taller de altas energías 2023

eth zurich

The Standard Model of particle Physics

Taller de Altas Energias 2023

Adrian Carmona Bermudez

September 2023
Slide 1/141
The Standard Model of Particle Physics
The Standard Model in a Nutshell
• The Standard Model describes incredibly well how Nature works to
very short distances ∼ 10−19 m (or equivalently, to very high
energies)

• The discovery of the Higgs boson at CERN in 2012 confirmed the

existence of its last missing piece (and lead to the Nobel prize)

• However, we know that there has to be something else since it does

not explain e.g.

• Dark Matter

• The matter-antimatter asymmetry of the Universe

• Why gravity is so weak

• Neutrino masses

• Why fermion masses are so different

see Javi’s lectures!

References:
Textbooks:
• Paul Langacker (2017). The Standard Model and Beyond. Taylor &
Francis. doi: 10.1201/b22175
• Matthew D. Schwartz (2013). Quantum Field Theory and the
Standard Model. Cambridge University Press. doi:
10.1017/9781139540940
• John F. Donoghue, Eugene Golowich, and Barry R. Holstein (2014).
Dynamics of the Standard Model : Second edition. Cambridge
University Press. doi: 10.1017/CBO9780511803512
• Pierre Ramond (1999). Journeys beyond the standard model.
Perseus Books
• Ta-Pei Cheng and Ling-Fong Li (1984). Gauge Theory of
Elementary Particle Physics. Oxford, UK: Oxford University Press
• S. Pokorski (2000). Gauge Field Theores. Cambridge University
Press. doi: 10.1017/CBO9780511612343
• E. Leader and E. Predazzi (1996). An Introduction to gauge
theories and modern particle physics. Vol. 1. Cambridge University
Press. doi: 10.1017/CBO9780511622595
References:
More Textbooks:
• R. Keith Ellis, W. James Stirling, and B. R. Webber (1996). QCD
and collider physics. Cambridge University Press. doi:
10.1017/CBO9780511628788
• Guido Altarelli (2017). Collider Physics within the Standard Model:
A Primer. Ed. by James Wells. Vol. 937. Springer. doi:
10.1007/978-3-319-51920-3
• Gustavo C. Branco, Luis Lavoura, and Joao P. Silva (1999). CP
Violation. Clarendon Press

Lectures:
• José Ignacio Illana and Alejandro Jimenez Cano (2022). “Quantum
field theory and the structure of the Standard Model”. In: doi:
10.22323/1.406.0314. arXiv: 2211.14636 [hep-ph]
• Antonio Pich (2012). “The Standard Model of Electroweak
Interactions”. In: arXiv: 1201.0537 [hep-ph]
• A. Pich (1995). “Chiral perturbation theory”. In: doi:
10.1088/0034-4885/58/6/001. arXiv: hep-ph/9502366
References:

Lectures:
• G. P. Salam (2020). “Elements of QCD for hadron colliders”. In:
CERN Yellow Rep. School Proc. 5. Ed. by M. Mulders and
J. Trân Thanh Vân, pp. 1–56. doi: 10.23730/CYRSP-2020-005.1
• Peter Skands (2013). “Introduction to QCD”. In: Theoretical
Advanced Study Institute in Elementary Particle Physics: Searching
for New Physics at Small and Large Scales, pp. 341–420. doi:
10.1142/9789814525220_0008. arXiv: 1207.2389 [hep-ph]

AND MANY MORE!

What is the SM?
[Glashow 1961; A. Salam 1968; S. Weinberg 1967]

The Standard Model is so complex it would be

hard to put it on a T-shirt — though not
impossible; you’d just have to write kind of small

S. Weinberg to Quanta Magazine

What is the SM?

The Standard Model is so complex it would be

hard to put it on a T-shirt — though not
impossible; you’d just have to write kind of small

S. Weinberg to Quanta Magazine

What is the SM?

The SM is a local quantum field theory (QFT) defined by

1 its (gauge) symmetry group
2 its matter content
3 the condition of renormalizability
together with a specific pattern of spontaneous symmetry breaking.

We will come back to all these ingredients one by one.

Global AND gauge symmetries
Global AND gauge symmetries
SPACE-TIME symmetries
Relativistic fields in QFT span infinite-dimensional vector spaces
transforming under irreps of the Poincaré group (Lorentz group +
translations) ISO(1, 3)

1 𝑘𝑙𝑚 𝑙𝑚
[𝑃 𝜇 , 𝑃 𝜈 ] = 0, [𝑃 𝜇 , 𝐽 𝜌𝜎 ] = 𝑖(𝑔𝜇𝜌 𝑃 𝜎 − 𝑔𝜇𝜎 𝑃 𝜌 ), 𝐽𝑘 = 𝜖 𝐽
2
[𝐽 𝜇𝜈 , 𝐽 𝜌𝜎 ] = 𝑖(𝑔𝜈𝜌 𝐽 𝜇𝜎 − 𝑔𝜇𝜌 𝐽 𝜈𝜎 − 𝑔𝜈𝜎 𝐽 𝜇𝜌 + 𝑔𝜇𝜎 𝐽 𝜈𝜌 ), 𝐾 = 𝐽 0𝑘 = −𝐽 𝑘0
𝑘

It is known that finite-dimensional representations of a simple

non-compact Lie group are not unitary. This is what happens e.g. for the
Poincaré group, whose unitary representations are infinite-dimensional.

The Lorentz group is locally isomorphic to 𝑆𝑈 (2) ⊗ 𝑆𝑈 (2) (same

algebra), whose irreps can be labelled by (𝑗− , 𝑗+ ), 𝑗± ∈ {0, 1/2, 1, …}
SPACE-TIME symmetries
Once quantized, these fields will act on the Fock space of multiparticle
states. One can consider the Hilbert subspace of one-particle states
invariant under Poincaré transformations, and study the resulting irreps

|0⟩ (vacuum), 𝑎†p,𝑠 |0⟩ (1 particle), †

𝑏p,𝑠 |0⟩ (1 antiparticle)

These irreps need to be unitary, to make the scalar products invariant

under a change of reference system

⟨𝜓1 |𝜓2 ⟩ = ⟨𝜓1 |ℛ† ℛ|𝜓2 ⟩.

This implies in particular that such irreps will be infinite-dimensional.

We can classify the one-state particles by the Casimir operators

1
𝑚2 ≡ 𝑃𝜇 𝑃 𝜇 𝑊𝜇 𝑊 𝜇 , 𝑊 𝜇 = − 𝜖𝜇𝜈𝜌𝜎 𝐽𝜈𝜌 𝑃𝜎 .
2
since their eigenvalues will label the different ’irreps’.
SPACE-TIME symmetries

Case 𝑚 ≠ 0, we choose 𝑃 𝜇 = (𝑚, 0, 0, 0), which leads to

𝑚 𝑖𝑗𝑘0 𝑗𝑘 𝑚
𝑊 0 = 0, 𝑊 𝑖 = − 𝜖 𝐽 = 𝜖𝑖𝑗𝑘 𝐽 𝑗𝑘 = 𝑚𝐽 𝑖 ⇒ 𝑊𝜇 𝑊 𝜇 = −𝑚2 𝑗(𝑗+1)
2 2
The irreps are labeled by 𝑚 and 𝑗 and the vectors by |𝑗3 = −𝑗, … , 𝑗⟩
Massive particles of spin 𝑗 have 2𝑗 + 1 dof and 𝑆𝑈 (2) is the little group

Case 𝑚 = 0, we choose 𝑃 𝜇 = (𝜔, 0, 0, 𝜔), which leads to

−𝑊 0 = 𝑊 3 = 𝜔𝐽 3 , 𝑊 1,2 = 𝜔(𝐽 1 ± 𝐾 2 )
⇒ 𝑊𝜇 𝑊 = −𝜔2 [(𝐽 1 + 𝐾 2 )2 + (𝐽 2 − 𝐾 1 )2 ]
𝜇

Now the little group is 𝑆𝑂(2) and the irreps are unidimensional and are
labeled by the helicity ℎ ∈ {0, ±1/2, ±1, …}.
SPACE-TIME symmetries

If we want to build a QFT with a masless vector field 𝑉𝜇 = 4 (= 1 ⊕ 3

under the rotation group 𝑆𝑂(3)), we need to be sure that just propagate
transverse polarizations.

Indeed polarizations related by a multiple of 𝑝𝜇 are related by a Poincaré

transformation and should be considered equivalent

𝜖𝜇 → 𝜖𝜇 + 𝛼𝑝𝜇

This looks very similar to

𝐴𝜇 → 𝐴 𝜇 + 𝜕 𝜇 𝛼

Gauge redundancy! It is a redundancy of our QFT but not a symmetry.

There is no conserved charge.
Global symmetries
For the sake of concreteness, let us consider a free-fermion theory
̄ 𝜕
− 𝑚)Ψ,
ℒ = 𝜓(𝑖 𝜕
≡ 𝛾 𝜇 𝜕𝜇 , 𝜓 = 𝜓† 𝛾 0

This Lagrangian density and the resulting action are invariants under a
transformation
𝜓(𝑥) → 𝑒−𝑖𝑞𝜃 𝜓(𝑥), 𝑞, 𝜃 ∈ ℝ

Noether theorem tell us that there is a conserved current

𝑗𝜇 = 𝑞 𝜓𝛾̄ 𝜇 𝑞, 𝜕𝜇 𝑗 𝜇 = 0

as well as a Noether charge [𝑄, 𝐻] = 0

d3 𝑝
𝑄 = 𝑞 ∫ d3 𝑥 ∶ 𝜓𝛾̄ 0 𝜓 ∶ = 𝑞 ∫ ∑ (𝑎†p,𝑠 𝑎p,𝑠 − 𝑏p,𝑠
†
𝑏p,𝑠 )
(2𝜋)3 𝑠=1,2

𝑄𝑎†k,𝑠 |0⟩ = +𝑞𝑎†k,𝑠 |0⟩ (particle), †

𝑄𝑏k,𝑠 †
|0⟩ = −𝑞𝑏k,𝑠 |0⟩ (antiparticle),
Global symmetries

In general, global symmetries will be described by 𝑁 -dimensional

compact Lie groups
𝑎 𝑎
𝑔 ∈ 𝐺, 𝑔(𝜃) = 𝑒−𝑖𝑇 𝜃
𝜃𝑎 ∈ ℝ, 𝑎 = 1, … , 𝑁 ,

where 𝑇 𝑎 are the generators, satisfying the Lie algebra

1 𝑎𝑏
[𝑇 𝑎 , 𝑇 𝑏 ] = 𝑖𝑓 𝑎𝑏𝑐 𝑇 𝑐 , Tr(𝑇 𝑎 ⋅ 𝑇 𝑏 ) = 𝛿 .
2
They will have associated a conserved current and fields will transform

𝜓1
⎛ 𝜓2 ⎞
𝜓(𝑥) → 𝑈 (𝜃)𝜓(𝑥) = exp(−𝑖𝑇 𝑎 𝜃𝑎 )𝜓(𝑥), 𝜓(𝑥) = ⎜
⎜
⎜ ⋮ ⎟
⎟
⎟
𝜓
⎝ 𝑑⎠

where 𝑇 𝑎 is represented by an hermitian 𝑑-dimensional square matrix.

The global symmetries of the SM
The SM has the following global symmetry

𝑆𝑈 (3)𝐶 ⊗ 𝑆𝑈 (2)𝐿 ⊗ 𝑈 (1)𝑌

- 𝑈 (1): 1 generator (𝑞), one-dimensional irreps only

- 𝑆𝑈 (2): 3 generators, 𝑓 𝑎𝑏𝑐 = 𝜖𝑎𝑏𝑐 , Levi-Civita symbol

• Fundamental representation (𝑑 = 2), 2, 𝑇 𝑎 = 𝜎𝑎 /2, 𝑎 = 1, 2, 3
• Adjoint representation (𝑑 = 𝑁 = 3), 3, (𝑇 𝑎 )𝑏𝑐 = −𝑖𝑓 𝑎𝑏𝑐

Note that the fundamental is pseudo-real since if 𝜙 ∼ 2, 𝑖𝜎2 𝜙∗ ∼ 2

- 𝑆𝑈 (3): 8 generators
√
3 147
𝑓 123 = 1, 𝑓 458 = 𝑓 678 = ,𝑓 = 𝑓 156 = 𝑓 246 = 𝑓 247 = 𝑓 345 = −𝑓 367
2
• Fundamental representation (𝑑 = 3), 3, 𝑇 𝑎 = 𝜆𝑎 /2, 𝑎 = 1, … , 8
• Adjoint representation (𝑑 = 𝑁 = 8), 8, (𝑇 𝑎 )𝑏𝑐 = −𝑖𝑓 𝑎𝑏𝑐
What about gauge 'symmetries'?
One can turn any global symmetry of the theory into something local by
correcting the difference from point 𝑥 to point 𝑦 with a connection.
When we talk about gauge theories we just mean that the connection is
indeed physical.
In practice,

𝜃𝑎 = 𝜃𝑎 (𝑥) ⇒ 𝑈 (𝜃) = exp (−𝑖𝑇 𝑎 𝜃𝑎 (𝑥)), 𝜓(𝑥) → exp (−𝑖𝑇 𝑎 𝜃𝑎 (𝑥))𝜓(𝑥)

The covariant derivative of the field

𝐷𝜇 𝜓(𝑥) = (𝜕𝜇 − 𝑖𝑔𝑇 𝑎 𝐴𝑎𝜇 (𝑥))𝜓(𝑥) = (𝜕𝜇 − 𝑖𝑔𝐴𝜇̂ (𝑥))𝜓(𝑥)

needs to transform as the field, i.e., 𝐷𝜇 𝜙(𝑥) → 𝑈 𝐷𝜇 𝜙(𝑥). This is true if

𝑖 1
𝐴𝜇̂ (𝑥) → 𝑈 𝐴𝜇̂ (𝑥)𝑈 † − (𝜕𝜇 𝑈 )𝑈 † , 𝐴𝑎𝜇 → 𝐴𝑎𝜇 − 𝑓 𝑎𝑏𝑐 𝐴𝑏𝜇 𝜃𝑐 − 𝜕𝜇 𝜃𝑎 .
𝑔 𝑔

Then, 𝜓𝑖̄ 𝐷𝜓 is invariant under a local transformation 𝑈 .

Dynamics of the gauge fields

Since we said that the connection is physical, we should add kinetic

terms for them. This is the so-called Yang-Mills Lagrangian [Yang and
Mills 1954]
1 ̂ 𝐴𝜇𝜈̂ ) = − 1 𝐴𝑎 𝐴𝑎𝜇𝜈
ℒYM = − Tr(𝐴𝜇𝜈
2 4 𝜇𝜈
where
̂ = 𝐴𝑎 𝑇 𝑎 = 𝐷 𝐴 ̂ − 𝐷 𝐴 ̂ = 𝜕 𝐴 ̂ − 𝜕 𝐴 ̂ − 𝑖𝑔[𝐴 ̂ , 𝐴 ̂ ]
𝐴𝜇𝜈 𝜇𝜈 𝜇 𝜈 𝜈 𝜇 𝜇 𝜈 𝜈 𝜇 𝜇 𝜈
⇒ 𝐴𝑎𝜇𝜈 = 𝜕𝜇 𝐴𝑎𝜈 − 𝜕𝜈 𝐴𝑎𝜇 + 𝑔𝑓 𝑎𝑏𝑐 𝐴𝑏𝜇 𝐴𝑐𝜈

Then, we can write the first terms of our SM t-shirt

1 1 𝑖 1
ℒgauge = − 𝐺𝑎𝜇𝜈 𝐺𝑎𝜇𝜈 − 𝑊𝜇𝜈 𝑊 𝑖𝜇𝜈 − 𝐵𝜇𝜈 𝐵𝜇𝜈
4 4 4
where 𝐺𝑎𝜇 , 𝑊𝜇𝑖 , 𝐵𝜇 with 𝑎 = 1, … , 8, 𝑖 = 1, 2, 3, are the gauge bosons of
𝑆𝑈 (3)𝐶 , 𝑆𝑈 (2)𝐿 and 𝑈 (1)𝑌 , respectively.
Dynamics of the gauge fields
The Yang-Mills Lagrangian includes cubic and quartic self interactions
1
ℒkin = − (𝜕𝜇 𝐴𝑎𝜈 − 𝜕𝜈 𝐴𝑎𝜇 )(𝜕 𝜇 𝐴𝑎𝜈 − 𝜕 𝜈 𝐴𝑎𝜇 )
4
1
ℒcubic = − 𝑔𝑓 𝑎𝑏𝑐 (𝜕𝜇 𝐴𝑎𝜈 − 𝜕𝜈 𝐴𝑎𝜇 )𝐴𝑏𝜇 𝜇 𝐴
𝑐𝜈
2
1
ℒquartic = − 𝑔2 𝑓 𝑎𝑏𝑒 𝑓 𝑐𝑑𝑒 𝐴𝑎𝜇 𝐴𝑏𝜈 𝐴𝑐𝜇 𝐴𝑑𝜈
4
Dynamics of the gauge fields
Is this everything? What happens with terms like the one below?

𝑔𝑠2 ̃𝑎𝜇𝜈 , ̃𝑎𝜇𝜈 = 1 𝜖𝜇𝜈𝜌𝜎 𝐺𝑎

ℒ𝜃 = 𝜃 𝐺𝑎 𝐺 𝐺
32𝜋2 𝜇𝜈 2 𝜌𝜎

The short answer is that it depends. One can write down such a term
like the derivative of a current
𝑔𝑠2 ̃𝑎𝜇𝜈 = 𝜕 𝒦𝜇
ℒ𝜃 = 𝜃 𝐺𝑎 𝐺
32𝜋2 𝜇𝜈 𝜇

When this is the case, due to the generalized Stokes theorem

𝑆𝜃 = ∫ d4 𝑥 ℒ𝜃 = ∮ 𝑑𝑆 𝒦𝜇 𝑛𝜇
ℳ4 𝜕ℳ4

If 𝒦𝜇 goes to zero fast enough when go to the infinite, there is no

contribution and we can forget about such term. We will come back to
this later (strong CP problem).
Quantizing a gauge theory

Within the path-integral formalism, Green functions are obtained by

derivating the generating functional

𝒵[𝐽 ] = 𝒩 ∫ 𝒟𝜙 exp {𝑖 ∫ d4 𝑥 [ℒ(𝜙) + 𝐽 𝜙]}, 𝒵[0] = 1

(one should define 𝒵[𝐽 ] in the Euclidean space and then continue it
analytically to the Minkowsky space). In particular

1 𝛿 𝛿 𝛿
𝐺(𝑁) (𝑥1 , 𝑥2 , … , 𝑥𝑁 ) = ⋯ 𝒵[𝐽 ]∣
𝑖𝑁 𝛿𝐽 (𝑥1 ) 𝛿𝐽 (𝑥2 ) 𝛿𝐽 (𝑥𝑁 ) 𝐽=0

Then, cross-sections are obtained through the LSZ theorem [Lehmann,

Symanzik, and Zimmermann 1955].
Quantizing a gauge theory

In gauge theories, redundancy makes everything more complicated.

Indeed, since 𝑆YM [𝐴𝜇̂ ] = 𝑆YM [𝐴𝑈
̂
𝜇 ], where

̂
𝐴𝑈 ̂ † 𝑖 †
𝜇 = 𝑈 𝐴𝜇 𝑈 − (𝜕𝜇 𝑈 )𝑈
𝑔

Therefore, in evaluating

𝒵[𝐽 ] = 𝒩 ∫ 𝒟𝐴𝜇 exp(𝑖𝑆YM [𝐴𝜇̂ ] + 𝑖 ∫ 𝑑4 𝑥 𝐽𝜇𝑎 𝐴𝑎𝜇 )

we would like to integrate over just one representative 𝐴𝑈

𝜇 of each
equivalence class defined by gauge transformations, where

∫ 𝒟𝐴𝜇 = ∫ 𝒟𝑈 ∫ 𝒟𝐴𝑈
𝜇.
Quantizing a gauge theory

Therefore, we fix a particular gauge

𝐹 [𝐴𝜇 ] = 0

with 𝐹 [𝐴𝜇 ] any functional of 𝐴𝜇 , like e.g. 𝐹 [𝐴𝜇 ] = 𝜕𝜇 𝐴𝜇 . Then

𝒟𝐴𝑈 𝑈
𝜇 = 𝒟𝐴𝜇 𝛿(𝐴𝜇 ∼ 𝐴𝜇 ) = 𝒟𝐴𝜇 𝛿[𝐹 [𝐴𝜇 ]]det𝑀

with
𝛿𝐹 [𝐴𝜇 (𝑥)]
𝑀 (𝑥, 𝑦) = ∣ ,
𝛿𝑈 (𝑦) 𝐹 =0
analogously to
𝜕𝑓
𝛿(𝑥 − 𝑥0 ) = 𝛿(𝑓(𝑥)) ∣ ∣ .
𝜕𝑥 𝑓(𝑥0 )=0
Quantizing a gauge theory
If we consider gauge-fixing conditions of the form

𝐹 [𝐴𝜇 ] − 𝐶(𝑥) = 0

with 𝐶 an arbitrary, independent function of 𝐴𝜇 , det𝑀 becomes

independent of 𝐶(𝑥) and we can write

∫ 𝒟𝐴𝑈
𝜇 = ∫ 𝒟𝐴𝜇 𝛿[𝐹 [𝐴𝜇 ] − 𝐶(𝑥)] det𝑀

= 𝒩 ∫ 𝒟𝐴𝜇 det𝑀 ∫ 𝒟𝐶 𝛿[𝐹 [𝐴𝜇 ] − 𝐶(𝑥)]𝐺[𝐶]

= 𝒩 ∫ 𝒟𝐴𝜇 det𝑀 𝐺[𝐹 [𝐴𝜇 ]]

with 𝐺[𝐶] some arbitrary functional like

𝑖
𝐺[𝐶] = exp{ − ∫ 𝑑4 𝑥 𝐶 2 (𝑥)}.
𝜉
Quantizing a gauge theory

Therefore, we can avoid summing over equivalent gauge configurations

by replacing
𝑖
∫ 𝒟𝐴𝜇 → ∫ 𝒟𝐴𝜇 det𝑀 exp{ − ∫ 𝑑4 𝑥 𝐹 2 [𝐴𝜇 ]}.
𝜉

This expression can be further simplified by taking into account that

det𝑀 = ∫ 𝒟𝜂 ̄ 𝒟𝜂 exp{ − 𝑖 ∫ d4 𝑥 d4 𝑦𝜂(𝑥)𝑀

̄ (𝑥, 𝑦)𝜂(𝑦)},

𝛿𝐹 [𝐴𝜇 (𝑥)]
𝑀 (𝑥, 𝑦) = ∣ ∣
𝛿𝑈 (𝑦) 𝐹 =0

where 𝜂 and 𝜂 ̄ are Grassmann variables.

Quantizing a gauge theory
This leads to [Faddeev and Popov 1967]

𝒵[𝐽 ] = 𝒩′ ∫ 𝒟𝐴𝜇 𝒟𝜂 ̄ 𝒟𝜂 exp{𝑖 ∫ 𝑑4 𝑥[ℒtotal + 𝐽 𝜇 𝐴𝜇 ]},

1 1
ℒtotal = ℒYM + ℒGF + ℒFP = − 𝐴𝑎𝜇𝜈 𝐴𝑎𝜇𝜈 − 𝐹 2 [𝐴𝜇 ] − 𝜂𝑀
̄ 𝜂,
4 2𝜉

where 𝜂 and 𝜂 ̄ are ghosts, anti-conmuting scalar fields. They are

unphysical but need to be included in loops. For the Lorentz gauge,
1
𝐹 [𝐴𝑎𝜇 ] = 𝜕𝜇 𝐴𝑎𝜇 ⇒ ℒGF = − ∑ (𝜕 𝜇 𝐴𝑎𝜇 )2
𝑎
2𝜉𝑎

and
1
𝐹 [𝐴𝑎𝜇 ] ↦ 𝐹 [𝐴𝑎𝜇 ] − 𝑓 𝑎𝑏𝑐 𝜕 𝜇 (𝐴𝑏𝜇 𝜃𝑐 ) − □𝜃𝑎
𝑔
𝑎 𝑎
where we have used that under a gauge transformation 𝑈 = 𝑒−𝑖𝑇 𝜃 (𝑥)
,
1
𝐴𝑎𝜇 ↦ 𝐴𝑎𝜇 − 𝑓 𝑎𝑏𝑐 𝐴𝑏𝜇 𝜃𝑐 − 𝜕𝜇 𝜃𝑎 .
𝑔
Quantizing a gauge theory

In this case, the Faddeev-Popov determinant reads

𝛿𝐹 [𝐴𝑎𝜇 ] 1 1
𝑀𝑎𝑏 = 𝑏
= 𝑓 𝑎𝑏𝑐 (𝜕 𝜇 𝐴𝑐𝜇 + 𝐴𝑐𝜇 𝜕 𝜇 ) − 𝛿 𝑎𝑏 □ = − (𝛿 𝑎𝑏 □ − 𝑔𝑓 𝑎𝑏𝑐 𝐴𝑐𝜇 𝜕 𝜇 ).
𝛿𝜃 𝑔 𝑔

We can get rid of the 1/𝑔 factor by redefining 𝜂 and 𝜂,̄ leading to

det𝑀 = ∫ 𝒟𝜂 ̄ 𝒟𝜂 exp{ − 𝑖 ∫ d4 𝑥 𝜂𝑎̄ (𝛿 𝑎𝑏 □ − 𝑔𝑓 𝑎𝑏𝑐 𝐴𝑐𝜇 𝜕 𝜇 )𝜂𝑏 }

IBP
= ∫ 𝒟𝜂 ̄ 𝒟𝜂 exp{𝑖 ∫ d4 𝑥 [𝜕 𝜇 𝜂𝑎̄ 𝜕𝜇 𝜂𝑎 − 𝑔𝑓 𝑎𝑏𝑐 (𝜕 𝜇 𝜂𝑎̄ )𝜂𝑏 𝐴𝑐𝜇 ]}

Then

ℒFP = (𝜕 𝜇 𝜂𝑎̄ )(𝜕𝜇 𝜂𝑎 − 𝑔𝑓 𝑎𝑏𝑐 𝜂𝑏 𝐴𝑐𝜇 ) = (𝜕 𝜇 𝜂𝑎̄ )(𝐷𝜇Adj )𝑎𝑏 𝜂𝑏 .

Quantizing a gauge theory
The full Lagrangian for a Yang-Mills theory

ℒtot = ℒYM + ℒGF + ℒFP

is invariant under the following BRS transformations [Becchi, Rouet, and

Stora 1976] (with 𝜃 a Grassmann variable, 𝜃2 = 0)

1 1 1 𝜈 𝑎 1 𝑎𝑏𝑐 𝑏 𝑐
𝛿𝐴𝑎𝜇 = − 𝜃(𝐷𝜇Adj )𝑎𝑏 𝜂𝑏 , 𝛿 𝜂𝑎̄ = 𝜃 𝜕 𝐴𝜈 , 𝛿𝜂𝑎 = 𝜃𝑓 𝜂 𝜂 ,
𝑔 𝑔 𝜉 2

where 𝜑 → 𝜑 + 𝛿𝜑 = 𝜑 + 𝜃Δ(𝜑), 𝜑 = 𝐴𝑎𝜇 , 𝜂𝑎 , 𝜂𝑎̄ . BRS invariance of the

effective action Γ (generation functional of all 1PI diagrams – see later)

𝛿Γ 𝛿Δ(𝑆)
Δ(Γ) = 0 = ∫ 𝑑4 𝑥 [ 𝛿𝜑 + …] ⇒ ∣ =0 ST identities.
𝛿𝜑 𝛿𝜑𝑗 ⋯
𝜑=0

Slavnov-Taylor (ST) [Slavnov 1972; Taylor 1971] identities are relations

between vertex functions. They summarize all Ward identities and are
essential for renormalization (BRS do not change in 𝐷 = 4 − 2𝜀).
The full SM gauge Lagrangian
Then, the complete SM gauge Lagrangian reads (calling 𝑐𝑔𝑎 , 𝑐𝑤
𝑖
, 𝑐𝑏 the FP
ghost fields of 𝑆𝑈 (3)𝐶 , 𝑆𝑈 (2)𝐿 and 𝑈 (1)𝑌 , respectively)
1 𝑎 𝑎𝜇𝜈 1 𝑖 1
ℒtot
𝐺 = ℒgauge + ℒGF + ℒFP = − 𝐺𝜇𝜈 𝐺 − 𝑊𝜇𝜈 𝑊 𝑖𝜇𝜈 − 𝐵𝜇𝜈 𝐵𝜇𝜈
4 4 4
1 𝜇 𝑎 2 1 𝜇 𝑖 2 1 𝜇 2
− (𝜕 𝐺𝜇 ) − (𝜕 𝑊𝜇 ) − (𝜕 𝐵𝜇 )
2𝜉𝑎 2𝜉𝑖 2𝜉
+ (𝜕 𝜇 𝑐𝑔𝑎̄ )(𝜕𝜇 𝑐𝑔𝑎 − 𝑔𝑠 𝑓 𝑎𝑏𝑐 𝑐𝑔𝑏 𝐺𝑐𝜇 ) + (𝜕 𝜇 𝑐𝑤
𝑖 𝑖
̄ )(𝜕𝜇 𝑐𝑤 − 𝑔𝜖𝑖𝑗𝑘 𝑐𝑤
𝑖
𝑊𝜇𝑘 )

+ 𝜕 𝜇 𝑐𝑏̄ 𝜕𝜇 𝑐𝑏
Matter content
Matter content
The SM building blocks

The matter of the SM is made of chiral fermions, transforming under

𝑆𝑈 (3)𝐶 ⊗ 𝑆𝑈 (2)𝐿 ⊗ 𝑈 (1)𝑌

Multiplet Quantum Numbers I II III 𝑄

𝑢𝐿 𝑐𝐿 𝑡𝐿 +2/3
𝑖
𝑞𝐿 (3, 2, + 16 ) ( ) ( ) ( )
𝑑𝐿 𝑠𝐿 𝑏𝐿 −1/3
𝑢𝑖𝑅 (3, 1, + 32 ) 𝑢𝑅 𝑐𝑅 𝑡𝑅 +2/3
𝑖
𝑑𝑅 (3, 1, − 31 ) 𝑑𝑅 𝑠𝑅 𝑏𝑅 −1/3
𝜈𝑒 𝜈𝜇 𝜈𝜏 0
𝑖
ℓ𝐿 (1, 2, − 12 ) ( 𝐿) ( 𝐿) ( 𝐿)
𝑒𝐿 𝜇𝐿 𝜏𝐿 −1
𝑒𝑖𝑅 (1, 1, −1) 𝑒𝑅 𝜇𝑅 𝜏𝑅 −1

As we will see later, 𝑄 = 𝑇𝐿3 + 𝑌

The SM building blocks
The SM matter content
With these ingredients we can now write
𝑖 𝑖 𝑖̄ 𝑖̄
ℒferm = 𝑞𝐿
̄ 𝑖𝐷𝑞𝐿 + 𝑢̄𝑖𝑅 𝑖𝐷𝑢𝑖𝑅 + 𝑑𝑅 𝑖
𝑖𝐷𝑑𝑅 + ℓ𝐿 𝑖
𝑖𝐷ℓ𝐿 + 𝑒𝑖𝑅
̄ 𝑖𝐷𝑒𝑖𝑅

This Lagrangian is invariant under unitary rotations of the fields

𝒢ferm = 𝑈 (3)5 = 𝑈 (1)5 ⊗ 𝒢𝑞 ⊗ 𝒢𝑙

where

𝒢𝑞 = 𝑆𝑈 (3)𝑞𝐿 ⊗ 𝑆𝑈 (3)𝑢𝑅 ⊗ 𝑆𝑈 (3)𝑑𝑅 𝒢𝑙 = 𝑆𝑈 (3)ℓ𝐿 ⊗ 𝑆𝑈 (3)𝑒𝑅

This flavor group will be broken by the Higgs interactions (see later)
Mass terms are fordbidden by the global symmetries of the SM. E.g.,

−𝑚𝑞𝐿̄ 𝑢𝑅 + h.c.

is not invariant under 𝑆𝑈 (3)𝐶 ⊗ 𝑆𝑈 (2)𝐿 ⊗ 𝑈 (1)𝑌 .

The SM matter content

The covariant derivatives include the fermion interactions with the

different gauge bosons of the SM

𝑖 𝜆𝑎 𝑎 𝜎𝑖 1
̄ 𝑖𝛾 𝜇 [𝜕𝜇 − 𝑖𝑔𝑠
ℒferm = 𝑞𝐿 𝐺𝜇 − 𝑖𝑔 𝑊𝜇𝑖 − 𝑖𝑔′ ( )𝐵𝜇 ]𝑞𝐿
𝑖
2 2 6
𝜆𝑎 2
+ 𝑢̄𝑖𝑅 𝑖𝛾 𝜇 [𝜕𝜇 − 𝑖𝑔𝑠 𝐺𝑎𝜇 − 𝑖𝑔′ ( )𝐵𝜇 ]𝑢𝑖𝑅
2 3
𝑖̄ 𝜆𝑎 1
+ 𝑑𝑅 𝑖𝛾 𝜇 [𝜕𝜇 − 𝑖𝑔𝑠 𝐺𝑎𝜇 − 𝑖𝑔′ ( − )𝐵𝜇 ]𝑢𝑖𝑅
2 3
𝑖̄ 𝜎𝑖 1
+ ℓ𝐿 𝑖𝛾 𝜇 [𝜕𝜇 − 𝑖𝑔 𝑊𝜇𝑖 − 𝑖𝑔′ ( − )𝐵𝜇 ]ℓ𝐿 𝑖
2 2
+ 𝑒𝑖𝑅
̄ 𝑖𝛾 𝜇 [𝜕𝜇 − 𝑖𝑔′ ( − 1)𝐵𝜇 ]𝑒𝑖𝑅

where 𝜎𝑖 , 𝑖 = 1, 2, 3 are the Pauli matrices and 𝜆𝑎 , 𝑎 = 1, 2, … , 8 are

the Gell-Mann ones (or any other equivalent representation of the
corresponding algebras!).
The SM matter content

The covariant derivatives include the fermion interactions with the

different gauge bosons of the SM
Renormalization
Renormalizability
[’t Hooft and Veltman 1972]

Each operator of our Lagrangian contributes to the action (for a process

with energies 𝐸 ≪ Λ):
𝑘−4
𝐸
𝛿𝑆(𝑖,𝑘) ∼ 𝒞(𝑖,𝑘) ( ) ,
Λ
with 𝑘 being the operator dimension. Operators are then classified by
their importance at low energies (𝐸 → 0):

k low-energy behavior classical renormalizability name

<4 grows super-renormalizable relevant
=4 constant renormalizable marginal
>4 decreases non-renormalizable irrelevant

The SM by definition only includes operators with 𝑘 ≤ 4. Including terms

with 𝑘 > 4 leads to the SM Effective Field Theory (SMEFT).
Renormalizability
[’t Hooft and Veltman 1972]

The condition of renormalizability allows us to compute the counter-term

for a given parameter only once at a given order in perturbation theory.

It is useful to remember that

3
[𝜓] = , [𝜙] = 1, [𝜕𝜇 ] = 1, [𝐴𝜇 ] = 1
2
Then, for instance, terms like the ones below will be forbidden

𝜎𝑖 𝑗 𝑖
𝑘 𝜇𝜎 𝑙
𝑖
(𝑞𝐿 𝑗
̄ 𝛾𝜇 𝑞𝐿 )(𝑢𝑘𝑅 𝛾 𝜇 𝑢𝑙𝑅 ), 𝑖
(𝑞𝐿
̄ 𝛾𝜇 𝑞𝐿 )(𝑞𝐿 𝛾 𝑞 ) 𝑓 𝑎𝑏𝑐 𝐺𝑎𝜇 𝜈 𝐺𝑏𝜈 𝜌 𝐺𝑐𝜌 𝜇 .
2 2 𝐿
Spontaneous symmetry breaking
Spontaneous symmetry breaking
Spontaneous symmetry breaking

If the symmetry group of the SM is not broken, fermions and gauge

bosons will be massless. A way out is given by spontaneous symmetry
breaking (SSB).
Our experimental knowledge tell us that the breaking pattern has to be

𝑆𝑈 (3)𝐶 ⊗ 𝑆𝑈 (2)𝐿 ⊗ 𝑈 (1)𝑌 ↦ 𝑆𝑈 (3)𝐶 ⊗ 𝑈 (1)𝑄

Spontaneous symmetry breaking

Imagine a simple theory of a scalar field 𝜙 with a potential 𝑉 (𝜙)

1
ℒ= (𝜕 𝜙)(𝜕 𝜇 𝜙) − 𝑉 (𝜙)
2 𝜇
The generating functional 𝒵[𝐽 ] will read

𝒵[𝐽 ] = 𝑒𝑖𝒲[𝐽] = 𝒩 ∫ 𝒟𝜙 exp{𝑖 ∫ 𝑑4 𝑥 [ℒ(𝜙) + 𝐽 (𝑥)𝜙(𝑥)]}

We define the vacuum expectation value (vev) of 𝜙(𝑥) in presence of 𝐽

̄ 𝛿𝒲[𝐽 ] 1
𝜙(𝑥) = = ∫ 𝒟𝜙 𝜙(𝑥) exp{𝑖 ∫ 𝑑4 𝑥 [ℒ(𝜙) + 𝐽 (𝑥)𝜙(𝑥)]}
𝛿𝐽 (𝑥) 𝒵[𝐽 ]
⟨0|𝜙(𝑥)|0⟩
=[ ]
⟨0|0⟩ 𝐽

which is the ’conjugated’ of 𝐽 . We assumme that we can invert the

relationship between 𝜙 ̄ and 𝐽 , i.e., 𝜙 ̄ = 𝜙[𝐽
̄ ] and 𝐽 = 𝐽 [𝜙].̄
The effective potential

Using 𝐽 = 𝐽 [𝜙],̄ we can thus define the Legendre transform of 𝒲[𝐽 ]

Γ[𝜙]̄ = 𝒲[𝐽 ] − ∫ 𝑑4 𝑥 𝐽 (𝑥)𝜙(𝑥)

̄

One can see that,

𝛿Γ[𝜙]̄ 𝛿𝑊 [𝐽 ] 𝛿𝐽 (𝑦) 𝛿𝐽 (𝑦) ̄ ̄

𝛿 𝜙(𝑦)
= ∫ 𝑑4 𝑦 − ∫ 𝑑4 𝑦 { 𝜙(𝑦) + 𝐽 (𝑦) }
̄
𝛿 𝜙(𝑥) ̄
𝛿𝐽 (𝑦) 𝛿 𝜙(𝑥) ̄
𝛿 𝜙(𝑥) ̄
𝛿 𝜙(𝑥)
̄ 𝛿𝐽 (𝑦) − ∫ 𝑑4 𝑦 { 𝛿𝐽 (𝑦) 𝜙(𝑦)
= ∫ 𝑑4 𝑦 𝜙(𝑦) ̄ + 𝐽 (𝑦)𝛿(𝑦 − 𝑥)} = −𝐽 (𝑥)
̄
𝛿 𝜙(𝑥) ̄
𝛿 𝜙(𝑥)

and thus
𝛿Γ[𝜙]̄
𝐽 (𝑥) = −
̄
𝛿 𝜙(𝑥)
It can be shown that, while 𝒲[𝐽 ] is the generating functional of the
connected diagrams, Γ[𝜙]̄ generates 1PI-diagrams.
The effective potential

1 particle irreducible (1PI) diagrams are those which can not be

separated in two by cutting a single line

Diagrams which are not 1PI can be generated via 1PI ones, thus Γ is
enough to generate the 𝑆 matrix elements. Moreover

𝑖ℏ 𝛿 2 𝑆[𝜙]
Γ[𝜙]̄ = 𝑆[𝜙]̄ + Tr [log ∣ ] + 𝒪(ℏ2 ).
2 𝛿𝜙(𝑥)𝛿𝜙(𝑦) 𝜙=𝜙 ̄
The effective potential
̄ 𝜙 ̄ = 0. We will have SSB if
In the case of 𝐽 = 0 we know that 𝛿Γ[𝜙]/𝛿

𝛿Γ[𝜙]̄
∣ =0
̄
𝛿 𝜙(𝑥) ̄
𝜙(𝑥)=⟨𝜙(𝑥)⟩≠0

̄
for a non-vanishing configuration 𝜙(𝑥)| 𝐽=0 = ⟨𝜙(𝑥)⟩. In general, if we
expand in derivatives,
1
Γ[𝜙]̄ = ∫ 𝑑4 𝑥 [ − 𝑉eff (𝜙)̄ + (𝜕𝜇 𝜙)(𝜕
̄ 𝜇 𝜙)𝑍(
̄ 𝜙)̄ + … ]
2

and assuming a translationally invariant vev, i.e., ⟨𝜙(𝑥)⟩ = ⟨𝜙⟩, this

condition is translated to
𝑑𝑉eff (𝜙)̄
∣ = 0.
𝑑𝜙 ̄ 𝜙=⟨𝜙⟩ ̄
The effective potential

It can be seen that

̄
𝑉eff (𝜙)∣ = ℰ0
̄
𝜙=⟨𝜙⟩

where ℰ0 is the energy density of the ground state |0⟩, i.e.,

ℰ0 = ⟨0|ℋ|0⟩.

It can also be shown that

ℏ 𝑑4 𝑝𝐸
𝑉eff (𝜙)̄ = 𝑉 (𝜙)̄ + ∫ 2
log(𝑝𝐸 ̄ + 𝒪(ℏ2 )
+ 𝑉 ′′ (𝜙))
2 (2𝜋)4

This is the so-called Coleman-Weinberg potential [Coleman and

E. J. Weinberg 1973]
Example I
Consider a real scalar field with
1 2 2 𝜆 4
𝑉 (𝜙) = 𝜇 𝜙 + 𝜙 , 𝜆>0
2 4
invariant under 𝜙 ↦ −𝜙. Then
1
ℋ= [(𝜕0 𝜙)2 + (∇𝜙)2 ] + 𝑉 (𝜙)
2
For 𝜇2 > 0 the minimum of the potential happens for ⟨𝜙⟩ = 0, while for
𝜇2 < 0 we have
−𝜇2
⟨𝜙⟩ = 𝑣 = ±√
𝜆
Example I

In order to have a quantum field with no vev (such that 𝑎|0⟩ = 0), we do

𝜙(𝑥) = 𝑣 + 𝜂(𝑥) ⟨𝜂⟩ = 0

At the quantum level,

1 𝜆
ℒ= (𝜕𝜇 𝜂)(𝜕 𝜇 𝜂) − 𝜆𝑣2 𝜂2 − 𝜆𝑣𝜂3 − 𝜂4
2 4
which is not invariant under 𝜂 ↦ −𝜂.

Even if ℒ feature some symmetry, it can happen that the parameters are
such that the ground state of the Hamiltonian is not symmetric ⇔ SSB
Example II
Let us consider now a complex scalar field 𝜙(𝑥) with Lagrangian

ℒ = (𝜕𝜇 𝜙† )(𝜕 𝜇 𝜙) − 𝜇2 |𝜙|2 − 𝜆|𝜙|4

invariant under 𝑈 (1) rotations 𝜙 ↦ 𝑒𝑖𝛼 𝜙. If we assume 𝜆 > 0, 𝜇2 < 0

and we obtain
𝑣 −𝜇2
⟨0|𝜙|0⟩ = √ , |𝑣| = √
2 𝜆
Example II

If we take 𝑣 > 0, we can write

1
𝜙(𝑥) = √ [𝑣 + 𝜂(𝑥) + 𝑖𝜒(𝑥)], ⟨0|𝜂|0⟩ = ⟨0|𝜒|0⟩ = 0
2
and
1 1
ℒ= (𝜕 𝜂)(𝜕 𝜇 𝜂) + (𝜕𝜇 𝜒)(𝜕 𝜇 𝜒) − 𝜆𝑣2 𝜂2 − 𝜆𝑣𝜂(𝜂2 + 𝜒2 )
2 𝜇 2
𝜆 2 𝜆 √
− (𝜂 + 𝜒2 )2 + 𝑣4 , 𝑚𝜂 = 2𝜆𝑣, 𝑚𝜒 = 0.
4 4
𝑈 (1) is no longer respected and one scalar remains massless ⇒
Goldstone theorem [Goldstone 1961; Nambu 1960]
Goldstone theorem
Consider a theory with 𝑛 real scalar fields, 𝜙𝑖 , 𝑖 = 1, … , 𝑛 featuring some
global continuous symmetry 𝐺, generated by charges 𝑄𝑎 . Under 𝐺

𝜙𝑖 → 𝜙𝑖′ ≈ 𝜙𝑖 − 𝑖Θ𝑎 𝑇𝑖𝑗𝑎 𝜙𝑗 ⟺ [𝑄𝑎 (𝑡), 𝜙𝑖 (x, 𝑡)] = −𝑇𝑖𝑗𝑎 𝜙𝑗 (x, 𝑡)

where
𝜕ℒ
𝑗𝜇𝑎 (𝑥) = −𝑖 𝑇 𝑎 𝜙 = −𝑖(𝜕𝜇 𝜙𝑖 )𝑇𝑖𝑗𝑎 𝜙𝑗 , 𝑄𝑎 (𝑡) = ∫ 𝑑3 x 𝑗0 (x, 𝑡).
𝜕(𝜕 𝜇 𝜙𝑖 ) 𝑖𝑗 𝑗
We also assume that ⟨0|𝜙𝑖 |0⟩ = 𝑣𝑖 ≠ 0, 𝑖 = 1, … , 𝑛, and consider
𝐺𝑎𝜇,𝑘 (𝑥 − 𝑦) = ⟨0|𝑇 {𝑗𝜇𝑎 (𝑥)𝜙𝑘 (𝑦)}|0⟩ = 𝜃(𝑥0 − 𝑦0 )⟨0|𝑗𝜇𝑎 (𝑥)𝜙𝑘 (𝑦)|0⟩+
𝜃(𝑦0 − 𝑥0 )⟨0|𝜙𝑘 (𝑦)𝑗𝜇𝑎 (𝑥)|0⟩
such that
𝜕𝑥𝜇 𝐺𝑎𝜇,𝑘 (𝑥 − 𝑦) = 𝛿(𝑥0 − 𝑦0 )⟨0|[𝑗0𝑎 (𝑥), 𝜙𝑘 (𝑦)]|0⟩.
On the other hand, assuming translational invariance,

[𝑗0𝑎 (x, 𝑡), 𝜙𝑘 (y, 𝑡)] = −𝑇𝑘𝑗

𝑎
𝜙𝑗 (x, 𝑡)𝛿(x − y)
⇒ 𝜕𝑥𝜇 𝐺𝑎𝜇,𝑘 (𝑥 − 𝑦) = −𝛿(𝑥 − 𝑦)𝑇𝑘𝑗
𝑎
⟨0|𝜙𝑗 (0)|0⟩
Goldstone theorem
If we introduce its Fourier transform
𝑑4 𝑝 ̃ (𝑝),
𝐺𝑎𝜇,𝑘 (𝑥 − 𝑦) = ∫ exp( − 𝑖𝑝𝜇 (𝑥 − 𝑦)𝜇 )𝐺𝑎𝜇,𝑘
(2𝜋)4

we obtain
̃ (𝑝) = 𝑇 𝑎 ⟨0|𝜙 (0)|0⟩.
𝑖𝑝𝜇 𝐺𝑎𝜇,𝑘 𝑘𝑗 𝑗

̃ (𝑝) = 𝑝 𝐹 𝑎 (𝑝2 ) and thus

Lorentz invariance imposes that 𝐺𝑎𝜇,𝑘 𝜇 𝑘

1
𝐹𝑘𝑎 (𝑝2 ) = −𝑖𝑇𝑘𝑗
𝑎
⟨0|𝜙𝑗 (0)|0⟩
𝑝2

Therefore, ⟨0|𝜙𝑗 (0)|0⟩ = 𝑣𝑗 ≠ 0 ⇔ we have poles at 𝑝2 = 0.

There is one massless boson for each ’broken’ generator, i.e., 𝑄𝑎 |0⟩ ≠ 0.
In the case of gauge symmetries, gauge-fixing requires the specification
of some four-vector 𝑛𝜇 and 𝐺𝑎𝜇,𝑘 (𝑝) ≠ 𝑝𝜇 𝐹𝑘𝑎 (𝑝2 ). Loophole!
Goldstone theorem
For a global symmetry given by the Lie group 𝐺, one should have

[𝑄𝑎 (𝑡), 𝐻] = 0, 𝑎 = 1, … , 𝑛, 𝐻 = ∫ 𝑑3 x ℋ(x, 𝑡).

By definition

𝐻|0⟩ = 0 ⇒ 𝐻(𝑄𝑎 |0⟩) = 𝑄𝑎 𝐻|0⟩ = 0, 𝑎 = 1, … , 𝑛.

Therefore, 𝑄𝑎 |0⟩ is also a vacuum state. There are two possibilites:

1 𝑄𝑎 |0⟩ = 0, ∀𝑎. There is just one vacuum.

2 ∃ 𝐴 ⊂ {1, … , 𝑛} | ∀𝑎′ ∈ 𝐴, 𝑄′𝑎 |0⟩ ≠ 0. Then
𝑎′
𝑒−𝑖𝑄𝑎′ Θ |0⟩

are degenerate minimum, and the excitations between them cost no

energy ⇒ Goldstone bosons!
The gauge case: example II revisited
Let us consider a complex scalar field charged under some 𝑈 (1) (sQED),

1
ℒ = − 𝐹𝜇𝜈 𝐹 𝜇𝜈 + (𝐷𝜇 𝜙)† (𝐷𝜇 𝜙) − 𝜇2 |𝜙|2 − 𝜆|𝜙|4 , 𝐷𝜇 = 𝜕𝜇 − 𝑖𝑒𝐴𝜇
4
the gauged version of Example II. The Lagrangian is invariand under
1
𝜙(𝑥) → 𝜙′ (𝑥) = 𝑒−𝑖𝜃(𝑥) 𝜙(𝑥), 𝐴𝜇 (𝑥) → 𝐴𝜇 (𝑥) − 𝜕𝜇 𝜃(𝑥).
𝑒
In the case of 𝜇2 < 0, 𝜆 > 0 we can again write
1
𝜙(𝑥) = √ [𝑣 + 𝜂(𝑥) + 𝑖𝜒(𝑥)], 𝜇2 = −𝜆𝑣2 ,
2
so we obtain
1 1 1 𝜆
ℒ = − 𝐹𝜇𝜈 𝐹 𝜇𝜈 + (𝜕𝜇 𝜂)2 + (𝜕𝜇 𝜒)2 − 𝜆𝑣2 𝜂 − 𝜆𝑣𝜂(𝜂2 + 𝜒2 ) − (𝜂2 + 𝜒2 )2
4 2 2 4
2 2 2
1 4 𝑒 𝑣 𝑒
+ 𝜆𝑣 − 𝑒𝑣𝐴𝜇 𝜕 𝜇 𝜒 + 𝑒𝐴𝜇 (𝜒⃡⃡⃡⃡⃡⃡⃡⃡
𝜕 𝜇 𝜂) + 𝐴𝜇 𝐴𝜇 + 𝐴𝜇 𝐴𝜇 [𝜂2 + 𝜒2 + 2𝑣𝜂]
4 2 2
The gauge case: example II revisited
Some highlights:
• The boson 𝐴𝜇 becomes massive, 𝑚𝐴 = |𝑒𝑣|.
√
• The scalar 𝜂 gets a mass 𝑚𝜂 = 2𝜆𝑣.
• The scalar 𝜒 is massless but has a kinetic mixing with 𝐴𝜇 , 𝐴𝜇 𝜕 𝜇 𝜒.

In order to remove the mixing we add the following gauge-fixing term

1 2
ℒGF = − (𝜕𝜇 𝐴𝜇 + 𝜉𝑚𝐴 𝜒)
2𝜉
Indeed,

IBP 1 𝑚2 1 2
ℒ + ℒGF = − 𝐹𝜇𝜈 𝐹 𝜇𝜈 + 𝐴 𝐴𝜇 𝐴𝜇 − (𝜕 𝐴𝜇 )
4 2 2𝜉 𝜇
1 𝜉𝑚2𝐴 2
+ (𝜕𝜇 𝜒)(𝜕 𝜇 𝜒) − 𝜒 +…
2 2
√
𝜒 gets a gauge-dependent mass, 𝑚𝜒 = 𝜉𝑚𝐴 . It is an unphysical field.
The gauge case: example II revisited
This can be made simpler by using a smarter parametrization of the
complex field
1 𝑎(𝑥)
𝜙(𝑥) = √ [𝑣 + 𝜂(𝑥)]𝑒𝑖 𝑣 .
2
We can then gauge away the exponential by doing the following gauge
transformation
1
𝜙(𝑥) → 𝜙′ (𝑥) = 𝑒−𝑖𝑎(𝑥)/𝑣 𝜙(𝑥) = √ [𝑣 + 𝜂(𝑥)].
2
This is the unitary gauge (⇔ 𝜉 → ∞), where

1 1 𝜆
ℒ = − 𝐹𝜇𝜈 𝐹 𝜇𝜈 + (𝜕𝜇 𝜂)(𝜕 𝜇 𝜂) − 𝜆𝑣2 𝜂 − 𝜆𝑣𝜂3 − 𝜂4
4 2 4
1 4 𝑒2 𝑣2 𝑒 2
+ 𝜆𝑣 + 𝐴𝜇 𝐴𝜇 + 𝐴𝜇 𝐴𝜇 [𝜂2 + 2𝑣𝜂]
4 2 2
Goldstone bosons are decoupled and no need to take them into account
in loop calculations. However, gauge propagators are more complicated.
Brout-Englert-Higgs mechanism

[…] has it ever occurred to you, that, […] this

could be a-a-a-a lot more, uh, uh, uh, uh, uh, uh,
complex, I mean, it’s not just, it might not be just
such a simple... uh, you know?

The Dude to The Big Lebowski

Theorem (Brout-Englert-Higgs mechanism)

The gauge bosons associated with the spontaneously broken generators
become massive, the corresponding would-be Goldstone bosons are
unphysical and can be absorbed, the remaining massive scalars (Higgs
bosons) are physical.

[Englert and Brout 1964; Guralnik, Hagen, and Kibble 1964; Higgs 1964]
The Higgs mechanism in the SM

We introduce a complex scalar

1 𝜙+ 1 0
Φ ∼ (1, 2, ), Φ=( ), ⟨0|𝜙|0⟩ = √ ( )
2 𝜙0 2 𝑣
with the following Lagrangian

𝜎𝑖 1
ℒ𝛷 = |𝐷𝜇 Φ|2 − 𝜇2 |Φ|2 − 𝜆|Φ|4 , 𝐷𝜇 Φ = (𝜕𝜇 − 𝑖𝑔𝑊𝜇𝑖 − 𝑖𝑔′ 𝐵𝜇 )Φ.
2 2
We want to break 𝑆𝑈 (2)𝐿 ⊗ 𝑈 (1)𝑌 → 𝑈 (1)𝑄 so we define 𝑄 = 𝑇𝐿3 + 𝑌 ,
where 𝑌 𝜑 = 𝑦𝜑 𝜑, ∀𝜑. Then

𝜎3 1 1 0 0
𝑄|0⟩ = [𝑇𝐿3 + 𝑌 ]|0⟩ = [ + ]|0⟩ = ( ) |0⟩ = 𝑄 ( ) = 0
2 2 0 0 𝑣

The term |𝐷𝜇 Φ|2 will generate masses for 𝑊𝜇1,2 and one linear
combination of 𝑊𝜇3 and 𝐵𝜇 .
The Higgs mechanism in the SM

We can write √ +
1 2𝜙 (𝑥)
Φ(𝑥) = √ ( )
2 𝑣 + ℎ(𝑥) + 𝑖𝜒(𝑥)
It is useful to define the following linear combinations

𝑍 𝑐 −𝑠𝑊 𝑊3
( 𝜇) = ( 𝑊 ) ⋅ ( 𝜇) 𝑠𝑊 = sin 𝜃𝑊 , 𝑐𝑊 = cos 𝜃𝑊
𝐴𝜇 𝑠𝑊 𝑐𝑊 𝐵𝜇

where tan 𝜃𝑊 = 𝑔′ /𝑔. We also define

1 1
𝑊 ± = √ [𝑊𝜇1 ∓ 𝑖𝑊𝜇2 ], 𝑇𝐿± = √ [𝜎1 ± 𝑖𝜎2 ].
2 2 2

Then, defining 𝑒 = 𝑔𝑠𝑊 = 𝑔′ 𝑐𝑊 = 𝑔𝑔′ /√𝑔2 + 𝑔′2 , we can write

𝑔
𝐷𝜇 Φ = [𝜕𝜇 − 𝑖𝑔𝑊𝜇± 𝑇𝐿± − 𝑖 𝑍 (𝑇 3 − 𝑠2𝑊 𝑄) − 𝑖𝑒𝐴𝜇 𝑄]Φ.
𝑐𝑊 𝜇 𝐿
The Higgs mechanism in the SM
Then, taking into account that
1 0 1 1 0 0
𝑇𝐿+ = √ ( ), 𝑇𝐿− = √ ( ),
2 0 0 2 1 0

we obtain
√
1 2𝜕𝜇 𝜙+ (𝑥) 𝑖𝑔 𝑊 + (𝑥)[𝑣 + ℎ(𝑥) + 𝑖𝜒(𝑥)]
𝐷𝜇 Φ(𝑥) = √ ( )− ( 𝜇 √ )
2 𝜕𝜇 ℎ(𝑥) + 𝑖𝜕𝜇 𝜒(𝑥) 2 2𝑊𝜇− (𝑥)𝜙+ (𝑥)
√ √
𝑖𝑔 ( 1 − 𝑠2𝑊 ) 2𝜙+ (𝑥) 1 2𝜙+ (𝑥)
−√ 𝑍𝜇 ( 12 ) − 𝑖𝑒𝐴𝜇 √ ( )
2𝑐𝑊 − 2 [𝑣 + ℎ(𝑥) + 𝑖𝜒(𝑥)] 2 0

and
𝑔2 𝑣2 + −𝜇 1 𝑔2 𝑣2 𝑖𝑔𝑣 𝑔𝑣
ℒ𝛷 ⊃ 𝑊𝜇 𝑊 + 𝑍𝜇 𝑍 𝜇 ∓ 𝜕𝜇 𝜙∓ 𝑊 ±𝜇 + 𝜕 𝜒𝑍 𝜇 + … .
4 2
2 4𝑐𝑊 2 2𝑐𝑊 𝜇

Then 𝑚𝑊 = 𝑔𝑣/2, 𝑚𝑍 = 𝑚𝑊 /𝑐𝑊 and

1 1 1
ℒGF = − (𝜕 𝐴𝜇 )2 − |𝜕 𝑊 +𝜇 − 𝑖𝜉𝑊 𝑚𝑊 𝜙+ |2 − (𝜕 𝑍 𝜇 − 𝜉𝑍 𝑚𝑍 𝜒)2
2𝜉𝛾 𝜇 𝜉𝑊 𝜇 2𝜉𝑍 𝜇
The Higgs mechanism in the SM
Then, the propagators for the electroweak (EW) bosons become

𝛾 𝑖 𝑘𝜇 𝑘𝜈
𝐷̃ 𝜇𝜈 (𝑘) = [ − 𝑔𝜇𝜈 + (1 − 𝜉𝐴 ) 2 ]
𝑘2 + 𝑖𝜀 𝑘
𝑖 𝑘𝜇 𝑘𝜈
𝐷̃ 𝜇𝜈
𝑍
(𝑘) = 2
[ − 𝑔𝜇𝜈 + (1 − 𝜉𝑍 ) 2 ],
2
𝑘 − 𝑚𝑍 + 𝑖𝜀 𝑘 − 𝜉𝑍 𝑚2𝑍
𝑖 𝑘𝜇 𝑘𝜈
𝐷̃ 𝜇𝜈
𝑊
(𝑘) = 2
[ − 𝑔𝜇𝜈 + (1 − 𝜉𝑍 ) 2 ],
2
𝑘 − 𝑚𝑊 + 𝑖𝜀 𝑘 − 𝜉𝑊 𝑚2𝑊
𝑖
𝐷̃ 𝜒 (𝑘) = 2 2 + 𝑖𝜀
𝑘 − 𝜉𝑍 𝑚𝑍
𝑖
𝐷̃ 𝜙 (𝑘) = .
𝑘 − 𝜉𝑊 𝑚2𝑊 + 𝑖𝜀
2

t’Hooft-Feynman gauge: 𝜉𝐴 = 𝜉 𝑍 = 𝜉 𝑊 = 1
Unitary gauge: 𝜉𝑍 = 𝜉 𝑊 → ∞
The Higgs mechanism in the SM
Let’s see what happens to ghosts. The gauge-fixing term (for the EW
part of the SM) could be writen as
1 2 1 1 2 1 2 1 1 2
ℒGF = − 𝐹 − |𝐹 |2 − 𝐹 =− 𝐹 − (𝐹 2 +𝐹𝑊
2
)− 𝐹
2𝜉𝛾 𝛾 𝜉𝑊 + 2𝜉𝑍 𝑍 2𝜉𝛾 𝛾 2𝜉𝑊 𝑊1 2 2𝜉𝑍 𝑍
√
where (𝜙+ = 1/ 2(𝜙1 − 𝑖𝜙2 ))
𝐹𝑊1 = 𝜕𝜇 𝑊 1𝜇 − 𝜉𝑊 𝑚𝑊 𝜙2 , 𝐹𝑊2 = 𝜕𝜇 𝑊 2𝜇 + 𝜉𝑊 𝑚𝑊 𝜙1
𝐹𝑍 = 𝜕𝜇 𝑍 𝜇 − 𝜉𝑍 𝑚𝑍 𝜒 𝐹𝛾 = 𝜕 𝜇 𝐴𝜇 .
Defining analoguous linear combinations for the ghost fields,

3 3 1 1 2
𝑐𝛾 = 𝑠 𝑊 𝑐𝑤 + 𝑐 𝑊 𝑐𝑏 , 𝑐 𝑍 = 𝑐 𝑊 𝑐𝑤 − 𝑠 𝑊 𝑐𝑏 , 𝑐± = √ [𝑐𝑤 ∓ 𝑖𝑐𝑤 ]
2
1,2,3
we can write (where 𝑐1,2,3 = 𝑐𝑤 , 𝑐4 = 𝑐𝑏 , 𝑈 (𝜃) = exp(−𝑖𝑇𝐿𝑖 𝜃𝑖 ) and
𝑈 (𝜃4 ) = exp(−𝑖𝑌 𝜃4 ))
4
𝛿𝐹+ 𝛿𝐹− 𝛿𝐹𝛾 𝛿𝐹𝑍
ℒghost |EW = ∑ [𝑐+̄ + 𝑐−̄ + 𝑐𝛾̄ + 𝑐𝑍̄ ]𝑐
𝑖=1
𝛿𝜃𝑖 𝛿𝜃𝑖 𝛿𝜃𝑖 𝛿𝜃𝑖 𝑖
The Higgs mechanism in the SM
At the end of the day, we obtain

ℒFP = (𝜕𝜇 𝑐𝛾̄ )(𝜕 𝜇 𝑐𝛾 ) + (𝜕𝜇 𝑐𝑍̄ )(𝜕 𝜇 𝑐𝑍 ) + (𝜕𝜇 𝑐+̄ )(𝜕 𝜇 𝑐+ ) + (𝜕𝜇 𝑐−̄ )(𝜕 𝜇 𝑐− )

− 𝜉𝑍 𝑚2𝑍 𝑐𝑍̄ 𝑐𝑍 − 𝜉𝑊 𝑚2𝑊 𝑐+̄ 𝑐+̄ − 𝜉𝑊 𝑚2𝑊 𝑐−̄ 𝑐−

+ interactions with EW gauge bosons + interactions with Φ(𝑥)

with propagators
𝑖 𝑖 𝑖
𝐷̃ 𝑐𝛾 (𝑘) = , 𝐷̃ 𝑐𝑍 (𝑘) = 2 , 𝐷̃ 𝑐± (𝑘) = 2
𝑘2 + 𝑖𝜀 𝑘 − 𝜉𝑍 𝑚2𝑍 + 𝑖𝜀 𝑘 − 𝜉𝑊 𝑚2𝑊 + 𝑖𝜀
Custodial symmetry
One can notice that the Higgs potential 𝑉 (Φ) is invariant under 𝑆𝑂(4)
rotations, broken after EWSB to 𝑆𝑂(3), since the Higgs could get its
vev in any of its four real degrees of freedom.

Since 𝑆𝑈 (2) ⊗ 𝑆𝑈 (2) is the double cover of 𝑆𝑂(4), we can also describe
this breaking as 𝑆𝑈 (2)𝐿 × 𝑆𝑈 (2)𝑅 → 𝑆𝑈 (2)𝑉 . For this, it can useful to
to write

Σ = (Φ̃ Φ) , where Φ̃ = 𝑖𝜎2 Φ∗

obtaining
1 1 𝜆 2
ℒΦ = Tr ((𝒟𝜇 Σ)† (𝒟Σ)) − 𝜇2 Tr (Σ† Σ) + [Tr (Σ† Σ)] ,
2 2 4
where
𝜎𝑖 𝜎3
𝒟𝜇 Σ = 𝜕𝜇 Σ − 𝑖𝑔𝑊𝜇𝑖 Σ + 𝑖𝑔′ Σ 𝐵𝜇 .
2 2
Custodial symmetry
One can see, that indeed, the Higgs potential is invariant under a
𝑆𝑈 (2)𝐿 ⊗ 𝑆𝑈 (2)𝑅 symetry – dubbed custodial symmetry – under which

Σ → 𝑈𝐿 Σ 𝑈𝑅† .
After EWSB,
1 𝑣 0
⟨Σ⟩ = ( )
2 0 𝑣
so that 𝑆𝑈 (2)𝐿 ⊗ 𝑆𝑈 (2)𝑅 → 𝑆𝑈 (2)𝑉 as anticipated.

This is however not a symmetry of the entire SM Lagrangian. Indeed,

already 𝒟Σ breaks such symmetry, since 𝑆𝑈 (2)𝑅 will mix components
with different hypercharge. The difference in quark masses will also
violate 𝑆𝑈 (2)𝑅 as well.

One thus expect large radiative corrections coming from top loops (since
the breaking of custodial symmetry should be proportional to 𝑚𝑡 − 𝑚𝑏 )
𝑚2𝑊 3 𝐺 𝑚2
𝜌= 2 2
≈ 1 + √𝐹 2𝑡 .
𝑚𝑍 𝑐𝑊 8 2 𝜋
RECAP
RECAP

Right now, our SM Lagrangian consists of the following terms

ℒSM ⊃ ℒgauge + ℒGF + ℒFP + ℒferm + ℒΦ

These terms provide:

1 Massive gauge bosons 𝑊𝜇 , 𝑍𝜇 with their longitudinal dof

2 Massless gauge bosons 𝐺𝑎𝜇 , 𝐴𝜇
3 Ghosts required for loop calculations
𝑖
4 Chiral fermions 𝑞𝐿 , 𝑢𝑖𝑅 , 𝑑𝑅
𝑖 𝑖
, ℓ𝐿 , 𝑒𝑖𝑅 interacting with SM gauge bosons
5 An additional scalar dof, the Higgs ℎ(𝑥)

However, at the moment, fermions are still massless!

Giving fermions a mass

We can give masses to fermions through the so-called Yukawa Lagrangian

𝑖 ̃ 𝑗 𝑗 𝑖̄
ℒYuk = −(Y𝑢 )𝑖𝑗 𝑞𝐿 𝑖
̄ Φ𝑢𝑅 − (Y𝑑 )𝑖𝑗 𝑞𝐿
̄ Φ𝑑𝑅 − (Y𝑒 )𝑖𝑗 ℓ𝐿 Φ𝑒𝑗𝑅 + h.c.

After EWSB, in the unitary gauge, we obtain

1 𝑖̄ 𝑗
ℒYuk = − √ (𝑣 + ℎ)[(Y𝑢 )𝑖𝑗 𝑢̄𝑖𝐿 𝑢𝑗𝑅 + (Y𝑑 )𝑖𝑗 𝑑𝐿 𝑑𝑅 + (Y𝑒 )𝑖𝑗 𝑒𝑖𝐿̄ 𝑒𝑗𝑅 + h.c.].
2
We can define mass matrices
𝑣 𝑣 𝑣
ℳ𝑢 = √ Y𝑢 , ℳ 𝑑 = √ Y𝑑 , ℳ 𝑒 = √ Y𝑒 .
2 2 2
• Note that neutrino are massless in the SM. Neutrino masses
constitute physics beyond the SM.
Giving fermions a mass

We can diagonalize fermion masses through a singular value

decomposition

𝑢𝐿 = 𝒰𝑢 𝑢′𝐿 , ′
𝑑𝐿 = 𝒰𝑑 𝑑𝐿 , 𝑢𝑅 = 𝒱𝑢 𝑢′𝑅 , ′
𝑑 𝑅 = 𝒱 𝑑 𝑑𝑅 ,
𝑒𝐿 = 𝒰𝑒 𝑒′𝐿 , 𝑒𝑅 = 𝒱𝑒 𝑒′𝑅 ,

with 𝒰𝑢,𝑑,𝑒 and 𝒱𝑢,𝑑,𝑒 unitary satisfying

𝒰†𝑢 ℳ𝑢 𝒱𝑢 = 𝜆𝑢 , 𝒰†𝑑 ℳ𝑑 𝒱𝑑 = 𝜆𝑑 , 𝒰†𝑒 ℳ𝑒 𝒱𝑒 = 𝜆𝑒 ,

with 𝜆𝑢,𝑑,𝑒 diagonal. These rotations do not affect fermion kinetic terms
nor the neutral currents because they are family universal
𝑔
𝜓 ̄ 𝑖𝛾 𝜇 𝐷𝜇 𝜓 ̄ ⊃ 𝜓𝐿̄ 𝑖𝛾 𝜇 [ − 𝑖 𝑍 (𝑇 3 − 𝑠2𝑊 𝑄𝜓 ) − 𝑖𝑒𝐴𝜇 𝑄𝜓 ]𝜓𝐿 + (𝐿 ↔ 𝑅)
𝑐𝑊 𝜇 𝐿

Neutral currrents do not change flavour

Charged currents
Charged currents however are different
𝑔
𝑞𝐿̄ 𝑖𝛾 𝜇 𝐷𝜇 𝑞𝐿 ⊃ −𝑖𝑔𝑊𝜇± 𝑞𝐿̄ 𝑇𝐿± 𝑞𝐿 = √ 𝑊𝜇+ 𝑢̄𝐿 𝛾 𝜇 𝑑𝐿 + h.c.
2
𝑔 + ′ † 𝜇 ′
= √ 𝑊𝜇 𝑢̄ (𝒰𝑢 𝒰𝑑 )𝛾 (1 − 𝛾5 )𝑑𝐿 + h.c.
2 2

The matrix VCKM = 𝒰†𝑢 𝒰𝑑 is the so-called Cabibbo-Kobayashi-Maskawa

(CKM) matrix. – btw, henceforth we will drop the primes.
[Cabibbo 1963; Kobayashi and Maskawa 1973]
Charged currents

In the lepton sector we will have something similar

𝑔
ℓ𝐿̄ 𝑖𝛾 𝜇 𝐷𝜇 ℓ𝐿 ⊃ −𝑖𝑔𝑊𝜇± ℓ𝐿̄ 𝑇𝐿± ℓ𝐿 = √ 𝑊𝜇+ 𝜈𝐿̄ 𝛾 𝜇 𝑒𝐿 + h.c.
2
𝑔 + 𝜇
→ √ 𝑊𝜇 𝜈 ̄ 𝒰𝑒 𝛾 (1 − 𝛾5 )𝑒𝐿 + h.c.
2 2
However, since 𝜈 are massless (in the SM), we could rotate 𝜈𝐿 to make
the interaction diagonal
𝜈𝐿 → 𝒰𝑒 𝜈𝐿 .
This is a consequence of the

𝒢𝑙 = 𝑆𝑈 (3)ℓ𝐿 ⊗ 𝑆𝑈 (3)𝑒𝑅

global symmetry of ℒferm .

The CKM matrix

In general, a 𝑛 × 𝑛 unitary matrix has 𝑛2 real parameters. However,

some phases can be rotated away, leading to (𝑛 − 1)2 real parameters,

𝑛(𝑛 − 1)/2 moduli and (𝑛 − 1)(𝑛 − 2)/2 phases.

The standard parametrization of the CMK matrix gives

𝑐12 𝑐13 𝑠12 𝑐13 𝑠13 𝑒−𝑖𝛿

VCKM ⎛
= ⎜−𝑠12 𝑐23 − 𝑐12 𝑠23 𝑠13 𝑒𝑖𝛿 𝑐12 𝑐23 − 𝑠12 𝑠23 𝑠13 𝑒𝑖𝛿 𝑠23 𝑐13 ⎞⎟
𝑖𝛿
⎝ 𝑠12 𝑠23 − 𝑐12 𝑐23 𝑠13 𝑒 −𝑐12 𝑠23 − 𝑠12 𝑐23 𝑠13 𝑒𝑖𝛿 𝑐23 𝑐13 ⎠

where 𝑐𝑖𝑗 = cos 𝜃𝑖𝑗 > 0, 𝑠𝑖𝑗 = sin 𝜃𝑖𝑗 > 0.

𝛿 is the only source of CP violation in the SM (modulo ℒ𝜃 ).

The last missing piece
For 𝑆𝑈 (𝑛) gauge theories, with 𝑛 ≥ 2, there are gauge configurations
that do not vanish fast enough to be ignored. In Euclidean space,
keeping only the gauge part of the QCD lagrangian, we obtain
1 ̂ ),
̂ 𝒢𝜇𝜈
𝒮𝐸 = − ∫ 𝑑4 𝑥̂ Tr(𝒢𝜇𝜈
2𝑔𝑠2

where

𝒢𝜇̂ ≡ −𝑖𝑔𝑠 𝐺𝑎𝜇̂ 𝑇 𝑎 , ̂ 𝑇 𝑎,

̂ ≡ −𝑖𝑔𝑠 𝐺𝑎𝜇𝜈
𝒢𝜇𝜈

and

𝑥0 = −𝑖𝑥4̂ , 𝑥𝑖 = 𝑥𝑖̂ , 𝑥2 = −𝑥2̂ , 𝑥𝜇̂ = (𝑥0̂ , 𝑥𝑖̂ ), 𝑥𝜈̂ = 𝛿 𝜇𝜈 𝑥𝜈̂ .

𝐴0 = 𝑖𝐴4̂ , 𝐴𝑖 = −𝐴𝑖̂ , 𝐴𝜇̂ = (𝐴0̂ , 𝐴𝑖̂ ), 𝐴𝜇̂ = 𝐴𝜈̂ 𝛿 𝜇𝜈 .

with 𝑥𝜇 = (𝑥0 , 𝑥𝑖 ) and 𝐴𝜇 = (𝐴0 , 𝐴𝑖 ) any Lorentz vector. Henceforth,

we will drop the hat, besides for 𝑥.̂
The last missing piece
There are non-trivial gauge configurations having |𝒮𝐸 | < ∞ and
therefore lim|𝑥|→∞
̂ 𝒢𝜇𝜈 = 0. This implies

lim 𝒢𝜇 (𝑥)̂ = 𝒢𝜇 (𝑥)̂ = −(𝜕𝜇 𝑈 (𝑥))𝑈

̂ (𝑥)̂ −1 = 𝑈 (𝑥)𝜕
̂ 𝜇 𝑈 (𝑥)̂ −1
|𝑥|→∞
̂

For 𝑆𝑈 (2), they are (with 𝑟 = |𝑥|)

̂ [Belavin et al. 1975]
𝑟2 𝑥4̂ + 𝑖𝑥𝑎̂ 𝜏 𝑎
𝒢𝜇 (𝑥)̂ = − (𝜕 𝑈 )𝑈 −1 , 𝑈= , 𝜆 ∈ ℝ+ .
𝑟 2 + 𝜆2 𝜇 𝑟
and satisfy
2
1 1 ̃ ) = 8𝜋 𝑛[𝐺] ,
𝒮𝐸 = − 2
∫ 𝑑4 𝑥̂ Tr(𝒢𝜇𝜈 𝒢𝜇𝜈 ) = − 2 ∫ 𝑑4 𝑥̂ Tr(𝒢𝜇𝜈 𝒢𝜇𝜈
2𝑔𝑠 2𝑔𝑠 𝑔𝑠2
with
1 ̃𝜇𝜈 ) = 1 ∮ 𝑑𝜎𝜇 𝜀𝜇𝜈𝜌𝜆 Tr(𝒢𝜈 𝒢𝜌 𝒢𝜆 )
𝑛[𝐺] = − 2
∫ 𝑑4 𝑥̂ Tr(𝒢𝜇𝜈 𝒢
16𝜋 24𝜋2 𝑆∞
3

1
=− ∮ 𝑑𝜎 𝜀 Tr((𝜕𝜈 𝑈 )𝑈 −1 (𝜕𝜌 𝑈 )𝑈 −1 (𝜕𝜆 𝑈 )𝑈 −1 ).
24𝜋2 𝑆∞ 𝜇 𝜇𝜈𝜌𝜆
3
The last missing piece. Instantons

The funtion 𝑛[𝐺] is called the winding number. The aforementioned

solution has 𝑛[𝐺] = 1 and it is called instanton. Therefore,

8𝜋2
𝒮𝐸 = .
𝑔𝑠2

It can be proven that 𝑛[𝐺1 𝐺2 ] = 𝑛[𝐺1 ] + 𝑛[𝐺2 ]. It is a topological

charge that classifies different homotopy classes in 𝜋3 (𝑆 3 ) = ℤ, since
these configurations are maps

𝑆 3 → 𝑆𝑈 (2) ≅ 𝑆 3 .

Bott’s theorem tell us that an arbitrary mapping 𝑆 3 → 𝐺 can be

deformed into a mapping 𝑆 3 → 𝑆𝑈 (2) and this are also solutions of the
QCD gauge group 𝑆𝑈 (3).
The QCD vacua
The Euclidean QCD partition function reads
𝒵[𝑡] = ⟨Ω|𝑒−𝐻𝑡 |Ω⟩
where |Ω⟩ is the vacuum of the theory. We assume the existence of
infinite vacuum states, |𝑛⟩, with winding numbers 𝑛 ∈ ℤ. Moreover,
there exist a correspondence
𝐺𝜇[𝑛] ⟷ 𝑈𝑛 , 𝑈𝑛 acting on the fock space.
For 𝑛 = 1,
𝑈1 |𝑛⟩ = |𝑛 + 1⟩.
A gauge-invariant vacuum state should have contributions from all
classes, so it makes sense to define a coherent superposition
|𝜃⟩ = ∑ 𝑒𝑖𝑛𝜃 |𝑛⟩
𝑛∈ℤ

where 𝜃 is an arbitrary parameter. Then, this vacuum is gauge-invariant

up to an overall phase
𝑈1 |𝜃⟩ = 𝑒−𝑖𝜃 |𝜃⟩.
The QCD vacua
Consider now a gauge invariant operator 𝐵, [𝐵, 𝑈1 ] = 0, then
′
0 = ⟨𝜃|[𝐵, 𝑈1 ]|𝜃′ ⟩ = (𝑒−𝑖𝜃 − 𝑒−𝑖𝜃 )⟨𝜃|𝐵|𝜃′ ⟩

and ⟨𝜃|𝐵|𝜃′ ⟩ = 0 if 𝜃 ≠ 𝜃′ . Therefore, each |𝜃⟩ is the vacuum of a

separate sector of states, unconnected by any gauge-invariant operator.
In particular,

⟨𝜃|𝑒−𝐻𝑡 |𝜃′ ⟩ = 2𝜋𝛿(𝜃 − 𝜃′ )𝑒−𝐸𝜃 𝑡

and

𝒵(𝑡) = ⟨𝜃|𝑒−𝐻𝑡 |𝜃⟩ = ∑ 𝑒−𝑖𝑛𝜃 ⟨𝑛 + 𝑚|𝑒−𝐻𝑡 |𝑚⟩ →

𝑡→∞
𝑛,𝑚∈ℤ

𝑖𝑔𝑠2 𝜃 ̃ )
∑ 𝑒−𝑖𝑛𝜃 ∫[𝒟𝐺𝜇 ](𝑛) 𝑒−𝒮𝐸 = ∫[𝒟𝐺𝜇 ]exp( − 𝒮𝐸 − ∫ d𝑑 𝑥̂ 𝐺𝑎𝜇𝜈 𝐺𝑎𝜇𝜈
𝑛∈ℤ
32𝜋2

where [𝒟𝐺𝜇 ](𝑛) implies that we only integrate over gauge configurations
with winding number 𝑛.
The QCD vacua

Therefore, this effect can be parametrized by adding a term to the

Euclidean action
𝑖𝑔𝑠2 𝜃 ̃
∫ d𝑑 𝑥̂ 𝐺𝑎𝜇𝜈 𝐺𝑎𝜇𝜈
32𝜋2
which correspond in Minkowski space, to

𝑔𝑠2 ̃ , ̃𝑎𝜇𝜈 = 1 𝜀𝜇𝜈𝜌𝜎 𝐺𝑎 .

ℒ𝜃 = 𝜃 𝐺𝑎 𝐺𝑎𝜇𝜈 𝐺
32𝜋2 𝜇𝜈 2 𝜌𝜎

Chiral rotations can also induce a similar term, so at the end of the day,
the physical combination is

𝜃 ̄ = 𝜃 + argdetℳ

where ℳ is the mass matrix after combining up and down-type quark

matrices in one.
The strong CP problem

Such a term, in particular, induces an electric dipole moment of the

neutron of the size of 𝑑𝑛 = 𝐶EDM 𝑒𝜃,̄ where 𝐶EDM = 2.4(1.0) × 10−16 cm .

Experimental bounds on the latter leads to |𝑑𝑛 | < 1.8 × 10−26 𝑒 cm and

|𝜃|̄ ≲ 10−10 .

This is what is known as the strong CP problem.

In the 𝑆𝑈 (2)𝐿 ⊗ 𝑈 (1)𝑌 → 𝑈 (1)𝑄 case, due to the Higgs sector, the
parameter is unphysical and we can rotate it away.
The whole thing
The whole SM Lagrangian

Now, we have the whole SM Lagrangian

ℒSM = ℒgauge + ℒGF + ℒFP + ℒferm + ℒΦ + ℒYuk + ℒ𝜃

Some EW pheno
Input parameters

The input parameters of the EW sector are:

𝑔, 𝑔′ , 𝑣, 𝜆, 𝑚𝑓 [×9], CKM physical parameters [×4]

We can trade the first four parameters by

𝑒2 𝑔2 𝑠2𝑊 1 𝑚𝑊 √
𝛼= = , 𝑚𝑊 = 𝑔𝑣, 𝑚𝑍 = , 𝑚𝐻 = 2𝜆𝑣,
4𝜋 4𝜋 2 𝑐𝑊

or 𝛼, 𝑚𝑍 , 𝐺𝐹 , 𝑚𝐻 .
The muon decay
The muon decay is very well measured experimentally

𝑔2 𝑔2 √ 1 𝐺2𝐹 𝑚5𝜇
2
≈ 2
= 4 2𝐺𝐹 , = Γ(𝜇 → 𝜈𝜇 𝑒− 𝜈𝑒 ) ≈ ,
𝑚𝑊 − 𝑞 2 𝑚𝑊 𝜏𝜇 192𝜋2

leading to √ 1
𝑣 = ( 2𝐺𝐹 )− 2 = 246 GeV.

This relation will (most likely) change for new physics models!
 𝑒+ 𝑒− → 𝑓 𝑓 ̄
Measured at PEP, PETRA, TRISTAN, …, LEP1, SLD

𝐺1 (𝑠) = 𝑄2𝑒 𝑄2𝑓 + 2𝑄𝑒 𝑄𝑓 𝑣𝑒 𝑣𝑓 Re(𝜉𝑍 (𝑠)) + (𝑣𝑒2 + 𝑎2𝑒 )|𝜉𝑍 (𝑠)|2 ,
𝐺2 (𝑠) = (𝑣𝑒2 + 𝑣𝑓2 )𝑎2𝑓 |𝜉𝑍 (𝑠)|2 ,
𝐺3 (𝑠) = 2𝑄𝑒 𝑄𝑓 𝑎𝑒 𝑎𝑓 Re(𝜉𝑍 (𝑠)) + 4𝑣𝑒 𝑣𝑓 𝑎𝑒 𝑎𝑓 |𝜉𝑍 (𝑠)|2

where
𝑠
ℒ𝑍 = 𝑒𝑓𝛾̄ 𝜇 (𝑣𝑓 − 𝑎𝑓 𝛾5 )𝑓𝑍𝜇 , 𝜉𝑍 (𝑠) = ,
𝑠 − 𝑚2𝑍 + 𝑖𝑚𝑍 Γ𝑍
𝛽𝑓 = √1 − 4𝑚2𝑓 /𝑠
 𝑒+ 𝑒− → 𝑓 𝑓 ̄
If we integrate over the whole solid angle we obtain
2𝜋𝛼2
𝜎(𝑠) = 𝑁𝑐𝑓 𝛽 [(3 − 𝛽𝑓 )𝐺1 (𝑠) − 3(1 − 𝛽𝑓2 )𝐺2 (𝑠)]
3𝑠 𝑓
Z pole observables
At the 𝑍 pole we can neglect the 𝛾 − 𝑍 interference. Then (neglecting
𝑚𝑓 )

Γ(𝑒+ 𝑒− )Γ(ℎ𝑎𝑑) 𝛼𝑚𝑍 2

𝜎had = 12𝜋 , Γ(𝑍 → 𝑓 𝑓)̄ = 𝑁𝑐𝑓 (𝑣𝑓 + 𝑎2𝑓 )
𝑚2𝑍 Γ2𝑍 3
Γ(𝑏𝑏)̄ Γ(𝑐𝑐)̄ Γ(ℎ𝑎𝑑)
𝑅𝑏 = , 𝑅𝑐 = , 𝑅ℓ = .
Γ(ℎ𝑎𝑑) Γ(ℎ𝑎𝑑) Γ(ℓ+ ℓ− )

Some asymmetries can be very usefull

𝜎(cos 𝜃 > 0) − 𝜎(cos 𝜃 < 0) 3 𝐴 + 𝑃𝑒

𝐴𝑓FB = = 𝐴𝑓 𝑒 ,
𝜎(cos 𝜃 > 0) + 𝜎(cos 𝜃 < 0) 4 1 + 𝑃 𝑒 𝐴𝑒
𝜎 − 𝜎𝑅
𝐴LR = 𝐿 = 𝐴 𝑒 𝑃𝑒 ,
𝜎𝐿 + 𝜎𝑅

where 𝑃𝑒 is the initial electron polarization and

2𝑣𝑓 𝑎𝑓
𝐴𝑓 = .
𝑣𝑓2 + 𝑎2𝑓
Some other observables

• 𝑊 pair production (LEP2), 𝑊 production (Tevatron/LHC)

• Top quark production
• Higgs production and decay
• Higgs pair production, …
GIM mechanism
[Glashow, Iliopoulos, and Maiani 1970]

We just saw that there is no flavour-changing-neutral currents at

tree-level. We are also protected at the loop level,

It is proportional to

∑(VCKM )𝑘𝑖 (V∗CKM )𝑘𝑗 𝐹 (𝑚𝑢𝑘 )

𝑘

In the limit of identical masses this sum goes to zero. In general, it is

going to be suppressed ⇒ GIM mechanism
Hierarchical masses and mixing angles
Fermion masses display huge hierarchies

𝑚𝑢 [GeV] 𝑚𝑑 [GeV] 𝑚𝑠 [GeV] 𝑚𝑐 [GeV] 𝑚𝑏 [GeV] 𝑚𝑡 [GeV]

−3 −3
2.16 ⋅ 10 4.67 ⋅ 10 93.4 1.27 4.18 172.69

𝑚𝑒 [GeV] 𝑚𝜇 [GeV] 𝑚𝜏 [GeV]

0.511 ⋅ 10−3 0.105 1.78

Leptons Quarks

e µ τ u c t

νe νµ ντ
d s b
Hierarchical masses and mixing angles
The same happens with the entries in the CKM matrix

1 − 𝜆2 /2 𝜆 𝐴𝜆3 (𝜚 − 𝑖𝜂)
VCKM ⎛
=⎜ −𝜆 2
1 − 𝜆 /2 𝐴𝜆2 ⎞
⎟ + 𝒪(𝜆4 )
3 2
⎝𝐴𝜆 (1 − 𝜚 − 𝑖𝜂) −𝐴𝜆 1 ⎠

with 𝜆 = |(VCKM )𝑢𝑠 | ≈ 0.22 and 𝐴, 𝜚, 𝜂 = 𝒪(1).

It also helps with flavour

∑(VCKM )𝑘𝑖 (V∗CKM )𝑘𝑗 𝐹 (𝑚𝑢𝑘 ) = (VCKM )𝑢𝑖 (V∗CKM )𝑢𝑗 𝐹 (𝑚𝑢 )
𝑘

+ (VCKM )𝑐𝑖 (V∗CKM )𝑐𝑗 𝐹 (𝑚𝑐 ) + (VCKM )𝑡𝑖 (V∗CKM )𝑡𝑗 𝐹 (𝑚𝑡 )

∼ [(VCKM )𝑢𝑖 (V∗CKM )𝑢𝑗 + (VCKM )𝑐𝑖 (V∗CKM )𝑐𝑗 ]𝐹 (0) + (VCKM )𝑡𝑖 (V∗CKM )𝑡𝑗 𝐹 (𝑚𝑡 )

∼ (VCKM )𝑡𝑖 (V∗CKM )𝑡𝑗 [𝐹 (𝑚𝑡 ) − 𝐹 (0)]

The unitarity triangle
The unitarity of the CKM matrix implies in particular (henceforth, we
drop the CKM subscript for simplicity)

V𝑢𝑑 V∗𝑢𝑏 + V𝑐𝑑 V∗𝑐𝑏 + V𝑡𝑑 V∗𝑡𝑏 = 0

or
V𝑢𝑑 V∗𝑢𝑏 V𝑡𝑑 V∗𝑡𝑏
∗ + + 1 = 0 ⇔ [𝜚 ̄ + 𝑖𝜂]̄ + [(1 − 𝜚)̄ − 𝑖𝜂]̄ − 1 = 0
V𝑐𝑑 V𝑐𝑏 V𝑐𝑑 V∗𝑐𝑏
where
V∗𝑢𝑏 V𝑢𝑑 𝜆2 𝜆2
𝜚 ̄ + 𝑖𝜂 ̄ = − , 𝜚 ̄ = 𝜚(1 − ) + 𝒪(𝜆4 ), 𝜂 ̄ = 𝜂(1 − ) + 𝒪(𝜆4 )
V∗𝑐𝑏 V𝑐𝑑 2 2
The unitarity triangle

A global fit leads to

𝜆 = 0.22500 ± 0.00067, 𝐴 = 0.826+0.018

−0.015 ,
𝜚 ̄ = 0.159 ± 0.010, 𝜂 ̄ = 0.348 ± 0.010,

and

𝛼 + 𝛽 + 𝛾 = (173 ± 6)∘ , 𝐽 = ℑ(V𝑢𝑠 V𝑐𝑏 V∗𝑢𝑏 V∗𝑐𝑠 ) = (3.08+0.15 −5

−0.13 ) × 10 ,

with 𝐽 being twice the area of all the unitarity triangles.

The unitarity triangle
The chiral anomaly
Let us start with QED
1 2
∫ 𝒟𝜓 ̄ 𝒟𝜓 𝒟𝐴𝜇 exp[𝑖 ∫ 𝑑4 𝑥( − 𝐹𝜇𝜈 ̄
+ 𝑖𝜓𝐷𝜓)]
4

The integrand is unvariant under 𝜓 → 𝑒𝑖𝛼 𝜓 and 𝜓 → 𝑒𝑖𝛽𝛾5 𝜓 with 𝐴𝜇

unchanged. However, under a local axial transformation the path-integral
measure is no longer invariant (𝐴𝜇 is also invariant)

𝒟𝜓 ̄ 𝒟𝜓 → |𝒥|−2 𝒟𝜓 ̄ 𝒟𝜓, 𝒥 = det (𝑒𝑖𝛽(𝑥)𝛾5 ) = exp tr log (𝑒𝑖𝛽(𝑥)𝛾5 )

Moreoever,
𝒥 = exp (𝑖 ∫ 𝑑4 𝑥𝛽(𝑥)Tr[𝛾5 ]) → ∞

Regulating the integral leads to [Fujikawa 1979]

𝑒2 𝜇𝜈𝜌𝜎
𝒥 = exp [ − 𝑖 ∫ 𝑑4 𝑥(𝛽(𝑥) 𝜀 𝐹𝜇𝜈 (𝑥)𝐹𝛼𝛽 (𝑥))]
32𝜋2
The chiral anomaly

Therefore,

∫ 𝒟𝜓 ̄ 𝒟𝜓 𝒟𝐴𝜇 exp[𝑖 ∫ 𝑑4 𝑥ℒQED ] →

𝑒2 𝜇𝜈𝜌𝜎
∫ 𝒟𝜓 ̄ 𝒟𝜓 𝒟𝐴𝜇 exp[𝑖 ∫ 𝑑4 𝑥ℒQED − 𝐽𝜇5 𝜕 𝜇 𝛽(𝑥) + 𝛽 𝜀 𝐹𝜇𝜈 𝐹𝛼𝛽 ]
32𝜋2
and
𝑒2 𝜇𝜈𝛼𝛽
𝜕𝜇 𝐽 5𝜇 = − 𝜀 𝐹𝜇𝜈 𝐹𝛼𝛽
16𝜋2
The current is no longer conserved!
The chiral anomaly
In the SM
𝑔′2 𝜇𝜈𝛼𝛽
𝜕𝜇 𝐽𝑌5𝜇 = ( ∑ 𝑌𝐿𝐻 − ∑ 𝑌𝑅𝐻 ) 𝜀 𝐵𝜇𝜈 𝐵𝛼𝛽 ∶ 𝑈 (1)3𝑌
𝐿𝐻 𝑅𝐻
16𝜋2

which vanishes for the hypercharge assignments of the SM! For the
non-abelian part, something similar happens but involve

Tr[𝑇𝑅𝑎 {𝑇𝑅𝑏 , 𝑇𝑅𝑐 }] = 𝐴(𝑅)𝑑𝑎𝑏𝑐 , Tr[𝑇𝐿𝑎 {𝑇𝐿𝑏 , 𝑇𝐿𝑐 }] = 𝐴(𝐿)𝑑𝑎𝑏𝑐 ,

All these anomalies cancel in the SM, but just for the case of 3
generations!!
Violation of baryon number
Instanton transitions violates 𝐵 and 𝐿 number in three units. Since the
tunneling rate of the transition is proportional to exp(−𝒮E ), it leads to
2
− 8𝜋2
Γ ∼ 𝑒−𝒮E (𝑛=1) = 𝑒 𝑔𝑠 ∼ 10−173 .

At high temperature (finite 𝑇 QFT), thermal fluctuations dominate

(sphalerons). Lattice calculations provide the best approximation to the
rate of sphaleron transitions in the symmetric phase (before EWSB)

Γ = (18 ± 3)𝛼5𝑊 𝑇 4 ≈ (8.0 ± 1.3) × 10−7 𝑇 4 , 𝛼𝑊 = 𝑔2 /(4𝜋).

After EWSB, these transitions become exponentially suppressed

7
𝐸sph
Γ ∼ 𝐴(𝛼𝑊 𝑇 )4 ( ) exp( − 𝐸sphal (𝑇 )/𝑇 ).
𝑇

At hight temperatures, sphaleron transitions provide sizable violation of

𝐵-number. Enough for baryogenesis!
Baryogenesis in the SM

Sakharov conditions for successful baryogenesis require

1 𝐵 violation. 3
2 Loss of thermal equilibrium. 7
3 𝐶, 𝐶𝑃 violation. 7

𝐶𝑃 violation is given by 𝛿 and 𝜃 ̄ but both are constrained to be too

small to give a sizable violation of 𝐶𝑃 .
Sphalerons are in thermal equilibrium for 𝑇 ∈ [132, 1012 ] GeV. One can
show that if the SM undergoes a sufficiently strong first order phase
transition (EW baryogenesis), this could provide the required departure
from thermal equilibrium. However, this is not the case.
So, the SM can not account for the baryon asymmetry of the universe.
Running in the SM
Running in the SM
All divergences are local, and can be thus absorbed by counterterms. For
instance, in the case of massless QED
1 (0) (0)𝜇𝜈 ̄ 𝑖𝛾 𝜇 𝐷𝜇(0) 𝜓(0) = − 1 𝑍𝐴 𝐹𝜇𝜈 𝐹 𝜇𝜈 + 𝑍𝜓 𝑖𝜓𝛾̄ 𝜇 𝜕𝜇 𝜓
ℒ1loop
QED = − 𝐹𝜇𝜈 𝐹 + 𝜓(0)
4 4
− 𝜇−𝜀 √𝑍𝐴 𝑍𝑒 𝑍𝜓 𝑒𝜓𝛾̄ 𝜇 𝜓𝐴𝜇 , 𝑍𝜓,𝐴 = 1 + 𝒪(1/𝜀), 𝐷 = 4 − 2𝜀,

−1/2
where 𝑍𝑒 = 𝑍𝐴 to all orders due to the Ward-Takahashi identity.
Since 𝑒0 = 𝑒𝜇−𝜀 𝑍𝑒 , we have

𝑒 = 𝑒0 𝜇𝜀 𝑍𝑒−1 = 𝑒0 𝜇𝜀 √𝑍𝐴 .

In particular, in the MS scheme

𝑑 𝑑 𝑑 𝑒3
𝜇 𝑒0 = 𝜇 [𝜇𝜀 𝑒𝑍𝑒 ] = 0 ⇒ 𝛽(𝑒) ≡ 𝜇 𝑒 = −𝜀𝑒 +
𝑑𝜇 𝑑𝜇 𝑑𝜇 12𝜋2

Couplings run!
QCD
The QCD Lagrangian
We have already seen the QCD Lagrangian
1 1
ℒQCD = − 𝐺𝑎𝜇𝜈 𝐺𝑎𝜇𝜈 − 𝑎 (𝜕 𝜇 𝐺𝑎𝜇 )2 + (𝜕 𝜇 𝑐𝑔𝑎̄ )(𝜕𝜇 𝑐𝑔𝑎 − 𝑔𝑠 𝑓 𝑎𝑏𝑐 𝑐𝑔𝑏 𝐺𝑏𝜇 )
4 2𝜉
𝑔2
+ 𝜓𝑖̄ [𝑖𝛾 𝜇 𝐷𝜇 − 𝑚𝑖 ] 𝜓𝑖 + 𝜃 ̄ 𝑠 2 𝐺𝑎𝜇𝜈 𝐺𝑎𝜇𝜈
̃
32𝜋
If we forget about the EW part the covariant derivatives look like
𝜆𝑎 𝑎
𝐷𝜇 = 𝜕𝜇 − 𝑖𝑔𝑠 𝑇 𝑎 𝐺𝑎𝜇 = 𝜕𝜇 − 𝑖𝐺 , 𝑎 = 1, … , 8
2 𝜇
where 𝜆𝑎 are the Gell-Mann matrices (3 × 3 matrices).
Color algebra

In the limit of large 𝑁𝑐 (yes, it is only 3!) we can represent

Since 3 ⊗ 3̄ = 1 ⊕ 8 and 3 ⊗ 3 = 3̄ ⊕ 6, we can build color singlest as

1
𝑞 𝑞 ′̄ ∼ 1 ∈ 3 ⊗ 3̄ ⇒ mesons √ 𝛿 𝛼𝛽 |𝑞 (𝛼) 𝑞 ′(𝛽)
̄ ⟩
3
1
𝑞𝑞 ′ 𝑞 ′′ ∼ 1 ∈ 3 ⊗ 3 ⊗ 3 ⇒ baryons √ 𝜖𝛼𝛽𝛾 |𝑞 (𝛼) 𝑞 ′(𝛽) 𝑞 ′′(𝛾) ⟩
6
We can not build 𝑞𝑞 ′ invariants (but we can do tetraquarks and
pentaquarks!).
Color factors
We can define useful color factors for a given 𝔰𝔲(𝑁 ) representation,
𝑇𝑅𝑎 , 𝑎 = 1, … , 𝑁 2 − 1:

Tr[𝑇𝑅𝑎 𝑇𝑅𝑏 ] = 𝑇𝑅 𝛿 𝑎𝑏 , (𝑇𝑅𝑎 𝑇𝑅𝑎 )𝑖𝑗 = 𝛿𝑖𝑗 𝐶𝑅

where 𝑇𝑅 is the index of the irrep and 𝐶𝑅 the quadratic Casimir. In

particular for the fundamental (F) and the adjoint (A)

One can show that 𝑇𝐹 = 1/2, 𝐶𝐹 = (𝑁 2 − 1)/2𝑁 and 𝐶𝐴 = 𝑁 .

The QCD running
When renormalizing such theory non-abelian theory with 𝑆𝑈 (𝑁 ) and 𝑛𝑓
flavors (at 1-loop) we obtain

𝑔𝑠3 11 4
𝛽(𝑔𝑠 ) = −𝜀𝑔𝑠 − 2
[ 𝐶𝐴 − 𝑛𝑓 𝑇𝐹 ]
16𝜋 3 3
Then, for 𝑁 = 3 and 𝛼𝑠 = 𝑔𝑠2 /4𝜋 we obtain (at 𝜀 = 0)

𝑑 𝛼2 2𝑛𝑓
𝛽(𝛼𝑠 ) = 𝜇 𝛼𝑠 = − 𝑠 𝛽0 , 𝛽0 = 11 −
𝑑𝜇 2𝜋 3
As long as 𝑛𝑓 < 17 we obtain at one loop that [Gross and Wilczek 1973]

𝛽0 > 0 and 𝛼𝑠 (𝜇) decreases with energy!

Solving the equation at one-loop results in

2𝜋 𝜇 −1
𝛼𝑠 (𝜇) = [ log ( )]
𝛽0 ΛQCD

where ΛQCD is the position of the QCD Landau pole. The coupling
constant gets weaker at high energy ⇒ asymptotic freedom!
Asymptotic freedom
At higher orders, we get something similar
Asymptotic freedom
Different regimes of QCD

• At high scales: coupling becomes small, quarks and gluons are

almost free and the strong interactions become weak
• At low scales: coupling becomes large, quarks and gluons interact
strongly and perturbation theory fails
QCD at low energies
At energies ≲ ΛQCD perturbation theory fails. We can try to solve QCD
numerically: lattice QCD. We can also try to take advante of the
symmetries of the QCD Lagrangian at these scales.
For the case of 𝑛𝑓 = 3 active flavors, neglecting quark masses, we can
write
1
ℒ0QCD = − 𝐺𝑎𝜇𝜈 𝐺𝜇𝜈 𝑎 + 𝑖𝑞𝐿̄ 𝛾 𝜇 𝐷𝜇 𝑞𝐿 + 𝑖𝑞𝑅̄ 𝛾 𝜇 𝐷𝜇 𝑞𝑅 , 𝑞 𝑇 = (𝑢, 𝑑, 𝑠),
4
which is invariant under a global 𝑆𝑈 (3)𝐿 ⊗ 𝑆𝑈 (3)𝑅

𝑞𝐿 → 𝑔𝐿 𝑞𝐿 , 𝑞𝑅 → 𝑔𝑅 𝑞𝑅 , 𝑔𝐿,𝑅 ∈ 𝑆𝑈 (3)𝐿,𝑅 .

This symmetry will be spontaneously broken by the vacuum condensate

(and explicitly by the different quark masses)

⟨0|𝑞𝑖̄ 𝑞𝑗 |0⟩ ∝ 𝛿𝑖𝑗 Λ3QCD

making 𝑆𝑈 (3)𝐿 ⊗ 𝑆𝑈 (3)𝑅 → 𝑆𝑈 (3)𝑉 and delivering thus 32 − 1 = 8

(pseudo)Goldstone bosons.
Chiral perturbation theory
We can write a non-linear realization of this symmetry breaking through
the Goldstone matrix
1 𝑎𝑏
𝑈 (𝜋𝑎 ) = exp(2𝑖𝜋𝑎 𝑇 𝑎 /𝑓), Tr(𝑇 𝑎 ⋅ 𝑇 𝑏 ) = 𝛿 ,
2
†
transforming as 𝑈 → 𝑔𝑅 𝑈 𝑔𝐿 , where

√1 𝜋 0 𝜂8
2
+ √
6
𝜋+ 𝐾+
𝜆𝑎 1 ⎛ 𝜋− − √12 𝜋0 + 𝜂8
𝐾0 ⎟
⎞
𝜋𝑎 𝑇 𝑎 = 𝜋𝑎 =√ ⎜
⎜ √
⎟
⎟.
2⎜
6
2 0 2
⎝ 𝐾− 𝐾 − 6 𝜂8 ⎠
√

We can write the most general Lagrangian which is compatible with the
symmetries of QCD organized in increasing powers of derivatives. At
lowest order
𝑓2
ℒChPT = Tr(𝜕𝜇 𝑈 † 𝜕 𝜇 𝑈 )
4
Chiral perturbation theory
Chiral perturbation theory: external sources

The ChPT effective Lagrangian becomes much more powerful if we

asumme that QCD is coupled to some external classical fields:

ℒQCD = ℒ0QCD + 𝑞𝛾̄ 𝜇 (𝑣𝜇 + 𝑎𝜇 𝛾5 )𝑞 − 𝑞(𝑠

̄ − 𝑖𝛾5 𝑝)𝑞

where 𝑣𝜇 , 𝑎𝜇 , 𝑠, 𝑝 are in principle 3 × 3 hermitian matrices. This

Lagrangian is invariant under the following set of local
𝑆𝑈 (3)𝐿 ⊗ 𝑆𝑈 (3)𝑅 transformations

𝑞𝐿 → 𝑔𝐿 (𝑥)𝑞𝐿 , 𝑞𝑅 → 𝑔𝑅 (𝑥)𝑞𝑅 , 𝑠 + 𝑖𝑝 → 𝑔𝑅 (𝑥)(𝑠 + 𝑖𝑝)𝑔𝐿 (𝑥)† ,

ℓ𝜇 → 𝑔𝐿 (𝑥)ℓ𝜇 𝑔𝐿 (𝑥)† + 𝑖𝑔𝐿 (𝑥)𝜕𝜇 𝑔𝐿 (𝑥)† ,
𝑟𝜇 → 𝑔𝑅 (𝑥)𝑟𝜇 𝑔𝑅 (𝑥)† + 𝑖𝑔𝑅 (𝑥)𝜕𝜇 𝑔𝑅 (𝑥)† ,

where

𝑟𝜇 = 𝑣𝜇 + 𝑎𝜇 , ℓ 𝜇 = 𝑣 𝜇 − 𝑎𝜇 .
Chiral perturbation theory: external sources
The way to incorporate the external sources is through the covariant
derivative

𝐷𝜇 𝑈 = 𝜕𝜇 𝑈 − 𝑖𝑟𝜇 𝑈 + 𝑖𝑈 ℓ𝜇 , 𝐷𝜇 𝑈 † = 𝜕𝜇 𝑈 † + 𝑖𝑈 † 𝑟𝜇 − 𝑖ℓ𝜇 𝑈 † .

Then, to lowest order,

𝑓2 𝑓2
ℒ2 = Tr (𝐷𝜇 𝑈 † 𝐷𝜇 𝑈 ) + Tr (𝑈 † 𝜒 + 𝜒† 𝑈 ) , 𝜒 = 2𝐵0 (𝑠 + 𝑖𝑝).
4 4
In the case of QCD,

𝑠 = ℳ + … = diag(𝑚𝑢 , 𝑚𝑑 , 𝑚𝑠 ) + … , 𝑝 = 0,
1
𝑟𝜇 = 𝑒𝒬𝐴𝜇 + … , 𝒬 = diag(2, −1, −1).
3
0 𝑉𝑢𝑑 𝑉𝑢𝑠
𝑒
ℓ𝜇 = 𝑒𝒬𝐴𝜇 + √ (𝑊𝜇+ 𝑇 + + h.c.) + … , 𝑇+ = ⎛
⎜0 0 0 ⎞⎟.
2𝑠𝑊
⎝0 0 0 ⎠
Chiral perturbation theory: meson masses

We obtain for the different mesons:

𝑚2𝜋± = 2𝑚𝐵
̂ 0, 𝑚2𝜋0 = 2𝑚𝐵
̂ 0 − 𝛿 + 𝒪(𝛿 2 ),
𝑚2𝐾 ± = (𝑚𝑢 + 𝑚𝑠 )𝐵0 , 𝑚2𝐾 0 = (𝑚𝑑 + 𝑚𝑠 )𝐵0 ,
2
𝑚2𝜂8 = (𝑚̂ + 2𝑚𝑠 )𝐵0 + 𝛿 + 𝒪(𝛿 2 )
3
where
1 𝐵0 (𝑚𝑢 − 𝑚𝑑 )2
𝑚̂ = (𝑚 + 𝑚𝑑 ), 𝛿= .
2 𝑢 4 (𝑚𝑠 − 𝑚)̂

However, we are just describing the lightest mesons, which are pNGBs

⇒ Resonances and baryons are missing

QCD at high energies
At high energies, we can rely on perturbation theory. We will concentrate
on (hadron) colliders. However, even in this case, stuff is complicated.
QCD at high energies
At colliders, very different energy scales are at work. Neither perturbation
theory nor lattice QCD can give a full solution for QCD at colliders.
QCD at high energies
The common approach relies on perturbative QCD ⊕ non-perturbative
modelling/factorization.
Factorization
Factorization in Deep Inelastic Scattering (DIS) [Collins and Soper 1987]

The cross-section can be written in factorised form:

𝑑𝜎̂ ℓ𝑖→𝑓 (𝑥𝑖 , Φ𝑓 , 𝑄2𝐹 )
𝜎ℓℎ = ∑ ∑ ∫ 𝑑𝑥𝑖 ∫ 𝑑Φ𝑓 𝑓𝑖/ℎ (𝑥𝑖 , 𝑄2𝐹 )
𝑖 𝑓
𝑑𝑥𝑖 𝑑Φ𝑓
Factorization
We assume that an analogous factorization works for hadron collisions

𝑑𝜎 𝑑𝜎̂𝑎𝑏→𝑓 (𝑥𝑎 , 𝑥𝑏 , 𝑓, 𝑄2𝑖 , 𝑄2𝑓 )

= ∑ ∑ ∫ 𝑓𝑎 (𝑥𝑎 , 𝑄2𝑖 )𝑓𝑏 (𝑥𝑏 , 𝑄2𝑖 ) ×
𝑑𝑋 𝑎,𝑏 𝑓 𝑋̂ 𝑓 𝑑𝑋̂ 𝑓
× 𝐷(𝑋̂ 𝑓 → 𝑋, 𝑄2𝑖 , 𝑄2𝑓 )

⋆ 𝑓𝑎 (𝑥𝑎 , 𝑄2𝑖 ) Parton Distribution Function (PDF). It gives the probability

of finding a quark/gluon 𝑎 inside the incoming hadron, carrying a
fraction 𝑥𝑖 of the incoming momentum. Determined experimentally.
̂ 𝑋̂ 𝑓 Differential Partonic Hard Scattering. Computed in
⋆ 𝑑𝜎/𝑑
perturbation theory.
⋆ 𝐷(𝑋̂ 𝑓 → 𝑋, 𝑄2𝑖 , 𝑄2𝑓 ) Fragmentation Functions. Connect high-scale
processes with final state hadrons. Phenomenological models.
Stuff is even more complicated in experiments like ATLAS and CMS
since we do not identify hadrons but ’jets’ from the activity in the
hadronic calorimeter.
Factorization
To understand better factorization we study 𝑒+ 𝑒− collisions with
hadronic final state.
Start with 𝛾 ∗ → 𝑞 𝑞 ̄

Now let’s radiate a gluon

Factorization

After some algebra, if we make the gluon soft, i.e., 𝑘 ≪ 𝑝1 , 𝑝2 , we obtain

𝑝1 ⋅ 𝜖∗ 𝑝 ⋅ 𝜖∗
̄ 1 )𝑖𝑒𝑄𝑞 𝛾𝜇 𝑇 𝑎 𝑣(𝑝2 )𝑔𝑠 (
𝑖ℳ𝑞𝑞𝑔̄ ≈ 𝑢(𝑝 − 2 )
𝑝1 ⋅ 𝑘 𝑝2 ⋅ 𝑘
Then
2
𝑝1 ⋅ 𝜖∗ 𝑝 ⋅ 𝜖∗
∑ |ℳ𝑞𝑞𝑔̄ |2 = ∑ ∣𝑢(𝑝
̄ 1 )𝑒𝑄𝑞 𝛾𝜇 𝑇 𝑎 𝑣(𝑝2 )𝑔𝑠 ( − 2 )∣
𝑎,𝜖 𝑎,𝜖
𝑝1 ⋅ 𝑘 𝑝2 ⋅ 𝑘
2
𝑝1 𝑝 2𝑝1 ⋅ 𝑝2
= −|ℳ𝑞𝑞 |̄ 2 𝐶𝐹 𝑔𝑠2 ( − 2 ) = |ℳ𝑞𝑞 |̄ 2 𝐶𝐹 𝑔𝑠2
𝑝1 ⋅ 𝑘 𝑝2 ⋅ 𝑘 (𝑝1 ⋅ 𝑘)(𝑝2 ⋅ 𝑘)
Factorization

If we include phase space, we obtain

𝑑3 k 2𝑝1 ⋅ 𝑝2
𝑑Φ𝑞𝑞𝑔̄ |ℳ𝑞𝑞𝑔̄ |2 ≈ (𝑑Φ𝑞𝑞 |ℳ
̄
2
𝑞𝑞 |̄ ) 𝐶 𝑔2
2𝐸k (2𝜋)3 𝐹 𝑠 (𝑝1 ⋅ 𝑘)(𝑝2 ⋅ 𝑘)
2
≈ (𝑑Φ𝑞𝑞 |ℳ
̄ 𝑞𝑞 |̄ )𝑑𝒮

̂
where the soft-gluon emission piece can be written as (with 𝜃 = (k, p1 ))

𝑑𝜙 2𝛼𝑠 𝐶𝐹 2𝑝1 ⋅ 𝑝2 2𝛼𝑠 𝐶𝐹 𝑑𝐸 𝑑𝜃 𝑑𝜙

𝑑𝒮 = 𝐸𝑑𝐸𝑑 cos 𝜃 =
2𝜋 𝜋 (𝑝1 ⋅ 𝑘)(𝑝2 ⋅ 𝑘) 𝜋 𝐸 sin 𝜃 2𝜋

which diverges for 𝐸 = 0 (soft or IR) and for sin 𝜃 → 0 (collinear).

So we have seen an example of factorization of soft and/or collinear

divergences, which is indeed a very general feature of QCD.
Soft/collinear divergences
Kinoshita-Lee-Nauenberg theorem tells that if we sum over all the
allowed states, the result should be finite!

Thus, if we regularize the divergences with dimensional regularization

(dimreg), with 𝐷 = 4 − 2𝜀, 1/𝜀 terms of the real part should be
cancelled by 1/𝜀 terms of the virtual part

3 𝛼𝑠 𝐶𝐹
𝜎tot = 𝜎𝑞𝑞 ̄ (1 + + 𝒪(𝛼2𝑠 ))
4 𝜋
Total cross-section

• Corrections to 𝜎tot come from hard large-angle gluons, as well as

from large virtualities: physics at short distance.
• Since soft gluons are emitted on long timescales relative to the
collision scale they can not influence the cross section.
• Hadronization also happens on a long timescale and is thus
factorized.
Infrared and collinear-safe observables

We can also avoid the problem of IR divergences if we define IR and

collinear-safe obervables 𝒪. Then, if

𝑑𝜎Born = ℬ(Φ𝐵 )𝑑Φ𝐵 , 𝑑𝜎NLO = [ℬ(Φ𝐵 ) + 𝒱(Φ𝐵 )] 𝑑Φ𝐵 + ℛ(Φ𝑅 )𝑑Φ𝑅

with 𝑑Φ𝑅 = 𝑑Φ𝐵 𝑑Φrad = 𝑑Φ𝐵 𝑑𝐸𝑑 cos 𝜃𝑑𝜙,

⟨𝒪⟩ = ∫ [ℬ(Φ𝐵 ) + 𝒱(Φ𝐵 ) + ∫ 𝒞(Φ𝑅 )𝑑Φrad ] 𝒪(Φ𝐵 )𝑑Φ𝐵

+ [ℛ(Φ𝑅 )𝒪(Φ𝑅 ) − 𝒞(Φ𝑅 )𝒪(Φ𝐵 )]𝑑Φ𝑅

with 𝒞(Φ𝑅 ) → ℛ(Φ𝑅 ) in the soft/collinear limit. Both, the □ and the □
pieces are independently finite.
Sudakov form factor
Factorization of soft/collinear emissions allows us to write

𝛼𝑠 (𝑞) 𝑑𝑞 𝑑𝜙
𝑑𝜎𝑛+1 = 𝑑𝜎𝑛 (Φ𝑛 )𝒫(Φrad )𝑑Φrad , 𝒫(Φrad )𝑑Φrad ≈ 𝑃 (𝑧, 𝜙)𝑑𝑧
𝜋 𝑞 2𝜋
Then, we can introduce the so-called Sudakov form factor
𝑞2 1
𝛼𝑠 (𝑞) 𝑑𝑞
Δ𝑆 (𝑞1 , 𝑞2 ) = exp {− ∫ ∫ 𝑃 (𝑧)𝑑𝑧}
𝑞1 𝜋 𝑞 𝑧0

which gives the probability of no emission between the scales 𝑞1 and 𝑞2 .

This is the principle of monte-carlo parton-showers!
Showering and hadronization

Taken from G. P. Salam 2020.

References I

’t Hooft, Gerard and M. J. G. Veltman (1972). “Regularization and

Renormalization of Gauge Fields”. In: Nucl. Phys. B 44, pp. 189–213.
doi: 10.1016/0550-3213(72)90279-9 (cit. on pp. 40, 41).
Altarelli, Guido (2017). Collider Physics within the Standard Model:
A Primer. Ed. by James Wells. Vol. 937. Springer. doi:
10.1007/978-3-319-51920-3 (cit. on p. 5).
Becchi, C., A. Rouet, and R. Stora (1976). “Renormalization of
Gauge Theories”. In: Annals Phys. 98, pp. 287–321. doi:
10.1016/0003-4916(76)90156-1 (cit. on p. 30).
Belavin, A. A. et al. (1975). “Pseudoparticle Solutions of the
Yang-Mills Equations”. In: Phys. Lett. B 59. Ed. by J. C. Taylor,
pp. 85–87. doi: 10.1016/0370-2693(75)90163-X (cit. on p. 77).
References II

Branco, Gustavo C., Luis Lavoura, and Joao P. Silva (1999). CP

Violation. Clarendon Press (cit. on p. 5).
Cabibbo, Nicola (1963). “Unitary Symmetry and Leptonic Decays”.
In: Phys. Rev. Lett. 10, pp. 531–533. doi:
10.1103/PhysRevLett.10.531 (cit. on p. 73).
Cheng, Ta-Pei and Ling-Fong Li (1984). Gauge Theory of Elementary
Particle Physics. Oxford, UK: Oxford University Press (cit. on p. 4).
Coleman, Sidney R. and Erick J. Weinberg (1973). “Radiative
Corrections as the Origin of Spontaneous Symmetry Breaking”. In:
Phys. Rev. D 7, pp. 1888–1910. doi: 10.1103/PhysRevD.7.1888
(cit. on p. 49).
References III

Collins, John C. and Davison E. Soper (1987). “The Theorems of

Perturbative QCD”. In: Ann. Rev. Nucl. Part. Sci. 37, pp. 383–409.
doi: 10.1146/annurev.ns.37.120187.002123 (cit. on p. 122).
Donoghue, John F., Eugene Golowich, and Barry R. Holstein (2014).
Dynamics of the Standard Model : Second edition. Cambridge
University Press. doi: 10.1017/CBO9780511803512 (cit. on p. 4).
Ellis, R. Keith, W. James Stirling, and B. R. Webber (1996). QCD
and collider physics. Cambridge University Press. doi:
10.1017/CBO9780511628788 (cit. on p. 5).
Englert, F. and R. Brout (1964). “Broken Symmetry and the Mass of
Gauge Vector Mesons”. In: Phys. Rev. Lett. 13. Ed. by J. C. Taylor,
pp. 321–323. doi: 10.1103/PhysRevLett.13.321 (cit. on p. 60).
References IV

Faddeev, L. D. and V. N. Popov (1967). “Feynman Diagrams for the

Yang-Mills Field”. In: Phys. Lett. B 25. Ed. by Jong-Ping Hsu and
D. Fine, pp. 29–30. doi: 10.1016/0370-2693(67)90067-6 (cit. on
p. 28).
Fujikawa, Kazuo (1979). “Path Integral Measure for Gauge Invariant
Fermion Theories”. In: Phys. Rev. Lett. 42, pp. 1195–1198. doi:
10.1103/PhysRevLett.42.1195 (cit. on p. 98).
Glashow, S. L. (1961). “Partial Symmetries of Weak Interactions”.
In: Nucl. Phys. 22, pp. 579–588. doi:
10.1016/0029-5582(61)90469-2 (cit. on p. 7).
Glashow, S. L., J. Iliopoulos, and L. Maiani (1970). “Weak
Interactions with Lepton-Hadron Symmetry”. In: Phys. Rev. D 2,
pp. 1285–1292. doi: 10.1103/PhysRevD.2.1285 (cit. on p. 92).
References V

Goldstone, J. (1961). “Field Theories with Superconductor

Solutions”. In: Nuovo Cim. 19, pp. 154–164. doi:
10.1007/BF02812722 (cit. on p. 53).
Gross, David J. and Frank Wilczek (1973). “Ultraviolet Behavior of
Nonabelian Gauge Theories”. In: Phys. Rev. Lett. 30. Ed. by
J. C. Taylor, pp. 1343–1346. doi: 10.1103/PhysRevLett.30.1343
(cit. on p. 109).
Guralnik, G. S., C. R. Hagen, and T. W. B. Kibble (1964). “Global
Conservation Laws and Massless Particles”. In: Phys. Rev. Lett. 13.
Ed. by J. C. Taylor, pp. 585–587. doi:
10.1103/PhysRevLett.13.585 (cit. on p. 60).
Higgs, Peter W. (1964). “Broken Symmetries and the Masses of
Gauge Bosons”. In: Phys. Rev. Lett. 13. Ed. by J. C. Taylor,
pp. 508–509. doi: 10.1103/PhysRevLett.13.508 (cit. on p. 60).
References VI

Illana, José Ignacio and Alejandro Jimenez Cano (2022). “Quantum

field theory and the structure of the Standard Model”. In: doi:
10.22323/1.406.0314. arXiv: 2211.14636 [hep-ph] (cit. on p. 5).
Kobayashi, Makoto and Toshihide Maskawa (1973). “CP Violation in
the Renormalizable Theory of Weak Interaction”. In: Prog. Theor.
Phys. 49, pp. 652–657. doi: 10.1143/PTP.49.652 (cit. on p. 73).
Langacker, Paul (2017). The Standard Model and Beyond. Taylor &
Francis. doi: 10.1201/b22175 (cit. on p. 4).
Leader, E. and E. Predazzi (1996). An Introduction to gauge
theories and modern particle physics. Vol. 1. Cambridge University
Press. doi: 10.1017/CBO9780511622595 (cit. on p. 4).
References VII

Lehmann, H., K. Symanzik, and W. Zimmermann (1955). “On the

formulation of quantized field theories”. In: Nuovo Cim. 1,
pp. 205–225. doi: 10.1007/BF02731765 (cit. on p. 23).
Nambu, Yoichiro (1960). “Quasiparticles and Gauge Invariance in the
Theory of Superconductivity”. In: Phys. Rev. 117. Ed. by J. C. Taylor,
pp. 648–663. doi: 10.1103/PhysRev.117.648 (cit. on p. 53).
Pich, A. (1995). “Chiral perturbation theory”. In: doi:
10.1088/0034-4885/58/6/001. arXiv: hep-ph/9502366 (cit. on
p. 5).
Pich, Antonio (2012). “The Standard Model of Electroweak
Interactions”. In: arXiv: 1201.0537 [hep-ph] (cit. on p. 5).
Pokorski, S. (2000). Gauge Field Theores. Cambridge University
Press. doi: 10.1017/CBO9780511612343 (cit. on p. 4).
References VIII

Ramond, Pierre (1999). Journeys beyond the standard model.

Perseus Books (cit. on p. 4).
Salam, Abdus (1968). “Weak and Electromagnetic Interactions”. In:
Conf. Proc. C 680519, pp. 367–377. doi:
10.1142/9789812795915_0034 (cit. on p. 7).
Salam, G. P. (2020). “Elements of QCD for hadron colliders”. In:
CERN Yellow Rep. School Proc. 5. Ed. by M. Mulders and
J. Trân Thanh Vân, pp. 1–56. doi: 10.23730/CYRSP-2020-005.1
(cit. on pp. 6, 131).
Schwartz, Matthew D. (2013). Quantum Field Theory and the
Standard Model. Cambridge University Press. doi:
10.1017/9781139540940 (cit. on p. 4).
References IX

Skands, Peter (2013). “Introduction to QCD”. In: Theoretical

Advanced Study Institute in Elementary Particle Physics: Searching
for New Physics at Small and Large Scales, pp. 341–420. doi:
10.1142/9789814525220_0008. arXiv: 1207.2389 [hep-ph]
(cit. on p. 6).
Slavnov, A. A. (1972). “Ward Identities in Gauge Theories”. In:
Theor. Math. Phys. 10, pp. 99–107. doi: 10.1007/BF01090719
(cit. on p. 30).
Taylor, J. C. (1971). “Ward Identities and Charge Renormalization
of the Yang-Mills Field”. In: Nucl. Phys. B 33, pp. 436–444. doi:
10.1016/0550-3213(71)90297-5 (cit. on p. 30).
Weinberg, Steven (1967). “A Model of Leptons”. In: Phys. Rev.
Lett. 19, pp. 1264–1266. doi: 10.1103/PhysRevLett.19.1264
(cit. on p. 7).
References X

Yang, Chen-Ning and Robert L. Mills (1954). “Conservation of

Isotopic Spin and Isotopic Gauge Invariance”. In: Phys. Rev. 96.
Ed. by Jong-Ping Hsu and D. Fine, pp. 191–195. doi:
10.1103/PhysRev.96.191 (cit. on p. 20).