The Wu experiment

Hold a beta-decay experiment up to a mirror. Every law of physics anyone had written down before 1957 promised that the reflection would be a perfectly good experiment too — indistinguishable, in its statistics, from the original. That promise is the conservation of parity, and for the strong and electromagnetic interactions it holds exactly. In 1956 Lee and Yang noticed that no one had ever actually tested it for the weak interaction, the force behind beta decay. Chien-Shiung Wu built the test. She took nuclei of cobalt-60, lined up their spins by cooling a paramagnetic crystal to about $0.01\ \mathrm{K}$ , and watched which way the emitted electrons flew. The nuclear spin $\mathbf J$ is an axial vector: reflect the world and it is unchanged. The electron's momentum $\mathbf p$ is a polar vector: reflect the world and it reverses. Their product $\mathbf J\cdot\mathbf p$ therefore changes sign under reflection — it is a pseudoscalar, exactly the mirror-odd quantity a parity-symmetric world must average to zero. Wu found it was not zero: the electrons came out preferentially opposite to the nuclear spin, and the mirror-image experiment would have shown the opposite preference. Parity was violated, and not by a little — by the maximum the kinematics allow. That single spin-momentum correlation forced the weak interaction into its $V\!-\!A$ (vector-minus-axial) form, revealed that the neutrino is one-handed, and rewrote the Standard Model's treatment of the weak force. This article rebuilds the argument from the pseudoscalar up.

Key takeaways

Parity $\mathsf P$ inverts space ( $\mathbf x\to-\mathbf x$ ); it is exact in QED and QCD but the Wu experiment (1957) proved it is violated by the weak interaction.
The observable is a pseudoscalar, the spin-momentum correlation $\mathbf J\cdot\mathbf p$ : an axial vector times a polar vector, which flips sign under reflection — nonzero only if parity is broken.
Polarized $^{60}\mathrm{Co}$ nuclei ( $J^P=5^+$ ) beta-decay by a Gamow-Teller transition; the electrons are emitted preferentially antiparallel to the nuclear spin.
The angular distribution is $W(\theta)=1+A_\beta\,(v/c)\,P\cos\theta$ with beta-asymmetry $A_\beta=-1$ for pure Gamow-Teller — the coefficient a parity-conserving theory would force to zero.
The asymmetry traces to the weak vertex $\frac{-ig_w}{2\sqrt2}\gamma^\mu(1-\gamma^5)$ : the emitted electron is left-handed ( $\mathcal P_e=-v/c$ ), the antineutrino right-handed, and the violation is maximal.
Key idea: If you watch a process in a mirror and it behaves differently from reality, parity is violated — and a spin-momentum correlation (helicity) is exactly such a mirror-asymmetric quantity, which the weak V−A interaction produces while QED and QCD cannot.

Where this lives today

Parity violation stopped being a shock and became a tool. The maximal $V\!-\!A$ structure Wu's asymmetry pointed to is now a load-bearing wall of the electroweak Standard Model, and precision parity-violating electron scattering has grown into an industry: the Qweak experiment at Jefferson Lab measured the proton's weak charge $Q_W^p$ by the tiny left-right asymmetry in scattering polarized electrons, the MOLLER experiment aims to pin the weak mixing angle $\sin^2\theta_W$ at the $0.1\%$ level from electron-electron scattering, and the P2 experiment at Mainz pursues the same quantity with a complementary target. The PREX and CREX programs turned parity-violating scattering into a nuclear-structure probe, weighing the neutron skin of lead and calcium. In atomic physics, parity violation in cesium — a mirror-odd mixing of atomic states — tests the same electroweak couplings at near-zero momentum transfer. And the frontier still lives where Wu worked: modern cold-neutron beta-decay experiments (PERKEO, PERC, Nab) measure the neutron's beta-asymmetry parameter to a few parts in $10^4$ , hunting for right-handed currents or other departures from pure $V\!-\!A$ that would signal physics beyond the Standard Model. Every one of these is a descendant of a single number Wu measured as her crystal warmed.

The physical picture

The whole experiment is a machine for measuring one number: the average of $\mathbf J\cdot\hat{\mathbf p}$ , the projection of the electron's flight direction onto the nuclear spin. To make it visible you need two things — nuclei whose spins all point the same way, and a way to count electrons above versus below that axis. Wu polarized the spins by adiabatic demagnetization of a cerium-magnesium-nitrate crystal doped with $^{60}\mathrm{Co}$ , reaching about $0.01\ \mathrm{K}$ , and monitored the polarization through the anisotropy of the accompanying gamma rays. The electrons then reported the answer: they streamed out preferentially in the hemisphere opposite to the spin. The first figure sets the observed decay beside its own mirror image, which is where the violation becomes undeniable.

Figure

As observed, the electrons from spin-up cobalt-60 fly preferentially opposite to the nuclear spin J, a distribution W(θ) = 1 + A(v/c)cosθ with A = −1, β ≈ 0.6 at T ≈ 0.01 K. Reflecting the experiment leaves the axial spin J unchanged but reverses the polar momentum p, so the mirror world predicts electrons preferentially along J — the opposite result, so the two worlds are distinguishable and parity is violated

Why should the electrons care which way the nucleus points? The answer is angular-momentum bookkeeping run through the weak interaction's peculiar handedness. The decay $^{60}\mathrm{Co}\,(5^+)\to{}^{60}\mathrm{Ni}^*\,(4^+)+e^-+\bar\nu_e$ drops the nuclear spin by one unit, so the electron and antineutrino must carry off one unit of spin between them, aligned — a Gamow-Teller transition. Their spins therefore both point along the original nuclear spin. Now the weak interaction insists the electron be left-handed and the antineutrino right-handed: a left-handed particle has its momentum opposite to its spin. So the electron's momentum points against $\mathbf J$ , and the antineutrino's points along it. The second figure traces this chain from the spin ledger to the flight directions.

Figure

Angular-momentum bookkeeping in the Gamow-Teller decay ⁶⁰Co(5⁺) → ⁶⁰Ni(4⁺) + e⁻ + ν̄: the lost unit of nuclear spin is carried by the electron and antineutrino spins both aligned along J. The weak V−A interaction makes the electron left-handed (helicity −v/c ≈ −0.6) so its momentum points opposite J, and the antineutrino right-handed (helicity +1) so its momentum points along J; the beta endpoint is 0.31 MeV

The core equations

Parity begins as a statement about coordinates and ends as a statement about an observable. Spatial inversion sends a polar vector to its negative and leaves an axial vector alone; the electron momentum is polar, the nuclear spin axial. Their dot product is the pseudoscalar the whole experiment is built to catch.

\mathsf P:\ \mathbf p\to-\mathbf p,\quad \mathbf J\to\mathbf J,\qquad \langle\,\mathbf J\cdot\mathbf p\,\rangle\ \xrightarrow{\ \mathsf P\ }\ -\langle\,\mathbf J\cdot\mathbf p\,\rangle \qquad (1)

where: $\mathbf p$ — the electron momentum, a polar (ordinary) vector; $\mathbf J$ — the nuclear spin, an axial (pseudo) vector; $\langle\,\mathbf J\cdot\mathbf p\,\rangle$ — the ensemble-averaged spin-momentum correlation. Because the two vectors transform oppositely, their product is a pseudoscalar: a parity-invariant world must give it zero, so any nonzero value is a direct signal of parity violation. The correlation is read out as an angular distribution. The rate of electrons at angle $\theta$ to the nuclear spin carries a term linear in $\cos\theta$ — the pseudoscalar in disguise — weighted by the electron speed and the degree of nuclear polarization.

W(\theta)=1+A_\beta\,\frac{v}{c}\,P\,\cos\theta \qquad (2)

where: $\theta$ — angle between the nuclear spin $\mathbf J$ and the electron momentum; $A_\beta$ — the beta-asymmetry parameter fixed by the decay type ( $A_\beta=-1$ for pure Gamow-Teller); $v/c$ — the electron speed; $P=\langle J_z\rangle/J$ — the nuclear polarization. A parity-conserving interaction forces $A_\beta=0$ , killing the $\cos\theta$ term; the measured nonzero coefficient is the violation. The asymmetry parameter is not free — it is dictated by the Lorentz structure of the weak charged current. The leptonic vertex mixes a vector ( $\gamma^\mu$ ) with an axial vector ( $\gamma^\mu\gamma^5$ ) in equal measure, the $V\!-\!A$ form, which is what a pseudoscalar observable requires.

\mathcal M \propto \big[\,\bar u_e\,\gamma^\mu\tfrac12(1-\gamma^5)\,v_{\bar\nu}\,\big]\,J^{\rm nucl}_\mu,\qquad \text{vertex}=\frac{-ig_w}{2\sqrt2}\,\gamma^\mu(1-\gamma^5) \qquad (3)

where: $\bar u_e,\,v_{\bar\nu}$ — the electron and antineutrino spinors; $\tfrac12(1-\gamma^5)$ — the left-chiral projector; $J^{\rm nucl}_\mu$ — the nuclear (Gamow-Teller) current; $g_w$ — the weak coupling. Mixing $V$ with $A$ in the ratio $1:-1$ is what makes parity maximally violated; a pure $V$ (like QED) or pure $A$ theory would conserve it. The projector in equation (3) forces the emitted leptons into definite handedness. In the relativistic limit chirality equals helicity, so the electron emerges longitudinally polarized with its spin lagging its momentum by exactly $v/c$ .

\mathcal P_e=\frac{N_\uparrow-N_\downarrow}{N_\uparrow+N_\downarrow}=-\frac{v}{c},\qquad \tfrac12(1-\gamma^5)u \ \xrightarrow{\,E\gg m\,}\ \text{left-handed} \qquad (4)

where: $\mathcal P_e$ — the electron's longitudinal polarization (helicity expectation); $N_\uparrow,N_\downarrow$ — counts with spin along or against the momentum; the sign is negative because $\beta^-$ electrons are left-handed. The antineutrino carries the opposite, maximal value $+1$ : it is purely right-handed. None of this is observable unless the nuclei are aligned, and that is a thermodynamic battle. The polarization follows the Boltzmann competition between the nuclear Zeeman energy $\mu B_{\rm eff}$ and the thermal energy $kT$ , small until the sample is made very cold.

P=\frac{\langle J_z\rangle}{J},\qquad x=\frac{\mu B_{\rm eff}}{kT},\qquad P\simeq\frac{J+1}{3J}\,x\ \ (x\ll1),\qquad P\to1\ \ (x\gg1) \qquad (5)

where: $\mu$ — the $^{60}\mathrm{Co}$ magnetic moment ( $\approx3.8\,\mu_N$ ); $B_{\rm eff}$ — the effective (hyperfine) field at the nucleus; $k$ — Boltzmann's constant; $x$ — the ratio of magnetic to thermal energy. Order-unity polarization needs $x\gtrsim1$ , i.e. $\mu B_{\rm eff}\sim kT$ ; at ordinary fields this demands $T\sim0.01\ \mathrm{K}$ , which is why the experiment is a cryogenics tour de force.

Derivation AWhy a spin-momentum correlation cannot survive parity conservation

The goal is to show, with no reference to any particular interaction, that the mere appearance of a $\cos\theta$ term in the electron angular distribution is logically equivalent to parity violation — that the pseudoscalar $\mathbf J\cdot\mathbf p$ is forced to zero in any theory whose Hamiltonian commutes with the parity operator, so a measured asymmetry is a theorem-level proof, not a model-dependent inference. We build the observable, act on it with parity, and let the invariance of the dynamics do the rest.

Step 1. Form the observable. The decay rate into solid angle about $\hat{\mathbf p}$ for a nucleus polarized along $\hat{\mathbf J}$ can only depend on the one scalar you can build from the two available directions, $\hat{\mathbf J}\cdot\hat{\mathbf p}=\cos\theta$ . Expand: $W(\theta)=a_0+a_1\cos\theta+a_2\cos^2\theta+\cdots$ . Why: rotational invariance about the spin axis leaves $\cos\theta$ as the only variable, so the entire physics is carried by the coefficients $a_n$ .

Step 2. Classify each term under parity. Under $\mathsf P$ , momentum reverses ( $\hat{\mathbf p}\to-\hat{\mathbf p}$ ) while the axial spin is untouched ( $\hat{\mathbf J}\to\hat{\mathbf J}$ ), so $\cos\theta\to-\cos\theta$ . The even powers $a_0,a_2,\dots$ are parity-even; the odd term $a_1\cos\theta$ is parity-odd — it is the pseudoscalar. Why: only the odd-power terms distinguish the process from its mirror image, so they are the sole carriers of any parity signal.

Step 3. Impose parity invariance. If the interaction conserves parity, $[\hat H,\hat P]=0$ , then the rate must be invariant under $\mathsf P$ : $W(\theta)=W(\pi-\theta)$ . But $\cos(\pi-\theta)=-\cos\theta$ , so invariance demands $a_1\cos\theta=-a_1\cos\theta$ for all $\theta$ , forcing $a_1=0$ . Equivalently, $\langle\mathbf J\cdot\mathbf p\rangle=-\langle\mathbf J\cdot\mathbf p\rangle=0$ . Why: a parity-symmetric world cannot tell $\theta$ from $\pi-\theta$ , so the up-down asymmetry it predicts is identically zero — there is no room for a preferred hemisphere.

Step 4. Read the contrapositive. Wu measured $a_1\neq0$ : the electrons picked a hemisphere. By Step 3 the only escape is $[\hat H,\hat P]\neq0$ — the weak Hamiltonian does not commute with parity. No detail of nuclear structure or of the weak coupling was needed; the existence of the correlation alone settles it. Why: this is what makes the experiment decisive rather than merely suggestive — a single nonzero pseudoscalar rules out every parity-conserving theory at once.

Go deeper · Derivation B · 2 minThe electron asymmetry from V minus A and the emitted lepton helicities

Derivation BThe electron asymmetry from V minus A and the emitted lepton helicities

The aim is to trace the sign and size of $A_\beta$ from the $V\!-\!A$ vertex to the flight direction of the electron, showing that the left-chiral projector plus Gamow-Teller angular-momentum bookkeeping forces the electron opposite the nuclear spin with the maximal coefficient $A_\beta=-1$ . We project the leptons into chirality eigenstates, convert chirality to helicity in the relativistic limit, and then balance angular momentum in the transition.

Step 1. Project the leptons. The current $\bar u_e\gamma^\mu\tfrac12(1-\gamma^5)v_{\bar\nu}$ couples only the left-chiral electron field and (through the charge-conjugate structure) the right-chiral antineutrino. For $E\gg m$ chirality equals helicity, so the electron is emitted with helicity $\mathcal P_e=-v/c$ and the antineutrino with helicity $+1$ . Why: the single projector $\tfrac12(1-\gamma^5)$ is the whole source of handedness — it hard-wires one chirality into every weak vertex.

Step 2. Set up the Gamow-Teller spin ledger. The transition $^{60}\mathrm{Co}\,(J=5)\to{}^{60}\mathrm{Ni}\,(J=4)$ lowers the nuclear spin by one unit with no orbital change, so the electron-antineutrino pair must carry total spin $S=1$ with its projection along the original spin axis, i.e. both lepton spins point along $\hat{\mathbf J}$ . Why: Gamow-Teller means the leptons leave in a spin-triplet, and conserving $J_z$ pins that triplet's spin along the nuclear polarization.

Step 3. Convert spin to momentum. A left-handed electron has its spin antiparallel to its momentum; with the spin fixed along $\hat{\mathbf J}$ by Step 2, its momentum must point against $\hat{\mathbf J}$ . The right-handed antineutrino, spin also along $\hat{\mathbf J}$ , flies along it. Why: helicity is the bridge — once the spin direction is set by angular momentum and the handedness by $V\!-\!A$ , the momentum direction is no longer free.

Step 4. Assemble the distribution and the coefficient. Weighting the amplitude over electron directions gives $W(\theta)=1+A_\beta(v/c)\cos\theta$ with $A_\beta=-1$ for this pure Gamow-Teller $J\to J-1$ decay; the factor $v/c$ is precisely the electron's own longitudinal polarization from Step 1. In the massless limit $v/c\to1$ the asymmetry saturates — parity violation is maximal, not perturbative. Why: the same $-v/c$ that measures the electron's handedness reappears as the strength of the up-down asymmetry, tying the two halves of the physics into one number.

Go deeper · Derivation C · 2 minPolarizing cobalt-60 by adiabatic demagnetization

Derivation CPolarizing cobalt-60 by adiabatic demagnetization

The purpose here is quantitative: to show why the nuclei must be cooled to the millikelvin scale before any asymmetry can appear, by writing the polarization as a Boltzmann average over the nuclear Zeeman ladder and estimating the temperature at which it becomes of order unity. We set up the level populations, extract the low-field polarization, and plug in the numbers for $^{60}\mathrm{Co}$ .

Step 1. Write the Zeeman ladder. A nucleus of spin $J$ and moment $\mu$ in an effective field $B_{\rm eff}$ has $2J+1$ equally spaced levels of energy $E_m=-\mu B_{\rm eff}\,m/J$ for $m=-J,\dots,J$ . The states are populated by the Boltzmann factor $e^{-E_m/kT}$ . Why: the only thing that aligns the spins is that the lowest-energy orientation ( $m=+J$ , spin along the field) is thermally favoured, and that advantage is set by $E/kT$ .

Step 2. Form the polarization. $P=\langle J_z\rangle/J=\frac{1}{J}\sum_m m\,e^{-E_m/kT}\big/\sum_m e^{-E_m/kT}$ , which is the Brillouin function of $x=\mu B_{\rm eff}/kT$ . For $x\ll1$ it linearizes to $P\simeq\frac{(J+1)}{3J}x$ ; for $x\gg1$ it saturates at $P\to1$ . Why: the competition is entirely captured by the single dimensionless ratio $x$ — magnetic energy over thermal energy — so estimating $x$ estimates the whole experiment's feasibility.

Step 3. Put in the numbers. With $\mu\approx3.8\,\mu_N=1.2\times10^{-7}\ \mathrm{eV/T}$ and thermal energy $kT=8.6\times10^{-7}\ \mathrm{eV}$ at $T=0.01\ \mathrm{K}$ , reaching $x\sim1$ needs $B_{\rm eff}\sim kT/\mu\approx7\ \mathrm{T}$ ; the hyperfine field at the cobalt nucleus in the polarized crystal is of order ten tesla, so $x$ is order unity to a few and $P$ climbs toward $\sim0.6$ . Why: this is the crux of the experimental difficulty — at $T=1\ \mathrm{K}$ one would need $x\sim1$ to survive a hundred-fold hotter bath, requiring $\sim700\ \mathrm{T}$ , which is why demagnetization to $\sim0.01\ \mathrm{K}$ is not a luxury but a necessity.

Step 4. Note the clock. There is no cooling power at these temperatures, so once demagnetized the crystal slowly warms by heat leak; as $T$ rises, $x$ falls and $P\to0$ , and with it the asymmetry. Wu saw the effect fade over about six minutes — the same signal appearing and then vanishing as the polarization relaxed, the cleanest possible control that the asymmetry tracked the spins and nothing else. Why: the warm-up doubles as a null test — a spurious instrumental asymmetry would not politely decay on the thermal-relaxation timescale of the nuclear spins.

Limiting cases

Case	Result	Meaning
Relativistic electron $v/c\to1$	$A_\beta(v/c)\to-1$	asymmetry saturates; electron fully left-handed, parity violation maximal
Non-relativistic electron $v/c\to0$	$\cos\theta$ term $\to0$	slow electrons carry little helicity, so the asymmetry is speed-suppressed
Unpolarized nuclei $P\to0$	$W(\theta)\to1$	isotropic emission — parity is still violated, but with no spin axis the pseudoscalar is unobservable
Parity-conserving limit (pure $V$ or pure $A$ )	$A_\beta=0$	no $\cos\theta$ term; QED- or QCD-like couplings give a mirror-symmetric decay
Pure Fermi ( $\Delta J=0$ ) vs Gamow-Teller ( $\Delta J=1$ )	GT drives the $^{60}$ Co asymmetry	the spin-flip channel is what correlates lepton spins with $\mathbf J$
Warm-up $T\uparrow$	$x=\mu B_{\rm eff}/kT\downarrow$ , asymmetry $\to0$	the signal decays over $\sim6$ min as the spins relax — a built-in null test

Common trapThe aligning magnetic field is what breaks the left-right symmetry, so of course there is an asymmetry

It does not, and this is the subtlety that makes the experiment airtight rather than circular. A magnetic field is an axial vector: under spatial inversion it does not reverse, exactly like the nuclear spin it aligns. So a parity-conserving theory, handed both the field and the resulting polarization, still predicts equal electron emission along and against $\mathbf J$ — the field defines an axis but not a sense, and there is no parity-even way to convert an axial axis into a preferred polar direction. The only object in the problem that reverses under reflection is the electron momentum, a polar vector; correlating a polar vector with an axial one ( $\mathbf J\cdot\mathbf p$ ) is the single pseudoscalar available, and its nonvanishing cannot be manufactured by any parity-symmetric apparatus, however it is wired. That is why the six-minute warm-up matters: the asymmetry appeared and faded with the nuclear polarization, not with the applied field or the detector geometry, which never changed. The field aligns the spins; the weak interaction, and only the weak interaction, tells the electrons which way to go.

Worked examples

1 · The temperature needed to polarize the nuclei (fully worked)

Cobalt-60 has spin $J=5$ and magnetic moment $\mu\approx3.8\,\mu_N$ . Estimate the temperature at which the nuclear polarization becomes of order $60\%$ in an effective field $B_{\rm eff}\approx15\ \mathrm{T}$ , and contrast it with what $T=1\ \mathrm{K}$ would give.

Step 1. Compute the magnetic energy scale. With $\mu_N=3.15\times10^{-8}\ \mathrm{eV/T}$ , the moment is $\mu=3.8\times3.15\times10^{-8}=1.2\times10^{-7}\ \mathrm{eV/T}$ , so $\mu B_{\rm eff}=1.2\times10^{-7}\times15\approx1.8\times10^{-6}\ \mathrm{eV}$ . The Zeeman energy is $\mu B_{\rm eff}\approx1.8\ \mu\mathrm{eV}$ .

Step 2. Set it against the thermal energy at $0.01\ \mathrm{K}$ . Here $kT=(8.62\times10^{-5}\ \mathrm{eV/K})(0.01\ \mathrm{K})=8.6\times10^{-7}\ \mathrm{eV}$ , so the ratio is $x=\mu B_{\rm eff}/kT\approx1.8\times10^{-6}/8.6\times10^{-7}\approx2.1$ . $x\approx2$ — magnetic energy comfortably exceeds thermal energy.

Step 3. Convert to polarization. At $x\approx2$ the Brillouin function for $J=5$ is well past its linear regime and gives $P\sim0.6$ ; the low-field formula $P\simeq\frac{J+1}{3J}x=\frac{6}{15}(2.1)\approx0.84$ overestimates because saturation has set in, but both confirm order- $60\%$ alignment is reached near $T\approx0.01\ \mathrm{K}$ .

Step 4. Show why $1\ \mathrm{K}$ fails. At $T=1\ \mathrm{K}$ , $kT=8.6\times10^{-5}\ \mathrm{eV}$ , so $x=1.8\times10^{-6}/8.6\times10^{-5}\approx0.02$ and $P\simeq\frac{6}{15}(0.02)\approx0.008$ — under $1\%$ . To recover $x\sim2$ at $1\ \mathrm{K}$ you would need $B_{\rm eff}\sim1500\ \mathrm{T}$ , far beyond any laboratory field. A hundred-fold colder bath is the whole reason the experiment demands adiabatic demagnetization; at $1\ \mathrm{K}$ the electrons would emerge essentially isotropic.

2 · Reading parity violation off the asymmetry (fully numerical)

For the Gamow-Teller decay take $A_\beta=-1$ and an electron speed $v/c=0.6$ , with polarization $P=1$ . Compute the forward/backward count ratio and the up-down asymmetry, and state what a parity-conserving theory would predict.

Step 1. Write the distribution. $W(\theta)=1+A_\beta(v/c)P\cos\theta=1-0.6\cos\theta$ , with $\theta$ measured from the nuclear spin. Evaluate along the spin ( $\theta=0$ ) and against it ( $\theta=180^\circ$ ): $W(0^\circ)=1-0.6=0.4$ and $W(180^\circ)=1+0.6=1.6$ . Electrons are four times as likely to fly against the spin as along it.

Step 2. Form the up-down asymmetry. $\mathcal A=\dfrac{W(180^\circ)-W(0^\circ)}{W(180^\circ)+W(0^\circ)}=\dfrac{1.6-0.4}{2.0}=0.60$ , which is just $|A_\beta|(v/c)$ . The measured asymmetry equals $A_\beta\,v/c=0.6$ in magnitude, and its sign says the electrons prefer the anti-spin hemisphere.

Step 3. Compute the mean correlation. $\langle\cos\theta\rangle=\dfrac{\int(1-0.6\cos\theta)\cos\theta\,d\Omega}{\int(1-0.6\cos\theta)\,d\Omega}=-\dfrac{0.6}{3}=-0.2$ , so $\langle\mathbf J\cdot\hat{\mathbf p}\rangle=-0.2\neq0$ . The pseudoscalar is nonzero and negative — the smoking gun.

Step 4. State the parity-conserving prediction. Such a theory forces $A_\beta=0$ , giving $W(\theta)=1$ , $\mathcal A=0$ , and $\langle\cos\theta\rangle=0$ : perfect up-down symmetry. The observed $\mathcal A\approx0.6$ (before dilution by partial polarization and finite geometry) is impossible in that world. A single nonzero asymmetry, decaying as the sample warms, refutes every parity-symmetric weak interaction.

Physics of the portrait

The Wu experiment is remembered for a result, but it was won by an experimentalist of exceptional care, working at the edge of what cryogenics could then do, on a question most of her colleagues thought was not worth asking.

Origins · 2 minChien-Shiung Wu and how this was discovered

Chien-Shiung Wu

1912–1997 · the experimental discovery of parity violation in the weak interaction

Chien-Shiung Wu was the foremost experimental nuclear physicist of her generation, and the beta-decay expert Tsung-Dao Lee turned to when he and Chen-Ning Yang realized, in 1956, that parity conservation had simply never been tested in the weak interaction. Born near Shanghai and trained at Berkeley under Ernest Lawrence and Emilio Segrè, Wu had already established herself as the person whose beta-spectrum measurements could be trusted when others could not — she had shown, against the textbooks, that Fermi's theory of beta decay was correct once the electron scattering was handled properly. When Lee described the parity puzzle to her, she grasped the stakes at once, postponed a planned trip, and designed the cobalt-60 measurement. It was brutally hard: it required polarizing nuclei by adiabatic demagnetization of a cerium-magnesium-nitrate crystal to about $0.01\ \mathrm{K}$ , keeping thin radioactive sources cold and aligned, and reading the electron asymmetry before the crystal warmed — a collaboration with Ernest Ambler's low-temperature group at the National Bureau of Standards. Over the winter of 1956–57 the effect emerged unmistakably: the electrons came out preferentially opposite to the nuclear spin, and the asymmetry decayed exactly as the polarization relaxed. The result, published in Physical Review in February 1957, overturned a symmetry physicists had assumed was as sacred as energy conservation. Lee and Yang received the Nobel Prize that same year for the theory; Wu, whose experiment supplied the proof, was passed over — one of the most criticized omissions in the prize's history. She continued to do decisive weak-interaction experiments for decades, confirmed the conserved-vector-current hypothesis, and became the first woman to preside over the American Physical Society. "It is shameful," she once said, "that there are so few women in science."

History

1956

— the θ–τ puzzle: two particles alike in mass and lifetime decay to opposite-parity final states

1956

— Lee and Yang show parity had never been tested in the weak interaction and propose the cobalt-60 test

1957

— Wu, Ambler, Hayward, Hoppes and Hudson observe the cobalt-60 asymmetry: parity is violated

1957

— Garwin, Lederman and Weinrich, and independently Friedman and Telegdi, confirm it in pion and muon decay

1957

— Lee and Yang receive the Nobel Prize in Physics

1958

— Goldhaber, Grodzins and Sunyar measure the neutrino helicity; Feynman–Gell-Mann and Sudarshan–Marshak formulate V−A

Going further

Texts. Griffiths, Introduction to Elementary Particles, 2nd ed., §4.6 and §9.1 (parity, the θ–τ puzzle, the V−A vertex) · Thomson, Modern Particle Physics, §11.1–11.3 (parity, its violation, and the weak charged current) · Halzen & Martin, Quarks and Leptons, §12.1–12.2 (the V−A interaction and parity violation) · Perkins, Introduction to High Energy Physics, 4th ed., §7 (weak interactions and parity) · Peskin & Schroeder, §20.2 (chiral structure of the weak interaction); Schwartz, Quantum Field Theory and the Standard Model, §29 (the electroweak sector). The key numerical checks in this article are: (1) $\cos\theta\to-\cos\theta$ under $\mathsf P$ forces $A_\beta=0$ in a parity-conserving theory; (2) $A_\beta=-1$ for a pure Gamow-Teller $J\to J-1$ decay; (3) the electron helicity $\mathcal P_e=-v/c$ ; (4) the polarization estimate $x=\mu B_{\rm eff}/kT\approx2$ at $T=0.01\ \mathrm{K}$ , $B_{\rm eff}\approx15\ \mathrm{T}$ , $\mu\approx3.8\,\mu_N$ , giving $P\sim0.6$ ; (5) the asymmetry $\mathcal A=|A_\beta|\,v/c=0.6$ at $v/c=0.6$ . The cobalt-60 quantum numbers ( $5^+\to4^+$ , endpoint $0.31\ \mathrm{MeV}$ ) and half-life ( $5.27\ \mathrm{yr}$ ) are standard tabulated values.

Prerequisites. Intrinsic Parity of Dirac Fermions · Dirac Adjoint and Bilinear Covariants · The Weak Charged Current. Leads to. The V−A Structure · Parity Violation · Helicity in Pion Decay.

Ask the tutor: Ask me to walk through why $\mathbf J\cdot\mathbf p$ is a pseudoscalar, or to trace the Gamow-Teller spin bookkeeping that sends the left-handed electron opposite the nuclear spin.

Why cobalt-60 specifically — what made it the right nucleus?

Several properties line up. Cobalt-60 has a large nuclear spin ( $J=5$ ) and a sizeable magnetic moment, so it can be strongly polarized in an achievable effective field at millikelvin temperatures. Its beta decay is a pure Gamow-Teller transition ( $5^+\to4^+$ , a one-unit spin change with no parity change), which produces the maximal electron asymmetry $A_\beta=-1$ — a mixed Fermi/Gamow-Teller decay would dilute the effect. And the decay is followed by two gamma rays whose angular distribution provided an independent, real-time monitor of how well the nuclei were polarized. A convenient $5.27$ -year half-life meant the source did not decay away during the run. Few nuclei combine high spin, pure Gamow-Teller character, a built-in polarization gauge, and a manageable lifetime.

How were the nuclei actually aligned, and why is that so hard?

By adiabatic demagnetization. A paramagnetic salt (cerium-magnesium-nitrate) doped with the cobalt-60 is magnetized at low temperature, then the field is slowly removed; the electronic spins disorder, drawing heat from the lattice and cooling the crystal to about $0.01\ \mathrm{K}$ . At that temperature the nuclear Zeeman energy $\mu B_{\rm eff}$ in the internal hyperfine field finally exceeds the thermal energy $kT$ , so the nuclei preferentially settle into the lowest-energy orientation and a net polarization builds up. The difficulty is that there is essentially no cooling power at these temperatures — the crystal warms from any heat leak within minutes, so the whole measurement is a race against relaxation. Wu's team saw the asymmetry appear and then fade over about six minutes, which doubled as proof that the signal tracked the spins.

Doesn't the applied magnetic field itself create a left-right asymmetry?

No, and this is the crucial subtlety. A magnetic field is an axial vector — under spatial reflection it is unchanged, just like the nuclear spin it aligns. So a parity-conserving theory, even knowing the field's direction, predicts equal numbers of electrons emitted along and against it: the field sets an axis but not a preferred sense. The only quantity that reverses under reflection is the electron momentum, a polar vector, and there is no parity-even way to correlate a polar direction with an axial one. The observed correlation $\mathbf J\cdot\mathbf p$ is a genuine pseudoscalar that no parity-symmetric apparatus could fake. The clinching control is that the asymmetry decayed with the nuclear polarization while the field and detectors were untouched.

Is parity violation the same thing as CP violation?

No — they are different and independent. Parity (P) violation, which Wu discovered, means the mirror-image process occurs at a different rate; the weak interaction violates it maximally. For a while physicists hoped that the combined operation CP — mirror reflection together with swapping particles for antiparticles — might still be exact, restoring a kind of symmetry. But in 1964 CP violation was found in neutral kaon decays, at a tiny level, and it is a distinct phenomenon governed by the complex phase of the CKM matrix rather than by the $V\!-\!A$ handedness. Maximal P violation is a statement about chirality; CP violation is a much subtler statement about the difference between matter and antimatter. The weak interaction does both, but for different reasons.

What does it mean that parity is violated "maximally"?

It means the weak interaction uses only one chirality, not a lopsided mixture of two. The vertex $\gamma^\mu(1-\gamma^5)$ contains the vector and axial-vector pieces with coefficients $+1$ and $-1$ — exactly equal in magnitude — so the left-chiral projector $\tfrac12(1-\gamma^5)$ appears with full weight and the right-chiral component is entirely absent. The consequence is that the emitted electron reaches its largest possible longitudinal polarization ( $-v/c$ , which tends to $-1$ as the electron becomes relativistic) and the antineutrino is purely right-handed. Had parity been only partially violated, the vector and axial coefficients would differ and the asymmetry would fall short of its kinematic maximum. Wu's asymmetry, extrapolated for polarization and speed, was consistent with the maximal value — which is why the weak interaction is written as $V\!-\!A$ and not, say, $V-0.8\,A$ .

How does the electron's helicity connect to the up-down asymmetry?

They are two faces of the same handedness. The $V\!-\!A$ interaction forces the electron to be left-handed, meaning its spin points opposite its momentum, with a definite polarization $\mathcal P_e=-v/c$ . In the Gamow-Teller decay of cobalt-60, angular-momentum conservation makes the lepton spins point along the nuclear spin. Combine the two: a left-handed electron whose spin is along $\mathbf J$ must have its momentum pointing against $\mathbf J$ — hence the electrons fly preferentially opposite to the spin. The same factor $v/c$ that sets the electron's helicity is exactly the coefficient of $\cos\theta$ in the angular distribution. So measuring the up-down asymmetry is, at root, measuring the electron's helicity indirectly — and later experiments (Frauenfelder, and Goldhaber's neutrino-helicity measurement) confirmed the helicities directly.

The Wu experiment

Where this lives today

The physical picture

The core equations

Limiting cases

Worked examples

1 · The temperature needed to polarize the nuclei (fully worked)

2 · Reading parity violation off the asymmetry (fully numerical)

Physics of the portrait

Going further

Before this

Where this leads