Isospin symmetry

Isospin is the oldest of the internal symmetries, and it began as a coincidence nobody could ignore. The proton and neutron weigh almost exactly the same — 938.272 MeV against 939.565 MeV, a difference of barely 0.14% — and the strong nuclear force between them ignores their electric charge entirely: $V_{pp}\approx V_{np}\approx V_{nn}$ . In 1932, the year the neutron was discovered, Heisenberg proposed the audacious reading of that coincidence: switch off electromagnetism and there is no way to tell a proton from a neutron, so the two are not different particles but two states of one entity, the nucleon — exactly as spin-up and spin-down are two states of one electron. He gave the nucleon a new quantum number, total isospin $I=\tfrac12$ with $I_3=\pm\tfrac12$ , and recast charge independence as invariance under rotations in an abstract "isospin space." The idea outlived its nuclear origins and became a statement about quarks. QCD couples to colour, not flavour, so the strong force treats up and down quarks identically; the only things that break the symmetry are the tiny up–down mass difference ( $m_d-m_u\approx2.5$ MeV against the QCD scale of a few hundred MeV) and electromagnetism. Writing $u=\binom10$ and $d=\binom01$ and demanding invariance under $2\times2$ unitary mixing yields the group $SU(2)$ , generated by three traceless Hermitian matrices $\hat T=\tfrac12\boldsymbol\tau$ whose algebra is identical to that of angular momentum. So the entire machinery of spin — multiplets, ladder operators, Clebsch–Gordan coefficients — transfers wholesale and turns a near-coincidence of masses into hard, parameter-free predictions for hadron spectra and scattering rates.

Key takeaways

The proton and neutron are nearly degenerate ( $\Delta m=1.29$ MeV, 0.14%) and the nuclear force is charge-independent, so they behave as two states of one nucleon.
Heisenberg's isospin treats $p,n$ (equivalently the $u,d$ quarks) as an $I=\tfrac12$ doublet with $I_3=\pm\tfrac12$ .
Mathematically it is an internal $SU(2)$ symmetry, $\hat U=e^{i\boldsymbol\alpha\cdot\hat T}$ with generators $\hat T=\tfrac12\boldsymbol\tau$ obeying $[\hat T_i,\hat T_j]=i\epsilon_{ijk}\hat T_k$ — the angular-momentum algebra.
Hadrons fall into multiplets of $2I+1$ near-degenerate states: nucleon doublet, pion triplet, $\Delta$ quadruplet.
It is only approximate: $m_u\neq m_d$ and electromagnetism split the multiplets by $\lesssim1\%$ , and the Gell-Mann–Nishijima relation $Q=I_3+\tfrac12(B+S)$ ties $I_3$ to charge.
The sharpest test: at the $\Delta(1232)$ resonance isospin forces $\sigma(\pi^+p):\sigma(\pi^-p)=3:1$ , confirmed to a few percent.
Key idea: the strong force barely distinguishes up from down (proton from neutron), so those flavours can be rotated into one another like the components of a spin — an $SU(2)$ symmetry mathematically identical to angular momentum, whose multiplets and Clebsch–Gordan coefficients predict hadron masses and reaction rates.

The core insight

The strong force barely cares about the difference between up and down (or proton and neutron), so these flavours can be rotated into one another like the components of a spin — and that rotational freedom is an $SU(2)$ symmetry mathematically identical to ordinary angular momentum, whose multiplet and Clebsch–Gordan machinery then fixes hadron spectra and scattering ratios.

Where this lives today

Isospin is far from a historical relic; it is a working tool across nuclear and particle physics. The most striking modern development is that the once-mysterious $1.29$ MeV neutron–proton mass splitting has been computed from first principles: lattice-QCD collaborations (notably the Budapest–Marseille–Wuppertal group, Science 2015) combined QCD with QED on the lattice and reproduced $m_n-m_p=1.51(28)$ MeV, disentangling the competing effects — the up–down quark-mass difference pushes the neutron up by $\sim+2.5$ MeV while electromagnetic self-energy pushes it down by $\sim-1$ MeV, and the mass difference wins. In nuclear structure, isospin organizes mirror nuclei (like $^{13}$ C and $^{13}$ N, related by $p\leftrightarrow n$ ) and isobaric analog states, whose near-degeneracy is a routine spectroscopic diagnostic. In hadron spectroscopy, amplitude analyses at LHCb, BESIII, and GlueX use isospin Clebsch–Gordan relations to connect channels and assign the isospin of new exotic states — pentaquarks and tetraquarks are classified partly by whether their charge multiplets fill an isospin representation. And in precision physics, isospin-breaking corrections are a leading systematic in extractions of $V_{ud}$ from superallowed beta decays.

The physical picture

Strip away electric charge and the proton and neutron become interchangeable: the strong interaction binds them, scatters them, and builds nuclei out of them without ever consulting which is which. That indifference is a symmetry, and its natural language is borrowed wholesale from spin — just as an electron's two spin states sit at the same energy until a magnetic field splits them, the nucleon's two "isospin states" sit at (almost) the same mass until electromagnetism splits them. The first figure sets the two pictures side by side — the same abstract $SU(2)$ doublet, once in real space, once in internal flavour space — and pins the numbers to the nucleon: $m_p=938.27$ MeV, $m_n=939.57$ MeV, a splitting of only $0.14\%$ .

Figure

Isospin borrows the whole structure of spin ½. Left: an electron's up and down states are degenerate until a field splits them. Right: the nucleon is an isospin doublet — proton uud at I₃ = +½ and neutron udd at I₃ = −½ — degenerate to 0.14 percent, mp = 938.27 MeV and mn = 939.57 MeV, split only by the u,d mass difference and electromagnetism

The symmetry is not confined to two particles: it organizes the entire low-lying hadron spectrum into families. Quark–antiquark pairs build the pion triplet ( $\pi^+,\pi^0,\pi^-$ at $I_3=+1,0,-1$ ); three quarks build the $\Delta$ quadruplet ( $\Delta^{++},\Delta^+,\Delta^0,\Delta^-$ ). Within each family the members are nearly degenerate, and — more powerfully — their interactions are locked together by Clebsch–Gordan coefficients. The most incisive consequence is in pion–nucleon scattering through the $\Delta(1232)$ resonance, where isospin conservation forces the total cross sections into the fixed ratio $\sigma(\pi^+p):\sigma(\pi^-p)=3:1$ . The second figure is that prediction against the data: two resonance bumps whose peak heights stand in a ratio of nearly three.

Figure

The sharpest test of isospin. Pion-nucleon total cross sections through the Δ(1232) resonance at pion lab energy near 195 MeV. The Δ has I = 3/2, so isospin forces the peak ratio σ(π⁺p):σ(π⁻p) = 9:(2+1) = 3:1. Measured peaks about 195 mb and 70 mb give 2.8, confirming the parameter-free prediction of 3 to a few percent

The core equations

Isospin is defined by three generators that rotate the flavour doublet and close on the angular-momentum algebra. Everything downstream — multiplets, ladder operators, Clebsch–Gordan coefficients — follows from this one commutation relation, which is why isospin is often called "the spin that has nothing to do with spin."

\hat T=\tfrac12\boldsymbol\tau,\qquad [\hat T_i,\hat T_j]=i\epsilon_{ijk}\hat T_k \qquad (1)

where: $\hat T_i$ — the three isospin generators ( $i=1,2,3$ ); $\boldsymbol\tau=(\tau_1,\tau_2,\tau_3)$ — the Pauli matrices acting in the abstract $u$ – $d$ (or $p$ – $n$ ) space, not in real space; $\epsilon_{ijk}$ — the Levi-Civita symbol. This is exactly the $\mathfrak{su}(2)$ algebra of orbital and spin angular momentum, so every result about $\hat J$ carries over verbatim to $\hat T$ . A finite isospin transformation exponentiates the generators; demanding that the strong Hamiltonian be invariant under all of them is the precise statement of the symmetry.

\hat U=e^{i\boldsymbol\alpha\cdot\hat T},\qquad \binom{u}{d}\to\hat U\binom{u}{d},\qquad [\hat H_{\text{strong}},\hat U]=0 \qquad (2)

where: $\boldsymbol\alpha$ — three real rotation angles in isospin space; $\binom{u}{d}$ — the fundamental $I=\tfrac12$ doublet (identically the nucleon doublet $\binom pn$ ). Invariance $[\hat H_{\text{strong}},\hat U]=0$ means the strong interaction cannot distinguish rotated flavour states — the mathematical content of charge independence. Because the algebra is that of angular momentum, states are labelled by a total isospin $I$ (from the Casimir $\hat T^2$ ) and a third component $I_3$ , and each multiplet holds $2I+1$ members.

\hat T^2\,|I,I_3\rangle=I(I+1)\,|I,I_3\rangle,\qquad \hat T_3\,|I,I_3\rangle=I_3\,|I,I_3\rangle \qquad (3)

where: $I$ — total isospin ( $\tfrac12$ nucleon, $1$ pion, $\tfrac32$ delta); $I_3\in\{-I,\dots,+I\}$ — the third component, one value per multiplet member. Exact isospin predicts all $2I+1$ members degenerate in mass and identical in their strong couplings. The ladder operators step between members of a multiplet — physically, they turn a $d$ into a $u$ (a neutron into a proton) inside the strong sector.

\hat T_\pm=\hat T_1\pm i\hat T_2,\qquad \hat T_\pm\,|I,I_3\rangle=\sqrt{I(I+1)-I_3(I_3\pm1)}\;|I,I_3\pm1\rangle \qquad (4)

where: $\hat T_+$ raises $I_3$ by one (e.g. $n\to p$ ), $\hat T_-$ lowers it. These are the operators whose matrix elements are the Clebsch–Gordan coefficients that lock scattering and decay rates together across a multiplet. Finally, the third component is tied to observable electric charge by the Gell-Mann–Nishijima relation, the bridge between the internal symmetry and the real world.

Q=I_3+\tfrac12 Y=I_3+\tfrac12(B+S) \qquad (5)

where: $Q$ — electric charge in units of $e$ ; $Y=B+S$ — hypercharge; $B$ — baryon number; $S$ — strangeness (extended by charm, bottom, top for heavy flavours). For a $u$ quark ( $I_3=+\tfrac12$ , $B=\tfrac13$ , $S=0$ ) this returns $Q=+\tfrac23$ ; for $d$ , $Q=-\tfrac13$ .

Derivation AFrom charge independence to an internal SU(2) symmetry

The goal is to turn the two empirical facts — that the proton and neutron are nearly degenerate and that the nuclear force is charge-independent, $V_{pp}\approx V_{np}\approx V_{nn}$ — into a precise group-theoretic statement. We package the two particles into a single two-component object, ask what transformations leave the strong Hamiltonian unchanged, and show that the largest such group factorizes into a $U(1)$ of baryon-number phase and an $SU(2)$ whose three generators are the Pauli matrices obeying the angular-momentum algebra. That $SU(2)$ is isospin.

Step 1. Package the proton and neutron into a doublet $N=\binom pn$ . Charge independence plus mass degeneracy means the strong Hamiltonian carries no label distinguishing the two components: in the internal space it is proportional to the identity, $\hat H_{\text{strong}}=m\,\mathbb{1}+\dots$ , with the same $m$ for $p$ and $n$ . Why: "the force does not care which is which" is the plain-language version of "the Hamiltonian is the unit matrix in the internal space," which is the algebraic seed of the whole symmetry.

Step 2. The most general linear transformation that preserves the total probability is $N\to\hat U N$ with $\hat U^\dagger\hat U=\mathbb{1}$ , i.e. $\hat U\in U(2)$ . Since $\hat H_{\text{strong}}\propto\mathbb{1}$ in this space, it commutes with every such $\hat U$ : $[\hat H_{\text{strong}},\hat U]=0$ . Why: a unit matrix commutes with everything, so exact charge independence is invariant under the entire unitary group of the doublet — that group is the symmetry.

Step 3. Factor the group as $U(2)=U(1)\times SU(2)$ : write $\hat U=e^{i\alpha}\,\hat V$ with $\det\hat V=1$ . The overall phase $e^{i\alpha}$ multiplies $p$ and $n$ alike and simply expresses baryon-number conservation; all the physics of mixing $p$ and $n$ lives in the determinant-one part $SU(2)$ . Why: the $U(1)$ phase rotates nothing between the components, so the flavour-mixing content — the interesting symmetry — is entirely the $SU(2)$ piece.

Step 4. An $SU(2)$ element is $\hat V=\exp(i\boldsymbol\alpha\cdot\hat T)$ with three Hermitian, traceless generators $\hat T=\tfrac12\boldsymbol\tau$ . Direct computation gives $[\tfrac12\tau_i,\tfrac12\tau_j]=i\epsilon_{ijk}\tfrac12\tau_k$ , i.e. $[\hat T_i,\hat T_j]=i\epsilon_{ijk}\hat T_k$ . Why: three generators with exactly the angular-momentum algebra means the Casimir $\hat T^2$ , the eigenvalue ladder $I(I+1)$ , the raising and lowering operators, and the $(2I+1)$ -plets all transfer verbatim — isospin is spin's algebraic twin, though it is an internal rather than a spatial rotation.

Go deeper · Derivation B · 1 minBuilding the isospin multiplets with ladder operators

Derivation BBuilding the isospin multiplets with ladder operators

The aim is to show that, because isospin obeys the angular-momentum algebra, hadrons must organize into multiplets of $2I+1$ near-degenerate states connected by the ladder operators $\hat T_\pm$ , and to construct the three families that matter most — the nucleon doublet, the pion triplet, and the $\Delta$ quadruplet — directly from the fundamental $u,d$ doublet using the same Clebsch–Gordan addition rules as for spin. The payoff is a prediction of degenerate families whose measured splittings then measure the quality of the symmetry.

Step 1. Take the fundamental doublet $|u\rangle=|\tfrac12,+\tfrac12\rangle$ , $|d\rangle=|\tfrac12,-\tfrac12\rangle$ , with $\hat T_3$ eigenvalues $\pm\tfrac12$ and $\hat T_+|d\rangle=|u\rangle$ , $\hat T_-|u\rangle=|d\rangle$ . These two states are the $I=\tfrac12$ doublet, realized physically as $(p,n)$ or $(u,d)$ . Why: the smallest nontrivial representation is the doublet, and it is the building block from which every larger multiplet is assembled.

Step 2. Combine a quark and an antiquark, $\tfrac12\otimes\tfrac12=1\oplus0$ . The symmetric $I=1$ triplet has $I_3=+1,0,-1$ — the pions $\pi^+,\pi^0,\pi^-$ — while the antisymmetric $I=0$ singlet is an isoscalar. Why: the Clebsch–Gordan series for combining two isospins is identical to adding two spins, so $q\bar q$ states must appear as an isotriplet plus an isosinglet, exactly the observed pion multiplet structure.

Step 3. Combine three quarks, $\tfrac12\otimes\tfrac12\otimes\tfrac12=\tfrac32\oplus\tfrac12\oplus\tfrac12$ . The fully symmetric $I=\tfrac32$ representation has $2I+1=4$ members: $\Delta^{++}(uuu)$ , $\Delta^+(uud)$ , $\Delta^0(udd)$ , $\Delta^-(ddd)$ at $I_3=+\tfrac32,+\tfrac12,-\tfrac12,-\tfrac32$ ; the $I=\tfrac12$ pieces are the nucleon-like states. Why: the top three-quark multiplet is the isospin quadruplet, and a symmetric bound state $(uuu)$ forces a whole degenerate family down to $(ddd)$ — the strong analog of " $uud$ implies a partner $ddu$ ."

Step 4. State the degeneracy prediction: with exact isospin every member of a multiplet has the same mass and the same strong couplings. The measured splittings — nucleon $0.14\%$ , the four $\Delta$ states within $\sim1\%$ , the pions $3.3\%$ — are small isospin-breaking corrections from $m_d\neq m_u$ and electromagnetism. Why: the concrete physical value of the symmetry is precisely this prediction of near-degenerate families, and the residual splitting is a direct, quantitative meter of how good the symmetry is.

Go deeper · Derivation C · 1 minClebsch–Gordan coefficients and the pion-nucleon cross-section ratios

Derivation CClebsch–Gordan coefficients and the pion-nucleon cross-section ratios

The objective is to show that isospin conservation reduces every pion–nucleon scattering channel to just two reduced amplitudes, $A_{3/2}$ and $A_{1/2}$ , and that at the $\Delta(1232)$ resonance — where only the $I=\tfrac32$ amplitude is large — the total cross sections are locked into the parameter-free ratio $\sigma(\pi^+p):\sigma(\pi^-p)=3:1$ . This converts the abstract symmetry into a hard number that a bubble-chamber experiment can check.

Step 1. The pion has $I=1$ and the nucleon $I=\tfrac12$ , so $\pi N$ states span $1\otimes\tfrac12=\tfrac32\oplus\tfrac12$ . Isospin conservation makes the strong $\hat T$ -matrix diagonal in $I$ , with just two numbers $A_{3/2}$ and $A_{1/2}$ . The channel $\pi^+p=|1,+1\rangle|\tfrac12,+\tfrac12\rangle=|\tfrac32,+\tfrac32\rangle$ is a pure $I=\tfrac32$ state. Why: conservation of isospin collapses the entire scattering problem to two reduced amplitudes, and the maximally-charged channel isolates $A_{3/2}$ with no contamination.

Step 2. Decompose the mixed channels at $I_3=-\tfrac12$ using the Clebsch–Gordan coefficients for $1\otimes\tfrac12$ : $\;\pi^-p=\sqrt{\tfrac13}\,|\tfrac32,-\tfrac12\rangle-\sqrt{\tfrac23}\,|\tfrac12,-\tfrac12\rangle\;$ and $\;\pi^0n=\sqrt{\tfrac23}\,|\tfrac32,-\tfrac12\rangle+\sqrt{\tfrac13}\,|\tfrac12,-\tfrac12\rangle$ . Why: the mixed-charge channels are superpositions of both isospins, so their amplitudes interfere with coefficients fixed purely by the algebra, not by dynamics.

Step 3. Project initial and final states onto the two amplitudes. Elastic $\pi^+p\to\pi^+p$ has amplitude $A_{3/2}$ ; elastic $\pi^-p\to\pi^-p$ has $\tfrac13 A_{3/2}+\tfrac23 A_{1/2}$ ; charge exchange $\pi^-p\to\pi^0n$ has $\tfrac{\sqrt2}{3}A_{3/2}-\tfrac{\sqrt2}{3}A_{1/2}$ . Why: sandwiching the diagonal $\hat T$ -matrix between the decomposed states gives every physical channel as a known linear combination of $A_{3/2}$ and $A_{1/2}$ .

Step 4. At the $\Delta(1232)$ the $I=\tfrac32$ amplitude resonates and dominates, $A_{1/2}\approx0$ . The three cross sections become $|A_{3/2}|^2:\tfrac29|A_{3/2}|^2:\tfrac19|A_{3/2}|^2=9:2:1$ , so summing the two $\pi^-p$ channels gives $\sigma_{\text{tot}}(\pi^+p):\sigma_{\text{tot}}(\pi^-p)=9:(2+1)=3:1$ . Why: a single resonant amplitude sets all three channels at once, turning the symmetry into a parameter-free prediction that is confirmed at the resonance peak.

Limiting cases

Case	Result	Meaning
Exact isospin ( $m_u=m_d$ , no EM)	multiplets perfectly degenerate	all $2I+1$ members equal in mass and strong coupling; $\sigma$ ratios exact
Real world	splittings $\lesssim1\%$	$n{-}p$ 0.14%, $\Delta$ states $<$ 1%, pions 3.3% — measures the breaking
Charged-current / EM probe	symmetry broken explicitly	electromagnetism sees $Q$ , so it distinguishes $I_3$ and lifts the degeneracy
Add strangeness ( $u,d,s$ )	$SU(2)\subset SU(3)$ flavour	octets and decuplets, but breaking grows to $\sim20$ – $30\%$ since $m_s\approx95$ MeV
Heavy quarks ( $c,b,t$ )	no isospin partner	$m_c,m_b,m_t\gg\Lambda_{\text{QCD}}$ , so there is no near-degenerate flavour to rotate into
High isospin $I$	$2I+1$ members	e.g. $\Delta$ has $I=\tfrac32\Rightarrow4$ charge states $\Delta^{++}\dots\Delta^-$

Common trapIsospin is a kind of spin, and any isospin rotation is physically allowed

Both halves are traps. First, isospin is not spin: it carries no angular momentum, points in no direction of real space, and couples to no magnetic field. It only shares the $SU(2)$ algebra — the name "spin" is a mathematical analogy, nothing more. A nucleon at rest with $I_3=+\tfrac12$ is a proton, full stop; it is not "spinning" in any physical space. Second, only the strong interaction is isospin-symmetric, so a general continuous rotation $e^{i\boldsymbol\alpha\cdot\hat T}$ that mixes $u$ and $d$ (or $p$ and $n$ ) is not an exact symmetry of nature — it changes electric charge, which electromagnetism and the vacuum most certainly notice. What is (approximately) good is the discrete relabelling $u\leftrightarrow d$ , $p\leftrightarrow n$ , and the multiplet structure it enforces on strong processes. Treating an arbitrary isospin-rotated superposition as a physical particle would violate charge conservation.

Worked examples

1 · Charges and multiplets from Gell-Mann–Nishijima (confidence builder)

Use $Q=I_3+\tfrac12(B+S)$ to (a) recover the up and down quark charges from their isospin, (b) confirm the pion triplet is $I=1$ , and (c) place $\Delta^{++}=uuu$ in its multiplet.

Step 1. Quarks have $B=\tfrac13$ , $S=0$ . For $u$ at $I_3=+\tfrac12$ : $Q_u=+\tfrac12+\tfrac12\cdot\tfrac13=+\tfrac12+\tfrac16=+\tfrac23$ . For $d$ at $I_3=-\tfrac12$ : $Q_d=-\tfrac12+\tfrac16=-\tfrac13$ . $Q_u=+\tfrac23,\ Q_d=-\tfrac13$ , the standard quark charges, recovered from isospin alone.

Step 2. Pions are $q\bar q$ with $B=0,S=0$ , so $Q=I_3$ directly. The triplet $I_3=+1,0,-1$ maps to $\pi^+,\pi^0,\pi^-$ with masses $139.57,134.98,139.57$ MeV — degenerate to a splitting of $4.6$ MeV, or $3.3\%$ . A charge- $(+1,0,-1)$ triplet is exactly the $2I+1=3$ members of $I=1$ .

Step 3. $\Delta^{++}=uuu$ has $B=1$ , $S=0$ , $Q=+2$ , so $Q=I_3+\tfrac12\Rightarrow I_3=+\tfrac32$ . That is the top of an $I=\tfrac32$ quadruplet, whose other members ( $\Delta^+,\Delta^0,\Delta^-$ ) must exist by the ladder — and do. $I_3(\Delta^{++})=+\tfrac32$ , confirming $I=\tfrac32$ with four charge states.

2 · The Δ(1232) cross-section ratio against data (fully worked)

At the $\Delta(1232)$ resonance the measured pion–nucleon total cross sections are $\sigma(\pi^+p)\approx195$ mb and $\sigma(\pi^-p)\approx70$ mb (elastic plus charge exchange). Compare with the isospin prediction and interpret the residual.

Step 1. From Derivation C, at the pure- $I=\tfrac32$ resonance the ratio is fixed with no free parameters: $\sigma(\pi^+p):\sigma(\pi^-p)=9:(2+1)=3:1$ , i.e. the prediction is exactly 3.

Step 2. The measured ratio is $\dfrac{195\ \text{mb}}{70\ \text{mb}}=2.79$ . This sits within about 7% of the predicted value 3 — striking agreement for a relation that uses only the algebra of $SU(2)$ and the dominance of one resonance.

Step 3. Split the $\pi^-p$ number: isospin predicts elastic : charge-exchange $=1:2$ , so of the $\approx70$ mb roughly $23$ mb is elastic $\pi^-p\to\pi^-p$ and $47$ mb is charge exchange $\pi^-p\to\pi^0n$ . The few-percent deviation from a clean $3:1$ is the small non-resonant $A_{1/2}$ background and genuine isospin breaking. Prediction 3, data 2.8 — isospin confirmed to the few-percent level.

Physics of the portrait

Isospin is a rare case of a symmetry proposed before its underlying substance was known: Heisenberg wrote it down in 1932 to describe forces between protons and neutrons, decades before quarks explained why the two are nearly identical. The portrait is his.

Origins · 1 minWerner Heisenberg and how this was discovered

Werner Heisenberg

1901–1976 · German physicist, architect of matrix mechanics and of isospin

Werner Heisenberg is remembered first for the uncertainty principle and matrix mechanics, the 1925 breakthrough that earned him the 1932 Nobel Prize and founded quantum theory. But 1932 was a doubly decisive year: Chadwick's discovery of the neutron in February handed Heisenberg a new puzzle, and within months he published a three-part paper on nuclear structure that introduced what he called isotopic spin. His reasoning was bold for the time — he treated the proton and neutron as two states of a single particle and wrote the nuclear force as a formal exchange interaction in an abstract two-dimensional space, deliberately modelled on the spinor space of an electron. The analogy was purely mathematical: isospin has nothing to do with rotation in real space, yet it obeys the identical $SU(2)$ algebra, so all the hard-won apparatus of angular momentum could be imported at once. Eugene Wigner sharpened the language and coined "isotopic spin" in 1937; Cassen and Condon built the charge-independent formalism; and two decades later Gell-Mann and Nishijima tied isospin to strangeness and charge, setting the stage for the quark model in which isospin became the up–down flavour symmetry of QCD. Heisenberg's leap — reading a coincidence of masses as a symmetry — is the template for every internal symmetry in modern particle physics.

History

1932

— Chadwick discovers the neutron, revealing the proton's near-twin

1932

— Heisenberg introduces isotopic spin to describe the charge-independent nuclear force

1937

— Wigner coins the term isotopic spin and develops the multiplet formalism

1953

— Gell-Mann and Nishijima relate isospin, strangeness, and charge

1964

— Gell-Mann and Zweig propose quarks, recasting isospin as up-down flavour symmetry

Going further

Texts. Thomson, Modern Particle Physics, §9.2 (Flavour symmetry) and §9.2.1–9.2.2 · Griffiths, Introduction to Elementary Particles, 2nd ed., Ch. 4 §4.1–4.5 (symmetries, groups, isospin, the eightfold way) · Halzen and Martin, Quarks and Leptons, Ch. 2 (symmetries and quarks) · Perkins, Introduction to High Energy Physics, Ch. 4 · Peskin and Schroeder, An Introduction to Quantum Field Theory, §20.1 (global symmetries) · Schwartz, Quantum Field Theory and the Standard Model, Ch. 28. The key results to keep in mind are: (1) $[\hat T_i,\hat T_j]=i\epsilon_{ijk}\hat T_k$ from $\hat T=\tfrac12\boldsymbol\tau$ ; (2) the multiplet decompositions $\tfrac12\otimes\tfrac12=1\oplus0$ and $\tfrac12^{\otimes3}=\tfrac32\oplus\tfrac12\oplus\tfrac12$ ; (3) $Q=I_3+\tfrac12(B+S)$ giving $Q_u=+\tfrac23$ , $Q_d=-\tfrac13$ , $Q_s=-\tfrac13$ ; (4) the Clebsch–Gordan ratio $9:2:1$ and $\sigma(\pi^+p)/\sigma(\pi^-p)=3$ ; (5) the mass splittings $m_n-m_p=1.29$ MeV (0.14%) and $m_{\pi^\pm}-m_{\pi^0}=4.6$ MeV (3.3%).

Prerequisites. Symmetries in Quantum Mechanics · Non-relativistic QM and Angular Momentum Review · The Particle Zoo. Leads to. Flavour Symmetry · $SU(3)$ Flavour · Combining Quarks into Baryons · Weak Isospin and Hypercharge.

Ask the tutor: Ask me to walk through the Clebsch–Gordan bookkeeping behind the $9:2:1$ ratio, or to show how isospin fixes any other pion–nucleon reaction rate you name.

Is isospin the same thing as spin?

No — they share only their mathematics. Isospin carries no angular momentum, defines no axis in real space, and does not couple to a magnetic field. What it shares with spin is the $SU(2)$ algebra $[\hat T_i,\hat T_j]=i\epsilon_{ijk}\hat T_k$ , and that is enough for the entire apparatus of angular momentum — total isospin $I$ , third component $I_3$ , ladder operators, $(2I+1)$ -member multiplets, and Clebsch–Gordan coefficients — to apply unchanged. So when you "add isospins" you use the same tables you use for spins, but a state with $I_3=+\tfrac12$ is simply a proton (or an up quark); nothing is physically rotating.

Why is it called isotopic spin if it has nothing to do with spin?

Historical analogy. When Heisenberg introduced it in 1932, the only two-state quantum system with a full formalism was the spin- $\tfrac12$ electron, with its up and down states and its Pauli matrices. The proton and neutron looked like another two-state system, so he borrowed the spinor language wholesale and called the new label "isotopic spin" (later shortened to "isospin," though strictly it is "isobaric spin," as $p$ and $n$ are isobars). The name records the parentage, not any physical spinning — the same algebra simply describes different physics.

Why is isospin only approximate?

Two effects break it. First, the up and down quarks are not truly degenerate: $m_d-m_u\approx2.5$ MeV. Second, they carry different electric charges, so electromagnetism treats them differently. Both effects are small compared with the QCD scale of a few hundred MeV that sets hadron masses, which is why isospin is a good approximate symmetry — multiplet splittings are typically well under a percent. But they are not zero, and they show up as the neutron–proton mass difference, the charged–neutral pion splitting, and the small deviations of scattering ratios from their exact isospin values. The size of the breaking is itself a useful measurement.

If the down quark is heavier than the up, why is the neutron heavier than the proton and not the other way round?

Two effects compete. The neutron ( $udd$ ) has an extra down quark relative to the proton ( $uud$ ), and since $m_d>m_u$ the quark-mass difference pushes the neutron up by roughly $+2.5$ MeV. Electromagnetism pulls the other way: the proton's larger self-energy would push it up by roughly $1$ MeV. The quark-mass effect wins, so the neutron is heavier by the observed $1.29$ MeV. This near-cancellation is profound — had electromagnetism won, the proton would be unstable and hydrogen could not exist. Lattice QCD+QED now reproduces the balance from first principles.

How does isospin relate to SU(3) flavour and the eightfold way?

Isospin is the $SU(2)$ acting on the two lightest quarks. Include the strange quark and the flavour symmetry enlarges to $SU(3)$ , whose multiplets are the octets and decuplets of the eightfold way — with isospin sitting inside as the $SU(2)$ subgroup that rotates $u$ and $d$ while leaving $s$ alone. Each $SU(3)$ multiplet decomposes into isospin sub-multiplets stacked by strangeness: the baryon octet holds the nucleon isodoublet, the $\Sigma$ isotriplet, the $\Lambda$ isosinglet, and the $\Xi$ isodoublet. Because the strange quark is much heavier ( $m_s\approx95$ MeV), $SU(3)$ is broken far worse than isospin — tens of percent rather than fractions of a percent — so isospin remains the sharper of the two symmetries.

What is the Gell-Mann–Nishijima relation actually good for?

The relation $Q=I_3+\tfrac12(B+S)$ is the dictionary between the internal isospin label $I_3$ and the electric charge you actually measure. Practically it does three jobs: it fixes the charges of the quarks from their isospin and hypercharge ( $Q_u=+\tfrac23$ , $Q_d=-\tfrac13$ ); it lets you assign $I_3$ (and hence place a particle in its multiplet) directly from its charge and its baryon number and strangeness; and it exposes that the "vertical" quantum number splitting multiplets by strangeness (hypercharge) is independent of the "horizontal" isospin one — the two-dimensional bookkeeping that becomes the weight diagram of $SU(3)$ . It is the hinge between the abstract symmetry and the particle data book.

Does isospin still matter now that we understand quarks and QCD?

Very much. QCD explains isospin — the strong force couples to colour and is blind to flavour — but it does not replace it as a working tool. Isospin gives exact, parameter-free relations between amplitudes that would otherwise demand a full non-perturbative calculation: cross-section ratios, decay branching ratios, and the classification of new states all follow from Clebsch–Gordan algebra alone. Modern amplitude analyses of exotic hadrons, CKM extractions from beta decay, and the hadronic corrections to the muon anomaly all lean on isospin relations and their controlled breaking. It remains one of the most economical predictive tools in hadron physics.