An Introduction to the Standard Model
Transcription
An Introduction to the Standard Model
An Introduction to the Standard Model Gianluigi Fogli Dipartimento di Fisica & INFN - Bari Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 Outline 1. A critical approach to the Standard Model 2. The Electromagnetic Interaction 3. Weak Interactions 4. Electroweak Interactions 5. Gauge Symmetry 6. The Standard Model 7. Renormalization and Running Coupling Constants Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 2 Reference textbooks Main text of reference for these lectures: Francis Halzen and Alan D. Martin, “Quarks and Leptons: An Introductory Course in Modern Particle Physics” A more theoretical approach: Giovanni Costa and Gianluigi Fogli, “Symmetries and Group Theory in Particle Physics” A more experimental approach: Alessandro Bettini, “Introduction to Elementary Particle Physics” Texts on Quantum Field Theory: James D. Bjorken and Sidney Drell, “Relativistic Quantum Fields” Michael E. Peskin and Daniel V. Schroeder, “An Introduction to Quantum Field Theory” Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 3 1. A critical approach to the Standard Model Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 4 1. About Fundamental Interactions • Scope of Fundamental Physics: description all natural phenomena in terms of basic theoretical laws, able to reproduce and predict experimental observations. • At the microscopic level all natural phenomena understood in terms of three basic fundamental interactions Electromagnetic Weak Strong interactions • All these three interactions studied within the same framework characterized by Quantum Mechanics Special Relativity Gianluigi Fogli Local Relativistic Quantum Field Theory An Introduction to the Standard Model, Canfranc, July 2013 5 2. The Gauge Invariance Principle Each particle is considered point-like and is described by a field with specific transformation properties under the Lorentz Group. The properties of the three interactions are understood within a common, general principle, the Gauge Invariance Principle In other words, they are asked to satisfy a “gauge” symmetry invariance, which means that they are supposed to be invariant under phase transformations that rotate the basic internal degrees of freedom (internal quantum numbers), with rotation angles dependent on the specific space-time coordinates (Local Gauge Invariance). Field Theories with gauge symmetry in a 4-dimensional space-time are “renormalizable”, which means that they are completely determined in terms of the gauge group of symmetry and the representations of the interacting fields. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 6 3. “The Standard Model” Indeed, a complete description of the three interactions is possibile in terms of a well-defined gauge theory, characterized by 12 gauged non-commuting charges known as “The Standard Model” of the fundamental interactions. It is remarkable that at present only a subgroup of the Standard Model symmetry is reflected by the spectrum of the physical states. The part of the electroweak symmetry related to the Higgs mechanism and to the spontaneous symmetry breaking of the gauge symmetry is still hidden to us. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 7 4. Where is Gravity ? For all material bodies on the Earth and in all astrophysical and cosmological phenomena a fourth interaction, the gravitational interaction has an important role, however negligible in atomic and nuclear physics. The theory of General Relativity is the classical (non quantistic) description of gravity, going beyond the statical approximation described by the Newton Law and including dynamical phenomena (for example, gravitational waves). As one can easily understand, the problem of formulating a quantum theory of gravitational interactions is one of the central problems of the contemporary theoretical physics. On the other hand, quantum effects in gravity are expected to become important only for energy concentrations in space-time that in practice are not accessible to the experimentation in laboratories. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 8 5. The Planck Mass So, gravity has no measurable effects on a subatomic scale and no manifestations that can guide us to a quantum field theory of it. It is possible however to estimate the distance r at which the gravitational force between two particles become significant. Let us remind how we obtain a dimensionless measure of the strenght of electromagnetic interactions: we compare the electrostatic energy of repulsion between two electrons at a distance equal to the natural unit of lenght (the Compton wavelenght) with the rest mass energy of the electron: 1 e2 α = 4π ħ/mc mc2 1 e2 = 4π ħc 1 ~ ~ 137 4π adopting the rationalized Heaviside-Lorentz system: it reduces Maxwell eqs. to their simplest form and is usual in particle physics which represents the so-called fine structure constant. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 9 In a similar way, we can estimate that gravitational effects become of order 1 (the interaction becomes a strong interaction) when the masses of the two particles that interact gravitationally (m1 = m2 = M) are such that the gravitational potential is comparable to the rest mass energies of the particles, i.e. when GM2 r Mc2 ~ 1 for masses separated by a natural unit of lenght r = ħ Mc This gives Mc2 = ħc5 G 1/2 = 1.22 x 1019 GeV Planck Mass The Planck Mass is the only dimensional quantity appearing in gravity. It indicates the mass scale at which gravitational effects become significant. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 10 It is interesting to compare the Planck mass to the mass scale at present under observation in the LHC experiments at CERN: 7-14 TeV (later we will compare also with the mass scale coming from GUT’s, the Grand Unified Theories). Instead of considering the mass scale, we can look to the distance at which the interaction takes place, making use of the Heisemberg indetermination principle, p Δr ≥ ħ. In the case of the LHC, we are testing the interaction at distances of the order rLHC ≥ 10-18 cm On the basis of the experiments performed until now, we can say that down to distances of this order of magnitude the subatomic particles do not show an appreciable internal structure and behave as elementary and point-like. Using the Planck Mass to estimate the distances at which the quantum effects due to gravity become significant, we find rPM ≥ 10-33 cm Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 11 6. String theories At these distances, r ~ 10-33 cm, the particles so far appeared as point-like could well reveal a structure, like strings, and could require a completely different theoretical framework. A theoretical framework of which the local quantum field theory description of the Standard Model would be just a “low energy”/”large distance” limit. At present the most complete and plausible description of quantum gravity is a theory formulated in terms of non-point-like objects, called “strings”, extended over distances much shorter than those experimentally accessible, objects that live in a space-time with 10 or 11 dimensions. The additional dimensions beyond the usual 4 are, typically, compactified. In principle, string theories constitute an all-comprehensive framework that suggests a unified description of all interactions. However, at present they represent only a purely speculative framework, not accessible to experiment. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 12 7. More on implications of a description in terms of relativistic local fields The description in terms of relativistic local fields has important implications that we attempt to review here. Since the photon emerges in a natural way from the quantization of the Maxwell field Aµ(x), it is reasonable to ask whether also the other particles observed in nature, primarily the electron, are also related to force fields by the same quantization procedure. When this point of view is assumed, it becomes natural to associate with each kind of observed particles a field φ(x) which satisfies an assumed wave equation. The particle interpretation of the field φ(x) is obtained when we apply the canonical quantization program. An implication of such a program is that we are led to a theory with differential wave propagation. The fields φ(x) are continuous functions of continuous x and t, and the values of φ(x) at x are determined by properties of the fields infinitesimally close to x. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 13 For most wave fields (for example, sound waves or the vibrations of strings and membranes) such a description is an idealization which is valid for distances larger than the characteristic lenght which measures the granularity of the medium (the air for sound waves). For smaller distances these theories are then modified in a profound way. Apparently, the electromagnetic field is a notable exception. Indeed, until the special theory of relativity obviated the necessity of a mechanical interpretation, physicists made great efforts to discover evidence for such a mechanical description of the radiation field. After the requirement of an “ether” propagating light waves had been abandoned, it has been considerably less difficult to accept the same idea when the observed wave properties of the electron suggested the introduction of a new field ψ(x). From that moment, the present description of the subatomic world in terms of relativistic local fields follows. However, it is a gross and profound extrapolation of our present experimental knowledge to assume that a wave description successful at “large” distances (as we have seen, 10-18 cm at LHC) may be extended to distances an indefinite number of order of magnitude smaller. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 14 In the relativistic theory the assumption that the field description is correct in arbitrarily small space-time intervals leads – in perturbation theory – to divergent expressions for the electron self-energy and the “bare charge”. These divergence difficulties have been sidestepped with the renormalization theory. However, it is widely felt that the divergences, that enter through the estimate of the quantum effects of the theory, are symptomatic of a chronic desease in the small-distance behaviour of the theory. At this point it is legitime to ask why local field theories, that is theories of fields which can be described by differential laws of wave propagation, have been so extensively used and accepted. There are several reasons of that. An important reason is that with their aid we have found a significant agreement with observations. But the foremost reason is brutally simple. There exists no convincing form of a theory which avoids differential field equations. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 15 The Standard Model is indeed a theory based on a relativistic local field description. Because of the existence of creation and annihilation processes, it is at once a theory of the many-body problem. The prescription of a quantization strongly involves the existence of a Hamiltonian H. However, since it generates infinitesimal time displacements, we are led to a description with differential development in time. Lorentz invariance then requires a differential development in space as well. The notion of a Lorentz invariant microscopic description in terms of continuous coordinates x and t implies that the influence of the interaction should not propagate through space-time with velocity faster than c. This notion of “microscopic causality” strongly forces us again into the field concepts. Of course, a Hamiltonian may well not exist for a non-local “granular” theory: if it does not, the link connecting us with the quantization method of non-relativistic theories is fatally broken. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 16 There is no concrete experimental evidence of a “granularity” at small distances. There is likewise nothing but positive evidence that special relativity is correct in the high-energy domain, and, furthermore, there is positive evidence that the notion of microscopic causality is a correct hypothesis. Since there exists no alternative theory which is any more convincing, we are forced to restrict ourselves to the formalism of relativistic local causal fields. It is undoubtedly true that a modified theory must have local field theory as an appropriate large-distance approximation or correspondence. However, we again emphasize that the formalism of the Standard Model may well describe only the large-distance limit (at present, distances ≥ 10-18 cm) of a physical world of considerably different submicroscopic properties. But let us now abandon these speculations and go to the concrete formulation of the Standard Model. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 17 2. The electromagnetic interaction Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 18 1. A spinless electron in an electromagnetic field A free “spinless” electron satisfies the Klein-Gordon equation (∂µ∂µ + m2 ) φ(x) = 0 which is nothing else that the relativistic energy-momentum relation E2 + → p2 = m2 with the introduction of the differential quantum operators pµ → iħ ∂ . ∂xµ In classical electrodynamics the motion of a particle of charge –e in an electromagnetic potential Aµ is obtained by the substitution pµ → pµ + eAµ Gianluigi Fogli quantum-mechanically corresponding to i∂µ → i∂µ + eAµ An Introduction to the Standard Model, Canfranc, July 2013 19 so that the Klein-Gordon equation becomes (∂µ∂µ + m2 ) φ(x) = -Vφ(x) sign chosen according to the relative sign of kinetic energy and potential in the Schrodinger equation with V = -ie (∂µAµ + Aµ∂µ) – e2A2 e related to the coupling: in natural units e2 α = 4π In the usual non-relativistic perturbation theory the transition amplitude for the scattering of a “spinless” electron from a state φi to the state φf off an electromagnetic potential Aµ is given by Tfi = -i ∫ φf*(x) V(x) φi(x) d4x = i ∫ φf*(x) ie (∂µAµ + Aµ∂ ) φi(x) d4x µ Integrating by parts ∫ φf* ∂µ(Aµ φi) d4x = - ∫ ∂ (φf*) Aµ φi d4x µ Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 20 we obtain φi Tfi = -i ∫ jµfi(x) Aµ(x) d4x jµfi φf with jµfi = - ie [φf*∂µ(φi) - ∂µ(φf*) φi] electromagnetic current Aµ If the ingoing and outgoing electrons have momenta pi and pf, respectively, we can write, with Ni and Nf normalization constants, φi = Ni e-ip x i φf = Nf Gianluigi Fogli e-ipf x jµfi = - e NiNf (pi + pf)µ ei(pf-pi) x An Introduction to the Standard Model, Canfranc, July 2013 21 1. “Spinless” electron-muon scattering We are now able to calculate the scattering of the electron from another particle, for example a muon. The graph is simple: pA and pC are the initial and final momenta of the electron, pB and pD of the muon. jµ(1) epA epC The calculation is similar to the previous one: we have to identify Aµ with its source, the charged “spinless” muon of the lower vertex. The identification is performed through the Maxwell equations q Aµ = jµ(2) pB µ- pD jµ(2) Gianluigi Fogli with µ- From jµ(2) = - e NBND (pD + pB)µ ei(pD-pB) x eiqx = -q2eiqx An Introduction to the Standard Model, Canfranc, July 2013 22 it follows Aµ = - 12 jµ(2) q with q = pD - pB In conclusion Tfi = -i ∫ jµ (1) (x) - 12 q jµ(2) (x) d4x Using the currents introduced before we can write Tfi = -i NANBNCND (2π)4 δ(4)(pD+pc-pA-pB)M with M = ie (pA+pC) µ Gianluigi Fogli -i gµν q2 ie (pB+pD) ν An Introduction to the Standard Model, Canfranc, July 2013 invariant amplitude 23 The invariant amplitude is a relativistic quantity and represents the Feynman diagram of the process. It contains the dynamics of the interaction. In M we distinguish -i e- -i jµ(1) ie (pA+pC) µ gµν q q2 µ- Gianluigi Fogli e- ie (pB+pD) ν jν(2) µ- gµν q2 is the propagator of the photon (a spin 1 particle) exchanged between electron and muon. The photon is “virtual”, or off-mass-shell: the changed particle carries the quantum number of the photon, but not the mass. Each vertex factor contains the electromagnetic coupling e and the fourvector index of a current. In conclusion, the graph describes the exchange of a virtual photon between the two currents associated to the two particles, electron and muon. An Introduction to the Standard Model, Canfranc, July 2013 24 1. Electrodynamics of spin- 21 particles 1 It is relatively simple to extend the previous approach to spin 2 particles. A free electron of four-momentum pµ is described by a four-dimensional Dirac spinor ψ (x) = u(p) e-ipx which satisfies the Dirac equation (γµpµ ‒ m) ψ = 0 The substitution pµ → pµ + eAµ describes the electron in an electromagnetic field Aµ: we write the equation in the form (γµpµ ‒ m) ψ = γ0V ψ with γ0V = - e γµAµ where γ0 is introduced to make the equation relativistically invariant. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 25 We can repeat the previous procedure and calculate the scattering amplitude: Tfi = -i ∫ ψf†(x) V(x) ψi(x) d4x = ie ∫ ψf†(x) γµAµ ψi(x) d4x = -i with jµfi = -e ψf γµ ψi = - e uf γµ ui ei(p -p )x f ∫ jµfiAµ d4x ψ = ψ†γ0 where i Jµfi The vertex factor is now a 4x4 matrix in spin space 1 Jµfi 1 ui uf ie(pf+pi)µ ieγµ The Gordon decomposition of the current shows that the spin interacts via both its charge and its magnetic moment: uf γ µ ui = Gianluigi Fogli 1 uf (pf+pi)µ + iσµν(pf-pi)ν ui 2m with 1 2 electron σµν = i (γµγν – γνγµ) An Introduction to the Standard Model, Canfranc, July 2013 2 26 We can now estimate the scattering amplitude of the process e-µ- → e-µ- Tfi = -i ∫ jµ(1) (x) - 12 q jµ(2) (x) d4x which implies Tfi = -i (- e uC γµ uA) - 12 q (- e uD γµ uB) (2π)4 δ(4) (pA+pB-pC-pD) so that the invariant amplitude is -iM = (ie uC γµ uA) -i gµν (ie uD γν uB) 2 q e- gµν -i 2 q µ- Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 jµ(1) e- ieγµ ieγν jν(2) µ- 27 Now we have to estimate |M|2 for all possible spin configurations, which means average over the spins of the incoming particles and sum over the spins of the particles in the final state: |M|2 → |M|2 = 1 (2sA+1)(2sB+1) ∑ all spin states |M|2 Introducing explicitly the particle momenta M = - e2 u(k’) γµ u(k) 12 u(p’) γµ u(p) q e- e- k we can write |M|2 = e4 q µνL muon L µν 4 e where Leµν = 1 ∑ [ u(k’) γµ u(k)][u(k’) γν u(k)]* 2 e spins gµν -i 2 q p µ- k’ ieγµ q ieγν P’ µ- with a similar expression for Lmuon µν . Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 28 The spin summation requires some care, but can be done with some trace techniques. First of all, let us write (using γν γ0 = γ0γν) [u(k’) γν u(k)] * = [u(k’) γ0 γν u(k)] = [u(k) γν γ0 u(k’)] = [u(k)γνu(k’)] The sum in Leµν can now be done. Writing explicitly the indeces and using the completeness relations of the Dirac spinors (m being the electron mass) Leµν = (s’) (s) (s) (s’) 1 ∑ uα (k’) γµαβ ∑ uβ (k) uγ (k) γνγδ uδ (k’) 2 s s’ (k’ + m)δα i.e. Leµν Gianluigi Fogli = (k + m)βγ 1 Tr [(k’ + m) γµ (k + m) γν ] 2 An Introduction to the Standard Model, Canfranc, July 2013 29 Traces are calculated making use of the commutation algebra of the γ matrices γµγν + γνγµ = 2gµν Accordingly Leµν = 1 Tr 2 (k’ γµ k γν ) + 1 m2 Tr 2 (γµ γν ) = 2 [k’µkν + k’νkµ – (k’ k – m2) gµν] • In similar way muon Lµν = 1 Tr 2 (p’γµ p γν ) + 1 2 M2 Tr (γµ γν ) = 2 [p’µpν + p’νpµ – (p’ p – M2) gµν] • So that, in conclusion |M|2 Gianluigi Fogli = e4 8 4 [(k’•p’)(k•p) + (k’•p)(k•p’) – m2(p’•p) – M2(k’•k) + 2m2M2] q An Introduction to the Standard Model, Canfranc, July 2013 30 3. Weak Interactions Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 31 1. The V-A structure of weak interactions Typical weak processes β decay n → p + e- + νe mean life 820 s µ- decay µ- → e- + νe + νµ τ = 2.2 x 10-6 s π- decay π- → µ- + νµ τ = 2.6x 10-8 s Fermi explanation of β-decay (1932) inspired by the structure of the electromagnetic interactions: the β-decay in its “crossed” form p + e- → n + νe no propagator: point-like int. described by the invariant amplitude M = G ( un γµ up) ( uν γµ ue) Fermi constant Gianluigi Fogli charged weak currents An Introduction to the Standard Model, Canfranc, July 2013 p jµ(1) e- n νe jµ(2) 32 2. Parity violation In 1956 Lee and Yang made a critical survey of all the weak interaction data and argued persuasively that parity was not conserved in weak interactions. The Wu experiment in the same year studied β-transition of polarized cobalt nuclei 60Co → 60Ni* + e- + νe z -νR and observed that the electron is emitted preferentially in the direction opposite to that of the spin of the 60Co nucleus. The observation is consistent with the explanation that the required Jz = 1 is formed by a right-handed antineutrino νR and a left-handed electron eL. Gianluigi Fogli + eLJz = 5 Jz = 4 Jz = 1 60Co 60Ni* - R + (e-)L+ (ν) An Introduction to the Standard Model, Canfranc, July 2013 33 The cumulative evidence of many experiments led to the conclusion that only νL and -νR are involved in weak interactions. The absence of νR and νL is a clear indication of parity violation: Γ(π+ → µ+νL) ≠ Γ(π+ → µ+νR) = 0 P violation Not only parity is maximally violated in weak interaction, but also charge conjugation, i. e. the interchange particle-antiparticle. Indeed, C transforms a νL into a νL so that Γ(π+ → µ+νL) ≠ Γ(π- → µ-νL) = 0 C violation However, the combination of the two symmetry operations is not violated, at least in principle, in weak interactions Γ(π+ → µ+νL) = Γ(π- → µ-νR) CP invariance We will discuss later the problem of CP violation in weak interactions. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 34 At the end, we arrive to the conclusion that weak interactions phenomena are described by a V-A current-current interaction with a universal coupling G. Accordingly for β- and µ-decay M( p → n e+νe) = G √2 [ un γµ (1-γ5) up][uν γµ(1-γ5)ue] M(µ- → e- νeνµ) = [ uνµ γµ (1-γ5) uµ][ue γµ (1-γ5) uν e] G √2 e Weak interactions are then of the general form M = 4G Jµ Jµ √2 Jµ = uν γµ 1 (1-γ5)ue with 2 Jµ = ue γµ 1 (1-γ5)uν 2 Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 charge-raising current charge-lowering current 35 3. Interpretation of the coupling G: the intermediate vector boson theory The comparison between electromagnetic and weak amplitudes shows that G essentially replaces e2/q2. Thus G has dimensions GeV-2. It is tempting to extend the analogy by assuming that weak interactions are characterized by the exchange of charged vector bosons, W±, with an amplitude of the form (for the µ-decay) -eνµ) = M(µ- → e- ν g g 1 uνµ γµ (1-γ5) uµ ue γµ (1-γ5) uνe 2 2 MW – q √2 √2 νµ where g/√2 is a dimensionless variable and q the momentum carried by the vector boson W±. g W± √2 g µ- √2 e- -e ν At the present level the introduction of W± simply leads to a reinterpretation of the Fermi constant G. This is the so-called “intermediate vector boson hypothesis”. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 36 In situations in which q2 « M2W (e.g. β-decay and µ-decay) we have G √2 = g2 8M2W and the propagator disappears: the interaction is point-like. From the previous equation we are induced to suspect that weak interactions are weak since M2W is large, while it is reasonable to expect g ≈ e. With the previous structure of the invariant amplitude, a large number of processes can be explicitly calculated and compared with the experimental measurements. Let us only mention an important experimental results, concerning the constant G as measured in µ-decay and in β-decay. One obtains Gµ = (1.16632 ± 0.00002) × 10-5 GeV-2 Gβ = (1.136 ± 0.003) × 10-5 GeV-2 The reason of this difference is important and will be discussed later. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 37 3. Weak neutral currents Detection in 1973 of neutrino events of the type - νµ e - → - νµ e - clean events, but small cross-section νµ N → νµ X - νµ N deep-inelastic scattering (DIS) → - νµ X ! neutral currents A quantitative comparison with the strenght of neutral currents (NC) to charged currents (CC) weak processes was performed. For example, for DIS processes the following experimental values were obtained Rν ≡ R-ν ≡ Gianluigi Fogli σNC(ν) σCC(ν) σNC(ν) - σCC(ν) ≡ ≡ σ (νµ N → νµ X) σ (νµ N → µ- X) -µ N → -νµ X) σ (ν -µ N → µ+ X) σ (ν = 0.31 ± 0.01 = 0.38 ± 0.02 An Introduction to the Standard Model, Canfranc, July 2013 38 Data can be understood in terms of neutral current-current interactions of amplitude (at the quark level) M= GN √2 [ uν γµ (1- γ5) uν] [uq γµ (cVq - cAq γ5)uq] GN cV cA new parameters The conventional normalization of the weak neutral current is then of the type NC µ = 4GN 2 JNC JNC µ M = 4G 2ρ JNC J µ µ √2 √2 Jµ NC (ν) = 1 2 with Jµ NC (q) = -uν γµ 1 (1- γ ) u 5 ν 2 q q u-q γµ 1 (cV - cAγ5) uq 2 Neutral currents, unlike charged currents, are not pure V-A (cV ≠ cA): they have right-handed components. However the neutrino is left-handed with cV=cA=1/2. As we will see, in the Standard Model all the couplings cV, cA are all given in terms of only one parameter, sin2θW. The parameter ρ “measures” the relative strenght of neutral to charged current processes. In the minimal version of the Standard Model (MSM) ρ = 1. This value is confirmed by the experiment within the experimental errors. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 39 4. The Cabibbo angle We have seen that charged currents are constructed considering transition between states coupled in the following pairs of (left-handed) fermionic doublets: νe e- u d νµ µ- They are coupled with the same universal coupling G. However, there are also processes that implies transitions u → s. For example the decay u K+ → µ+ νµ µ+ K+ s similarly to u µ+ d νµ π+ νµ In 1963 Cabibbo suggests to accomodate observation as K+ decay maintaining universality but modifying the quark doublets: charged currents couple “rotated” quark states as u u d’ = d cosθc + s sinθc with d' s' s’ = -d sinθ + s cosθ c Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 c 40 The quarks d and s are mixed in terms of the arbitrary parameter θc, the Cabibbo angle, to be measured experimentally. This gives rise to a comparison of ΔS = 0 and ΔS = 1 processes, with a suppression ~ sin2θc of the decay of strange particles (being θc ≈ 13°). The introduction of Cabibbo angle leads to a new form of charged currents, always in the form M = 4G Jµ Jµ √2 CC but with now - ) 1 γµ (1- γ5) U d Jµ = ( u c s 2 being U = cosθc sinθc - sinθc cosθc It is now clear the reason of the small difference between Gµ and Gβ: Gβ includes the cosine of the Cabibbo angle. The relation is Gβ = Gµ cosθc Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 41 4. The GIM Mechanism The original motivation of the proposal of the GIM mechanism has been to understand why there are no transitions s ⇔ d, which change flavor but not charge. Indeed, the experimental evidence for the absence of strangeness-changing neutral currents is compelling. Absent or strongly suppressed are decays of strange particles as K0 → µ+µ- K+ → π+e+e- K+ → π+νν This leads to the conclusions that direct transitions s ⇔ d d µ+ s µ- K0 are forbidden. When the GIM mechanism has been proposed (1970), only the three lightest quarks (u, d, s) were supposed to exist. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 42 However, even excluding direct neutral current transitions s ⇔ d, the K0 → µ+µdecay would occur at a rate far in excess of what is observed Γ(K0 → µ+µ-) Γ(K0 → all) = (9.1 ± 1.9) × 10-9 because of charged current u ⇔ d’ transition due to the box diagram in which the transition is mediated by a quark u. d u s µ- νµ W µ+ M ~ cosθcsinθc The GIM proposal was to introduce a fourth quark c, of charge 2/3, such that a second box diagram occurs, which would cancel exactly the first if it were not for the mass difference between c and u. d The existence of the quark c was predicted together with an estimate of its mass, in good agreement with the next estimates from experiment. s Gianluigi Fogli W W c An Introduction to the Standard Model, Canfranc, July 2013 µ- νµ W µ+ M ~ -sinθccosθc 43 It is important to understand how the GIM mechanism works, as an effect of the unitarity of the matrix U which describes the quark rotation in the charged current structure d’ s’ = cosθc -sinθc sinθc cosθc d s that we rewrite synthetically d’i = ∑j Uijdj - Starting from neutral currents of the form d’id’i and summing up we have - - - ∑i d’id’i = ∑ijk djU jiUikdk = ∑j djdj which shows that only diagonal transitions are allowed. Q: Why the mixing is taken in the down-type quarks? A: The reason is historical. The mixing could equally be taken in the up-type quarks. Q: Why no Cabibbo mixing in the leptonic sector? A: The reason is the mass degeneracy of neutrinos, that are massless in the SM. Indeed, as well known, there is a mixing also in the leptonic sector. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 44 4. The Kobayashi Maskawa matrix As well-known, there exists a third generation of leptons and quarks, so that we have to consider three quark doublets c t u d b s with their mixing. In the case of two doublets the mixing is described in terms on only one parameter, the Cabibbo mixing angle θc. How many parameters we have in the general case of N doublets? The number of parameter (angles and phases) of a N×N unitary matrix U is N2. We can subtract (2N-1) phases since we can absorb them in the 2N quark fields, apart from 1 overall phase. We can also subtract the number of parameters (angles) that characterize an orthogonal N × N matrix. In conclusion we have N2 – (2N - 1) - Gianluigi Fogli 1 2 N (N – 1) = 1 (N 2 – 1)(N – 2) An Introduction to the Standard Model, Canfranc, July 2013 residual phases 45 There are no phases for N=2, but there is a phase if N=3. Then, for three generations we have 3 angles and 1 phase factor, eiδ. We rewrite charged currents in terms of all the three quark doublets - -- Jµ = ( u c t ) 1 γµ (1- γ ) 5 2 U d s b The unitary matrix U describes the charged currents transitions according to Uud Uus Uub Ucd Ucs Ucb Utd Uts Utb c1 = -s1c3 -s1s3 s1c2 c1c2c3-s2s3eiδ c1c2s3-s2c3eiδ s1s2 c1s2c3+c2s3eiδ c1s2s3+c2c3eiδ where one of the most usual parametrizations is also indicated (with ci, si cosine and sine of the angles θ1, θ2, θ3). Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 46 5. CP violation An analysis of CP invariance requires the comparison of the amplitudes of a generic process and of the process conjugate under CP. Let us consider the charged current process ab → cd. The invariant amplitude is given by a M ∼ Jca Jµ bd ∼ uc γµ (1- γ5) Uca ua u-b γµ (1- γ5) Ubd ud = c W µ = Uca Udb* uc γµ (1- γ5) ua ud γµ (1- γ5) ub Uca b Udb * d On the basis of the correspondence between particle solutions going backward in time and antiparticle solutions going forward in time, the previous amplitude describes both the two processes ab → cd Gianluigi Fogli -- -- cd → ab An Introduction to the Standard Model, Canfranc, July 2013 47 Let us now consider the process -ab- → -cdIts amplitude is obtained taking the conjugate currents, so that it corresponds to M ∼ µ Jca µ Jµ bd ∼ U * U ca db ua γ (1- γ5) uc ub γµ (1- γ5) ud -a Uca* -c W - b - Udb d This process is described by the same Hamiltonian H: indeed H is hermitian, so it contains M + M . In order to understand if the process ab → cd is invariant under CP, we have to calculate the process conjugate under CP and compare its invariant amplitude, MCP, with M . If MCP = M occurs, then the process ab → cd is CP invariant. Otherwise CP is violated. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 48 We omit the explicit calculations, essentially based on the transformation properties under C and P of Dirac spinors and gamma matrices. We find MCP ∼ Uca Udb* ua γµ (1- γ5) uc ub γµ (1- γ5) ud The comparison shows that, provided that the elements of the mixing matrix U are real, we find MCP = M This is the case in which U is a 2×2 matrix and only 4 quarks (u, d, c, s) exist. If however we have three generations of quarks, the mixing matrix becomes the 3×3 matrix of Kobayashi-Maskawa. It now contains a complex phase eiδ. Then in general it is MCP ≠ M and CP invariance is violated. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 49 4. Electroweak Interactions Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 50 1. Weak Isospin and Hypercharge The Standard Model leads to a description of fundamental interactions in terms of a gauge invariant renormalizable theory. A first step in this direction is to find within the weak interaction phenomenology an underlying symmetry group. Let us start from the charged currents written in terms of the left-handed fields νe Jµ = Jµ(+) = uν γµ 1 (1-γ5) ue = ν γµ 1 (1-γ5) e = νL γµ eL W+ 2 2 e- Jµ = Jµ(-) = u-e γµ 1 2 - γµ (1-γ5) uν = e 1 2 (1-γ5) ν = eL γµ νL e- W- νe By introducing the doublet Gianluigi Fogli χL = νe e- L and the usual “step-up” and “step-down” operators An Introduction to the Standard Model, Canfranc, July 2013 τ± = 1 2 (τ1 ± iτ2) 51 the two currents can be rewritten in a two-dimensional form -Lγµτ+χL Jµ(+) (x) = χ Jµ(-) (x) = χLγµτ-χL If now we add a third current in the form of a neutral current Jµ(3) (x) = -χLγµ 1τ χ 2 3 L = 1 2 -νLγµνL - 1 -eLγµeL 2 e- (νe) W0 e- (νe) we have thus constructed an “isospin” triplet of weak currents Jµ(i) (x) = χLγµ 1τ χ 2 i L (i = 1, 2, 3) whose corresponding “charges” T i = ∫ J0(i)(x) d3x satisfy an SU(2)L algebra [Ti,Tj] = ieijkTk and can be assumed as generators of a new quantum number, the “Weak Isospin”. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 52 Question: can Jµ(3) (x) be identified with the weak neutral current introduced in the study of the Deep Inelastic Scattering νµ N → νµ X ? The answer is negative, since Jµ(3) (x) is a pure left-handed current, whereas the observed phenomenological current contains a (small) right-handed component. However, we know a current which contains a right-handed component, the vectorial electromagnetic current, which contains a right- as well a left-handed component. Taking apart, for simplicity, the multiplicative factor e (the electric charge), we can write the current - µe = - -eLγµeL - -eRγµeR jµem(x) = - eγ so that the current appearing in the electromagnetic interaction cam be written as - µQψ jµ = e jµem = e ψγ where Q is the electric charge generator, with eigenvalue Q = -1 for the electron. In other words, Q is the generator of the U(1)em symmetry group of the electromagnetic interactions. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 53 At this point, we have two symmetry groups SU(2)L U(1)em with generators T i and currents Jµ(i) with generator Q and current jµem But we have also a phenomenological neutral currents JµNC which needs to be interpreted in terms of them. The two currents JµNC and jµem do not “respect” the SU(2)L symmetry, but we can expect that this is the case for a suitable linear combination of them. This combination is a neutral current that can be identified with the member Jµ(3) of the isospin triplet. The corresponding orthogonal combination is independent of SU(2)L, i.e. it must be an isospin singlet under SU(2)L and can be connected to the generator of a new U(1) group, whose generator is a linear combination of T(3) and Q. The explicit form of this linear combination of T(3) and Q is quite arbitrary. We can adopt the same relation which characterizes the third component of the isotopic spin and the electric charge in the definition of the hypercharge through the Gell Mann-Nishijima scheme (i.e. the scheme adopted for the arrangement of the strange particles in the SU(2)I symmetry of the hadronic isospin multiplets). Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 54 So, we adopt the well-known Gell Mann-Nishijima relation Q = T (3) + Y 2 to connect the third component of the “weak isospin” to a new generator, the “weak hypercharge” Y. Going to the currents associated to these generators, we can write jµem = Jµ(3) + 1 jµY 2 The hypercharge generator Y can be taken as the generator of a new abelian group U(1)Y, so that the complete symmetry group is now enlarged to SU(2)L ⊗ U(1)Y U(1)em will appear as a subgroup of it, to be properly identified as the subgroup whose generator is the electric charge Q. Note that we do not have a simple group, but the product of two groups, which means that we need to introduce, in addition to the electric charge, another coupling constant. Or, which is equivalent, we have to introduce two couplings, one for SU(2)L and one for U(1)Y, by identifying the electric charge as a proper combination of them. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 55 The SU(2)L ⊗ U(1)Y proposal was first made by Glashow in 1961, long before the discovery of the weak neutral currents, and, as we will see later, was extended to accomodate massive vector bosons (W±, Z0) by Weinberg in 1967 and Salam in 1968. This is the Standard Model of the electroweak interactions. Since we have to do with the product of two symmetry groups, the generator Y must commute with the generators T(i). Accordingly, all the members of an isospin multiplet must have the same value of the hypercharge. For example, for the electron doublet (ν e-) it is jµY = 2jµem + 2Jµ(3) = -2 (e-RγµeR + e-LγµeL) – (ν-LγµνL – e-LγµeL) = -2 e-RγµeR – 1 χ-LγµχL so that the left-handed doublet (ν e-)L has hypercharge -1 and the isospin singlet eR has hypercharge -2. We are now able to provide the assignment of weak isospin and hypercharge to leptons and quark of the first generation. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 56 Weak isospin and hypercharge assignment to leptons and quarks of the first generation. Particle T T(3) Q Y νe 1 2 1 2 0 -1 e-L 1 2 1 2 -1 -1 e-R 0 0 -1 -2 uL 1 2 1 2 2 3 1 3 dL 1 2 1 3 uR 0 0 2 3 1 3 4 3 dR 0 0 1 3 - 23 Gianluigi Fogli - - 1 2 - An Introduction to the Standard Model, Canfranc, July 2013 57 2. The basic electroweak interaction We can now modify the current-current form of the weak interactions assuming that it corresponds to an effective interaction resulting from the exchange of massive vector bosons (in the limit of small momentum transfer). Exactly in the same way used to develop QED from the basic interaction - i e jµem Aµ coupling current vector boson We can introduce an isotriplet of vector bosons Wµ(i) and a single vector boson Bµ, with coupling g and g’, respectively, to describe the SU(2)L ⊗ U(1)Y interaction, which takes the form - i g Jµ(i) Wµ(i) - - i g’ Wµ(i) Gianluigi Fogli g Jµ(i) 1 2 j µY B µ Bµ 1 2 g’ An Introduction to the Standard Model, Canfranc, July 2013 j µY 58 The electromagnetic interaction, described in terms of the photon Aµ, and the phenomenological neutral current interaction, described in terms of the vector boson Zµ, are embedded in the previous form of the electroweak interaction. We have only to extract them explicitly, taking into account that they must appear as two orthogonal combinations of the neutral vector bosons associated to Jµ(3) and jµY. Accordingly, we can introduce the physical states (the mass eigenstates) in the form Aµ = Bµ cosθw + Wµ(3) sinθw (massless) Zµ = - Bµ sinθw + Wµ(3) cosθw (massive) The parameter θw, the Weinberg angle, “measures” the mixing. It is a phenomenological parameter, to be determined experimentally. Introducing explicitly Aµ and Zµ in the structure of the electroweak interaction we obtain - i g Jµ(3) Wµ(3) - i g’ 1 Y µ j B 2 µ -i Gianluigi Fogli 1 2 ( g cosθw Jµ(3) - 21 = - i ( g sinθw Jµ(3) + g’ cosθw jµY ) Aµ + g’ sinθw jµY ) Zµ An Introduction to the Standard Model, Canfranc, July 2013 59 The first term is to be identified with the electromagnetic interaction: using e jµem = e (Jµ(3) + 1 Y j ) 2 µ it follows the important relation between couplings and mixing angle g sinθw = g’ cosθw = e tg θw = g’ g The second term contains the neutral current JµNC. It can be rewritten 1 2 - i ( g cosθw Jµ(3) - g’ sinθw jµY ) Zµ = - i = -i g cosθw ( Jµ(3) - 1 2 g’ g sinθw cosθw jµY ) Zµ = g ( Jµ(3) – sin2θw jµem ) cosθw So, one identifies Jµ Gianluigi Fogli NC = Jµ (3) – sin2θ w jµ em coupled to Zµ with coupling An Introduction to the Standard Model, Canfranc, July 2013 Zµ = - i g JµNC cosθw Zµ g cosθw 60 3. The effective current-current interactions Let us recall the charged current invariant amplitude Mcc = 4G Jµ Jµ √2 CC with, in the isospin notation, -Lγµτ+χL Jµ = Jµ(+) (x) = χ On the other hand, introducing the charged vector bosons, we can rewrite the basic interaction in the form -i g √2 which leads to Mcc = Jµ Jµ Wµ(+) + Jµ Wµ(-) g √2 Jµ 1 M2w g √2 Jµ CC where 1 2 is the approximation of the W propagator at low q2. Mw Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 W(±) Jµ 61 Comparing the two expressions of Mcc we obtain G √2 = g2 8M2w In an analogous way for the neutral current interaction in terms of Z0 exchange MNC = g cosθw Jµ NC 1 M2Z g cosθw JµNC comparing with the corresponding current-current form 4GN MNC = 4G 2ρ Jµ NC JNC µ = 2 Jµ NC NC µ √2 √2 we can identify g2 G ρ = 2 2 8M z cos θw √2 JµNC Z0 JµNC By comparing the expressions for charged and neutral currents, we find for the parameter ρ, which measures the relative strenght of the two weak interactions ρ = M2w M2z cos2θw We shall see that the minimal version of the Standard Model (MSM) predicts ρ = 1. This value is confirmed experimentally within small errors. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 62 4. Feynman rules for electroweak interactions For the electromagnetic interaction - µQψ ) Aµ -i e jµem Aµ = -i e ( ψγ γ vertex factor f -i e Qf γµ - f For the charged current interaction -i ( χ-Lγµτ+χL ) Wµ(+) = -i √2 ( -νLγµeL ) Wµ(+) g g √2 W(+) e+ νe -i g √2 ( χ-Lγµτ-χL ) Wµ(-) = -i ( e-LγµνL ) Wµ(-) √2 g W(-) -i g ( Jµ(3) – sin2θw jµem ) Zµ = cosθw Z0 g ψfγµ 1 (1-γ5)T(3) – sin2θw Q ψf Zµ 2 cosθw Gianluigi Fogli e- g √2 γµ 1 (1-γ5) 2 -νe For the neutral current interaction -i -i An Introduction to the Standard Model, Canfranc, July 2013 f f -i g cosθw γµ 1 (c f-γ cf ) 2 V 5 A 63 It follows that all the couplings cV and cA are determined in the Standard Model in terms of only one parameter, sin2θw, being for each fermion f f f cA = T(3)f cV = T(3)f – 2 sin2θw Qf - The explicit values of the different Z → f f vertex factors are reported in the Table (sin2θw = 0.234) fermion Qf cAf cVf νe , νµ , … 0 1 2 1 2 e- , µ- , … -1 Gianluigi Fogli - 1 3 1 2 1 2 2 3 u,c,… d,s,… - - 1 2 - 1 + 2 sin2θw 2 ≈ - 0.03 1 - 4 sin2θw ≈ + 0.19 2 3 - 21 + 2 sin2θw ≈ - 0.34 An Introduction to the Standard Model, Canfranc, July 2013 3 64 A large number of experiments have been performed in the ‘80 to test the Standard Model, more specifically to measure sin2θw in different processes. Let us mention: Deep inelastic scattering Neutrino-electron scattering One-pion production from neutrino scattering Electroweak interference in e+e- annihilation Parity violating effects in atomic transitions Parity violating asymmetry in the inelastic scattering of longitudinally polarized electrons or muons In all these processes, a perfect agreement with the Standard Model has been found. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 65 5. Gauge symmetries Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 66 1. Lagrangians and single particle wave equation The fundamental belief is that all particle interactions may be dictated by the socalled “local gauge symmetries”. We shall see that this is intimately connected with the idea that conserved physical quantities (such as electric charge, color, …) are conserved in local regions of space and not just globally. The framework in which these principles are discussed is the Lagrangian Field Theory. It is well-known that in classical mechanics the particle equations of motion can be obtained from the Lagrange equations d dt ∂L . ∂ qi - ∂L = 0 ∂ qi with where L=T-V with qi generalized coordinates . qi = dqi/dt T kinetic energy V potential energy It is possible to extend the formalism to a continuous system, that is a system φ with continuously varying coordinates: φ( x , t). Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 67 The Lagrangian becomes a Lagrangian density . L ( qi, qi, t) → L ( φ, ∂φ , xµ) ∂xµ with L = ∫ L d 3x which satisfies the equation of motion ∂ ∂xµ ∂L ∂φ ∂ µ ∂x - ∂∂Lφ =0 What is the relation between the Lagrangian approach and the perturbative method based on Feynman rules? To each Lagrangian there corresponds a set of Feynman rules, so that, once we identify these rules, the connection is established. The identification proceeds as follows: We associate to the various terms in the Lagrangian a set of propagators and vertex factors. The propagators are determined by the terms quadratic in the fields, i.e. the - ψ, etc. terms in the Lagrangian containing for example φ2, ψ The other terms in L are associated to interaction vertices. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 68 Let us start from the Lagrangian - γµ∂µ ψ ‒ m ψ- ψ L = iψ The Euler-Lagrange equation is easily derived to be the Dirac equation of a free particle. The Lagrangian is clearly invariant under the phase transformation ψ(x) → eiα ψ(x) with α real constant The family of phase transformations U(α) = eiα forms a unitary Abelian group U(1) This invariance, through the Noether theorem, implies the existence of a conserved current and, at the integral level, of a conserved “generalized charge”, which acts as the generator of the group U(1). Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 69 By estimating δL and imposing the invariance in the form δL = 0, one easily obtain ∂ µj µ = 0 jµ(x) = ie with 1 2 ∂L - γµ ψ - ∂L ψ–ψ = e ψ ∂(∂µψ) ∂(∂µψ) where the proportionality factor has been chosen in such a way to match up the electromagnetic current of an electron of charge –e. From the conservation of the current one easily derives that Q= ∫ j0 (x) d3x is a conserved quantity, the electromagnetic charge. In conclusion, the existence of an invariance under this kind of transformation, which is referred to as a “global phase transformation”, leads to a conserved current and a conserved generalized charge. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 70 2. U(1) local gauge invariance and QED A more subtle, and more general, invariance can be invoked by requiring invariance even when α differs from space-time point to point, i.e. α = α(x), under the transformation ψ(x) → eiα(x) ψ(x) This is a “local gauge invariance” requirement. However, it is easy to verify that the Lagrangian is not invariant, because of the derivative term ∂µ ψ. In order to insist in the requirement of invariance, we have to modify the derivative term, going from the ordinary derivative to the so-called ∂µ → Dµ = ∂µ - ieAµ covariant derivative defined in terms of an arbitrary vector field Aµ(x), by requiring a well-defined transformation property for Aµ(x) Aµ → Aµ + 1e ∂µα(x) Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 71 It is easily checked that Dµψ → eiα(x) Dµψ so that the Lagrangian is invariant if expressed in terms of the covariant derivative. But with the introduction of the covariant derivative the Lagrangian becomes - ψ = -ψ (i γµ∂µ ‒ m ) ψ + e -ψγµψ Aµ L = iψ γµDµ ψ ‒ m ψ which is the Lagrangian of an electron in the presence of the electromagnetic field Aµ: in other words, the Lagrangian of QED. The result is to some extent rather surprising: the requirement of “local gauge invariance” applied to the Lagrangian of a free particle leads to a Lagrangian which describes the interaction of the particle with a vector field (gauge field) that we can interpret as the electromagnetic field Aµ. In other words, we can derive QED from the requirement of local gauge invariance. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 72 Regarding the new field Aµ as the physical photon field, we add to the Lagrangian a term corresponding to the kinetic energy of the photon field. It is usually expressed in terms of the gauge invariant field strenght tensor Fµν = ∂µ Aν - ∂ν Aµ Accordingly, the complete QED Lagrangian becomes L= ψ (i γµ∂µ ‒ m ) ψ + e ψγµψ Aµ - 1 4 FµνFµν It is important to observe that the addition of a mass term for Aµ 1 2 M2 AµAµ is prohibited by gauge invariance because of the assumed transformation properties of the field Aµ. The gauge vector boson must be massless. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 73 3. Non-Abelian gauge invariance and QCD We attempt now to extend the requirement of local gauge invariance to non-Abelian gauge groups. For example, we can take the SU(3) group of phase transformations on the quark color fields. The free Lagrangian is (for a single quark q) L0 = qi (i γµ∂µ ‒ m ) qi (i = 1, 2, 3) with q1, q2 , q3 the three color fields of a quark of given flavor. The phase transformations are of the form q(x) → U q(x) = eiα a (x)T a q(x) (a = 1, …, 8) with U arbitrary 3×3 unitary matrix. Ta are the eight generators of SU(3) in the three-dimensional representation (the Gell Mann matrices λa/2). They satisfy the commutation relations [Ta, Tb] = i fabc Tc Gianluigi Fogli fabc structure functions of SU(3) An Introduction to the Standard Model, Canfranc, July 2013 74 We can repeat the step of the Abelian case: by using infinitesimal phase transformations we have q(x) → [1+i αa(x)Ta] q(x) ∂µ q → (1+i αaTa) ∂µ q + i Ta q ∂µ αa We introduce the (eight) gauge fields Gµa transforming according to 1 Gµa → Gµa - g ∂µ αa (where g is the coupling) and form the covariant derivative Dµ = ∂µ + i g TaGµa Then we make the replacement ∂µ → Dµ in the Lagrangian and obtain - γµTa q) Gµa L0 → L = q (i γµ∂µ ‒ m ) q – g ( q Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 75 But for non-Abelian gauge transformations this does not appear an invariant Lagrangian, since - γµ (TaTb – TbTa) q → ( -q γµTa q) + fabcαb ( -q γµTc q) ( q- γµTa q) → ( q γµTa q) + iαb q as an effect of the commutator of the generators Ta. In order to obtain gauge invariance, we have to re-write the transformation properties of the gauge fields in the form 1 Gµa → Gµa - g ∂µ αa ‒ fabc ab Gµc Adding to L the gauge invariant kinetic energy term of the gauge fields we have for the QCD gauge invariant Lagrangian the final form a L= q (i γµ∂µ ‒ m ) q ‒ g ( q γµTa q) Gµa ‒ Gµν Gµν a with a a a b c Gµν = ∂µGν - ∂νGµ ‒ g fabcGµ Gν in order to satisfy the requirement of gauge invariance of the kinetic energy of the gauge fields. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 76 The QCD Lagrangian describes the interaction between colored quarks q and vector gluons Gµ with coupling g. Gauge invariance requires the eight gluons to be massless. a Note that the particular expression of the terms Gµν, due to the non-Abelian character of the gauge group, introduces, together with the kinetic energy terms, also a self-interaction effect between the gauge fields. In a symbolic form we can distinguish L = “qq” + “G2” + g ”qqG” + g ”G3” + g2 ”G4” qa qa qa gab ab gab gab gab qb qa gc a gc a gc a gacgc a gac gab gbc gbagab gba gab qb g bc The first three terms have their analogue in QED. They describe the free propagators of quarks and gluons and the quark-gluon interaction. The remaining two terms correspond to three and four gluon vertices and reflect the fact that gluons themselves carry color. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 77 4. Massive gauge bosons ? The gauge bosons are required to be massless in gauge invariant theories. How can we justify the existence of massive vector bosons, as they appear in weak and electroweak interactions? Of course, a possibility is to forget about gauge invariance and include the vector boson mass terms as well. However, the point is that the introduction or not of mass terms is not merely an aesthetic problem. The introduction of mass terms not only spoils gauge invariance, but also introduces unrenormalizable divergences that make the theory meaningless. This is essentially due to the more divergent behaviour of the propagators of massive vector bosons. They are of the form i Gianluigi Fogli - gµν + Q2 - qµ qν M2 M2 qµ qν ~ q2 → ∞ q2M2 An Introduction to the Standard Model, Canfranc, July 2013 78 This behaviour for q2 → ∞ no longer prevents the loop integrals ∫ d4q (propagator) …. from diverging for large loop momenta. Even the introduction of a cut-off does not work, since the inspection of the diagrams containing more loops shows that new, even more severe divergences appear in each order, and ultimately an infinite number of unknown parameters has to be introduced. The point is: it is possible to introduce masses without breaking gauge invariance? The answer is yes, it is possible through an intriguing procedure known as Higgs mechanism Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 79 5. Spontaneous Symmetry Breaking Let us consider the following simple Lagrangian describing the self interaction of a scalar particle associated to the field φ(x) L=T–V= 1 2 ∂µφ∂µφ ‒ ( 21 µ2φ2 + 41 λφ4) with λ > 0 The symmetry here is a discrete reflection symmetry: φ -φ We have two possible forms of the potential depending on the parameter µ2. The case µ2 > 0 is rather familiar. It describes a scalar field with mass µ in a confining potential. V(q) The ground state (“vacuum” in the quantum language) corresponds to φ = 0. The solution satisfies the reflection symmetry of the Lagrangian. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 q 80 The case µ2 < 0 is more interesting. The Lagrangian has a mass term with a wrong sign, moreover the potential has two minima. From V(q) ∂V = φ (µ2 + λφ4) = 0 ∂φ we see that they correspond to φ=±v with v= - √ µ2 -v v q λ The extremum φ = 0 is not a minimum. Perturbative calculations should involve expansion from a minimum of the potential, either φ = v or φ = - v. We have to choose one of them. Note that this does not implies a loss of generality, since the other minimum can be always reached through a symmetry operation: the symmetry is not lost. Let us choose φ0 = v and expand φ(x) around the minimum by writing φ(x) = v + η(x) Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 81 We perform the substitution of φ(x) in the Lagrangian (shift of the field) and discuss the new Lagrangian in terms of the new field η(x), whose minimum corresponds to η0 = 0. After a few calculations L’= 1 ∂ η ∂ µη 2 µ ‒ λ v 2 η2 ‒ λ v η3 - 1 4 λ η4 + const We see that the new field η(x) has a mass term of the right sign, with mη = √ 2λv2 = √-2µ2 whereas the higher order terms in η(x) represent the interaction of the field η(x) with itself. What about L and L ’ ? The two Lagrangians describe the same physics. If we were able to solve the two Lagrangians exactly, we would find the same results. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 82 But we are forced to use a perturbative approach and calculate the fluctuations around the vacuum, the point where we have the minimum energy. Under this profile, L ’ seems the right choice: we expand around a minimum and we can derive the spectrum of states. In particular the scalar particles under study when described by the Lagrangian L ’ have a well defined mass! What about the reflection symmetry? Since the two Lagrangians are equivalent, the symmetry is not lost. Whereas in L the symmetry is manifest, in L ’ it is hidden, but still present. In particular we find exactly the same results independently of the specific choice of the minimum: φ = v or φ = - v. We say that a “spontaneous symmetry breaking” has occurred, which has “generated” the mass of the particles. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 83 5. Spontaneous Breaking of a global gauge symmetry We can adopt the previous approach in the case of a complex scalar field described by the Lagrangian L = (∂µφ)*(∂µφ) – µ2φ*φ – λ(φ*φ)2 with φ= 1 √2 (φ1 ± i φ2) invariant under the global symmetry φ → eiα φ Considering the case λ > 0 and µ2 < 0, we can re-write L in terms of φ1 and φ2 L= 1 2 (∂µφ1)2 + 1 ( φ )2 µ 2 2 ∂ – 1 2 2 µ (φ1 2 + φ22 ) – 41 λ(φ21 + φ22 )2 with the potential which is now a function of the two fields φ1 and φ2: V = V(φ1, φ2) Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 84 For µ2 < 0 there is now a circle of minima of V(φ) in the plane (φ1,φ2) of radius v, such that V(q) 2 µ 2 =2 2 2 v φ 1 + φ2 = v with λ q2 Without loss of generality we choose the point φ1 = v q1 φ2 = 0 j d as the minimum of V(φ) and expand L around the vacuum in terms of the fields η(x) and ξ(x) through the substitution φ(x) = so obtaining L’= Gianluigi Fogli 1 2 ∂ µξ 2 + 1 2 1 √2 [v + η(x) + i ξ(x)] 2 ∂µη + µ 2η2 + const + cubic and quartic terms in η and ξ An Introduction to the Standard Model, Canfranc, July 2013 85 The third term in L ‘ has the form of a mass term leading to 1 2 mη2η2 for the field η(x), mη = √- 2µ2 The first term in L ‘ is the kinetic term of ξ(x), but there is no a corresponding mass term for ξ(x). This is the result of the Goldstone theorem which states that massless scalar particles occur whenever a continuous symmetry of a physical system is “spontaneously broken”. In conclusion, the attempt of generating a massive gauge boson through the spontaneous symmetry breaking of a global continuous symmetry leads to a theory “plagued” by the presence of massless scalar particles to worry about. Neverthless, let us proceed from a global to a local gauge theory. A miracle is about to happen. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 86 5. The Higgs mechanism We can now consider the spontaneous symmetry breaking of a “local gauge symmetry”. The simplest case is the U(1) gauge symmetry φ(x) → eiα(x) φ(x) with φ= 1 √2 (φ1 ± i φ2) We introduce in the lagrangian L = (∂µφ)*(∂µφ) – µ2φ*φ – λ(φ*φ)2 the covariant derivative ∂µ → Dµ = ∂µ – i e Aµ with the gauge field Aµ transforming according to Aµ → Aµ + 1e ∂µ α(x) Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 87 The Lagrangian takes then the form L = (∂µ+ i e Aµ) φ* (∂µ - i e Aµ) φ – µ2φ*φ – λ(φ*φ)2 - 1 4 FµνFµν If µ2 > 0, then this is just the QED Lagrangian for a charged scalar particle of mass m, with the addition of a φ4 self-interaction term. But we take µ2 < 0, since we want to generate mass terms through the spontaneous symmetry breaking mechanism. In this case we have to translate the field φ(x) to the ground state. With the same substitution as before φ(x) = the Lagrangian becomes L’= 1 2 ∂ µξ Gianluigi Fogli 2 + 1 2 1 √2 [v + η(x) + i ξ(x)] mass term 2 ∂ µ η - v 2 λ η2 + 1 2 mass term e2v2AµAµ - e vAµ∂µξ - strange off-diagonal term 1 F Fµν 4 µν An Introduction to the Standard Model, Canfranc, July 2013 + interaction terms 88 The particle spectrum in L ‘ contains a massless Goldstone boson ξ(x) mξ = 0 a massive scalar field η(x) mη = √- 2µ2 = a massive vector field Aµ(x) mA = ev √2λv2 We have obtained a massive vector field, but we have still the occurrence of a massless Goldstone boson. However, the occurrence of a strange term off-diagonal in the vector field induces to be careful in the interpretation of L ’. Indeed, giving mass to Aµ we have raised the number of degrees of freedom of the system: the polarization degrees of freedom go from 2 to 3 (addition of the longitudinal polarization of the massive vector boson) whereas the number of degrees of freedom of the scalar fields seems to be the same. Since we have one more degree of freedom in L ’, this means that the fields in L ’ not all correspond to distinct physical particles. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 89 Which field in L ‘ is unphysical ? Can we make use of gauge invariance to find a particular gauge transformation that eliminates one degree of freedom from the Lagrangian ? Let us note that at the lowest order in ξ(x)/v we can write φ(x) = 1 √2 [v + η(x) + i ξ(x)] ≅ 1 √2 [v + η(x)] e i ξ(x)/v This suggests the use of a different specific set of fields in the original Lagrangian: h(x) , θ(x) , Aµ(x) assuming for φ(x) and Aµ(x) the following transformation properties φ(x) → 1 √2 [v + h(x) + i θ(x)] ≅ √21 [v + h(x)] e i θ(x)/v Aµ → Aµ + 1e ∂µ θ(x) Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 90 This corresponds to a specific choice of the gauge, chosen in such a way to make the field h(x) real. Gauge invariance, i.e. the arbitrariness in the choice of α(x), is lost since we have identified the, in principle arbitrary, α(x) with the specific field θ (x) which makes h (x) real. The substitution in the original Lagrangian L leads to L = 1 2 ∂ µh 2 - λv 2h2 + 1 2 2 e v A µA µ 2 - λv h3 - 1 4 λh4 + 1 2 e2v2AµAµ h2 + e2vAµAµ h - 1 F Fµν 4 µν The Goldstone boson has disappeared. The extra degree of freedom is indeed spurious, it corresponds to the freedom of making a gauge transformation. L describes two interacting massive particles, a massive scalar field h(x) and a massive gauge vector boson Aµ(x). The field h(x) is called “Higgs particle”. The unwanted massless Goldstone boson has been turned into the longitudinal polarization of the massive vector boson. It is a “would be” Goldstone boson and it has been “eated” by the vector boson. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 91 6. Spontaneous Symmetry Breaking of a local SU(2) gauge symmetry We start from the Lagrangian L = (∂µφ) (∂µφ) – µ2φ φ – λ(φ φ)2 where φ(x) is an SU(2) doublet of complex fields φ(x) = φα φβ = 1 √2 φ1 + i φ2 φ3 + i φ4 L is invariant under the global SU(2) phase transformations φ → φ = ei α a τa/2 φ But if we require invariance under a local phase transformation, then we have to introduce the covariant derivative τa a ∂µ → Dµ = ∂µ + i g Wµ 2 a τa in terms of a triplet of vector fields Wµ and the coupling constant g, the being 2 the SU(2) generators in the doublet representation. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 92 Under infinitesimal local gauge transformations φ(x) → φ’(x) = 1 + i αa(x) and τa 2 φ(x) → → → → Wµ → Wµ – g1 ∂µ→ α – α × Wµ last term due to the nonAbelian character of SU(2) The Lagrangian then becomes L= ∂ µφ + i g → 1→ τ • Wµφ 2 ∂ µφ + i g → 1→ τ • Wµ φ 2 – V(φ) - 1 → →µν Wµν•W 4 with the usual potential term V(φ) = µ2φ φ – λ(φ φ)2 and the kinetic energy of the vector fields given by → → → → → Wµν = ∂µWν - ∂νWµ ‒ g Wµ × Wν Gianluigi Fogli last term due to the nonAbelian character of SU(2) An Introduction to the Standard Model, Canfranc, July 2013 93 We are interested to the spontaneous symmetry breaking of L. Accordingly we assume λ > 0 and µ2 < 0. The minimum of V(φ) is assumed in all the points that satisfy φ φ= 1 2 2 2 2 2 ( φ1 + φ2 + φ3 + φ4 ) = - µ2 λ We choose a specific minimum 2 φ3 φ1 = φ2 = φ4 = 0 µ2 == v2 λ The choice introduces the spontaneous breaking of the SU(2) symmetry. We expand around the specific vacuum φ0 = 1 √2 0 v Gauge invariance, with the specific choice of the gauge, allows to simply substitute in L the expansion φ(x) = 1 √2 0 v + h(x) with only the Higgs field surviving. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 94 It may appear rather surprising, but, according with the approach used before in the U(1) case, we can imagine to parametrize the fluctuations around φ0 in terms of four real fields θ1(x), θ2(x), θ3(x) and h(x) using the expression φ(x) = 1 √2 e 0 →→ i τ•θ(x)/v v + h(x) Let us expand the expression in terms of small perturbations: we obtain φ(x) = 1 √2 1 + iθ3/v i(θ1 - i θ2)/v 0 i(θ1 + i θ2)/v 1 - iθ3/v v + h(x) = 1 √2 θ2 + i θ1 v + h – iθ3 so fully parametrizing the deviation from the vacuum. On the other hand, the Lagrangian is locally invariant: we can then “gauge” the three would-be Goldstone bosons θ1(x), θ2(x), θ3(x), with the appropriate choice of the gauge. We arrive to the previous form of the “shift”, which is so justified φ(x) = 1 √2 0 v + h(x) At this point, in L ‘ there remains only the Higgs field without any trace of the gauge fields. They have been “gauged away” and disappear from the theory. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 95 The three gauge vector bosons are now massive. We can isolate the mass terms ig 1 2 →→ τ•Wµφο 2 = g2 8 (3) Wµ (1) Wµ (1) (2) (2) Wµ – i Wµ - + i Wµ (3) Wµ 0 v 2 = g2v2 8 ∑ (Wµ(i) ) 2 i and find their common mass to be MW = 1 2 gv In conclusion, we have a theory characterized by one massive scalar field h(x) and three massive gauge fields. The Goldstone bosons have been “eated” by the gauge fields when they become massive. The choice of the scalar field representation is crucial. If φ(x) instead than a complex doublet is chosen to be a SU(2) triplet of real scalar fields, then, always for λ > 0 and µ2 < 0, we find that only two gauge bosons acquire a mass, whereas the third remain massless. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 96 6. The Weinberg-Salam Standard Model Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 97 1. Revisiting electroweak interactions Electromagnetic amplitudes have been described as due to the interaction - µQ ψ) Aµ - i e jµ Aµ = - i e (ψγ em Q generator of U(1)em We have seen that the interaction derives from demanding gauge invariance of the Lagrangian of a free fermion - (iγ ∂µ ‒ m) ψ L=ψ µ under U(1)em local gauge transformations ψ → ψ’= eia(x)Q ψ We then obtain the Lagrangian - (iγ ∂µ ‒ m) ψ ‒ e (ψ- γ Q ψ) Aµ L=ψ µ µ kinetic energy and mass of ψ Gianluigi Fogli 1 4 interaction An Introduction to the Standard Model, Canfranc, July 2013 FµνFµν kinetic energy of Aµ 98 In order to include electroweak interactions, we have to introduce the interactions due to SU(2)L and U(1)Y. The first through the coupling of the isotriplet of lefthanded weak currents Jµ(i) with a triplet of vector bosons Wµ(i) - i g Jµ•Wµ = - i g χLγµT•Wµ χL → → → → The second interaction with the hypercharge current coupled to a vector boson Bµ - i g’ 1 Y µ j B = 2 µ - γ 1 Y ψ Bµ -i g’ ψ µ 2 The transformation properties of left-handed and right-handed fields are then → → χL → χL’ = eiα(x)•T + iβ(x)Y χL ψR → ψR’ = eiβ(x)Y ψL For example, in the case of the (νe e-) lepton pair we have νe eψR = eR χL = Gianluigi Fogli L isospin doublet with T = 1 , Y = - 1 2 isospin singlet with T = 0 , Y = - 2 An Introduction to the Standard Model, Canfranc, July 2013 99 The electromagnetic interaction is embedded in SU(2)L ⊗ U(1)Y being Q = T(3) + 1 Y 2 jµem = Jµ(3) + and then 1 Y j 2 µ Accordingly, the neutral current of the SU(2)L ⊗ U(1)Y interaction can be rewritten by introducing Wµ(3) and Bµ in terms of Aµ and Zµ through the mixing angle θw - i g Jµ(3) Wµ(3) - i g’ 1 Y µ j B 2 µ = - i ( g sinθw Jµ(3) + - i ( g cosθw Jµ(3) = - i e jµem Aµ - i 1 g’ cosθw jµY ) Aµ + 2 1 g’ sinθw jµY ) Zµ = 2 g JµNC cosθw Zµ so obtaining JµNC = Jµ(3) – sin2θw jµem Gianluigi Fogli and g sinθw = g’ cosθw = e An Introduction to the Standard Model, Canfranc, July 2013 100 From the requirement of gauge invariance under SU(2)L ⊗ U(1)Y it is possible to derive the electroweak gauge invariant lagrangian. Always in the case of the (νe e-) lepton pair, we have L1 = χLγµ i ∂µ - g 21 → τ•Wµ - g (→ 1 ) 2 Bµ χL + eRγµ i ∂µ - g (- 1) Bµ eR - → → 1 W Wµν µν 4 - 1 B Bµν 4 µν where the values of the hypercharge have been explicitly inserted. The last two terms correspond to kinetic energy and self-coupling of the Wµ(i) fields and to the kinetic energy of the Bµ field, respectively. L neutral. Gauge invariance requires the Lagrangian to be neutral, i.e. it must transform as a singlet under U(1)em in order to imply charge conservation. Fermion masses. L1 describes massless gauge bosons. Mass terms for the gauge bosons are not gauge invariant and cannot be added. The same is true also for the fermions fields: a mass term, for example for the electron, is given by - e = - m e- me e e 1 2 (1 – γ5) + 1 (1 2 - e + -e e ) + γ5) e = - me (e R L L R But this term breaks gauge invariance since eL is the member of an isospin doublet and eR is an isospin singlet. The fermion masses will be introduced by the spontaneous symmetry breaking of the theory. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 101 2. The Higgs field We want to introduce the Higgs mechanism in such a way to have W± and Z0 massive, whereas the photon Aµ remains massless. This is obtained adding to L1 a second SU(2)L ⊗ U(1)Y invariant Lagrangian containing the scalar fields necessary to induce the Higgs mechanism itself: L2 = →→ i ∂µ ‒ g T•Wµ - g 1 2 YBµ φ 2 - V(φ) Minimal Standard Model (Weinberg 1967) where φ(x) = φ+ φ0 = 1 √2 φ1 + i φ2 φ3 + i φ4 is an isospin doublet with weak hypercharge Y = 1. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 102 V(φ) is the usual potential. We assume λ > 0 and µ2 < 0 choosing the vacuum expectation value φ0 = 1 √2 0 v with T= 1 2 , Y=1 , Q=0 which breaks T(3) and Y, but leaves Q unbroken: Q φ0 = 0 φ0 → φ0 = eiα(x) Q φ0 = φ0 Since the vacuum φ0 is still left invariant by some subgroup of the gauge group, then the gauge bosons associated with this subgroup will remain massless. With the previous choice, this is indeed the case of U(1)em with the photon expected massless after the occurrence of the spontaneous symmetry breaking. The other three vector bosons associated to the remaining three generators of SU (2)L ⊗ U(1)Y will become massive and will be identified with W± and Z0 . Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 103 2. Masses of the gauge bosons The gauge boson masses come from L2 when the usual shift is applied φ(x) → 1 √2 0 = φ0 + v + h(x) 1 √2 0 h(x) From the terms proportional to φ0 we obtain (3) gWµ + g’ Bµ → 2 1 ig 1 → τ•Wµ - i g’ 1 Bµ φ0 = 2 2 8 = = Gianluigi Fogli (2) (Wµ(1)) 2 + (Wµ )2 1 v2g2 2 1 2 vg (1) g (Wµ + 2 (-) (+) Wµ W µ + (2) i Wµ ) (1) (2) g (Wµ - i Wµ ) - (3) gWµ + g’ Bµ 0 v 2 = + v2 (g’Bµ – gWµ(3) ) (g’Bµ – gW (3) µ) = 1 2 (3) v ( W µ 2 Bµ ) g2 - gg’ - gg’ g’2 An Introduction to the Standard Model, Canfranc, July 2013 Bµ W (3) µ 104 By comparing the first term with the expectation for the mass of a charged vector boson M2 W (+)W µ(-) we find W µ MW = 1 v g 2 The second term can be rewritten 1 8 v2 g2 Wµ(3) Wµ(3) - 2 g g’ Wµ(3)Bµ + g’2 BµBµ = 1 2 v 8 (3) gWµ - g’ Bµ 2 (3) + 0 g’ Wµ + g Bµ 2 which shows that the eigenvalue of the combination (g’ Wµ + g Bµ) is zero. The orthogonal combination has eigenvalue different from zero. We identify the first combination as the massive Zµ and the orthogonal one as the photon, with masses 1 M 2 Z 2 + 1 M 2 A 2 Z µ A µ 2 2 Normalizing the fields Aµ = Zµ = Gianluigi Fogli g’ Wµ(3) + g Bµ √ g2 + g’2 g Wµ(3) - g’ Bµ √g2 + g’2 with MA = 0 with MZ = An Introduction to the Standard Model, Canfranc, July 2013 1 v 2 √g2 + g’2 105 By taking into account that g sinθw = g’ cosθw = e i.e. tg θw = g’ g we find Aµ = Bµ cosθw + Wµ(3) sinθw Zµ = - Bµ sinθw + Wµ(3) cosθw and MW = cosθw MZ The two vector boson masses are different, and this appears an effect of the mixing. Note that the previous relation has been obtained in the Minimal Standard Model (one Higgs doublet) and is specific of the Minimal Standard Model. Different choices of the Higgs sector lead in general to different relations, even though the existence of more than one Higgs doublet leads to the same numerical results. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 106 Making use of the relation which connects the vector boson masses to the low energy Fermi coupling constant G √2 = g2 8M2w and comparing the the result obtained for MW, MW = 1 2v2 = g2 8M2w = G √2 1 2 v g, one derives Making use of the experimental value of the Fermi coupling constant one obtains for the vev v v = 246 GeV Similarly, making use of the value at low energy of θw one can predict the vector boson masses MW = 37.3 sinθw GeV MW = 74.6 sin2θw GeV These relations are well verified experimentally. All the phenomenology strongly supports the Standard Model in its minimal version (even though more Higgs doublets leads to the same relations). Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 107 3. The parameter ρ We have just established the relation between MZ and MW, that we can rewrite in the form MW = cosθw MZ Let us remind that the parameter ρ, introduced to specify the relative strenght of the neutral and charged current, is expressed in terms of the vector boson masses. Accordingly, for the Minimal Standard Model ρ = M2w M2z cos2θw = 1 whereas for a general Higgs contributions ρ Gianluigi Fogli = M2w M2z cos2θw = ∑ v i2 Ti(Ti+1) - ∑ 1 2 1 Y2 2 i v i2 Y i2 An Introduction to the Standard Model, Canfranc, July 2013 vi , Ti , Yi are vev, weak isospin and weak hypercharge of the generic Higgs representation 108 4. Masses for the fermions A direct fermion mass term – m ψψ cannot be assumed since it breaks gauge invariance. But terms corresponding to fermion masses come from the piece of the Lagrangian L3 which couples fermions to the Higgs sector. Assuming the minimal version of the Standard Model, i.e. only one Higgs doublet, we have for the fermion doublet (νe e-) the following gauge invariant couplings L3 = - Ge -e e- ) L (ν φ+ φ0 eR + -R (φ - -φ0) e νe e L The spontaneous symmetry breaking, introduced through the usual shift φ(x) → 1 √2 0 v + h(x) = φ0 + 1 √2 0 h(x) transforms L3 into L3 = Gianluigi Fogli 1 Gv √2 e - LeR + e-ReL) (e - 1 G (e LeR + √2 e -eReL) h An Introduction to the Standard Model, Canfranc, July 2013 109 i.e. -e L3 = - me e me e e h v Since v = 246 GeV, the coupling of the electron to the Higgs particle is very small. with me = 1 Gv √2 e eR (T= 0 , Y= -2) eL (T= he+e- vertex factor -i h0 (T= 1 , Y= 1) 2 Ge √2 1 , Y= -1) 2 = -i g 2 me MW Quark masses are generated in the same way. In order to give mass also to the upper member of the doublet (the u quark) we have to introduce the Higgs doublet conjugate to φ - φ0 v + h(x) 1 φc = -i τ2 φ* = √2 φ 0 after SSB which transforms as φ under SU(2), but has opposite weak hypercharge. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 110 We can then construct the following Lagrangian L4 = - Gd ( u d) L - φ+ - - φ0 d - Gu ( u d)L u + h.c. φ- R φ0 R which can be rewritten in the form: - L4 = - md d d – mu -u u - md mu ddh uuh v v Of course, the formalism can be extended to the case of more generations of quarks and leptons. In the case of quarks, because of the mixing, the Yukawa couplings take a matrix form. It is however interesting to observe that the Higgs coupling to the different generations is always flavor conserving because of the GIM mechanism. In conclusion, no flavor changing neutral currents (FCNC) appear in the theory. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 111 4. The Higgs mass The Higgs mass cannot be predicted from the theory. Starting from the potential and considering it as an effective potential V(φ) = µ2φ φ – λ(φ φ)2 + …. the use of the first two terms of it leads to mh2 = 2v2λ Since v is fixed, large values of mh corresponds to large value of λ. A meaningful perturbative approach requires λ < 1 and then the Higgs mass cannot larger than a few hundred GeV. On the other hand, mh cannot be smaller than, say, ~10 GeV, otherwise radiative corrections would wash out the minimum at v ≠ 0. It is matter of facts that the Higgs mass has been calculated through its quantum effects well before its recent measurement at LHC. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 112 7. Renormalization and running coupling constants Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 113 1. A loop diagram and the photon propagator Let us consider the Feynman graph for the Rutherford scattering of an electron from a static charge Ze (a nucleus, typically). The same procedure adopted in the study of the scattering e-µ- → e-µ- leads to the invariant amplitude Jµ e- ui -iM = (ie uf γµ ui) -i gµν → (i j ( ν q )) q2 -i i.e. e- ieγµ gµν uf q2 • -i(Ze,0) -i (- i Ze) -iM = (ie uf γ0 ui) q2 being for a static charge j0( x ) = ρ( x ) = Ze δ( x ) → Gianluigi Fogli → → and →→ j( x ) = 0 An Introduction to the Standard Model, Canfranc, July 2013 114 Let us now introduce a loop along the photon propagator. We can apply the Feynman rules and obtain (-1)n for a diagram with n fermion loops Jµ e- ui e- ieγµ gµρ -i 2 q uf q -iM = (-1) (ie uf γµ ui) • ∫ (2π) • -i ieγρ q-p p ieγσ gσν -i 2 q •→ -i jν( q ) Gianluigi Fogli d 4p 4 gµρ q2 -i (ieγρ)αβ • i(p + m)βλ p2 – m2 (ieγσ)λτ i(q - p + m)τα (q-p)2 – m2 • gσν → ( i j ( q )) ν q2 where p is the four-momentum circulating around the loop. Since p is not observable, we have to sum over all possible values of p. An Introduction to the Standard Model, Canfranc, July 2013 115 Comparing with the lowest order, the effect of the loop can be regarded as a modification of the propagator: we can write -i gµν q2 -i gµν q2 + -i gµρ ρσ gσν I -i q2 q2 -i gµν q2 (-i) (-i) + q2 Iµν q2 where Iµν (q2) = (-1)1 d 4p ∫ (2π) Tr (ieγρ) 4 i(p + m) p2 – m2 (ieγσ) i(q - p + m) (q-p)2 – m2 Symbolically - Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 116 In the integral we have terms that diverge as p → ∞. It can be shown that Iµν can be written (after a lenghty calculation) Iµν = -i gµν q2 I(q2) + (terms ∼ qµqν) these terms vanish when coupled with external currents because of charge conservation with I(q2) = α 3π ∫ ∞ m2 dp2 p2 - 2α π ∫ 0 1 q2z(1-z) dz z (1-z) ln 1 m2 where m is the electron mass. In I(q2): The first term is divergent, but only logarithmically because of the “conspiracy” of the divergent terms, with cancellation of the terms quadratically divergent. The second term is a finite part, properly parametrized. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 117 In the following we will be interested to the contribution of Iµν only for small values and large values of (-q2). So, with M introduced as a cut-off to “cure” the divergence, we have for (-q2) for (-q2) small I(q2) 2 α q M2 α = ln + 15π m2 m2 3π large I(q2) -q2 α M2 α = ln ln 3π m2 m2 3π = α M2 ln 3π -q2 Apparently, the divergence spoils of any meaning the result. But let us consider the invariant amplitude including the loop contribution in the small (-q2) limit -iM = - γ0 u) (ie u (-i) q2 2 α 1ln M2 3π m 2 α q + O (e4) (-iZe) 2 15π m We can re-write it in the form -iM = Gianluigi Fogli - γ0 u) (ieR u (-i) q2 1- eR2 60 π2 q2 (-iZeR) m2 An Introduction to the Standard Model, Canfranc, July 2013 118 if we assume eR = e 1 - e2 12π2 ln M2 1 2 m2 It is easy verified that the two expressions of the invariant amplitude are equivalent at the O(e2). We assume eR as the physically measured electric charge in any long range Coulomb experiment as the Thomson scattering or the Rutherford scattering γ γ e e Thomson scattering So, interpreting e e • Z√α Rutherford scattering eR2 α = 4π we introduce a measurable parameter into the play, in terms of which the invariant amplitude is finite. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 119 2. The g – 2 Can we verify that this approach is meaningfull and leads to a physically observable effect? The answer is yes. There are several examples. One of the most significant concerns the magnetic moment of the electron, the so-called g-2. Coming back to the Gordon decomposition uf γ µ ui = 1 uf (pf + pi)µ + iσµν(pf - pi)ν ui 2m with σµν = i (γµγν – γνγµ) 2 it is possible to verify that the σµνqν term represents the contribution of the magnetic moment of the electron → → → → e → µ= - e σ often written as µ =-g S with S = 1 → σ 2 2m 2m g is called gyromagnetic ratio, it satisfy g=2 Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 120 If we now consider the correction to the vertex related to the loop diagram in the figure, it is possible to verify that the diagram modifies the structure of the electromagnetic current in the following form - e u-fγµui → -f γµ -eu q2 m 3 α 1+ ln mγ 8 3π m2 - α 1 i σµνqν 2π 2m ui This implies a contribution of order α to the gyromagnetic ratio coming from the second term: i.e. → α α → µ= - e 1+ σ g=2+ π 2m 2π We can now calculate higher order contributions to g – 2: we find g–2 2 theor g–2 2 exper = 1 2 α π - 0.32848 α π 2 + (1.49 ± 0.2) α π 3 + … = (1159655.4 ± 3.3) × 10-9 = (1159657.7 ± 3.5) × 10-9 which proves relativistic quantum field theory at the level of quantum effects. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 121 3. Renormalization in QED Renormalization is a rather complex procedure and cannot be developed here. The various diagrams lead to the redefinition of charge, mass and wave function of the electron, by re-absorbing the divergent parts of the diagrams. The radiative corrections come from the finite terms and can be tested experimentally. In particular the loop diagrams in the photon propagator, known as vacuum polarization modify the charge of the electron. We can include the loops at higher order in the form e e0 1 - = e Gianluigi Fogli 2 + … e0 An Introduction to the Standard Model, Canfranc, July 2013 122 which corresponds to the expansion e2 = e20 1 – I(q2 ) + O(e40 ) The geometric series can be summed up e e0 1 = e e0 1 + and we can write e2 = e20 Gianluigi Fogli 1 1 + I(q2 ) An Introduction to the Standard Model, Canfranc, July 2013 123 This shows that we can introduce a renormalized coupling constant α(Q2) = e2(Q2) for different values of Q2 = - q2, where Q2 is the physical value of the momentum transfer of the process we are considering. The coupling behaves as a 1 α(Q2) = α0 1+ running coupling constant I(q2 ) In the large Q2 = - q2 limit we can use the expression obtained for I(q2) and write α0 2 α(Q ) = α0 Q2 1ln 3π M2 There remains the dependence on the cut-off M2. But it can be eliminated by taking a renormalization momentum µ as reference and subtracting α(µ2) from α(Q2): α(Q2) = α(µ2) α(µ2) 1ln 3π Q2 µ2 The renormalization procedure is completed: only physical measurable quantities appear. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 124 We interpret the result of a running coupling constant a(Q2) as the effect of the distance at which we measure the charge of the electron. The presence of the pairs e+e- in the perturbative expansion of the electron propagator gives rise to a electromagnetic screening when we “measure” the electric charge of the electron with a test charge. - + + R - - + - + + test charge + Therefore, the closer one approaches the electron, the larger is the charge one measures. One expects a behaviour of the Coulomb charge as that shown in the figure. electron charge high energy probe low energy probe α = Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 1 137 R 125 3. Running coupling constant in QCD A similar approach can be applied in QCD, but with very different results. This depends on the effects of the additional graphs that contain the self-interaction effects of gluons. The running coupling constant as(Q2) for QCD is characterized by a coefficient αs(µ2) 4π 2 n –5 3 f + 16 to be compared with the QED coefficient α(µ2) 4π - 43 The first term is essentially the same: in the QED case we have the photon which - to be can fluctuate in the pair e+e-, in QCD the gluon can fluctuate in the pair qq, multiplied for the number of flavors. Moreover, there is a factor 2 in the definition of the two coupling constants. The factor – 5 comes from the fermion loops of transverse gluons. It can be shown that these terms contribute always with a negative sign. Finally, we have a factor + 16, related to the introduction of “ghost” particles that cancel the unphysical polarizations but do not lead to the production of physical particles, and then contribute with a positive term. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 126 In conclusion, the QCD “running coupling constant” is given by αs(Q2) = αs(µ2) αs(µ2) 1+ 33 – 2nf ln 12π Q2 µ2 The coefficient is positive. It needs to have more than 16 quark flavors to change the sign and then the behaviour of as(Q2), which decreases with increasing Q2 and becomes small for short-distance interactions. We refer to this property as asymptotic freedom Conversely, at sufficiently low Q2 the coupling becomes large, i.e. of order O(1). It is usual to denote the value of Q2 at which this occurs as Λ2. It follows that Λ2 = µ2 exp Gianluigi Fogli - 12π (33 - 2nf) as(µ2) An Introduction to the Standard Model, Canfranc, July 2013 127 It follows that αs(Q2) can be written in terms of Λ2 in the form αs(Q2) = 12π 33 – 2nf ln Q2 Λ2 For Q2 much larger than Λ2, αs(Q2) is small and a perturbative description of quarks and gluons is possible. This is the regime of perturbative QCD. For Q2 of order Λ2, the perturbative approach does not make sense: quarks and gluons arrange themselves into hadrons. This is the regime of confinement. color charge high energy probe confinement region αs = 1 1 Fermi R Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 128 4. Grand Unification Is it possible to go beyond the Standard Model searching for an unifying group G ⊂ SU(3)c ⊗ SU(2)L ⊗ U(1)Y ?? The dependence on Q of the three couplings, g , g’ , gs (where we assume αs = g2s /4π) seems to agree with the possibility that for some large-momentum (or short-distance) scale Q = MX the three couplings merge into a single coupling gG, so that the group G describes a unified interactions with coupling gG(Q) at Q = MX Indeed, the two non-Abelian groups are asymptotically free, whereas the coupling of the Abelian group increases with Q, so suggesting a convergence towards a common value. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 129 Let us identify the couplings according to running coupling constants g1(Q) = C g’(Q) α3 g2(Q) = g(Q) and g3(Q) = gs(Q) gi2 αi = 4π α2 α1 MX Z mass where C is a Clebsh-Gordan coefficient of the group G. 102 Energy Scale (GeV) 1015 It is convenient to make use of the quantities 1/αi = 1/g2i since they depend linearly on ln Q. For example, in the case of g3 = gs we obtain, by rewriting the relation previously found for αs, 1 g32(µ) Gianluigi Fogli = 1 g32(Q) + 2b3 ln Q µ with b3 = 1 4π2 An Introduction to the Standard Model, Canfranc, July 2013 2 3 nf – 11 130 Similar expressions can be found for the other two couplings, so that, by identifying Q with MX and gi(MX) = gG, we can write 1 gi2 (µ) = 1 gG2 + 2bi ln MX µ with b1 = 1 4π2 2n 3 f b2 = 1 4π2 22 3 - + b1 b3 = 1 4π2 – 11 + b1 A straightforward calculation allows to eliminate nf and gG from the three equations. We have M 6π ln X = µ 11(1 + 3C2) 1 α - 1 + C2 αs with MX expressed in terms of α and αs estimated at the momentum of reference µ, C being dependent on the unification group G. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 131 Assuming µ = 10 GeV, we can use the actual values α = 1/137 and αs = 0.1. Concerning C, it can be taken C2 = 5/3, the value relative to the unification group SU(5). We obtain MX = 5 × 1014 GeV with a weak dependence on the different parameters. The Weinberg angle is determined in a grand unified theory, since it is given by sin2 θw = g12(Q) g12(Q) + C2g22 (Q) It follows sin2θw = 3/8 at Q = MX. On the other hand, it is possible to derive, with a simple calculation, the value of sin2θw at Q = 10 GeV, to be compared with its actual value. We have sin2 θW = 1 1 + 3C2 1 + 2C2 α αs which gives sin2θw = 0.2, close to the experimental value. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 132 4. SU(5) and proton decay The simplest grand unification group is SU(5), proposed by Georgi and Glashow in 1974. SU(5) can accomodate all the known fermions (leptons and quarks) in two distinct irreducible representations of the group, according to - (1 , 1) + (3 , 1) + (3 , 2) - 5 = (1 , 2) + (3 , 1) = (νe , e-)L + dL 10 = = e+L + uL + (u ,d)L in terms of IR’s of SU(3)c ⊗ SU(2)L On the other hand, in SU(5) there are 24 vector bosons. We distinguish - 24 = (8 + 1) + (1 , 3) + (1 , 1) + (3 , 2) + (3 , 2) gluons W±, Z0, γ leptoquarks X, Y The leptoquarks X and Y are two colored heavy gauge bosons, SU(2)L doublets. They mediate interactions that turn quarks into leptons and viceversa, with violation of leptonic and baryonic number. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 133 In particular leptoquarks are expected to mediate proton decay. We can estimate the decay rate, by comparing the process to the µ-decay: from G √2 = g2 νµ 8M2w it follows Γ(µ- → e-νeνµ) = …. G2mµ5 = …. mµ5 4 MW g µ- √2 W± e- g ν e √2 In a similar way we expect GG √2 = gG2 u 2 8MX so that we can estimate 2 5 Γ(p → πe-) = …. GGmp = …. Gianluigi Fogli gG X p mp5 MX4 gG e+ - u d d d An Introduction to the Standard Model, Canfranc, July 2013 π0 134 We can compare in the range of values of αs going from 0.1 to 0.2 the corresponding values of MX, sin2θw, and τp. We find it difficult to reconcile the results: either sin2θw is too small, or it is the proton which decays too fast. as MX (GeV) sin2θw τp (years) 0.1 5 × 1014 0.21 ~ 1027 0.2 2 × 1016 0.19 ~ 1034 However, an accurate prediction of the proton lifetime requires a more sophisticated calculation. The measurement of the proton lifetime would be an important experimental result in favor of GUT’s, since this process is allowed in any theory of grand unification. As said before, there are other, larger, groups that are even better candidates for the unification. In particular, they have the possibility of including also right handed neutrinos. But it is clear that we need more experimental information to say something of more convincing on grand unification. Gianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 135 5. The quantization of the electric charge This is one of the more interesting results related to GUT’s, in particular to SU(5). Indeed, within SU(5) the photon is one of the gauge bosons of the group and, as a consequence, the electric charge Q is one of the generators. Since the group is a simple group, the trace of each generator, and then also of the electric charge Q, is zero for any representation of the group. - For the representation 5 this means that Tr Q = 3Q d- + Qν + Qe- = 0 Qd = 1 3 Q e- an amazing results, since implies that charge is quantized. A similar calculation for the representation 10 leads to Qu = -2Qd The combination of the two results solves one of the most intriguing mistery of particle physics: why it is Qp = - Q eGianluigi Fogli An Introduction to the Standard Model, Canfranc, July 2013 136