Normalize the generating functional by its value with all sources zero, and take the connected part of the time-ordered correlation function. Source differentiation of the interaction exponential is equivalent to the Dyson series followed by Wick's theorem. Write and temporarily, so that every contraction has an unambiguous meaning.
The second-order Dyson series contributesThere are two connected external scalar field attachments:They become equal after exchanging the integration variables , cancelling . The connected fermion contractions have the closed fermion loop sign:The minus sign can also be checked directly: the pairing of the first with the fourth field and the second with the third has positive pairing parity, but . Multiplying it by leaves a positive result:A contraction of the two external fields with each other times a vacuum bubble cancels against the normalization. Separate fermion tadpole integrals at and belong to a product of one-point functions, removed by taking the connected part.
To match the source's generating functional convention, distinguish its displayed kernels from contractions. The positive exponent gives . With the source insertion order in equation (5), the positive exponent similarly gives . The two scalar propagators and two Dirac propagators supply a total phase . Thus, writing as in the requested formula,The matrix trace closes the spinor indices; no extra factor of two or additional symmetry factor remains for a Dirac field with this Yukawa interaction.
This cannot be the sole symmetry-breaking sector of the Standard Model. Its gauge-boson mass matrix has two zero eigenvalues, so besides the electromagnetic photon it predicts an extra massless neutral gauge boson. Every linear combination of and , including the combination normally called , is massless. The observed massive Z boson is a concrete contradiction. Choosing differently cannot change these two zero eigenvalues.
A real scalar triplet can occur as an additional sector, but a further vacuum expectation value, for example that of a Higgs doublet, is then necessary to obtain the observed neutral gauge boson mass pattern. This does not rescue the triplet-only model posed here.
Let and be the two gauge couplings. In unitary gauge, and the weak hypercharge term in the gauge covariant derivative vanishes. Using , the kinetic term becomesThus, in the basis , the gauge-boson mass matrix and its eigenvalues areHere . The two massless neutral gauge bosons may be rotated to any orthonormal basis; no neutral mass term selects a unique mixing angle.
Expanding the same kinetic term to first order in givesFunctional differentiation gives the Feynman rulesand, in the real basis, for . All such three-point Feynman vertices involving or vanish. The factor two in the real basis comes from the two identical gauge boson fields; in the charged basis it is already explicit in .
Put . The potential has its minimum atChoose the vacuum expectation value . Since the weak hypercharge is zero, on this real scalar triplet, and . Neither nor annihilates . The stabilizer subgroup is consequentlyIn particular, spontaneous symmetry breaking has not reduced the two neutral gauge generators to a single one: annihilates every component and remains unbroken independently of .
The scalar mass matrix isWith canonical kinetic normalization , the radial Higgs boson therefore hasThe two tangential zero eigenvalues are Goldstone modes, absorbed by the two massive gauge bosons through the Higgs mechanism.
There are two tree-level Feynman diagrams, with intermediate photon and gluon respectively. Their color charge tensors are and . The pure photon factor isTheir interference vanishes after the color charge sum because of photon-gluon colour orthogonality:Thus there is no term proportional to , and with the of part (iii),Here . The absence of interference concerns the specified averaged and summed result; a fixed choice of external color charge need not have zero interference.
The exchanged particle is now a gluon, and the incoming up quark mass replaces . DefineThe Dirac trace and spin average are unchanged. Normalize the fundamental gauge generators by . Averaging over the initial color charge choices and summing the final color charge choices givesReplace the factor from part (ii) by . ThereforeKeeping is allowed; if the usual negligible up quark mass approximation is wanted, simply set . The factor is essential: initial quark color charge states are averaged, whereas final ones are summed.
Let , , and . The single tree-level Feynman diagram is annihilation through a virtual photon. Up to an irrelevant common phase, its scattering amplitude isThere are two initial spin angular momentum states for each particle, so the spin average is . The final color charge sum gives . There is no additional photon polarization average for a virtual photon and no separate polarization label beyond the Dirac spinor spin sums.
The spin sums and the gamma-matrix trace identities giveHere . Contracting givesSince and , this is the massive vector-current annihilation kernelThus the absolutely normalized result isAfter taking out the displayed prefactor, . Because the source uses a proportionality sign, only their ratio is fixed until that prefactor is specified. Equivalently, after taking out just , their values are . No mass has been neglected. At and , the bracket yields the finite isotropic threshold value of the scattering amplitude; the vanishing threshold scattering cross-section comes from relativistic two-body phase space.
Use the Mandelstam variables , and . The external on shell relations are and . Expanding the three Mandelstam variables and using four-momentum conservation, , gives
There are four tree-level Feynman diagrams: three photon exchange graphs with pairings , , , and the four-photon contact interaction. They are drawn in the bottom row of the figure. With no external fermions or Faddeev-Popov ghost fields, an internal matter or ghost line would have to close into a loop; it cannot add a tree-level Feynman diagram here.
Use all-incoming null four-momenta with . Let be the external photon polarization vectors, including the complex conjugation for physical outgoing photons, and impose . For one pairing , put . At the first cubic Feynman vertex, the terms proportional to and vanish, leavingAt the other end, the incoming internal four-momentum is , giving . The Feynman gauge photon propagator then givesThe last expression is evaluated at generic kinematics, with singular channel limits taken after cancellation. Summing the three pairings produces , whereThe quartic Feynman vertex contributes . ThusThis is the tree photon-scattering cancellation in a quadratic Abelian gauge: the individual gauge fixing interactions cancel in the physical scattering amplitude. The argument uses transverse external photon polarization vectors; it does not assert that an arbitrary off shell four-point correlation function vanishes. Physical light-by-light scattering first occurs through charged-particle loop Feynman diagrams.
Expand the gauge fixing square:The second term is a derivative cubic photon interaction; the third is a quartic photon contact interaction. Both arise from nonlinear Abelian gauge fixing, rather than an additional physical electromagnetic force.
For , differentiating the quartic term with respect to four identical gauge fields gives eight copies for each pairing, henceWith the Fourier convention and all four-momenta incoming, the differentiated cubic term givesmatching the supplied Feynman rules. Reversing the Fourier convention reverses this cubic rule's sign; each exchange graph contains two such Feynman vertices, so its contribution is unchanged.
For fixed , the delta functional sets . Therefore its weighted functional integral isSince the action enters as , the gauge fixing contribution isThe Faddeev-Popov determinant does not depend on and remains in the path integral. The limit enforces the original condition sharply; for nonzero this is a weighted average of the gauge slices.
An ordinary bosonic Gaussian functional integral produces an inverse determinant. An integral over independent Grassmann fields instead produces a determinant, allowing a local action to represent the Faddeev-Popov determinant. With one convenient antighost sign convention,Multiplying the operator by the field-independent factor changes only the normalization of its determinant. After integration by parts, an equivalent density isThus the Faddeev-Popov ghost fields have a massless kinetic term and a photon-ghost-antighost interaction proportional to . They are auxiliary anticommuting scalar fields, not physical external particles. Closed ghost loops carry the minus sign of Grassmann variables. Replacing by reverses both displayed signs without changing any observable; both the kinetic and interaction conventions must be changed together.
A gauge-invariant path integral counts all representatives of each gauge orbit. In a perturbative region with a regular gauge fixing slice, insert the Faddeev-Popov gauge-orbit identityThe Faddeev-Popov determinant is the functional Jacobian converting integration over the gauge transformation parameter into integration over the condition. The redundant gauge group volume then factors out of normalized correlation functions.
Choose . Its infinitesimal variation givesThere is no derivative of in this expression: only is varied in . For , the determinant is field independent and can be absorbed into normalization. For , it is a functional of the gauge field, so dropping it changes the path integral, even though the gauge group is Abelian. Constant residual gauge transformations are removed by boundary conditions or by dividing out their volume, as usual in perturbative gauge fixing.
For fundamental-doublet breaking of SU(2), assume and . The minima of the scalar potential satisfy , giving a three-sphere vacuum manifold. A chosen vacuum expectation value is . No nonzero generator annihilates : its squared norm is . At the group level, an matrix fixing the nonzero vector must be the identity. Thus all three generators of the SU(2) gauge theory are broken in the usual perturbative Higgs mechanism description.
Make the radial and angular fields explicit:A local gauge transformation removes the three angular fields . In unitary gauge, the Pauli matrices identity givesThe mixed derivative term cancels between a term and its complex conjugate. The scalar potential becomesConsequently the complete physical Lagrangian density isReading the quadratic terms with their canonical factors givesAll physical scalar-vector and scalar self-interactions are displayed in the second line of . In particular the new cubic couplings proportional to are and . Their Feynman vertices, together with the quartic ones, areThe three-vector and four-vector interactions are already contained in the non-Abelian gauge field strength term. There are no further physical fields or interactions in this specified model. Before fixing unitary gauge, the quadratic angular term is . It exhibits three zero-curvature angular directions of the scalar potential, but also their mixing with the gauge fields. They supply the third polarization of each massive gauge boson, rather than massless physical particles.
The degrees of freedom count is before breaking and afterwards. If instead the symmetry were global, as in the decoupled scalar theory, the three angular fields would be physical massless Goldstone bosons. This is consistent with Goldstone's theorem. The language of a broken local symmetry refers to this gauge-fixed expansion; a gauge-dependent vacuum expectation value is not itself a gauge-invariant observable.
Goldstone's theorem states that in a relativistic quantum field theory, spontaneous breaking of an exact continuous global internal symmetry produces a massless Goldstone boson for each independent broken generator. More precisely, a conserved global current whose charge acts nontrivially on the vacuum expectation value has a massless pole in an appropriate current-field correlation function. The usual counting assumes a Lorentz invariant vacuum state and the standard locality and positivity hypotheses.
A local gauge symmetry is a gauge redundancy, so it does not by itself imply massless physical Goldstone bosons. In the Higgs mechanism, the corresponding scalar directions provide the longitudinal polarizations of massive gauge bosons.
For massless quark Compton scattering, the color average contributes one, since . The spin average and incoming photon polarization average supply . The Ward identities permit use of for each photon polarization completeness relation, and the massless fermion spin sum is .
LetReversing the gamma matrices order gives its Dirac-conjugate matrix . The squared scattering amplitude isFor clarity evaluate the four traces before dividing by their denominators. The Mandelstam variables obey , with , , and . For the squared -channel contribution, the gamma matrix trace identities implyHere and . For the squared -channel contribution, givesOne interference trace reduces tosince . The other interference trace is its complex conjugate and also vanishes. ThereforeEquivalently in terms of the fine-structure constant. In the physical noncollinear region , , so the result is positive. The singular massless endpoint reflects a collinear divergence; it is not a negative squared scattering amplitude.
Put , , , and suppress the common colour factor . Define by the product of Feynman rules. With massless external Dirac spinors the two graphs giveThe massless Dirac equation gives and . ThusBy four-momentum conservation, , soThe two terms cancel:This is the Ward identity for the incoming photon; replacing its photon polarization vector by its four-momentum produces zero. The outgoing photon obeys the analogous identity. Each exchange graph separately fails this test, so the cancellation requires their sum. Consequently the physical scattering amplitude is invariant under .
The two tree-level Feynman diagrams are the -channel and -channel quark exchange graphs, drawn in the middle row of the figure. Along the directed quark line the incoming photon is absorbed before the outgoing photon is emitted in the first graph; those two insertions are reversed in the second. Their internal four-momenta are and , respectively. There is no three-photon Feynman vertex in ordinary QED, and no two-photon-two-quark contact interaction in the Dirac field Lagrangian density.
The massive Dirac propagator isUse the gamma matrix trace identities and . They giveThe integrand is odd under , and dimensional regularization preserves that symmetry. The same vanishing holds for . One must not assert that itself is zero: its term proportional to generally is nonzero and divergent, but its trace against a single gamma matrix is zero. This one-photon fermion loop is also the simplest instance of Furry's theorem.
First fix a convention that is implicit in the source. For the stated free generating functionalthe time-ordered Wick contractions are and . These follow directly by the source differentiation rules in part (i). Thus the kernels called propagators in the question are distinguished here from the contractions, which already include their factors of .
At first order in the QED interaction, write for its integration point. The normalized three-point correlation function isThe zeroth-order term vanishes because the free photon one-point correlation function is zero. There are two nonzero Wick contractions of the four fields:The first term pairs the two external fermions directly and closes the two fields at the interaction into a fermion loop; its minus sign is the closed fermion loop sign. The second term joins the external fermions through the interaction, giving the connected contribution. This is also obtained by differentiating the quartic term in the free generating functional; the two choices of bilinear cancel its factor .
Insert and . Since , the result is exactlyThe trace term is a disconnected product of a photon tadpole diagram with the free external fermion propagator; the other term is the connected three-point correlation function. In this QED calculation the disconnected term vanishes, as established next. With a convention using themselves as the symbols for the propagators, the equivalent formula retains the prefactor and the connected-minus-trace ordering above; mixing these two conventions would give an erroneous overall phase.
There are unlisted articles, also show them or only show them.