W
Figure 1.11: Diagramas de Feynman para las corrientes cargadas
Figure 1.12: diagrama de Feynman para el decaimiento µ− → νµ e− ν̄e

Este resultado se entiende fácilmente cuando se compara con el propagador de una partı́culas escalar
masiva 1/(q 2 − M 2 ) → −1/M 2 . Las componentes espaciales de Wµ con µ = 1, 2, 3, a bajas energı́as
tienen el mismo propagador que el de una partı́cula escalar, mientras W0 , tiene el signo opuesto.
El Lagrangiano efectivo para el decaimiento del muón, µ− → νµ e− ν̄e es entonces
g2 gµν
L= [ν̄µ γ µ (1 − γ5 )µ] 2 [ēγ ν (1 − γ5 )νe ]
8 mW
2
g
= 2 [ν̄µ γ µ (1 − γ5 )µ] [ēγ ν (1 − γ5 )νe ]
8mW
GF
= √ [ν̄µ γ µ (1 − γ5 )µ] [ēγµ (1 − γ5 )νe ] , (1.591)
2
donde
GF g2
√ = 2
2 8mW
g24
= 2 2
8g v
1
= 2, (1.592)
2v
y
√ −1/2
v= 2GF . (1.593)
De otro lado, para el decaimiento β, n → pe− ν̄e , de acuerdo a la figura 1.13, tenemos
Gβ
L = √ [p̄γ µ (1 − 1.26γ5 )n] [ēγµ (1 − γ5 )νe ] . (1.594)
2
con GF dado en la ec. (1.575) y Gβ = 1.10 × 10−5 GeV2 . La corriente hadrónica tiene la forma
V–1.26A. El factor 1.26 puede entenderse como debido a las correcciones a nivel hadrónico de una
corriente que es de la forma V–A a nivel del quarks, como en la ec. (1.574). A nivel de quarks el
decaimiento del neutrón (udd) al protón (uud) corresponde al decaimiento de uno de los quarks down
del neutrón d → ue− ν̄e
GF
L = √ V11 [ūγ µ (1 − γ5 )d] [ēγµ (1 − γ5 )νe ] . (1.595)
2
Figure 1.13: Decaimiento del neutrón.
De modo que Gβ = GF V11 = GF cos θC , donde θC es el ángulo de Cabbibo. Una vez se tienen en
cuenta correcciones electrodébiles se obtiene el valor |V11 | = 0.97418(27)[?]. Las magnitudes de los
elementos de la matriz CKM son[?]
 
0.97419 0.2257 0.0359
V ≈  0.2256 0.97334 0.0415  ∼ 1 (1.596)
0.00874 0.0407 0.999133
Chapter 2
Second quantization
2.1 Introduction
Two key ingredients to formulate the Quantum Field Theory (QFT) are the quantization of systems
in which the particles can be created and destroyed (quamtum theory of radiation) and the behavior
of relativistic systems. When both ingredients are present the particles can be understood as the
excited modes of certain field. When the particles in a system are not relativistic, the formalism of
creation and annihilation operators is just an alternative method to describe the Hamiltonian of the
Schrödinger equation. In relativistic systems however, the existence of negative energy states force
the construction of new quantum states, the Fock states, in order to have proper defined probabilities
for the states of the system. In section xx we start by building the Fock states associated to a massless
not relativistic scalar field. Then we generalize the results to a massive scalar field satisfying the
Klein-Gordon equation.
Some parts of the discussion were based in some topics of chapters 4-6 of [2].
In general, the formalism of second quantization is usefull to describe the states of an unde-
termined number of particles and interactions which do not conserve particle number. In addition
to high-energy physics where any number of particles may be created or annihilated during a col-
lision process, in statistitical physics it becomes useful to describe a macroscopic body using the
grand-canonical statistical ensemble, in which the number of particles is allowed to fluctuate. In
condensed-matter states the interactions may modify also the number of various excitation quanta,
such as phonons. A more general formalism to discuss this systems is developed in Appendix ??.
121
122 CHAPTER 2. SECOND QUANTIZATION
2.2 Quantization of the nonrelativistic string

2.2.1 The clasical string
In conventional quantization the energy of one state is interpreted as the possible eigenstates of an
Hamiltonian operator acting on the states of the system.
b
H|Statei = E|Statei (2.1)
One step further is to consider the wave function as the eigenstate of the operator–field acting on
certain Fock states
b
Φ|Fock Statei = Φ|Fock Statei , (2.2)
Like that usual quantum mechanical observable, the wave function will have an uncertainty. The
Fock states are the states under which the classical wave function can be obtained with a small
uncertainty
b
Φ ± ∆Φ = hFock State|Φ|Fock Statei (2.3)
This happens when the number of quanta of the Fock state is big enough. In fact, a state with a
definite number of quanta has a infinity uncertainty [19].
Eq. (2.2) is the basis for the calculation of cross section and decay widths in quantum field theory.
Now we will study how to define a such Fock state for a scalar field.
We have already see in Chapter 1 of [1] that a string have a collective wave motion that is
described by a continuous field, which satisfies the familiar one-dimensional wave equation
1 ∂ 2φ ∂ 2φ
− 2 =0 (2.4)
v 2 ∂t2 ∂z
This equation can be derived following two different paths. The first is to decomposing the string
into individual oscillators for which the usual Lagrangian formalism can be used. The second is just
by formulating certain Lagrangian density from which the equation of motion can be obtained by
using the Euler-Lagrange equation

∂L ∂L
∂µ − = 0. (2.5)
∂(∂µ φ) ∂φ
In the first approach, the string is considered to be composed of N oscillators coupled together by
springs with a spring constant k. At certain time t, the displacement of the oscillator i at time t is
represented by φi (t). In Table 2.1 it is displayed the corresponding macroscopic quantities. Note also
2.2. QUANTIZATION OF THE NONRELATIVISTIC STRING 123
micro macro
l L = Nl
m µ = m/l
k T = kl
φi (t) = φ(zi , t) φ(z, t)
Table 2.1: From micro to macro
that 1/v 2 = µ/T . It is worth to stress that the Lagrangian is the sum of each individual oscillator
Lagrangian, which in turn is the sum of the kinetic and potential energy of each oscillator. However,
it turns out that the full Lagrangian density only have the kinetic term for the scalar field
1 0
L= ∂ φ∂0 φ + ∂ 3 φ∂3 φ
2
−→ 21 ∂ µ φ∂µ φ . (2.6)
v→c
where x0 = vt. Note that only in the case v = c this Lagrangian density can be written in a covariant
form. Moreover, the scalar field φ(z, t) have nothing to do with the individual oscillators. An specific
solution for φ(z, t) would represent one specific oscillation mode of the string. It turn out that this
specific frequency mode corresponds to an particle state, that does not have connection with the
physical particles in the string.
The most general discrete solution to the wave equation (2.4) is the Fourier decomposition
X v
φ(t, z) = √ an e−i(ωn t−kn z) +a∗n ei(ωn t−kn z) (2.7)
n
2ωn L
where the dispersion relation is
ωn2 = v 2 kn2 (2.8)
where ωn is positive-definite:
p
ωn = +|v| |kn | (2.9)
To satisfy the boundary conditions we must have

2πn
kn = (2.10)
L
Note that
k−n = −kn . (2.11)

Therefore
ω−n = ωn . (2.12)
We now will check the origin of the normalization factor. By using eq. (2.7)
∞
X v
φ(z, t) = √ [an φn (z, t) + a∗n φ∗n (z, t)] , (2.13)
n=−∞
2ω n
1
[E] = [a]
[E]1/2 [E]−1
=E 1/2 [a] (2.14)
[a] = [E]1/2 (2.15)
we define
1
φn (z, t) = √ e−i(ωn t−kn z) , (2.16)
L
and the functions φn satisfy the following normalization conditions

Z L Z
1 L
dz φ∗n (z, t)φm (z, t) = d z ei(ωn t−kn z) e−i(ωm t−km z)
0 L 0
Z
1 L
= d z exp{i[(ωn − ωm )t − (kn − km )z]}
L 0
Z
ei(ωn −ωm )t L
= d z e−i(kn −km )z
L 0
(2.17)
When n = m
Z L Z L
1
dz φ∗n (z, t)φm (z, t) = dz
0 L 0
=1 (2.18)
For n 6= m, 2(n − m) is an even integer and then

Z L L
∗ ei(ωn −ωm )t e−i(kn −km )z
d z φn (z, t)φm (z, t) =
0 L −i(kn − km ) 0
ei(ωn −ωm )t 1
= e−i2π(n−m) −1
L −i(kn − km )
=0 (2.19)
In this way Z L
d z φ∗n (z, t)φm (z, t) = δnm . (2.20)
0
Moreover Z L
d z φn (z, t)φm (z, t) = δn,−m e−2iωn t . (2.21)
0
From the analysis of the Theorem of Noether in chapter 1 of [1] we have, that in a similar way to the
usual Lagrangian formulation, where the canonical conjugate variable is used to define the Legendre
transformation
H = pq̇ − L , (2.22)
the Hamiltonian density, H, can be obtained from the Lagrangian density trough the Legendre
transformation
∂L
H = T00 = φ̇ − L (2.23)
∂ φ̇
∂φ(x)
= Π(x) − L. (2.24)
∂t
where
∂L
Π(x) =
∂(∂φ(x)/∂t)
1 ∂φ
= 2 , (2.25)
v ∂t
is the canonical conjugate variable (conjugate momentum) of the canonical variable φ(x).
In this way
1 ∂φ ∂φ 1 ∂φ ∂φ 1 ∂φ ∂φ
H= 2
− 2 +
v ∂t ∂t 2v ∂t ∂t 2 ∂z ∂z
1 ∂φ ∂φ 1 ∂φ ∂φ
= 2 + . (2.26)
2v ∂t ∂t 2 ∂z ∂z
We have then,
Z L
H= H dz
0
Z L Z
1 ∂φ ∂φ 1 L ∂φ ∂φ
H= 2 dz + dz
2v 0 ∂t ∂t 2 0 ∂z ∂z
∞
X
= ωn a∗n an (2.27)
n=−∞
Demonstration:
∞
X
∂φ v
= √ [−iωn an φn (z, t) + iωn a∗n φ∗n (z, t)] ,
∂t n=−∞ 2ωn
X∞
−ivωn
= √ [an φn (z, t) − a∗n φ∗n (z, t)] , (2.28)
n=−∞
2ωn
X∞
1 ∂φ ∂φ −ωn ωm
2
= √ [an φn (z, t) − a∗n φ∗n (z, t)] [am φm (z, t) − a∗m φ∗m (z, t)] (2.29)
v ∂t ∂t n,m=−∞ 2 ωn ωm
∞
X −ωn ωm
= √ [an am φn φm − a∗n am φ∗n φm − an a∗m φn φ∗m + a∗n a∗m φ∗n φ∗m ] (2.30)
n,m=−∞
2 ω n ωm
∞
X
∂φ v
= √ [ikn an φn (z, t) − ikn a∗n φ∗n (z, t)] ,
∂z n=−∞ 2 ωn
X∞
ivkn
= √ [an φn (z, t) − a∗n φ∗n (z, t)] , (2.31)
n=−∞
2 ωn
X∞
∂φ ∂φ −v 2 kn km
= √ [an φn (z, t) − a∗n φ∗n (z, t)] [am φm (z, t) − a∗m φ∗m (z, t)] (2.32)
∂z ∂z n,m=−∞ 2 ωn ωm
∞
X −v 2 kn km
= √ [an am φn φm − a∗n am φ∗n φm − an a∗m φn φ∗m + a∗n a∗m φ∗n φ∗m ] (2.33)
n,m=−∞
2 ω n ω m
Using eqs. (2.20), and (2.21)

∞ Z L
1 X −ωn ωm
H= dz √ [an am φn φm − a∗n am φ∗n φm − an a∗m φn φ∗m + a∗n a∗m φ∗n φ∗m ]
2 n,m=−∞ 0 2 ωn ωm
∞ Z L
1 X −v 2 kn km
+ dz √ [an am φn φm − a∗n am φ∗n φm − an a∗m φn φ∗m + a∗n a∗m φ∗n φ∗m ]
2 n,m=−∞ 0 2 ωn ωm
∞
1 X −ωn ωm
= √ an am δn,−m e−2iωn t −a∗n am δn,m − an a∗m δn,m + a∗n a∗m δn,−m e2iωn t
2 n,m=−∞ 2 ωn ωm
∞
1 X −v 2 kn km
+ √ an am δn,−m e−2iωn t −a∗n am δn,m − an a∗m δn,m + a∗n a∗m δn,−m e2iωn t
2 n,m=−∞ 2 ωn ωm
∞
1 X −ωn ω−n −2iωn t ωn ωn ∗ ∗ ωn ω−n ∗ ∗ 2iωn t
= √ an an e + √ (a an + an an ) − √ a a e
2 n=−∞ 2 ωn ω−n 2 ωn ωn n 2 ωn ω−n n −n
∞
1 X −v 2 kn k−n −2iωn t kn kn ∗ ∗ kn k−n ∗ ∗ 2iωn t
+ √ an an e + √ (a an + an an ) − √ a a e (2.34)
2 n=−∞ 2 ωn ω−n 2 ωn ωn n 2 ωn ω−n n −n
Since ωn = ω−n and kn = −k−n

∞
1 X 1
H= (−ωn2 + v 2 kn2 )an an e−2iωn t +(ωn2 + v 2 kn2 )(a∗n an + an a∗n )
2 n=−∞ 2ωn

+ (−ωn2 + v 2 kn2 )a∗n a∗−n e2iωn t (2.35)
Finally, using eq. (2.8)
∞
1 X
H= ωn (a∗n an + an a∗n ) (2.36)
2 n=−∞
Since an and a∗n are classical quantities that commutates, the Hamiltonian is
∞
X ∞
X
H= ωn a∗n an = ωn |an |2 (2.37)
n=−∞ n=−∞
√
In this way, the factor 2ωn in eq. (2.13), is a convenient choice of normalization for the coefficients
an which guarantees the proper Hamiltonian.
To quantize the string, we need to promote H to an operator. In canonical quantization we need
to identify the proper conjugates variables. For this purpose it is convenient to write eq. (2.37) as
the Hamiltonian of an harmonic oscillator.
2.2.2 Quantization of the string

This Hamiltonian can be rewritten as a sum of independent oscillators Hamiltonians. Consider an
harmonic oscillator of frequency ω. The equation of motion for F = −kq is
k
q̈ + q =0
m
q̈ + ω 2 q =0 (2.38)
with
k
ω2 = (2.39)
m
This equation of motion can be obtained from the Lagrangian
1
L = T − V = [mq̇ 2 − kq 2 ] (2.40)
2
And the Hamiltonian can be obtained from eq. (2.22)
H =pq̇ − L
p2 1 p2
= − + kq 2
m 2 m
1 p2 2 2
= + mω q
2 m
1
= p2 + m 2 ω 2 q 2
2m
(2.41)
For a set of independent oscillators we have

X 1
H= p2n + m2 ωn2 qn2
n
2m
X 1 mωn 2

2
H= ωn pn + qn (2.42)
n
2mω n 2
Comparing eq. (2.42) with Eq. (2.37) we see that the complex number an can be written as (~ = 1)
an =c1 qn + ic2 pn (2.43)

a∗n an = c21 qn2 + c22 p2n (2.44)
√
mωn mωn 1
c1 = √ =√ c2 = √ (2.45)
2 2mωn 2mωn
mωn qn + i pn
an = √
2mωn
mωn qn − i pn
a∗n = √
2mωn
(2.46)
mωn √ 1
an + a∗n =2 √ qn = 2mωn qn ⇒ qn = √ (an + a∗n )
2mωn 2mωn
√
∗ 2i i 2mωn imωn
an − an = √ pn = pn ⇒ pn = − √ (an − a∗n ) (2.47)
2mωn mωn 2mωn
In quantum mechanics the classical objects qn and pn are promoted to operators which satisfy the
commutation relation
qn , pbm ] =iδmn
[b qn , qb †m ] = [pbn , pb †m ] =0 .
[b (2.48)
This implies that the objects an and a∗n , are also operators
imωm
qn , pbm ] = − √
[b {[b am ] − [b
an , b a†m ] + [b
an , b a†n , b a†n , b
am ] + [b a†m ]}
2mωn 2mωm
imωm
qn , pbm ] = − √
[b {[b am ] − 2[b
an , b a†m ] + [b
an , b a†n , b
a†m ]} (2.49)
2 mωn mωm
If the operators b
an and ba†n satisfy the commutation relations
† †
b a†m = δn,m
an , b an , b
[b am ] = b
an , b
am = 0 , (2.50)
then we recover equations (2.48). The scalar field is now an operator
∞
X v h i
φb = √ b a †n ei(ωn t−kn z) ,
an e−i(ωn t−kn z) +b (2.51)
n=−∞
2ωn L
an and ab †n the Hamiltonian from eq. (2.36) can be written as

In terms of operators b
X∞
b =1
H a †nb
ωn (b an + ban ab †n )
2 n=−∞
1 X
∞ h i
= ωn 2b a †nb
an + b an , ab †n
2 n=−∞
X∞
† 1
= ωn ab nb an + (2.52)
n=−∞
2
Since
∞
X 1
ωn → ∞, (2.53)
n=−∞
2
it is convenient to renormalize the Hamiltonian as

∞
X X∞
b † 1 1
:H: = ωn ab nban + − ωn
n=−∞
2 n=−∞
2
∞
X
= ωn ab †nb
an (2.54)
n=−∞
This procedure is consistent since the related physics quantities arise from energy differences, no
from absolute energy determinations.
h i ∞
X
b b † 1
H, am = ωn ab nb an + ,b
am
n=−∞
2
X∞ h i
= ωn ab †nb
an , b
am (2.55)
n=−∞
By using the identity
[AB, C] = [A, C] B + A [B, C] (2.56)

we have
h i ∞
X h i
b
H, b
am = ωn ab †n , b
am ban + ab †n [b
an , b
am ]
n=−∞
∞
X
=− ωn δnmb
an
n=−∞
= − ωm b
am (2.57)
h i ∞
X h i h i
b †
H, ab m = † †
ωn ab n , ab m b †
an + ab n b †
an , ab m
n=−∞
X∞
= ωn ab †n δnm
n=−∞
=ωm ab †m (2.58)
b with eigenvalue En
If |mn i is an eigenstate of H
b n i = En |mn i
H|m (2.59)
then

bb
H an |mn i = b b − ωnb
an H an |mn i
= (En − ωn ) b
an |mn i
(2.60)
an |mn i is also an eigenstate with eigenvalue En − ωn . Moreover,

b

b † † b †
H ab n |mn i = ab n H + ωn ab n |mn i
= (En + ωn ) ab †n |mn i
(2.61)
ab †n |mn i is also an eigenstate with eigenvalue En + ωn .

As stablished in [4]
In other words, the operator b

an seems to annihilate a quantum of energy, of amount ~ωn ,
from the state. On the other hand, ab †n creates a quantum of energy ~ωn . In this sense,
they are the are the annihilation and the creation operators, respectively. [...]
The ground state can be denoted by |0i = |0n i. Since this is state of lowest energy, the
annihilation operator ab †n , acting on it, cannot produce a state of lower energy. Thus, this
state must be totally annihilated by the operation of b an :
an |0i =0
b
a †n =0 ,
h0|b (2.62)
such that
h0|0i = 1 (2.63)
The energy if the ground state can be fixed to zero:
b : |0i = 0 .
:H (2.64)
We define the state whose energy is larger tha the energy of |0i by one quantum ~ωn by
a †n |0i
|1n i ≡b
h1n | =h0|b an (2.65)
|1n i is an Hamiltonian eigenstate of energy ωn :
b : |1n i =ωn a† an |1n i

:H n
= ωn |1n i
= ωn · 1|1n i , (2.66)
where we have made explicit that we have a quantum of energy ~ω. The normalized state is
an ab †n |0i
h1n |1n i =h0|b
h i
=h0| ban , ab †n |0i
=h0|0i
=1 . (2.67)
Similarly, the state with energy 2~ω is
1 2
√ ab †n |0i =|2n i
2
1
an )2 =h2n |
h0| √ (b (2.68)
2
with normalization
1
h2n |2n i = h0|b an ab †n ab †n |0i
an b
2
1
an ab †n |1n i
= h1n |b
2
1 h i
= h1n | b an , ab †n + ab †nb
an |1n i
2
1
= (h1n |1n i + h0|0i)
2
=1 . (2.69)
By induction we get
1 † m
√ ab n |0i =|mn i (2.70)
m!
From here we have
1 m−1
√ ab †n ab †n |0i =|mn i
m!
p m−1
(m − 1)! 1
√ p ab n ab †n
†
|0i =|mn i
m! (m − 1)!
s
(m − 1)! †
ab |(m − 1)n i =|mn i
m(m − 1)! n
r
1 †
ab |(m − 1)n i =|mn i
m n
√
ab †n |(m − 1)n i = m|mn i
√
ab †n |mn i = m + 1|(m + 1)n i (2.71)
or,
√
hmn |b
a n = m + 1h(m + 1)n | (2.72)
From this expressions we can check that number operator can be defined from:
a n ab †n |mn i =(m + 1)h(m + 1)n |(m + 1)n i

hmn |b
hmn |1 + ab †n ab n |mn i =(m + 1)
hmn |1 + N bn |mn i =(m + 1) (2.73)
In this way, the number operator as

bn = ab † b
N n an (2.74)
bn |mn i = c|mn i, where c must be real because N

If N bn is Hermitian
1+c=m+1 (2.75)
and
bn |mn i = mn |mn i
N (2.76)
From here, we can calculate the eigenvalues of b an . Since

h i
Nbnb
an = N bn , ban + b an Nbn
h i
= ab †n , b
an ban + ban Nbn + ab † [b
n an , b
an ]
=−b bn
an N
an + b

=b b
an Nn − 1 (2.77)
bnb
N an |mn i =(mn − 1)b
an |mn i (2.78)
Since the state

bn |mn − 1i =(mn − 1)|mn − 1i
N (2.79)
has the same eigenvalue, therefore
an |mi = C− |mn − 1i
b (2.80)
where C− is a number to be determined from the normalization condition

a †n ab n |mn i = |C− |2 hmn − 1|mn − 1i
hmn |b
hmn |N bn |mn i = |C− |2
|C− |2 = mn (2.81)
√
an |mn i =
b mn |mn − 1i (2.82)
such that
hmn |mn i = 1 (2.83)
Eq. (2.54) can be rewritten as
∞
X
b: =
:H bn
ωn N (2.84)
n=−∞
Noting also that

Nn |ml i = 0 n 6= l , (2.85)
we have that
b : |mn i = mn ωn = mn ~ωn .
hmn | : H (2.86)
Therefore, once we have proper normalized states and renormalized Hamiltonian, the energy of an
state with m quantum ( of frequency ωn ) is just m times the energy of the one quanta of energy ~ωn .
Note that
b : |0i = 0 .
h0| : H (2.87)
The general procedure to redefine the zero of energy such that the vacuum energy vanishes is called
normal ordering. We define a normal-ordered product by moving all annihilation operators to the
right of all creation operators. For an operator X,b its normal-ordered product will be denoted as
:Xb : . Using this algorithm on the expression of eq. (2.52), we find that
∞
b 1 X
: H: = a †nb
ωn : (b an ab †n ) :
an + b
2 n=−∞
∞
1 X
= a †nb
ωn (b an + ab †nb
an )
2 n=−∞
∞
X
= ωn ab †nb
an (2.88)
n=−∞
From [20] (pag. 121):

These idea carry over to quantum field theory, but with a different interpretation. In
quantum mechanics we are talking about a single particle state |mn i and energy levels
En = ω(n + 1/2). The creation and annihilation operators move the state of the particle
up and down in energy from the ground.
In quantum field theory, we take the notion of “number operator” literally. The state |ni
is not a state of a single particle, rather is an state of the field with N particles present.
The background state which is also the lowest energy state is a state of the field with 0
particles (but the field is still there). The creation operator ab †n adds a single quantum
(a particle) to the field, while the annihilation operator b an destroys a single quantum
(removes a single particle) from the field. As we will see, in general there will be creation
operators and annihilation operators for particles as well as for antiparticles.
These operators will be functions of momentum. The fields will become operators which
will be written as sums over annihilation and creation operators.
2.2.3 Generalization to three dimensions

Taking into account that En = ~ωn = ωn , when ~ = 1, the most general solution to the generalization
to three dimensions of the wave equation with velocity of propagation c = 1
∂ µ ∂µ φ = 0 , (2.89)
obtained from the three dimension Lagrangian
L = 12 ∂ µ φ∂µ φ , (2.90)
is
X 1
−ipn ·x ∗ ipn ·x
φ(t, x) = √ a n e +a n e ,
n
2En L3
X 1
= p a(nx ,ny ,nz ) exp{−i[E(nx ,ny ,nz ) t − px x − py y − pz z]}
nx ,ny ,nz
2E(nx ,ny ,nz ) L3
o
+a∗(nx ,ny ,nz ) exp{i[E(nx ,ny ,nz ) t − px x − py y − pz z]} , (2.91)
where in natural units the wave number can be identified with the momentum, p = k. In eq. (2.91)
2π
En =p0n pi = ni (2.92)
L
2.3. QUANTIZATION OF THE KLEIN-GORDON FIELD 137
where p0 = En , and the solution satisfies the dispersion relation
p2n = p2n = c2 En = En2 . (2.93)
The Energy will always be chosen to be positive
2π q 2
En = nx + n2y + n2z (2.94)
L
Since the Action is dimensionless,
Z
S = d4 x m2 φ2 →[1] = [E]−4 [E]2 [φ]2
[φ] =([S]/[E]2 )1/2 = [E] , (2.95)
this solution φ must have units of energy in natural units. To obtain the dimensions of an , we just
check the dimensions in both sides of eq. (2.91)
1
[E] = p [an ]
[E][E]−3
=[E][an ] , (2.96)
and therefore an is dimensionless.

The canonical quantization in eqs. (2.50) can be generalized to
† †
b a†m = δn,m
an , b an , b
[b am ] = b
an , b
am = 0 , (2.97)
2.3 Quantization of the Klein-Gordon field

It is convenient to put the system into a box of size L, so that the total volume is finite. According
eq. (2.10), in this case the frequency is discret. However particles like the photon or electron have
frequencies in a continuum range. Therefore we need to establish relations that allows extrapolate
the discrete results into the continuum, and also we will need to take the limit of infinite volume.
The Klein-Gordon equation for a real scalar field φ (Chapter 3. [1])
(∂ µ ∂µ + m2 )φ = 0 , (2.98)
can be obtained from the Lagrangian
L = 12 ∂ µ φ∂µ φ − 12 m2 φ2 , (2.99)
The solution is the same that for the case m = 0 in eq. (2.91), but the new dispersion relation is
En2 = p2n + m2 . (2.100)
and therefore m can be interpreted as the mass of field φ.

We assume that φ can have frequencies in the continuum. In this way the most general solution
is obtained after replacing the summatory by an integral
Z X X 2π X 2π X
dp → ∆p = pn+1 − pn = n+1−n= (2.101)
n n
L n L n
X 3 Z
L
→ d3 p (2.102)
n
2π
From
Z
d3 pδ (3) (p − q) =1 (2.103)
and taking into account that

X
δn,m = δn,m =1 , (2.104)
n
where
2π 2π
pi = ni qi = mi , (2.105)
L L
we have
Z X
d3 pδ (3) (p − q) = δnm
n
X 2π 3 X
δ (3) (p − q) = δn,m
n
L n
3
2π
δ (3) (p − q) =δn,m . (2.106)
L
In this way
3
(3) L
δ (p − q) = δn,m , (2.107)
2π
and we get that in the continuum limit

3
L
δn,m → δ (3) (p − q) (2.108)
2π
In particular, this implies that
(2π)3 δ (3) (p = 0) → L3 = V (2.109)
V
δ 3 (0) = . (2.110)
(2π)3
This expression can be also obtained from the definition

Z
3 1 3 −ip·x
δ (p) = lim dxe , (2.111)
V →∞ (2π)3 V
before taking the limit to infinity.

Therefore, in the continuum the solution in eq. (2.91) can be written as
3 Z
L 1
φ(t, x) = d3 p p ap e−ip·x +a∗p eip·x
2π 2Ep L3
Z √
L3 −ip·x ∗ ip·x

= d3 p p a pe +a p e (2.112)
(2π)3 2Ep
Using eq. (2.108), we can write the commutation relations (2.50) in the continuum as
3
2π † †
b a†q =
ap , b δ (3) (p − q) ap , b
[b aq ] = b
ap , b
aq = 0 . (2.113)
L
Note that again ap is dimensionless. It is customary to write the general solution (2.112) with
√
a0p = L3 ap . (2.114)
Then
Z
1 ∗
φ(t, x) = d3 p p a0p e−ip·x +a0p eip·x . (2.115)
(2π)3 2Ep
and the commutation relations in eq. (2.113) can be written as
0 0† 0 0 0† 0†
b aq = (2π)3 δ (3) (p − q)
ap , b b
ap , b
aq = b
ap , b
aq = 0 . (2.116)
In what follows we will drop out the prime in b a0p .
The basic principle of canonical quantization is to promote the field φ and its conjugate momentum
to operators, and to impose the equal time commutation relation
h i
b x), Π(t,
φ(t, b y) = iδ (3) (x − y)
h i h i
b b b b
φ(t, x), φ(t, y) = Π(t, x), Π(t, y) = 0 . (2.117)
We will now check that the commutation relations in eq. (2.116) will just generate the equal time
commutation relations in eq. (2.117).
Promoting the real field φ to a hermitian operator means to promote ap to an operator; thus
Z
b 1 −ip·x † ip·x
φ(t, x) = d3 p p bap e +b
a p e (2.118)
(2π)3 2Ep
with
h i h i
bap , ab †q = (2π)3 δ (3) (p − q) aq ] = ab †p , ab †q = 0 .
ap , b
[b (2.119)
The conjugate momentum can be obtained from the Klein-Gordon Lagrangian in eq. (2.99), by
using eq. (2.25)
b ∂ h 1 b 2i
Π(x) = (∂ φ)
b 2 0
∂(∂0 φ)
=∂0 φb
Z
1 −ip·x † ip·x

= d3 p p −iE pba pe +iE pba p e
(2π)3 2Ep
Z r
3 i Ep −ip·x † ip·x

= dp −b
a p e +ba p e
(2π)3 2
(2.120)
b and Π,
Using the expressions for φ, b in terms of b a†p , the commutation relation (2.117) reads
ap , b
s
h i Z Z
i Ep0 h i
b b
φ(t, x), Π(t, y) = d p d3 p0
3
ap e−ip·x +b
b a†p eip·x , −b
0
ap0 e−ip ·y +b
0
a†p0 eip ·y
2(2π)6 Ep
Z Z s
i Ep0
= d3 p d3 p0 ×
2(2π)6 Ep
nh 0
i h 0
io
bap e−ip·x +ba†p eip·x , −b ap0 e−ip ·y + b ap e−ip·x +b a†p0 eip ·y
a†p eip·x , b
Z Z s
i Ep0 nh −ip0 ·y
i
= d3 p d3 p0 × b
a p e −ip·x
, −b a p 0 e
2(2π)6 Ep
h i h i h io
† ip·x −ip0 ·y −ip·x † ip0 ·y † ip·x † ip0 ·y
+ b ap e , −b ap e
0 + b ap e ,bap0 e + b ap e , b ap0 e
Z Z s
i Ep0 n −i(p·x+p0 ·y)
= d p d3 p0
3
× −e ap , b
[b ap0 ]
2(2π)6 Ep
0 † 0
h i 0
h io
− ei(p·x−p ·y) b ap0 + e−i(p·x−p ·y) b
ap , b a†p0 + ei(p·x+p ·y) b
ap , b a†p0
a†p , b . (2.121)
Taking into account the eqs. (2.119), then
s
h i Z Z
i Ep0 n −i(p·x−p0 ·y) h i † o
b b
φ(t, x), Π(t, y) = d p d3 p0
3
e b
ap , b
0
a†p0 − ei(p·x−p ·y) b
ap , b
ap0
2(2π)6 Ep
Z Z s
i Ep0 h −i(p·x−p0 ·y) (3) 0
i
= d3 p d3 p0 × e δ (p − p0 ) + ei(p·x−p ·y) δ (3) (p0 − p)
2(2π)3 Ep
Z Z s
i Ep0 (3) h 0 0
i
= d3 p d3 p 0 δ (p − p0 ) e−i(p·x−p ·y) + ei(p·x−p ·y) . (2.122)
2(2π)3 Ep
δ (3) (p − p0 ) forces p = p0 , which also means Ep = Ep0 , and since x0 = y 0 = t, we have

s
h i Z Z
i Ep0 (3)
b x), Π(t,
φ(t, b y) = d3 p d3 p0 δ (p − p0 )×
2(2π) 3 Ep
h 0 0
i
e−i[t(Ep −Ep0 )−p·x+p ·y] + ei[t(Ep −Ep0 )−p·x+p ·y]
Z
i −i(−p·x+p·y)
= d3 p 3
e + ei(−p·x+p·y)
2(2π)
Z
i ip·(x−y)
= d3 p 3
e + e−ip·(x−y) . (2.123)
2(2π)
Since
Z
(3) d3 p −ip·(x−y)
δ (x − y) = e
(2π)3
Z Z
(3) d3 p −ip·(−x+y) d3 p ip·(x−y)
=δ (−x + y) = e = e , (2.124)
(2π)3 (2π)3
then
h i
b x), Π(t,
φ(t, b y) = iδ (3) (x − y) . (2.125)
The same expression is obtained for the original field operator in eq. (2.112) if the commutation
relations (2.113) are used. Moreover eq. (2.125) is covariant [4].
Note that the commutation relations for the real scalar field in (2.119) are equivalent to that of
a collection of independent harmonic oscillators, with one oscillator for each value of the momentum
p.
Previous equations for the Hamiltonian still holds.
Z
b 1
H= d3 p Ep ab †pb
ap + bap ab †p (2.126)
2
h i
b ap = −Epb
H, b ap
h i
b ab †p = +Ep ab †p
H, (2.127)
The analogy between the simple harmonic oscillator and the field is now complete. Therefore ab †p
b while b
creates the quanta of momentum p of the field φ, ap is the annihilation operator for a field
quantum with momentum p. From [4]:
What was the positive energy component of the classical field now annihilates the quan-
tum, and the negative energy component now creates the quantum. This quantum is
what we call particle of positive energy.
2.4. FOCK SPACE 143
2.4 Fock space

Given the Hilbert space of single-particles H, to construct the space of states of variable particle
number, consider the collection of all possible spaces of n identical particles for either bosons (ν = 1
or fermions ν = −1, Hνn . In particular the one-dimensional H0 space is defined by
H0 = {λ|0i; λ ∈ C} (2.128)
where |0i is called the vacuum state. A state in which the number of particles is not fixed, e.g n → ∞,
is given by the sequences (|0i = |Φ(0)i)
|Φi = {|Φ(n)i}n , |Ψi = {|Ψ(n)i}n , (2.129)
with properties
|Φi + |Ψi = {|Φ(n)i + |Ψ(n)i}n

X∞
hΦ|Ψi = hΦ(n)|Ψ(n)i . (2.130)
n=0
The collection of all vector of the form (2.129) which are of finite norm
∞
X
hΦ|Φi = hΦ(n)|Φ(n)i < ∞ , (2.131)
n=0
forms a Hilbert space Fν (H) called Fock space.

The operator A b acting on Fock space is defined by
∞
X
b
A|Φi = b
A(n)|Φ(n)i . (2.132)
n=0
Suppose you have a system described by a Hilbert space H, for example a single particle. The
Hilbert space of two non-interacting particles associated to the same field φ as that described by H
is simply the tensor (aka direct) product
H2 = H ⊗ H (2.133)
More generally, for a system of m particles as above, the Hilbert space for the m-excitations of
the field ψ is
H m := H
| ⊗ ·{z
· · ⊗ H} (2.134)
m times
In QFT there are operators that intertwine the different H m , that is, create and annihilate
particles. Typical examples are the creation and annihilation operators. Instead of defining them in
terms of their action on each pair of H n and H m , one is allowed to give a comprehensive definition
on the larger Hilbert space
Γ(H) := C ⊕ H ⊕ H 2 ⊕ · · · ⊕ H N ⊕ · · · (2.135)
known as the Fock Hilbert space of H.

From a physical point of view, the general definition above of Fock space is immaterial. Identical
particles are known to observe a definite (para) statistics that will reduce the actual Hilbert space
(by symmetrization/antisymmetrization for the bosonic/fermionic case etc...).
In eq. (2.70) we write out the ith bosonic state occupied by mi particles. Written as a Fock state
we have
1 † m
√ ab i | . . . , 0i , . . .i =| . . . , mi , . . .i . (2.136)
m!
The complete Fock space for a system
Y
k
1 m
|m1 , . . . , mi , . . . , mk i = √ âj j |0, . . . , 0k i (2.137)
j=1
mj !
As an example,
Further:
2.5 Fock space for the harmonic oscillator

We can now construct the Fock space following the standard procedure for the harmonic oscillator:
we interpret bap as destruction operators and ba†p as creation operators, and we define a vacuum state
|0i as the state annihilated by all destruction operators, so for all p
ap |0i = 0 .
b (2.138)
We normalize the vacuum with h0|0i = 1. The vacuum is the state which contains no particles and
no antiparticles either,
The normal ordered Hamiltonian is
Z
b : = d3 p Ep ab †pb
:H ap (2.139)
2.5. FOCK SPACE FOR THE HARMONIC OSCILLATOR 145
such that, as in discrete case
b : |0i = 0 .
h0| : H (2.140)
A possible normalization factor for the Fock one-particle state is (|pi ≡ |1p i)
1 †
|pi = √ b ap |0i
V
1
ap √
hp| =h0|b (2.141)
V
This state contains one quantum of the field φb with momenta pµ = (Ep , p). Such states have positive
norm, since
1
hp|p0 i = h0|b a†p0 |0i
ap b
V
1
= h0|b a†p0 − b
ap b a†p0 b
ap |0i
V
1
= h0|[b a†p0 ]|0i
ap , b
V
(2π)3 (3)
= δ (p − p0 )
V
3
2π
= δ (3) (p − p0 ) (2.142)
L
With this normalization, the limit to discrete case is straightforward:
h1n |1m i = δn,m (2.143)
The results are summarized in Table 2.2.

Similarly we can define many particle states. If a state has N particles with all different momenta
p1 , p2 , . . . , pN , it is defined by
1
|p1 , . . . , pN i = ab † · · · ab †pN |0p1 , . . . , 0pN i
V N/2 p1
1
≡ N/2 ab †p1 · · · ab †pN |0i
V
(2.144)
Discret Continuum Continuum ab 0p

P L 3
R 3 L 3
R 3
n 2π
dp 2π
dp
2π
3 2π
3
δ δ (3) (p − q) δ (3) (p − q)
hn,m i hL i hL i
b
φ(x), b
Π(y) = iδ (3) (x − y) b
φ(x), b
Π(y) = iδ (3) (x − y) b
φ(x), b
Π(y) = iδ (3) (x − y)
3 (3)
an , ab †m ] = δn,m
[b ap , ab †q ] = 2π
[b L
δ (p − q) ap , ab †q ] = (2π)3 δ (3) (p − q)
[b
|1n i = ab †n |0i a†p |0i
|pi = b |pi = √1
a†p |0i
b V

2π 3
3 (3)
h1n |1m i = δn,m hp|qi = L
δ (3) (p − q) hp|qi = 2π
L
δ (p − q)
Table 2.2: From discret to continuos, where pi = 2πni /L, and qi = 2πmi /L ,
On the other hand, if we want to construct a state with m particles of momentum p, we must have
a Fock state similar to (??)
11 † m
|mp i = m/2 √ ab p |0i (2.145)
V m!
From[4]
The vacuum, together with single particles states (2.141) and all multi–particle states
(2.144), (2.145), constitute a vector space which is called the Fock space. The creation
and annihilation operators act on this space.
It is convenient to define:
b
φ(x) = φb+ (x) + φb− (x) (2.146)
where
Z
1
φb+ (x) = d3 p p ap e−ip·x
b
(2π)3 2Ep
Z
1
φb− (x) = d3 p 3
p a†p eip·x .
b (2.147)
(2π) 2Ep
The effect of the operator field, φb± (x), on the one particle state |pi
φ± (x)|pi (2.148)
2.5. FOCK SPACE FOR THE HARMONIC OSCILLATOR 147
will be important for the evaluation of S–matrix elements in Chapter 5. In fact, as established in
Sec. 2.2, it is convenient to work in the discrete limit where (2.110)
V
δ 3 (0) = . (2.149)
(2π)3
Now we can write down the action of various field operators on different one particles states.
From now on we will denote the particle state with the initial of the name of the particle as a
function of its momentum. For example, the creation of a Higgs of momentum k, will be denoted as
1 †
|H(k)i = √ b ak |0i . (2.150)
V
Using the Fourier decomposition of the scalar field in eq. (2.147), and taking into account that
ap |0i = 0, we have
Z
1
φ+ (x)|H(k)i = d3 p p ap e−ip·x |H(k)i
b
(2π)3 2ωp
Z
1 1 †
= d3 p 3
p ap e−ip·x √ b
b ak |0i
(2π) 2ωp V
Z
1
= d3 p p e−ip·x [b a†k ]|0i .
ap , b (2.151)
(2π)3 2ωp V
By using the commutation relations in eq. (2.119) we have
Z
δ (3) (p − k) −ip·x
φ+ (x)|H(k)i = d3 p p e |0i (2.152)
2ωp V
1
φ+ (x)|H(k)i = √ e−ik·x |0i (2.153)
2ωk V
Similarly, we have initial one-particles states on left and initial one-particles states on right
1 1
φ+ (x)|H(k)i = √ e−ik·x |0i, hH(k)|φ− (x) =h0| √ eik·x . (2.154)
2ωk V 2ωk V
Another choice of normalization is the Lorentz invariant one, to be used later. In this case, the
Fock state of N particles with all different momenta p1 , p2 , . . . , pN , is obtained acting on the vacuum
with the creation operators,
|p1 , . . . , pn i ≡ (2Ep1 )1/2 . . . (2Epn )1/2 b
a†p1 . . . b
a†pn |0i . (2.155)
The factors (2Ep1 )1/2 are a convenient choice of normalization. In particular, the one-particle states
are
|pi = (2Ep )1/2 b

a†p |0i . (2.156)
From the commutations relations and eq. (2.119) we find that
hp|qi = (2Ep )1/2 (2Eq )1/2 h0|b a†q |0i

ap b

= (2Ep )1/2 (2Eq )1/2 h0| b a†q |0i
ap , b
= (2Ep )1/2 (2Eq )1/2 (2π)3 δ (3) (p − q)
=2Ep (2π)3 δ (3) (p − q) . (2.157)
The factors (2Ep )1/2 in eq. (2.156) have been chosen so that in the above product the combination
Ep δ (3) (p − q) appears, which is Lorentz invariant. To see this perform a boost along z–axis. Since
the transverse components of the momentum are no affected we must consider only Ep δ(pz − kz ).
Use the form of the Lorentz transformation of Ep , pz , together with the property of the Dirac delta
δ(f (x)) = δ(x − x0 )/f 0 (x0 ) [2].
Using (2.108) we have in a finite box
hp|qi =2En L3 δn,m

=2En V δn,m . (2.158)
2.6 Propagator
With conventions
b
φ(x) = φb+ (x) + φb− (x) (2.159)
where
Z
1
b
3
(2π) 2Ep
Z
1
φb− (x) = d3 p p a†p eip·x .
b (2.160)
(2π)3 2Ep

∂ µ ∂µ + m2 φ(x) = J(x)
2.7. QUANTIZATION OF FERMIONS 149
2.6.1 Complex scalar field

With conventions
b
φ(x) = φb+ (x) + φb− (x) (2.161)
where
Z
1
b
3
(2π) 2Ep
Z
1
φb− (x) = d3 p p bb† eip·x .
p (2.162)
3
(2π) 2Ep
2.7 Quantization of Fermions

We consider now the Dirac equation
(iγ µ ∂µ − m)ψ(x) = 0 (2.163)
that can be obtained from the Lagrangian
L = iψγ µ ∂µ ψ − mψψ (2.164)
where
ψ = ψ†γ 0 (2.165)
and the γ matrices satisfy the Dirac algebra
{γ µ , γ ν } = 2g µν 1 (2.166)
See [1]. If we assume a plane wave solution like the wave function of the Scrödinger equation
ψ ∝ e−iEt , after sustition in eq. (2.163), we have
iγ 0 (−iE) − m =0
γ 0 E − m =0 (2.167)
From the Dirac matrices properties we have

† 2
γ 0 =γ 0 γ0 =1 Tr γ 0 =0 . (2.168)
Moreover, we know that if γ µ satisfy the Dirac algebra, the matrices obtained after the unitary
transformation
eµ =U † γ µ U
γ s.t U † =U −1 (2.169)
also satisfy the Dirac algebra. To check this note that

eν } = U † γ µ U, U † γ ν U
γ µ, γ
{e
=U † {γ µ , γ ν } U
=2g µν U † U
=2g µν (2.170)
In this way we can always choose U such that γ 0 be diagonal. Because the restrictions in eq. (2.168)
this implies that in this representation we have

0 1 0
γ = (2.171)
0 −1
where the 1 and 0 are the 2 × 2 identity and null matrix respectively. Replacing back in eq. (2.167)
we have

E−m 0
=0
0 −E − m
E = ± m. (2.172)
so that from the four wave functions that compose the full Dirac spinor ψ, two of them are of positive
energy and the other two of negative energy. The Dirac spinor has four components, in this way we
expect four independent solutions. Let us represent solutions in the form
 
u1 (p)e−ip·x
u2 (p)e−ip·x 
ψ(x) ∝  
 v1 (p)eip·x  = ψ+ (x) + ψ− (x) , (2.173)
v2 (p)eip·x
where
ψ+ (x) ∝ us (p)e−i(Et−p·x) ψ− (x) ∝ vs (p)ei(Et−p·x) (2.174)
with
   
u1 (p) 0
u2 (p)  0 
us (p) = 
 0 
 vs (p) = 
v1 (p)
 (2.175)
0 v2 (p)
Checking this solutions to eq. (2.163) we have
(iγ 0 ∂0 + iγ i · ∂i − m)ψ+ (x) =0

(iγ 0 ∂0 + iγ · ∇ − m)ψ+ (x) =0
(γ 0 E − γ · p − m)ψ+ (x) =0
(γ µ pµ − m)ψ+ (x) =0
(p − m)ψ+ (x) =0
(p − m)us (p) =0 (2.176)
and
(p + m)vs (p) =0 (2.177)
This equations can also be written as
[(p − m)us (p)]† =0
u†s (p)(γµ† pµ − m) =0
u†s (p)γµ† γ 0 pµ − mu†s (p)γ 0 =0
u†s (p)γ 0 γµ pµ − mu†s (p)γ 0 =0
ūs (p)(p − m) =0 (2.178)
v̄s (p)(p + m) =0 (2.179)

At zero momentum, E = m and
ūs (0)(p − m) =0
u†s (0)γ 0 (γ 0 p0 + γ i pi − m) =0
u†s (0)γ 0 (γ 0 p0 − m) =0
u†s (0)(E − γ 0 m) =0
u†s (0)m =u†s (0)γ 0 m
u†s (0) =u†s (0)γ 0 ,
therefore
γ 0 us (0) = + us (0) γ 0 vs (0) = − vs (0) (2.180)

From [6]
Consider the matrix γ0 . It is a 4 × 4 matrix, so it has four eigenvalues and eigenvectors.

It is hermitian, so the eigenvalues are real. In fact, from Eq. (2.166) we know that its
square is the unit matrix, so that its eigenvalues can only be ±1. Since γ0 is traceless, as
we have proved in §3, there must be two eigenvectors with eigenvalue +1 and two with
−1
Eq. (2.180) shows that at zero momentum, the u–spinors and the v–spinors are simply eigenstates
of γ0 with eigenvalues +1 and −1. Of course this guaranteses that
us (0)vs0 (0) = 0 (2.181)
since the belong to differente eigenvalues. Note that the two us (0) and the two vs (0) are degenerate.
We define
us (0) ∝ξs vs (0) ∝ η−s (2.182)
where the munis sign in η−s is just a convention. We define the normalized eigenvectors ξ and η such
that
ξs† ξs0 =δss0 ηs† ηs0 =δss0

ξs† ηs0 = 0 (2.183)
In this way we have

       
1 0 0 0
0 1 0 0
ξ1/2 = 
0
 ξ−1/2 = 
0 η1/2 = 
1 η−1/2 =
0
 (2.184)
0 0 0 1
To obtain the spinors for any value of p we know that they must satisfy eqs. (2.176), (2.177),
and, reduce to eq. (2.182) when p → 0. The result is
us (p) =Np (p + m) ξs

vs (p) =Np (−p + m) η−s (2.185)
Choosing
1
Np = √ (2.186)
E+m
we obtain
u†s (p)us0 (p) = vs† (p)vs0 (p) = 2Eδss0 (2.187)
u†s (−p)vs0 (p) = vs† (−p)us0 (p) = 0 (2.188)
In terms of the conjugate spinors
ūs (p)us0 (p) =2mδss0

v̄s (p)vs0 (p) = − 2mδss0 (2.189)
ūs (p)vs0 (p) = v̄s (p)us0 (p) = 0 (2.190)
The spinors also satisfy some completeness relations (For details see [5]: Sec. 5.3)
X
us (p)ūs (p) = p + m (2.191)
s
X
vs (p)v̄s (p) = p − m (2.192)
s
The solutions to the free Dirac equations are

1
ψparticle (x) = p us (p)e−ip·x
2Ep V
1
ψantiparticle (x) = p vs (p)eip·x (2.193)
2Ep V
We define the projection operators as

±p + m
Λ± (p) = ,
2m
with properties
[Λ± (p)]2 = Λ± (p) .

Λ− Λ+ = Λ+ Λ− =0
Λ− + Λ+ =1 .
Λ+ us (p) =us (p)

Λ+ vs (p) =0
We define

Σ = σ 23 σ 31 σ 12
Σ·p
Σp =
|p|
1 ± Σp
Π± (p) = ,
2
where
[Π± (p)]2 = Π± (p) .
Moreover
[Λ± , Π± ] =0 ,
and
Σp us (p) =sus (p)

Σp vs (p) = − svs (p) .
We define also the chiral operator

2
1 ± γ5
L± =
2
2
(L± ) =L±
L± + L∓ =1
L± L∓ =0 .
To build the spin projector is a generalization applicable to a particle of p = 0

p · ŝ
n0 =
m
n0 p
n =ŝ + .
E+m
With them we can define
1
p↑ = 1 + γ 5
n
2
1
p↓ = 1 + γ 5n .
2
As with the scalar field, we write the Dirac field as an integral over momentum space of the plane
wave solutions, with creation and annihilation operators as coefficients,
Z X
1
ψ(x) = d3 p 3
p as (p)us (p)e−ip·x + b†s (p)vs (p)eip·x (2.194)
(2π) 2Ep s=1,2
Z
H= d3 xH
Z
3 ∂L
= dx ψ̇ − L
∂ ψ̇
Z

= d3 x iψγ 0 ∂0 ψ − iψγ 0 ∂0 ψ − iψγ i ∂i ψ + mψψ
Z

= d3 x −iψγ i ∂i ψ + mψψ
Z
= d3 xψ (−iγ · ∇ + m) ψ . (2.195)
Since that
Z X
1
(−iγ · ∇ + m) ψ = d3 p 3
p as (p) (−iγ · ∇ + m) us (p)e−ip·x
(2π) 2Ep s=1,2

+b†s (p) (−iγ · ∇ + m) vs (p)eip·x
Z X
1
= c d3 p 3
p as (p) (γ · p + m) us (p)e−ip·x
(2π) 2Ep s=1,2

+b†s (p) (−γ · p + m) vs (p)eip·x . (2.196)
From eqs. (2.176), and (2.177) we have

γ0 p0 + γi pi − m us (p) =0

γ0 p0 + γi pi + m vs (p) =0 , (2.197)
!
X
γ0 Ep − γ i pi − m us (p) =0
i
!
X
γ0 Ep − γ i pi + m vs (p) =0 , (2.198)
i
(−γ · ∇ − m) us (p) = − γ0 Ep us (p)

(−γ · ∇ + m) vs (p) = − γ0 Ep vs (p) , (2.199)
(γ · ∇ + m) us (p) =γ0 Ep us (p)

(γ · ∇ − m) vs (p) =γ0 Ep vs (p) . (2.200)
Replacing back in eq. (2.196), we have
Z X
1 −ip·x †

(−iγ · ∇ + m) ψ = d3 p p a s (p)γ0 Ep us (p)e − b s (p)γ0 Ep vs (p)eip·x
. (2.201)
(2π)3 2Ep s=1,2
Therefore
Z
H= d3 xψ (−iγ · ∇ + m) ψ
Z Z X h † i
3 3 0 1 0 † 0 ip0 ·x 0 † 0 −ip0 ·x
= dx dp 3
p as0 (p )us0 (p )e + bs0 (p )vs0 (p )e γ 0 (−iγ · ∇ + m) ψ
(2π) 2Ep0 s0 =1,2
Z Z X h † i
1 0 † 0 ip0 ·x 0 † 0 −ip0 ·x
= d3 x d3 p0 p a 0 (p )u 0 (p )e + b s0 (p )v 0 (p )e γ0
(2π)3 2Ep0 s0 =1,2 s s s
Z X
1 −ip·x †

× d3 p p a s (p)γ0 Ep us (p)e − b s (p)γ 0 Ep vs (p)e
ip·x
(2π)3 2Ep s=1,2
Z Z Z s
d3 x 3 0 3 Ep X h † 0 † 0 ip0 ·x 0 † 0 −ip0 ·x
i
= d p d p a 0 (p )u 0 (p )e + b s 0 (p )v 0 (p )e
2(2π)6 Ep0 s,s0 =1,2 s s s

× as (p)us (p)e−ip·x − b†s (p)vs (p)eip·x
Z Z Z s
d3 x 3 0 3 Ep X
= d p d p
2(2π)6 Ep0 s,s0 =1,2
h 0 0
i
× a†s0 (p0 )as (p)u†s0 (p0 )us (p)ei(p −p)·x − bs0 (p0 )b†s (p)vs†0 (p0 )vs (p)ei(p−p )·x
Z Z s
d 3
p Ep X
= d3 p0
2(2π)3 Ep0 s,s0 =1,2
Z Z
† 0 † 0 d3 x i(p0 −p)·x 0 † † 0 d3 x i(p−p0 )·x
× as0 (p )as (p)us0 (p )us (p) e − bs0 (p )bs (p)vs0 (p )vs (p) e
(2π)3 (2π)3
Z Z s
d 3
p Ep X
= d3 p0
2(2π)3 Ep0 s,s0 =1,2
h i
× a†s0 (p0 )as (p)u†s0 (p0 )us (p)ei(Ep0 −Ep )t δ (3) (p − p0 ) − bs0 (p0 )b†s (p)vs†0 (p0 )vs (p)ei(Ep −Ep0 )t δ (3) (p − p0 )
Z
d3 p X h † † † †
i
= a 0 (p)a s (p)u 0 (p)u s (p) − b s 0 (p)b s (p)v 0 (p)v s (p)
2(2π)3 s,s0 =1,2 s s s
Z 3 Xh † i
dp
= (2π)3 as0 (p)as (p) − bs0 (p)b†s (p) . (2.202)
E p s=1,2
In order to obtain the quantization relations could see that if commutation relations are used we
could get
Z
d3 p X † †

:H: = E p a s (p)a s (p) − b s (p)b s (p) (2.203)
(2π)3 s=1,2
The minus sign arise from the anticommutation relations, so that a real spinor field, where bs (p) =
as (p) is automatically zero. Even after normal ordering, this Hamiltonian could give to arise
negative energy eigenvalues, which is a serious problem. If instead we assume that the creation and
annihilation operators satisfy anticommutation relations

ar (p), a†s (q) = br (p), b†s (q) = (2π)3 δrs δ (3) (p − q) (2.204)
With this relations and taking into account
∂L
Πψ (x) = = iψγ 0 = iψ † (2.205)
∂(∂0 ψ)
we obtain
{ψ(x, t), Πψ (y, t)} = iδ (3) (x − y) (2.206)
With the anticommutators the normal–ordered Hamiltonian is
Z
d3 p X † †

:H: = Ep as (p)a s (p) + b s (p)b s (p) (2.207)
(2π)3 s=1,2
Moreover
Z
: Q : =q d3 x : ψ † ψ :
Z X
=q d3 p a†s (p)as (p) − b†s (p)bs (p) (2.208)
s=1,2
With this definition a†s (p) creates particles of charge q, while b†s (p) creates antiparticles of charge
−q. In a similarly way to eq. (2.146), the most general free particle solution to Dirac equation is
b
ψ(x) = ψb+ (x) + ψb− (x) (2.209)
Z X
1
ψb+ (x) = d3 p
3
p as (p)us (p)e−ip·x
(2π) 2Ep s=1,2
Z X
1
ψb− (x) = d3 p p b†s (p)vs (p)eip·x (2.210)
(2π)3 2Ep s=1,2
The Fourier expansion for antiparticles is

Z X
b 1
ψ + (x) = d3 p p bs (p)v̄s (p)e−ip·x
(2π)3 2Ep s=1,2
Z X
b 1
ψ − (x) = d3 p 3
p a†s (p)ūs (p)eip·x (2.211)
(2π) 2Ep s=1,2
In this way a†s and as are the creation and annihilation operators for particles, while b†s and bs are
the creation and annihilation operators for antiparticles.
It is clear then that the one particle state is
r
1 †
|e− (p, s)i ≡ a (p)|0i (2.212)
V s
while the one antiparticle state is
r
1 †
|e+ (p, s)i ≡ b (p)|0i . (2.213)
V s
2.7.1 Quantization of Weyl spinors

We start from the Euler-Lagrange equation for a left-spinor:
The proposed solution is:
From which we obtains
(2.214)
These results are applicable both to spin sums and helicity sums, and hold for both massive and
massless spin-1/2 fermions
This results can be combined as
The relevant basis of two-component spinors χs are eigenstates of (1/2)σ · s
and therefore
In the following we need an explicit form for the internal Lorentz transformations
where:
After some calculations
X X
xα (p, s)x†β̇ (p, s) =p · σαβ̇ , x†α̇ (p, s)xβ (p, s) =p · σ α̇β
s s
X X
yα (p, s)yβ̇† (p, s) =p · σαβ̇ , y †α̇ (p, s)y β (p, s) =p · σ α̇β (2.215)
s s
X X
xα (p, s)y β (p, s) =mδαβ , yα (p, s)xβ (p, s) = − mδαβ
s s
X X
y †α̇
(p, s)χ†β̇ (p, s) =mδβ̇α̇ , x†α̇ (p, s)yβ̇† (p, s) = − mδβ̇α̇ (2.216)
s s
XZ d3 p
−ip·x † ip·x
eL → ξα = p x α (s, p) as e + y α (s, p) bs e
s
(2π)3 2Ep
X Z h i
d3 p
(eL )† → ξα̇† = p y †
α̇ (s, p) b s e−ip·x
+ x †
α̇ (s, p) a † ip·x
s e
s
(2π)3 2Ep
†
XZ d3 p α
α −ip·x α † ip·x
(eR ) → η = p x (s, p) b se + y (s, p) a s e
s
(2π)3 2Ep
XZ d3 p †α̇
†α̇ −ip·x †α̇ † ip·x
eR → η = p y (s, p) a se + x (s, p) b s e . (2.217)
(2π) 3 2E
s p
from here we can define the one–particle states

r r
1 1 †α̇
|eL (p, s)i ≡ xα (s, p) a†s (p)|0i, |eR (p, s)i ≡ y (s, p) a†s (p)|0i
V V
r r
† 1 † † † 1 α
| (eL ) (p, s)i ≡ yα̇ (s, p) bs (p)|0i, | (eR ) (p, s)i ≡ x (s, p) b†s (p)|0i , (2.218)
V V
By following similar steps

1 † 1
ξ+ (x)|eL (p, s)i = p x(s, p)e−ip·x |0i, heL (p, s)|ξ− (x) =h0| p x† (s, p)eip·x
2Ep V 2Ep V
† 1 1
η+ (x)|eR (p, s)i = p y † (s, p)e−ip·x |0i, heR (p, s)|η− (x) =h0| p y(s, p)eip·x
2Ep V 2Ep0 V
1 1
η+ (x)| (eR )† (p, s)i = p x(s, p)e−ip·x |0i, h(eR )† (p, s)|η− †
(x) =h0| p x† (s, p)eip·x
2Ep V 2Ep V
1 1
†
ξ+ (x)| (eL )† (p, s)i = p y † (s, p)e−ip·x |0i, h(eL )† (p, s)|ξ− (x) =h0| p y(s, p)eip·x . (2.219)
2Ep V 2Ep V
One additional identity is the one relating the two-component spinor momentum space wave
function to the four-component spinor wave functions
Chapter 3
Quantization of the electromagnetic field
3.1 Preliminaries
If we impose charge conservation: ∂µ J µ = 0, the Proca Equations can be written without lost of
generality as (§ 2.4 of [1])
( + m2 )Aµ = J µ . (3.1)
where Aµ = (φ, A).

The right side of the equation can be obtained after replacing the quantities in the equation for
energy-momentum conservation
E 2 − p2 = m2 , (3.2)
pµ = i∂ u . (3.3)
This suggest that quantum mechanics is a key ingredient to understand the local conservation of
electric charge, as we will see later.
For the scalar part we have the Klein-Gordon equation of an real scalar field:
( + m2 )φ = ρ (3.4)
which can be obtained from the Lagrangian (§ 3.1 of [1])
L = Lfree + Lint , (3.5)
163
164 CHAPTER 3. QUANTIZATION OF THE ELECTROMAGNETIC FIELD
1 1
Lfree = ∂ µ φ∂µ φ − m2 φ2
2 2
Lint =ρφ , (3.6)
where ρ is the charge density of the field which is the source for φ, and Lint is the interaction
Lagrangian.
In the same section it is shown that this Lagrangian give to arise to the Yukawa interaction
1 e−mr
V (r) = − . (3.7)
4π r
In a similar way, when the Lorentz force
F = qE + qv × B, (3.8)
is interpreted in terms of quantum mechanical operators (§ 3.3 of [1]) we have the canonical momen-
tum
∂ µ → Dµ = ∂ µ + iqAµ . (3.9)
Now, if we force the Scrödinger equation to be invariant under local phase changes (§ 3.4 of [1]),
we need to replace the normal derivate by the covariant derivate which must transform as the wave
equation:
Dµ ψ → (Dµ ψ)0 = eiθ(x) Dµ ψ . (3.10)
This suggest to make the minimum replacement
Dµ = ∂ µ + iqAµ , (3.11)
where Aµ is a new field that compensates the changes form the derivate. From this it can be shown
that the same identity is valid for all of the powers
n
[(Dµ ψ)0 ] = eiθ(x) (Dµ ψ)n . (3.12)
From eqs. (3.10) and (3.11) the tranformation of Aµ can be obtained:
1
Aµ → Aµ0 = Aµ − ∂ µ θ . (3.13)
q
Therefore, the new field tranform like the electromagnetic field, and the modified Lagrangian for the
fields ψ and Aµ
1 i ∗ 1
L= (Dψ)∗ · Dψ − ψ ∗ D0 ψ − D0 ψ ψ F µν Fµν , (3.14)
2m 2 4
3.2. QUANTIZATION OF THE ELECTROMAGNETIC FIELD 165
give to arise to the Scrödinger equation in presence of the electromagnetic field plus the Maxwell
with the explicit current
(
−qψ ∗ ψ ν=0
j ν = iq ∗ ∗ ∗
. (3.15)
2m
[(∇ψ) ψ − ψ ∇ψ − 2iqψ ψA] ν = i
We now turn to tue quantization of the electromagnetic field
3.2 Quantization of the electromagnetic field

Here we follow closelly [19] chapter 2.
in the electromagnetic Lagrangian the generalized momentum conjugate to the time component
of the four-vector potential is zero
∂L
π0 = 0 = 0. (3.16)
∂A
∂
∂t
Therefore, we cannot quantizate the A0 field.

The arbitrariness associated with the gauge freedom (3.13) must be removed so that the field can
be uniquely specified everywhere. Two popular choices for the gauge fixing, are the Lorentz gauge
∂µ Aµ = 0 , (3.17)
and Coulomb gauge
∇ · A = 0. (3.18)
With the Lorentz gauge a new term is added to the Lagrangian which contains the time derivative
of A0 , while in the Coulomb gauge the quantity A0 may be eliminated from the Lagrangian.
In the Coulomb gauge, we have for the A0 component
∇2 A0 = −ρ . (3.19)
For the Ai component

∂i ∂j
A (r , t), π(r, t) = i δij − 2 δ 3 (r − r0 ) ≡ δ 3T (r − r0 ) .
j 0
(3.20)
∇
These commutation relations between the creation and annihilation involve only the independent
degrees of freedom.
The Hamiltonian obtainded from the Lagrangian is
X
b=
H ωn a†n,α an,α , (3.21)
n,α
while the momentum operator is

X
b=
p kn a†n,α an,α , (3.22)
n,α
We now shown that the particles which emerges from the quantization of the electromagnetic field
(the photons) have spin one. To obtain these results, it is necessary to discuss the behavior of these
fields under rotations.
To follow the non-relativistic part of this course we recommend now go directly from section 4.1
to 4.3 where the S–matrix is defined and the probability calculated. In Section 4.5 there is the
general formula for decay. In section 5.1 the perturbative expansion of the S–matrix is presented.
Finally, in section ?? an application for the interaction of a non-relativistic atom with radiation, is
given in the context of radiative decay.
The expansion in the Coulomb gauge is
XZ eλ (p)
†

3 −ip·x ip·x
A= dp p aλ (p) e +aλ (p) e (3.23)
λ
(2π)3 2Ep
where λ = 1, 2 and eλ (p) is the polarization vector
eλ (p)eλ0 (p) = δλλ0 (3.24)

3.2. QUANTIZATION OF THE ELECTROMAGNETIC FIELD 167
See http://bolvan.ph.utexas.edu/~vadim/Classes/2015f/propagator.pdf and figures/propagator_

notes.pdf figures/propagator_notes.tex
Chapter 4
S–matrix
We will use the S–matrix formulation to obtain the decay rates and cross section formulas.
4.1 The S–matrix

The Scrödinger equation for the wave function of some state a
|a, ti ≡ ψa (t) , (4.1)
is
∂
i |a, ti = HS |a, ti , (4.2)
∂t
where HS is said to be in the Schrödinger picture where the time dependence is carried out by the
states |a, ti. In this way HS is independent of time. Therefore, the solution to this equation is
|a, ti =e−iHS (t−ti ) |a, ti i , (4.3)
since
∂
i |a, ti =i(−i)HS eHS (t−ti ) |a, ti i
∂t
=HS |a, ti . (4.4)
Defining the time-evolution operator as
U (t, ti ) = e−iHS (t−ti ) , (4.5)
169
170 CHAPTER 4. S–MATRIX
we have that in the Scrödinger picture defined by eq.(4.3), the state of a system evolves with time
|a, ti =U (t, ti )|a, ti i (4.6)
|a, ti =U (t, ti )|ai |a, ti = e−iHS (t−ti ) |ai ,
where |a, ti i, at an initial time ti , is an eigenstate of a set of conmuting operators, and is denoted
simply by |ai. Similarly |bi = |b, tf i at an final time tf .
We have then
hb, tf |a, tf i =hb|a, tf i
=hb|e−iHS (tf −ti ) |a, ti i
=hb|e−iHS (tf −ti ) |ai , (4.7)
is the amplitude for the process in which the initial state |ai evolves into the final state |bi. In the
limit tf − ti → ∞, the operator e−iHS (tf −ti ) is called the S–matrix. Therefore S is an operator that
maps an initial state to a final state
|ai → S|ai , (4.8)
an the scattering amplitudes are given by its matrix elements, hb|S|ai. Observe that
ha| → ha|S † , (4.9)
ha|ai = 1 → ha|S † S|ai = 1 , (4.10)

so that SS † = S † S = 1.
More rigorously, if ha|ai = 1, and |ni is a complete set of states, the probability that |ai evolves
into |ni, summed over all |ni, must be 1,
X
|hn|S|ai|2 = 1. (4.11)
n
On the other hand we can write

X X
|hn|S|ai|2 = ha|S † |nihn|S|ai
n n
!
X
=ha|S † |nihn| S|ai
n
†
=ha|S S|ai
=1 , (4.12)
4.2. RELATIVISTIC AND NO RELATIVISTIC NORMALIZATIONS 171
and we conclude that SS † = S † S = 1. The unitarity of the S–matrix express the conservation of
probability. It is also convenient to define the T matrix, separating the identity operator,
S = 1 + iT (4.13)
Consider a generic S–matrix element
hp1 . . . pm |S|k1 . . . kn i (4.14)
For notational simplicity the states are just labeled by their momenta, but all our considerations can
be generalized to the case in which the spin is taken into account. We have also defined the operator
T from S = 1 + iT . We assume that none of the initial momenta pj coincides a final momentum
ki . This eliminates processes in which one of the particles behaves as a “spectator” and does not
interact with the other particles. In the language of Feynman diagrams to be explained later, this
means that we will consider only connected diagrams. Therefore, if we restrict to the situation in
which no initial and final momenta coincide, the matrix element of the identity operator between
these states vanishes, and we need actually to compute the matrix element of iT
hp1 . . . pm |iT |k1 . . . kn i (4.15)
In explicit calculations there will be an overall Dirac delta factor imposing energy–momentum con-
servation. In order not to write explicitly the Dirac delta each time we compute a matrix element of
iT , it is convenient to define a matrix element Mf i from the matrix element,
!
X X
hp1 . . . pm |iT |k1 . . . kn i = (2π)4 δ (4) pj − kj iMf i . (4.16)
j j
The labels i, f refer to the initial and final states. Explicitly
Mf i = M (p1 , . . . , pm ; k1 , . . . , kn ) . (4.17)
More generally, the initial and final states are labeled also by the spin states of the initial and final
particles.
So, instead of S or T , the quantity to be calculated is Mf i , but this need first to be relativistically
normalized, in which case it will be denoted as Mf i .
4.2 Relativistic and no relativistic normalizations

We first consider a system in a cubic box with spatial volume V = L3 . At the end of the computation
V will be sent to infinity. It is sometimes convenient to put the system into a box of size L, so that the
total volume V = L3 is finite. This procedure regularizes divergences coming from the infinite-volume
limit or, equivalently, from the small momentum region, and is an example of an infrared cutoff. In
a finite box of size L, imposing periodic boundary conditions on the fields, the momenta take the
discrete values p = 2πn/L with n = (nx , ny , nz ) a vector with integer components. In non-relativistic
quantum mechanics a one-particle state with momentum p in the coordinate representation is given
by a plane wave
ψp (x) = Ceip·x , (4.18)
and the normalization constant is fixed by the condition that there is one particle in the volume V ,
Z Z
2
1= 3
d x |ψp (x)| = d3 x ψp∗ (x)ψp (x)
V V Z
=|C|2 d3 x
V
=|C|2 V , (4.19)
and
1
ψp (x) = √ eip·x . (4.20)
V
Wave functions with different momenta are orthogonal, and therefore
Z
d3 x ψp∗ 1 (x)ψp2 (x) = δp1 ,p2 (4.21)
V
R
Writing ψp (x) = hx|pi and using the completeness relation V d3 x|xihx| = 1, we can write this as
Z
NR
hp1 |p2 i =hp1 | d3 x|xihx|p2 i
Z V
3
= d xhp1 |xihx|p2 i
ZV
= d3 x ψp∗ 1 (x)ψp2 (x)
V
=δp1 ,p2 . (4.22)
The superscript NR reminds us that the states have been normalized according to the conventions
of non-relativistic quantum mechanics.
4.2. RELATIVISTIC AND NO RELATIVISTIC NORMALIZATIONS 173
In relativistic QFT this normalization is not the most convenient, because the spatial volume
V is not relativistically invariant, and therefore the condition “one-particle per volume V ” is not
invariant. A more convenient Lorentz invariant form was introduced in eq. (2.158)
hp1 |p2 iR = 2Ep1 V δp1 ,p2 (4.23)
Therefore the difference between the relativistic and non-relativistic normalization of the one-particle
states is, comparing eqs. (4.22) and (4.23)
|piR = (2Ep V )1/2 |piNR (4.24)

and of course for a multiparticle state
" #
Y
n
|p1 , . . . , pn iR = (2Ep V )1/2 |p1 , . . . , pn iNR (4.25)
i=1
We denote by Mf i , defined in eq. (4.16), the scattering amplitude between the initial state with
momenta p1 , . . . , pn and the final state with momenta k1 , . . . , kn , with non-relativistic normalization
of the states, and by Mf i the same matrix element with relativistic normalization of the states. Then
from eq. (4.16)
!
X X
4 (4)
(2π) δ pi − ki iMf i =hk1 . . . km |iT |p1 . . . pn iR
i i
Y
n Y
m
1/2
= (2Epi V )1/2 2Ekj V hk1 . . . km |iT |p1 . . . pn iNR
i=1 j=1
!
Y
n
1/2
Y
m
1/2 X X
= (2Epi V ) 2Ekj V (2π)4 δ (4) pi − ki iMf i
i=1 j=1 i i
(4.26)
Therefore
Y
n
−1/2
Y
m
−1/2
Mf i = (2Epi V ) 2Ekj V Mf i (4.27)
i=1 j=1
the matrix element,

!
Y
n
−1/2
Y
m
−1/2 X X
hk1 . . . km |iT |p1 . . . pn i = (2Epi V ) 2Ekj V (2π)4 δ (4) pj − kj iMf i .
i=1 j=1 j j
(4.28)
4.3 Process probability

Consider the matrix element of iT in (4.16)
!
X X
NR 4 (4)
hk1 . . . km |iT |p1 . . . pn i = (2π) δ pi − kj iMf i (4.29)
i j
Assume for the moment that all particles are indistinguishable. The rules of quantum mechanics tell
us that the probability of this process is obtained by taking the square module of the amplitude
! 2
X X
2
hk1 . . . km |iT |p1 . . . pn iNR = (2π)4 δ (4)
pi − kj iMf i (4.30)

i j
and we are confronted with the square of the delta function. To compute it, we recall that we are
working in a finite spatial volume and, from eq. (2.109)
(2π)3 δ (3) (0) = V (4.31)
Similarly we regularize also the time interval, saying that the time runs from −T /2 to T /2 so that
(2π)4 δ (4) (0) = V T (4.32)
Then
! 2
X
2
hk1 . . . km |iT |p1 . . . pn iNR = (2π)4 δ (4) p −
ki iMf i

j
!
X
=(2π)4 δ (4) p − kj V T Mf i (4.33)
i
Moreover we must sum over all final states. In the discrete limit, since we are working in a finite
volume V , the sum over all final states corresponds to the sum over the possible discrete values of
the momenta of the final particles

 x
2πnj nj = −∞, . . . , −1, 0, 1, . . . ∞
kj = , nyj = −∞, . . . , −1, 0, 1, . . . ∞ (4.34)
L 
 z
nj = −∞, . . . , −1, 0, 1, . . . ∞
4.3. PROCESS PROBABILITY 175
X XXX
= (4.35)
kj nx
j nyj nzj
In the large-volume limit for each particle we can write, using eq. (2.102)
X Z
V
→ 3
d3 kj , (4.36)
k
(2π)
j
The decay probability in (4.33) can be written as
X X
ω= ... hk1 . . . km |iT |p1 . . . pn iNR 2
k1 km
!
X X X X
= ... (2π)4 δ (4) pi − kj V T |Mf i |2
k1 km i i
Z Z !
V d k1 3
V d km 3 X X
= ... 3
... 3
(2π)4 δ (4) pi − kj V T |Mf i |2
(2π) (2π)
Z Z ! i i
X X 2
Ym
V d3 kj
4 (4)
= . . . (2π) δ pi − kj V T |Mf i | . (4.37)
i i j=1
(2π)3

Z Z ! !2
X X Ym
V d3 kj Y
n
−1/2
Y
m
−1/2
ω= ... (2π)4 δ (4) pi − kj V T |Mf i |2 (2Epi V ) 2Ekj V
i i j=1
(2π)3 i=1 j=1
Z Z !
X X Y
n
1 Y
m
d3 kj
= ... (2π)4 δ (4) pi − kj V T |Mf i |2 (4.38)
i i i=1
2Epi V j=1
(2π)3 2Ekj
The probability for the process of an initial particle decaying into n final particles is then
Z Z !
X 1 Y
n
d3 kj
ω1 = . . . (2π)4 δ (4) p − kj T |Mf i |2 (4.39)
i
2Ep j=1 (2π)3 2Ekj
On the other hand the probability for a process with two initial particles colliding into n final particles
is
Z Z !
X 1 1 Y
n
d3 kj
ω2 = . . . (2π)4 δ (4) p1 + p2 − kj V T |Mf i |2 3 2E
(4.40)
i
2E p 1 V 2Ep 2 V j=1
(2π) k j
A
σ
Figure 4.1: Cross section probability
4.4 Cross Section

Consider a large number point-like projectiles directed to an area A that includes a solid target of
area σ, as displayed in Fig. , such that all the points the fill the area A randomly. Assuming that
an interaction will occur (with 100% probability) if the projectile hits the solid, and not at all (0%
probability) if it misses, the total interaction probability for the single projectile will be
σ
PS = . (4.41)
A
Now suppose we have a parallel beam with density of particles n and velocity v towards the target.
In time t, this beam fills a volume
V = Avt . (4.42)
Choosing t such that the volume contains just one particle, we can write
n = 1/V , (4.43)
or
1
A= . (4.44)
nvt
replacing back in (4.41) we have
Ps
σ = PS A = . (4.45)
nvt
4.4. CROSS SECTION 177
Ps is just the decay probability in eq. (4.40). Therefore

Z Z !
ω2 1 X 2 1 1 Y
n
d3 kj
4 (4)
σ= = . . . (2π) δ p 1 + p2 − kj V T |Mf i |
nvT nvT i
2Ep1 V 2Ep2 V j=1 (2π)3 2Ekj
Z Z !
1 X 1 1 Y
n
d3 kj
= . . . (2π)4 δ (4) p1 + p2 − kj |Mf i |2 3 2E
. (4.46)
nvV i
2E p1 2E p2 j=1
(2π) kj
The density of particles of the incident state is normalized to one particle in the entire volume, so
that n = 1/V . Therefore
Z Z !
1 X 2 1 1 Y
n
d3 kj
4 (4)
σ= . . . (2π) δ p1 + p2 − kj |Mf i | 3 2E
. (4.47)
v i
2Ep1 2E p2 j=1
(2π) kj
In general, as both particles may be moving we could use the relative velocity between them, vrel ,
Z Z !
1 X 2 1 1 Y
n
d3 kj
4 (4)
σ= . . . (2π) δ p1 + p2 − kj |Mf i | 3 2E
. (4.48)
vrel i
2Ep1 2E p2 j=1
(2π) k j
Since
p =γmv , E =γm , (4.49)
then
p
v= . (4.50)
E
In a frame where p1 and p2 are along the same line, this reduces to

p1 p 2
vrel = − . (4.51)
E1 E2
In fact, for not relativistic particles, where Ei = mi , this coincides with the usual relative velocity

m1 v1 m2 v2

vrel = −
E1 E2
= |v1 − v2 | . (4.52)
The most general formula for the relative velocity is
I
vrel = (4.53)
E1 E2
where
q
I = (p1 · p2 )2 − m21 m22 (4.54)
In general
q
I = (E1 E2 − p1 · p2 )2 − m21 m22
q
= E12 E22 + (p1 · p2 )2 − 2E1 E2 p1 .p2 − m21 m22 (4.55)
Since
m21 m22 =(E12 − p21 )(E22 − p22 )
=(E12 E22 − p21 E22 − E12 p22 + p21 p22 ) (4.56)
q
I = p21 E22 − 2E1 E2 p1 · p2 + E12 p22 + (p1 · p2 )2 − p21 p22 (4.57)
If
(p1 · p2 )2 − p21 p22 = 0 (4.58)
that implies that p1 and p2 are colineals,
q
I = p21 E22 − 2E1 E2 p1 · p2 + E12 p22
p
= (p1 E2 − p2 E1 )2
=|p1 E2 − p2 E1 | (4.59)

I p1 p 2
vrel = = − (4.60)
E1 E2 E1 E2
To simplify the notation we set Ei = Epi =, and Ef = Epf . Moreover, the differential cross
section is
2
!
X X 1 Y d3 kf
dσ =(2π)4 δ (4) pi − pf |Mf i |2
i=1 f
4vrel E1 E2 f
(2π)3 2Ef
1
= (2π)4 |Mf i |2 dΦn (p1 , p2 ; k1 , . . . , kn ) (4.61)
4vrel E1 E2
where
!
X Y
n
d3 kj
(n) (4)
dΦ (p1 , p2 ; k1 , k2 , . . . , kn ) = δ p− kj 3 2E
. (4.62)
j j=1
(2π) k j
We keep the diferential notation both for dσ, and dΦ until the last integration have been made.
4.4.1 2–to–2 cross section

The the 2–to–2 cross section is
(2π)4
dσ = |Mf i |2 dΦ2 (p1 , p2 ; p01 , p02 )
4vrel E1 E2
24 π 4
= 8 6 28 π 6 |Mf i |2 dΦn (p1 , p2 ; k1 , . . . , kn )
2 π 4vrel E1 E2
1
= 6 2 |Mf i |2 4(2π)6 dΦ2 (p1 , p2 ; p01 , p02 )
2 π vrel E1 E2
1
= 2
|Mf i |2 4(2π)6 dΦ2 (p1 , p2 ; p01 , p02 ) (4.63)
64π vrel E1 E2
where, as in eq. (4.116)
3 0 3 0
4(4π)6 (4) 0 d p1 d p2
4(4π)6 dΦ(2) (p1 , p2 ; p01 , p02 ) = δ (p 1 + p 2 − p 0
1 − p 2 )
4(2π)6 E10 E20
d3 p0 d3 p0
= δ (4) (p1 + p2 − p01 − p02 ) 4(2π)6 0 1 0 2 (4.64)
E1 E2
We now will find an expression for cross section in the center of mass frame (CM)
The center of mass (CM) frame is defined by the condition
p1 + p2 = 0 (4.65)
The δ–function in Eq. (4.64)
δ (4) (p + p2 − p01 − p02 ) = δ (3) (p1 + p2 − p01 − p02 )δ(E1 + E2 − E10 − E20 ) (4.66)
In the CM frame
δ (4) (p + p2 − p01 − p02 ) = δ (3) (p01 + p02 )δ(E1 + E2 − E10 − E20 ) (4.67)
Mf i in integration does not depend on |p01 | or |p02 | as the final momentum is fixed by the initial
momentum whenever the final states have only two particles. In this way the integration on p02 can
be evaluated directly for dΦ(2) . Replacing back in Eq. (4.63)
d3 p0 d3 p0
4(2π)6 dΦ(2) =δ (3) (p01 + p02 )δ(E1 + E2 − E10 − E20 ) 0 1 0 2
E1 E2
3 0 Z
dp d3 p 0
=δ(E1 + E2 − E10 − E20 ) 0 1 δ (3) (p01 + p02 ) 0 2
E1 E2
3 0
dp
=δ(E1 + E2 − E10 − E20 ) 0 10 (4.68)
E1 E2
p01 2 d|p01 |dΩ

4(2π)6 dΦ(2) = δ(E1 + E2 − E10 − E20 ) (4.69)
E10 E20
As
q
|p01 | = E10 2 − m1 2 (4.70)
d|p01 | 2E10
= q
dE10
2 E10 2 − m1 2
E10
= (4.71)
|p01 |
In this way, we can write, in general
|p| d|p| = E dE (4.72)
and
|p01 |E10 dE10
4(2π)6 dΦ(2) = δ(E1 + E2 − E10 − E20 ) dΩ
E10 E20
|p0 |dE 0
= δ(E1 + E2 − E10 − E20 ) 1 0 1 dΩ (4.73)
E2
From the δ–function in Eq. (??) we have that in the CM frame
p1 + p2 =0 , p01 + p02 =0 . (4.74)
Squaring the first expression, and taking into account that

q
p1 = E10 2 − m01 2
0
(4.75)
we have
2 2
p01 =p02
2 2 2 2
E10 − m01 =E20 − m02 , (4.76)
q
E20 = E10 2 − m01 2 + m02 2 (4.77)
In this way we can express E20 in terms of E10 in Eq. (4.73). Moreover, we can define the center of
mass energy as
s = (p1 + p2 )2 . (4.78)
By using the center of mass condition (4.65)
p1 = −p2 , (4.79)
then
s =p21 + p22 + 2p1 · p2

=p21 + p22 + 2 (E1 E2 − p1 · p2 )
=E1 − p21 + E2 − p22 + 2E1 E2 − 2p1 p2
=E1 − p21 + E2 − p21 + 2E1 E2 + 2p21
=E12 + 2E1 E2 + E22
= (E1 + E2 )2 . (4.80)
So that
√
s = E1 + E2 (4.81)
Using The energy part of δ–function in Eq. (??) can be written as

q
0 0
√ 0 2 2 2
δ (E1 + E2 − E1 − E2 ) = δ s − E1 − E10 − m01 + m02 . (4.82)
As established before, Mf i in this case in independent of |p01 |, and the integration on E10 can be done
directly only for dΦ(2) . The integral is easily performed using the identity
X δ(z − zn )
δ (f (z)) = (4.83)
n
|f 0 (zn )|
where zn are the zeroes of f (z). In this case, this δ–function is a function of the integration variable
E10 , with only one zero
δ(x − x0 )
δ (f (x)) = (4.84)
|f 0 (x0 )|
where
q
√
f (x) = s−x− x2 − m01 2 + m02 2 (4.85)
The zero is obtained from
q
√
s − x0 − x20 − m01 2 + m02 2 = 0
√ 2 2
s − 2 s x0 + x20 = x20 − m01 + m02
√ 2 2
s − 2 s x0 = −m01 + m02 , (4.86)
with solution
s + m01 2 − m02 2
x0 = √ . (4.87)
2 s
Therefore
Z
6 (2) δ(x − x0 ) |p01 (x)|
4(2π) dΦ =dΩ dx
|f 0 (x0 )| E20 (x)
1 |p01 (x0 )|
=dΩ 0 , (4.88)
|f (x0 )| E20 (x0 )
where from Eqs. (4.75), (4.77),
q q
p1 (x0 ) = x20 − m01 2
0
E20 (x0 ) = x20 − m01 2 + m02 2 (4.89)
As (See deltaxn.nb for additional details)
x
f 0 (x) = − q −1 (4.90)
x2 − m01 2 + m02 2
we have
0 m01 2 − m02 2 + s
f (x0 ) = − r −1
√ (−m01 2 +m02 2 +s)
2
s s
−m01 2 + m02 2 − s
= −1
−m01 2 + m02 2 + s
−m01 2 + m02 2 − s + m01 2 − m02 2 − s
=
−m01 2 + m02 2 + s
−2s
= , (4.91)
s + m02 2 − m01 2
and

s + m02 2 − m01 2
δ(f (E10 )) = δ(E10 − x0 ) (4.92)
2s
Replacing the expression for x0 in (4.87) into Eq. (4.89) we have (See deltaxn.nb for additional
details)
p
0 [s − (m01 − m02 )2 ][s − (m01 + m02 )2 ]
|p1 (x0 )| = √
2 s
0 s − m01 2 + m02 2
E2 (x0 ) = √ (4.93)
2 s
Replacing Eqs. (4.91), and (4.93) in Eq. (4.88) we have
q
1 x20 − m01 2
6 (2)
4(2π) dΦ = dΩ 0 q
|f (x0 )|
x20 − m01 2 + m02 2
p
s − m01 2 + m02 2 [s − (m01 − m02 )2 ][s − (m01 + m02 )2 ]
= dΩ
2s s − m01 2 + m02 2
p
[s − (m01 − m02 )2 ][s − (m01 + m02 )2 ]
= dΩ (4.94)
2s
Defining the kinematic two particle function
λ(a, b, c) ≡ (a − b + c)2 − 4ac (4.95)
and taking into account that

2
2 2 2
s − m02 + m01 − 4sm01 = [s − (m01 − m02 )2 ][s − (m01 + m02 )2 ] (4.96)
we have
6 (2) λ1/2 (s, m02 2 , m01 2 )

4(2π) dΦ = dΩ (4.97)
2s
Moreover
λ1/2 (s, m02 2 , m01 2 )
p01 = √ (4.98)
2 s
To further evaluate Eq. (4.63), we need to express vrel and E1 E2 in terms of s and the masses.
Concerning vrel , from Eq. (4.60), evaluated in CM frame

p1 p2
E1 E2 vrel =E1 E2 −
E1 E2

p1 p1
=E1 E2 +
E1 E2
= |p1 | (E1 + E2 )
√
= |p1 | s (4.99)
Replacing back Eqs. (4.94), and (4.99) into Eq. (4.63), we have
1 2 6 (2)

dσ = |Mf i | 4(2π) dΦ (4.100)
64π 2 E1 E2 vrel
p
dσ 1 [s − (m01 + m02 )2 ][s − (m01 − m02 )2 ]
= 2
|Mf i |2 (4.101)
dΩ 64π E1 E2 vrel 2s
By using Eq. (4.99)

q
dσ 1 2
[s − (m01 + m02 )][s − (m01 2 − m02 2 )]
= √ |Mf i | (4.102)
dΩ 64π 2 |p1 | s 2s
In the CM frame
√
s =E1 + E2
q q
= p21 + m21 + p22 + m22
q q
= p1 + m1 + p21 + m22
2 2
(4.103)
q
s =2p21
+ + m21 m22 + 2 p41 + (m21 + m22 )p21 + m21 m22
q
s − (2p21 + m21 + m22 ) =2 p41 + (m21 + m22 )p21 + m21 m22 (4.104)
4.5. DECAY RATES 185
s2 − 2s(2p21 + m21 + m22 ) + [2p21 + (m21 + m22 )]2 =4(p41 + (m21 + m22 )p21 + m21 m22 )
s2 − 2s(2p21 + m21 + m22 ) + 4p41 + 4p21 (m21 + m22 ) + (m21 + m22 )2 =4(p41 + (m21 + m22 )p21 + m21 m22 )
−4sp21 + s2 − 2s(m21 + m22 ) + (m21 + m22 )2 =4m21 m22
−4sp21 + s2 − 2sm21 − 2sm22 + m41 + m42 + 2m21 m22 =4m21 m22
−4sp21 + s2 − 2sm21 − 2sm22 + m41 + m42 − 2m21 m22 =0 (4.105)
(s − m21 − 2m2 m1 − m22 ) (s − m21 + 2m2 m1 − m22 )

p21 = (4.106)
4s
p
[s − (m1 + m2 )2 ][s − (m1 − m2 )2 ]
|p1 | = √
2 s
λ1/2 (s, m22 , m21 )
= √ (4.107)
2 s
Replacing Eq. (4.107) back in Eq. (4.99) we have
1p
E1 E2 vrel = [s − (m1 + m2 )2 ][s − (m1 − m2 )2 ] (4.108)
2
Replacing Eqs. (4.108), and (4.94) in Eq. (4.63)
s
1 2 dΩ [s − (m01 + m02 )2 ][s − (m01 − m02 )2 ]
dσ = |M fi | (2) (4.109)
64π 2 2s [s − (m1 + m2 )2 ][s − (m1 − m2 )2 ]
and, finally
1/2
dσ 1 [s − (m01 + m02 )2 ][s − (m01 − m02 )2 ]
= |M|2 , (4.110)
dΩ 64π 2 s [s − (m1 + m2 )2 ][s − (m1 − m2 )2 ]
or, in terms of the kinematic function defined in eq. (4.145)
dσ 1 λ1/2 (s, m02 2 , m01 2 )
= |M|2 (4.111)
dΩ 64π 2 s λ1/2 (s, m22 , m21 )
4.5 Decay Rates

Consider the matrix element of iT in (4.16)
!
X
hk1 . . . kn |iT |piNR = (2π)4 δ (4) p − ki iMf i (4.112)
i
where the initial state is a single particle of momentum p and mass M , while the final state is
given by n particles of momenta ki and masses mi , i = 1, . . . , n. We are therefore considering a decay
process.
Z Z ! Z
X 2 1
Yn
d3 kj
4 (4)
ω = . . . (2π) δ p− kj dt |Mf i | (4.113)
i
2Ep j=1 (2π)3 2Ekj
Therefore the differential probability is

!
X 1 2
Yn
d3 kj
4 (4)
dω =(2π) δ p− kj |Mf i | dt (4.114)
i
2Ep j=1
(2π)3 2Ekj
Finally we define the decay rate dΓ as the decay probability in which in the final state the j–th
particle has momentum between kj and kj + dkj per unit time
!
dω X 1 2
Yn
d3 kj
4 (4)
dΓ ≡ = (2π) δ p− kj |Mf i |
dt j
2Ep j=1
(2π)3 2Ekj
(2π)4
= |Mf i |2 dΦ(n) (p; k1 , k2 , . . . , kn ) (4.115)
2Ep
where
!
X Y
n
d3 kj
(n) (4)
dΦ (p; k1 , k2 , . . . , kn ) = δ p− kj . (4.116)
j j=1
(2π)3 2Ekj
and the differential decay width in the center of mass frame
(2π)4
dΓ = |Mf i |2 dΦ(n) (p; k1 , k2 , . . . , kn ) (4.117)
2Ep
4.5.1 Two body decays

We now consider the decay of particle of mass M decaying into two particles of 4–momenta p1 , p2
and masses m1 , m2 . In the CM frame the initial momentum satisfy
p =0 ⇒ M = Ep (4.118)
Therefore
(2π)4
dΓ = 6
|Mf i |2 4(2π)6 dΦ(2) (p; p1 , p2 )
2M [4(2π) ]
1
= 3 |Mf i |2 4(2π)6 dΦ(2) (p; p1 , p2 )
2 M (2π)2
1
= 2
|Mf i |2 4(2π)6 dΦ(2) (p; p1 , p2 ) (4.119)
32π M
where
d3 p 1 d3 p2
4(2π)6 dΦ(2) (p; p1 , p2 ) = δ (4) (p − p1 − p2 ) (4.120)
E1 E2
The Dirac delta in eq. (4.116) can be written in the CM frame as
δ (4) (p − p1 − p2 ) =δ (3) (p − p1 − p2 )δ(E − E1 − E2 )

=δ (3) (p1 + p2 )δ(M − E1 − E2 ) (4.121)
and,
d3 p1 d3 p2
4(2π)6 dΦ(2) (p; p1 , p2 ) =δ(M − E1 − E2 )δ (3) (p1 + p2 ) (4.122)
E1 E2
1
We now proceed with a calculation similar to the one leading to eq. (4.68)
d3 p1 d3 p 2
4(2π)6 dΦ(2) =δ (3) (p1 + p2 )δ(M − E1 − E2 )
E1 E2
Z
d3 p1 d3 p2
=δ(M − E1 − E2 ) δ (3) (p1 + p2 )
E1 E2
3
d p1
=δ(M − E1 − E2 ) (4.123)
E1 E2
p1 2 d|p1 |dΩ
4(2π)6 dΦ(2) = δ(M − E1 − E2 ) (4.124)
E1 E2
As
p
|p1 | = E1 2 − m1 2 (4.125)
√
1
we see that the two quantities are the same after the replacing s → M , p01 → p1 and p02 → p2 .
d|p1 | 2E1
= p
dE1 2 E1 2 − m1 2
E1
= (4.126)
|p1 |
In this way, we can write, in general
|p| d|p| = E dE (4.127)
and
|p1 |E1 dE1
4(2π)6 dΦ(2) = δ(M − E1 − E2 ) dΩ
E1 E2
|p1 |dE1
= δ(M − E1 − E2 ) dΩ (4.128)
E2
From the δ–function in Eq. (4.66) we have that in the CM frame
CM
p1 − p2 = 0 ⇒ p1 = −p2 (4.129)
Squaring the first expression, and taking into account that

p
p1 = E1 2 − m1 2 (4.130)
we have
p1 2 =p2 2
E1 2 − m1 2 =E2 2 − m2 2 , (4.131)
p
E2 = E1 2 − m1 2 + m2 2 (4.132)
In this way we can express E2 in terms of E1 in Eq. (4.73).

Using The energy part of δ–function in Eq. (4.66) can be written as
p
δ M − E1 − E1 2 − m1 2 + m2 2 (4.133)
As established before, Mf i in this case in independent of |p1 |, and the integration on E1 can be done
directly only for dΦ(2) .
The integral is easily performed using the identity

δ(x − x0 )
δ (f (x)) = (4.134)
|f 0 (x0 )|
where
p
f (x) = M − x − x2 − m1 2 + m2 2 (4.135)
and the root is given by (4.87)
M 2 + m1 2 − m2 2
x0 = (4.136)
2M
Therefore
Z
6 (2) δ(x − x0 ) |p1 (x)|
4(2π) dΦ =dΩ dx
|f 0 (x0 )| E2 (x)
1 |p1 (x0 )|
=dΩ 0 , (4.137)
|f (x0 )| E2 (x0 )
where
q q
p1 (x0 ) = x20 − m1 2 E2 (x0 ) = x20 − m1 2 + m2 2 (4.138)
As (See deltaxn.nb for additional details)

x
f 0 (x) = − √ −1 (4.139)
x2 − m1 2 + m2 2
we have as in (4.91)
0 −2M 2
f (x0 ) = 2 , (4.140)
M + m2 2 − m1 2
and

M 2 + m2 2 − m1 2
δ(f (E1 )) = δ(E1 − x0 ) (4.141)
2M 2
Replacing the expression for x0 in into Eq. (4.138) we have (See deltaxn.nb for additional details)
p
[M 2 − (m1 − m2 )2 ][M 2 − (m1 + m2 )2 ]
|p1 (x0 )| =
2M
M 2 − m1 2 + m2 2
E2 (x0 ) = (4.142)
2M
Replacing Eqs. (4.140), and (4.142) in Eq. (4.137) we have

p
6 (2) 1 x20 − m1 2
4(2π) dΦ = dΩ 0 p
|f (x0 )| x20 − m1 2 + m2 2
2 p 2
M − m1 2 + m2 2 [M − (m1 − m2 )2 ][M 2 − (m1 + m2 )2 ]
= dΩ
2M 2 M 2 − m1 2 + m2 2
p
[M 2 − (m1 − m2 )2 ][M 2 − (m1 + m2 )2 ]
= dΩ (4.143)
2M 2
Since
2
M − (m1 − m2 )2 M 2 − (m1 + m2 )2

= M 4 − M 2 m21 + 2m1 m2 + m22 − M 2 m21 − 2m1 m2 + m22 + [(m1 − m2 ) (m1 + m2 )]2
2
= M 4 − M 2 m21 − 2M 2 m1 m2 − M 2 m22 − M 2 m21 + 2M 2 m1 m2 − M 2 m22 + m21 − m22
= M 4 − 2M 2 m21 − 2M 2 m22 + m41 − 2m21 m22 + m42
= M 4 − 2M 2 m21 + 2M 2 m22 + m41 + m42 − 2m21 m22 − 4M 2 m22
2
= M 4 − 2M 2 m21 − m22 + m21 − m22 − 4M 2 m22
2
= M 2 − m21 − m22 − 4M 2 m22
2
= M 2 − m21 + m22 − 4M 2 m22 . (4.144)
we can define the kinematic two particle function
λ(a, b, c) ≡ (a − b + c)2 − 4ac . (4.145)
Then we have
λ1/2 (M 2 , m2 2 , m1 2 )
4(2π)6 dΦ(2) = dΩ (4.146)
2M 2
Finally we get a similar result than in eq. (4.97)
λ1/2 (M 2 , m22 , m21 )
4(2π)6 dΦ(2) = dΩ (4.147)
2M 2
Replacing back in eq. (4.119)
1/2
dΓ 1 2 λ (M 2 , m22 , m21 )
= |M f i |
dΩ 32π 2 M 2M 2
1
= |Mf i |2 λ1/2 (M 2 , m22 , m21 ) (4.148)
64π 2 M 3
By using eq. (4.142) we can write this expression also as

dΓ 1
= |Mf i |2 2M |p1 |
dΩ 64π 2 M 3
|p1 |
= 2 2
|Mf i |2 (4.149)
32π M
as usually written in several texts.
Recommended extra material: [7, 8]
4.5.2 Example
Energy violation of the neutron decay in the rest frame. In absence of neutrino we can calculate the
energy of the final estate electron. From (4.142) we have
p
[M 2 − (mp − me )2 ][M 2 − (mp + me )2 ]
|pe (x0 )| = |pp (x0 )| =
2M
M 2 − mp 2 + me 2
Ee (x0 ) = (4.150)
2M
where pp (pe ) is the proton (electron) momentum. Moreover
pe =γme v
2 m2e v 2
pe =
1 − v2
pe (1 − v ) =m2e v 2
2 2
p2e − p2e v 2 =m2e v 2
Therefore
(p2e + m2 )v 2 =p2e
pe
v =p
p2e+ m2e
import math as np
M=939.57 #MeV
mp=938.28 #MeV
me=0.511 #MeV
pe=np.sqrt( (M**2- (mp-me)**2 )*(M**2 - (mp+me)**2 ) )/(2.*M)

print(’p_e=’,pe)
Ee=(M**2-mp**2+me**2)/(2.*M)
print(’E_e=’,Ee)
print(’m_e=’,np.sqrt(Ee**2-pe**2))
ve=pe/(np.sqrt(pe**2+ me**2 ) )
print(’v_e=’,ve)
p_e= 1.183660978265827
E_e= 1.2892533930415913
m_e= 0.5110000000000403
v_e=0.92 #c
|pe (x0 )| =1.184 MeV

Ee (x0 ) =1.289 MeV (4.151)
so that me = 0.511 MeV as a crosschek. In this way, the speed of the electron is v = 0.92c.
What is measured is the knetic energy of the electron2
me
Ke = − me = 0.78 MeV . (4.152)
1 − ve2
What is observed however is that the kinetic energy of the electron from neutron decay is a
distribution with a tail at 0.78MeV as in Fig. 4.2, taken from http://hyperphysics.phy-astr.gsu.
edu/hbase/Particles/proton.html. This imply that decay must involve an additional particle: the
electronic-antineutrino.
4.6 Backup
Perturbation theory is developed more easily using the Hamiltonian formalism. We therefore consider
a general field theory with a Hamiltonian
H = H0 + Hint (4.153)
2
The non-relativistic limit is K ≈ m(1 + v 2 /2) − m = mv 2 /2.
4.6. BACKUP 193
Figure 4.2: Electron properties in neutron decay in the rest frame. From http://hyperphysics.
phy-astr.gsu.edu/hbase/Particles/proton.html
where H0 is the free Hamiltonian and Hint is the interaction term. The interaction term will be
considered small. For instance in QED
Z Z
3
Hint = d x Hint = − d3 x Lint (4.154)
with
Lint = −eAµ ψγ µ ψ (4.155)
The smallnes of the interaction follows from the fact that the parameter which turns out to be
relevatn for the perturbation expansion is α = e2 /4π ≈ 1/137.
SS † =(1 + iT )(1 − iT † )
= 1 + i(T − T † ) + T T † = 1 ,
(4.156)
T T † = −i(T − T † ) . (4.157)
Inserting a complete set of states we have
hb|T T † |ai = − i(hb|T |ai − hb|T † |ai)

!
X h i
hb|T |nihn| T † |ai = − i hb|T |ai − (ha|T |bi)†
n
X
hb|T |niha|T |ni† = − i (hb|T |ai − ha|T |bi∗ )
n
X
∗ ∗
Tbn Tan = − i (Tba − Tab ).
n
(4.158)
if a = b
|Tan |2 = −i Im Taa . (4.159)

Chapter 5
Feynman Rules
When the case of interacting fields are considered, the particles can be created, destroyed and scat-
tered. In essence this requires solving the coupled non-linear field equations for given conditions.
This is an extremely difficult problem which has only been solved in perturbation theory.
In the Heisenberg picture, which we have so far been using, this program is still very complex, and
it was decisive for the successful development of the theory to work instead in the interaction picture.
In section 5.1 we write the S–matrix expansion derived in Chapter 4, in the interaction picture. In
section 5.4 we show how to use the Wick expansion to calculate S–matrix elements involving scalars
and spinors.
The considerations which allow to write a formal series of the evolution operator are completely
general and apply just as well the perturbative calculation of the scattering operator as to the Green
functions of the many-body problem in Appendix ??.
5.1 Interaction picture

This part is based in [3].
The system is decribed by a Hamiltonian H = H0 + HI , where H0 is the non-perturbated
Hamiltonian and HI is the interaction Hamiltonian (which can depend on time).
In the Schrödinger Picture (SP) the time dependence is carried by the states according to the
Scrödinger equation
∂ d
i |a, tiS = i |a, tiS = H|a, tiS (5.1)
∂t dt
With the solution given in Eq. (4.7)
|a, tiS = U (t, ti )|aiS . (5.2)
195
196 CHAPTER 5. FEYNMAN RULES
where U is the unitary operator [see Eq. (4.5)]
U ≡ U (t, ti ) = e−iH(t−ti ) . (5.3)
Given the state |a, tiS in the SP, in the Heisenberg picture (HP) we defined the state
|aiH = U † |a, tiS = |aiS (5.4)
Si OS in an operator in the SP, the corresponding Heisenberg operator is defined as
OH (t) = U † OS U (5.5)
Hence, the transformation from HP to SP is unitary. At t = ti , states and operators in the two
pictures are the same. We see from Eq. (5.4) that in the HP state vectors are constant in time, while
from Eq. (5.5) the Heisenberg operators evolve with time. It is convenient to keep the temporal label
in the Heisenberg states
|aiH = |a, ti iH (5.6)
Eq. (5.5) ensures the invariance of matrix elements and commutation relations:
S hb, t| O
S
|a, tiS = S hb, t| U OH (t)U † |a, tiS = H hb, ti |OH (t)|a, ti iH (5.7)

OS , P S = c ⇒ OH (t), P H (t) = c (5.8)
where c is a constant.
Differentiation of Eq. (5.5)

d H d † d
O (t) = U OS U + U † OS U
dt dt dt
† S † S
=iH U O U + U O U (−iH)
= − i(OH H − HOH ) , (5.9)
gives the Heisenberg equation of motion
d H
i O (t) = OH (t), H (5.10)
dt
The interaction picture (IP) arises if the Hamiltonian is split into two parts
H = H0 + Hint . (5.11)
5.1. INTERACTION PICTURE 197
In quantum field theory Hint will describe the interaction between fields1 , themselves described by
H0
IP is related to the SP by the unitary transformations
U0 ≡ U0 (t, ti ) = e−iH0 (t−ti )

Uint ≡ Uint (t, ti ) = e−iHint (t−ti ) , (5.12)
along with2 ,
|a, tiI = U0† |a, tiS , (5.13)
and
OI (t) = U0† OS U0 . (5.14)
Therefore, the relation between IP and SP is similar to that between HP and SP, but with the unitary
transformation U0 involving only the non–interacting Hamiltonian H0 . Note that both the vector
states as the operators in the IP are time-dependent.
Differentiating Eq. (5.14) gives the differential equation of motion operators in the IP:
d I
i O (t) = OI (t), H0 (5.15)
dt
Substituting Eq. (5.13) into the Scrödinger Eq. (5.1), one obtains the equation of motion of state
1
Corresponding to a Lagrange density with more than two fields
2
Colors in the hamiltonians are dropped from here.
vectors in the IP, If the system is described by a time-dependent state vector |Φ(t)i
d
i |a, tiS =H S |a, tiS
dt
d
i (U0 |Φ(t)i) =H S U0 |Φ(t)i
dt
d d
i U0 |Φ(t)i + iU0 |Φ(t)i =H S U0 |Φ(t)i
dt dt
d
H0 U0 |Φ(t)i + iU0 |Φ(t)i =H S U0 |Φ(t)i
dt
d S
H0 U0 |Φ(t)i + iU0 |Φ(t)i =(H0 + Hint )U0 |Φ(t)i
dt

d S
H 0 |Φ(t)i + iU0
0 U |Φ(t)i =
H 0 |Φ(t)i + Hint U0 |Φ(t)i
0 U
dt
d S
iU0 |Φ(t)i =Hint U0 |Φ(t)i
dt
d S
i |Φ(t)i =U0 Hint U0 |Φ(t)i (5.16)
dt
d I
i |Φ(t)iI = Hint |Φ(t)iI , (5.17)
dt
where, as in Eq. (5.14)
I S S
S −iH0 (t−ti )
Hint = eiH0 (t−ti ) Hint e (5.18)
S
is the interaction Hamiltonian in the IP, with Hint and H0S being the interaction and free-field Hamil-
tonian in the SP. From now on we shall omit the labels I, used in the equations to distinguish the
IP, as we shall be working exclusively in the IP in what follows.
Eq. (5.17) is a Scrödinger-like equation with the time dependent Hamiltonian Hint (t). With the
interaction switched off (i.e. we put Hint = 0), the state vector is constant in time. The interaction
leads to the state |Φ(t)i changing with time. Given that the system is in a state |ii at an initial time
t = ti , i.e.
|Φ(ti )i = |ii , (5.19)
the solution of Eq. (5.17) with this initial condition, gives the state |Φ(t)i of the system at any other
time t. It follows from the Hermicity of the operator Hint (t) that the time development of the state
|Φ(t)i according to Eq. (5.17) is a unitary transformation. Accordingly it preserves the normalization
of states
hΦ(t)|Φ(t)i = const. (5.20)
and, more generally, the scalar product.

Clearly the formalism which we are developing here is not appropriate for the description of
bound states but it is particularly suitable for scattering processes. In a collision processes the state
vector |ii will define an initial state, long before the scattering occurs (ti = −∞), by specifying a
definite number of particles, with definite properties and far apart from each other so that they do
not interact. (For example |ii would specify a definite number of electrons, and positrons with given
momenta and spins). In the scattering process, the particles will come close together, collide (i.e
interact) and fly apart gain. Eq. (5.17) determines the state |Φ(t)i into which the initial state
|Φ(−∞)i = |ii , (5.21)
evolves at t = ∞, long after the scattering is over and all particles are for apart again. The S–matrix
relates |Φ(∞)i to Φ(−∞) and is defined by
|Φ(∞)i = S|Φ(−∞)i = S|ii , (5.22)
A collision can lead to many different final states |f i, and all these possibilities are constrained
within |Φ(∞)i.
The transition probability is given by
|hf |Φ(∞)i|2 = |hf |S|ii|2 ≡ Sf2i , (5.23)
where Sf i is the corresponding probability amplitude.

In order to calculate the S–matrix we must solve Eq. (5.17) for the initial condition (5.19). These
equations can be combined into the integral equation
d|Φ(t)i = − i d t Hint (t)|Φ(t)i

Z |Φ(t)i Z t
d|Φ(t)i = − i d t1 Hint (t1 )|Φ(t1 )i
|Φ(−∞)i ∞
Z t
|Φ(t)i − |Φ(−∞)i = − i d t1 Hint (t1 )|Φ(t1 )i
∞
(5.24)
Z t
|Φ(t)i = |ii − i d t1 Hint (t1 )|Φ(t1 )i . (5.25)
−∞
In the limit t → ∞
Z ∞
(0)
|Φ(∞)i = S |ii − i d t1 Hint (t1 )|Φ(t1 )i . (5.26)
−∞
where
S (0) = 1 . (5.27)
From Eq. (5.25) we can obtain |Φ(t1 )i at next order:

Z t1
|Φ(t1 )i =|ii − i d t2 Hint (t2 )|Φ(t2 )i . (5.28)
−∞
This equation then can be solved iteratively. If Hint is small we can solve this equation by iteration
Z t Z t Z t1
2
|Φ(t)i = |ii + (−i) d t1 Hint (t1 )|ii + (−i) d t1 d t2 Hint (t1 )Hint (t2 )|Φ(t2 )i . (5.29)
−∞ −∞ −∞
Z ∞ Z ∞ Z t1
(0) 2
|Φ(∞)i = S + (−i) d t1 Hint (t1 ) |ii + (−i) d t1 d t2 Hint (t1 )Hint (t2 )|Φ(t2 )i
−∞ −∞ −∞
Z ∞ Z t1
(0) (1)
2
= S +S |ii + (−i) d t1 d t2 Hint (t1 )Hint (t2 )|Φ(t2 )i , (5.30)
−∞ −∞
where
Z ∞
(1)
S = (−i) d t1 Hint (t1 ) . (5.31)
−∞
For an infinity hipervolume we have

Z
S (1)
= (−i) d4 x Hint (t1 ) . (5.32)
The next order of Eq. (5.29) is

Z t Z t Z t1
2
|Φ(t)i =|ii + (−i) d t1 Hint (t1 )|ii + (−i) d t1 d t2 Hint (t1 )Hint (t2 )
−∞ −∞ −∞
Z t2 Z t2 Z t3
2
× |ii + (−i) d t3 Hint (t3 )|ii + (−i) d t3 d t4 Hint (t3 )Hint (t4 )|Φ(t4 )i (5.33)
−∞ −∞ −∞
Z t Z t Z t1
2
|Φ(t)i =|ii + (−i) d t1 H1 (t1 )|ii + (−i) d t1 d t2 Hint (t1 )Hint (t2 )|ii
−∞ −∞ −∞
Z t Z t1 Z t2
3
+ (−i) d t1 d t2 d t3 Hint (t1 )Hint (t2 )H1 (t3 )|ii
−∞ −∞ −∞
Z t Z t1 Z t2 Z t3
4
+ (−i) d t1 d t2 d t3 d t4 Hint (t1 )Hint (t2 ) Hint (t3 )Hint (t4 )|Φ(t4 )i (5.34)
−∞ −∞ −∞ −∞

|Φ(t)i = S (0) + S (1) + S (2) + S (3) |ii
Z ∞ Z t1 Z t2 Z t3
4
+ (−i) d t1 d t2 d t3 d t4 Hint (t1 )Hint (t2 ) Hint (t3 )Hint (t4 )|Φ(t4 )i (5.35)
−∞ −∞ −∞ −∞
where
Z ∞ Z t1
(2) 2
S =(−i) d t1 d t2 Hint (t1 )Hint (t2 )
−∞ −∞
Z ∞ Z t1 Z t2
S (3) =(−i)3 d t1 d t2 d t3 Hint (t1 )Hint (t2 )H1 (t3 ) (5.36)
−∞ −∞ −∞
and so on we obtain the S–matrix

∞
X
S= S (n)
n=0
∞
X Z ∞ Z t1 Z tn−1
n
=1 + (−i) d t1 d t2 . . . d tn Hint (t1 )Hint (t2 ) . . . Hint (tn ) . (5.37)
n=1 −∞ −∞ −∞
5.2 Useful identities

The following identities can be used to simplify products of σ, σ matrices
σαµα̇ σ β̇β β β̇
µ =2δα δ β ,
(σ µ )α̇α (σ µ )β̇β =αβ α̇β̇ (5.38)
Tr [σ µ σ ν ] =2g µν , (5.39)
Tr [σ µ σ ν σ ρ σ κ ] =2 (g µν g ρκ − g µρ g νκ + g µκ g νρ + iµνρκ ) . (5.40)
We use the notation
a · b = aµ b µ . (5.41)
X X
xα (p, s)x†β̇ (p, s) =p · σαβ̇ , x†α̇ (p, s)xβ (p, s) =p · σ α̇β
s s
X X
yα (p, s)yβ̇† (p, s) =p · σαβ̇ , y †α̇ (p, s)y β (p, s) =p · σ α̇β (5.42)
s s
X X
xα (p, s)y β (p, s) =mδαβ , yα (p, s)xβ (p, s) = − mδαβ
s s
X X
y †α̇ (p, s)χ†β̇ (p, s) =mδβ̇α̇ , x†α̇ (p, s)yβ̇† (p, s) = − mδβ̇α̇ (5.43)
s s
For the demonstration of the last identity we follow the steps in [26].
5.3 Dirac delta function

Consider the next identity, i, j, l, m = 1, 2, and where we will simplify the δ-function by eliminating
q
Z Z
0 0
d4 x1 d4 x2 e−i(−q−pi +pj )·x1 e−i(q−pl +pm )·x2 =δ 4 (−q − p0i + pj )δ 4 (q − p0l + pm )
=δ 4 (−p0i + pj − p0l + pm )
=δ 4 (pj + pm − p0i − p0l ). (5.44)
5.4. YUKAWA INTERACTION 203
5.4 Yukawa interaction

As a concrete example, we take a theory with a fermion field and scalar field, which interact via the
Standard Model Yukawa interaction [4].
h
Lint = − √ φΨΨ
2
h i
† †
= −yφ (ψR ) ψL + (ψL ) ψR . (5.45)
Let the quantum of the field φ be denoted by H, since the particle is a Higgs. The quanta of
the fermionic field f will be called electrons. The mass of H is M , and the mass of the electron
by m. Suppose M > 2m, so that kinematically it is possible to have the H particle decay into an
electron-positron pair. The process is denoted by
H(k) → e− (p) + e+ (p0 ) , (5.46)
where k, p, p0 are the 4–momenta of the particles.

For the interaction Hamiltonian, since it does not depend explicitly in time, we have
Lint = −Hint . (5.47)
h
Hint = √ : (ψR )† ψL φ : (5.48)
2
where the required ordered product will be explained in next section. The term linear in the
interaction Hamiltonian in the S–matrix. It is
Z
h
(1)
S = − i√ d4 x : (ψR )† ψL φ : . (5.49)
2
Z
h
S (1)
= −i √ d4 x : [(ψR )†+ + (ψR )†− ](ψL + + ψL − )(φ+ + φ− ) : . (5.50)
2
where + or − denote the annihilation or creation operators respectively
φ+ |nφ i ∝|n − 1φ i hnφ |φ+ ∝hn + 1φ | (5.51)
and
φ− |nφ i ∝|n + 1φ i hnφ |φ− ∝hn − 1φ | (5.52)
with similar expressions for fermion or vector fields.

Expanding the ordered product in the interaction Hamiltonian-densyty in eq. (5.48) in terms of
the + and − components of the several scalar and fermions fields, we have

Hint =h : (ψR )†+ + (ψR )†− ψL + + ψL − (φ+ + φ− ) :
=h : (ψR )†+ ψL + φ+ + (ψR )†+ ψL + φ− + (ψR )†+ ψL − φ+ + (ψR )†+ ψL − φ− + (ψR )†− ψL + φ+
+ (ψR )†− ψL + φ− + (ψR )†− ψL − φ+ + (ψR )†− ψL − φ− : . (5.53)
The ordered product refers to the different from zero terms when the interaction Hamiltonian density
is operated between the initial and final states, which for the decay processes under study correspond
to
|ii =|0(ψR )† , 0ψL , 1φ i and, hf | =h1(ψR )† , 1ψL , 0φ | (5.54)
Evaluating the expectation values for each term in eq. (5.53) for these initial and final states, we
have
h1(ψR )† , 1ψL , 0φ |(ψR )†+ ψL + φ+ |0(ψR )† , 0ψL , 1φ i ∝ h2(ψR )† , 2ψL , 0φ |0(ψR )† , 0ψL , 0φ i = 0
h1(ψR )† , 1ψL , 0φ |(ψR )†+ ψL + φ− |0(ψR )† , 0ψL , 1φ i ∝ h2(ψR )† , 2ψL , 0φ |0(ψR )† , 0ψL , 2φ i = 0
h1(ψR )† , 1ψL , 0φ |(ψR )†+ ψL − φ+ |0(ψR )† , 0ψL , 1φ i ∝ h2(ψR )† , 0ψL , 0φ |0(ψR )† , 0ψL , 0φ i = 0
h1(ψR )† , 1ψL , 0φ |(ψR )†+ ψL − φ− |0(ψR )† , 0ψL , 1φ i ∝ h2(ψR )† , 0ψL , 0φ |0(ψR )† , 0ψL , 2φ i = 0
h1(ψR )† , 1ψL , 0φ |(ψR )†− ψL + φ+ |0(ψR )† , 0ψL , 1φ i ∝ h0(ψR )† , 2ψL , 0φ |0(ψR )† , 0ψL , 0φ i = 0
h1(ψR )† , 1ψL , 0φ |(ψR )†− ψL + φ− |0(ψR )† , 0ψL , 1φ i ∝ h0(ψR )† , 2ψL , 0φ |0(ψR )† , 0ψL , 2φ i = 0
h1(ψR )† , 1ψL , 0φ |(ψR )†− ψL − φ+ |0(ψR )† , 0ψL , 1φ i ∝ h0(ψR )† , 0ψL , 0φ |0(ψR )† , 0ψL , 0φ i =
6 0
h1(ψR )† , 1ψL , 0φ |(ψR )†− ψL − φ− |0(ψR )† , 0ψL , 1φ i ∝ h0(ψR )† , 0ψL , 0φ |0(ψR )† , 0ψL , 2φ i = 0
In this way, we filter
h1(ψR )† , 1ψL , 0φ |(ψR )†− ψL − φ+ |0(ψR )† , 0ψL , 1φ i ∝ h0(ψR )† , 0ψL , 0φ |0(ψR )† , 0ψL , 0φ i = 1 . (5.55)
which is consistent with the definition
: (ψR )† ψL φ :=(ψR )†− ψL − φ+ . (5.56)

Figure 5.1: Feynman diagrams for B → e+ e−
In general, we have the following result: The ordered product of a set of operators is such that all
the creation operators are on the left and destruction operators are in the right while keeping the
original order consistent with the initial and final states.
h1, 1, · · · , 0, 0| : φ1 φ2 · · · φn−1 φn : |0, 0, · · · , 1, 1i ∝ h1, 1, · · · , 0, 0|a†1 a†2 · · · an−1 an |0, 0, · · · , 1, 1i (5.57)
The only term that contributes to the matrix element of these process is therefore
Z
h
−i √ d4 x(ψR )†− ψL − φ+ . (5.58)
2
In the language of second quantization it is said that at the interaction point in Fig. 5.1, the
scalar is destroyed by φ+ and one electron (ψL − ) and a positron ((ψR )†− ) are created.
Let us define the one–particle states as in eq. (2.141)
r
1 †
|H(p)i ≡ a |0i (5.59)
V p
For the two Weyl spinors forming a Dirac spinor we have the solutions in eq. (2.217), from here we
can define the one–particle states were defined as in eq. (5.103)
r r
1 † 1 †α̇
|eL (p, s)i ≡ xα (s, p) as (p)|0i, |eR (p, s)i ≡ y (s, p) a†s (p)|0i
V V
r r
† 1 † † † 1 α
| (eL ) (p, s)i ≡ yα̇ (s, p) bs (p)|0i, | (eR ) (p, s)i ≡ x (s, p) b†s (p)|0i . (5.60)
V V
Using the commutation relations, our states are then normalized as
(2π)3 3
hH(p)|H(p0 )i = δ (p − p0 )
V
0 0 (2π)3
heL,R (p, s)|eL,R (p , s )i = δss0 δ 3 (p − p0 )
V
3
(2π)
h(eL,R )† (p, s)| (eL,R )† (p0 , s0 )i = δss0 δ 3 (p − p0 ) (5.61)
V
As established in Sec. 2.2, it is convenient to work in the discrete limit where (2.110)
V
δ 3 (0) = . (5.62)
(2π)3
Now we can write down the action of various field operators on different one particles states. Using
the Fourier decomposition of the scalar field in eq. (2.147), and taking into account that ap |0i = 0,
we have
Z
1
φ+ (x)|H(k)i = d3 p p ap e−ip·x |H(k)i
b
(2π)3 2ωp
Z
1 1 †
= d3 p 3
p ap e−ip·x √ b
b ak |0i
(2π) 2ωp V
Z
1
= d3 p p e−ip·x [b a†k ]|0i .
ap , b (5.63)
(2π)3 2ωp V
By using the commutation relations in eq. (2.119) we have
Z
δ (3) (p − k) −ip·x
φ+ (x)|H(k)i = d3 p p e |0i (5.64)
2ωp V
1
φ+ (x)|H(k)i = √ e−ik·x |0i (5.65)
2ωk V
Similarly, we have initial one-particles states on left and initial one-particles states on right
1 1
φ+ (x)|H(k)i = √ e−ik·x |0i, hH(k)|φ− (x) =h0| √ eik·x
2ωk V 2ωk V
1 † 1
ξ+ (x)|eL (p, s)i = p x(s, p)e−ip·x |0i, heL (p, s)|ξ− (x) =h0| p x† (s, p)eip·x
2Ep V 2Ep V
1 1
η+ (x)| (eR )† (p, s)i = p x(s, p)e−ip·x |0i, h(eR )† (p, s)|η− †
(x) =h0| p x† (s, p)eip·x
2Ep V 2Ep V
1 1
†
ξ+ (x)| (eL )† (p, s)i = p y † (s, p)e−ip·x |0i, h(eL )† (p, s)|ξ− (x) =h0| p y(s, p)eip·x
2Ep V 2Ep V
† 1 1
η+ (x)|eR (p, s)i = p y † (s, p)e−ip·x |0i, heR (p, s)|η− (x) =h0| p y(s, p)eip·x . (5.66)
2Ep V 2Ep0 V
where ωk and Ep represent the energies of the scalar and the electron for the 3-momenta in the
subscripts.
In the lowest order the only term which contributes to the matrix element is the term shown in
Eq. (5.58). The matrix element at first order in eq. (5.31), between the initial and the final state is
then
Z hD E
(1) h X 4 † 0 † †
Sf i = −i √ d x 0, eL (p), (eR ) (p ) ξ− (x)η− (x)φ+ (x) H(k), 0, 0 +
2 s,s0
D i

0, (eL )† (p), eR (p0 ) ξ− (x)η− (x)φ+ (x) |H(k), 0, 0i (5.67)
Using Eqs. (5.66) we obtain

(1) h Xh † α̇
i
Sf i =(−i) √ xα̇ (s, p)x† (s0 , p0 ) + y α (s, p)yα (s0 , p0 )
2 s,s0
Z
0 1 1 1
× d4 x ei(p+p −k)·x √ p p . (5.68)
2ωk V 2Ep V 2Ep0 V
Since
Z
0
d4 x ei(p+p −k)·x = (2π)4 δ 4 (k − p − p0 ) , (5.69)
we obtain
" #
(1) 1 1 1
Sf i = √ p p (2π)4 δ 4 (k − p − p0 )
2ωk V 2Ep V 2Ep0 V
h Xh † α̇
i
× (−i) √ xα̇ (s, p)x† (s0 , p0 ) + y α (s, p)yα (s0 , p0 ) . (5.70)
2 s,s0
Comparing with Eq. (4.28) we have therefore that the relativistic amplitude is
h Xh † α̇
i
iMf i = −i √ xα̇ (s, p)x† (s0 , p0 ) + y β (s, p)yβ (s0 , p0 ) . (5.71)
2 s,s0
Check the specific calculation in ??

The equation for a two body decays, assuming that the amplitude (5.71) does not depends in the
final momentum, is given in eq. (4.148) is
dΓ 1
= |Mf i |2 λ1/2 (M 2 , m22 , m21 ) (5.72)
dΩ 64π 2 M 3
where, from (4.145)
λ(a, b, c) = [a − (b − c)]2 − 4ac . (5.73)
For our specific problem, the final state masses of electron and positron is the same, however, we
will obtain the results for the most general case of different masses. In this way, let m (m0 = m) the
mass of eL (eR ).
The missing part is to calculate the square of the amplitude
|M|2 =M† M =
h2 X h † α̇
i† h α̇
i
= xα̇ (s, p)x† (s0 , p0 ) + y β (s, p)yβ (s0 , p0 ) x†α̇ (s, p)x† (s0 , p0 ) + y β (s, p)yβ (s0 , p0 )
2 s,s0
h2 X h α 0 0 ih α̇
i
= x (s , p )xα (s, p) + yβ̇† (s0 , p0 )y †β̇ (s, p) x†α̇ (s, p)x† (s0 , p0 ) + y β (s, p)yβ (s0 , p0 ) ,
2 s,s0
(5.74)
which includes (using anticommuting spinors)
" #
X α̇
X X α̇
α 0
x (s , p 0
)xα (s, p)x†α̇ (s, p)x† (s0 , p0 ) = α 0
x (s , p ) 0
xα (s, p)x†α̇ (s, p) x† (s0 , p0 )
s,s0 s0 s
X α̇
= xα (s0 , p0 ) (p · σαα̇ ) x† (s0 , p0 )
s0
X α̇
= (p · σαα̇ ) x† (s0 , p0 )xα (s0 , p0 )
s0
0

= (p · σαα̇ ) p · σ α̇α
= Tr [(p · σ) (p0 · σ)]
= pµ p0ν Tr [σ µ σ ν ]
= 2g µν pµ p0ν
= 2p · p0 . (5.75)
X X
yβ̇† (s0 , p0 )y †β̇ (s, p)y β (s, p)yβ (s0 , p0 ) = p · σ β̇β yβ̇† (s0 , p0 )yβ (s0 , p0 )
s,s0 s,s0
X
= p · σ β̇β yβ (s0 , p0 )yβ̇† (s0 , p0 )
s,s0

= p · σ β̇β p0 · σβ β̇
=2p · p0 . (5.76)
" #
X X X
xα (s0 , p0 )xα (s, p)y β (s, p)yβ (s0 , p0 ) = xα (s0 , p0 ) xα (s, p)y β (s, p) yβ (s0 , p0 )
s,s0 s0 s
X
= mδα β
yβ (s0 , p0 )xα (s0 , p0 )
s0

= mδα (−m0 δβ α )
β
= Tr [−mm0 11]
=mm0 Tr [1]
= − 2mm0 . (5.77)
X α̇

yβ̇† (s0 , p0 )y †β̇ (s, p)x†α̇ (s, p)x† (s0 , p0 ) 0 α̇
β̇
= mδ α̇ −m δβ̇
s,s0
= − 2mm0 . (5.78)
Therefore
h2
|M|2 = 4 (p · p0 − mm0 ) (5.79)
2
From the conservation of the momentum-energy in the center of mass frame
k µ = (M, 0)
pµ = (E, p)
µ
p0 = (E 0 , p0 ) , (5.80)
as assummed in (5.72), we have

µ
k µ = pµ + p 0 , , (5.81)
which implies for each component that
M =E + E 0
|p| = |p0 |. (5.82)
From the first equation

2
M 2 = E 2 + E 0 + 2EE 0 . (5.83)
Therefore
0M 2 − E 2 − E 02
EE = (5.84)
2
p · p0 − mm0 = EE 0 − p · p0 − mm0
= EE 0 + p2 − mm0
M 2 − E 2 − E 02
= + p2 − mm0
2
1 2 2

= M − m2 − p2 − m0 − p2 + p2 − mm0
2
1 2 2

= M − m2 − m0 − 2mm0
2
1 2
= M − (m − m0 )2
2
Therefore, the scattering amplitude is
X
|M|2 = h2 M 2 − (m + m0 )2 (5.85)
s1 ,s2
Since this does not depends explicitly on the final state tri-momentum, is was consistent to already
make the tri-momentum integrals. Moreower, Replacing back in eq. (5.72)
dΓ h2 2
= 2 3
λ1/2 (M 2 , m0 , m2 ) M 2 − (m + m0 )2 (5.86)
dΩ 64π M
R
After the integration dΩCM = 4π 3 we have
h2 1/2 2 02 2
2 0 2

Γ= λ (M , m , m ) M − (m + m ) (5.87)
16πM 3
For m = m0 = mf , and using eq. (5.73)
h 2 i1/2
λ1/2 (M 2 , m2f , m2f ) =M 2 − m2f + m2f − 4M 2 m2f
1/2
= M 4 − 4M 2 m2f
1/2
4
m2f
= M 1−4 2
M
1/2
2
m2f
=M 1 − 4 2 , (5.88)
M
and therefore
1/2
h2 2
m2f 2 2

Γ(H → f f ) = M 1 − 4 M − 4m f
16πM 3 M2
1/2
h2 4
m2f m2f
= M 1−4 2 1−4 2
16πM 3 M M
3/2
h2 4m2f
= M 1− (5.89)
16π M2
In the case of the standard model Higgs with mass MH = M decaying to fermion pair, and using
√ √ −1/2
mf = hv/ 2, v = 2GF
3
R 2π Rπ
0
dφ 0
sin θdθ = 4π
3/2
h2 v 2 4m2f
Γ(H → f f ) = MH 1 − 2
16v 2 π MH
3/2
h2 v 2 1 4m2f
= MH 1 − 2
2 8v 2 π MH
√ 3/2
2 2GF 4m2f
=mf MH 1 − 2
8π MH
3/2
MH m2f GF m2f
= √ 1−4 2 , (5.90)
4π 2 MH
In the limit mf MH this expression reduces to
MH m2f GF
Γ(H → f f ) = √ . (5.91)
4π 2
For the tau4 , taking into account that 1.166 × 10−5 GeV−2 , MH = 125 GeV, and mτ = 1.777 GeV,
~
Γ(H → τ + τ − ) = 0.26 MeV → τ = = 2.53 × 10−21 seconds . (5.92)
Γ
5.5 Wick Theorem

5.5.1 Basics concepts
Return back to the interpretation of interaction in terms of one exchanged particle, as illustrated
again in Fig. 5.2
From [4]. The normal ordering procedure involved putting all the annihilation operators to
the right of all creation operators so that it annihilates the vacuum. But the time ordering raises
complications because in it all operators at earlier times must be further to the right. So creation
operators at later times would be to the right of annihilation operators at later times, contrary to what
we need for normal ordering. The advantage of normal ordered products is that their expectation
values vanish in the vacuum.
4
to the top quarks is kinematically forbidden, to the bottom quark a further factor Nc = 3 must be considered
5.5. WICK THEOREM 213
t (a) (b) t (c)
x2 x1
x1 x2
t(x1)<t(x2) t(x2)<t(x1)
x x
Figure 5.2: Electromagnetic repulsion. The diagrams (a) and (b) are summarized in the diagram (c)
If Hint contains an even number of fermion factors, we can use the time–ordered product T{. . .}
of n factors to write this expression in the equivalent form. For S (2) we have for example
Z ∞ Z ∞ Z ∞ Z ∞
d t1 d t2 T{Hint (t1 )Hint (t2 )} = d t1 d t2 θ(t2 − t1 )Hint (t2 )Hint (t1 )
−∞ −∞ −∞ −∞
Z ∞ Z ∞
+ d t1 d t2 θ(t1 − t2 )Hint (t1 )Hint (t2 ) (5.93)
−∞ −∞
where, for t2 and t1 fixed respectively
( (
0 t1 > t2 0 t2 > t1
θ(t2 − t1 ) = , θ(t1 − t2 ) = . (5.94)
1 t1 < t2 1 t2 < t1
In this way.
Z ∞ Z ∞ Z ∞ Z t2
d t1 d t2 T{Hint (t1 )Hint (t2 )} = d t2 d t1 θ(t2 − t1 )Hint (t2 )Hint (t1 )
−∞ −∞ −∞ −∞
Z ∞ Z t1 ((((
(((
+ (2(− t1 )Hint (t2 )Hint (t1 )
d t2 (d(t1(θ(t ( ( (
−∞ ( ( (
t2
Z( Z
( (
∞ t1
+ d t1 d t2 θ(t1 − t2 )Hint (t1 )Hint (t2 )
−∞ −∞
Z ∞ Z t2 ((((
((((
+ d t1 d t2(θ(t1 − t2 )Hint (t1 )Hint (t2 ) .
(((( (5.95)
((((
(−∞
((( t1
The integral in d t1 is for t1 > t2 , with t2 fixed, so that θ(t2 − t1 ) = 0 for t1 > t2 . Similarly, The
integral in d t2 is for t2 > t1 , with t1 fixed, so that θ(t1 − t2 ) = 0 for t2 > t1 . Therefore,
Z ∞ Z ∞ Z ∞ Z t2 Z ∞ Z t1
d t1 d t2 T{Hint (t1 )Hint (t2 )} = d t2 d t1 Hint (t2 )Hint (t1 ) + d t1 d t2 Hint (t1 )Hint (t2 )
−∞ −∞ −∞ −∞ −∞ −∞
Z ∞ Z t1 Z ∞ Z t1
= d t1 d t2 Hint (t1 )Hint (t2 ) + d t1 d t2 Hint (t1 )Hint (t2 )
−∞ −∞ −∞ −∞
Z ∞ Z t1
=2 d t1 d t2 Hint (t1 )Hint (t2 ) . (5.96)
−∞ −∞
In terms of the ordering operator all the integrals are between −∞ to ∞:

X∞ Z Z ∞ Z ∞
(−i)n ∞
S =1 + d t1 d t2 . . . d tn T{Hint (t1 )Hint (t2 ) . . . Hint (tn )} , (5.97)
n=1
n! −∞ −∞ −∞
In terms of the Hamiltonian density, we have

X∞ Z Z
(−i)n
S =1+ · · · d4 x1 d4 x2 . . . d4 xn T{Hint (x1 )Hint (x2 ) . . . Hint (xn )} , (5.98)
n=1
n!
In the above perturbation formalism the states |ii and |f i are, as usual, eigenstates of the unperturbed
free-field Hamiltonian H0 . As such can be introduced inside the integrals
Sf i =hf |S|ii
X∞ Z Z
(−i)n
=1 + · · · d4 x1 d4 x2 . . . d4 xn hf | T{Hint (x1 )Hint (x2 ) . . . Hint (xn )}|ii . (5.99)
n=1
n!
5.5. WICK THEOREM 215
For example, at first order

(1)
Sf i =hf |S (1) |ii
Z
=hf | − i d4 x1 T{Hint (x1 )}|ii
Z
= − i d4 x1 hf | : Hint (x1 ) : |ii . (5.100)
In order to evaluate this integrals we need to write the time ordered product in terms of the fields.
This can done by induction. We start by considering the simple no trivial case with two scalar fields
5.5.2 Wick theorem

To evaluate the time-ordered product we need first to show that
T{φ(x1 )φ(x2 )} =: φ(x1 )φ(x2 ) : +h0| T{φ(x1 )φ(x2 )}|0i , (5.101)
where the last term
i∆F (x1 − x2 ) =h0| T{φ(x1 )φ(x2 )}|0i (5.102)
corresponds to the scalar propagator. The demostration is displayed below.
We can rewrite the scalar propagator in terms of the Wick contraction, which is defined just as:
φ(x1 )φ(x2 ) =h0| T{φ(x1 )φ(x2 )}|0i . (5.103)
We can then stablish the Wick theorem as

T{φ(x1 )φ(x2 )} =: φ(x1 )φ(x2 ) : + φ(x1 )φ(x2 ) . (5.104)
For the demostration

since
φ(x) = φ+ (x) + φ− (x) , (5.105)
T [φ(x1 ), φ(x2 )] = θ(t1 − t2 )φ(x1 )φ(x2 ) + θ(t2 − t1 )φ(x2 )φ(x1 ) (5.106)
h0| T [φ(x1 ), φ(x2 )] |0i =θ(t1 − t2 )h0|φ(x1 )φ(x2 )|0i + θ(t2 − t1 )h0|φ(x2 )φ(x1 )|0i
=θ(t1 − t2 )h0|φ+ (x1 )φ− (x2 )|0i + θ(t2 − t1 )h0|φ+ (x2 )φ− (x1 )|0i
=h0|θ(t1 − t2 )φ(x1 )+ φ− (x2 )|0i + h0|θ(t2 − t1 )φ+ (x2 )φ− (x1 )|0i (5.107)
with
Z Z
1 1
φ+ (x) = 3
d p 3
p ap e−ip·x
b φ− (x) = d3 p p a†p eip·x ,
b (5.108)
(2π) 2ωp (2π)3 2ωp
we have
Z Z
d3 p1 −ip1 ·x1 d3 p2
h0|T (φ(x1 )φ(x2 ))|0i =h0| 3
p θ(t 1 − t )â
2 p1 e 3
p â†p2 e−ip2 ·x2 |0i
(2π) 2Ep1 (2π) 2Ep2
Z 3 Z
d p2 −ip2 ·x2 d3 p1
+ h0| 3
p θ(t 2 − t )â
1 p2 e 3
p â†p1 e−ip1 ·x1 |0i
(2π) 2Ep2 (2π) 2Ep1
Z Z 3 3
d p1 d p2
= p p θ(t1 − t2 )e−ip1 ·x1 e−ip2 ·x2 h0|âp1 â†p2 |0i
(2π)6 2Ep1 2Ep2
Z Z
d3 p2 d3 p1
+ p p θ(t2 − t1 )e−ip2 ·x2 e−ip1 ·x1 h0|âp2 â†p1 |0i , (5.109)
(2π)6 2Ep2 2Ep1
In this way
Z Z
d 3 p 1 d 3 p2
h0|T (φ(x1 )φ(x2 ))|0i = 6
p θ(t1 − t2 )e−ip1 ·x1 e−ip2 ·x2 h0|âp1 â†p2 |0i
(2π) 2Ep1 2Ep2

+θ(t2 − t1 )e−ip2 ·x2 e−ip1 ·x1 h0|âp2 â†p1 |0i (5.110)
and the final result is that

Z
d4 p e−ip·(x1 −x2 )
h0|T (φ(x1 )φ(x2 ))|0i = (5.111)
(2π)4 p2 − m2 − iε
Generalizing the results for n scalar or fermion fields, but with an even number of fermions fields,
we have the Wick theorem
T{Φ(x1 )Φ(x2 )Φ(x3 ) · · · Φ(xn )} = : Φ(x1 )Φ(x2 )Φ(x3 ) · · · Φ(xn ) : +

+ Φ(x1 )Φ(x2 ) : Φ(x3 ) · · · Φ(xn ) :
+ : Φ(x1 ) Φ(x2 )Φ(x3 ) · · · Φ(xn ) : + · · ·
= : Φ(x1 )Φ(x2 )Φ(x3 ) · · · Φ(xn ) : +
h i
+ : Φ(x1 )Φ(x2 )Φ(x3 ) · · · Φ(xn ) : +perm
h i
+ : Φ(x1 )Φ(x2 ) Φ(x3 )Φ(x4 )Φ(x5 ) · · · Φ(xn ) : +perm
+ ···
(5.112)
For details of the full result see for example [4].
5.6 Mandelstam variables

We can generalize the definition of s in eq.(4.78) to the full Mandelstam variables http://bolvan.
ph.utexas.edu/~vadim/Classes/2011f/STU.pdf, (See also [12])
2
s = (p1 + p2 )2 = (p01 + p02 )
2 2
t = (p1 − p01 ) = (p2 − p02 )
2 2
u = (p1 − p02 ) = (p01 − p2 ) . (5.113)
The explicit form for s is
s = (p1 + p2 )2
=p21 + p22 + 2p1 · p2
=m21 + m22 + 2p1 · p2 , (5.114)
5.7. SCATTERING 225
from which the product p1 · p2 can be obtained. By following similar steps for t and u, we have
2p1 · p2 =s − m21 − m22
2 2
2p01 · p02 =s − m01 − m02
2
2p1 · p01 =m21 + m01 − t
2
2p2 · p02 =m22 + m02 − t
2
2p1 · p02 =m21 + m02 − u
2
2p2 · p01 =m22 + m01 − u . (5.115)
Moreover,
2 2
s + t + u = m21 + m22 + m01 + m02 . (5.116)
In the non-relativistic limit
s = m21 + m22 + 2p1 · p2 ≈ (m1 + m2 )2 + m1 m2 (v1 − v2 ) > 0 ,
2
t = m21 + m01 − 2p1 · p01 ≈ (m1 − m01 )2 − m1 m01 (v1 − v10 ) > 0 ,
2
u = m21 + m02 − 2p1 · p02 ≈ (m1 − m02 )2 − m1 m02 (v1 − v20 ) > 0 . (5.117)
In the center of mass frame (p1 + p1 = 0)
s = (E1 + p1 + E2 + p2 )2 = (E1 + p1 + E2 − p1 )2 = (E1 + E2 )2 . (5.118)
5.7 Scattering
From the previous calculation we have
Z Z
(n) (−i)n
S = · · · d4 x1 d4 x2 . . . d4 xn T{Hint (x1 )Hint (x2 ) . . . Hint (xn )} . (5.119)
n!
5.7.1 Higgs mediated scattering

The relevant term for the scattering that we consider now is the t-channel
e− − − 0 − 0
L (p1 ) + eL (p2 ) → eR (p1 ) + eR (p2 ) (5.120)
As always, we use the left-handed Weyl spinors
eL →ξα (eR )† →η α , (5.121)
with Yukawa Lagrangian
Ly =y (eR )† eL h0 + y (eL )† eR h0
=yη α ξα h0 + yη †α̇ ξα̇† h0 . (5.122)
We will make the calculation for any Standard Model fermion, but keep the notation just for the
electron with mass m1 . To keep the calculation in general, we left open the possibility that the
second vertex involves a different fermion of mass m2 . The Feynman diagrams for the electronic case
is displayed in figure 5.3. Therefore
Z Z
(2) (−i)2
S = d4 x1 d4 x2 T{Hint (x1 )Hint (x2 )}
2!
Z Z
(−iy)2
= d4 x1 d4 x2 T{ η α ξα h0 x1 η α ξα h0 x2 }
2!
Z Z Z Z
(−iy)2 4 4 0
0
(−iy)2
= α α
d x1 d x2 : η ξα h x1 η ξα h x2 : + d4 x1 d4 x2 : (η α ξα h0 )x1 (η α ξα h0 )x2 : + · · ·
2! 2!
(5.123)
The first term corresponds to two disconnected Feynman diagrams that does not contribute to the S–
matrix. For the process at hand, we want terms where four fermionic operators are not contracted,
corresponding to the particles in the initial and final states. The second term in the previous
expansion of the Wick theorem is the only satisfying this requirement. In this way
Z Z
(2) + − + − (−iy)2
S (e e → e e ) = d4 x1 d4 x2 h(x1 )h(x2 ) : (η α ξα )x1 (η β ξβ )x2 : (5.124)
2!
On the other hand
: (η α ξα )x1 (η β ξβ )x2 := : η α (x1 )ξα (x1 )η β (x2 )ξβ (x2 ) : . (5.125)

5.7. SCATTERING 227
The specific calculation by the scattering mediated by the Higgs is

Z Z
(−iy)2 X
Sf i = d4 x1 d4 x2 h(x1 )h(x2 )
2! spins
heR (p01 ) , eR (p02 )| : η α (x1 )η β (x2 )ξβ (x2 )ξα (x1 ) : |eL (p1 ) , eL (p2 )i .
Z Z
(−iy)2 X
= d4 x1 d4 x2 i∆F (x1 − x2 )
2! spins
β
heR (p01 ) , eR (p02 )| η−
α
(x1 )η− (x2 )ξβ+ (x2 )ξα+ (x1 ) |eL (p1 ) , eL (p2 )i .
Z Z Z
(−iy)2 X d4 q
= 4
d x1 d x24
i∆F (q) eiq·(x1 −x2 )
2! spins (2π)4
β
heR (p01 ) , eR (p02 )| η−
α
(x1 )η− (x2 )ξβ+ (x2 )ξα+ (x1 ) |eL (p1 ) , eL (p2 )i . (5.126)
In the following, to avoid clutter in the expressions we ignore the spin part during the intemediate
calculations. The two particle Fock state is, after proper normalization
1
|eL (p2 )eL (p1 )i = √ a†s1 (p1 )a†s2 (p2 )|0i (5.127)
V2
Therefore
Z Z
β d3 k d3 k 0 0
α
ξ+ (x1 )ξ+ (x2 )|e− −
L (p2 )eL (p1 )i = 3
√ 3
√ xα (s, k)xβ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V
0 †
× as (k)as0 (k )as1 (p1 )a†s2 (p2 )|0i
Z Z
d3 k 0 d3 k 0
= 3
√ 3
√ xα (s, k)xβ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V

× as (k)as0 (k )as1 (p1 )a†s2 (p2 ) − a†s1 (p1 )a†s2 (p2 )as (k)as0 (k0 ) |0i
0 †
Z Z
d3 k d3 k 0 α β 0 0 −ik·x1 −ik0 ·x2
= √ √ x (s, k)x (s , k ) e e
(2π)3 2Ek V (2π)3 2Ek0 V

× as (k)as0 (k0 ), a†s1 (p1 )a†s2 (p2 ) |0i. (5.128)
[AB, CD] = A[B, C]D − [A, C]BD + CA[B, D] − C[A, D]B , (5.129)

as (k)as0 (k0 ), a†s1 (p1 )a†s2 (p2 ) =as (k) as0 (k0 ), a†s1 (p1 ) a†s2 (p2 ) − as (k), a†s1 (p1 ) as0 (k0 )a†s2 (p2 )

+ a†s1 (p1 )as (k) as0 (k0 ), a†s2 (p2 ) − a†s1 (p1 ) as (k), a†s2 (p2 ) as0 (k0 )

= as0 (k0 ), a†s1 (p1 ) as (k)a†s2 (p2 ) − as (k), a†s1 (p1 ) as0 (k0 )a†s2 (p2 )

+ as0 (k0 ), a†s2 (p2 ) a†s1 (p1 )as (k) − as (k), a†s2 (p2 ) a†s1 (p1 )as0 (k0 )
(5.130)
Therefore,

as (k)as0 (k0 ), a†s1 (p1 )a†s2 (p2 ) |0i

= as0 (k0 ), a†s1 (p1 ) as (k)a†s2 (p2 )|0i − as (k), a†s1 (p1 ) as0 (k0 )a†s2 (p2 )|0i

= as0 (k0 ), a†s1 (p1 ) as (k), a†s2 (p2 ) |0i − as (k), a†s1 (p1 ) as0 (k0 ), a†s2 (p2 ) |0i

= (2π)6 δ (3) (k0 − p1 )δ (3) (k − p2 ) − δ (3) (k − p1 )δ (3) (k0 − p2 ) |0i (5.131)
β
α
ξ+ (x1 )ξ+ (x2 )|e− −
L (p2 )eL (p1 )i
1 α
=√ √ x (p2 )xβ (p1 ) e−ip2 ·x1 e−ip1 ·x2 −xα (p1 )xβ (p2 ) e−ip1 ·x1 e−ip2 ·x2 |0i . (5.132)
2E1 V 2E2 V
This implies also that
†α̇ †β̇
he− −
L (p2 )eL (p1 )|ξ− (x1 )ξ− (x2 )
1 h i
= h0| √ √ x†α̇ (p2 )x†β̇ (p1 ) eip2 ·x1 eip1 ·x2 −x†α̇ (p1 )x†β̇ (p2 ) eip1 ·x1 eip2 ·x2 . (5.133)
2E1 V 2E2 V
Since
1
|eR (p2 )eR (p1 )i = √ a†s1 (p1 )a†s2 (p2 )|0i (5.134)
V2
we have
Z Z
†α̇ †β̇ d3 k d3 k 0 −ik·x1 −ik0 ·x2
η+ (x1 )η+ (x2 )|e− −
R (p2 )eR (p1 )i = √ √ y †α̇
(s, k)y †β̇ 0
(s , k 0
) e e
(2π)3 2Ek V (2π)3 2Ek0 V
× as (k)as0 (k0 )a†s1 (p1 )a†s2 (p2 )|0i . (5.135)
5.7. SCATTERING 229
Following similar steps, we find
†α̇ †β̇
η+ (x1 )η+ (x2 )|e− −
R (p2 )eR (p1 )i
1 h i
=√ √ y †α̇ (p2 )y †β̇ (p1 ) e−ip2 ·x1 e−ip1 ·x2 −y †α̇ (p1 )y †β̇ (p2 ) e−ip1 ·x1 e−ip2 ·x2 |0i
2E1 V 2E2 V
(5.136)
β
he− − α
R (p2 )eR (p1 )|η− (x1 )η− (x2 )
1 α
= h0| √ √ y (p2 )y β (p1 ) eip2 ·x1 eip1 ·x2 −y α (p1 )y β (p2 ) eip1 ·x1 eip2 ·x2 . (5.137)
2E1 V 2E2 V
Replacing back we have
Z Z Z
(−iy)2 d4 q
Sf i = 4
d x1 d x24
i∆F (q) eiq·(x1 −x2 )
2! (2π)4
β
× heR (p01 ) , eR (p02 )| η− α
(x1 )η− (x2 )ξβ+ (x2 )ξα+ (x1 ) |eL (p1 ) , eL (p2 )i
2 Z Z Z
(iy) 1 d4 q
= p 0 p 0 √ √ 4
d x1 d x2 4
4
i∆F (q) eiq·(x1 −x2 )
2 2E1 V 2E2 V 2E1 V 2E2 V (2π)
h 0 0 0 0
i
× y α (p01 )y β (p02 ) eip1 ·x1 eip2 ·x2 −y α (p02 )y β (p01 ) eip1 ·x2 eip2 ·x1

× xα (p1 )xβ (p2 ) e−ip1 ·x1 e−ip2 ·x2 −xα (p2 )xβ (p1 ) e−ip2 ·x1 e−ip1 ·x2 . (5.138)
Expanding all the terms,
Z Z Z
(iy)2 1 d4 q
Sf i = p 0 p 0 √ √ 4
d x1 d x24
4
i∆F (q) eiq·(x1 −x2 )
2 2E1 V 2E2 V 2E1 V 2E2 V (2π)
h 0 0
× y α (p01 )y β (p02 )xα (p1 )xβ (p2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2
0 0
− y α (p02 )y β (p01 )xα (p1 )xβ (p2 ) eip1 ·x2 eip2 ·x1 e−ip1 ·x1 e−ip2 ·x2
0 0
0 0
i
+y α (p02 )y β (p01 )xα (p2 )xβ (p1 ) eip1 ·x2 eip2 ·x1 e−ip2 ·x1 e−ip1 ·x2 . (5.139)
Interchanging α ↔ β and moving terms around

Z Z Z
(iy)2 1 d4 q
Sf i = p 0 p 0 √ √ 4
d x1 d x24
i∆F (q) eiq·(x1 −x2 )
2 2E1 V 2E2 V 2E1 V 2E2 V (2π)4
h 0 0
× y α (p01 )y β (p02 )xα (p1 )xβ (p2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2
0 0
+ y β (p02 )y α (p01 )xβ (p2 )xα (p1 ) eip1 ·x2 eip2 ·x1 e−ip2 ·x1 e−ip1 ·x2
0 0
0 0
i
−y β (p01 )y α (p02 )xβ (p2 )xα (p1 ) eip1 ·x1 eip2 ·x2 e−ip2 ·x1 e−ip1 ·x2 . (5.140)
We can now get the factor xα (p1 )xβ (p2 )

Z Z Z
(iy)2 1 1 d4 q
Sf i = p 0 p 0 √ √ 4
d x1 d x24
4
i∆F (q) eiq·(x1 −x2 )
2 2E1 V 2E2 V 2E 1 V 2E 2 V (2π)
n h 0 0 0 0
i
× y α (p01 )y β (p02 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2 + eip1 ·x2 eip2 ·x1 e−ip2 ·x1 e−ip1 ·x2
h 0 0 0 0
io
−y α (p02 )y β (p01 ) eip1 ·x2 eip2 ·x1 e−ip1 ·x1 e−ip2 ·x2 + eip1 ·x1 eip2 ·x2 e−ip2 ·x1 e−ip1 ·x2 xα (p1 )xβ (p2 ).
(5.141)
For the second term we have

Z Z Z
d4 q 0 0
4 4
d x1 d x2 4
i∆F (q) eiq·(x1 −x2 ) eip1 ·x2 eip2 ·x1 e−ip2 ·x1 e−ip1 ·x2
(2π)
Z Z Z
4 4 d4 q iq·(x2 −x1 ) ip01 ·x1 ip02 ·x2 −ip2 ·x2 −ip1 ·x1
= d x2 d x1 i∆F (q) e e e e e
(2π)4
Z Z Z
d4 (−q) 0 0
= 4 4
d x1 d x2 4
i∆F (−q) eiq·(x1 −x2 ) eip1 ·x1 eip2 ·x2 e−ip2 ·x2 e−ip1 ·x1
(2π)
Z Z Z
d4 (q) 0 0
= 4 4
d x1 d x2 4
i∆F (q) eiq·(x1 −x2 ) eip1 ·x1 eip2 ·x2 e−ip2 ·x2 e−ip1 ·x1 . (5.142)
(2π)
Similarly, for the fourth term

Z Z Z
d4 q 0 0
4 4
d x1 d x2 4
i∆F (q) eiq·(x1 −x2 ) eip1 ·x1 eip2 ·x2 e−ip2 ·x1 e−ip1 ·x2
(2π)
Z Z Z
d4 q 0 0
d4 x1 d4 x2 4
i∆F (q) eiq·(x1 −x2 ) eip1 ·x2 eip2 ·x1 e−ip2 ·x2 e−ip1 ·x1 , (5.143)
(2π)
5.7. SCATTERING 231
where we have used

Z Z Z Z Z Z
4
d (−q) = d(−q0 ) d(−q1 ) d(−q2 ) d(−q3 ) = d4 q , (5.144)
and ∆F (−q) = ∆F (q).

In this way
Z Z Z
(iy)2 d4 q
Sf i = p 0 p 0 √ √ 4 4
d x1 d x2 i∆F (q) eiq·(x1 −x2 )
2E1 V 2E2 V 2E1 V 2E2 V (2π)4
h i
0 0 ip01 ·x1 ip02 ·x2 0 0 ip01 ·x2 ip02 ·x1
α β
× y (p1 )y (p2 ) e e α β
−y (p2 )y (p1 ) e e xα (p1 )xβ (p2 ) e−ip1 ·x1 e−ip2 ·x2 .
(5.145)
Performing the integrations in x1 and x2
Z
(2π)8 (iy)2 d4 q
Sf i = p 0 p 0 √ √ i∆F (q)
2E1 V 2E2 V 2E1 V 2E2 V (2π)4

× y α (p01 )y β (p02 )δ (4) (−q − p01 + p1 ) δ (4) (q − p02 + p2 )

−y α (p02 )y β (p01 )δ (4) (−q − p02 + p1 ) δ (4) (q − p01 + p2 ) xα (p1 )xβ (p2 ) (5.146)
After the final integration we have

(2π)4 (iy)2
Sf i = p 0 p 0 √ √
2E1 V 2E2 V 2E1 V 2E2 V

× y α (p01 )y β (p02 )i∆F (p1 − p01 )δ (4) (−q − p01 + p1 ) δ (4) (q − p02 + p2 )

−y α (p02 )y β (p01 )i∆F (p1 − p02 )δ (4) (−q − p02 + p1 ) δ (4) (q − p01 + p2 ) xα (p1 )xβ (p2 ) . (5.147)
We can simplify the δ-functions by eliminating q:
δ (4) (−q − p01 + p1 ) δ (4) (q − p02 + p2 ) =δ (4) (−p02 + p2 − p01 + p1 ) = δ (4) (p1 + p2 − p01 − p02 )
δ (4) (−q − p02 + p1 ) δ (4) (q − p01 + p2 ) =δ (4) (−p01 + p2 − p02 + p1 ) = δ (4) (p1 + p2 − p01 − p02 ) . (5.148)
In this way
(2π)4 (iy)2 δ (4) (p1 + p2 − p01 − p02 ) α 0 β 0

Sf i = p 0 p 0 √ √ y (p1 )y (p2 )i∆F (p1 − p01 )xα (p1 )xβ (p2 )
2E1 V 2E2 V 2E1 V 2E2 V

−y α (p02 )y β (p01 )i∆F (p1 − p02 )xα (p1 )xβ (p2 ) . (5.149)
e− (p1) x1 e− (p′2) e− (p1) x1 e− (p′1)
B(p1 − p′2 ) B(p1 − p′1 )
x2 x2
e− (p2) e− (p′1) e− (p2) e− (p′2)
Figure 5.3: fermion scattering (Not Weyl fermion notation yet)
As expected, the final result can be written in term of three different factors: the momentum
conservation, normalization, and the relativistic amplitude
!
(2)
X X Y 1 Y 1
Sf i = i(2π)4 δ 4 pi − p0f √ q Mf i (5.150)
i=1,2 f =1,2 i=1,2
2E i V f =1,2 2E 0
V
f
where

Mf i = (iy)2 y α (p01 )y β (p02 )i∆F (p1 − p01 )xα (p1 )xβ (p2 ) − y α (p02 )y β (p01 )i∆F (p1 − p02 )xα (p1 )xβ (p2 )
(5.151)
The two contributions are displayed in Fig. 5.3

To calculate the corresponding cross section, we must start by calculating

MM† =y 4 y α (p01 )y β (p02 )∆F (p1 − p01 )xα (p1 )xβ (p2 ) − y α (p02 )y β (p01 )∆F (p1 − p02 )xα (p1 )xβ (p2 )
h i
† † 0 †β̇ 0 †α̇ 0 † † 0 †β̇ 0 †α̇ 0
× xβ̇ (p2 )xα̇ (p1 )∆F (p1 − p1 )y (p2 )y (p1 ) − xβ̇ (p2 )xα̇ (p1 )∆F (p1 − p2 )y (p1 )y (p2 )
(5.152)
For the first term in the expansion and taking into account summming over spins, we have
X X
|M1 | ⊃y 4 ∆2F (p1 − p01 ) y α (p01 )y β (p02 )xα (p1 )xβ (p2 )x†β̇ (p2 )x†α̇ (p1 )y †β̇ (p02 )y †α̇ (p01 ) (5.153)
spins spins
5.7. SCATTERING 233
Using (5.42)
X X
|M1 |2 ⊃y 4 ∆2F (p1 − p01 ) p2 · σβ β̇ x†α̇ y α (p01 )y β (p02 )xα (p1 )(p1 )y †β̇ (p02 )y †α̇ (p01 )
spins spins
X
=y 4
∆2F (p1 − p01 ) p2 · σβ β̇ p1 · σαα̇ y α (p01 )y β (p02 )y †β̇ (p02 )y †α̇ (p01 )
spins
X
=y 4 ∆2F (p1 − p01 ) p2 · σβ β̇ p02 · σ β̇β p1 · σαα̇ y α (p01 )y †α̇ (p01 )
spins
=y 4 ∆2F (p1 − p01 ) p2 · σβ β̇ p02 · σ β̇β p1 · σαα̇ p01 · σ α̇α

=y 4 ∆2F (p1 − p01 ) Tr [p1 · σ p01 · σ] Tr [p2 · σ p02 · σ] , (5.154)
By using (5.39)
|M1 |2 ⊃y 4 ∆2F (p1 − p01 ) pµ1 p01 Tr [σ µ σ ν ] pµ2 p02 Tr [σ µ σ ν ]
ν ν
(p1 · p01 ) (p2 · p02 ) .

=2y 4 2 (5.155)
(p1 − p01 )2 − m2h
where we also have used
1
∆F (q) = (5.156)
q2 − m2
The partial amplitude can be written in terms of the Mandelstam variables (5.113) and (5.115)
2
2

2 y 4 m21 + m01 − t m22 + m02 − t
|M1 | ⊃ 2 (5.157)
2 [t − m2h ]
For generic standard model Higgs couplings we have at least (in the specific case in consideration all
the fermionic masses are the same)
m1 =m01 , m2 =m02 , (5.158)
therefore
By using the general formula for the 2 → 2 cross section in eq. (??)
dσ 1 λ1/2 (s, m02 2 , m01 2 )
= |M|2
dΩ 64π 2 s λ1/2 (s, m22 , m21 )
1
= |M|2
64π 2 s
y4 m21 + m01 2 − t m22 + m02 2 − t
= 2 (5.159)
128π 2 s [t − m2h ]
dσ y 4 (2m21 − t) (2m22 − t)
= 2 (5.160)
dΩ 128π 2 s (t − m2h )
We consider two limits here for the partial contribution to the differential cross section:
• For massless intial an final fermions we have

dσ y4 t2
= (5.161)
dΩ 128π 2 s (t − m2h )2
• For some heavy fermion in the non-relativistic limit we have
t ≈(m1 − m1 )2 = 0 , s ≈(m1 + m2 )2 , (5.162)
and therefore
2
dσ y4 m1 m2
= . (5.163)
dΩ 32π 2 m4h m1 + m2
5.7.2 Other two-particle states

When the two particles are distinguishable: For example if the spin of qL is s2 and s0 and the spin
of eL is s1 and s, then
Z Z
α β d3 k d3 k 0 α β 0 0 −ik·x1 −ik0 ·x2
ξe+ (x1 )ξq+ (x2 )|qL (p2 )eL (p1 )i = √ √ x e (s, k)x q (s , k ) e e
(2π)3 2Ek V (2π)3 2Ek0 V
× as (k)as0 (k0 )a†s1 (p1 )a†s2 (p2 )|0i
Z Z
d3 k 0 d3 k α β 0 0 −ik·x1 −ik0 ·x2
= √ √ x e (s, k)x q (s , k ) e e
(2π)3 2Ek V (2π)3 2Ek0 V

× as (k)as0 (k0 )a†s1 (p1 )a†s2 (p2 ) − a†s1 (p1 )a†s2 (p2 )as (k)as0 (k0 ) |0i
Z Z
d3 k d3 k 0 0
= 3
√ 3
√ xαe (s, k)xβq (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V

× as (k)as0 (k ), as1 (p1 )a†s2 (p2 ) |0i.
0 †
(5.164)
5.7. SCATTERING 235

as (k)as0 (k0 ), a†s1 (p1 )a†s2 (p2 ) =as (k)(a( 0 (†(((
), as1 (p1 ) a†s2 (p2 ) − as (k), a†s1 (p1 ) as0 (k0 )a†s2 (p2 )
(
s0 (k
((

+ a†s1 (p1 )as (k) as0 (k0 ), a†s2 (p2 ) − a†s1 (p1 )(a( s (k),
((a(
† ((( 0
s2 (p2 ) as0 (k )

= − as (k), a†s1 (p1 ) as0 (k0 )a†s2 (p2 ) + as0 (k0 ), a†s2 (p2 ) a†s1 (p1 )as (k)
(5.166)
Therefore,

as (k)as0 (k0 ), a†s1 (p1 )a†s2 (p2 ) |0i

= − as (k), a†s1 (p1 ) as0 (k0 )a†s2 (p2 )|0i

= − as (k), a†s1 (p1 ) as0 (k0 ), a†s2 (p2 ) |0i
= −(2π)6 δ (3) (k − p1 )δ (3) (k0 − p2 )|0i (5.167)
and
β 1
α
ξe+ (x1 )ξq+ (x2 )|qL (p2 )eL (p1 )i = − √ √ xα (s, p1 )xβq (s0 , p2 ) e−ip1 ·x1 e−ip2 ·x2
2E1 V 2E2 V e
†α̇ 1
hqL (p2 )eL (p1 )|ξ †β̇ (x2 )ξ− (x1 ) = − √ √ x†qβ̇ (s0 , p2 )x†eα̇ (s, p1 ) eip1 ·x1 eip2 ·x2 (5.168)
2E1 V 2E2 V
Others possibilities for two initial particle states includes

1
|eR (p2 )eL (p1 )i = √ a†s1 (p1 )a†s2 (p2 )|0i (5.169)
V2
Following similar steps, we find for
†α̇ α
η+ (x1 )ξ+ (x2 )|eR (p2 )eL (p1 )i
1 †α̇
=√ √ y (p2 )xα (p1 ) e−ip2 ·x1 e−ip1 ·x2 −y †α̇ (p1 )xα (p2 ) e−ip1 ·x1 e−ip2 ·x2 |0i (5.170)
2E1 V 2E2 V
†α̇ α
heR (p2 )eL (p1 )|ξ− (x1 )η− (x2 )
1 †α̇
= h0| √ √ x (p2 )y β (p1 ) eip2 ·x1 eip1 ·x2 −x†α̇ (p1 )y α (p2 ) eip1 ·x1 eip2 ·x2 . (5.171)
2E1 V 2E2 V
Consider now
1
| (eL )† (p2 )eL (p1 )i = √ a†s1 (p1 )b†s2 (p2 )|0i (5.172)
V2
Therefore, the spin of (eL )† is s2 and s0 and the spin of eL is s1 and s, then
Z Z
d3 k d3 k 0 0
α
ξ+ †α̇
(x1 )ξ+ x(x2 )| (eL )† (p2 )eL (p1 )i = 3
√ 3
√ xα (s, k)y †α̇ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V
× as (k)as0 (k )bs1 (p1 )b†s2 (p2 )|0i
0 †
Z Z
d3 k 0 d3 k 0
= 3
√ 3
(2π) 2Ek V (2π) 2Ek0 V

× as (k)bs0 (k )as1 (p1 )b†s2 (p2 ) − a†s1 (p1 )b†s2 (p2 )bs (k)as0 (k0 ) |0i
0 †
Z Z
d3 k d3 k 0 0
= 3
√ 3
(2π) 2Ek V (2π) 2Ek0 V

0 †
× as (k)bs0 (k ), as1 (p1 )b†s2 (p2 ) |0i. (5.173)

as (k)bs0 (k0 ), a†s1 (p1 )b†s2 (p2 ) =as (k)(b( 0 (†(((
), as1 (p1 ) b†s2 (p2 ) − as (k), a†s1 (p1 ) bs0 (k0 )b†s2 (p2 )
(
s0 (k
((

+ a†s1 (p1 )as (k) bs0 (k0 ), b†s2 (p2 ) − a†s1 (p1 )(a(s (k),
((b(
† ((( 0
s2 (p2 ) bs0 (k )

= − as (k), a†s1 (p1 ) bs0 (k0 )b†s2 (p2 ) + bs0 (k0 ), b†s2 (p2 ) a†s1 (p1 )as (k)
(5.175)
Therefore,

as (k)bs0 (k0 ), a†s1 (p1 )b†s2 (p2 ) |0i

= − as (k), a†s1 (p1 ) bs0 (k0 )b†s2 (p2 )|0i

= − as (k), a†s1 (p1 ) bs0 (k0 ), b†s2 (p2 ) |0i
= −(2π)6 δ (3) (k − p1 )δ (3) (k0 − p2 )|0i (5.176)
Therefore
1
ξ+α †α̇
(x1 )ξ+ (x2 )| (eL )† (p2 )eL (p1 )i = − √ √ xα (s, p1 )y †α̇ (s0 , p2 ) e−ip1 ·x1 e−ip2 ·x2
2E1 V 2E2 V
D 1
† α †α̇
(eL ) (p2 )eL (p1 ) ξ− (x2 )ξ− (x1 ) = − √ √ y α (s0 , p2 )x†α̇ (s, p1 ) eip1 ·x1 eip2 ·x2 (5.177)
2E1 V 2E2 V
5.7. SCATTERING 237
5.7.3 Vector interactions

Bahba scattering
We consider the t-channel interaction
eL + (eL )† → eL + (eL )† (5.178)
mediated by Aµ , with couplings
L = eξα̇† (σ µ )α̇α ξα Aµ + eη α σαµα̇ η †α̇ Aµ (5.179)
The specific calculation by the scattering mediated by the Aµ is

Z Z
(−ie)2
Sf i = d4 x1 d4 x2 Aµ (x1 )Aν (x2 )
2!
D E

(eL )† (p02 ) , eL (p01 ) : ξα̇† (x1 ) (σ µ )α̇α ξα (x1 )ξβ̇† (x2 ) (σ ν )β̇β ξβ (x2 ) : (eL )† (p2 ) , eL (p1 )
(5.180)
Z Z
(−ie)2
= d4 x1 d4 x2 Dµν (x1 − x2 )
2!
D E

(eL )† (p02 ) , eL (p01 ) ξβ− (x2 )ξα̇−
† †
(x1 ) (σ µ )α̇α (σ ν )β̇β ξα+ (x1 )ξβ̇+ (x2 ) (eL )† (p2 ) , eL (p1 ) . (5.181)
By using (5.177), we have an expression similar to (5.145)

Z Z
(−ie)2 1
Sf i = p 0 p 0 √ √ d4 x1 d4 x2 Dµν (x1 − x2 )
2! 2E1 V 2E2 V 2E1 V 2E2 V
0 0
yβ (s02 , p02 ) x†α̇ (s01 , p01 ) (σ µ )α̇α (σ ν )β̇β xα (s1 , p1 ) yβ̇† (s2 , p2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2 . (5.182)
and following the same steps
(2π)4 (ie)2 δ (4) (p1 + p2 − p01 − p02 ) h † 0 µ α̇α 0 † ν β̇β 0

i
Sf i = p 0 p 0 √ √ xα̇ (p1 ) (σ ) xα (p1 )iDµν (p1 − p1 ) yβ̇ (p2 ) (σ ) yβ (p2 )
2E1 V 2E2 V 2E1 V 2E2 V
(2π)4 (ie)2 δ (4) (p1 + p2 − p01 − p02 ) † 0 µ
=p 0 p 0 √ √ x (p1 )σ x(p1 )iDµν (p1 − p01 ) y † (p2 )σ ν y(p02 )
2E1 V 2E2 V 2E1 V 2E2 V
(2π)4 (ie)2 δ (4) (p1 + p2 − p01 − p02 )
=p 0 p 0 √ √ x(p1 )σ µ x† (p01 )iDµν (p1 − p01 ) y † (p2 )σ ν y(p02 ) , (5.183)
2E1 V 2E2 V 2E1 V 2E2 V
where we use the identity for commuting spinors.

In this way

iM =e2 x(p1 )σ µ x† (p01 )iDµν (p1 − p01 ) y † (p2 )σ ν y(p02 ) (5.184)
If in addition we include the other three possibilites in the t-channel, by defining
e = (eR )† , e† =eR , (5.185)
then, we have the full contributions
eL + (eL )† →eL + (eL )† , eR + (eR )† →eR + (eR )† ,

eL + (eR )† →eL + (eR )† , eR + (eL )† →eR + (eL )† , (5.186)
which give the full t-channel amplitude [26] eq. (6.3.6), pag. 51:

iM =e2 x(p1 )σ µ x† (p01 ) + y † (p1 )σ µ y(p01 ) iDµν (p1 − p01 ) x(p2 )σ ν x† (p02 ) + y † (p2 )σ ν y(p02 )

=e2 x(p1 )σ µ x† (p01 ) + y † (p1 )σ µ y(p01 ) iDµν (t) x(p2 )σ ν x† (p02 ) + y † (p2 )σ ν y(p02 )

=e2 x(p1 )σ µ x† (p01 )iDµν (t)x(p2 )σ ν x† (p02 ) + x(p1 )σ µ x† (p01 )iDµν (t)y † (p2 )σ ν y(p02 )

+y † (p1 )σ µ y(p01 )iDµν (t)x(p2 )σ ν x† (p02 ) + y † (p1 )σ µ y(p01 )iDµν (t)y † (p2 )σ ν y(p02 ) . (5.187)
By using
−igµν
iDµν = , (5.188)
t
we have that
−ie2
iM = x(p1 )σ µ x† (p01 )x(p2 )σµ x† (p02 ) + x(p1 )σ µ x† (p01 )y † (p2 )σ µ y(p02 )
t
+y † (p1 )σ µ y(p01 )x(p2 )σµ x† (p02 ) + y † (p1 )σ µ y(p01 )y † (p2 )σ µ y(p02 ) . (5.189)
By using (5.38) we have for example
x(p1 )σ µ x† (p01 )y † (p2 )σ µ y(p02 ) =x(p1 )α σαµα̇ x†α̇ (p01 )yβ̇† (p2 )σ β̇β 0
µ yβ (p2 )
=2x(p1 )α δα β δ β̇ α̇ x†α̇ (p01 )yβ̇† (p2 )yβ (p02 )

=2x(p1 )α x†α̇ (p01 )yα̇† (p2 )yα (p02 )
=2x(p1 )y(p02 )x† (p01 )y † (p2 )
=2x(p1 )y(p02 )y † (p2 )x† (p01 ) . (5.190)
5.7. SCATTERING 239
In the same way
y † (p1 )σ µ y(p01 )x(p2 )σµ x† (p02 ) =2y † (p1 )x† (p02 )x(p2 )y(p01 ) , (5.191)
x(p1 )σ µ x† (p01 )x(p2 )σµ x† (p02 ) =2xα (p1 )x†α̇ (p01 )xβ (p2 )x†β̇ (p02 )αβ α̇β̇
=2xα (p1 )αβ xβ (p2 )α̇β̇ x†α̇ (p01 )x†β̇ (p02 )
= − 2xα (p1 )αβ xβ (p2 )β̇ α̇ x†α̇ (p01 )x†β̇ (p02 )
= − 2xα (p1 )xα (p2 )x†β̇ (p01 )x†β̇ (p02 )
= − 2x(p1 )x(p2 )x† (p01 )x† (p02 ) (5.192)
y † (p1 )σ µ y(p01 )y † (p2 )σ µ y(p02 ) = − 2y † (p1 )y † (p2 )y(p01 )y(p02 ) (5.193)
Therefore
−i2e2 †
iM = y (p1 )x† (p02 )x(p2 )y(p01 ) + x(p1 )y(p02 )y † (p2 )x† (p01 )
t
−x(p1 )x(p2 )x† (p01 )x† (p02 ) − y † (p1 )y † (p2 )y(p01 )y(p02 ) (5.194)
X 4e4 X h † 0† i
|M|2 = iM (iM)† = y x x y
2 1
0
+ x y 0 † 0†
y
1 2 2 1 x − x x x
1 2 1 2
0† 0†
x − y † † 0 0
y y y
spins
t2 spins 1 2 1 2 1 2
h i
× y10† x†2 x02 y1 + x01 y2 y20† x†1 − x02 x01 x†2 x†1 − y20† y10† y2 y1 (5.195)
We have for example

X X
y1† x0† 0 0† † 0
2 x2 y 1 y 1 x 2 x2 y 1 = y1†α̇ x0† α̇ α 0 0† †β̇ 0β
2 x2 y1α y1β̇ x2 x2 y1β
spins spins
X
=p01 · σαβ̇ y1†α̇ x0†α̇ α †β̇ 0β
2 x2 x2 x2 y1β
spins
X
=p01 · σαβ̇ y1†α̇ x0†α̇ †β̇ α 0β
2 x2 x2 x2 y1β
spins
X
=p01 · σαβ̇ p2 · σ β̇α y1†α̇ x0†α̇ 0β
2 x2 y1β
spins
X
=p01 · σαβ̇ p2 · σ β̇α p02 · σ α̇β y1†α̇ y1β
spins
X
=p01 · σαβ̇ p2 · σ β̇α p02 · σ α̇β y1β y1†α̇
spins
=p01 · σαβ̇ p2 · σ β̇α p02 · σ α̇β p1 · σβ α̇

= Tr [p01 · σp2 · σ] Tr [p1 · σp02 · σ]
=p01µ p2ν Tr [σ µ σ ν ] p1ρ p02η Tr [σ ρ σ η ]
=4 (p01 · p2 ) (p1 · p02 ) . (5.196)
for the crossed terms

X X
y1† x0† 0 0 0† †
2 x2 y1 x1 y2 y2 x1 = y1† x0† α 0 0β 0† †
2 x2 y1α x1 y2β y2 x1
spins spins
X
= y1† x0† α β 0† †
2 x2 −me δα y2β y2 x1
spins
→0 , when me → 0 . (5.197)
In this way all the crossed products disappear and only the squared of each term remains. The full
contributions are
X h † 0† i
y1 x2 x2 y10 y10† x†2 x02 y1 + x1 y20 y2† x0†
1 1x 0
y 2 y 0† †
x
2 1 + x 1 x 2 x 0† 0† 0 0 † †
x x x x
1 2 2 1 2 1x + y † † 0 0 0† 0†
y y y y y
1 2 1 2 2 1 y 2 y 1
spins
= 4 [(p01 · p2 ) (p1 · p02 ) + (p1 · p02 ) (p2 · p01 ) + 2 (p1 · p2 ) (p01 · p02 )]
= 8 [(p1 · p02 ) (p2 · p01 ) + (p1 · p2 ) (p01 · p02 )] , (5.198)
5.7. SCATTERING 241
and we get
X 32e4
|M|2 = [(p1 · p2 ) (p01 · p02 ) + (p1 · p02 ) (p2 · p01 )]
spins
t2
32e4 2 2

= s + u
4t2
(s2 + u2 )
=8e4 . (5.199)
t2
Therefore
dσ 1 λ1/2 (s, m02 2 , m01 2 )
= |M|2
dΩ 64π 2 s λ1/2 (s, m22 , m21 )
1
= |M|2
64π 2 s
e4 (s2 + u2 )
= 2 .
8π s t2
2 2
2 e2 (s + u2 )
=
s 4π t2
2α2 (s2 + u2 )
= . (5.200)
s t2
Fermion quark scattering

We consider the current of a Weyl fermion mediated by Z µ , with a coupling given by the Lagrangian
L = aL ξα̇† (σ µ )α̇α ξα Zµ . (5.201)
We now consider the t-channel interaction
e L + qL → e L + qL (5.202)
as displayed in figure 5.4.

We have then the two particle states (5.168) with distinguishable particles:
eL →ξe qL → ξq . (5.203)
We label the initial (final) momentum as p1 , p2 (p01 , p02 ) such that
me =m1 = m01 , mq =m2 = m02 . (5.204)

eL eL
qL qL
Figure 5.4: t-channel scattering of Weyl fermions
The specific calculation by the scattering mediated by the Z µ is

Z Z
(−iaL )2 X
Sf i = d4 x1 d4 x2 Zµ (x1 )Zν (x2 )
2! spins
hqL (p02 ) , eL (p01 )| : ξe†α̇ (x1 ) (σ µ )α̇α ξeα (x1 )ξq†β̇ (x2 ) (σ ν )β̇β ξqβ (x2 ) : |qL (p2 ) , eL (p1 )i
Z Z
(−iaL )2 X
= d4 x1 d4 x2 Dµν (x1 − x2 )
2! spins
hqL (p02 ) , eL (p01 )| ξe†α̇− (x1 )ξq†β̇− (x2 ) (σ µ )α̇α (σ ν )β̇β ξeα+ (x1 )ξqβ+ (x2 ) |qL (p2 ) , eL (p1 )i . (5.205)
By using (5.168) we have an expression similar to (5.183)
(2π)4 (iaL )2 δ (4) (p1 + p2 − p01 − p02 ) h † i

Sf i = p 0 p 0 √ √ xeα̇ (p01 ) (σ µ )α̇α xeα (p1 )iDµν (p1 − p01 ) x†qβ̇ (p02 ) (σ ν )β̇β xqβ (p2 )
2E1 V 2E2 V 2E1 V 2E2 V
(2π)4 (iaL )2 δ (4) (p1 + p2 − p01 − p02 ) † 0 µ
= p 0 p 0 √ √ xe (p1 )σ xe (p1 )iDµν (p1 − p01 ) x†q (p02 )σ ν xq (p2 ) . (5.206)
2E1 V 2E2 V 2E1 V 2E2 V
In this way
M =(iaL )2 x†e (p01 )σ µ xe (p1 )iDµν (t)x†q (p02 )σ ν xq (p2 ) (5.207)
By using
−igµν
iDµν = (5.208)
t − m2Z
5.7. SCATTERING 243
and the identity (5.38)

(−i) †
M =(iaL )2 x (p0 ) (σ µ )α̇α xeα (p1 )x†qβ̇ (p02 ) (σ µ )β̇β xqβ (p2 )
t − m2Z eα̇ 1
(−i) †
=(iaL )2 x (p0 )αβ α̇β̇ xeα (p1 )x†qβ̇ (p02 )xqβ (p2 )
t − m2Z eα̇ 1
(−i) βα
= − (iaL )2 xeα (p1 )xqβ (p2 )x†eα̇ (p01 )xq†α̇ (p02 )
t − m2Z
(−i) β
= − (iaL )2 x (p1 )xqβ (p2 )x†eα̇ (p01 )xq†α̇ (p02 )
t − m2Z e
(−i)
= − (iaL )2 x (p1 )xq (p2 )x†e (p01 )x†q (p02 ) .
2 e
(5.209)
t − mZ
We now can square the amplitude and sum over spins to obtain from (5.42) and (5.39)
X a4L X
|M|2 = 2 xe (p1 )xq (p2 )x†e (p01 )x†q (p02 )xq (p02 )xe (p01 )x†q (p2 )x†e (p1 )
spins
(t − m2Z ) spins
a4L X
= 2 Tr [p02 · σp01 · σ] xe (p1 )xq (p2 )x†q (p2 )x†e (p1 )
(t − m2Z ) spins
a4L
= 2 Tr [p02 · σp01 · σ] Tr [p2 · σp1 · σ]
(t − m2Z )
4a4L
= 2 (p1 · p2 ) (p01 · p02 ) . (5.210)
(t − m2Z )
The products of momentum can be written in terms of the Mandelstam variables (5.115)
X a4L 2
|M|2 = 2 s − m2e − m2q . (5.211)
spins
(t − m2Z )
The cross section is

dσ 1 λ1/2 (s, m02 2 , m01 2 )
= |M|2
dΩ 64π 2 s λ1/2 (s, m22 , m21 )
1
= |M|2
64π 2 s
1 a4L 2

2 2
= s − m e − mq . (5.212)
64π 2 s (t − m2Z )2

t ≈(me − me )2 ≈ 0 , s ≈(me + mq )2 , (5.213)
and, in this way,
2
dσ 1 a4L m2e + m2q + 2me mq − m2e − m2q
=
dΩ 64π 2 m4Z (me + mq )2
2
a4L me mq
= . (5.214)
16π 2 m4Z me + mq
Dark-matter like calculation

The following calculation is expected to be self-contained with every step worked out in detail.
To have the Higgs portal direct dark matter detection cross section The relevant term for the
scattering that we consider now is the t-channel
S(p1 ) + qL− (p2 ) → S(p01 ) + qR− (p02 ) (5.215)
As always, we use the left-handed Weyl spinors
qL →ξα , (qR )† →η α , (5.216)
We need the interaction of a singlet scalar with the standard model Higgs, h(x), and the standard
modelo Yukawa
L =λvS 2 (x)h(x) + y (qR )† qL (x)h0 + y (qL )† qR (x)h(x)
=λvS 2 (x)h(x) + yη α (x)ξα (x)h(x) + yη †α̇ (x)ξα̇† (x)h(x) . (5.217)
hence
q
S
Hint = − λvS 2 (x)h(x) , Hint = − yη α (x)ξα (x)h(x) , (5.218)
The S-matrix expansion includes
Z Z
(2) (−i)2 q
S = d4 x1 d4 x2 T{Hint
S
(x1 )Hint (x2 )}
2!
Z Z
(−i)2
= λyv d4 x1 d4 x2 T{(SSh)x1 (η α ξα h)x2 }
2!
Z Z
(−i)2
= λyv d4 x1 d4 x2 : (SSh)x1 (η α ξα h)x2 : + · · · (5.219)
2!
5.7. SCATTERING 245
In this way
Z Z
(2) (−iy)2
S (Sq → Sq) = d4 x1 d4 x2 h(x1 )h(x2 ) : (SS)x1 (η β ξβ )x2 : (5.220)
2!
On the other hand
: (SS)x1 (η β ξβ )x2 := : S(x1 )S(x1 )η β (x2 )ξβ (x2 ) : . (5.221)
We label the initial (final) momentum as p1 , p2 (p01 , p02 ) such that
mS =m1 = m01 , mq =m2 = m02 . (5.222)
The specific calculation by the scattering mediated by the Higgs is

Z Z
λvyq X
Sf i = d4 x1 d4 x2 h(x1 )h(x2 )
2! spins
hqR (p02 ) , S (p01 )| : S(x1 )S(x1 )η β ξβ (x2 )(x2 ) : |qL (p2 ) , S (p1 )i
Z Z
λmq X
=√ d4 x1 d4 x2 i∆F (x1 − x2 )
2 spins
β
× hqR (p02 ) , S (p01 )| S− (x1 )η− (x2 )S+ (x1 )ξβ+ (x2 ) |qL (p2 ) , S (p1 )i . (5.223)
√
where mq = yd v/ 2. This means that S(p1 ) (qL (p2 )) is destroyed in x1 (x2 ), while S(p01 ) (qL (p02 ))
is created in x1 (x2 ).
When the two particles are distinguishable: For example if the spin of qL is s2 and s0 , then
S+ (x1 )ξβ+ (x2 )|qL (p2 )S(p1 )i =

Z Z
d3 k d3 k 0 0
3
√ a(k) 3
√ xβ (s0 , k0 )as0 (k0 ) e−ik·x1 e−ik ·x2 |qL (p2 ) , S (p1 )i (5.224)
(2π) 2Ek V (2π) 2Ek0 V
The two particle Fock state is, after proper normalization
1
|qL (p2 )S(p1 )i = √ a† (p1 )a†s2 (p2 )|0i. (5.225)
V2
Therefore
Z Z
d3 k d3 k 0 0
S+ (x1 )ξβ+ (x2 )|qL (p2 )S(p1 )i = 3
√ 3
√ xβ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V
× a(k)as0 (k )a (p1 )a†s2 (p2 )|0i
0 †
Z Z
d3 k 0 d3 k 0
= 3
√ 3
√ xβ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V

× a(k)as0 (k )a (p1 )a†s2 (p2 ) − a† (p1 )a†s2 (p2 )a(k)as0 (k0 ) |0i
0 †
Z Z
d3 k d3 k 0 0
= 3
√ 3
√ xβ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V

× a(k)as0 (k ), a (p1 )a†s2 (p2 ) |0i.
0 †
(5.226)

a(k)as0 (k0 ), a† (p1 )a†s2 (p2 ) =a(k)(a( 0 (†(((
), a (p1 ) a†s2 (p2 ) − a(k), a† (p1 ) as0 (k0 )a†s2 (p2 )
(
s0 (k
((

+ a† (p1 )a(k) as0 (k0 ), a†s2 (p2 ) − a† (p1 )(a(k),
((( a†s( (((
(p 0
2 ) as0 (k )
2

= − a(k), a† (p1 ) as0 (k0 )a†s2 (p2 ) + as0 (k0 ), a†s2 (p2 ) a† (p1 )a(k) (5.228)
Therefore,

a(k)as0 (k0 ), a† (p1 )a†s2 (p2 ) |0i

= − a(k), a† (p1 ) as0 (k0 )a†s2 (p2 )|0i

= − a(k), a† (p1 ) as0 (k0 ), a†s2 (p2 ) |0i
= −(2π)6 δ (3) (k − p1 )δ (3) (k0 − p2 )|0i (5.229)
and
1
S+ (x1 )ξβ+ (x2 )|qL (p2 )S(p1 )i = − √ √ xβ (s0 , p2 ) e−ip1 ·x1 e−ip2 ·x2 |0i . (5.230)
2E1 V 2E2 V
Following similar steps, we find
†β̇ 1 −ip01 ·x1 −ip02 ·x2
S+ (x1 )η+ (x2 )|qR (p2 )S(p01 )i = − p p y †β̇ 0
(s , p 0
2 ) e e |0i (5.231)
2E10 V 2E20 V
β 1 0 0
hqR (p02 )S(p01 )|S− (x1 )η− (x2 ) = − h0| p 0 p 0 y β (s0 , p02 ) eip1 ·x1 eip2 ·x2 . (5.232)
2E1 V 2E2 V
5.7. SCATTERING 247
Hence
β
hqR (p02 )S(p01 )|S− (x1 )η− (x2 )S+ (x1 )ξβ+ (x2 )|qL (p2 )S(p1 )i
1 0 0 1
= h0| p 0 p 0 y β (s0 , p02 ) eip1 ·x1 eip2 ·x2 √ √ xβ (s0 , p2 ) e−ip1 ·x1 e−ip2 ·x2 |0i
2E1 V 2E2 V 2E1 V 2E2 V
1 1 0 0
=√ √ p 0 p 0 y β (s0 , p02 ) xβ (s0 , p2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2 h0|0i
2E1 V 2E2 V 2E1 V 2E2 V
1 1 0 0
=√ √ p 0 p 0 y β (s0 , p02 ) xβ (s0 , p2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2 . (5.233)
2E1 V 2E2 V 2E1 V 2E2 V
Checking back eq. (5.223), we see that

λmq 1 1 X
Sf i = √ √ √ p 0 p 0 y β (s0 , p02 ) xβ (s0 , p2 ) I, (5.234)
2 2E1 V 2E2 V 2E1 V 2E2 V spins
where
Z Z
0 0
I= d4 x1 d4 x2 i∆F (x1 − x2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2 . (5.235)
The integral is easily eavaluated if we use the Fourier expansion

Z
d4 q iq·(x1 −x2 )
i∆F (x1 − x2 ) = 4 i∆(q) e , (5.236)
(2π)
such that
Z Z
0 0
I= d4 x1 d4 x2 i∆F (x1 − x2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2
Z Z Z
d4 q 0 0
= 4 i∆(q) d4 x1 d4 x2 eiq·(x1 −x2 ) eip1 ·x1 eip2 ·x2 e−ip1 ·x1 e−ip2 ·x2 (5.237)
(2π)
Z Z Z
4 4 d4 x1 d4 x2 −i(−q−p01 +p1 )·x1 −i(q−p02 +p2 )·x2
= (2π) d q i∆(q) e e
(2π)4 (2π)4
Z
4
= (2π) d4 q i∆(q) δ 4 (−q − p01 + p1 ) δ 4 (q − p02 + p2 ). (5.238)
By using the first δ-function to make the final integration on q, we have
I = (2π)4 i∆(p1 − p01 ) δ 4 (p1 + p2 − p01 − p02 ) . (5.239)

Replacing back in eq. (5.234)
(2π)4 λmq δ (4) (p1 + p2 − p01 − p02 ) h 0β i

Sf i = p 0 p 0 √ √ y2 i∆F (p1 − p01 )x2β
2E1 V 2E2 V 2E1 V 2E2 V
(5.240)
where
y20β ≡y β (s0 , p02 ), x2β ≡xβ (s0 , p2 ). (5.241)
Comparing with the general expression for the 2 → 2 processes, we get that
iM = λmq ∆F (p1 − p01 )y20α x2α . (5.242)
As an initial step to calculate the probability we must squared this amplitude

X X
|M|2 =λ2 m2q ∆2F (p1 − p01 ) y20α x2α x†2α̇ y20†α̇
spins spins
X
=λ2 m2q ∆2F (p1 − p01 ) x2α x†2α̇ y20†α̇ y20α
spins
α̇α
=λ m2q ∆2F (p1 − p01 ) (p2 · σ)αα̇ (p02 · σ)
2
=2λ2 m2q ∆2F (p1 − p01 ) Tr (p2 · σp02 · σ)

=2λ2 m2q ∆2F (p1 − p01 ) Tr (pµ2 σµ p0ν
2 σν )
=2λ2 m2q ∆2F (p1 − p01 ) Tr (σµ σ ν ) pµ2 p0ν
2
2 2 2 0 µ 0ν
=4λ mq ∆F (p1 − p1 )gµν p2 p2
=4λ2 m2q ∆2F (p1 − p01 ) (p2 · p02 ) (5.243)
The partial amplitude can be written in terms of the Mandelstam variables (5.113) and (5.115)
X
2
|M|2 ==4λ2 m2q ∆2F (p1 − p01 ) m22 + m02 − t , (5.244)
spins
where
2 2
t = (p1 − p01 ) = (p2 − p02 ) . (5.245)
5.7. SCATTERING 249
By using
−i −i
i∆F (p1 − p01 ) = 0 2 2
= , (5.246)
(p1 − p1 ) − mh t − m2h
and, since mS = m1 = m01 , mq = m2 = m02

X λ2 m2q
|M|2 = 2 2m2q − t . (5.247)
spins
(t − m2h )
Therefore
dσ 1 λ1/2 (s, m02 2 , m01 2 )
= |M|2
dΩ 64π 2 s λ1/2 (s, m22 , m21 )
1/2
1 [s − (m01 + m02 )2 ][s − (m01 − m02 )2 ]
= |M|2 (5.248)
64π 2 s [s − (m1 + m2 )2 ][s − (m1 − m2 )2 ]
1/2
1 [s − (mS + mq )2 ][s − (mS − mq )2 ]
= |M|2 (5.249)
64π 2 s [s − (mS + mq )2 ][s − (mS − mq )2 ]
And we have
dσ 1
= |M|2
dΩ 64π 2 s
1 λ2 m2q 2

= 2 2 2mq − t . (5.250)
64π s (t − mh )
2
t ≈(mS − mS )2 ≈ 0 , s ≈(mS + mq )2 , (5.251)
dσ λ2 m4q
=
dΩ 32π 2 (mS + mq )2 m4h
λ2 m4q m2S
=
32π 2 m4h (mS + mq )2 m2S
2
dσ λ2 m2q mS mq
= . (5.252)
dΩ 32π 2 m4h m2S mS + mq
5.7.4 Example: Direct detection cross sections

Scalar portal
We repeat here the calculation of the Higgs mediated scattering in sec. 5.7.1 but for the explicit case
of distinguishable particles. To be more specific, we now consider the t-channel interaction
eL + qL → eR + qR , (5.253)
mediated by h(x), with couplings
L = ye (eR )† eL h0 + y (eL )† eR h0 + yq (qR )† qL h0 + y (qL )† qR h0 . (5.254)
The specific calculation by the scattering mediated by the Higgs, h(x), is

Z Z
(ye yq ) X
Sf i = d4 x1 d4 x2 h(x1 )h(x2 )
2! spins
hqR (p02 ) , eR (p01 )| : ξeα (x1 )ηeα (x1 )ξqβ (x2 )ηqβ (x2 ) : |qL (p2 ) , eL (p1 )i
Z Z
(ye yd ) X
= d4 x1 d4 x2 i∆F (x1 − x2 )
2! spins
β
hqR (p02 ) , eR (p01 )| ηeα− (x1 )ηqβ− (x2 )ξe+
α
(x1 )ξq+ (x2 ) |qL (p2 ) , eL (p1 )i . (5.255)
By using the standard procedure
(2π)4 ye yq δ (4) (p1 + p2 − p01 − p02 ) h 0 0 i

Sf i = p 0 p 0 √ √ y1α y2β i∆F (p1 − p01 )xα1 xβ2
2E1 V 2E2 V 2E1 V 2E2 V
(5.256)
iM = ye yq ∆F (t)x1 y10 x2 y20 (5.257)
X X
|M|2 =ye yq ∆F (t) x1 y10 x2 y20 y20† x†2 y10† x†1
spins spins
=4ye yq ∆F (t) (p2 · p02 ) (p1 · p01 ) (5.258)
By using
−i
i∆F (t) = (5.259)
t − m2h
5.8. DIRAC FERMION CASE 251
Since me = m1 = m01 , mq = m2 = m02

X ye yq
|M|2 = 2 2m2e − t 2m2q − t (5.260)
spins
(t − m2h )
dσ 1 λ1/2 (s, m02 2 , m01 2 )

= |M|2
dΩ 64π 2 s λ1/2 (s, m22 , m21 )
1
= |M|2
64π 2 s
1 ye2 yq2 2
2

= 2 2 2m e − t 2mq − t . (5.261)
64π s (t − m2h )
t ≈(me − me )2 ≈ 0 , s ≈(me + mq )2 , (5.262)
2
dσ ye2 yq2 me mq
=
dΩ 16π 2 m4h me + mq
2
m2e m2q me mq
= , (5.263)
16π 2 v 4 m4h me + mq
where mf = yf v.
Vector portal
See full treatment in https://indico.cern.ch/event/746178/contributions/3384110/attachments/
1849289/3035359/SUSY2019_SatomiOkada.pdf
5.8 Dirac fermion case

For the interaction of a fermion pair with Wµ± , we know from the standard model Lagrangian [1],
that
g2
√ ψγ µ (1 − γ5 )ψ (5.264)
2 2
Therefore in this case
Γ = γ µ (1 − γ5 ) (5.265)
2
For p MW the analysis is similar to the previous one with
W µ (x1 )W ν (x2 ) =h0| T{W µ (x1 )W ν (x2 )}|0i

Z
d4 q g µν −iq·(x1 −x2 )
≈ 2
e (5.266)
(2π)4 MW
For the process
e− (p) + νµ (k) → µ− (p0 ) + νe (k 0 ) (5.267)

2
the global coupling for p MW is
g2 GF
2
=√ (5.268)
8MW 2
√
After the replacement GF / 2 ≡ h2 /m2 , we have
(2) 1 1 1 1
Sf i = i(2π)4 δ 4 (p1 + p2 − p01 − p02 ) √ √ p 0 p 0 Mf i (5.269)
2E1 V 2E2 V 2E1 V 2E2 V
where
GF
Mf i = √ ūνe (p02 )Γue (p1 )ūµ (p01 )Γuνµ (p2 ) (5.270)
2
The corresponding Feynman diagram is shown in Fig. 5.5 Therefore we have
GF
Mf i = √ ūνe (p02 )γ µ (1 − γ5 )ue (p1 )ūµ (p01 )γµ (1 − γ5 )uνµ (p2 ) (5.271)
2
We now must sqaure the scattering amplitude, M, and summing up over final spin states, and
averaging over the intial spin states, as we did in Eq. (C.9). The result that will be obtained in detail
in Chapter D for the muon–decay is
|M|2 = 64 G2F (p1 · p2 )(p01 · p02 ) (5.272)

5.8. DIRAC FERMION CASE 253
e− (p1) x1 νe−(p′2)
W µ (p1 − p′2 )
νµ−(p2) x2 µ− (p′1)
Figure 5.5: scattering with four fermions
From Eq. (??)

dσ 1 s − m2µ
= |M|2 (5.273)
dΩ 64π 2 s s − m2e
The center of mass (CM) frame is defined by the condition in Eq. (??):
p1 + p2 = 0 (5.274)
The δ–function in Eq. (5.269)
δ (4) (p1 + p2 − p01 − p02 ) = δ (3) (p1 + p2 − p01 − p02 )δ(E1 + E2 − E10 − E20 ) (5.275)
implies
(
p1 = −p2
CM
p1 + p2 − p01 − p02 =0⇒ (5.276)
p01 = −p02
Moreover
√
s = E1 + E2 (5.277)
In the CM frame
s = (E1 + E2 )2
q q 2
= p21 + m21 + p22 + m22
q q 2
2 2 2 2
= p1 + me + p1 + mνe (5.278)
Therefore
E2 = |p1 | (5.279)
We already have the expression for |p1 | as given in eq. (??). In this case m2 = 0, and m1 = me , so
that
s − m2
|p1 | = √ e (5.280)
2 s
From (5.279)
√
s =E1 + E2
=E1 + |p1 | (5.281)
√
E1 = s − |p1 |
√ −s + m2e
= s+ √
2 s
2s − s + m2e
= √
2 s
s + m2
= √ e (5.282)
2 s
Then, by using Eqs. (5.276), (5.279) and (5.280), and (5.282), we have
p1 · p2 =E1 E2 − p1 · p2
=E1 |p1 | + p21
(s − m2e )(s + m2e ) (s − m2e )2
= +
4s 4s
(s − m2e )
= (s + m2e + s − m2e )
4s
1
= (s − m2e ) (5.283)
2
5.9. EXERCISES 255
As p22 = p02 2 = 0, we have from δ–function
(p1 + p2 )2 =(p01 + p02 )2

(p1 + p2 )2 =(p01 + p02 )2
2 2
p21 + 2p1 · p2 + p22 =p01 + 2p01 · p02 + p02
2
p21 + 2p1 · p2 =p01 + 2p01 · p02
m2e + 2p1 · p2 =m2µ + 2p01 · p02 (5.284)
1
p01 · p02 = p1 · p2 − (m2µ − m2e ) (5.285)
2
1
p01 · p02 = (s − m2µ ) (5.286)
2
Replacing back in Eq. (5.272) and then in Eq. (5.273) we have

dσ 1 s − m2µ 1 1
= 2 2
64G2F (s − m2e ) (s − m2µ ) (5.287)
dΩ 64π s s − me 2 2
dσ G2 (s − m2µ )2
= F2 (5.288)
dΩ 4π s
G2F (s − m2µ )2
σ= (5.289)
π s
Note that σ ∝ s.
5.9 Exercises
Display possible diagrams that correct the mass of the scalar in a theory with λφ4 , with and without
a φ3 -term.
Chapter 6
Neutrino masses
6.1 Standard model particle content

In the following discussion we use the following doublets
+
H νLi
H= , Li = − . (6.1)
H0 eLi
corresponding to the Higgs doublet and the lepton doublets respectively.
6.2 Weinberg operator

See Weinberg, PRL43(1979)1566:
yij
L= Li · HLj · H + h.c
Λ
yij
= ab cd Lai H b Lcj H d + h.c .
Λ
where yij are adimensional couplings and Λ some high-scale.

For a discussion around this operator see for example http://bit.ly/neutrinoreview.
The treatment in four notation component is given in Sec. B.1.
This operator breaks Lepton number (L) by two units, where the L = −1 and 0 for Li and
H respectively. In the same paradigm of the standard model, it is expected that the mass terms
associated with the Λ scale would be related to the spontaneous breaking of some symmetry.
257
258 CHAPTER 6. NEUTRINO MASSES
Field U(1)L
Li −1
H 0
NR −1
S +2
Table 6.1: Lepton number flux in green
H H
L NR NR L
Figure 6.1: Seesaw mechanism
6.3 Type-I seesaw

To keep the paradigm of the standard model where the masses of the fundamental fields must be
generated from the spontaneous breaking of some symmetry, we assume that at some scale Λ the
lepton number is still preserved even after the introduction of right-handed neutrinos
Lν = yi (NR )† Li · H + 12 hR NR NR S + h.c . (6.2)
where yi and hR are adimensional couplings, while S is an scalar singlet which can acquire a vacuum
expectation value hSi, such that the right handed neutrino can acquire a Majorana mass
MR = hR hSi . (6.3)
The assignment of L is given in Table 6.1

We can build the diagram displayed in fig. 6.1, where the arrows indicate the flux of lepton
number.
After the spontaneous symmetry breaking, if we assume to have just one right-handed neutrino
NR
Lν = yi (NR )† Li · H + 12 MR NR NR + h.c . (6.4)
To track down the relevant term beyond the standard model, it is convenient to reassign L such
that the resulting lepton number violating terms involve some of standard model fields. In fact, by
6.3. TYPE-I SEESAW 259
reassigning L = 0 to NR , it turns out that the only lepton number violating terms are the ones
involving the Yukawa coupling yi in eq (6.4). Therefore
mν ∝ y i , (6.5)
and we recover massless neutrinos when yi → 0.

In addition to the previous Majorana mass contribution MR , after the spontaneous symmetry
breaking the neutral states νLi and NR acquire a Dirac mass contribution
1
Lν ⊃MDi (νLi )† NR + MR NR NR + h.c
2
1 i 1 1
= MD (νLi ) NR + MDi NR (νLi )† + MR NR NR + h.c
†
2 2 2
1 03×3 MD (νL )†
= (νL )† NR + h.c
2 MDT MR NR
1
≡ χT Mχ χ + h.c , (6.6)
2
where the Dirac neutrino mass vector MD has components
yiv
MDi = √ , (6.7)
2
T T
(νL )† T = νL1 νL2 νL3 , χT = (νL )† NR , and

03×3 MD
Mχ = (6.8)
MDT MR
which has eigenvalues

1h 1/2 i
Mχ ∓ = MR ∓ MR2 + 4MD MDT
2
1 h 1/2 i
= MR 1 ∓ 1 + 4MD MR−2 MDT . (6.9)
2
When MR2 MD MDT
1
Mχ ∓ ≈ MR 1 ∓ 1 + 2MD MR−2 MDT (6.10)
2
Therefore
Mν → Mχ− ≈ − MD MR−1 MDT

Mχ+ ≈MR . (6.11)
The effective neutrino mass matrix, Mν , in this case is just a rank one matrix
Mν → − MD MR−1 MDT
(Mν )ij ≈ − MDi MR−1 MDj . (6.12)
It is possible to write the final result in terms of a Majorana fermion defined from the Weyl spinors
in the diagonal by blocks basis, denoted with a zero superindex be low. We have
0
νL Mν 0
ν= † , mν = . (6.13)
(νL0 ) 0 Mν
Therefore
Lν ⊃ 21 ν c mν ν
=···
= 21 νL0 Mν νL0 + h.c . (6.14)
Note that Abelian Lepton number is not longer a conserved symmetry of the Lagrangian. For
example by assigning Lepton −1 for NR , the Majorana term NR NR violates Lepton number by two
units. However we can identify a remnant Z2 symmetry under which the Leptons are odd and all the
other SM particles are even, including the Higgs doublet. This symmetry is called Lepton parity.
6.4 Dirac operator

If neutrinos are Dirac particles, the Standard Model (SM) particle content must be extended with
right-handed neutrinos, and some symmetry must be imposed to prevent their Majorana mass terms.
At least a Z3 symmetry is required to forbid both the tree-level Yukawa terms and Majorana mass
terms
1
D
Lν = yαi (νRα )† Li · H + mR νR νR + h.c , (6.15)
2
with Li ·H = ab Lai H b , where Li are the lepton doublets, H is the SM Higgs doublet with hypercharge
Y = 1, and yD is the matrix of neutrino Yukawa couplings. A possible assignment to forbid only the
6.5. DIRAC TYPE-I SEESAW 261
Field U(1)L
Li −1
H 0
νR −4
S +3
Table 6.2: Lepton number flux in green
Majorana mass term is obtained if the SM fields that transform non-trivially under Z3 as: L ∼ ω,
(eR )† ∼ ω 2 and (νR )† ∼ ω 2 , with ω 3 = 1. At this level, the neutrino mass problem is not longer a
phenomenological issue but a theoretical one, in which it is necessary to explain the smallness of the
Yukawa couplings in yD , which must be of order 10−11 .
To do so, we assume that the symmetry allows for the 5-dimensional operator with total lepton
number conservation [?]
yαi
L5 = (νRα )† Li · H S + h.c. , (6.16)
Λ
where yαi are adimensional couplings and Λ is the new physics scale. To be compatible with neutrino
oscillation data [?], y should be at least of order 2 × 3.
6.5 Dirac Type-I seesaw

T
We need to assume the existence of a heavy Dirac fermion, F = FL FR , where FLR are Weyl
fermions, and one real scalar singlet S. The most general Lagrangian is [27]
Lν = yi (FR )† Li · H + MF (FL )† FR + hi (FL )† SνRi + h.c . (6.17)
The assignment of L is given in Table 6.2

We can build the diagram displayed in fig. 6.2, where the arrows indicate the flux of lepton
number.
In addition to the previous Dirac mass contribution MF , after the full spontaneous symmetry
breaking the neutral states νLi and νRi acquire a Dirac mass contributions MDi and MSi respectively
Lν ⊃MDi (νLi )†i FR + MF (FL )† FR + MSi (FL )† νRi + h.c .

† †
03×3 MD νR
= (νL ) (FL ) T + h.c
MS MF FR
≡ (χL )† Mχ χR + h.c , (6.18)
H S
L FR FR νR
Figure 6.2: Seesaw mechanism
where the Dirac neutrino mass vectors MD,S have components
yiv hi hSi
MDi = √ , MSi = √ , (6.19)
2 2
T
T T
νLR = νLR1 νLR2 νLR3 , χT
LR = νLR FLR , and

03×3 MD
Mχ = (6.20)
MST MF
which has eigenvalues

1h 1/2 i
Mχ ∓ = MF ∓ MF2 + 4MD MST . (6.21)
2
When MF2 MD MST we have effective neutrino mass matrix with a single light neutrino mass
Mχ− → (Mν )ij ≈ − MDi MF−1 MSj , (6.22)
and a heavy mass Mχ+ ≈ MF .

Note that the recipe to pass from Majorana to Dirac is just to replace MDT → MST .
For go to the diagonal basis, we define
νR0 =VR χR , νL0 =VL χL . (6.23)
Therefore
Lν ⊃ (χL )† Mχ χR + h.c

⊃ (χL )† VL† VL Mχ VR† VR χR + h.c
†
⊃ νL0 Mν νR0 + h.c , (6.24)
6.5. DIRAC TYPE-I SEESAW 263
where
Mν ≡ VL Mχ VR† = diag(Mχ− , Mχ+ ) . (6.25)
It is convenient to write the final result in terms of a Dirac fermion defined from the Weyl spinors
in the diagonal basis. We have
0
νL Mν 0
ν= , mν = . (6.26)
νR0 0 Mν
Therefore
Lν ⊃ ν mν ν
0 1 M 0
0
νL
† † ν
= (νL0 ) (νR0 )
1 0 0 Mν νR0
0 M ν 0
0 † 0 † ν L
= (νL ) (νR )
Mν 0 νR0
M ν 0
0 † 0 † ν R
= (νL ) (νR )
Mν νL0
†
= νL0 Mν νR0 + h.c . (6.27)
Chapter 7
Leptogenesis
In ./Leptogenesis/ [PDF] by Katherine Builes Londoño, the lepton asymmetry is calculated by

using two-component spinor techniques.
265
266 CHAPTER 7. LEPTOGENESIS
Chapter 8
One-loop neutrino masses
Here we calculate the one-loop neutrino mases in the mass-basis
8.1 Extra particles

We assume to have α-scalar and n-fermions in the mass eigenstate basis in Weyl spinor notation
L = (yinα νi χn Sα + mn χn χn + h.c) + 21 m2α Sα2 . (8.1)
From left to right (clockwise) in Figure 8.1
Z
d4k
−iΣνij (p) = (yinα ) iSF (k) (yjnα ) i∆F (p + k)
(2π)4
Z 4
yinα yjnα dk k + mχn
= (8.2)

16π 2 iπ k 2 − m2χn (p + k)2 − m2Sα
2
χn χn
k
p p+k p
νL S νL
Figure 8.1: Generic one-loop neutrino mass contribution
267
268 CHAPTER 8. ONE-LOOP NEUTRINO MASSES
In the limit p → 0
yinα yjnα 2 2

Mijν = − m χ n B0 0; mχ , mS . (8.3)
16π 2 n α

where B0 0; m2χn , m2Sα is the B0 Passarino-Veltman function [?] in dimensional regularization
Z
(2πµ) 1
B0 0; m2χn , m2Sα = dd k . (8.4)
iπ 2 k2 − m2χn k 2 − m2Sα
where µ is the subtraction point of dimensional regularization and d = 4 − .
By following Romao notes: http://porthos.tecnico.ulisboa.pt/Public/textos/one-loop.
pdf1 it will be shown below that this can be decomposed as
1
B0 0; m2χn , m2Sα = A0 m2
χn − A0 m2
Sα , (8.5)
m2χn − m2Sα
where
Z
2
(2πµ) 1
A0 m = dd k . (8.6)
iπ 2 k2 − m2
To evaluate the integral through the Passarino-Veltman functions, we start with the general
integral with two denominators
Z Y1
2 2 2
(2πµ) d 1
B0 r10 , m0 , m1 = d k (8.7)
iπ 2
i=0
(k + ri )2 − m2i
In particular
Z
dd k 1
I=
(2π)d [(k + p)2 − m21 + i] [k 2 − m22 + i]
Z 1 Z
dd k 1
= dx
0 (2π) [k + 2p · kx + p x − m21 x − m22 (1 − x) + i]2
d 2 2
Z 1 Z
dd k 1
= dx
0 (2π) [k + 2P (x) · k − M 2 (x) + i]2
d 2
Z 1 Z
dd k 1
= dx (8.8)
0 (2π)d {[k + P (x)]2 − P 2 (x) − M 2 (x) + i}2
1
Old version, now is a part of ”Advanced Quantum Field Theory” http://porthos.ist.utl.pt/Public/textos/
tca.pdf
8.1. EXTRA PARTICLES 269
where
P (x) =xp , M 2 (x) = − xp2 + m21 x + m22 (1 − x). (8.9)
Changing variable k → k − P
Z 1 Z
dd k 1
I= dx (8.10)
0 (2π) [k − C(x) + i]2
d 2
where
C(x) = P 2 (x) + M 2 (x) (8.11)
In general, the integral to be calculated with dimensional regularization is

Z r
dd k k2
Ir,m = . (8.12)
(2π)d [k 2 − C + i]m
Making a Wick rotation, this can be written as
Z r
r−m dd kE kE2
Ir,m = i(−1) m (8.13)
(2π)d [kE2 + C]
where kE = (kE0 , k) is an euclidean vector with

2
k 0 =ikE0 , and kE2 = kE0 + |k|2 . (8.14)
The integral can be evaluated to give

r−m+ d2 (−1)r−m Γ r + d2 Γ m − r − d2
Ir,m = iC d
. (8.15)
(4π) 2 Γ d2 Γ(m)
By using d = 4 −
2
(−1)r−m 4π 2+r−m Γ 2 + r − 2 Γ m − r − 2 + 2
Ir,m =i C . (8.16)
(4π)2 C Γ 2 − 2 Γ(m)
We are interested in
2
i 4π
I0,2 = Γ . (8.17)
(4π)2 C 2
where
2
Γ = + ψ(1) + O() (8.18)
2
where
d
ψ(z) = ln Γ(z) (8.19)
dz
ψ(1) = −γ . (8.20)
So that
i
I0,2 = (∆ − ln C) (8.21)
16π 2
where
2
∆ = − γ + ln 4π , (8.22)

where γ is the Euler-Mascheroni constant.
Replacing back C(x) from eq (8.11) in (8.7)
Z 1
2
−x(1 − x)p2 + xm21 + (1 − x)m20
B0 p , m20 , m21 = ∆ − dx ln (8.23)
0 µ2
In[1]:= Integrate[Log[a x], x]
Out[1]= -x + x Log[a x]
In[2]:= Dt[-x + x Log[a x], x, Constants -> a]

Out[3]= Log[a x]
Taking into account that
i
I0,1 = C (1 + ∆ − ln C) , (8.24)
16π 2
we can define
Z
2
(2πµ) 1
A0 m = dd k . (8.25)
iπ 2 k 2 − m2
8.1. EXTRA PARTICLES 271
such that

A0 m2 = m2 ∆ + 1 − ln m2 /µ2
In this way, from (8.23) we can get

m20 m21
m20 ln µ2
− m21 ln µ2
B0 0, m20 , m21 =∆ + 1 − 2
m0 − m21
A0 (m20 ) − A0 (m21 )
= . (8.26)
m20 − m21
Replacing back in eq. (8.5), we have

1 2
B0 0; m2χn , m2Sα = m χn ∆ + 1 − ln m 2
χn /µ2
− m2
Sα ∆ + 1 − ln m 2
Sα /µ2
m2χn − m2Sα

m2Sα ln m2Sα /µ2 − m2χn ln m2χn /µ2
=(∆ + 1) +
m2χn − m2Sα

m2Sα ln m2Sα − ln (µ2 ) − m2χn ln m2χn − ln (µ2 )
=(∆ + 1) +
m2χn − m2Sα
2 2
2 2

m S ln mS − m χ ln m χ
= ∆ + 1 + ln µ2 + α α n n
m2χn − m2Sα

m2Sα ln m2Sα − m2χn ln m2χn
= cte(∞) + . (8.27)
m2χn − m2Sα
and replacing back in eq. (8.3):
yinα yjnα 2

Mijν = − mχ n cte(∞) + f mχ n , mSα , (8.28)
16π 2
where

m2Sα ln m2Sα − m2χn ln m2χn
f m2χn , m2Sα = (8.29)
m2χn − m2Sα
See ./FeynCalc/radiativeseesaw.nb for an example based in the Radiative seesaw below

To obtain in detail could be useful the Feynman parameters for the one-loop neutrino integral in
Fig. ??
8.2 Applications
8.3 Radiative seesaw
It is based in the inert-doublet scalar dark matter model where in order to have a viable scalar dark
matter particle we need a sizeable splitting between the real an imaginary part of the inert doublet
 
η+
η =  ρ0 + iA0  (8.30)
√
2
2 2
V =µ21 H † H + µ22 η † η + λ1 H † H + λ2 η † η + λ3 H † H η † η + λ4 H † η η † H
h 2 i
+ λ5 η † H + H.c. (8.31)
The terms relevant for the mass terms are

µ21 2 λ1 4 2 + − µ22 02 µ22 02 λ3 2 + − 1 02 02

V ⊃ (h + v) + (h + v) + µ2 η η + ρ + A + (h + v) η η + ρ +A
2 4 2 2 2 2
λ4 v 2 02 λ5 v 2 02
+ ρ + A02 + ρ − A02 . (8.32)
4 4
From here, in addition to the known standard model Higgs mass, we obtain
m2η± =µ22 + λ3 v 2 ,
v2
m2ρ =µ22 + (λ3 + λ4 + λ5 ) ,
2
v2
m2A =µ22 + (λ3 + λ4 − λ5 ) .
2
(8.33)
If we assign zero lepton number to NR and −1 to η (see below) the only Lepton number violating
term is the one with λ5 . In this way we expect small neutrino masses for small λ5 . In the limit λ5 →,
mρ = mA and the neutrinos are massless. To explain the smallness of λ5 we may follow the same
idea than in [28]
S
λ5 η † H → λ5 η † H (8.34)
M
The diagram of charge flux is displayed in figure 8.2
8.3. RADIATIVE SEESAW 273
Figure 8.2: Radiative seesaw with small λ5 (top S can also be S ∗ if required)
Fields U(1)global
L l= 0
H h= 0
NR n = −1
η η = −1
S s = −2
Table 8.1: Global solution, see ./scotolocal.nb, to diagram of charge flux in figure 8.2
The relevant Lagrangian must be consistent with
l + η =n , 2n =s , 2h = 2η ± s . (8.35)
The solution for minus sign in the last equation, corresponding to S ∗ in the top part of figure 8.2.,
with a global U(1)global [28]2 is displayed in table 8.1.
Therefore must include the fermion terms (matrix notation for family indices)
L ⊃ h (NR )† L · η + yNR NR S + h.c . (8.36)
After the spontaneous braking of the global symmetry we have the relevant terms
h i
e5 η † H 2 + h.c + µ2 η † η + λ3 H † H η † η + λ4 H † η η † H ,
L ⊃ h (NR )† L · η + MR NR NR + λ 2
(8.37)
2
Local by just adding NL with same charge −1
Figure 8.3: Contribution to neutrino mass in mass basis
where
MR =yhSi , e5 =λ5 hSi .

λ (8.38)
M
If we assign zero lepton number to NR we can assign 0 (+1) lepton number to η such that the term
with coupling h (λe5 ) is violates lepton number. Therefore the light neutrino masses are proportional
e5 . In which follows we will drop the tilde in λ5 .
to the product of y λ
The Lagrangian in the mass basis, relevant for the generation of neutrino masses is
h 1 1
L ⊃ √ (NR )† νL ρ0 + iA0 + MR NR NR + m2ρ ρ02 + m2A A02 . (8.39)
2 2 2
which generate the two contribution to the neutrino mass displayed in figure 8.3
h2
Mijν = − 2
MR f MR2 , m2ρ − f MR2 , m2A . (8.40)
16π
8.4 Singlet-doublet fermions with scalar singlets

There [?]
yinα =hiα N3n (8.41)
and
3
X hiα hjα X
Mijν = (N3n )2 mχn f m2Sα , m2χn . (8.42)
α
16π 2 n=1
8.4. SINGLET-DOUBLET FERMIONS WITH SCALAR SINGLETS 275
8.4.1 Zee
√
In the Zee model we can work in the Higgs-basis with hH1 i = v/ 2, hH2 i = 0 [?]. In that basis the
scalar potential is
1
V =µ21 H1† H1 + µ22 H2† H2 − [µ23 H1† H2 + H.c.] + λ1 (H1† H1 )2
2
1
+ λ2 (H2 H2 ) + λ3 (H1 H1 )(H2 H2 ) + λ4 (H1† H2 )(H2† H1 )
† 2 † †
2

1 † 2 † † †
+ λ5 (H1 H2 ) + [λ6 (H1 H1 ) + λ7 (H2 H2 )]H1 H2 + H.c.
2
+ µ2h |h+ |2 + λh |h+ |4 + λ8 |h+ |2 H1† H1 + λ9 |h+ |2 H2† H2
+ λ10 |h+ |2 (H1† H2 + H.c.) + µαβ H1α H2β h− . (8.43)
We define

G+ √ H+ √
H1 = H2 = (8.44)
Re (H10 ) + iG0 / 2 Re (H20 ) + iA0 / 2
The minimum equations come from

∂V ∂ 2

0 2 1
0 4
= µ Re H1 + λ1 Re H1 + · · ·
∂ Re (H10 ) ∂ Re (H10 ) 1 2
1 3
=2µ21 Re H10 + λ1 4 Re H10 + · · · (8.45)
2
∂V 2 0

0 3
= − µ 3 2 Re H1 + λ6 2 Re H1 + ··· (8.46)
∂ Re (H20 )
From the conditions

∂V

=0 , (8.47)
∂ Re H1,2
0
√
hH1 i→v/ 2,hH2 i→0
we have
1 1
µ21 + λ1 =0 −µ3 + λ6 =0 (8.48)
2 2
or
1 1
µ21 = − λ1 µ3 = λ6 . (8.49)
2 2
Replacing back in the potential, we have for the charged scalars that In the basis Φ† = (G− , H − , h− )
the squared-mass matrix for the charged Higgs states is given by
Lcharged = Φ† M2C Φ + · · · , (8.50)
where
 
0 0 0√
M2C = 0 MH2 ±√ −µv/ 2 , and (8.51)
0 −µv/ 2 M233
1
MH2 ± =µ22 + v 2 λ3
2
M233 =µ2h + v 2 λ8 . (8.52)

By defining the mass basis state as S = G− , h.1 , h−
2 , then after the rotation
S = RΦ , (8.53)
where
 
1 0 0
R = 0 cos ϕ sin ϕ  . (8.54)
0 − sin ϕ cos ϕ
Then,
− 2 + −
Lcharged = M12 h+
1 h1 + M2 h2 h2 + · · · (8.55)
The relevant terms for neutrino masses in the mass eigenstate basis reads
Lν =Π02 cos ϕ (eR )† νL h− 0 † − + +

1 − Π2 sin ϕ (eR ) νL h2 + fji (eL )j (νL )i sin ϕh1 + fji (eL )j (νL )i cos ϕh2
cl (eR )† eL + h.c + M 2 h+ h− + M 2 h+ h−
+M (8.56)
1 1 1 2 2 2
Therefore
ckk fjk sin ϕ
Mijν ∝ (Π02 )ik cos ϕM (8.57)
8.4. SINGLET-DOUBLET FERMIONS WITH SCALAR SINGLETS 277
In that case n corresponds to the usual leptons labeled with i, j, k, . . ., and α = 1, 2
L =νLj Ojk eRk R1α h+

α+
νLi T C (2fik ) eLk R2α h+
α+
mk eLk eRk + h.c , (8.58)
Therefore
yikα =Oik R1α

yjkα =2fik Rα1
(8.59)
where

cos ϕ − sin ϕ
R= (8.60)
sin ϕ cos ϕ
X X 2Oik R1α fjk R2α

Mijν = − mk cte(∞) + f m2χn , m2Sα
k α
16π 2
X Oik fjk
2 2 2 2
=− m k R 11 R 21 cte(∞) + R11 R 21 f m k , M1 + R 12 R22 cte(∞) + R12 R22 f mk , M 2
k
8π 2
X Oik fjk
2 2 2 2
=− m k cos ϕ sin ϕ f m k , M 1 − f m k , M 2
k
8π 2
X Oik fjk
2 2
≈− m k sin 2ϕ f 0, M 1 − f 0, M 2
k
16π 2
X Oik fjk 2 2
M1 M2 ln (M12 ) − M12 M22 ln (M22 )
=− mk sin 2ϕ
k
16π 2 M12 M22
X fjk mk Oik sin 2ϕ M 2
2
= 2
ln 2
, (8.61)
k
(4π) M 1
To compare with [?]. There is factor 2?

sin 2ϕ X
Mijν = − 2 (fik mk Ojk + fjk mk Oik ) f m2k , M12 − f m2k , M22 (8.62)
(4π) k
In the generic basis, The important factor is

√ tan β 1 √ tan β 1
fik mk Ojk + fjk mk Oik =fik mk − 2 mk + Π2jk + fjk mk − 2 mk + Π2ik
v cos β v cos β
√ tan β 1
=− 2 fik m2k + m2k fjk + fik mk Π2jk + Π2ik mk fjk (8.63)
v cos β
Seem to be that there is a mistake in [?]w wiht tan β.
8.4.2 Minimal Zee

According to [?], the famous Zee-Wolfenstein matrix by settin Π2 = 0 in eq. (8.63)
This case corresponds to the limit Π2 = 0 of [?]
√ tan β
L=− 2 νLj mj eRj R1α h+
α+
v
νLi T C (2fik ) eLk R2α h+
α + h.c , (8.64)
Therefore
√ tan β
yikα = − 2 mk δik R1α
v
yjkα =2fik Rα1
(8.65)
X fjk mk δik sin 2ϕ

M22
Mijν = ln
k
(4π)2 M12

fji mi sin 2ϕ M22
= ln . (8.66)
(4π)2 M12
which is non-zero only for i 6= j.

Chapter 9
Incubator
In ./Incubator/ [PDF] new themes to include in the main book are started.
Currently:
1. Anomaly cancelation.
2. Discrete symmetries.
279
280 CHAPTER 9. INCUBATOR
Appendix A
Computational QFT
There are several tools which allows for the generation of models of particle physics models like
LanHEP [22]
http://theory.sinp.msu.ru/~semenov/lanhep.html,
or FeynRules [23]
http://feynrules.phys.ucl.ac.be/ .
This kind of programs are able to generate the output required for other programs which make
the calculation of Feynman diagrams and integration over multi-particle phase space. CalcHEP:
http://theory.sinp.msu.ru/~pukhov/calchep.html
for example, is able to calculate cross section and decays widths at tree level.
In this chapter we will illustrate the use LanHEP+CalcHEP
A.1 LanHEP
After download the source code from the http://theory.sinp.msu.ru/~pukhov/calchep.html to
some DIR,
• Note that the tar.gz file name depends on the current version. At the moment of this writing
this was lhep311.tar.gz. To directly download this file use:
$ wget http://theory.sinp.msu.ru/~semenov/lhep311.tar.gz
281
282 APPENDIX A. COMPUTATIONAL QFT
where \$ is to indicate that the command is to be written in the shell of your Linux session1 .
To uncompress the file:
$ tar -zxvf DIR/lhep311.tar.gz
• Go to the created directory

$ cd lanhep311
• To compile and create the executable of the program (lhep):
$ make
The input of LanHEP are files were the Lagrangian of some model is written in a symbolic way.
Then, the LanHEP executable process the input files and generates four outputfiles which are the
input for the CalcHEP program. For example, in the LanHEP dir
$ ./lhep stand.mdl
Here ./lhep command, search for the file in the defaul directory mdl/stand.mdl. If there are no erros
printed, for files are created:
ls *4.mdl
func4.mdl lgrng4.mdl prtcls4.mdl vars4.mdl
A.2 CalcHEP
The installation of CalcHEP is similar. In Ubuntu you must be sure to have libx11-dev package, in
addion to the C and Fortran compilers:
$ sudo apt-get install libx11-dev build-essential gfortran
In the CalcHEP directory:

$ make
1
An introduction to scientific computing is at http://gfif.udea.edu.co/cf
A.3. LANHEP/CALCHEP 283
To use CalcHEP you must first create a directory with the required files. This is achieved with
the CalcHEP command
$ ./mkUsrDir YourModel
A directory YourModel is created with several files and directories inside. By default, a models
directory is created with two set of .mdl files, corresponding to two versions of the Standard Model:
$ ls YourModel/models/
From the YourModel directory in CalcHEP, run the command

./calchep
A new window must appear with the info of CalcHEP and the loaded models in YourModel/models.
To navigate through this window, use the arrows keys and the <ESC> key to navigate back into the
menus.
A.3 LanHEP/CalcHEP
The sample .mdl files in the mdl directory of LanHEP must be modified in order to generate the
proper CalcHEP input files. From the LanHEP directory
$ mkdir sm
$ cd sm
$ wget --no-check-certificate \
https://github.com/rescolo/LanHEP/raw/release/sm/sm.mdl
$ wget --no-check-certificate \
https://github.com/rescolo/LanHEP/raw/release/sm/sm_tex.mdl
$ ../lhep sm.mdl
The four CalcHEP input files:

are then created.

From CalcHEP directoty:
$ ./mkUsrDir sm
$ cd sm/models
$ rm *
then copy the *1.mdl files to the sm/models, and from the sm CalcHEP directory run ./calchep.
In order to understand the structure of the LanHEP files consider the following skeleton:
1 model ’MODEL NAME’/N.
2 % The coments are either this way
3 /* or this other way */
4
5 use file_tex.
6
7 prtcprop pdg.
8
9 prtcformat fullname:’ Full Name ’,

10 name:’ P ’,
11 aname:’ aP ’,
12 pdg:’ number ’,
13 spin2, mass, width, color, aux,
14 texname:’> LaTeX P name <’,
15 atexname:’> LaTeX aP name < ’.
16
17 parameter VAR = VALUE : ’Description’.

18
19 particle_type
20 particle/Antiparticle: (’name’, property name=VALUE, ...).
21
22 lterm Write here the Lagrangian in a LaTeX--like format

23
24 prtcprop pdg:(Particle=PDF code,...).

25
26 SetEM(A,EE). %check charge conservation

27 CheckHerm.
In line 1, N is an integer that will identify the four output files. The file in line 5 will contain the
LATEX definitions of the used particles. In lines 7-15, the format of the table prtclN.mdl, as required
by CalcHEP, is defined: A new column with the PDG number for the particle. In line 17, the general
form to declarate a variable is established, while the lines 19-20 are the generic declaration for a
particle. The final commands in 26 and 27 is to check the consistency of the defined model. As a
simple illustration consider the simple case of QED:
model ’QED: e, mu tau’/1.
use qed3g_tex.
prtcprop pdg.
%prtc1.mdl is one of the output files of LanHEP. To make

% it compatible with CalcHEP we need to change their format
% to include the PDG particle number in the third column
prtcformat fullname:’ Full Name ’,

name:’ P ’,
aname:’ aP ’,
pdg:’ number ’,
spin2, mass, width, color, aux,
texname:’> LaTeX P name <’,
atexname:’> LaTeX aP name < ’.
parameter EE = 0.31333 : ’Electromagnetic coupling constant (<->1/128)’.
vector
A/A: (photon, gauge).
spinor e1:(electron),
e2:(muon, mass Mm = 0.1057),
e3:(’tau-lepton’, mass Mt = 1.777).
% fermion interaction with gauge fields
lterm anti(psi)*gamma*(i*deriv - EE*A)*psi

where
psi=e1;
psi=e2;
psi=e3.
% gauge bosons Lagrangian
lterm -F**2/4 where

F=deriv^mu*A^nu-deriv^nu*A^mu.
%set PDG particle numbers:
prtcprop pdg:(A=22,e1=11, e2=13, e3=15).

SetEM(A,EE). %check charge conservation

CheckHerm.
where the required file qed3g_tex.mdl is

SetTexName([e1=e,E1=’\\bar{e}’]).
SetTexName([e=e,E=’\\bar{e}’]).
SetTexName([’e1.c’=’e^c’,’E1.c’=’\\bar{e}^c’]).
SetTexName([’e.c’=’e^c’,’E.c’=’\\bar{e}^c’]).
SetTexName([e2=’\\mu’,E2=’\\bar{\\mu}’]).
SetTexName([e3=’\\tau’,E3=’\\bar{\\tau}’]).
SetTexName([m=’\\mu’,M=’\\bar{\\mu}’]).
SetTexName([l=’\\tau’,L=’\\bar{\\tau}’]).
SetTexName([EE=e]).
SetTexName([Me=’M_e’, Mm=’M_\\mu’, Mt=’M_\\tau’]).
Running with the option -tex:

../lhep qed3g.mdl
the following output is generated

• lgrng1.tex
Fields in the vertex Variational derivative of Lagrangian by fields

µ
ēa eb Aµ −e · γab
µ
µ̄a µb Aµ −e · γab
µ
τ̄ a τb Aµ −e · γab
• prtcls1.tex:
P aP Name Spin EM charge Color Comment

Aµ Aµ photon 1 0 1 gauge
ea ēa electron 1/2 1 1
µa µ̄a muon 1/2 1 1
τa τ̄ a tau-lepton 1/2 1 1
• vars1.tex
Figure A.1: CalcHEP welcome window
Parameter Value Comment

EE 0.31333 Electromagnetic coupling constant (1/128)
Mm 0.1057 mass of muon
Mt 1.777 mass of tau-lepton
With the command

$ ../lhep qed3g.mdl
the same files are generated by in the format of CalcHEP.

parameter is for constants to be exported to tables, while let is only for internal LanHEP variables.
In CalcHEP
\$ ./mkUsrDir qed3g
\$ cd qed3g/models
\$ rm *
#copy the files: func1.mdl, lgrng1.mdl prtcls1.mdl, vars1.mdl here
\$ cd ..
\$ ./calchep
The window in Fig. A.1 After hit <Enter>, the window with the model should appears as shown in
Fig. A.2 To test that the model was loaded without errors:
QED: e, mu tau -> Edit Model -> Check Model
A message with The model is OK, should popup.

After returning to the model window in Fig. A.2, we could calculate some process:
Figure A.2: CalcHEP model window
QED: e, mu tau -> Enter Process
and enter the process e1,E1 -> e2,E2 (e+ e− → µ+ µ− ) as shown in the Fig: A.3 After <Enter>, the
window to calculate the process should appears, as in Fig. A.4 In addition to View diagrams, we can
calculate the process with the sequence
Squaring technique -> Symbolic calculations -> C-compiler
Then a new window with the process details should appears, as displayed in Fig. A.5 After adjust
the input parameters at your convenience, we could just calculate the process with, in this case:
Easy 2-2, to obtain the result displayed in A.6 e.g, for center of mass energy of 14 TeV (7 TeV per
beam) we could have:
σ(e+ e− → µ+ µ− ) = 5 × 10−4 pb (A.1)
• Exercise: Repeat the previous calculation, but for one center of mass energy of 200 GeV.
For the Standard Model the Yukawa Lagrangian that couple the down fermions with the boson
scalar is written in the interaction basis:
−LY ∼ D0 MD PR D0 H + h.c , (A.2)
with D0 = V † D, we can write the eq. (A.2) in the mass eigenstates as
−LY ∼ DMDdia PR DH + h.c (A.3)

Figure A.3: CalcHEP process window
Figure A.4: CalcHEP calculation window

Figure A.5: CalcHEP calculation window
Figure A.6: CalcHEP results window

where
MDdia = V MD V † (A.4)
Investing the equation A.4 and replacing in (A.2) we can write in the interaction eigenstates:
−LY ∼ (D0 V † )(MDdia V )PR D0 H + h.c (A.5)
Expanding we get . . . which is just the expresion in the Standard Model file
Appendix B
Neutrinos
B.1 Weinberg operator

Based on Akhmedov, hep-ph/0001264; Langacker, arXiv:1112.5992
With the definition
T
ψ c =C ψ̄ T ψLc =CψL .
The Majorana mass term should be of the form νLc νL . Since νL has I3 = 1/2, the Majorana mass
T
term has I3 = 11 . With L = νL eL
Lc τ iτ2 L ∼ (3, −2) .
One would need an isotriplet scalar field ∆ ∼ (3, 2), which either elemental or composite. The term
H T τ iτ2 H ∼ (3, −2) ,

T
can play the role of the composite triplet, where H = H + H 0
Can the operator
f
Lc τ iτ2 L · H T τ iτ2 H
M
1
Any SU (2) spinor χ, χT iτ2 satisfy that
• both χ†1 χ2 and χT1 iτ χ2 are invariant.
• both χ†1 τ χ2 and χT1 iτ τ χ2 transform as vectors.
293
294 APPENDIX B. NEUTRINOS
be generated as an effective operator at some loop level?

Explicitly we have

c c
0 1 0 −i 1 0 eL
L τ iτ2 L = νL eL
c , ,
1 0 i 0 0 −1 −νL

−νL ν e
= νLc ecL ,i L , L
eL eL νL

= ecL eL − νLc νL , i νLc νL + ecL eL , νLc eL + ecL νL
Replacing back νLc → H + , ecL → H 0 , νL → H + , and eL → H 0 , we have

H T τ iτ2 H = H 0 H 0 − H + H + , i H 0 H 0 + H + H + , H + H 0 + H + H + .

Lc τ iτ2 L · H T τ iτ2 H

= 2 −ecL eL H + H + − νLc νL H 0 H 0 + νLc eL H + H 0 + ecL νL H + H 0

= −2 νLc H 0 − ecL H + H 0 νL − H + eL (B.1)
0
H νL
= −2 νLc ecL H 0 −H +
−H + eL
We have the Weinberg operator

f
L=− Lc τ iτ2 L · H T τ iτ2 H + h.c
2M
f c e ∗ e †
= LH H L + h.c
M
f
= νLc H 0 − ecL H + H 0 νL − H + eL + h.c
M
f
= 12 Lc1 H2 + 21 Lc2 H1 (12 L1 H2 + 21 L2 H1 ) + +h.c
M
f
= Lca Hc Lb Hd ac bd + h.c .
M
See Weinberg, PRL43(1979)1566.
Appendix C
Two body decays
In this chapter we use directly the Feynman rules for Fermions to carry out the calculation of the decay
of the standard model Higgs into a pair of fermions. In chapter 5 we will obtain the corresponding
Feynman rules from the S–matrix expansion.
C.1 Particle decays

Particle decay [9] is the spontaneous process of one elementary particle transforming into other
elementary particles. During this process, an elementary particle becomes a different particle with
less mass and an intermediate particle such as W boson in muon decay.
For a particle of a mass M , the differential decay width according Eq. (4.117), is
(2π)4
dΓn = |M|2 dΦ(n) (P ; p1 , p2 , . . . , pn ) (C.1)
2M
The phase space can be determined from Eq. (4.116)

!
X
n Y
n
d3 pi
dΦ(n) (P ; p1 , p2 , . . . , pn ) = δ 4 (P − pi ) . (C.2)
i=1 i=1
(2π)3 2Ei
We will keep the dΓ notation until all the integrals get evaluated.
The two-body decays in eq. (4.148) is
dΓ 1
= 2 3
|Mf i |2 λ1/2 (M 2 , m22 , m21 ) (C.3)
dΩ 64π M
295
296 APPENDIX C. TWO BODY DECAYS
C.2 Width decay

Reglas de Feynman: Time direction from left to right.
• Initial particle: u(p)
• Initial antiparticle: v̄(p)
• Final particle: ū(p)
• Final antiparticle: v(p)
We consider now a general Yukawa interaction term
Lint = hHf 1 f2 (C.4)
For the H → f 1 f2 decay. The interaction between the Higgs boson with fermions1 is given by the
Yukawa interaction term [1]
(v + H)
LHiggs = −Gf √ (f R fL + f L fR )
2
Gf v Gf H
= − √ ff − √ ff
2 2
√ 1/2
= −mf f f − mf GF 2 ff (C.5)
√
Such as the electro has acquired a √mass me = Gf ν/ 2. On the other hand the coupling to be
assigned to the process vertex is Gf 2 or mf /v =.
The decay process H → f f , is displayed in Fig. C.1
The Feynman rules, to be explained in Chapter 5 are indicated in Fig. C.2.
In this way the scattering amplitude is
√ 1/2
iM = −imf GF 2 u(s1 , p1 )v(s2 , p2 ). (C.6)
where p1 , s, p2 y s2 are the momentum and spines of fermion and anti–fermion respectively.
For the general case
iM = −ihu(s1 , p1 )v(s2 , p2 ). (C.7)

1
In this case we consider only electrons, by the formula is easy generalizable to other fermions
C.2. WIDTH DECAY 297
Figure C.1: Diagrama de proceso H → f f
Figure C.2: Reglas de Feynman del proceso H → f f
√ 1/2 †
h = mf GF 2 in the standard models Now, having into account that γ 0 = γ 0
(u(s1 , p1 )v(s2 , p2 ))†

= v † (s2 , p2 )(u(s1 , p1 ))†
= v † (s2 , p2 )(u† (s1 , p1 )γ 0 )†
†
= v † (s2 , p2 )(γ 0 u(s1 , p1 ))
= v † (s2 , p2 )(γ 0 u(s1 , p1 ))
= v(s2 , p2 )u(s1 , p1 ).
Squaring M, and summing over possible polarization states of final particles, we have
X X
|M|2 = h2 (u(s1 , p1 )v(s2 , p2 ))(v(s2 , p2 )u(s1 , p1 )). (C.8)
s1 ,s2 s1 ,s2
The several sums in Ec. (C.8) can be calculated by expressing the products uv y vu en in terms of
their components, as follow
X
(u(s1 , p1 )v(s2 , p2 ))(v(s2 , p2 )u(s1 , p1 ))
s1 ,s2
X
= (uα (s1 , p1 )vα (s2 , p2 ))(v β (s2 , p2 )uβ (s1 , p1 ))
s1 ,s2
X
= (uβ (s1 , p1 )uα (s1 , p1 ))(vα (s2 , p2 )v β (s2 , p2 ))
s1 ,s2
X X
= uβ (s1 , p1 )uα (s1 , p1 ) vα (s2 , p2 )v β (s2 , p2 )
s s2
= (p 1 + mf )βα (p 2 − mf )αβ

= Tr[(p 1 + mf )(p 2 − mf )]. (C.9)
Taking into account that Tr[γν ] = 0, and from the commutation relations for γµ matrices
Tr[γµ γν ] = Tr[−γν γµ + 2g µν ]
= Tr[−γν γµ ] + 2g µν Tr[1]
= Tr[−γµ γν ] + 2g µν 4 (Tr[AB] = Tr[BA])
µν
Tr[γµ γν ] = 4g .
In this way
Tr[(p 1 + m1 )(p 2 − m2 )]
= Tr[(γµ pµ1 + m1 )(γν pν2 − m1 )]
= Tr[γµ γν pµ1 pν2 − m2 γµ pµ1 + m1 γν pν2 − m1 m2 ]
= pµ1 pν2 Tr[γµ γν ] − 4m1 m2
= 4gµν pµ1 pν2 − 4m1 m2
= 4(p1 · p2 − m1 m2 ).
where m1 , m2 are the final masses, and
X
|M|2 = 4h2 (p1 · p2 − m1 m2 ).
s1 ,s2
From eq. (??)

M =E1 + E2
|p1 | = |p2 | (C.10)
C.2. WIDTH DECAY 299
Therefore
M 2 − E12 − E22
E1 E2 = (C.11)
2
p1 · p2 − m1 m2 = E1 E2 − p1 · p2 − m1 m2
= E1 E2 + p21 − m1 m2
M 2 − E12 − E22
= + p21 − m1 m2
2
1
= M 2 − m21 − p21 − m22 − p21 + p21 − m1 m2
2
1
= M 2 − m21 − m22 − 2m1 m2
2
1 2
= M − (m1 − m2 )2
2
Therefore, the scattering amplitude is
X
|M|2 = 2h2 M 2 − (m1 + m2 )2 (C.12)
s1 ,s2
Replacing back in eq. (C.3)
dΓ h2 1/2 2 2 2
2 2

= λ (M , m2 , m1 ) M − (m1 + m 2 ) (C.13)
dΩ 32π 2 M 3
R
After the integration dΩCM = 4π 2 we have
h2 1/2 2 2 2 2
Γ= 3
λ (M , m2 , m1 ) M − (m1 + m2 )2 (C.14)
8πM
For m1 = m2 = mf
1/2
1/2 2
4m2f
λ (M , m22 , m21 )
=M 1 − 2
M2

2 2
2
4m2f
M − (m1 + m2 ) =M 1 − (C.15)
M2
2
R 2π Rπ
0
dφ 0
sin θdθ = 4π
and therefore
3/2
h2 4m2f
Γ(H → f f ) = M 1− (C.16)
8π M2
In the case of the standard model Higgs with mass MH decaying to fermion pair, according to the
Lagrangian in eq. (C.5)
3/2
MH m2f GF m2f
Γ(H → f f ) = √ 1−4 2 , (C.17)
4π 2 MH
In the limit mf MH this expression reduces to
MH m2f GF
Γ(H → f f ) = √ . (C.18)
4π 2
C.3 e + e − → µ+ µ−
e2
L= v̄(k2 )γ λ u(k1 ) v̄(k2 )γ λ u(k1 ) (C.19)
s
Appendix D
Three body decays
D.1 Muon decay

For a three body decay we have from eq. (C.1)
1 1 2 4 d3 p1 d3 p2 d3 p3
dΓ = |M| δ (P − p 1 − p 2 − p 3 )
(2π)5 2M 2E1 2E2 2E3
3 Z
1 1 d p1 2 4 d3 p2 d3 p3
= |M| δ (P − p1 − p2 − p3 )
(2π)5 2M 2E1 2E2 2E3
(D.1)
D.1.1 Amplitude estimation

Since M is dimensionless, the amplitude averaged over spins for µ decay must be
|M|2 = CG2F m4µ . (D.2)
We use
1
C = (2 × 2 × 1 × 1) = 2 (D.3)
2
The first factor is for the initial average and the factor are for the number of spin states of µ, e and
the two neutrinos.
301
302 APPENDIX D. THREE BODY DECAYS
Consider first the integral

Z Z
4 d3 p1 d3 p2 d3 p1 d3 p2
δ (P − p1 − p2 ) = δ(E − E1 − E2 )δ 3 (P − p1 − p2 )
2E1 2E2 2E1 2E2
Z 3 Z
d p2
= δ(E − E1 − E2 ) δ 3 (P − p1 − p2 )d3 p1
4E1 E2
Z Z
d3 p2
= δ(E − E1 − E2 ) δ 3 (p1 + p2 − P)d3 p1
4E1 E2
Z Z
d3 p2
= δ(E − E1 − E2 ) δ 3 [p1 − (P − p2 )]d3 p1 (D.4)
4E1 E2
using
Z
δ(x − x0 )dx = 1 (D.5)
we have
Z Z
4 d3 p1 d3 p2 d3 p2
δ (P − p1 − p2 ) = δ(E − E1 − E2 )
2E1 2E2 4E1 E2
Z 3 3 Z
4 d p1 d p2 p22 d|p2 |dΩ
δ (P − p1 − p2 ) = δ(E − E1 − E2 ) (D.6)
2E1 2E2 4E1 E2
Since |p1 | = |p2 | we have
E = E1 + E2 =(m21 + p21 )1/2 + (m22 + p22 )1/2

=(m21 + p22 )1/2 + (m22 + p22 )1/2 (D.7)
differentiating this equation with respect to p2

dE 1 2|p2 | 2|p2 |
= +
d|p2 | 2 (m21 + p22 )1/2 (m22 + p22 )1/2

1 1
=|p2 | +
E E2
1
E1 + E2
=|p2 | (D.8)
E1 E2
Therefore

dE E1 E2
d|p2 | = (D.9)
|p2 | E1 + E2
D.1. MUON DECAY 303
replacing back in eq. (D.6)

Z Z
4 d3 p1 d3 p2 dE p22 dΩ E1 E2
δ (P − p1 − p2 ) = δ(E − E1 − E2 )
2E1 2E2 |p2 | 4E1 E2 E1 + E2
Z
|p2 | dΩ
= dE δ(E − E1 − E2 )
4(E1 + E2 )
Z
|p2 |
= dΩ (D.10)
4E
For a relativistic particle |p2 | ≈ E2 = E/2 and
Z
d3 p1 d3 p2
δ 4 (P − p1 − p2 ) = 2π (D.11)
E1 E2
Applying this result to eq. (D.1) we have
Z
1 1 d3 p1 2 4 d3 p2 d3 p3
dΓ = |M| δ (P − p 1 − p 2 − p 3 )
(2π)5 2M 2E1 2E2 2E3
3 Z
1 1 d p1 2 4 d3 p2 d3 p3
= |M| δ (P − p1 − p2 − p3 )
(2π)5 2M 8E1 E2 E3
3
1 1 d p1
= 5
|M|2 (2π)
(2π) 2M 8E1
Z
G2F m4µ 2
= p d|p1 | dΩ
8(2π)4 mµ E1 1
G2F m3µ
≈ 4
E12 dE1 (4π)
8(2π) E1
G2F m3µ
≈ E1 dE1 (D.12)
4(2π)3
As the maximum value of E1 is mµ /2
Z
G2F m3µ mµ /2
Γ≈ E1 dE1
4(2π)3 0
G2F m3µ m2µ
= , (D.13)
4(2π)3 8
or
3 G2F m5µ
Γ= . (D.14)
4 192π 3
D.1.2 Amplitude calculation

The Standard Model Lagrangian includes
√
2g
L=− (νeL γ µ eL Wµ+ + ēL γ µ νeL Wµ− + νµ L γ µ µL Wµ+ + µ̄L γ µ νµ L Wµ− )
2
g
= − √ (νe γ µ PL eWµ+ + ēγ µ PL νe Wµ− + νµ γ µ PL µWµ+ + µ̄γ µ PL νµ Wµ− )
2
g
= − √ (νe γ µ (1 − γ5 )eWµ+ + ēγ µ (1 − γ5 )νe Wµ− + νµ γ µ (1 − γ5 )µWµ+ + µ̄γ µ (1 − γ5 )νµ Wµ− )
2 2
(D.15)
where
(νe γ µ PL eWµ+ )† = e† γ µ PL † (νe )† Wµ− = e† PL γ µ† γ 0 νe Wµ− = e† γ 0 γ 0 γ µ† PR γ 0 νe Wµ−
= ēγ 0 γ µ† γ 0 PL νe Wµ− = ēγ µ PL νe Wµ− (D.16)
We can build the effective Lagrangian
Applying the Feynman rules to the diagram in fig. D.1 we have the amplitude
µν
−ig 2 2
−g + q µ q ν /MW
M= ū3 γµ (1 − γ5 )u1 2
ū4 γν (1 − γ5 )v2 (D.17)
8 q 2 − MW
where u (v) is for an incoming particle and ū (v̄) is for an ongoing particle (antiparticle).
The Dirac equations for spinors u and v are
(p − m)u = 0 (p + m)v = 0
ū(p − m) = 0 v̄(p + m) = 0 . (D.18)
In this way
1 1
γ q µ (1 − γ5 )u1 ū4 γν q ν (1 − γ5 ) = 2 (1 + γ5 )qu
2 µ − γ5 )
1 ū4 q(1
MW MW
mµ me
=− 2
(1 + γ5 )u1 ū4 (1 − γ5 ) (D.19)
MW
the term in q µ q ν can be safely neglected. The term q 2 is m2µ is small compared with m2W . Therefore
−ig 2
M= 2
ū3 γµ (1 − γ5 )u1 ū4 γ µ (1 − γ5 )v2
8MW
−iGF
= √ ū3 γµ (1 − γ5 )u1 ū4 γ µ (1 − γ5 )v2 (D.20)
2
D.1. MUON DECAY 305
Figure D.1: Tree level diagram for muon decay
M is a dimensionless scalar. The relevant coupling is

g2 GF
2
=√ . (D.21)
8MW 2
The conjugate is, following the same steps that in eq. (D.16)
ig 2
M∗ = 2
[ū3 γµ (1 − γ5 )u1 ]† [ū4 γ µ (1 − γ5 )v2 ]†
8MW
ig 2
M∗ = 2
[ū1 γµ (1 − γ5 )u3 ] [v̄2 γ µ (1 − γ5 )u4 ] . (D.22)
8MW
Multiplying M and M∗ we have
g4
|M|2 = 4
[ū3 γµ (1 − γ5 )u1 ū1 γν (1 − γ5 )u3 ]
64MW
× [ū4 γ µ (1 − γ5 )v2 v̄2 γ ν (1 − γ5 )u4 ]
g4
= 4
Lµν M µν (D.23)
64MW
where

Lµν = ū3α γµαβ (1 − γ5 )βγ uγ1 ū1δ γνδ (1 − γ5 )η uη3
µ

M µν = ūα4 γαβ (1 − γ5 )βγ v2γ v̄2δ γδ
ν
(1 − γ5 )η u4η (D.24)

Lµν = uη3 ū3α γµαβ (1 − γ5 )βγ uγ1 ū1δ γνδ (1 − γ5 )η

= (u3 ū3 )ηα γµαβ (1 − γ5 )βγ (u1 ū1 )γδ γνδ (1 − γ5 )η
= Tr [(u3 ū3 )γµ (1 − γ5 )(u1 ū1 )γν (1 − γ5 )] (D.25)
Using
X X
u(p, s)ū(p, s) = (p + m) v(p, s)v̄(p, s) = (p − m) (D.26)
s s
X
Lµν = Tr [(p 3 )γµ (1 − γ5 )(p 1 + mµ )γν (1 − γ5 )]
s
h i
β
=pα3 Tr γα γµ (1 − γ5 )(p1 γβ + mµ )γν (1 − γ5 )
h i
=pα3 Tr (γα γµ − γα γµ γ5 )(pβ1 γβ γν (1 − γ5 ) + mµ γν (1 − γ5 ))
h
=pα3 Tr pβ1 γα γµ γβ γν (1 − γ5 ) − pβ1 γα γµ γ5 γβ γν (1 − γ5 )
+mµ γα γµ γν (1 − γ5 ) − mµ γα γµ γ5 γν (1 − γ5 )]
h
=pα3 Tr pβ1 γα γµ γβ γν − pβ1 γα γµ γβ γν γ5 − pβ1 γα γµ γ5 γβ γν + pβ1 γα γµ γ5 γβ γν γ5
+mµ γα γµ γν − mµ γα γµ γν γ5 − mµ γα γµ γ5 γν + mµ γα γµ γ5 γν γ5 ] (D.27)
as the trace of an odd number of γ–matrices is zero, we have
X
Lµν =pα3 pβ1 Tr γα γµ γβ γν − γα γµ γβ γν γ5 − γα γµ γβ γν γ5 + γα γµ γβ γν γ52
s
=2pα3 pβ1 Tr [γα γµ γβ γν (1 − γ5 )] (D.28)
Similarly
X
M µν = 2p4δ p2 Tr γ δ γ µ γ γ ν (1 − γ5 ) (D.29)
s
D.1. MUON DECAY 307
substituting back in eq. (D.23) we have,
g4
|M|2 = 4
4pα3 pβ1 p4δ p2 Tr [γα γµ γβ γν (1 − γ5 )] Tr γ δ γ µ γ γ ν (1 − γ5 )
64MW
g4
= 4
4pα3 pβ1 p4δ p2 (64δδα δβ )
64MW
g4
= 4
4 × 64(p3 · p4 )(p1 · p2 )
64MW
4g 4
= 4 (p3 · p4 )(p1 · p2 )
MW
2
g2
=4 8 2
(p3 · p4 )(p1 · p2 )
8MW
2
GF
=4 8 √ (p3 · p4 )(p1 · p2 )
2
=128 G2F (p3 · p4 )(p1 · p2 )
G2
=256 F (p3 · p4 )(p1 · p2 ) . (D.30)
2
The demonstration of the used Tr×Tr identity can be found in Appendix B. of [11].
The spin–averaged differential decay width for µ− → νµ e− ν̄e is

1 1 d3 p3 1 X 2 4 d3 p 2 d3 p4
dΓ = |M| δ (p 1 − p 2 − p 3 − p 4 )
(2π)5 2E1 2E3 2 2E2 2E4
1 1 X 1 d p2 43 3
d p3 d3 p4
= |M|2 δ (p 1 − p 2 − p 3 − p 4 )
2E1 2 (2π)5 8E2 E3 E4
4 3
1 4g 1 d p2 4 d3 p3 d3 p4
= 4
(p 1 · p 2 )(p 3 · p 4 ) δ (p 1 − p 2 − p 3 − p 4 )
2 MW (2π)5 2E1 2E2 2E3 2E4
4
2g
= pβ pα d3 p4 Iαβ (D.31)
5
16(2π) MW 4
E1 E4 1 4
where the covariant integral Iαβ on the neutrino momentum is

Z
d3 p2 d3 p3
Iαβ = p3α p2β δ 4 (p − p2 − p3 ) . (D.32)
E2 E3
The variable p in ec. (D.32) is defined as p = p1 − p4 = p2 + p3 . Moreover

p2 =p22 + p23 + 2p2 · p3
=m2νe + m2νµ + 2p2 · p3
≈2p2 · p3
gαβ pα pβ =2gαβ pα3 pβ2
pα pβ =2pα3 pβ2 . (D.33)
Iαβ must have the form
Iαβ = gαβ A(p2 ) + pα pβ B(p2 ) . (D.34)
Defining the itegral I as follows
Z
d3 p2 d3 p3
I= δ 4 (p − p2 − p3 ) , (D.35)
E2 E3
Since
m2ν ≈ 0 =Eν2 − p2ν
Eν2 =p2ν (D.36)
and in addition the integral I is covariant, we choose to evaluate it in the rest frame of the two
neutrinos |p2 | = |p3 |, which implies E2 = E3 .
Z
3 d3 p 2 d3 p3
I = δ(E − E2 − E3 )δ (p − p2 − p3 )
E2 E3
Z Z
d3 p2
= δ(E − E2 − E3 ) δ 3 (p − p2 − p3 )d3 p3
E2 E3
| {z }
=1
Z
δ(E − 2E2 ) 2
= p2 d|p2 |dΩ
E22
Z
δ(E − 2E2 ) 2
= E2 dE2 (4π)
E22
Z
E
=4π δ 2 E2 − dE2
2
Z
1 E
=4π δ E2 − dE2
2 2
=2π (D.37)
D.1. MUON DECAY 309
then multiplying (6.13) by g αβ and pα pβ successively gives, using eq. (D.33)

Z
2 d3 p2 d3 p3 p2
αβ
g Iαβ = 4A + p B = p3 · p2 δ 4 (p − p2 − p3 ) = I = πp2
E2 E3 2
In order to compute pα pβ Iαβ , we make use of the fact that it is a Lorentz invariant quantity, so
that we may evaluate it in any reference frame. In particular, we can evaluate it in the rest frame of
the neutrinos involved in this process. This means that p = p2 + p3 = (p0 , 0) and E2 = E3
pα pβ Iαβ = p2 A + p4 B (D.38)
Z 3
d p2 d3 p3
α β
=p p p3α p2β δ 4 (p − p2 − p3 )
E2 E3
Z 3
d p2 d3 p3
= E3 p0 E2 p0 δ 4 (p − p2 − p3 )
E2 E3
Z
0 2
= (p ) d3 p2 d3 p3 δ 4 (p − p2 − p3 ) (D.39)
Z
0 2
= (p ) d3 p2 δ(p0 − 2E2 )
Z 2
2 2 1 p0 p2 p2
= (p) dE2 E2 dΩ δ( − E2 ) = 4π
2 2 2 2
4
πp
= (D.40)
2
where the usual tricks have been used to simplify the integrals, using the delta function inside.
Therefore
p2
A = (π − B) (D.41)
4
p4 πp4
(π − B) + p4 B =
4 2
π B π
− +B =
4 4 2
3B π
=
4 4
π
B = (D.42)
3
p2 π
A= (π − )
4 3
p2 2π
= ( )
4 3
πp2
= (D.43)
6
π
Iαβ = gαβ p2 + 2pα pβ . (D.44)
6
Substituting back in eq. (D.31) we have
2πg 4
dΓ = 5 4
pβ1 pα4 (gαβ p2 + 2pα pβ )d3 p4
16 × 6(2π) MW E1 E4
2g 4
dΓ = 4
[(p1 · p4 )p2 + 2(p · p1 )(p · p4 )]d3 p4
16 × 12(2π)4 MW E1 E4
2g 4
dΓ = 4
[(p1 · p4 )p2 + 2(p · p1 )(p · p4 )]d3 p4 (D.45)
192(2π)4 MW E1 E4
For further evaluation we will use the rest frame of the decaying muon. In this frame the four–
momentum are
p1 =(mµ , 0)
p4 =(E4 , p4 )
p =p1 − p4 = (mµ − E4 , −p4 )
p2 =E 2 − p2 = m2µ − 2mµ E4 + (E42 − p24 ) = m2µ + m2e − 2mµ E4 (D.46)
Moreover
p1 · p4 =mµ E4
p · p1 =m2µ − mµ E4
p · p4 =mµ E4 − E42 + p24 = mµ E4 − m2e
p24 = m2e = E42 − p24 ⇒ p24 = E42 − m2e
|p4 | =(E42 − m2e )1/2
d|p4 | 1 2E4 E4
⇒ = 2 2 1/2
=
dE4 2 (E4 − me ) |p4 |
E4
⇒ d|p4 | = dE4
|p4 |
d3 p4 =p24 d|p4 | dΩ = |p4 |E4 dE4 dΩ (D.47)
D.1. MUON DECAY 311
Substituting back in eq. (D.45) we have

2g 4
dΓ = 4 4
|p4 | dE4 dΩ[(m2µ + m2e − 2mµ E4 )mµ E4 + 2(m2µ − mµ E4 )(mµ E4 − m2e )] (D.48)
192(2π) MW mµ
Neglecting electron mass we have |p4 | = E4 , and
2g 4 (4π)
dΓ = 4
E4 dE4 [(m2µ − 2mµ E4 )mµ E4 + 2(m2µ − mµ E4 )mµ E4 ]
192(2π)4 MW mµ
2 × 2g 4
= 3 4
mµ E42 [m2µ − 2mµ E4 + 2m2µ − 2mµ E4 ] dE4
192(2π) MW mµ
4g 4 2
2
= 4
E4 3m µ − 4m µ E4 dE4
192(2π)3 MW

4g 4 2 2 E4
= 4
mµ E4 3 − 4 dE4
192(2π)3 MW mµ

4g 4 m4µ 2E4 2 2E4 mµ 2E4
= 4
3−2 d
192(2π)3 MW 4 mµ mµ 2 mµ
4 5 2
4g mµ 2E4 2E4 2E4
= 4
3−2 d (D.49)
192(2π)3 MW 8 mµ mµ mµ
E4 varies from 0 to E4max can be obtained from (me = 0)
p1 − p4 = p2 + p3 . (D.50)
The square of he factor on the left is
(p1 − p4 )2 =p21 + p24 − 2p1 · p4
=m2µ + m2e − 2p1 · p4 . (D.51)
We have then using eqs. (D.47)(D.50)
2p1 · p4 =m2µ + m2e − (p1 + p4 )2
2mµ E4 =m2µ + m2e − (p2 + p3 )2
≈m2µ − (p2 + p3 )2 . (D.52)
(p2 + p3 )2 is the invariant mass squared of the νµ + ν̄e system, which ranges from 0 to m2µ . For
(p2 + p3 )2 = mµ we have E4min = 0, while for (p2 + p3 )2 = 0 we have E4max = mµ /2. The missing
integration on dΓ is in the variable x such that
2E 2Emax
x= , xmin = 0 , xmax = = 1. (D.53)
mµ mµ
Therefore
Z
4g 4 m5µ 1 2
Γ= 4
x [3 − 2x] dx
192(2π)3 MW 8 0
4g 4 m5µ 1
= 4
192(2π)3 MW 8 2
4 5
g mµ
= 3 4
192π 8MW 4
g4 2
= m5
64MW 192π 3 µ
4
G2 2
= F 3
m5µ
2 192π
G2
= F 3 m5µ (D.54)
192π
From Eq. (D.49), we also have
4g 4
dΓ = 3 4
E42 3m2µ − 4mµ E4 dE4
192(2π) MW
g4 4 2
2
= 4
E 4 3m µ − 4m µ E4 dE4
32MW 6(2π)3
G2F 2 2
2
= E 4 3m µ − 4m µ E4 dE4
3 × 8π 3
G2 4 E4
= F 3 E42 3m2µ 1 − dE4 (D.55)
12π 3 mµ

dΓ G2F 2 2 4E4
= m E 3− (D.56)
dE4 12π 3 µ 4 mµ
D.2. THREE BODY DECAYS IN RADIATIVE SEESAW 313
Without neglect the electron mass we have

2g 4
dΓ = 4
|p4 | dE4 dΩ[(m2µ + m2e − 2mµ E4 )mµ E4 + 2(m2µ − mµ E4 )(mµ E4 − m2e )]
192(2π)4 MW mµ
4g 4
= 3 4
dE4 (E42 − m2e )1/2 [m3µ E4 + m2e mµ E4 − 2(mµ E4 )2
192(2π) MW mµ
+ 2m3µ E4 − 2(mµ E4 )2 − 2m2µ m2e + 2mµ m2e E4 ]
4g 4
dΓ = 3 4
dE4 (E42 − m2e )1/2 [3m3µ E4 + 3m2e mµ E4 − 4(mµ E4 )2 − 2m2µ m2e ] (D.57)
192(2π) MW mµ
The decay width, in terms of x = me /mµ is
Z mµ (1+x2 )/2
4g 4
Γ= 3 4
(E42 − m2e )1/2 [(3m2µ + 3m2e − 4mµ E4 )mµ E4 − 2m2µ m2e ] dE4
192(2π) MW mµ me
4g 4 m6µ
= 4
I (x) , I(x) = 1 − 8x2 − 24x4 ln(x) + 8x6 − x8
192(2π)3 MW mµ 16
G2F m5µ
= I (x)
192π 3
2 5
GF mµ
= √ I (x) , (D.58)
2 96π 3
D.2 three body decays in radiative seesaw

We have the Lagrangian [18]
L =ab hαj N j PL Laα η b + h.c.
=hαj N j PL L1α η 2 − hαj N j PL L2α η 1 + h.c.
=hαj N j PL να η 0 − hαj N j PL lα η + + h.c (D.59)
where
(N j PL lα η + )† = lα† PL γ 0 Nj η − = ¯lα PR γ 0 Nj η − (D.60)
Therefore
L =hαj N j PL να η 0 − hαj N j PL lα η + + h∗αj ν α PR Nj η ∗0 − h∗αj lα PR Nj η −
1
= hαj N j (1 − γ5 )να η 0 − hαj N j (1 − γ5 )lα η + + h∗αj ν α (1 + γ5 )Nj η ∗0 − h∗αj lα (1 + γ5 )Nj η −
2
(D.61)
Figure D.2: Tree level diagram for Nj decay
Applying Feynman rules to the diagram in fig.2 Nj (p1 ) → lα− (p3 )h+ , h+ → lβ+ (p2 ) + Ni (p4 ).
we have the amplitude

1
M = − ihαj ū3 (1 − γ5 )u1 hβi ū4 (1 − γ5 )v2
q 2 − Mη2

1
− ihβj ū3 (1 − γ5 )u1 hαi ū4 (1 − γ5 )v2
q 2 − Mη2
iHαβij
≈− ū3 (1 − γ5 )u1 ū4 (1 − γ5 )v2 (D.62)
Mη2
where
Hαβij = hαj hβi + hαi hβj (D.63)
iHαβij
M∗ = − 2
[ū3 (1 − γ5 )u1 ]† [ū4 (1 − γ5 )v2 ]†
Mη
iHαβij
=− [ū1 (1 + γ5 )u3 ][v̄2 (1 + γ5 )u4 ] . (D.64)
Mη2
Multiplying M and M∗ we have

2
Hαβij
2
|M| = 4
[ūα3 (1 − γ5 )αβ uβ1 ūγ1 (1 + γ5 )γδ uδ3 ][ūα4 (1 − γ5 )αβ v2β v̄2γ (1 + γ5 )γδ uδ4 ]
Mη
2
Hαβij
= 4
[uδ3 ūα3 (1 − γ5 )αβ uβ1 ūγ1 (1 + γ5 )γδ ][uδ4 ūα4 (1 − γ5 )αβ v2β v̄2γ (1 + γ5 )γδ ]
Mη
2
Hαβij
= [(u3 ū3 )δα (1 − γ5 )αβ (u1 ū1 )βγ (1 + γ5 )γδ ][(u4 ū4 )δα (1 − γ5 )αβ (v2 v̄2 )βγ (1 + γ5 )γδ ]
Mη4
2
Hαβij
= Tr[(u3 ū3 )(1 − γ5 )(u1 ū1 )(1 + γ5 )] Tr[(u4 ū4 )(1 − γ5 )(v2 v̄2 )(1 + γ5 )] (D.65)
Mη4
Using eq. (D.26), and neglecting charged fermion masses
2
2Hαβij
|M| = Tr[p 3 (1 − γ5 )(p 1 + Mj )(1 + γ5 )] Tr[(p 4 + Mi )(1 − γ5 )p 2 (1 + γ5 )]
Mη4
2
Hαβij
= LM (D.66)
Mη4
L = Tr[(p 3 − p 3 γ5 )(p 1 + p 1 γ5 + Mj + Mj γ5 ] (D.67)
L = Tr[p 3 p 1 + p 3 p 1 γ5 + Mj p 3 + Mj p 3 γ5 − p 3 γ5 p 1 − p 3 γ5 p 1 γ5 − p 3 γ5 Mj + Mj γ5 ]
=2 Tr[p 3 p 1 ]
=2pα3 pβ1 Tr[γα γβ ]
=8pα3 pβ1 gαβ
=8(p3 · p1 ) (D.68)
Similarly
M = 8(p4 · p2 ) (D.69)
Therefore
2
Hαβij
|M|2 = 64(p3 · p4 )(p1 · p2 )
Mη4
2
Hαβij
|M|2 = 4 × 64(p3 · p4 )(p1 · p2 )
4Mη4
(D.70)
In this way, comparing with eq. (D.30), the results for the moun decay can be directly used after the
replacements
2
g4 Hαβij
4
→
64MW 4Mη4
2
g4 16Hαβij
4
→
MW Mη4
mµ → Mj
me Mi
x= → . (D.71)
mµ Mj
The decay width is according eq. (D.58)
2
16Hαβij 4 Mj6
Γ(Nj → lα∓ lβ± Ni ) = I (x)
Mη4 192(2π)3 Mj 16
(hαj hβi + hαi hβj )2 Mj5
= I (x) (D.72)
2Mη4 192π 3
where
Mi
I(x) = 1 − 8x2 − 24x4 ln(x) + 8x6 − x8 , x= . (D.73)
Mj
Similarly the decay through η 0 is
(hαj hβi + hαi hβj )2 Mj5
Γ(Nj → να νβ Ni ) = I (x) (D.74)
2Mη40 192π 3
In this way, for example for N2
X X h2α2 h2β1 + h2α1 h2β2 + 2hα2 hα1 hβ2 hβ1 M 5
Γ(N2 → lα− lβ+ N1 ) = 4
2
3
I (x)
α α
2M η 192π
h22 h2β1 + h21 h2β2 + 2h2 · h1 hβ2 hβ1 M25
= I (x) (D.75)
2Mη4 192π 3
X h22 h21 + h21 h22 + 2(h2 · h1 )2 M25

Γ(N2 → lα− lβ+ N1 ) = I (x)
αβ
2Mη4 192π 3
h21 h22 + (h1 · h2 )2 M25
= I (x) (D.76)
Mη4 192π 3
In general
X
h2i h2j + (hi · hj )2 Mj5 Mi
Γ(Nj → lα− lβ+ Ni ) = I
αβ
Mη4 192π 3 Mj
X
h2i h2j + (hi · hj )2 Mj5 Mi
Γ(Nj → να νβ Ni ) = I (D.77)
αβ
Mη40 192π 3 Mj
For fix i and j

P − +
αβ Br(Nj → lα lβ Ni ) Mη40
P = 4 (D.78)
αβ Br(Nj → να νβ Ni ) Mη±
while for
P 4
αβ Br(N3 → να νβ N2 )h22 h23 + (h2 · h3 )2 Mη±
P − +
≈ I(M2 /M3 )
αβ Br(N3 → lα lβ N1 ) h21 h23 + (h1 · h3 )2 Mη40
P
Br(N3 → lα− lβ+ N2 ) h22 h23 + (h2 · h3 )2
Pαβ − +
≈ 2 2 I(M2 /M3 ) (D.79)
αβ Br(N3 → lα lβ N1 ) h1 h3 + (h1 · h3 )2
For N2 the total decay width is
" #
M25 M1 1 1
Γtot (N2 ) = h21 h22 + (h1 · h2 )2 I 4
+ 4 (D.80)
192π 3 M2 Mη± Mη0
And the individual branchings through η ± given by eq. (D.72).

For N3 we have several possibilities for signals with charged leptons. The cleanest one is when
N3 decay only through η ± through an intermediate N2 .
The branching of N3 to two charged leptons plus missing energy is either
Br(N3 → lα± lβ∓ N1 ) (D.81)
where the N3 is reconstructed, or
→ lα± lβ∓ N1 ) = Br(N3 → να νβ N2 ) × Br(N2 → lα± lβ∓ N1 )
Br(N3 |{z} (D.82)
η0
that seem to be very difficult to reconstruct. This also seem to be an irreducible background for
Br(N2 → lα± lβ∓ N1 ) (D.83)
To get rid of processes like the one in eq. (D.82) must be Br(N3 → να νβ N2 ) is suppressed. This
happens if
• I(M2 /M 3) 1. In this case the mutilepton signal for N3 is also suppressed. Clearly this
happens for M2 ≈ M3 as I(x) is a sharpest function which controls the kinematical suppression.
We show below for an specific point that even for M3 −M2 ≈ 20 GeV, we can have the Branching
in eq. (D.81) sufficiently large.
• Mη± Mη0
In appendix D.A, it is shown a full set of yukawas consistent with neutrino physics. For this
solution
Br(η + → N3 ) Br(η + → N2 )
≈0.61 ≈0.37
Br(η + → N1 ) Br(η + → N1 )
Br(η + → N1 ) ≈0.51 Br(η + → N2 ) ≈0.19 Br(η + → N3 ) ≈0.30 (D.84)
Below we estimate the branchings to N3 → lα− lβ+ N1 or N3 → να νβ N2 → να νβ lα− lβ+ N1 . For this we
need the Branchings for N2 → lα− lβ+ N1 compared with Branching to N2 → να νβ N1 . In general this is
From this, the visible decays are using eq. (D.78)
P − +
αβ Br(N2 → lα lβ N1 )
X
P ≈0.758 ⇒ Br(N2 → lα− lβ+ N1 ) = 0.431 (D.85)
αβ Br(N 2 → ν α ν β N1 ) αβ
On the other hand the chanels for N3 are N3 → lα− lβ+ N1 , N3 → να νβ N1 , N3 → lα− lβ+ N2 , and N3 →
να νβ N2 . From eqs. (D.78) (D.79)
P P − +
αβ Br(N3 → να νβ N2 ) αβ Br(N3 → lα lβ N2 )
P − +
≈ 0.0812 P − +
≈ 0.0615
αβ Br(N3 → lα lβ N1 ) αβ Br(N3 → lα lβ N1 )
P
Br(N3 → να νβ N1 )
Pαβ − +
≈ 1.320 (D.86)
αβ Br(N3 → lα lβ N1 )
X 1
Br(N3 → lα− lβ+ N1 ) ≈ = 0.406
αβ
1 + 0.0812 + 0.0615 + 1.320
X
Br(N3 → να νβ N1 ) ≈ 0.536
αβ
X
Br(N3 → να νβ N2 ) ≈ 0.030
αβ
X
Br(N3 → lα− lβ+ N2 ) ≈ 0.025 (D.87)
αβ
The expected background for N2,3 → lα− lβ+ N1 is

X X
Br(N3 → να νβ N2 ) × Br(N2 → lα− lβ+ N1 ) ≈ 0.030 × 0.431 = 0.013 (D.88)
αβ αβ
We have that
" #
M25 M1 1 1
Γtot (N2 ) = h21 h22 + (h1 · h2 )2 I +
192π 3 M2 Mη4± Mη40
X
Γvis (N2 → N1 ) ≡ Γ(N2 → lα− lβ+ N1 )
αβ

h21 h22 + (h1 · h2 )2 M25 M1
= I
Mη4± 192π 3 M2
X
Γvis (N3 → N1 ) ≡ Γ(N3 → lα− lβ+ N1 )
αβ

h21 h23 + (h1 · h3 )2 M35 M1
= I
Mη4± 192π 3 M3
X
Γinvis (N3 → N2 ) ≡ Γ(N3 → να νβ N2 )
αβ

h22 h23 + (h2 · h3 )2 M35 M2
= I .
Mη40 192π 3 M3
From above equations we can obtain the following observable:
Brinvis (N3 → N2 ) × Brvis (N2 → N1 )

Brvis (N3 → N1 )

h22 h23 +(h2 ·h3 )2 M35 M2 h21 h22 +(h1 ·h2 )2 M25 M1
4
M 0 192π 3 I M3
× 4
M ± 192π 3 I M2
= η
2 2 5
η
h1 h3 +(h1 ·h3 )2 M3
M 4± 192π 3
I M M3
1
Γtot (N2 )
η

h2 h2 + (h2 · h3 )2 M2 1
= 22 23 I
h1 h3 + (h1 · h3 )2 M3 4 1 1
Mη0 M 4 + M 4
η0 η±

h2 h2 + (h2 · h3 )2 M2 1
= 22 32 2
I
h1 h3 + (h1 · h3 ) M3 M 40
η
1 + M4
η±
D.A Sample point

write(32,*) (h(i,1),i=1,3),(h(i,2),i=1,3),(h(i,3),i=1,3)
-0.00188878597 0.000780236776 0.000248251388

-0.000352494763 -0.000180683976 -0.00122443053
0.000392272581 0.00120920029 -0.0012245638
So that
h21 ≈ 4.238 × 10−6 h22 ≈ 1.656 × 10−6

h23 ≈ 3.116 × 10−6 h1 · h2 ≈ 2.208 × 10−7
h1 · h3 ≈ 1.015 × 10−7 h2 · h3 ≈1.143 × 10−6 (D.89)
h21 h22 + (h1 · h2 )2 ≈7.067 × 10−12 h21 h23 + (h1 · h3 )2 ≈1.321 × 10−11
h22 h23 + (h2 · h3 )2 ≈6.465 × 10−12 (D.90)
The spectrum consistent with neutrino data is
M1 ≈6.16918656 KeV M2 ≈22.8695451 GeV M3 ≈43.126911 GeV

Mη0 ≈139.1382 GeV Mη± ≈149.1382 GeV (D.91)
I(M1 /M3 ) ≈1 I(M2 /M3 ) ≈0.126 (D.92)
D.B Preliminary discussion

One interesting possibility in view of the large invisible direct decay, like N3 → να νβ N1 , is to get the
observables from the missing plus one energetic lepton (coming from η + ) signal. May be decays like
η + → lα+ N3 → lα+ E T
η + → lα+ N2 → lα+ E T
(D.93)
0
Once ηR,I , or η ± are produced the full list of signals is: For η ± production. The decay to Nj is

+ +
3h2αj 1/2 2 2 2
Mj2 + m2α
Γ(η → lα Nj ) = λ Mη , Mj , mα 1 − (D.94)
16πMη Mη2
D.B. PRELIMINARY DISCUSSION 321
X
+
3h2j 1/2 Mj2 + m2α
Γ(η → lα+ Nj ) = λ 2 2 2
Mη , Mj , mα 1 − (D.95)
α
16πMη Mη2
with
h 2 i1/2
λ1/2 Mη2 , Mj2 , m2α = Mη2 + Mj2 − m2α − 4Mη2 Mj2 (D.96)
Neglecting mα with respect to N2,3 , we have for j = 2, 3

" 2 #1/2
Mj2 4Mj2
λ1/2 Mη2 , Mj2 , m2α ≈Mη2 1+ 2 −
Mη Mη2
1/2
2
Mj2 4Mj2
≈Mη 1 + 2 2 −
Mη Mη2
1/2
2
Mj2
≈Mη 1 − 2 2
Mη

2
Mj2
≈Mη 1 − 2 (D.97)
Mη
Therefore

 2 2
X  M
3h2j Mη j
1− 2 j = 2, 3
Γ(η + → lα+ Nj ) ≈ × Mη
16π 

α 1 j=1

2 
 Mj2
3hj Mη 1−2 2 j = 2, 3
≈ × Mη
16π 
1 j=1
(D.98)
In this way
X
Γtor (η + ) = Γ(η + → lα+ Nj )
αj

3Mη 2 2 M22 2 M32
≈ h1 + h2 1 − 2 2 + h3 1 − 2 2 (D.99)
16π Mη Mη
P + +
Br(η + → Nj ) α Γ(η → lα Nj )
= P
Br(η + → Ni ) + +
α Γ(η → lα Ni )
h2j 1 − 2Mj2 /Mη2
≈ 2
hi 1 − 2Mi2 /Mη2
h2j
≈ (1 − 2Mj2 /Mη2 )(1 − 2Mi2 /Mη2 )−1
h2i
h2j
≈ 2 (1 − 2Mj2 /Mη2 )(1 + 2Mi2 /Mη2 )
hi
2
h2j Mj − Mi2
≈ 2 1−2 (D.100)
hi Mη2
For three branchings we should have
a+b+c=1
b c 1
1+ + =
a a a
1
a= (D.101)
1 + b/a + c/a
In this way
1
Br(η + → N1 ) = Br(η + →N3 ) Br(η + →N3 )
(D.102)
1+ Br(η + →N1 )
+ Br(η + →N1 )
From eq.
Br (N3 → N1 )
= (D.103)
→ N2 )
Br(N3 |{z}
η±
D.3. DIAGONALIZATION OF A SYMMETRIC 2 × 2 MATRIX 323
D.3 Diagonalization of a symmetric 2 × 2 matrix

From http://scipp.ucsc.edu/~haber/ph116A/diag2x2_11.pdf
Physics 116A Winter 2011
Diagonalization of a 2 × 2 real symmetric matrix

Consider the most general real symmetric 2 × 2 matrix

a c
A= ,
c b
where a, b and c are arbitrary real numbers. In these notes, we will compute the
eigenvalues and eigenvectors of A, and then find the real orthogonal matrix that
diagonalizes A.
The eigenvalues are the roots of the characteristic equation:

a − λ c
= (a − λ)(b − λ) − c2 = λ2 − λ(a + b) + (ab − c2 ) = 0 .
c b − λ
The two roots, λ1 and λ2 , can be determined from the quadratic formula. Noting
that (a + b)2 − 4(ab − c2 ) = (a − b)2 + 4c2 , the two roots can be written as:
h p i h p i
λ1 = 21 a + b + (a − b)2 + 4c2 and λ2 = 21 a + b − (a − b)2 + 4c2 ,
(1)
where by convention we take λ1 ≥ λ2 .
Since (a − b)2 + 4c2 ≥ 0 (as the sum of two squares must be non-negative),
eq. (1) implies that λ1 and λ2 are real. We next work out the two eigenvectors
and demonstrate that they are orthogonal. It is convenient to define
p
D ≡ (a − b)2 + 4c2 (2)
We first solve the eigenvalue equation,

a c x 1 x
= 2 (a + b + D) ,
c b y y
This yields two equations:
ax + cy = 21 (a + b + D)x ,
cx + by = 21 (a + b + D)y ,
which can be rewritten as:

1
2
(a − b − D)x + cy = 0 , (3)
cx + 21 (b − a − D)y = 0 . (4)
1
One can show that eq. (4) is a multiple of eq. (3) [as it must be since the rank of
the matrix A − λ1 I is one]. Simply multiply eq. (4) by (a − b − D)/(2c) to obtain
1 (a − b − D)(b − a − D)y [D 2 − (a − b)2 ]y

2
(a − b − D)x + = 12 (a − b − D)x + = 0.
4c 4c
Using eq. (2), D 2 − (a − b)2 = 4c2 , and the above equation reduces to
1
2
(a − b − D)x + cy = 0 ,
which is equivalent to eq. (3). Solving for y yields
(b − a + D)x
y= ,
2c
which means that the eigenvector corresponding to eigenvalue λ1 is given by

x x 2c
= .
y 1 2c b − a + D
Since λ2 differs from λ1 by changing the sign of D, it follows without further

computation that the eigenvector corresponding to eigenvalue λ2 is given by

x x 2c
= .
y 2 2c b − a − D
To show that the two eigenvectors are orthogonal, we evaluate the dot product
of (x y)1 and (x y)2, which is equal to x1 x2 + y1 y2 . Inserting the corresponding
vector components, we end up with:
x2 2 x2 2 2 2
x2 2 2

4c + (b − a + D)(b − a − D) = 4c + (a − b) − D = 4c − 4c = 0,
4c2 4c2 c2
after making use of D 2 − (a − b)2 = 4c2 [cf. eq. (2)].
We now propose to find the real orthogonal matrix that diagonalizes A. The
most general 2 × 2 real orthogonal matrix S with determinant equal to 1 must
have the following form:

cos θ − sin θ
S= .
sin θ cos θ
Using this result, we shall determine θ in terms a, b and c such that

−1 λ1 0
S AS = ,
0 λ2
where λ1 and λ2 are the eigenvalues of A obtained in eq. (1). The most straight-
forward approach is to compute S −1 AS explicitly. Since the off-diagonal terms
2
must vanish, one obtains a constraint on the angle θ.

−1 cos θ sin θ a c cos θ − sin θ
S AS =
− sin θ cos θ c b sin θ cos θ

cos θ sin θ a cos θ + c sin θ −a sin θ + c cos θ
=
− sin θ cos θ c cos θ + b sin θ −c sin θ + b cos θ
!
a cos2 θ + 2c cos θ sin θ + b sin2 θ (b − a) cos θ sin θ + c(cos2 θ − sin2 θ)
=
(b − a) cos θ sin θ + c(cos2 θ − sin2 θ) a sin2 θ − 2c cos θ sin θ + b cos2 θ

λ1 0
= . (5)
0 λ2
The vanishing of the off-diagonal elements of S −1 AS implies that:
(b − a) cos θ sin θ + c(cos2 θ − sin2 θ) = 0 .
Using sin 2θ = 2 sin θ cos θ and cos 2θ = cos2 θ − sin2 θ, we can rewrite the above
equation as
1
2
(b − a) sin 2θ + c cos 2θ = 0 .
It follows that:
2c
tan 2θ = (6)
a−b
after writing tan 2θ = sin 2θ/ cos 2θ.
Let us now consider the range of the angle θ. You might think that 0 ≤ θ < 2π.
However, since
cos(θ + π) = − cos θ , and sin(θ + π) = − sin θ ,
it follows that shifting θ → θ + π simply multiplies S by an overall factor of −1.
Thus, S −1 AS is unchanged. Hence, without loss of generality, we may assume
that 0 ≤ θ < π. Unfortunately, eq. (6) does not distinguish between the two
intervals 0 ≤ θ ≤ π/2 or π/2 ≤ θ < π, since tan 2θ = tan(2θ + π) is unchanged if
θ → θ + π/2.
However, we have not yet used all the available information. In particular, the
diagonal elements of eq. (5) also provide some information on the possible values
of θ. Summing the diagonal terms of the matrices in eq. (5) yields:
λ1 + λ2 = (a cos2 θ + 2c cos θ sin θ + b sin2 θ) + (a sin2 θ − 2c cos θ sin θ + b cos2 θ)
= (a + b)(cos2 θ + sin2 θ) = a + b ,
which is independent of θ. This is not surprising since we know that
Tr A = λ1 + λ2 = a + b .
3
However, λ1 − λ2 does depend on θ:
λ1 − λ2 = (a cos2 θ + 2c cos θ sin θ + b sin2 θ) − (a sin2 θ − 2c cos θ sin θ + b cos2 θ)
= (a − b)(cos2 θ − sin2 θ) + 4c sin θ cos θ = (a − b) cos 2θ + 2c sin 2θ . (7)
From eqs. (1) and (7), we obtain
p
λ1 − λ2 = (a − b)2 + 4c2 = (a − b) cos 2θ + 2c sin 2θ . (8)
Using eq. (6) to write:
2c 2c cos 2θ
a−b= = ,
tan 2θ sin 2θ
and inserting this on the left hand side of eq. (8), the latter reduces to:
cos2 2θ 2c 2c
(a−b) cos 2θ+2c sin 2θ = 2c +2c sin 2θ = cos2 2θ + sin2 2θ = .
sin 2θ sin 2θ sin 2θ
Substituting this result back into eq. (8) and solving for sin 2θ, we find:
2c
sin 2θ = p (9)
(a − b)2 + 4c2
We can also obtain cos 2θ using eqs. (6) and (9):
a−b
cos 2θ = p (10)
(a − b)2 + 4c2
Eq. (9) tells us in which quadrant θ lives. If 0 < θ < 21 π, then sin 2θ > 0, which
implies that c > 0. If 21 π < θ < π, then sin 2θ < 0, which implies that c < 0.
Thus, the sign of c determines the quadrant of θ. Eq. (10) provides additional
information. For c > 0, the sign of a − b determines whether 0 < θ < 41 π or
1
4
π < θ < 12 π. The former corresponds to a − b > 0 while the latter corresponds
to a − b < 0. Likewise, if c < 0, the sign of a − b determines whether 21 π < θ < 34 π
or 43 π < θ < π. The former corresponds to a − b < 0 while the latter corresponds
to a − b > 0. The borderline cases are likewise determined:
c = 0 and a > b =⇒ θ = 0,
c = 0 and a < b =⇒ θ = 12 π ,
a = b and c > 0 =⇒ θ = 14 π ,
a = b and c < 0 =⇒ θ = 34 π .
If c = 0 and a = b, then A = I and it follows that S −1 AS = S −1 S = I, which is
satisfied for any invertible matrix S. Consequently, in this limit θ is undefined.
4
From the eigenvalue equations (1) we can easily check that in addition to trace invariance, we
have that
q
λ1 − λ2 = (a − b)2 − 4c2 . (D.104)
Replacing back in eq. (9)

2c
sin 2θ = . (D.105)
λ1 − λ2
Bibliography
[1] Diego Restrepo, “El Lagrangiano del Modelo Estándar.”, DOI:10.5281/zenodo.3960741 [Google
Scholar]
[2] M. Maggiore, “A Modern introduction to quantum field theory,” Oxford University Press, 2005.
(Oxford Series in Physics, 12. ISBN 0 19 852073 5)
[3] F. Mandl and G. Shaw, Chichester, Uk: Wiley ( 1984) 354 P. ( A Wiley-interscience Publica-
tion)
[4] A. Lahiri and P. B. Pal, “A first book of quantum field theory,” Harrow, UK: Alpha Sci. Int.
(2005) 380 p
[5] P. B. Pal, “Representation-independent manipulations with Dirac spinors,”

[arXiv:physics/0703214 [physics.ed-ph]].
[6] Pal, P. “Representation-independent manipulations with Dirac matrices and spinors”,

arXiv:physics/0703214v2 [physics.ed-ph]
[7] P. Langacker, “The standard model and beyond,”
[8] A. J. Larkoski, “Elementary Particle Physics: An Intuitive Introduction,”
[9] Wikipedia article: http://en.wikipedia.org/wiki/Particle_decay
[10] M. Klasen, Int. J. Mod. Phys. C 14 (2003) 1273 [arXiv:hep-ph/0210426].
[11] Electroweak Interactions, Peter Renton
[12] Michael E. Peskin and Daniel V. Schroeder. An introduction to Quantum Field Theory, Addison-
Wesley Publishing Company(1995), p. 101
[13] ibdg 1, p. 105
329
330 BIBLIOGRAPHY
[14] Francis Halzen and Alan D. Martin. Quarks & Leptons: An introductory Course in Modern
Particle Physics, John Wiley & Sons(1984), p. 89
[15] Chris Quigg. Gauge theory of the Strong, Weak and Electromagnetic Interactions, Westview
Press(1997), p. 110
[16] ibdg 4, p. 130.
[17] ibdg 5, p. 325.
[18] D. Aristizabal Sierra, J. Kubo, D. Restrepo, D. Suematsu and O. Zapata, “Radiative seesaw:
Warm dark matter, collider and lepton flavour violating signals,” Phys. Rev. D 79 (2009) 013011
[arXiv:0808.3340 [hep-ph]].
[19] Relativistic Quantum Mechanics and Field Theory, Franz Gross, John Wiley & Sons, INC. 1993
[20] D. McMahon, “Quantum field theory demystified: A self-teaching guide,” SPIRES entry New
York, USA: McGraw-Hill (2009) 299 p
[21] R. P. Feynman, “QED. The Strange Theory Of Light And Matter,” Princeton, Usa: Univ. Pr.
( 1985) 158 P. ( Alix G. Mautner Memorial Lectures)
[22] A. Semenov, “LanHEP - a package for the automatic generation of Feynman rules in field
theory. Version 3.0,” Comput. Phys. Commun. 180 (2009) 431 [arXiv:0805.0555 [hep-ph]], http:
//feynrules.phys.ucl.ac.be/.
[23] N. D. Christensen and C. Duhr, “FeynRules - Feynman rules made easy,” Comput. Phys. Com-
mun. 180, 1614 (2009) [arXiv:0806.4194 [hep-ph]], http://feynrules.phys.ucl.ac.be/.
[24] Christianto V., Smarandache F., and Lichtenberg F., “A Note of Extended Proca Equations
and Superconductivity,” Progress in Physics, 1 (2009) 40 [viXra:1003.0054]
[25] Martin, Many body quantum field theory
[26] H. K. Dreiner, H. E. Haber and S. P. Martin, “Two-component spinor techniques and

Feynman rules for quantum field theory and supersymmetry,” Phys. Rept. 494 (2010) 1
doi:10.1016/j.physrep.2010.05.002 [arXiv:0812.1594 [hep-ph]].
[27] M. Roncadelli and D. Wyler, “Naturally Light Dirac Neutrinos in Gauge Theories,” Phys. Lett.
133B (1983) 325. doi:10.1016/0370-2693(83)90156-9
BIBLIOGRAPHY 331
[28] D. Suematsu, “Dark matter stability and one-loop neutrino mass generation based on Pec-
cei–Quinn symmetry,” Eur. Phys. J. C 78 (2018) no.1, 33 doi:10.1140/epjc/s10052-018-5519-4
[arXiv:1709.02886 [hep-ph]].

Detailed Lecture Notes in Quantum Field Theory

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Detailed Lecture Notes in Quantum Field Theory

Uploaded by

Copyright:

Available Formats

Detailed lecture notes in Quantum Field

1 Classical Field Theory 3

1.11.5 Propiedades de las matrices de Dirac . . . . . . . . . . . . . . . . . . . . . . . 57

2 Second quantization 121

3 Quantization of the electromagnetic field 163

5 Feynman Rules 195

6 Neutrino masses 257

8 One-loop neutrino masses 267

A Computational QFT 281

C Two body decays 295

D Three body decays 301

Classical Field Theory

1.1 Lagrangian Formulation

δφ(x) = φ0 (x) − φ(x) (1.3)

De otro lado, con δx = x0 − x, la expansión de Taylor para f (x + δx) es

Para L, tenemos de la ec. (1.3)

Donde hemos aplicado el Teorema de Gauss

las ecuaciones de Euler-Lagrange (1.9)

1.1.2 Teorema de Noether para simetrı́as internas

Usando el mismo procedimiento, se obtiene

1.1.3 Teorema de Noether para simetrı́as externas

φ0 (x0 ) = φ0 (x + δa) (1.27)

Para una traslación, ∆φ(x) = 0, ver figura 1.1. De modo que

δφ = −(∂µ φ)δaµ , (1.32)

y la transformación del campo φ como consecuencia de la traslación es

φ(x) → φ0 (x) = φ(x) + δφ(x) = φ(x) − (∂µ φ(x))δaµ . (1.33)

Si aµ es constante (un análisis más general es hecho en [?])

Figure 1.1: Traslación de función y coordenadas en una dimensión: φ(x) = φ0 (x0 )

Generalizando a un campo complejo

1.2 Global gauge invariance

Que puede escribirse como  

δψ = ψ 0 − ψ = (eiθ − 1)ψ ≈ iθψ (1.55)

Que es un resultado bien conocido de la mecánica cuántica.

la condición de normalización en ec. (1.61) implica que |A|2 = 1/L3 , y

1.3 Local phase invariance in the Scrödinger’s Lagrangian

Yang-Mills, Physical Review, 1954

Xµ → Xµ0 = Xµ − i∂µ θ(x) . (1.81)

Replacing back in Eq. (1.80) we have

It is convenient to redefine Xµ in terms of Aµ :

such that the covariant derivative can be conveniently written as

iqAµ →iqA0µ = iqAµ − i∂µ θ(x)

1.4 Notación relativista

donde µ, ν = 0, 1, 2, 3, i = 1, 2, 3 y se asume suma sobre ı́ndices repetidos. Además

Finalmente la métrica usada se define como

donde {gµν } denota la forma matricial del tensor gµν .

• Example: Lorentz invariance

Como un ejemplo de Transformación de Lorentz considere un desplazamiento a lo largo del eje x

Denotaremos los cuadrivectores con ı́ndices arriba como

aµ = (a0 , a1 , a2 , a3 ) = (a0 , a) (1.109)

Entonces el correspondiente cuadrivector con ı́ndices abajo, usando la ec. (1.91), es

aµ = (a0 , a1 , a2 , a3 ) = (a0 , −a1 , −a2 , −a3 ) = (a0 , −a). (1.110)

1.4.1 Ejemplos de cuadrivectores

xµ =(x0 , x1 , x2 , x3 ) = (t, x, y, z) = (t, x) (1.111)

De la relatividad especial tenemos que

Por lo tanto, ya que v 2 = v2 = |v|2

El invariante de Lorentz asociado a pµ corresponde a la ecuación de momento energı́a una vez se

Aµ = (A0 , A) = (φ, A) (1.117)

Las derivada covariante en la ec, (1.85) es en términos de componentes:

1.4.2 Lorentz tranformation for fields

φ(x) → φ0 (x0 ) = φ(x) . (1.134)

By using eq. (1.133) we have

φ0 (x0 ) = φ(Λ−1 x0 ) . (1.135)

φ(x) → φ0 (x) = φ(Λ−1 x) . (1.136)

and the defintion of the Lorentz transformation itself:

Que puede escribirse como