The weak interactions of quarks

The weak force is famously egalitarian among leptons: the coupling at the electron, muon and tau vertices is one and the same number, $G_F^{(e)}=G_F^{(\mu)}=G_F^{(\tau)}$ . Turn that measurement on quarks and the tidy picture cracks. Nuclear $\beta$ -decay, which reads the $ud$ vertex, returns an effective coupling about 5% below the muonic value — $G_F^{(\beta)}=1.1066\times10^{-5}\ \mathrm{GeV^{-2}}$ against $G_F^{(\mu)}=1.1664\times10^{-5}\ \mathrm{GeV^{-2}}$ — and the strangeness-changing $us$ vertex misbehaves far worse: the leptonic decay $K^-\to\mu^-\bar\nu_\mu$ runs roughly twenty times slower than a universal coupling would demand. This article is the synthesis that rescues universality. Nicola Cabibbo's 1963 hypothesis is that the weak force has exactly the same intrinsic strength for quarks as for leptons, but that the states plugged into the charged current — the weak eigenstates $d'$ and $s'$ — are the mass eigenstates $d$ and $s$ rotated through a single angle, the Cabibbo angle $\theta_c\simeq13^\circ$ . That one rotation makes $\beta$ -decay carry a factor $\cos\theta_c$ and strangeness-changing decays a factor $\sin\theta_c$ , reconciling both anomalies at a stroke. Puzzles over the absence of flavour-changing neutral currents — why $K_L\to\mu^+\mu^-$ is a billion times rarer than a naive estimate — then forced a fourth quark into existence through the GIM mechanism, and the same algebra generalizes to the three-generation CKM matrix that carries all quark-flavour physics today.

Key takeaways

Nuclear $\beta$ -decay shows the $ud$ coupling ~5% below the leptonic $G_F$ , and the strangeness-changing $us$ coupling suppressed ~20-fold — universality appears broken.
The Cabibbo hypothesis restores it: the weak force is universal, but the down-type weak eigenstates $d',s'$ are the mass eigenstates rotated through one angle $\theta_c$ .
Amplitudes then carry $\cos\theta_c$ (strangeness-conserving) or $\sin\theta_c$ (strangeness-changing); rates go as $\cos^2\theta_c$ and $\sin^2\theta_c$ , so $\theta_c\simeq13^\circ$ fixes both anomalies.
Flavour-changing neutral currents are absent because a unitary rotation leaves the neutral current diagonal, $\bar{d'}d'+\bar{s'}s'=\bar dd+\bar ss$ .
The GIM mechanism completes the cancellation at loop level via the charm quark; the up and charm boxes for $K_L\to\mu^+\mu^-$ nearly cancel, leaving a residue $\propto(m_c^2-m_u^2)/M_W^2$ .
Key idea: the weak force couples universally, but quarks enter through weak eigenstates that are rotated combinations of mass eigenstates — one rotation explains the suppression of strangeness-changing decays and, with a fourth quark, the absence of flavour-changing neutral currents.

The core insight

The weak force couples with one universal strength, but quarks participate through weak eigenstates that are rotated combinations of the mass eigenstates; this single Cabibbo rotation simultaneously explains the apparent suppression of strangeness-changing decays and, once a fourth quark is added, the absence of flavour-changing neutral currents.

Where this lives today

Cabibbo's single angle is now the $2\times2$ corner of the $3\times3$ Cabibbo–Kobayashi–Maskawa (CKM) matrix, and pinning its elements is a central industry of flavour physics. $V_{ud}$ is fixed to five decimals by superallowed $0^+\to0^+$ $\beta$ -decays and $V_{us}$ by kaon decays; together they give the first-row unitarity test $|V_{ud}|^2+|V_{us}|^2+|V_{ub}|^2=1$ , whose current ~2–3σ shortfall — the "Cabibbo anomaly" — is a live hint of new physics. GIM remains the reason FCNC are the sharpest windows beyond the Standard Model: NA62 and KOTO hunt the ultra-rare $K^+\to\pi^+\nu\bar\nu$ and $K_L\to\pi^0\nu\bar\nu$ , computable precisely because GIM suppresses them, so any excess flags new virtual particles; at the LHC, LHCb scrutinizes the analogous $b\to s\ell^+\ell^-$ transitions. And the three-generation extension supplies, through its one irreducible phase, the only source of CP violation in the quark sector — the ingredient any theory of the cosmic matter asymmetry must reckon with.

The physical picture

Imagine measuring one universal weak coupling and getting different answers by transition: $\beta$ -decay ( $d\to u$ ) reads a little low, strangeness-changing $s\to u$ reads wildly low, the leptonic value is the reference. Cabibbo's resolution is not three couplings but one: the up quark's weak partner is a single fixed direction in the $(d,s)$ plane — the weak eigenstate $d'$ — and every rate is the projection of that direction onto the mass state you detect. The first figure is that rotation, drawn to scale at $\theta_c\simeq13^\circ$ , so you can read off both the large $\cos\theta_c$ (favoured) and small $\sin\theta_c$ (suppressed) couplings.

Figure

The Cabibbo rotation: the down-type weak eigenstate d′ that the u quark couples to is the mass eigenstates d and s rotated through θc ≈ 13.0°, so the strangeness-conserving u→d coupling is cos θc = 0.974 while the strangeness-changing u→s coupling is only sin θc = 0.225

The rotation resolves the suppression, but it does something worse before something better. Because the up quark now couples to both $d$ and $s$ , a second-order weak process can turn an $s$ into a $d$ with no change of charge — a flavour-changing neutral current. Taken literally, the single up-quark box predicts $K_L\to\mu^+\mu^-$ far larger than observed. Glashow, Iliopoulos and Maiani saw that a fourth quark, charm, coupling to the orthogonal combination $s'=-\sin\theta_c\,d+\cos\theta_c\,s$ , contributes a second box of equal magnitude and opposite sign. The two nearly annihilate, leaving a residue that vanishes as the charm and up masses become equal. The second figure is that cancellation — the content of GIM — with the measured rarity of the decay as its punchline.

Figure

The GIM cancellation for the flavour-changing decay KL → μ⁺μ⁻: the up-quark box and the charm-quark box carry equal and opposite amplitudes ± cos θc sin θc, so their sum collapses to a residue proportional to (mc² − mu²)/MW², leaving the measured branching ratio at only 6.84 × 10⁻⁹

The core equations

The whole subject follows from one hypothesis stated as one equation: the down-type weak eigenstates are a rotation of the mass eigenstates. Everything else — couplings, rates, FCNC suppression — is arithmetic on this rotation.

\begin{pmatrix} d'\\ s'\end{pmatrix}=\begin{pmatrix}\cos\theta_c & \sin\theta_c\\ -\sin\theta_c & \cos\theta_c\end{pmatrix}\begin{pmatrix} d\\ s\end{pmatrix} \qquad (1)

where: $d,s$ — the down-type mass eigenstates (definite mass, the states that propagate and hadronize); $d',s'$ — the weak eigenstates in the charged current; $\theta_c\simeq13^\circ$ — the Cabibbo angle. The matrix is orthogonal, so no probability is lost: the up quark's coupling is redistributed between $d$ and $s$ , never diminished. The charged current is built exactly as for leptons — a left-handed V−A current — but with the up-type quarks paired to the primed down-type states. This one line encodes quark weak interactions.

J_\mu^{+}=\bar u\,\gamma_\mu\tfrac12(1-\gamma^5)\,d'+\bar c\,\gamma_\mu\tfrac12(1-\gamma^5)\,s' \qquad (2)

where: $\bar u,\bar c$ — up-type quark fields; $\gamma_\mu\tfrac12(1-\gamma^5)$ — the V−A structure inherited unchanged from the leptonic current; $d',s'$ — the rotated states of equation (1). The coupling multiplying the whole current is the same $g_W$ as for leptons — the content of universality. Expanding $d',s'$ in mass eigenstates reads off the physical vertex couplings — the elements of the $2\times2$ Cabibbo matrix, ancestor of the CKM matrix.

V_{ud}=\cos\theta_c,\quad V_{us}=\sin\theta_c,\quad V_{cd}=-\sin\theta_c,\quad V_{cs}=\cos\theta_c \qquad (3)

where: each $V_{qq'}$ is the amplitude weight for the transition between up-type $q$ and down-type $q'$ . With $\cos\theta_c=0.974$ , $\sin\theta_c=0.225$ , the diagonal (within-generation) transitions $ud$ and $cs$ are favoured; the off-diagonal $us$ and $cd$ are Cabibbo-suppressed. Because a rate goes as the amplitude squared, these couplings map onto observable rates through their squares — making the angle measurable from lifetimes and branching ratios.

\Gamma(\Delta S=0)\propto\cos^2\theta_c,\qquad \Gamma(\Delta S=1)\propto\sin^2\theta_c,\qquad \frac{\Gamma(K^-\to\mu^-\bar\nu_\mu)}{\Gamma(\pi^-\to\mu^-\bar\nu_\mu)}\bigg|_{\text{coupling}}=\tan^2\theta_c \qquad (4)

where: $\Delta S$ — the change in strangeness; strangeness-conserving $ud$ transitions ( $\beta$ -decay, $\pi$ decay) carry $\cos^2\theta_c\approx0.95$ , strangeness-changing $us$ transitions ( $K$ , hyperon decay) carry $\sin^2\theta_c\approx0.05$ . The pure-coupling ratio $\tan^2\theta_c\approx1/18.5$ is dressed by phase space and $f_K/f_\pi$ in a real measurement. The neutral current tells the opposite story: written in the weak basis and rotated back, the flavour-changing cross terms cancel identically — the algebraic heart of GIM and the reason the $Z$ has no tree-level $\bar ds$ coupling.

\bar{d'}d'+\bar{s'}s'=\bar dd+\bar ss \qquad (5)

where: the left side is the neutral current in weak eigenstates; the right side, from substituting equation (1), contains no $\bar ds$ or $\bar sd$ term. Unitarity guarantees the off-diagonal pieces $\pm\sin\theta_c\cos\theta_c\,\bar ds$ cancel: neutral currents are flavour-diagonal at tree level. Flavour-changing neutral processes therefore appear only at loop level, where GIM suppresses them to a residue set by the up-type quark mass splitting.

\mathcal{M}(s\bar d\to\mu^+\mu^-)\;\propto\;\sin\theta_c\cos\theta_c\left(\,f(m_u)-f(m_c)\,\right)\;\xrightarrow{\ m_i\ll M_W\ }\;\sin\theta_c\cos\theta_c\,\frac{m_c^2-m_u^2}{M_W^2} \qquad (6)

where: $f(m_i)$ — the loop function for an internal up-type quark of mass $m_i$ ; the two contributions enter with opposite sign because $V_{cd}V_{cs}=-V_{ud}V_{us}$ . If $m_u=m_c$ the amplitude vanishes exactly; the residue $\propto(m_c^2-m_u^2)/M_W^2$ is how GIM predicted a light charm quark before it was found.

Derivation AReading the Cabibbo angle from two independent anomalies

The goal is to show that the single number $\theta_c$ is over-determined: the ~5% deficit in the $\beta$ -decay coupling and the ~20-fold suppression of strangeness-changing kaon decay are two measurements of the same angle, and must agree. We extract $\theta_c$ from each observable and check that the values coincide near $13^\circ$ — the consistency that turned Cabibbo's hypothesis from a guess into a law.

Step 1. Nuclear $\beta$ -decay runs on the $d\to u$ transition, so its effective coupling is the leptonic one reduced by the strangeness-conserving factor: $G_F^{(\beta)}=G_F^{(\mu)}\cos^2\theta_c$ , hence $\cos^2\theta_c=G_F^{(\beta)}/G_F^{(\mu)}=1.1066/1.1664=0.9488$ . Why: the $ud$ vertex projects the up quark's weak partner $d'$ onto the detected state $d$ , so the deficit from unity is a redirected coupling, not a broken one.

Step 2. Root and invert: $\cos\theta_c=0.974$ , so $\theta_c=13.1^\circ$ , with $\sin\theta_c=0.225$ , $\sin^2\theta_c=0.051$ . The "missing" 5% is precisely $\sin^2\theta_c$ — the fraction of $d'$ pointing along $s$ . Why: $\cos^2\theta_c+\sin^2\theta_c=1$ , so the deficit in one channel is exactly the strength gone into the other.

Step 3. Now use a different process. The strangeness-changing $K^-\to\mu^-\bar\nu_\mu$ runs on the $us$ vertex ( $\propto\sin^2\theta_c$ ) while $\pi^-\to\mu^-\bar\nu_\mu$ runs on $ud$ ( $\propto\cos^2\theta_c$ ), so their coupling ratio is $\tan^2\theta_c$ ; at $\theta_c=13.1^\circ$ this is $0.054\approx1/18.5$ . Why: the same two projections that shared out the $\beta$ -decay coupling now appear as the numerator and denominator of a decay ratio.

Step 4. Compare with experiment: the leptonic $K/\pi$ suppression is "roughly twenty-fold" once decay constants and phase space are stripped away, agreeing with $1/18.5$ from an angle fixed by an independent $\beta$ -decay measurement. Two anomalies, one angle. Why: a single parameter that predicts a second, unrelated number is a theory, not a fit — which is why $\theta_c\simeq13^\circ$ was believed.

Go deeper · Derivation B · 1 minWhy a unitary rotation forbids flavour-changing neutral currents

Derivation BWhy a unitary rotation forbids flavour-changing neutral currents

The aim is to prove equation (5) explicitly and expose the deepest structural fact of the quark weak sector: rotating the down-type states scrambles the charged current but leaves the neutral current perfectly flavour-diagonal, so there is no tree-level $Z\to\bar ds$ vertex and no $\Delta S=1$ neutral process. We write the neutral current in the weak basis, substitute the rotation, and watch the cross terms annihilate by orthogonality.

Step 1. The neutral current couples each weak eigenstate to itself: schematically $J_\mu^0\sim\bar{d'}\Gamma_\mu d'+\bar{s'}\Gamma_\mu s'$ , with $\Gamma_\mu$ the flavour-blind Lorentz-and-chirality structure. Factor it out and study the flavour bilinear $\bar{d'}d'+\bar{s'}s'$ . Why: the $Z$ cannot tell $d$ from $s$ , so the whole question is whether the sum over weak states leaves any flavour-changing cross term.

Step 2. Substitute equation (1): $d'=\cos\theta_c\,d+\sin\theta_c\,s$ and $s'=-\sin\theta_c\,d+\cos\theta_c\,s$ . The first bilinear is $\bar{d'}d'=\cos^2\theta_c\,\bar dd+\sin^2\theta_c\,\bar ss+\sin\theta_c\cos\theta_c(\bar ds+\bar sd)$ . Why: $d'$ is a genuine mixture, so squaring it produces flavour-off-diagonal $\bar ds$ pieces — at this stage FCNC looks unavoidable.

Step 3. The second bilinear is $\bar{s'}s'=\sin^2\theta_c\,\bar dd+\cos^2\theta_c\,\bar ss-\sin\theta_c\cos\theta_c(\bar ds+\bar sd)$ , whose cross term carries the opposite sign because $s'$ is orthogonal to $d'$ . Why: the minus sign is not an accident — it is the defining orthogonality of the two rows of a rotation matrix.

Step 4. Add them: the diagonal pieces combine as $(\cos^2\theta_c+\sin^2\theta_c)(\bar dd+\bar ss)=\bar dd+\bar ss$ , and the $\bar ds$ cross terms cancel identically. There is no neutral coupling that changes strangeness. Why: FCNC are forbidden not by fine-tuning but by the unitarity of mixing, so any observed FCNC must be a small loop effect.

Go deeper · Derivation C · 1 minThe charm box, the up box, and the mass that had to exist

Derivation CThe charm box, the up box, and the mass that had to exist

The purpose is to trace how the tree-level prohibition survives at loop level only because a fourth quark exists. Box diagrams generate an effective $s\to d$ neutral transition, and with only the up quark they predict $K_L\to\mu^+\mu^-$ far too large. We show that the charm quark supplies a second box with opposite coupling sign, that the two cancel exactly for equal masses, and that the residue bounds the charm mass — the logic by which GIM predicted charm in 1970.

Step 1. Build the box for $s\bar d\to\mu^+\mu^-$ with an internal up quark. Its two charged-current vertices carry $V_{us}^{*}V_{ud}=\sin\theta_c\cos\theta_c$ , so $\mathcal{M}_u\propto+\sin\theta_c\cos\theta_c\,f(m_u)$ , with $f$ the loop integral over the internal quark of mass $m_u$ . Why: an FCNC needs two charged-current insertions to move $s\to u\to d$ , so the amplitude is quadratic in the mixing and its sign is set by the product of Cabibbo factors.

Step 2. Add the charm box, identical in topology but with charm internal. Its vertices carry $V_{cs}^{*}V_{cd}=\cos\theta_c\cdot(-\sin\theta_c)=-\sin\theta_c\cos\theta_c$ — equal in magnitude, opposite in sign — so $\mathcal{M}_c\propto-\sin\theta_c\cos\theta_c\,f(m_c)$ . Why: charm couples to the orthogonal state $s'$ , and orthogonality flips the Cabibbo sign — the same unitarity that killed the tree-level term returns at one loop.

Step 3. Sum the boxes: $\mathcal{M}_u+\mathcal{M}_c\propto\sin\theta_c\cos\theta_c\,[\,f(m_u)-f(m_c)\,]$ . Equal internal masses would make the loop functions identical and the amplitude vanish exactly; unequal masses leave a residue set by the difference. Why: identical masses give perfect GIM cancellation, and reality gives a small controlled leak.

Step 4. Expand for $m_i\ll M_W$ : $f(m_i)\approx\text{const}+m_i^2/M_W^2$ , so $\mathcal{M}_u+\mathcal{M}_c\propto\sin\theta_c\cos\theta_c\,(m_c^2-m_u^2)/M_W^2$ . The smallness of $B(K_L\to\mu^+\mu^-)=6.84\times10^{-9}$ then bounds $m_c$ ; GIM put charm near $1.5\ \mathrm{GeV}$ , and the $J/\psi$ found in 1974 gave $m_c\approx1.27\ \mathrm{GeV}$ . Why: the absence of a process forecast the existence and rough mass of a new particle — one of the cleanest predictions in Standard Model history.

Limiting cases

Case	Result	Meaning
$\theta_c\to0$	$d'=d,\ s'=s$	no mixing; generations closed, strangeness-changing decays forbidden, $K^-\to\mu^-\bar\nu$ cannot occur
$\theta_c$ small (reality)	$\cos\theta_c\approx1,\ \sin\theta_c\approx0.22$	strong hierarchy: within-generation favoured, cross-generation suppressed by $\sin^2\theta_c\approx0.05$
$\theta_c\to45^\circ$	maximal mixing	$d$ and $s$ enter equally; not realized for quarks (it is for some neutrino mixings)
$m_c=m_u$ (exact GIM)	$\mathcal{M}_{\text{FCNC}}=0$	perfect box cancellation; FCNC vanish at this level
$m_c\gg m_u$ (real quarks)	residue $\propto m_c^2/M_W^2$	GIM suppression strong but incomplete; sets the tiny FCNC rates and once bounded $m_c$
Two generations, real matrix	one angle $\theta_c$ , no phase	no CP violation; that requires the third generation (CKM)

Common trapThe ud coupling is genuinely weaker than the leptonic coupling

It is not. The intrinsic weak coupling $g_W$ is one universal number, the same at every vertex. Nuclear $\beta$ -decay measures not a weaker force but a projection: only the component of the weak partner $d'$ along the detected state $d$ contributes, namely $\cos\theta_c$ . The 5% "deficit" in $G_F^{(\beta)}$ is exactly $\sin^2\theta_c$ , the coupling redirected into strangeness-changing $us$ transitions — add the $\beta$ -decay rate ( $\propto\cos^2\theta_c$ ) and the strange rate ( $\propto\sin^2\theta_c$ ) back and you recover the full universal strength. The coupling was never lost; it was rotated. Reading the 5% as broken universality misses the whole point of Cabibbo's hypothesis: to save universality by relocating the discrepancy into the definition of the quark states.

Worked examples

1 · Extracting the Cabibbo angle from the β-decay deficit (fully worked)

Nuclear $\beta$ -decay returns $G_F^{(\beta)}=1.1066\times10^{-5}\ \mathrm{GeV^{-2}}$ , muon decay gives $G_F^{(\mu)}=1.1664\times10^{-5}\ \mathrm{GeV^{-2}}$ . Find the Cabibbo angle, the vertex couplings, and predict the strangeness-changing suppression.

Step 1. The $\beta$ -decay coupling is the universal one reduced by the strangeness-conserving projection, $G_F^{(\beta)}=G_F^{(\mu)}\cos^2\theta_c$ . Form the ratio: $\cos^2\theta_c=1.1066/1.1664=\mathbf{0.9488}$ — the fraction of the $ud$ rate that survives, close to but below unity.

Step 2. Take the root: $\cos\theta_c=\sqrt{0.9488}=0.9740$ , so $\theta_c=\arccos(0.9740)=\mathbf{13.1^\circ}$ . The complementary coupling is $\sin\theta_c=\sqrt{1-0.9488}=\sqrt{0.0512}=\mathbf{0.226}$ .

Step 3. The vertex couplings are $V_{ud}=\cos\theta_c=\mathbf{0.974}$ (favoured) and $V_{us}=\sin\theta_c=\mathbf{0.225}$ (suppressed). The strangeness-changing rate is suppressed by $\tan^2\theta_c=(0.225/0.974)^2=\mathbf{0.054}\approx1/18.5$ .

Step 4. Sanity check against a different observable: the leptonic decay $K^-\to\mu^-\bar\nu$ is observed to run roughly twenty times slower than a universal coupling predicts, matching $1/\tan^2\theta_c\approx18.5$ once decay constants and phase space are removed. An angle fixed by $\beta$ -decay correctly predicts a kaon decay. θc = 13.1°, Vud = 0.974, Vus = 0.225, strangeness-changing suppression ≈ 1/18.5.

2 · The K → μν versus π → μν leptonic ratio

Both $K^-(u\bar s)\to\mu^-\bar\nu_\mu$ and $\pi^-(u\bar d)\to\mu^-\bar\nu_\mu$ are leptonic meson decays. From the Cabibbo couplings alone, by what factor is the kaon channel suppressed, and why is the observed ratio not exactly this number?

Step 1. The pion decay runs on $ud$ ( $\propto\cos\theta_c$ ), the kaon on $us$ ( $\propto\sin\theta_c$ ), so at the coupling level the rate ratio is $\Gamma_K/\Gamma_\pi\big|_{\text{coupling}}=\sin^2\theta_c/\cos^2\theta_c=\tan^2\theta_c$ .

Step 2. Numerically $\tan^2(13.1^\circ)=\mathbf{0.054}$ : the kaon leptonic channel is suppressed roughly 20-fold purely by the Cabibbo rotation — the most visible fingerprint of strangeness-changing suppression.

Step 3. The measured ratio differs because two non-weak factors intervene: the decay constants ( $f_K/f_\pi\approx1.19$ ) and phase space, which favours the heavier kaon ( $m_K\approx494\ \mathrm{MeV}$ versus $m_\pi\approx140\ \mathrm{MeV}$ ) — both enhance the kaon rate and partly offset the suppression. The clean coupling suppression is tan²θc ≈ 1/18.5; f_K/f_π and phase space dress the observed ratio.

3 · Cabibbo universality as a unitarity test

The mixing matrix is orthogonal, so its first row satisfies $|V_{ud}|^2+|V_{us}|^2=1$ . Using the modern values $|V_{ud}|=0.97373$ and $|V_{us}|=0.2243$ , test this and comment.

Step 1. Square and add: $|V_{ud}|^2=0.97373^2=0.94815$ , $|V_{us}|^2=0.2243^2=0.05031$ , so the sum is $\mathbf{0.99846}$ .

Step 2. The deficit from unity is $1-0.99846=\mathbf{0.00154}$ . In a strictly two-generation world this must be exactly zero; the shortfall is partly the tiny $|V_{ub}|^2\approx1.5\times10^{-5}$ opening the row to the third generation, and partly a genuine ~2–3σ tension known as the Cabibbo anomaly.

Step 3. What began as Cabibbo's device to save universality is now one of the most precise Standard Model tests: first-row CKM unitarity is measured to the fourth decimal, and any confirmed deviation would signal new physics. |Vud|² + |Vus|² = 0.99846, a 0.15% test of unitarity — the modern descendant of Cabibbo's 5% deficit.

Physics of the portrait

The Cabibbo angle is where the messy phenomenology of strange-particle decays became a single geometric idea — named for the theorist who saw the rotation before there were enough quarks to rotate.

Origins · 1 minNicola Cabibbo and how this was discovered

Nicola Cabibbo

1935-2010 · the weak interactions of quarks and quark flavour mixing

Nicola Cabibbo was an Italian theoretical physicist who, in a single 1963 paper, imposed order on the tangle of strange-particle weak decays. Experimenters had catalogued hyperon and kaon decays whose rates refused to match the universal coupling seen in muon and neutron decay; the strangeness-changing ones were suppressed by a consistent factor. Cabibbo's insight was to write the up quark's weak partner as a rotation, $d'=d\cos\theta_c+s\sin\theta_c$ , so one angle governed every rate at once — reducing a table of anomalies to a single number, $\theta_c\simeq13^\circ$ . He did this before the quark model was secure and years before the fourth quark existed even on paper. In 1970 Glashow, Iliopoulos and Maiani grafted charm onto his framework to explain the absence of flavour-changing neutral currents, and in 1973 Kobayashi and Maskawa generalized the single angle to the $3\times3$ CKM matrix, whose irreducible complex phase gives CP violation — with Cabibbo's rotation as its upper-left corner. That Kobayashi and Maskawa shared the 2008 Nobel Prize while Cabibbo, whose angle underpins the whole edifice, did not, is among the most debated omissions in the prize's history. He is remembered as the man who taught quarks to mix.

History

1963

— Cabibbo introduces the mixing angle θc, saving weak universality for quarks

1964

— Bjorken and Glashow propose a fourth quark, charm, on lepton-hadron symmetry grounds

1970

— Glashow, Iliopoulos and Maiani introduce the GIM mechanism to suppress FCNC

1973

— Kobayashi and Maskawa extend mixing to three generations, predicting CP violation

1974

— The J/ψ is discovered, confirming the charm quark at m ≈ 1.3 GeV

Going further

Texts. Thomson, Modern Particle Physics, §14.2 "The weak interactions of quarks" and §12.1 (universal leptonic weak coupling) · Halzen & Martin, Quarks and Leptons, §12 (the Cabibbo angle and GIM) · Griffiths, Introduction to Elementary Particles, 2nd ed., §9.6–9.7 · Peskin & Schroeder, An Introduction to Quantum Field Theory, §20.3 (quark mixing and the CKM matrix) · Schwartz, Quantum Field Theory and the Standard Model, §29 (flavour physics). Key numerical checks: (1) $\cos^2\theta_c=G_F^{(\beta)}/G_F^{(\mu)}=0.9488\Rightarrow\theta_c=13.1^\circ$ ; (2) $\bar{d'}d'+\bar{s'}s'=\bar dd+\bar ss$ with cross terms cancelling; (3) $\tan^2(13.1^\circ)=0.054\approx1/18.5$ ; (4) $|V_{ud}|^2+|V_{us}|^2=0.99846$ . The couplings $\cos\theta_c=0.974$ , $\sin\theta_c=0.225$ and the branching ratio $B(K_L\to\mu^+\mu^-)=6.84\times10^{-9}$ are standard tabulated values.

Prerequisites. The Weak Charged Current · The Fermi Constant · Two-Flavour Oscillations and Mixing. Leads to. The Cabibbo Angle · The GIM Mechanism · The CKM Matrix.

Ask the tutor: ask me how Cabibbo's single angle grows into the three-generation CKM matrix and where its one CP-violating phase comes from.

If the weak force is universal, why does β-decay measure a smaller coupling?

Because the coupling nuclear $\beta$ -decay reports is not the intrinsic weak coupling but its projection onto a quark transition. The up quark's weak partner is the fixed eigenstate $d'$ , pointing mostly along $d$ but slightly along $s$ . A strangeness-conserving decay picks up only the $d$ -component, weight $\cos\theta_c$ , so the rate carries $\cos^2\theta_c\approx0.95$ . The missing 5% is exactly $\sin^2\theta_c$ , the strength that reappears in strangeness-changing decays like $K^-\to\mu^-\bar\nu$ . Add both channels back and you recover the full coupling: universality is a statement about $g_W$ , not any single rate.

What is the difference between a weak eigenstate and a mass eigenstate?

A mass eigenstate is a quark with a definite mass — the state that propagates, hadronizes, and appears in the particle tables ( $d$ , $s$ , $b$ ). A weak eigenstate is the combination that couples to the $W$ at a charged-current vertex ( $d'$ , $s'$ , $b'$ ). Nature does not align the two: the weak eigenstates are the mass eigenstates rotated by the Cabibbo angle (two generations) or the full CKM matrix (three). So a $W$ emits a superposition of mass eigenstates, and the quark that later propagates can be a different flavour than nominally emitted. This mismatch is the entire origin of quark flavour mixing.

Why did the absence of KL → μ⁺μ⁻ require a whole new quark?

With only three quarks the decay is not suppressed enough. The Cabibbo current lets an $s$ turn into a $d$ through a virtual up quark in a box diagram, an effective FCNC that would make $K_L\to\mu^+\mu^-$ far more common than observed. Glashow, Iliopoulos and Maiani saw the natural fix was a second up-type quark — charm — coupling to the orthogonal combination $s'$ . Its box carries the opposite Cabibbo sign, so the two nearly cancel, leaving a residue $\propto(m_c^2-m_u^2)/M_W^2$ . Demanding this residue match the tiny observed rate even predicted roughly where the charm mass had to be, years before the $J/\psi$ was found in 1974.

Is the Cabibbo angle a fundamental constant of nature?

In the Standard Model it is a free parameter — not predicted but measured, like a quark mass. It arises from the misalignment between how quarks acquire mass (Yukawa couplings to the Higgs) and how they couple to the $W$ , and the theory offers no reason for its value of about $13^\circ$ . In three generations it becomes three angles plus one CP-violating phase — the four numbers of the CKM matrix. Why these angles are small and hierarchical is an open question, part of the unsolved "flavour puzzle" that also asks why the quark and lepton masses span so many orders of magnitude.

How does the two-generation Cabibbo picture become the CKM matrix?

By adding the third generation. Cabibbo's $2\times2$ rotation mixes $(d,s)$ through one angle; Kobayashi and Maskawa in 1973 showed three generations require a $3\times3$ unitary matrix with three angles and one irreducible complex phase that cannot be rotated away. That phase is the Standard Model's only source of quark-sector CP violation — which is why their extension predicted CP violation and a third generation before either was established. Cabibbo's angle survives as the dominant $1$ – $2$ mixing, with $V_{us}=\sin\theta_c\approx0.225$ still the largest off-diagonal element.

Does GIM suppression still matter now that all six quarks are known?

Enormously — it is why rare flavour-changing decays are precision probes of new physics. GIM guarantees neutral flavour-changing processes occur only through loops, suppressed by quark-mass differences over $M_W^2$ , so the Standard Model predicts their rates cleanly and small, and any excess is a sharp signal. Kaon experiments (NA62, KOTO) chase $K\to\pi\nu\bar\nu$ and LHCb studies $b\to s\ell^+\ell^-$ precisely because a new heavy particle in the loop would spoil the cancellation and lift the rate above the tiny prediction. GIM sets the baseline for today's searches beyond the Standard Model.