Spontaneous symmetry breaking

Symmetry is the deepest organising principle in particle physics — demand a symmetry locally and a force appears, insist on it globally and a conservation law follows. Yet the world we actually inhabit is not symmetric: the vacuum has picked out a definite state, the weak force distinguishes left from right, and the $W$ and $Z$ are heavy while the photon is not. Spontaneous symmetry breaking is the resolution of this tension. It is the observation that the laws of a theory — its Lagrangian — can be exactly invariant under a symmetry while the state of lowest energy that those laws admit is not. The symmetry is not violated; it is hidden. The mechanism is worth stating carefully because it is so easily misread. Consider a Lagrangian invariant under some group $G$ , with a potential $V(\phi)$ shaped so that its minimum is not the symmetric point $\phi=0$ but a whole family of degenerate states related to one another by the symmetry. Nature must settle into one of them. That single choice is invisible to the equations — every choice is equally valid, and $G$ maps them into each other — but once made, the excitations built on top of it no longer display the full symmetry. A ferromagnet cooled below its Curie point is the everyday picture: the laws of electromagnetism have no preferred direction, yet each domain freezes its spins along a particular axis. The same idea, transplanted from magnets and superconductors into relativistic field theory by Nambu, is what lets gauge theories carry the masses we measure without ever writing down a mass term that would wreck their gauge invariance. It is the indispensable prelude to the Higgs mechanism.

Key takeaways

Spontaneous symmetry breaking means the Lagrangian is symmetric but the vacuum is not — the ground state picks one of a degenerate family of minima, and $G$ maps those minima into each other.
The engine is a potential with a "Mexican hat" shape, $V(\phi)=\mu^2\phi^*\phi+\lambda(\phi^*\phi)^2$ with $\mu^2<0$ : the symmetric point $\phi=0$ is a local maximum, and the true minima form a circle $|\phi|=v/\sqrt2$ .
Expanding about a chosen minimum splits the field into a massive radial mode ( $m^2=2\lambda v^2=-2\mu^2$ ) and a massless angular mode — the Goldstone boson.
Goldstone's theorem: breaking a continuous global symmetry produces exactly one massless scalar per broken generator; breaking a discrete symmetry produces none.
Gauge (local) symmetry is the loophole: the would-be Goldstone boson is eaten by the gauge field, which acquires a mass $m_A=qv$ — the Higgs mechanism, developed in aq-pp-higgs-mechanism.
Key idea: a symmetry can be exact at the level of the laws yet absent in the actual state of the world; this hidden symmetry lets gauge theories acquire mass without explicitly violating the gauge invariance that built them.

The core insight

A symmetry can be exact at the level of the laws (the Lagrangian) yet absent in the actual state of the world (the vacuum). This hidden symmetry lets gauge theories acquire masses without explicitly violating the gauge invariance that built them — the vacuum, not the equations, breaks the symmetry.

Where this lives today

Spontaneous symmetry breaking is not a single result but a template that recurs across all of physics, and its electroweak instance is now under direct experimental assault. The discovery of the Higgs boson at the LHC in 2012 confirmed the radial mode of the electroweak potential; the frontier since has been to measure the shape of the potential itself. Its curvature at the minimum is already fixed by $m_H=125\ \mathrm{GeV}$ , but the cubic and quartic self-couplings — which encode the full Mexican-hat profile — are being pinned down through di-Higgs production, a headline target of the High-Luminosity LHC and any future collider, because deviations from the Standard-Model shape would signal new physics tied to how the vacuum broke. The same physics governs the early universe: above a critical temperature $T_c\sim160\ \mathrm{GeV}$ , roughly $10^{-11}\ \mathrm{s}$ after the Big Bang, the electroweak symmetry was restored ( $v=0$ ), and the cosmological electroweak phase transition through which it broke is a candidate arena for generating the matter–antimatter asymmetry and a potential source of a stochastic gravitational-wave background. Meanwhile the global version is everywhere below the particle scale: the chiral symmetry breaking of QCD makes the pions light (pseudo-Goldstone bosons) and dynamically generates roughly 99% of the mass of the proton and neutron — so most of the mass of ordinary matter comes from an SSB that has nothing to do with the Higgs field. And the original condensed-matter homes remain vibrant: superconductors, superfluids, ferromagnets, and Bose–Einstein condensates are all spontaneously broken vacua, and the deep question of whether our own electroweak vacuum is truly stable or merely long-lived metastable hinges on how the Higgs self-coupling $\lambda$ runs to high energy under the pull of the heavy top quark.

The physical picture

Picture a marble balanced at the crown of a sombrero. The hat is perfectly symmetric under rotation about its axis, and so is the marble's situation — for an instant. But that central perch is a point of maximum potential energy: the marble is in unstable equilibrium, and the faintest nudge sends it rolling down into the circular valley at the brim. The moment it comes to rest somewhere in that trough it has selected a direction, even though no direction was preferred. The laws (the shape of the hat) kept every rotation, but the marble's resting state singled one out. That is spontaneous symmetry breaking in one image, and the field-theory potential that realises it has exactly this cross-section. The first figure makes the two consequences of that shape concrete: an unstable symmetric point and a valley of degenerate vacua, whose curvature up the wall fixes a mass and whose flatness around the trough leaves a massless mode.

Figure

Cross-section of the Mexican-hat potential V(φ) with μ² < 0. The symmetric point φ = 0 is a maximum (unstable); the true minima lie at ±v where v = √(−μ²/λ). Rotating this curve about the V axis (inset) gives the full hat: climbing the wall costs energy and defines the massive Higgs mode with m² = 2λv², while moving around the flat circular valley costs nothing and is the massless Goldstone mode. With v = 246 GeV and λ ≈ 0.13 the wall curvature gives m = 125 GeV

The angular flatness is not an accident of the drawing — it is forced by the symmetry itself. Because the potential depends only on $|\phi|$ , moving a vacuum around the circle costs exactly zero energy: the group $G$ carries one minimum into a physically identical neighbour, so the direction along the degenerate family is perfectly flat and its excitation is massless. Any direction that climbs out of the family costs energy and is massive. This is the geometric heart of Goldstone's theorem, and it separates cleanly into "how many flat directions" (massless modes) and "how many steep ones" (massive modes). When the broken symmetry is not merely global but gauged, the flat direction is not a physical particle at all — it can be rotated away by a gauge transformation, reappearing as the extra polarisation state a massless gauge field needs to become massive. The second figure carries this through to the case that matters most: the electroweak vacuum, with the real masses it produces.

Figure

Electroweak spontaneous breaking SU(2)ₗ × U(1)ᵧ → U(1)ₑₘ. The Higgs doublet has four real scalar fields; its vacuum expectation value v = 246 GeV breaks three of the four generators. The three would-be Goldstone bosons are eaten to become the longitudinal polarisations of W⁺, W⁻ and Z, which acquire masses M_W = 80.4 GeV and M_Z = 91.2 GeV; the unbroken U(1)ₑₘ leaves the photon massless; the one remaining scalar is the physical Higgs at 125 GeV. Degrees of freedom are conserved: 12 before, 12 after

The core equations

The whole phenomenon is encoded in the shape of a single scalar potential. Take a complex field $\phi$ with a global $U(1)$ symmetry $\phi\to e^{i\alpha}\phi$ ; the most general renormalisable, symmetric potential is quadratic-plus-quartic. The sign of the quadratic coefficient decides everything.

\mathcal L = (\partial_\mu\phi)^*(\partial^\mu\phi) - V(\phi),\qquad V(\phi) = \mu^2\,\phi^*\phi + \lambda\,(\phi^*\phi)^2 \qquad (1)

where: $\phi$ — complex scalar field; $\mu^2$ — the quadratic parameter (a parameter, which may be negative, not a physical mass-squared); $\lambda>0$ — the quartic self-coupling, required positive so the potential is bounded below. Reading it: with $\mu^2>0$ the minimum sits at $\phi=0$ and the symmetry is manifest; the physics of SSB is the choice $\mu^2<0$ , which turns $\phi=0$ into a maximum. When $\mu^2<0$ the minimum is no longer at the origin. Minimising $V$ gives a circle of degenerate vacua in the complex plane — the brim of the hat — parametrised by a phase that the symmetry rotates freely.

\phi^*\phi\big|_{\min} = -\frac{\mu^2}{2\lambda} \equiv \frac{v^2}{2},\qquad v = \sqrt{\frac{-\mu^2}{\lambda}},\qquad \langle\phi\rangle = \frac{v}{\sqrt2}\,e^{i\alpha} \qquad (2)

where: $v$ — the vacuum expectation value (VEV), the radius of the trough; $\alpha$ — the arbitrary phase labelling which vacuum nature picks. Reading it: no value of $\alpha$ is preferred, and the $U(1)$ symmetry maps one choice into any other, but the theory must be built around one of them — and that single act of choosing is the symmetry breaking. Expand the field about the chosen vacuum, $\phi = \tfrac{1}{\sqrt2}(v+\eta+i\xi)$ , and the two real components acquire sharply different masses. This is the payoff — the mass spectrum of the broken theory.

m_\eta^2 = 2\lambda v^2 = -2\mu^2 \;>0,\qquad m_\xi^2 = 0 \qquad (3)

where: $\eta$ — the radial excitation, oscillating up the steep wall of the potential; $\xi$ — the angular excitation, sliding along the flat valley floor. Reading it: $\eta$ is massive (the Higgs-like mode, its mass set by the curvature of the wall), while $\xi$ is exactly massless — it is the Goldstone boson, and its masslessness is guaranteed by the flatness the symmetry enforces. Counting the massless modes in general is the content of Goldstone's theorem: one for each generator of $G$ that fails to leave the vacuum invariant.

n_{\text{Goldstone}} = \dim G - \dim H \qquad (4)

where: $G$ — the full symmetry group of the Lagrangian; $H\subset G$ — the subgroup that still leaves the chosen vacuum invariant (the unbroken symmetry); $\dim G-\dim H$ — the number of broken generators. Reading it: a broken continuous symmetry always leaves flat directions in the vacuum manifold, one massless Goldstone boson per flat direction; a discrete symmetry has no generators to break and produces none. The final equation is the loophole that makes SSB useful in the Standard Model. Promote the global $U(1)$ to a local gauge symmetry by introducing a gauge field $A_\mu$ through the covariant derivative; the would-be Goldstone boson is then not physical, and the gauge field swallows it to become massive.

D_\mu\phi = (\partial_\mu - iqA_\mu)\phi \;\Longrightarrow\; m_A = qv,\qquad M_W = \tfrac12 g v \qquad (5)

where: $q$ — the gauge coupling; $A_\mu$ — the gauge field; $m_A=qv$ — the mass it acquires from the VEV. Reading it: the massless gauge field (two polarisations) absorbs the massless Goldstone (one state) to become a massive vector (three polarisations) — degrees of freedom are conserved, and no gauge invariance was ever broken by hand. The full electroweak version, giving $M_W=\tfrac12 gv$ , is owned by aq-pp-higgs-mechanism and recalled here by name.

Derivation AThe vacuum of a complex scalar and the birth of the Goldstone boson

The goal is to take the symmetric Lagrangian of equation (1), impose $\mu^2<0$ , locate the degenerate ring of vacua, expand the field about one chosen point, and read off the mass spectrum — showing explicitly that one combination is massive and the orthogonal combination is exactly massless. This is the concrete calculation behind Goldstone's theorem, and it exhibits the massless mode as the direction along which the symmetry moves the vacuum for free.

Step 1. Locate the minimum. Write $V = \mu^2(\phi^*\phi)+\lambda(\phi^*\phi)^2$ and treat $s\equiv\phi^*\phi$ as the variable: $dV/ds = \mu^2+2\lambda s$ , which vanishes at $s = -\mu^2/2\lambda$ . For $\mu^2<0$ and $\lambda>0$ this is positive and is a genuine minimum, so $|\phi|_{\min}=v/\sqrt2$ with $v^2=-\mu^2/\lambda$ . Why: the minimum is a value of the modulus only — the phase is completely undetermined, which is precisely why the vacua form a degenerate circle rather than a point.

Step 2. Choose a vacuum and expand. Every point on the circle is equivalent, so pick the real one, $\langle\phi\rangle=v/\sqrt2$ , and write the field in terms of two real fluctuations $\phi = \tfrac{1}{\sqrt2}\big[(v+\eta) + i\xi\big]$ . Here $\eta$ measures displacement along the radius and $\xi$ measures displacement around the circle. Why: choosing real $\langle\phi\rangle$ is a convention — the physics is identical for any $\alpha$ — but making a choice is unavoidable, and it is exactly this choice that breaks the symmetry spontaneously.

Step 3. Substitute and isolate the quadratic terms. With $\phi^*\phi = \tfrac12[(v+\eta)^2+\xi^2]$ , the potential becomes $V = \tfrac12\mu^2[(v+\eta)^2+\xi^2] + \tfrac14\lambda[(v+\eta)^2+\xi^2]^2$ . Using $\mu^2=-\lambda v^2$ and dropping constants and cubic/quartic pieces, the quadratic part is $V^{(2)} = \lambda v^2\,\eta^2 + 0\cdot\xi^2$ . Why: the $\xi^2$ term cancels identically because the potential does not depend on the phase — moving along $\xi$ is moving around the flat valley, and a flat direction has no restoring force.

Step 4. Read the masses. A term $\tfrac12 m^2\varphi^2$ identifies the mass, so $m_\eta^2 = 2\lambda v^2 = -2\mu^2$ and $m_\xi^2 = 0$ . The radial mode $\eta$ is the massive "Higgs" excitation; the angular mode $\xi$ is the massless Goldstone boson. Why: the massless boson is not fine-tuned — it is protected by the exact symmetry of the Lagrangian, which forbids any potential term that would lift the valley floor, so it survives to all orders as long as the symmetry stays global and unbroken explicitly.

Go deeper · Derivation B · 2 minGoldstone's theorem — why every broken generator leaves a massless mode

Derivation BGoldstone's theorem — why every broken generator leaves a massless mode

The purpose here is to prove the count in equation (4) without committing to a specific potential: for any theory whose Lagrangian is invariant under a continuous group $G$ but whose vacuum is invariant only under a subgroup $H$ , the number of massless scalars equals the number of broken generators $\dim G-\dim H$ . The argument rests on a single fact — that the potential is constant along any direction generated by a symmetry — and turns that flatness into a vanishing entry of the mass matrix.

Step 1. State the invariance as a constraint on $V$ . Under an infinitesimal symmetry transformation the fields shift by $\delta\phi_a = i\epsilon\,(T)_{ab}\phi_b$ , and invariance of the potential means $V(\phi)=V(\phi+\delta\phi)$ for every generator $T$ . Expanding to first order gives the exact identity $\dfrac{\partial V}{\partial\phi_a}(T)_{ab}\phi_b = 0$ everywhere in field space. Why: this is the mathematical statement that the potential does not change as we move along a symmetry direction — the valley floor is level — and it holds as an identity, not just at special points.

Step 2. Differentiate and evaluate at the vacuum. Differentiate that identity with respect to $\phi_c$ and set $\phi=\langle\phi\rangle$ . The term containing $\partial V/\partial\phi_a$ vanishes because the vacuum is a minimum ( $\partial V/\partial\phi_a|_{\min}=0$ ), leaving $\dfrac{\partial^2 V}{\partial\phi_c\partial\phi_a}\Big|_{\min}(T)_{ab}\langle\phi_b\rangle = 0$ . Why: the second-derivative matrix evaluated at the minimum is exactly the mass matrix $M^2_{ca}$ , so this equation says the mass matrix annihilates the vector $(T)_{ab}\langle\phi_b\rangle$ .

Step 3. Interpret the null vectors. The object $(T)_{ab}\langle\phi_b\rangle$ is the change in the vacuum under the generator $T$ . If $T$ leaves the vacuum invariant (an unbroken generator, $T\in H$ ), this vector is zero and the equation is empty. If $T$ moves the vacuum (a broken generator), the vector is nonzero and is an eigenvector of $M^2$ with eigenvalue zero — a massless mode. Why: every direction in which the symmetry actually moves the vacuum is a flat direction of the potential, hence a zero of the mass matrix; this is the exact correspondence between broken generators and massless bosons.

Step 4. Count them. There is one such independent null vector for each broken generator, so the number of massless Goldstone bosons is $\dim G-\dim H$ , the dimension of the broken coset. Why: the theorem is completely general — it needs no perturbation theory and no special potential, only that the symmetry is continuous, global, and spontaneously (not explicitly) broken; a discrete symmetry has no generators and yields no Goldstone bosons.

Go deeper · Derivation C · 2 minGauging the symmetry — how the Goldstone boson becomes the gauge boson's mass

Derivation CGauging the symmetry — how the Goldstone boson becomes the gauge boson's mass

The aim is to show what changes when the broken symmetry is local: the massless Goldstone boson of the global case disappears from the physical spectrum, and its single degree of freedom reappears as the longitudinal polarisation of a now-massive gauge field, giving $m_A=qv$ with no violation of gauge invariance. We introduce the covariant derivative, go to the unitary gauge where the Goldstone is rotated away, and extract the gauge-boson mass from the kinetic term.

Step 1. Gauge the $U(1)$ . Replace $\partial_\mu\phi$ by $D_\mu\phi=(\partial_\mu-iqA_\mu)\phi$ and add the field-strength term $-\tfrac14 F_{\mu\nu}F^{\mu\nu}$ . The Lagrangian is now invariant under the local transformation $\phi\to e^{i\alpha(x)}\phi$ , $A_\mu\to A_\mu+\tfrac1q\partial_\mu\alpha(x)$ . Why: promoting the constant phase to a spacetime-dependent one is what turns the global symmetry into a gauge symmetry and forces the gauge field $A_\mu$ into existence, exactly as in the local gauge principle.

Step 2. Exploit the gauge freedom. Because $\alpha(x)$ is now arbitrary at every point, the phase of $\phi$ is pure gauge: choose $\alpha(x)$ to cancel the Goldstone field entirely, writing $\phi(x)=\tfrac{1}{\sqrt2}\big(v+h(x)\big)$ with no imaginary part. This is the unitary gauge. Why: the would-be Goldstone $\xi$ was the phase of $\phi$ , and a local phase can be removed by a gauge transformation — so unlike the global case, the Goldstone is not a physical particle here; it is a gauge artefact.

Step 3. Extract the gauge-boson mass. Substitute into the kinetic term: $|D_\mu\phi|^2 = \tfrac12(\partial_\mu h)^2 + \tfrac12 q^2(v+h)^2 A_\mu A^\mu$ . The piece quadratic in $A$ with $h\to0$ is $\tfrac12 q^2 v^2 A_\mu A^\mu$ , a mass term $\tfrac12 m_A^2 A_\mu A^\mu$ with $m_A=qv$ . Why: the vacuum value $v$ inside the covariant derivative feeds the gauge field a mass directly — the mass is generated dynamically by the VEV, never written into the Lagrangian by hand, so gauge invariance was never sacrificed.

Step 4. Balance the degrees of freedom. Before breaking: one complex scalar (2 real DOF) plus one massless gauge field (2 polarisations) = 4. After: one real scalar $h$ (1 DOF) plus one massive vector (3 polarisations) = 4. Why: the Goldstone boson was not destroyed — it was "eaten", becoming the longitudinal polarisation that a massless gauge field lacks and a massive one requires; the bookkeeping closes exactly, which is the sign the mechanism is consistent.

Limiting cases

Case	Result	Meaning
$\mu^2>0$ (unbroken)	unique minimum at $\phi=0$	symmetry manifest; excitations are a degenerate massive multiplet, no SSB
$\mu^2<0$ (broken)	ring of minima $\|\phi\|=v/\sqrt2$	one radial massive mode + one angular massless Goldstone
Discrete symmetry broken	two minima $\pm v$ , no flat direction	massive excitation $m^2=2\lambda v^2$ , no Goldstone boson; domain walls instead
Global continuous broken	$n_{\text{Goldstone}}=\dim G-\dim H$	one exactly massless scalar per broken generator (Goldstone's theorem)
Local (gauge) symmetry broken	Goldstone eaten, $m_A=qv$	gauge boson becomes massive, no physical massless scalar — the Higgs mechanism
$\lambda\to0^+$	potential flat, $v\to\infty$	theory unbounded/ill-defined; $\lambda>0$ is required for a stable vacuum
$v\to0$ (restoration)	symmetric vacuum returns	high-temperature phase; SSB switches off above the electroweak transition $T_c\sim160$ GeV

Common trapSpontaneous symmetry breaking destroys or violates the symmetry

It does neither. The single most common misreading is to hear "broken" and picture the symmetry being switched off or added-to-the-Lagrangian-with-a-minus-sign. Nothing of the sort happens. The Lagrangian remains exactly invariant under the full group $G$ at every stage — the symmetry is as real as ever and continues to constrain the theory rigidly: it dictates relations among couplings, guarantees the Goldstone bosons are massless, and forbids terms that would otherwise appear. What fails to respect $G$ is only the vacuum — one particular solution of the equations — and the excitations built on it. This is utterly different from explicit breaking, where a non-invariant term is genuinely inserted into the Lagrangian (as a small quark mass explicitly breaks chiral symmetry, giving the pion a small mass and making it a pseudo-Goldstone boson). Spontaneous breaking hides the symmetry; explicit breaking removes it. The tell is that a spontaneously broken symmetry still leaves fingerprints — degenerate vacua, massless Goldstone bosons, mass relations like $M_W=M_Z\cos\theta_W$ — that an explicitly broken one would not. And crucially, this is why SSB can give the $W$ and $Z$ mass: a hand-written mass term $\tfrac12 M^2 W_\mu W^\mu$ is not gauge invariant and would wreck renormalisability, whereas a mass generated by the vacuum keeps the gauge symmetry of the laws intact.

Worked examples

1 · The Higgs self-coupling and the vacuum value (fully worked)

The Fermi constant is $G_F=1.1664\times10^{-5}\ \mathrm{GeV}^{-2}$ and the Higgs boson mass is $m_H=125\ \mathrm{GeV}$ . Treating the electroweak vacuum as the SSB of a scalar potential $V=\mu^2\phi^*\phi+\lambda(\phi^*\phi)^2$ , find the vacuum expectation value $v$ , the quartic self-coupling $\lambda$ , and the parameter $\mu$ .

Step 1. Fix the VEV from the Fermi constant. The charged-current match of the low-energy theory gives $v=(\sqrt2\,G_F)^{-1/2}$ . Numerically $\sqrt2\,G_F=1.41421\times1.1664\times10^{-5}=1.6497\times10^{-5}\ \mathrm{GeV}^{-2}$ , so $v=1/\sqrt{1.6497\times10^{-5}}=\mathbf{246.2\ GeV}$ . This single number sets the scale of electroweak symmetry breaking.

Step 2. Extract the self-coupling from the Higgs mass. The radial (Higgs) mode has $m_H^2=2\lambda v^2$ , so $\lambda=\dfrac{m_H^2}{2v^2}=\dfrac{(125)^2}{2(246.2)^2}=\dfrac{15625}{121229}=\mathbf{0.129}$ . The Higgs self-interaction is weak-to-moderate — comfortably perturbative, but not tiny.

Step 3. Recover the potential parameter. From $v^2=-\mu^2/\lambda$ , $\mu^2=-\lambda v^2=-0.129\times(246.2)^2=-7813\ \mathrm{GeV}^2$ , so $|\mu|=\mathbf{88.4\ GeV}$ . As a check, $m_H^2=-2\mu^2=2\times7813=15625\ \mathrm{GeV}^2$ , giving $m_H=125\ \mathrm{GeV}$ — consistent.

Step 4. Interpret. Three measured quantities — $G_F$ , $m_H$ , and the mechanism — pin down the entire shape of the potential: its depth ( $\mu$ ), its steepness ( $\lambda$ ), and the radius of its trough ( $v$ ). The Higgs boson is nothing but the radial oscillation about the bottom of this hat, and its mass is a direct readout of the wall's curvature. The electroweak vacuum sits at $v=246\ \mathrm{GeV}$ with $\lambda\approx0.13$ and $|\mu|\approx88\ \mathrm{GeV}$ ; the Higgs at 125 GeV is the curvature of the potential wall.

2 · Counting Goldstone bosons in electroweak breaking

The electroweak group $G=SU(2)_L\times U(1)_Y$ is spontaneously broken by a Higgs doublet down to $H=U(1)_{\text{em}}$ . How many would-be Goldstone bosons appear, what happens to them, and how many physical scalars survive? Verify that degrees of freedom are conserved.

Step 1. Count the broken generators. $SU(2)$ has 3 generators and $U(1)$ has 1, so $\dim G=4$ ; the surviving electromagnetic $U(1)_{\text{em}}$ has $\dim H=1$ . By Goldstone's theorem the number of would-be Goldstone bosons is $\dim G-\dim H=4-1=\mathbf{3}$ .

Step 2. Track the scalars. The Higgs field is a complex $SU(2)$ doublet — four real scalar fields. Three of them are the would-be Goldstone bosons; the fourth is the radial mode. Three scalars are eaten, one physical Higgs remains.

Step 3. Track the gauge bosons. The three eaten Goldstones become the longitudinal polarisations of the three gauge bosons associated with the broken generators: $W^+$ , $W^-$ , and $Z$ all become massive ( $M_W=80.4$ , $M_Z=91.2\ \mathrm{GeV}$ ). The gauge boson of the unbroken generator — the photon of $U(1)_{\text{em}}$ — stays massless.

Step 4. Check degrees of freedom. Before: 4 scalar DOF $+$ 4 gauge fields $\times$ 2 polarisations $=12$ . After: 1 physical Higgs $+$ 3 massive vectors $\times$ 3 polarisations $+$ 1 photon $\times$ 2 $=1+9+2=12$ . The books balance exactly — 12 = 12 — with three Goldstones converted into the longitudinal modes of the $W^\pm$ and $Z$ .

Physics of the portrait

Spontaneous symmetry breaking did not begin in particle physics — it began in the theory of magnets and superconductors, and its migration into the vacuum of relativistic field theory is one person's decisive insight. The man who saw that a superconductor's energy gap and a particle's mass are the same phenomenon reshaped how we think about the origin of mass.

Origins · 1 minYoichiro Nambu and how this was discovered

Yoichiro Nambu

1921–2015 · spontaneous symmetry breaking in particle physics and the dynamical origin of mass

Yoichiro Nambu, working at Chicago at the end of the 1950s, was steeped in the just-completed BCS theory of superconductivity, and he noticed something others had missed: the energy gap of a superconductor is a spontaneously broken symmetry, the electromagnetic gauge symmetry hidden by the condensate of Cooper pairs. In 1960 he made the audacious leap of importing this idea into particle physics, proposing that the mass of the nucleon could arise not from a term in the Lagrangian but dynamically, from a symmetric theory whose vacuum breaks chiral symmetry — just as the gap arises in a superconductor. With Giovanni Jona-Lasinio he built an explicit model (the Nambu–Jona-Lasinio model, 1961) in which nucleons acquire mass this way, and the theory made a striking prediction: the breaking of the continuous chiral symmetry must leave behind a nearly massless particle, which he identified with the pion. This was Goldstone's theorem seen in the wild before it was named — the pion is the pseudo-Goldstone boson of broken chiral symmetry, light because the symmetry is only slightly broken explicitly by quark masses. Nambu's reframing — that mass can be an emergent property of an ordered vacuum rather than a fundamental constant — is the conceptual root of the entire Higgs programme. He shared the 2008 Nobel Prize "for the discovery of the mechanism of spontaneous broken symmetry in subatomic physics", awarded almost half a century after the work and four years before the Higgs boson that vindicated it was found.

History

1950

— Ginzburg and Landau describe superconductivity with an order parameter that hides electromagnetic gauge symmetry

1957

— BCS give the microscopic theory of the superconducting condensate

1960

— Nambu imports spontaneous symmetry breaking from superconductivity into particle physics

1961

— Goldstone conjectures that breaking a continuous symmetry produces a massless boson

1962

— Goldstone, Salam and Weinberg prove the theorem in relativistic field theory

1964

— Higgs, Englert and Brout find the gauge loophole that evades the massless boson

Going further

Texts. Thomson, Modern Particle Physics, §17.1–17.4 (spontaneous symmetry breaking, the Goldstone and Higgs mechanisms) · Halzen & Martin, Quarks and Leptons, Ch. 14 (spontaneous symmetry breaking and the Higgs mechanism) · Griffiths, Introduction to Elementary Particles, 2nd ed., §10.8 · Peskin & Schroeder, An Introduction to Quantum Field Theory, §11.1 and §20.1 (Goldstone's theorem and its gauged version) · Schwartz, Quantum Field Theory and the Standard Model, Ch. 28.

Provenance. The gauge-invariance constraint that forbids hand-written mass terms is recalled from aq-pp-gauge-boson-mass-problem and aq-pp-local-gauge-principle; the chiral structure that makes fermion mass terms non-invariant from aq-pp-weyl-chirality; and the Lagrangian and covariant-derivative machinery from aq-pp-lagrangians-gauge-invariance. This article establishes the mechanism; its application to the electroweak sector — the gauge-boson masses, $M_W=\tfrac12 gv$ , and the physical Higgs couplings — is developed in aq-pp-higgs-mechanism, aq-pp-higgs-couplings, and aq-pp-electroweak-unification. A verifier should confirm: (1) $v^2=-\mu^2/\lambda$ minimises (1); (2) $m_\eta^2=2\lambda v^2=-2\mu^2$ , $m_\xi^2=0$ from the expansion; (3) $v=(\sqrt2 G_F)^{-1/2}=246\ \mathrm{GeV}$ , $\lambda=m_H^2/2v^2=0.129$ ; (4) the DOF count $12=12$ across electroweak breaking; (5) $\dim G-\dim H=3$ broken generators for $SU(2)_L\times U(1)_Y\to U(1)_{\text{em}}$ .

Prerequisites. Lagrangians and Gauge Invariance · The Local Gauge Principle · Weyl Chirality · The Gauge Boson Mass Problem. Leads to. The Higgs Mechanism · Higgs Couplings · Electroweak Unification.

Ask the tutor: Ask me to roll the marble down the hat line by line — "expand the complex scalar about its vacuum and show me the Goldstone boson appear" — and we will minimise the potential, split the field into radial and angular modes, and watch one mass vanish while the other is set by the curvature of the wall.

What is the difference between spontaneous and explicit symmetry breaking?

They are opposite situations that share a word. In explicit breaking, you actually put a term into the Lagrangian that is not invariant under the symmetry — the laws themselves lose the symmetry, exactly and by construction. A small quark mass, for instance, explicitly breaks the chiral symmetry of QCD. In spontaneous breaking, the Lagrangian stays perfectly invariant; only the vacuum state fails to respect the symmetry, because it has to pick one option out of a degenerate family. The distinction has sharp consequences: a spontaneously broken continuous symmetry leaves behind massless Goldstone bosons and exact mass relations, whereas an explicitly broken one gives the "Goldstone" a mass (a pseudo-Goldstone boson) and relaxes those relations. The pion shows both at once — it is a pseudo-Goldstone boson, light because chiral symmetry is mostly spontaneously broken but slightly explicitly broken by quark masses.

If every vacuum on the circle is equivalent, which one does nature actually pick?

It picks one at random, and which one is not physically meaningful. Every point on the degenerate circle is related to every other by the symmetry, so they are genuinely indistinguishable — no experiment inside the theory can tell you the value of the phase $\alpha$ that labels the chosen vacuum. What is physical is that a choice was made: once the system settles into a particular minimum, all the physics (masses, couplings, which mode is the Goldstone) is fixed, and different choices give identical physics. In an infinite system the choice is also permanent — tunnelling between vacua is completely suppressed because it would require reorienting infinitely many degrees of freedom at once — which is why the broken vacuum is stable and the symmetry stays hidden rather than being restored by quantum fluctuations. In a finite system, by contrast, the "choice" can fluctuate, which is why true SSB strictly requires the infinite-volume limit.

What exactly is a Goldstone boson, and have we ever seen one?

A Goldstone boson is the massless, spinless excitation that appears when a continuous global symmetry is spontaneously broken — physically, it is the field that oscillates along the flat valley of the potential, the direction in which the symmetry moves the vacuum for free. Because that direction is exactly flat, the excitation costs vanishing energy at long wavelength and so has zero mass. We have seen them, though usually in slightly disguised, "pseudo" form. The clearest particle-physics example is the pion: it is the (pseudo-)Goldstone boson of spontaneously broken chiral symmetry in QCD, light precisely because that symmetry is almost exact. In condensed matter, Goldstone modes are ubiquitous — the spin waves (magnons) of a ferromagnet and the phonons of a crystal are Goldstone bosons of broken rotational and translational symmetry. In the electroweak theory the would-be Goldstone bosons are not seen as particles, because that symmetry is gauged and they are eaten by the $W$ and $Z$ .

Why does the photon stay massless while the W and Z get heavy?

Because the electroweak breaking is only partial. The Higgs vacuum breaks the group $SU(2)_L\times U(1)_Y$ down to a smaller surviving symmetry, $U(1)_{\text{em}}$ — the gauge symmetry of electromagnetism. The three gauge bosons associated with the broken generators eat Goldstone bosons and become massive: these are the $W^+$ , $W^-$ , and $Z$ . But the one generator that survives — electric charge — still leaves the vacuum invariant, so its gauge boson, the photon, has no Goldstone boson to eat and stays exactly massless. In the language of the mass matrix, the photon is the particular combination of neutral gauge fields for which the vacuum is neutral (uncharged), so the Higgs condensate is invisible to it. This is why electromagnetism remains a long-range force while the weak force is short-range: the vacuum broke the symmetry down to, but not below, the electromagnetic $U(1)$ .

Does spontaneous symmetry breaking explain the mass of everyday matter?

Mostly no, and this surprises people. The Higgs-related SSB gives mass to the elementary particles — the quarks, the charged leptons, the $W$ and $Z$ — but the mass of a proton or neutron is about 938 MeV, whereas the up and down quarks inside weigh only a few MeV each. More than 99% of a nucleon's mass is not Higgs mass at all: it is the energy of the gluon fields and quark motion bound up inside, generated by a different spontaneous symmetry breaking — the chiral symmetry breaking of QCD, which is dynamical and owes nothing to the Higgs field. So the atoms in your body get almost all of their mass from confined QCD energy (via $E=mc^2$ ), and only a sliver from the electroweak SSB. Both are spontaneous symmetry breaking; they are simply different instances of the same mechanism operating at different scales.

Is the Mexican-hat potential a real object or just a cartoon?

It is a real, quantitative feature of the theory, not merely a picture — though it is drawn schematically. The potential $V(\phi)=\mu^2\phi^*\phi+\lambda(\phi^*\phi)^2$ is a genuine term in the Standard-Model Lagrangian, and its parameters are measured: the depth is set by $|\mu|\approx88\ \mathrm{GeV}$ , the trough radius by the vacuum value $v=246\ \mathrm{GeV}$ , and the wall curvature by the Higgs mass $m_H=125\ \mathrm{GeV}$ , which fixes $\lambda\approx0.13$ . What is schematic is the drawing: the real field space is higher-dimensional (a complex doublet, four real directions), so the true "hat" cannot be pictured directly, and the vertical axis is energy density rather than a height. But the essential features — an unstable symmetric point, a degenerate ring of minima, a steep radial mode and flat angular modes — are exact consequences of the actual potential, and measuring the cubic and quartic self-couplings at colliders is a direct test of the hat's shape.

How does spontaneous symmetry breaking connect to phase transitions and the order parameter?

They are the same phenomenon in different languages. The vacuum expectation value $v=\langle\phi\rangle$ is an order parameter: it is zero in the symmetric phase and nonzero in the broken phase, exactly like the magnetisation of a ferromagnet, which vanishes above the Curie temperature and turns on below it. As you cool the system (or, in the early universe, as it expanded and cooled), a control parameter — effectively the sign of $\mu^2$ , which depends on temperature — changes, and the vacuum tips from the symmetric point into the broken ring. That is a phase transition, and the whole Ginzburg–Landau apparatus of order parameters, critical temperatures, and latent heat carries straight over. This is not a loose analogy: Nambu's original insight was precisely that the particle-physics vacuum and the superconductor are the same kind of ordered state, and the electroweak transition at $T_c\sim160\ \mathrm{GeV}$ in the early universe is a genuine cosmological phase transition of exactly this type.