Добавил:
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:

Supersymmetry. Theory, Experiment, and Cosmology

.pdf
Скачиваний:
79
Добавлен:
01.05.2014
Размер:
12.6 Mб
Скачать

Spontaneous breaking of symmetry 379

F is bounded and reaches a maximum for some element gˆ(x). Using δg = agta, we have

0 = δF g, x) = a (φ(x), gˆ(x)taφ0) = a gˆ1(x)φ(x), taφ0 .

(A.88)

ˆ 1 α

Hence φ(x) gˆ (x)φ(x) is orthogonal to the directions θ φ0 in field space: it parametrizes the (in general) massive excitations.

Since any element of G can be written as eαθα eiti and since gˆ(x) is in fact a

right coset, we may choose to represent it by

 

 

gˆ(x) = eα(x) θα .

(A.89)

Correspondingly, the field φ(x) is parametrized as

 

ˆ

α(x) θα ˆ

(A.90)

φ(x) = gˆ(x)φ(x) = e

φ(x),

which generalizes (A.70): the (dim G − dim H) fields γα(x) are the Goldstone bosons. We note [82] that they transform nonlinearly under G: γα → γα (x) defined by

g eα(x) θα = eα (x) θα ei(γα,g) ti .]

(A.91)

A.2.4 Spontaneous breaking of a local symmetry. Higgs mechanism

The conclusions that we have reached in the preceding section, especially the Goldstone theorem, are drastically changed when we turn to local transformations. Let us illustrate this on the case of a complex scalar field minimally coupled to an abelian gauge symmetry, a theory known as scalar electrodynamics. The Lagrangian reads

L = 41 F µν Fµν + DµφDµφ − V (φφ)

(A.92)

with Dµφ = µφ − igqAµφ and V (φφ) given by (A.61) with a = −m2. The scalar field kinetic term thus yields a term quadratic in the vector field:

DµφDµφ = µφµφ − igq φ∂µφ− φµφ Aµ + g2q2φφAµAµ.

 

(A.93)

Expressing the scalar field around its vacuum expectation value φ0 = e0

m2

as in

2

 

φ0A

µ

Aµ.

(A.67) thus yields what seems to be a mass term for the vector field: g2q"φ0

 

Let us pause for a while to discuss massive vector bosons. The free field action is

S =

d4x −

1

F µν Fµν +

1

M 2AµAµ

.

(A.94)

4

 

2

Gauge invariance is only recovered in the limit of vanishing mass M . The corresponding equation of motion is the Proca equation:

µF µν + M 2Aν = 0.

(A.95)

380 A review of the Standard Model and of various notions of quantum field theory

Acting with a derivative ν and using the antisymmetry of F µν , we infer from the Proca equation the Lorentz condition ν Aν = 0. Thus the massive vector field has 4 1 = 3 degrees of freedom: more precisely two transverse (as the massless photon) and one longitudinal. Using the Lorentz condition, we may finally write the Proca equation

( + M 2)Aν = 0,

(A.96)

which shows that, indeed, M is the mass of the gauge field.

Returning to our case of interest, we deduce that, once the gauge symmetry is

spontaneously broken (φ0 = 0), the gauge field acquires a mass

 

MA2 = 2g2q2φ0φ0

= 2g2q2

m2

.

(A.97)

 

 

 

λ

 

There seems, however, to be a discrepancy when we count the number of degrees of freedom. Using the gauge invariance built in the Lagrangian, we count two scalar degrees of freedom and two vector (transverse) degrees of freedom. However, once one translates the scalar field around its vev φ0 = 0, one seems to identify:

a massless Goldstone boson ϕ2 and a real scalar ϕ1 of mass 2m2 (as in Section A.2.1);

a vector field of mass MA, with three degrees of freedom (two transverse and one longitudinal).

It turns out that making the symmetry local has changed one of the scalar fields into a spurious degree of freedom. This is most easily seen by using the nonlinear parametrization (A.70) of the scalar degrees of freedom. We see that, in the local case, one can gauge away the field γ(x) by performing a local gauge transformation:

φ(x) → φ (x) = e−iqθ(x)φ(x) = ρ(x) +

m2

e0

(A.98)

λ

with the choice θ(x) = γ(x)/q. In other words, because the symmetry is local, the would be Goldstone boson is now a gauge artifact. It reappears in the theory as the longitudinal component of the massive vector particle. This is the famous Higgs mechanism.

The choice (A.98) corresponds to a gauge choice known as the unitary gauge: with this choice of gauge, all fields are physical and the S-matrix is unitary, but gauge symmetry is no longer apparent. One often prefers covariant gauges which retain some of the gauge invariance (through the dependence on some gauge parameters, which should drop from physical results) but include the spurious boson.

We may generalize this analysis to nonabelian gauge symmetries and will illustrate it on the example of SU (2) which provides the basis for the Standard Model. We

introduce a complex scalar field Φ(x) = φ1(x) which transforms as a doublet under

φ2(x)

SU (2) (cf. (A.37))

Φ(x)

Φ (x) = e−iαa(x)ta

Φ(x).

(A.99)

 

 

 

 

The Standard Model of electroweak interactions 381

Then the full Lagrangian reads:

L

= DµΦD

Φ

V Φ)

1 F aµν F a

(A.100)

µ

 

 

4

µν

 

where Fµνa is given in (A.52), and the potential is given by

V Φ) =

m2ΦΦ + λΦ)2.

(A.101)

 

The ground state is reached for Φ0Φ0 = m2/(2λ). We choose to orient it along the second component of the doublet:

 

0

 

 

m2

 

2

 

 

 

 

Φ0 =

v

, v =

 

λ

.

(A.102)

 

 

 

 

 

 

Then, studying as before the quadratic terms AaµAcoming from the term DµΦDµΦ, one concludes that all three gauge fields Aaµ, a = 1, 2, 3, acquire a mass

MA =

gv

(A.103)

2

through the Higgs mechanism. This requires three longitudinal degrees of freedom.

Reexpressing the scalar field around its vev as Φ = Φ + Φ0, and inspecting the scalar

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

potential, one finds three would be Goldstone bosons: (φ1

+ φ1)/

 

2, (φj − φj )/(i 2)

with j = 1, 2. A more transparent parametrization

is the nonlinear one:

 

 

 

 

 

 

 

 

 

 

 

 

 

0

 

 

 

 

 

 

 

(A.104)

2

 

 

 

 

Φ(x) = ea(x)ta v + η(x)

.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Since V Φ) = V ([v + η]2/2), the three fields γa(x) are the would-be Goldstone bosons. But a gauge transformation (A.99) with αa(x) = γa(x) shows that they are spurious degrees of freedom.

A.3 The Standard Model of electroweak interactions

From the very beginning, electrodynamics has inspired the theory of weak interactions. In electrodynamics, a vector current of matter fields couples to the photon. Electromagnetic interactions through photon exchange is thus described by a current–current interaction with a photon propagator to account for the action at distance. In nuclear beta decay, which we write here at the level of nucleons n → p + e+ νe, the range seemed to be zero or very small and [156, 157] proposed to keep a current–current structure but to make it a local one (in order to account for the zero range):

GF

 

 

 

µ

 

 

 

 

H =

 

[p(x)γ

 

n(x)] [e(x)γµνe(x)] + h.c.

(A.105)

2

 

where the letters p, n, . . . denote the corresponding spinor fields (proton, neutron, . . .) and GF is the Fermi constant:

GF

105

(A.106)

mp2 ,

where mp is the proton mass.

382 A review of the Standard Model and of various notions of quantum field theory

It was soon realized that the interaction (A.105) does not account for all nuclear beta decays (in particular, the so-called Gamow–Teller transitions) and that other processes are of the same type: in particular, the constant involved in the muon decay µ+ → e+ + νe + νµ is basically the Fermi constant. Weak interactions were discovered to be universal.

It was during the 1950s and 1960s that the full structure of low energy weak interactions was unravelled. A key stage was the discovery of parity violation [269,382] which showed that one had to include an axial vector into the current–current interaction. This was complemented by the measurement of the neutrino helicity by [197].

We take this opportunity to recall that under parity a spinor transforms as:

P

ψ(x)

P

= η

γ0ψ(x)

(A.107)

 

p

 

 

where ηp is a phase and xµ ≡ xµ. One deduces

scalar (S) (x)ψ(x)P= ψ(x)ψ(x)

pseudoscalar (P) (x)γ5ψ(x)P= −ψ(x)γ5ψ(x)

vector (V) (x)γµψ(x)P= ψ(x)γµψ(x)

axial (A)

P

ψ

(x)γµγ5ψ(x)

P

=

ψ

(x)γ γ5

ψ(x).

(A.108)

 

 

 

 

µ

 

 

Finally, the V-A theory was proposed by [346] and by [162]. It reads for the beta decay

GF

 

 

 

µ

 

 

 

 

 

[p(x)γ

(gV

+ gA γ5) n(x)] [e(x)γµ(1 − γ5)νe(x)] + h.c.

(A.109)

H =

 

 

 

2

 

where gV 1 (the conservation of the vector current – the so-called CVC hypothesis – ensures that strong interaction do not renormalize this coupling) whereas gA /gV 1.262 ± 0.005 (an e ect of strong interactions). Taking into account other weak processes such as muon decay and turning o strong interactions, one may write the pure weak interaction Hamiltonian density (this time at the quark level) as:

 

H

=

G

F

J

Jµ + h.c.

 

 

2 µ

 

where

 

 

 

 

 

 

J

(x) = J(h)(x) + J( )(x)

 

µ

µ

µ

 

 

 

 

Jµ(h)(x) = u(x)γµ(1 − γ5)d(x) + c(x)γµ(1 − γ5)s(x) + · · ·

Jµ( )(x) = νe(x)γµ(1 − γ5)e(x) + νµ(x)γµ(1 − γ5)µ(x) + · · ·

(A.110)

(A.111)

The main problem with the theory of weak interactions described by (A.111) is that it is nonrenormalizable: the Fermi coupling GF has negative mass dimension. This has some undesirable consequences at high energies: cross-sections increase monotonically with the energy available in the center of mass. There is thus an energy where the cross

The Standard Model of electroweak interactions 383

Fig. A.3 Tree-level Fermi interaction for nuclear beta decay.

sections have become so large that they are no longer compatible with the unitarity of the S-matrix (which expresses the conservation of probability). At such an energy the fundamental theory must be modified.

Since this type of reasoning will be repeated for gravity (another nonrenormalizable theory where the coupling, Newton’s constant GN , has also mass dimension 2), let us sketch the argument more precisely.

If we consider the tree-level process depicted in Fig. A.3, then, on purely dimensional grounds, we may deduce that the corresponding cross-section σ (given by the square of

the amplitude, and thus proportional to G2 ) behaves as G2

E2, where E is the energy

F

F

 

available in the center of mass. But

conservation of probability

(expressed through the unitarity of the S-matrix) imposes that σ E2. Thus at energies E GF 1/2, the theory is no longer compatible with first principles and must be replaced by a more complete theory (which involves new fields whose exchange modifies the high energy behavior of the cross-section). This is the so-called unitarity limit that we consider in more details in Section 1.2.1 of Chapter 1.

This behavior of the cross-sections for tree-level processes may be related here to

the nonrenormalizable character of the theory. If we compute the one-loop (second order) amplitude given in Fig. A.4, we obtain a contribution of order G2F EdE (to be compared with the tree-level GF as in Fig. A.3) which is infinite if we assume that the theory remains valid up to arbitrarily large energies, i.e. to arbitrarily small distances. The renormalization procedure may take care of a finite number of such infinities but, in the case considered, new infinities appear at each order of perturbation theory and thus perturbative renormalization cannot help.

These considerations show that, because the coupling GF of the weak theory has mass dimension 2, the theory at energy E is characterized by the dimensionless combination5 GF E2. When E reaches a scale of order GF 1/2 300 GeV, the dimensionless combination is of order one and the low energy weakly coupled theory is no longer valid: it must be replaced by a more complete fundamental theory6.

Since the problem occurs at short distance (i.e. when the two vertices of the graph of Fig. A.4 become arbitrarily close), an obvious way of curing the problem is to replace

 

F

 

 

 

E2

 

2 /E2

whereas the ratio of

5For example, one may write the tree-level cross-section as

G

F

 

the one-loop amplitude to the tree-level amplitude reads G

 

E2

 

where E

max

is the cut-o in the

 

 

max

 

 

 

 

 

integral over the energy.

6The precise value of scale may be found by imposing the unitarity bound on varied processes. The best limit thus obtained is 630 GeV.

384 A review of the Standard Model and of various notions of quantum field theory

Fig. A.4 One-loop contribution to beta decay.

the contact interaction between the four fermions in the Fermi interaction by a vector particle exchange as in Fig. A.5, much in the spirit of quantum electrodynamics. This (charged) vector particle was introduced long before the days of the Standard Model and called the intermediate vector boson. It had to be be massive in order to account for the short range interaction. The above arguments showed that its mass was expected to be smaller than a few hundred GeVs. Once it was realized that a nonvanishing mass for a vector boson was compatible with gauge symmetry (although spontaneously broken), it was natural to try to fit weak interactions into a gauge theory. We will now see that the minimal nontrivial possibility leads to the Standard Model.

A.3.1 Identifying the gauge structure and quantum numbers

We will try here to fit the current–current theory of weak interactions just described with the structure of a gauge theory. In other words, pursuing the analogy with QED which motivated Fermi, we try to identify the intermediate vector boson W ± as a gauge field. Our guiding principle will be here minimality: there is no a priori reason that the theory of weak interactions should be the minimal one but, as we will see, it turns out to be the case.

From the graph of Fig. A.5 (and a similar one for muon beta decay), we infer that the couplings of the intermediate vector boson to the quarks u and d and the leptons e, νe, µ, νµ must be of the form

L

g u¯ γµ

1 − γ5

d W + + g d¯ γµ

1 − γ5

u W

 

2

2

 

 

 

 

µ

 

µ

 

 

+ g ν¯e

γµ

1 − γ5

e Wµ+ + g e¯ γµ

1 − γ5

νe Wµ

 

 

 

2

 

 

2

 

 

 

 

+ g ν¯µ

γµ

1 − γ5

µ Wµ+ + g µ¯ γµ

1 − γ5

νµ Wµ.

(A.112)

 

 

2

 

 

2

 

 

 

Since a charged field is complex, we need to introduce two real gauge fields; minimality leads to consider an abelian symmetry U (1) × U (1), with a global symmetry between the two U (1), or a nonabelian gauge symmetry SU (2). The first choice is very close to QED. Since the physics of weak interactions seems quite di erent from electrodynamics, we pursue the second possibility.

The Standard Model of electroweak interactions 385

Fig. A.5 Intermediate vector boson exchange.

Following (A.55), we may rewrite the interaction term between a SU (2) gauge field

a

ψ1

(x)

Aµ(x) and a doublet fermion Ψ(x) =

ψ2

(x) as

¯

µ

a σa

a

 

 

Lint = g Ψγ

 

Aµ

 

Ψ = g J

 

Aµ

 

 

 

2

 

 

 

g

 

 

 

 

 

 

 

 

 

 

g

 

 

=

 

J1µ + iJ2µ Aµ1 − iAµ2 +

 

J1µ − iJ2µ Aµ1 + iAµ2

2

2

where we have introduced the complex gauge fields:

 

 

 

 

 

 

 

 

 

 

Aµ1 iAµ2

 

W ±.

 

 

 

 

 

 

 

 

 

 

 

 

 

µ

 

 

 

 

 

 

 

 

 

 

 

2

The associated currents are J±µ ≡ J1µ ± iJ2µ and we have

 

 

 

 

 

Lint =

g

J+µ Wµ+ + Jµ Wµ+ g J3µAµ3 .

 

 

 

 

 

 

 

 

 

 

 

 

2

 

+ g J3µA3µ

(A.113)

(A.114)

(A.115)

µ

¯

γ

µ

µ

¯

γ

µ

ψ1

, comparison with (A.112) tells us how the

Since J+

= ψ1

 

ψ2 and J

= ψ2

 

quarks and leptons must be grouped into SU (2) doublets. Indeed since, for example

u¯ γµ

1 − γ5

d = u¯

1 + γ5

γµ

1 − γ5

d = u¯L γµ dL ,

 

2

 

2

 

2

 

we conclude that only left-handed components of quarks and leptons behave as SU (2) doublets:

uL (x)

νeL (x)

,

νuL (x)

Ψ(x) = dL (x)

, eL (x)

µL (x) .

The right-handed chiralities do not transform under SU (2): they are singlets under SU (2).

We note immediately that this local SU (2) symmetry has nothing to do with the global SU (2) isospin symmetry of strong interactions that Yang and Mills attempted to make local. Strong isospin is a global symmetry of strong interactions and it does not

apply to leptons for example. It is true however that u(x) form a strong interaction

d(x)

386 A review of the Standard Model and of various notions of quantum field theory

doublet (note the absence of chiral indices though!). This has let to the regrettable use of the expression “weak isospin” for the SU (2) symmetry that we consider here : weak isospin is a local symmetry of (electro)weak interactions and leptons transform under it.

It remains to identify the nature of the real field A3µ which completes the gauge multiplet. Since it cannot be paired to another real gauge field, it is obviously neutral. The natural candidate would be the photon but the corresponding charge operator can

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3

=

1 ¯

be written now that we have identified the content of the doublets (Jµ

2 ψ1γµψ1

1 ¯

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 ψ2γµψ2):

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Q3

d3x J3 = d3x

1

 

 

 

 

 

 

+ νν

 

 

 

 

+ νν

 

 

 

uu

L

dd

 

 

ee

 

µL

µµ

2

 

 

 

 

 

0

 

 

L

L

L

eL

eL

L

 

L

 

µL

L L

which has nothing to do with the electric charge operator7

 

 

 

 

 

¯

(A.116)

 

 

Q

 

 

 

qi the charge of fermion ψi):

 

 

 

 

 

 

 

 

 

 

 

 

 

(Jµ

= #i qiψiγµψi, with

 

 

 

 

 

 

 

 

2

 

 

1

 

 

 

 

 

 

 

 

 

 

 

 

Q ≡

d3x J0Q

=

d3x

 

uu −

 

dd − ee

− µµ .

 

 

(A.117)

 

 

3

3

 

 

However, we may note that

Y2(Q − Q3) = d3x&13 uL uL + dL dL + 43 uR uR 23 dR dR

νeL νeL + eL eL eR2eR

ν

µL

ν

µL

+ µµ

L

2µµ

.

(A.118)

 

 

L

R

R '

 

In other words, all the elements of a given representation of SU (2) – whether a doublet of two left-handed fields, say uL and dL , or a singlet right-handed field, say uR – appear in the sum with the same coe cient. This means that out of the SU (2) gauge symmetry associated with the generator t3 and the U (1) QED gauge symmetry, one may form an abelian U (1) symmetry which commutes with the weak isospin SU (2) symmetry. The associated quantum number is noted y and called the weak hypercharge ; the gauge group is noted U (1)Y to distinguish it from quantum electrodynamics. From (A.118), we infer the following relation between the charge q, the weak isospin t3 and the weak hypercharge y of a fermion

q = t3 +

y

.

(A.119)

 

2

 

 

Table A.1 gives the quantum numbers of the low energy quarks and leptons.

We have introduced a right-handed neutrino which did not appear in the earliest formulations of the Standard Model, since neutrino were assumed to be massless. As noted in Chapter 1, this field is a gauge singlet under SU (2) × U (1)Y .

7Indeed, if the photon field Aµ is one of the gauge fields of SU (2), then the corresponding generator is traceless and the electric charges in a given representation must add to zero. One then needs to introduce extra leptons (and quarks) as in the Georgi and Glashow model [180] (see Exercise 9).

φ0
−φ

The Standard Model of electroweak interactions 387

Table A.1

Field

q

t3

y

uL

+2/3 +1/2 +1/3

dL

1/3 1/2 +1/3

uR

+2/3

0

+4/3

dR

1/3

0

2/3

νeL

0

+1/2

1

eL

1

1/2

1

eR

1

0

2

NR

0

0

0

The total gauge symmetry has now four gauge bosons: Aaµ, a = 1, 2, 3, for SU (2) and the abelian gauge field of U (1)Y which we will note Bµ. It follows from our construction that the photon field Aµ (associated with the charge operator Q) is a combination of A3µ and Bµ. The orthogonal combination is a real, thus neutral, vector field denoted by Zµ0. Its exchange leads to new low energy weak interactions known as neutral currents. It is their discovery in 1973 at CERN which led to the first experimental verification of the Standard Model.

In order to give a mass to the intermediate vector boson Wµ± (and to the new Zµ0 since there does not seem to exist a long range force associated with it), we must break spontaneously the gauge symmetry, while keeping the photon Aµ massless. In other words, we must break SU (2) × U (1)Y down to the U (1) symmetry of QED. We need at least three real scalar fields to provide for the three longitudinal degrees of freedom

WL±, ZL0 . The most economical choice is provided by a complex doublet Φ =

 

φ

+

with

0

quantum numbers given in Table A.2.

φ

 

After spontaneous breaking, there remains 43 = 1 degree of freedom: the neutral scalar field known as Higgs and actively searched at high energy colliders.

For future reference, we note that, since Φ transforms under SU (2) as a doublet

Φ(x) → e−iαa(x)ta Φ(x).

(A.120)

Φ which describes the charge conjugate does not field

 

Φ =

Φ = 2

transform as a doublet. It is the

(A.121)

which transforms as (A.120), as can be seen by using

 

σ2

σa

σ2 =

σa

(A.122)

 

 

.

2

2

This shows that, under SU (2), the representation 2 is equivalent to the conjugate

¯ representation 2. We note also that Φ has hypercharge y = 1.

We are now in a position to write the Lagrangian of the Standard Model and extract its consequences.

388 A review of the Standard Model and of various notions of quantum field theory

Table A.2

Field

q

t3

y

φ+

+1

+1/2

+1

φ0

0

1/2

+1

A.3.2 The Glashow–Weinberg–Salam or SU(2) × U(1) model

We start with the gauge and scalar sector of the theory. The Lagrangian simply reads

L = 41 F aµν Fµνa 41 Bµν Bµν + DµΦDµΦ − V Φ)

(A.123)

where Fµνa (a = 1, 2, 3) is the SU (2) covariant field strength given in (A.52), Bµν = µBν − ∂ν Bµ is the U (1)Y invariant field strength and Φ is the Higgs doublet. Its derivative DµΦ is covariant under both SU (2) and U (1)Y :

 

σa

g

 

DµΦ = µΦ − ig Aµa

 

Φ − i

 

yφ BµΦ,

(A.124)

2

2

where g is the SU (2) gauge coupling, g /2 is the U (1)Y gauge coupling and yφ = +1 is the Higgs hypercharge. This Lagrangian is invariant under local transformations of SU (2) × U (1)Y , i.e. infinitesimally, following (A.46), (A.31), (A.98) and (A.99),

1

 

 

 

 

 

δAµa =

 

 

 

µαa + εabc αb Aµc

 

g

 

2

 

 

 

 

δBµ =

 

µβ

 

g

 

 

 

 

 

 

σa

 

δΦ = −i αa

 

Φ − iyφβΦ

(A.125)

2

where αa, a = 1, 2, 3, are the parameters of the SU (2) transformation and β of the abelian U (1)Y . We note that the choice α1 = α2 = 0 and α3 = 2β = θ corresponds to the U (1) QED gauge transformation (A.29):

δφ+ = −iθφ+,

 

δφ0 = 0.

(A.126)

We take as usual a scalar potential

 

 

 

 

 

 

 

 

 

 

 

 

V Φ) =

m2ΦΦ + λΦ)2,

(A.127)

 

with the ground state8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0

 

 

 

 

 

 

 

 

 

 

 

 

v =

m2

 

 

Φ Φ0 =

v

 

 

,

 

.

(A.128)

 

 

λ

 

 

2

 

 

 

8We note that it is this choice of orientation which determines the exact nature of the residual symmetry, U (1)QED, and thus the electric charge of the fields. In order to remain simple in the presentation, we have worked backwards here.