Professional Documents
Culture Documents
Detailed Lecture Notes in Quantum Field Theory
Detailed Lecture Notes in Quantum Field Theory
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
v
vi CONTENTS
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
7 Leptogenesis 265
viii CONTENTS
9 Incubator 279
B Neutrinos 293
B.1 Weinberg operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
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
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.
donde d4 x = dt dx dy dz. Considere primero una variación sólo de los campos, tal que (x = xµ )
∂f
f (x + δx) = f (x) + δx + · · · (1.4)
∂x
3
4 CHAPTER 1. CLASSICAL FIELD THEORY
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
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. LAGRANGIAN FORMULATION 7
µ
xµ → x0 =xµ + δaµ
δxµ =δaµ (1.26)
tenemos
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
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)
8 CHAPTER 1. CLASSICAL FIELD THEORY
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
1.1. LAGRANGIAN FORMULATION 9
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)
10 CHAPTER 1. CLASSICAL FIELD THEORY
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
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
12 CHAPTER 1. CLASSICAL FIELD THEORY
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
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)
14 CHAPTER 1. CLASSICAL FIELD THEORY
Ejercicio: De la ec. (1.72) obtenega la densidad Hamiltoniana, y usando la ec. (1.41) encontrar la densidad
Lagrangiana (1.49).
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.
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
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)
1
Aµ ≡ Xµ , (1.84)
iq
Dµ = ∂µ + iqAµ . (1.85)
The transformation properties of Aµ can be obtained from the Xµ transformation in eq. (1.81):
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.
aν ≡ gµν aµ (1.91)
aν bν = gµν aµ bν = a0 b0 − a · b (1.93)
18 CHAPTER 1. CLASSICAL FIELD THEORY
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)
1.4. NOTACIÓN RELATIVISTA 19
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)
20 CHAPTER 1. CLASSICAL FIELD THEORY
Con esta notación, el producto escalar de cuadrivectores puede expresarse como el producto escalar
de los dos vectores de cuatro componente aµ y aµ .
E =γm
p =γmv . (1.113)
E 2 − p2 = γ 2 m2 (1 − v 2 ) = m2 . (1.114)
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
1.4. NOTACIÓN RELATIVISTA 21
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 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
22 CHAPTER 1. CLASSICAL FIELD THEORY
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
Therefore, for an arbitrary space-time point we have that the scalar field transforms under a Lorentz
transformation as
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
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)
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 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)
1.5. VECTOR FIELD LAGRANGIAN 27
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
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
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 )
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)
∂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,
∂µ J µ = 0 . (1.166)
∂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)
∂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)
T00 = −F 0λ (∂0 Aλ ) − L
1
= −F 0λ (∂0 Aλ ) + F µν Fµν + J µ Aµ
4
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
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)
32 CHAPTER 1. CLASSICAL FIELD THEORY
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
1 1
H = E2 + B2 + ∇ · (A0 E) . (1.180)
2 2
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)
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 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
34 CHAPTER 1. CLASSICAL FIELD THEORY
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
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. SCRÖDINGER EQUATION IN PRESENCE OF THE ELECTROMAGNETIC FIELD 35
If we define
D ≡ ∇ − iqA . (1.190)
we have in components:
D i = ∂i − iqAi
D i = ∂i + iqAi . (1.191)
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.
36 CHAPTER 1. CLASSICAL FIELD THEORY
• Ejemplo:
Demuestre que la ec. (1.196) es covariante:
Como
ψ → ψ 0 = eiθ(x) ψ (1.197)
Entonces
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)
38 CHAPTER 1. CLASSICAL FIELD THEORY
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.
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.
∂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
40 CHAPTER 1. CLASSICAL FIELD THEORY
∂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)
• 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
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)
∂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. (??),
42 CHAPTER 1. CLASSICAL FIELD THEORY
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.
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)
1.9. KLEIN-GORDON EQUATION 45
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 1 1 1
L = [∂ µ φ1 ∂µ φ1 − m2 φ21 ] + [∂ µ φ2 ∂µ φ2 − m2 φ22 ] (1.239)
2 2 2 2
46 CHAPTER 1. CLASSICAL FIELD THEORY
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
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)
( + m2 )φ = 0, (1.245)
1.9. KLEIN-GORDON EQUATION 47
( + 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)
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)
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
where S(Λ) is some spinorial representation of the Lorentz Group. We will check in next section if
a Action with a term like
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
∂
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
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)
50 CHAPTER 1. CLASSICAL FIELD THEORY
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)
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
1.11. DIRAC’S ACTION 51
b† = b
b2 = I
b㵆 b = 㵠(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
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)
52 CHAPTER 1. CLASSICAL FIELD THEORY
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)
∂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)
1.11. DIRAC’S ACTION 53
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)
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
1.11. DIRAC’S ACTION 55
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)
X 2 2
X
(A · θ)2 = Ai θi + Ai , Aj θi θj (1.302)
i i<j
56 CHAPTER 1. CLASSICAL FIELD THEORY
• 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
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)
{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)
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
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)
1.11. DIRAC’S ACTION 59
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)
L = iψ † γ 0 γ 0 ∂0 ψ + iψ † γ 0 γ i ∂i ψ = iψ † ∂0 ψ + iψ † αi ∂i ψ , (1.326)
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)
The Pauli matrices are set of matrices satisfying this commutation relations:
i j
τ τ τk
, = i ijk (1.330)
2 2 2
60 CHAPTER 1. CLASSICAL FIELD THEORY
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)
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)
[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)!
62 CHAPTER 1. CLASSICAL FIELD THEORY
1
Λ = 1 + ξi bi + θi ijk rjk , (1.349)
2
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
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
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)
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)
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
1.11. DIRAC’S ACTION 65
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
i
γ5 = µναβ γ µ γ ν γ α γ β (1.365)
24
and
ψ → ψ 0 = e−iθ(x)Q ψ
ψ̄ → ψ̄ 0 = ψ̄eiθ(x)Q , (1.367)
∂ 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
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
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)
j µ = −eQψ̄γ µ ψ. (1.374)
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).
68 CHAPTER 1. CLASSICAL FIELD THEORY
Figure 1.3: Electromagnetic repulsion. The diagrams (a) and (b) are summarized in the diagram (c)
1.13. CROMODINÁMICA CUÁNTICA 69
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− )
70 CHAPTER 1. CLASSICAL FIELD THEORY
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
1.13. CROMODINÁMICA CUÁNTICA 71
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
72 CHAPTER 1. CLASSICAL FIELD THEORY
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
Por consiguiente
µ
D0 U = U Dµ (1.394)
Dµ → (Dµ )0 = U Dµ U −1 (1.395)
Desarrollando a ambos lados
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
1.13. CROMODINÁMICA CUÁNTICA 73
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)
74 CHAPTER 1. CLASSICAL FIELD THEORY
De modo que
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
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
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)
1.13. CROMODINÁMICA CUÁNTICA 75
donde
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
1.13. CROMODINÁMICA CUÁNTICA 77
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
78 CHAPTER 1. CLASSICAL FIELD THEORY
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.
Sigiendo los mismos procedimientos anteriores debemos llegar a los siguientes resultados. Para el
campo Ψ
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)
80 CHAPTER 1. CLASSICAL FIELD THEORY
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. CROMODINÁMICA CUÁNTICA 81
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)
con
Λa , Λb = ifabc Λc . (1.435)
En componentes
Entonces las las Ecuaciones de Euler Lagrange para Gaµν , en (1.423) se pueden escribir como
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
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):
A · (B × C) = B · (C × A) = C · (A × B) , (1.445)
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 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
donde
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 .
84 CHAPTER 1. CLASSICAL FIELD THEORY
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
no puede interpretarse como un término de masa en el Lagrangiano dado por la ec. (1.451).
Consideremos ahora el potencial
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)
86 CHAPTER 1. CLASSICAL FIELD THEORY
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)
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)
1.14. SPONTANEOUS SYMMETRY BREAKING 87
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)
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.
88 CHAPTER 1. CLASSICAL FIELD THEORY
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)
[. . . ] 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.
Goldstone’s theorem states that for every spontaneously broken continuous symmetry, the
theory must contain a massless particle.
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
1.14. SPONTANEOUS SYMMETRY BREAKING 89
En el caso de la Acción invariante gauge local bajo el Grupo U (1), tenemos el Lagrangiano (??):
∗ 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
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
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.
90 CHAPTER 1. CLASSICAL FIELD THEORY
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
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)
Usando la ec. (1.321)
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)
92 CHAPTER 1. CLASSICAL FIELD THEORY
Entonces
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)
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
donde,
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
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
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
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)
1.16. STANDARD MODEL LAGRANGIAN 97
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
se reduce a
1 1
V (H) = µ2 (H + v)2 + λ(H + v)4 . (1.509)
2 4
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,
98 CHAPTER 1. CLASSICAL FIELD THEORY
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
1.16. STANDARD MODEL LAGRANGIAN 99
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)
100 CHAPTER 1. CLASSICAL FIELD THEORY
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
1.16. STANDARD MODEL LAGRANGIAN 101
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
102 CHAPTER 1. CLASSICAL FIELD THEORY
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)
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
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
Usando la ec. (1.533)
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)
104 CHAPTER 1. CLASSICAL FIELD THEORY
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:
1.16. STANDARD MODEL LAGRANGIAN 105
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
•
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
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-
106 CHAPTER 1. CLASSICAL FIELD THEORY
1.16.4 Self-interactions
El Lagrangiano gauge
µν
Lgauge = − 14 Gµν a 1 i 1 µν
a Gµν − 4 Wi Wµν − 4 B Bµν
(1.554)
1.16. STANDARD MODEL LAGRANGIAN 107
donde
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
− 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
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. STANDARD MODEL LAGRANGIAN 109
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
• Los bosones gauge cargados Wµ± decaen siempre a leptones de la misma generación y con la
misma intensidad.
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
110 CHAPTER 1. CLASSICAL FIELD THEORY
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)
1.16. STANDARD MODEL LAGRANGIAN 111
donde MD D D D
diag = diag(m1 , m2 , m3 ). Similarmente
2
D† †
VR MD MD VRD
= MD
diag . (1.567)
=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)
112 CHAPTER 1. CLASSICAL FIELD THEORY
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).
1.16. STANDARD MODEL LAGRANGIAN 113
Teniendo en cuenta estos resultados podemos escribir finalmente el Lagrangiano completo del
Modelo Estándar en la Gauge Unitario, para
X
LSM = if¯ (γ µ ∂µ − mf ) f
f
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
− 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 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.
114 CHAPTER 1. CLASSICAL FIELD THEORY
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)
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 [?]
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 → e+ e− ) 3.363(4)
= ≈ 0.222 (1.582)
Br(Z → b̄b) 15.12(5)
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)
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.
1.17. FENOMENOLOGÍA ELECTRODÉBIL 117
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.
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
118 CHAPTER 1. CLASSICAL FIELD THEORY
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
120 CHAPTER 1. CLASSICAL FIELD THEORY
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
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 ωn is positive-definite:
p
ωn = +|v| |kn | (2.9)
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)
we define
1
φn (z, t) = √ e−i(ωn t−kn z) , (2.16)
L
When n = m
Z L Z L
1
dz φ∗n (z, t)φm (z, t) = dz
0 L 0
=1 (2.18)
2.2. QUANTIZATION OF THE NONRELATIVISTIC STRING 125
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
2.2. QUANTIZATION OF THE NONRELATIVISTIC STRING 127
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.
128 CHAPTER 2. SECOND QUANTIZATION
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)
Comparing eq. (2.42) with Eq. (2.37) we see that the complex number an can be written as (~ = 1)
√
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
130 CHAPTER 2. SECOND QUANTIZATION
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
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=−∞
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 |0i =0
b
a †n =0 ,
h0|b (2.62)
such that
h0|0i = 1 (2.63)
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)
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)
2.2. QUANTIZATION OF THE NONRELATIVISTIC STRING 133
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!
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)
134 CHAPTER 2. SECOND QUANTIZATION
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:
1+c=m+1 (2.75)
and
bn |mn i = mn |mn i
N (2.76)
bnb
N an |mn i =(mn − 1)b
an |mn i (2.78)
an |mi = C− |mn − 1i
b (2.80)
2.2. QUANTIZATION OF THE NONRELATIVISTIC STRING 135
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.
∂ µ ∂µ φ = 0 , (2.89)
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
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
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)
(∂ µ ∂µ + m2 )φ = 0 , (2.98)
138 CHAPTER 2. SECOND QUANTIZATION
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
X 3 Z
L
→ d3 p (2.102)
n
2π
From
Z
d3 pδ (3) (p − q) =1 (2.103)
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
2.3. QUANTIZATION OF THE KLEIN-GORDON FIELD 139
In this way
3
(3) L
δ (p − q) = δn,m , (2.107)
2π
V
δ 3 (0) = . (2.110)
(2π)3
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)
140 CHAPTER 2. SECOND QUANTIZATION
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)
2.3. QUANTIZATION OF THE KLEIN-GORDON FIELD 141
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)
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
142 CHAPTER 2. SECOND QUANTIZATION
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
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)
with properties
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
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
144 CHAPTER 2. SECOND QUANTIZATION
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)
Y
k
1 m
|m1 , . . . , mi , . . . , mk i = √ âj j |0, . . . , 0k i (2.137)
j=1
mj !
As an example,
Further:
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
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
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)
146 CHAPTER 2. SECOND QUANTIZATION
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)
148 CHAPTER 2. SECOND QUANTIZATION
The factors (2Ep1 )1/2 are a convenient choice of normalization. In particular, the one-particle states
are
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
2.6 Propagator
With conventions
b
φ(x) = φb+ (x) + φb− (x) (2.159)
where
Z
1
φb+ (x) = d3 p p ap e−ip·x
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
where
Z
1
φb+ (x) = d3 p p ap e−ip·x
b
3
(2π) 2Ep
Z
1
φb− (x) = d3 p p bb† eip·x .
p (2.162)
3
(2π) 2Ep
where
ψ = ψ†γ 0 (2.165)
{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)
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)
2.7. QUANTIZATION OF FERMIONS 151
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
since the belong to differente eigenvalues. Note that the two us (0) and the two vs (0) are degenerate.
We define
where the munis sign in η−s is just a convention. We define the normalized eigenvectors ξ and η such
that
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
Choosing
1
Np = √ (2.186)
E+m
2.7. QUANTIZATION OF FERMIONS 153
we obtain
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
Λ− Λ+ = Λ+ Λ− =0
Λ− + Λ+ =1 .
We define
Σ = σ 23 σ 31 σ 12
Σ·p
Σp =
|p|
1 ± Σp
Π± (p) = ,
2
where
Moreover
[Λ± , Π± ] =0 ,
and
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)
156 CHAPTER 2. SECOND QUANTIZATION
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
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
2.7. QUANTIZATION OF FERMIONS 157
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
158 CHAPTER 2. SECOND QUANTIZATION
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
2.7. QUANTIZATION OF FERMIONS 159
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
(2.214)
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
160 CHAPTER 2. SECOND QUANTIZATION
and therefore
In the following we need an explicit form for the internal Lorentz transformations
where:
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
2.7. QUANTIZATION OF FERMIONS 161
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
One additional identity is the one relating the two-component spinor momentum space wave
function to the four-component spinor wave functions
162 CHAPTER 2. SECOND QUANTIZATION
Chapter 3
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)
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)
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
∂µ Aµ = 0 , (3.17)
∇ · 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)
∂i ∂j
A (r , t), π(r, t) = i δij − 2 δ 3 (r − r0 ) ≡ δ 3T (r − r0 ) .
j 0
(3.20)
∇
166 CHAPTER 3. QUANTIZATION OF THE ELECTROMAGNETIC FIELD
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
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
S–matrix
We will use the S–matrix formulation to obtain the decay rates and cross section formulas.
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
since
∂
i |a, ti =i(−i)HS eHS (t−ti ) |a, ti i
∂t
=HS |a, ti . (4.4)
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)
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)
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
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
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 .
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
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)
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
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)
Similarly we regularize also the time interval, saying that the time runs from −T /2 to T /2 so that
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
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
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
176 CHAPTER 4. S–MATRIX
A
σ
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
178 CHAPTER 4. S–MATRIX
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. CROSS SECTION 179
As
q
|p01 | = E10 2 − m1 2 (4.70)
d|p01 | 2E10
= q
dE10
2 E10 2 − m1 2
E10
= (4.71)
|p01 |
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
we have
2 2
p01 =p02
2 2 2 2
E10 − m01 =E20 − m02 , (4.76)
4.4. CROSS SECTION 181
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)
p1 = −p2 , (4.79)
then
So that
√
s = E1 + E2 (4.81)
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 )|
182 CHAPTER 4. S–MATRIX
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
4.4. CROSS SECTION 183
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
we have
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
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)
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.
By using eq. (4.39) we have
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
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
(2π)4
dΓ = |Mf i |2 dΦ(n) (p; k1 , k2 , . . . , kn ) (4.117)
2Ep
p =0 ⇒ M = Ep (4.118)
4.5. DECAY RATES 187
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
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 .
188 CHAPTER 4. S–MATRIX
d|p1 | 2E1
= p
dE1 2 E1 2 − m1 2
E1
= (4.126)
|p1 |
In this way, we can write, in general
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)
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)
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) .
4.5. DECAY RATES 189
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
190 CHAPTER 4. S–MATRIX
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
Therefore
(p2e + m2 )v 2 =p2e
pe
v =p
p2e+ m2e
import math as np
M=939.57 #MeV
mp=938.28 #MeV
192 CHAPTER 4. S–MATRIX
me=0.511 #MeV
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
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
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)
194 CHAPTER 4. S–MATRIX
if a = b
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 ??.
195
196 CHAPTER 5. FEYNMAN RULES
Given the state |a, tiS in the SP, in the Heisenberg picture (HP) we defined the state
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
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)
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
along with2 ,
and
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.
198 CHAPTER 5. FEYNMAN RULES
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.
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
5.1. INTERACTION PICTURE 199
|Φ(t)i according to Eq. (5.17) is a unitary transformation. Accordingly it preserves the normalization
of states
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
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
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)
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)
−∞ −∞ −∞
In the limit t → ∞
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)
−∞
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)
−∞ −∞ −∞ −∞
In the limit t → ∞
|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)
−∞ −∞ −∞
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].
=δ 4 (−p0i + pj − p0l + pm )
=δ 4 (pj + pm − p0i − p0l ). (5.44)
5.4. YUKAWA INTERACTION 203
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
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
and
204 CHAPTER 5. FEYNMAN RULES
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
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
h1(ψR )† , 1ψL , 0φ |(ψR )†− ψL − φ+ |0(ψR )† , 0ψL , 1φ i ∝ h0(ψR )† , 0ψL , 0φ |0(ψR )† , 0ψL , 0φ i = 1 . (5.55)
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
206 CHAPTER 5. FEYNMAN RULES
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
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
5.4. YUKAWA INTERACTION 207
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)
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
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)
" #
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)
5.4. YUKAWA INTERACTION 209
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)
210 CHAPTER 5. FEYNMAN RULES
M =E + E 0
|p| = |p0 |. (5.82)
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
5.4. YUKAWA INTERACTION 211
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
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
212 CHAPTER 5. FEYNMAN RULES
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
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
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)
−∞ −∞
( (
0 t1 > t2 0 t2 > t1
θ(t2 − t1 ) = , θ(t1 − t2 ) = . (5.94)
1 t1 < t2 1 t2 < t1
214 CHAPTER 5. FEYNMAN RULES
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 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
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
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)
216 CHAPTER 5. FEYNMAN RULES
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)
Generalizing the results for n scalar or fermion fields, but with an even number of fermions fields,
we have the Wick theorem
2
s = (p1 + p2 )2 = (p01 + p02 )
2 2
t = (p1 − p01 ) = (p2 − p02 )
2 2
u = (p1 − p02 ) = (p01 − p2 ) . (5.113)
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!
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!
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)
228 CHAPTER 5. FEYNMAN RULES
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
†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
†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
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)
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
− y α (p01 )y β (p02 )xα (p2 )xβ (p1 ) eip1 ·x1 eip2 ·x2 e−ip2 ·x1 e−ip1 ·x2
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)
230 CHAPTER 5. FEYNMAN RULES
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)
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
x2 x2
e− (p2) e− (p′1) e− (p2) e− (p′2)
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)
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
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:
and therefore
2
dσ y4 m1 m2
= . (5.163)
dΩ 32π 2 m4h m1 + m2
[AB, CD] = A[B, C]D − [A, C]BD + CA[B, D] − C[A, D]B , (5.165)
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
Consider now
1
| (eL )† (p2 )eL (p1 )i = √ a†s1 (p1 )b†s2 (p2 )|0i (5.172)
V2
236 CHAPTER 5. FEYNMAN RULES
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
√ xα (s, k)y †α̇ (s0 , k0 ) e−ik·x1 e−ik ·x2
(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
√ xα (s, k)y †α̇ (s0 , k0 ) e−ik·x1 e−ik ·x2
(2π) 2Ek V (2π) 2Ek0 V
0 †
× as (k)bs0 (k ), as1 (p1 )b†s2 (p2 ) |0i. (5.173)
[AB, CD] = A[B, C]D − [A, C]BD + CA[B, D] − C[A, D]B , (5.174)
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
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)
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)
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 )
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)
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)
240 CHAPTER 5. FEYNMAN RULES
→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
e L + qL → e L + qL (5.202)
eL →ξe qL → ξq . (5.203)
eL eL
qL qL
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)
In this way
By using
−igµν
iDµν = (5.208)
t − m2Z
5.7. SCATTERING 243
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 )
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!
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
1
|qL (p2 )S(p1 )i = √ a† (p1 )a†s2 (p2 )|0i. (5.225)
V2
246 CHAPTER 5. FEYNMAN RULES
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)
By using the identity
[AB, CD] = A[B, C]D − [A, C]BD + CA[B, D] − C[A, D]B , (5.227)
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
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)
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)
where
Comparing with the general expression for the 2 → 2 processes, we get that
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
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
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
250 CHAPTER 5. FEYNMAN RULES
eL + qL → eR + qR , (5.253)
X X
|M|2 =ye yq ∆F (t) x1 y10 x2 y20 y20† x†2 y10† x†1
spins spins
By using
−i
i∆F (t) = (5.259)
t − m2h
5.8. DIRAC FERMION CASE 251
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
Γ = γ µ (1 − γ5 ) (5.265)
2
For p MW the analysis is similar to the previous one with
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
e− (p1) x1 νe−(p′2)
W µ (p1 − p′2 )
νµ−(p2) x2 µ− (p′1)
The center of mass (CM) frame is defined by the condition in Eq. (??):
p1 + p2 = 0 (5.274)
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)
254 CHAPTER 5. FEYNMAN RULES
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
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.
256 CHAPTER 5. FEYNMAN RULES
Chapter 6
Neutrino masses
yij
L= Li · HLj · H + h.c
Λ
yij
= ab cd Lai H b Lcj H d + h.c .
Λ
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
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)
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)
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
1
Mχ ∓ ≈ MR 1 ∓ 1 + 2MD MR−2 MDT (6.10)
2
260 CHAPTER 6. NEUTRINO MASSES
Therefore
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.
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.
H S
L FR FR νR
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
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
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)
264 CHAPTER 6. NEUTRINO MASSES
Chapter 7
Leptogenesis
265
266 CHAPTER 7. LEPTOGENESIS
Chapter 8
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
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
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
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
270 CHAPTER 8. ONE-LOOP NEUTRINO MASSES
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]
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
m2χn − m2Sα
m2Sα ln m2Sα − m2χn ln m2χn
= cte(∞) + . (8.27)
m2χn − m2Sα
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
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)
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
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)
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
274 CHAPTER 8. ONE-LOOP NEUTRINO MASSES
where
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
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
∂V 2 0
0 3
= − µ 3 2 Re H1 + λ6 2 Re H1 + ··· (8.46)
∂ Re (H20 )
we have
1 1
µ21 + λ1 =0 −µ3 + λ6 =0 (8.48)
2 2
276 CHAPTER 8. ONE-LOOP NEUTRINO MASSES
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
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
Therefore
ckk fjk sin ϕ
Mijν ∝ (Π02 )ik cos ϕM (8.57)
8.4. SINGLET-DOUBLET FERMIONS WITH SCALAR SINGLETS 277
Therefore
where
cos ϕ − sin ϕ
R= (8.60)
sin ϕ cos ϕ
√ 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)
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:
$ 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
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/
func1.mdl lgrng1.mdl prtcls1.mdl vars1.mdl
func2.mdl lgrng2.mdl prtcls2.mdl vars2.mdl
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
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
19 particle_type
20 particle/Antiparticle: (’name’, property name=VALUE, ...).
21
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.
A.3. LANHEP/CALCHEP 285
prtcprop pdg.
vector
A/A: (photon, gauge).
spinor e1:(electron),
e2:(muon, mass Mm = 0.1057),
e3:(’tau-lepton’, mass Mt = 1.777).
SetTexName([EE=e]).
• prtcls1.tex:
• vars1.tex
A.3. LANHEP/CALCHEP 287
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
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:
• 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:
where
MDdia = V MD V † (A.4)
Investing the equation A.4 and replacing in (A.2) we can write in the interaction eigenstates:
Expanding we get . . . which is just the expresion in the Standard Model file
292 APPENDIX A. COMPUTATIONAL QFT
Appendix B
Neutrinos
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
One would need an isotriplet scalar field ∆ ∼ (3, 2), which either elemental or composite. The term
293
294 APPENDIX B. NEUTRINOS
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
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.
(2π)4
dΓn = |M|2 dΦ(n) (P ; p1 , p2 , . . . , pn ) (C.1)
2M
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
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
√ 1/2 †
h = mf GF 2 in the standard models Now, having into account that γ 0 = γ 0
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
298 APPENDIX C. TWO BODY DECAYS
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
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
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
300 APPENDIX C. TWO BODY DECAYS
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
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)
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
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
−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
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)
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
Similarly
X
M µν = 2p4δ p2 Tr γ δ γ µ γ γ ν (1 − γ5 ) (D.29)
s
D.1. MUON DECAY 307
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
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
310 APPENDIX D. THREE BODY DECAYS
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
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
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
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
iHαβij
M∗ = − 2
[ū3 (1 − γ5 )u1 ]† [ū4 (1 − γ5 )v2 ]†
Mη
iHαβij
=− [ū1 (1 + γ5 )u3 ][v̄2 (1 + γ5 )u4 ] . (D.64)
Mη2
D.2. THREE BODY DECAYS IN RADIATIVE SEESAW 315
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)
316 APPENDIX D. THREE BODY DECAYS
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
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
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
318 APPENDIX D. THREE BODY DECAYS
• 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
D.2. THREE BODY DECAYS IN RADIATIVE SEESAW 319
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
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±
320 APPENDIX D. THREE BODY DECAYS
So that
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)
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)
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
322 APPENDIX D. THREE BODY DECAYS
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
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
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)
ax + cy = 21 (a + b + D)x ,
cx + by = 21 (a + b + D)y ,
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
(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
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
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
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
328 APPENDIX D. THREE BODY DECAYS
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)
[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
[12] Michael E. Peskin and Daniel V. Schroeder. An introduction to Quantum Field Theory, Addison-
Wesley Publishing Company(1995), p. 101
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
[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]
[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]].