SU(3) flavour symmetry

The strong interaction is blind to flavour. The gluon couples to colour, not to whether a quark is an up, a down, or a strange — so if those three quarks had the same mass, the QCD Hamiltonian would be exactly invariant under any unitary rotation that mixes them. That invariance is a symmetry group, $SU(3)$ , and its consequences are geometric: hadrons are forced to arrange themselves into definite patterns — triangles, hexagons, a lone apex — plotted in the plane of two conserved labels, the third component of isospin $I_3$ and the hypercharge $Y$ . This is the framework that turned the chaotic "particle zoo" of the 1950s into a periodic table, and that let Gell-Mann point at an empty corner of a triangle and predict a particle, the $\Omega^-$ , that had never been seen. The symmetry is not exact. The strange quark is heavier than the up and down by roughly $100\ \mathrm{MeV}$ , so the classification is approximate — but the breaking is orderly, and its very orderliness is a second gift. Because the mass-splitting operator transforms as a definite piece of $SU(3)$ , the pattern of broken masses is itself predicted: the Gell-Mann–Okubo formula ties the nucleon, $\Lambda$ , $\Sigma$ , and $\Xi$ masses together to better than one percent, and the baryon decuplet splits into four equally spaced rungs whose spacing is the strange-quark mass. This article builds the group from the three quarks up, extracts the two good quantum numbers, decomposes quark and antiquark into the meson nonet $\mathbf 3\otimes\bar{\mathbf 3}=\mathbf 8\oplus\mathbf 1$ , and then cashes the whole apparatus out as numbers you can check against the Particle Data Group.

Key takeaways

The strong force treats $u$ , $d$ , $s$ as interchangeable, so QCD carries an approximate flavour symmetry with group $SU(3)$ , generated by the eight Gell-Mann matrices $\hat T_a=\tfrac12\lambda_a$ .
Only two generators, $\hat T_3$ and $\hat T_8$ , are simultaneously diagonal, so every hadron is labelled by two additive numbers: isospin $I_3$ and hypercharge $Y$ , tied to charge by Gell-Mann–Nishijima $Q=I_3+Y/2$ .
Quarks fill the fundamental $\mathbf 3$ and antiquarks the conjugate $\bar{\mathbf 3}$ ; combining them gives the meson nonet $\mathbf 3\otimes\bar{\mathbf 3}=\mathbf 8\oplus\mathbf 1$ , and three quarks give the baryons $\mathbf 3\otimes\mathbf 3\otimes\mathbf 3=\mathbf{10}\oplus\mathbf 8\oplus\mathbf 8\oplus\mathbf 1$ .
Symmetry breaking is orderly: the Gell-Mann–Okubo formula predicts $(m_N+m_\Xi)/2=(3m_\Lambda+m_\Sigma)/4$ , verified to $0.6\%$ , and the decuplet splits in equal $\sim147\ \mathrm{MeV}$ rungs.
The framework predicted the $\Omega^-$ ( $sss$ , $S=-3$ ) at $\approx1685\ \mathrm{MeV}$ before it was found at $1672\ \mathrm{MeV}$ — the definitive triumph of flavour $SU(3)$ .
Key idea: treating the three light quarks as interchangeable under an $SU(3)$ rotation forces hadrons into definite geometric patterns in the $(I_3,Y)$ plane — the octets, nonets, and decuplets seen in nature — and even the way the symmetry breaks is predicted.

The core insight

Treating the three light quarks as interchangeable under an $SU(3)$ rotation forces hadrons into definite geometric patterns in the $(I_3,Y)$ plane — the octets, singlets, and decuplets seen in nature — and because the breaking transforms as a fixed piece of the group, even the mass splittings are predicted.

Where this lives today

Flavour $SU(3)$ long ago stopped being a research frontier and became a working tool, quietly load-bearing across hadron physics. Lattice QCD now computes the octet and decuplet masses from first principles and reproduces the Gell-Mann–Okubo relations, with the small deviations tracing exactly to the strange-quark mass and confirming the first-order picture. In flavour physics, $SU(3)$ relations between the semileptonic decays of hyperons ( $\Lambda\to p e\nu$ , $\Sigma\to n e\nu$ , and their kin) let experimenters extract the CKM element $V_{us}$ , and the leading theoretical error there is the size of $SU(3)$ breaking — so the symmetry and its violation are both currency. The same group organises the spectrum of newly discovered exotic hadrons: LHCb's pentaquarks and tetraquarks are classified by asking which $SU(3)$ multiplet they occupy, and predicted partners are hunted seat by seat, exactly as the $\Omega^-$ once was. At a deeper level, the spontaneous breaking of the larger chiral $SU(3)_L\times SU(3)_R$ symmetry of massless QCD produces the pseudoscalar octet $(\pi,K,\eta)$ as pseudo-Goldstone bosons — the foundation of chiral perturbation theory, the effective field theory that governs low-energy strong interactions. Even the $\eta'$ 's stubborn refusal to be light, glimpsed in Example 3, became a landmark: its anomalously large mass is the fingerprint of QCD instantons and the axial- $U(1)$ anomaly. The triangles Gell-Mann drew on the $(I_3,Y)$ plane are still the first thing a hadron physicist sketches.

The physical picture

Take the three lightest quarks and stack them into a column, $q=(u,d,s)^{\mathsf T}$ . The claim of flavour symmetry is that the strong Hamiltonian does not care which basis you write this column in: rotate it by any $3\times3$ unitary matrix of unit determinant and the physics is unchanged, provided the three masses are equal. Such rotations form the group $SU(3)$ , and its eight generators are the operators that step you around a two-dimensional weight lattice — the plane whose axes are the two quantum numbers that survive as good labels, $I_3$ and $Y$ . The first figure plants the three quarks and their antiquarks on that plane; everything else in the theory is built by adding these triangles together.

Figure

The fundamental triplet 3 and antitriplet 3̄ on the weight plane. Each quark carries a fixed isospin I₃ and hypercharge Y — u at (½, ⅓), d at (−½, ⅓), s at (0, −⅔) — and the antiquarks sit at the mirrored points. Three sets of ladder operators, I-spin U-spin V-spin, step between the corners, and the Gell-Mann–Nishijima relation Q = I₃ + Y/2 recovers the fractional charges

The generators split into two kinds. Two of them, $\hat T_3$ and $\hat T_8$ , are diagonal: they measure without moving, and their eigenvalues are precisely $I_3$ and (up to a constant) $Y$ . The other six pair up into three ladder operators — call them $I_\pm$ , $U_\pm$ , $V_\pm$ — each of which slides a state from one corner of a triangle to another. $I_\pm$ swaps $u\leftrightarrow d$ (ordinary isospin), $V_\pm$ swaps $s\leftrightarrow u$ , and $U_\pm$ swaps $d\leftrightarrow s$ . These are the "roots" of the algebra, and they are why the multiplets are hexagons and triangles rather than arbitrary blobs: a multiplet is just the set of states you can reach by walking the ladders, and the ladders point along fixed directions in the $(I_3,Y)$ plane. Add a quark triangle to an antiquark triangle and the nine mesons fall out as a hexagon with a doubly-occupied centre — the octet — plus one extra state sitting alone, the singlet. Add three quark triangles and the ten spin- $\tfrac32$ baryons form a larger triangle. The second figure shows that baryon decuplet, and it carries the single most famous number in the subject: the four rows are equally spaced by about $147\ \mathrm{MeV}$ , and extrapolating one rung past the last known particle predicted the $\Omega^-$ .

Figure

The baryon decuplet 3⊗3⊗3 → 10, a triangle in the I₃–Y plane. Each row down adds one strange quark and drops the hypercharge by one unit, and the measured masses climb in nearly equal steps: Δ 1232, Σ 1385, Ξ 1533, and the apex Ω⁻ 1672 MeV. The equal spacing of about 147 MeV is the strange-quark mass showing through, and it let Gell-Mann predict the Ω⁻ near 1685 MeV before it was found

The core equations

The symmetry is a set of transformations acting on the flavour triplet. A general such rotation is the exponential of the eight generators, and the generators are half the Gell-Mann matrices — the $SU(3)$ analogue of the Pauli matrices — whose commutators define the algebra.

\hat U=\exp\!\Big(i\sum_{a=1}^{8}\alpha_a\hat T_a\Big),\qquad \hat T_a=\tfrac12\lambda_a,\qquad [\hat T_a,\hat T_b]=i f_{abc}\hat T_c \qquad (1)

where: $\hat U$ — a flavour rotation acting on $q=(u,d,s)^{\mathsf T}$ ; $\alpha_a$ — eight real parameters; $\hat T_a$ — the eight Hermitian generators; $\lambda_a$ — the Gell-Mann matrices, normalised so $\mathrm{Tr}(\lambda_a\lambda_b)=2\delta_{ab}$ ; $f_{abc}$ — the totally antisymmetric $SU(3)$ structure constants ( $f_{123}=1$ , $f_{458}=f_{678}=\sqrt3/2$ , and so on). Reading it: $SU(3)$ is what you get from a $3\times3$ unitary matrix ( $9$ parameters) after removing the overall phase ( $-1$ ), leaving $8$ generators. Of the eight generators, only two can be diagonalised at once — the algebra has rank two. Those two, $\hat T_3$ and $\hat T_8$ , are the measuring operators, and their eigenvalues are the two additive quantum numbers that label every state.

I_3=\hat T_3,\qquad Y=\frac{2}{\sqrt3}\,\hat T_8=\tfrac1{\sqrt3}\lambda_8,\qquad Q=I_3+\tfrac12 Y \qquad (2)

where: $I_3$ — third component of isospin; $Y$ — flavour hypercharge, related to baryon number $B$ and strangeness $S$ by $Y=B+S$ ; $Q$ — electric charge, in units of $e$ . Reading it: the last equality is the Gell-Mann–Nishijima relation, and it forces the quark charges $Q(u)=+\tfrac23$ , $Q(d)=Q(s)=-\tfrac13$ from their positions on the weight plane. The remaining six generators combine into three pairs of raising and lowering operators. Each pair generates an $SU(2)$ subgroup and steps between two of the three flavours, tracing out the six roots of the algebra.

I_\pm=\hat T_1\pm i\hat T_2,\qquad V_\pm=\hat T_4\pm i\hat T_5,\qquad U_\pm=\hat T_6\pm i\hat T_7 \qquad (3)

where: $I_\pm$ — isospin ladder, moves $u\leftrightarrow d$ ( $\Delta I_3=\pm1,\ \Delta Y=0$ ); $V_\pm$ — V-spin, moves $s\leftrightarrow u$ ( $\Delta I_3=\pm\tfrac12,\ \Delta Y=\pm1$ ); $U_\pm$ — U-spin, moves $d\leftrightarrow s$ ( $\Delta I_3=\mp\tfrac12,\ \Delta Y=\pm1$ ). Reading it: a multiplet is the orbit of these ladders, which is why hadron patterns are closed hexagons and triangles. Physical hadrons are built by tensoring the fundamental representations. A quark and an antiquark give the meson nonet; three quarks give the baryon multiplets. The Clebsch–Gordan series for $SU(3)$ tells you exactly which multiplets appear.

\mathbf 3\otimes\bar{\mathbf 3}=\mathbf 8\oplus\mathbf 1,\qquad \mathbf 3\otimes\mathbf 3\otimes\mathbf 3=\mathbf{10}\oplus\mathbf 8\oplus\mathbf 8\oplus\mathbf 1 \qquad (4)

where: $\mathbf 3$ — quark triplet; $\bar{\mathbf 3}$ — antiquark antitriplet; $\mathbf 8$ — octet; $\mathbf 1$ — flavour singlet; $\mathbf{10}$ — decuplet. Reading it: $3\times3=9=8+1$ mesons, and $3\times3\times3=27=10+8+8+1$ baryon states — the observed pseudoscalar/vector nonets and the spin- $\tfrac12$ octet and spin- $\tfrac32$ decuplet. When the symmetry is broken by $m_s>m_{u,d}$ , the breaking Hamiltonian transforms as the $Y$ -like ( $\hat T_8$ ) component of an octet. First-order perturbation theory then forces the masses within a multiplet to obey the Gell-Mann–Okubo relation — linear in mass for baryons, quadratic for mesons.

\frac{m_N+m_\Xi}{2}=\frac{3m_\Lambda+m_\Sigma}{4},\qquad 4m_K^2=m_\pi^2+3m_{\eta_8}^2 \qquad (5)

where: $m_N,m_\Lambda,m_\Sigma,m_\Xi$ — baryon-octet masses; $m_\pi,m_K,m_{\eta_8}$ — pseudoscalar-octet masses; $\eta_8$ — the pure octet state (which mixes with the singlet to give the physical $\eta,\eta'$ ). Reading it: both relations follow from the same single symmetry-breaking operator, and both are checked numerically in the worked examples below. The next three derivations build the framework in the order a student should meet it: first count what is actually conserved, then decompose the mesons, then predict the mass splittings.

Derivation ACounting the generators and finding the two good quantum numbers

The goal here is to explain why hadrons carry exactly two additive flavour labels, no more and no fewer, and why those two turn out to be isospin and hypercharge. The starting point is the bare statement that the strong Hamiltonian is invariant under unitary rotations of the $(u,d,s)$ column; the endpoint is the identification of a two-dimensional weight lattice on which every multiplet lives. The intermediate steps count the independent symmetry operators, isolate the ones that commute with everything we want to measure, and read off their physical meaning.

Step 1. Count the generators. A general $3\times3$ complex matrix has $18$ real entries; demanding unitarity ( $U^\dagger U=\mathbb 1$ ) imposes $9$ real constraints, leaving the $9$ parameters of $U(3)$ . Fixing the determinant to $+1$ removes the single overall phase, leaving $9-1=8$ independent generators. Why: the " $-1$ " is the $U(1)$ of baryon number, which is a separate symmetry; stripping it off is what isolates the non-trivial flavour rotations of $SU(3)$ .

Step 2. Ask how many generators can be simultaneously diagonal. The generators do not all commute — equation (1) shows $[\hat T_a,\hat T_b]=if_{abc}\hat T_c\neq0$ in general — so they cannot be measured at once. The largest mutually commuting set (the Cartan subalgebra) here has just two members, $\hat T_3$ and $\hat T_8$ ; one says $SU(3)$ has rank two. Why: only commuting operators share eigenstates, so the number of good additive quantum numbers a state can carry equals the rank, which is why every hadron sits at a point with exactly two coordinates.

Step 3. Read off the two labels. The eigenvalue of $\hat T_3=\tfrac12\lambda_3=\tfrac12\,\mathrm{diag}(1,-1,0)$ is $I_3$ : $+\tfrac12$ for $u$ , $-\tfrac12$ for $d$ , $0$ for $s$ — exactly ordinary isospin restricted to the top-left $SU(2)$ block. The eigenvalue of $\hat T_8=\tfrac1{2\sqrt3}\,\mathrm{diag}(1,1,-2)$ , rescaled to $Y=\tfrac{2}{\sqrt3}\hat T_8=\mathrm{diag}(\tfrac13,\tfrac13,-\tfrac23)$ , is the hypercharge. Why: this fixes the coordinates in the first figure — $u,d,s$ land at $(\tfrac12,\tfrac13),(-\tfrac12,\tfrac13),(0,-\tfrac23)$ — and the six off-diagonal generators become the ladder operators that connect them.

Step 4. Confirm consistency with charge. The Gell-Mann–Nishijima relation $Q=I_3+\tfrac12 Y$ must reproduce the known fractional quark charges. For $u$ : $\tfrac12+\tfrac12\cdot\tfrac13=+\tfrac23$ ; for $d$ : $-\tfrac12+\tfrac16=-\tfrac13$ ; for $s$ : $0-\tfrac13=-\tfrac13$ . Why: charge is not an independent input but a fixed linear combination of the two Cartan labels, which is the deep reason the electromagnetic and flavour structure of hadrons are locked together.

Go deeper · Derivation B · 2 minDecomposing quark ⊗ antiquark into an octet and a singlet

Derivation BDecomposing quark ⊗ antiquark into an octet and a singlet

The aim is to show explicitly how the nine quark–antiquark states organise into an $\mathbf 8$ and a $\mathbf 1$ , and in particular to construct the singlet — the one combination that is untouched by every $SU(3)$ rotation — and the neutral members of the octet. This is the content of $\mathbf 3\otimes\bar{\mathbf 3}=\mathbf 8\oplus\mathbf 1$ made concrete, and it is what underlies the observed pseudoscalar nonet $(\pi,K,\eta,\eta')$ . We place the states on the weight plane, apply the ladders, and then isolate the singlet by demanding total flavour neutrality.

Step 1. Place the nine states. An antiquark transforms in $\bar{\mathbf 3}$ , whose weights are the negatives of the quark weights: $\bar u$ at $(-\tfrac12,-\tfrac13)$ , $\bar d$ at $(+\tfrac12,-\tfrac13)$ , $\bar s$ at $(0,+\tfrac23)$ . Adding each antiquark weight to each quark weight gives nine points; six land on the corners of a hexagon (the charged/strange mesons $\pi^\pm,K^\pm,K^0,\bar K^0$ ) and three pile up at the centre $(0,0)$ . Why: the hexagon-plus-triple pattern is the visual signature of $\mathbf 8\oplus\mathbf 1$ — an octet is a hexagon with a doubly-filled centre, and the singlet supplies the third central state.

Step 2. Resolve the three central states. At $(I_3,Y)=(0,0)$ sit $u\bar u,\ d\bar d,\ s\bar s$ . The isospin ladder fixes one combination as the $I=1$ member $\pi^0=\tfrac1{\sqrt2}(u\bar u-d\bar d)$ . The other two are the $I=0$ states, and orthogonality plus the $SU(3)$ structure selects the octet and singlet combinations. Why: symmetry cannot mix states of different $I$ , so the $\pi^0$ separates cleanly; the remaining $2\times2$ block is where the octet and singlet, and later their physical mixing into $\eta$ and $\eta'$ , live.

Step 3. Build the singlet. The flavour singlet must be annihilated by every ladder operator — invariant under all of $SU(3)$ — which forces the symmetric, democratic combination $|\mathbf 1\rangle=\tfrac1{\sqrt3}(u\bar u+d\bar d+s\bar s)$ . The orthogonal $I=0$ octet state is $\eta_8=\tfrac1{\sqrt6}(u\bar u+d\bar d-2s\bar s)$ . Why: only the fully symmetric sum treats the three flavours identically, so only it can be a scalar under flavour rotations; the relative $-2$ in $\eta_8$ is exactly what makes it orthogonal to the singlet and part of the octet.

Step 4. Note the physical consequence. Because $SU(3)$ is broken, the singlet $\eta_1$ and octet $\eta_8$ are not mass eigenstates; they mix by an angle $\theta_P\approx-11^\circ$ to $-24^\circ$ into the observed $\eta(548)$ and $\eta'(958)$ . Why: this mixing is the first place the "approximate" in "approximate symmetry" bites — an exact $SU(3)$ would forbid it, so the $\eta$ – $\eta'$ system is a direct measure of how badly the symmetry is broken.

Go deeper · Derivation C · 2 minThe Gell-Mann–Okubo mass formula and the strange-quark splitting

Derivation CThe Gell-Mann–Okubo mass formula and the strange-quark splitting

The goal is to derive the relation that ties the baryon-octet masses together and to see why the decuplet splits into equal rungs — the quantitative payoff of the whole scheme. The physical input is that the symmetry-breaking term in the Hamiltonian is not arbitrary: because it comes from the strange-quark mass, which singles out the third flavour, it transforms as the $\hat T_8$ -like member of an $SU(3)$ octet. First-order perturbation theory in that single operator then constrains the masses. We write the most general such mass operator, evaluate it across a multiplet, and eliminate the unknown constants.

Step 1. Write the mass operator. To first order the mass of a state is $m=m_0+a\,Y+b\big[I(I+1)-\tfrac14 Y^2\big]$ , where $m_0,a,b$ are three unknown constants common to the whole multiplet. Why: these are the only two independent scalars ( $Y$ , and the combination $I(I+1)-\tfrac14Y^2$ ) that a $\hat T_8$ -like octet operator can contribute at first order, so any multiplet's masses are fixed by just three numbers — and any relation that eliminates all three is a parameter-free prediction.

Step 2. Apply it to the baryon octet. The four isospin multiplets have $(Y,I)$ : nucleon $(1,\tfrac12)$ , $\Lambda$ $(0,0)$ , $\Sigma$ $(0,1)$ , $\Xi$ $(-1,\tfrac12)$ . Substituting and forming the combination that cancels $m_0,a,b$ gives $\tfrac12(m_N+m_\Xi)=\tfrac14(3m_\Lambda+m_\Sigma)$ . Why: the left side collects the $Y=\pm1$ states (whose $a\,Y$ terms cancel in the sum) and the right side collects the $Y=0$ states, arranged so the $b$ -term matches — three unknowns, four masses, one clean prediction.

Step 3. Apply it to the decuplet. The decuplet has $I=Y/2+\text{const}$ across its rows, so the $b$ -term collapses and the formula reduces to $m=m_0'+a'\,Y$ — masses linear in $Y$ , i.e. equally spaced by $|a'|$ as you step down one unit of hypercharge (one extra strange quark). Why: the equal-spacing rule is not an accident but the decuplet's special geometry making the quadratic term vanish; the common spacing $\approx147\ \mathrm{MeV}$ is a direct read-out of the effective strange-quark mass inside a baryon.

Step 4. Turn it into a prediction. The equal-spacing rule extrapolated below the $\Xi^*$ demands an $I=0$ , $S=-3$ , $sss$ state one rung down — the $\Omega^-$ — at $m_{\Xi^*}+147\approx1685\ \mathrm{MeV}$ . Why: this converted a pattern into a falsifiable number with definite quantum numbers, and the $1964$ discovery at $1672\ \mathrm{MeV}$ turned $SU(3)$ from a classification scheme into established physics.

Limiting cases

Case	Result	Meaning
Exact $SU(3)$ ( $m_u=m_d=m_s$ )	all octet members degenerate	a hypothetical world with one light-quark mass; no $\Sigma$ – $\Lambda$ or $N$ – $\Xi$ splitting
Restrict to $u,d$ block	$SU(3)\to SU(2)$ isospin	the top-left generators $\hat T_1,\hat T_2,\hat T_3$ close on their own; the good, nearly-exact subsymmetry
First-order breaking	Gell-Mann–Okubo relations	$m_s>m_{u,d}$ enters linearly; predicts $(m_N+m_\Xi)/2=(3m_\Lambda+m_\Sigma)/4$ to $\sim0.6\%$
Heavy strange quark ( $m_s\to\infty$ )	$SU(3)\to SU(2)$ , $s$ decouples	strange hadrons lifted away; only the $u,d$ multiplets remain light
Add charm/bottom/top	$SU(3)$ not extended to $SU(4)+$	heavier quark masses split too far; $SU(4)$ flavour is badly broken and not a useful classification

Common trapthe misconception

It is tempting to treat flavour $SU(3)$ as if it were as good as isospin, and to read exact predictions off the multiplets. It is not. Isospin ( $SU(2)$ ) is broken only by the tiny $m_d-m_u\sim2\ \mathrm{MeV}$ difference plus electromagnetism, so it holds at the percent level or better. Flavour $SU(3)$ is broken by $m_s-m_{u,d}\sim100\ \mathrm{MeV}$ — fifty times larger — so its predictions are good to roughly $10$ – $25\%$ for masses and can fail badly for rates and mixings. The clearest symptom is the $\eta$ – $\eta'$ system: exact $SU(3)$ would keep the octet $\eta_8$ and singlet $\eta_1$ separate, but the real states mix substantially and differ in mass by more than $400\ \mathrm{MeV}$ . Use $SU(3)$ to organise and to predict patterns and orderings; treat any quantitative flavour- $SU(3)$ number as carrying a symmetry-breaking uncertainty unless you have gone to higher order or used lattice QCD.

Worked examples

1 · Testing Gell-Mann–Okubo on the baryon octet (fully worked)

Using the measured masses $m_N=938.9$ , $m_\Lambda=1115.7$ , $m_\Sigma=1193.2$ , $m_\Xi=1318.3\ \mathrm{MeV}$ , check the Gell-Mann–Okubo relation and quantify the residual symmetry breaking.

Step 1. Evaluate the left-hand side — the average of the two $Y=\pm1$ multiplets: $\tfrac12(m_N+m_\Xi)=\tfrac12(938.9+1318.3)=\tfrac12(2257.2)=\mathbf{1128.6\ \mathrm{MeV}}$ .

Step 2. Evaluate the right-hand side — the weighted average of the two $Y=0$ multiplets: $\tfrac14(3m_\Lambda+m_\Sigma)=\tfrac14(3\times1115.7+1193.2)=\tfrac14(3347.1+1193.2)=\tfrac14(4540.3)=\mathbf{1135.1\ \mathrm{MeV}}$ .

Step 3. Compare. The two sides differ by $1135.1-1128.6=6.5\ \mathrm{MeV}$ , a fractional discrepancy of $6.5/1128.6=\mathbf{0.6\%}$ . For a symmetry broken by an effect of order $m_s\sim100\ \mathrm{MeV}$ on masses of order $1000\ \mathrm{MeV}$ , one might have expected $\sim10\%$ errors; the relation does far better because it is constructed to cancel the first-order breaking, leaving only second-order ( $m_s^2$ ) residuals. GMO holds to $0.6\%$ — one of the most accurate predictions in hadron spectroscopy.

2 · Predicting the Ω⁻ from the decuplet (fully worked)

The spin- $\tfrac32$ decuplet masses are $m_\Delta=1232$ , $m_{\Sigma^*}=1385$ , $m_{\Xi^*}=1533\ \mathrm{MeV}$ . Predict the mass and quantum numbers of the missing apex, and compare to experiment.

Step 1. Extract the rung spacing. Going down the multiplet, $m_{\Sigma^*}-m_\Delta=1385-1232=153\ \mathrm{MeV}$ and $m_{\Xi^*}-m_{\Sigma^*}=1533-1385=148\ \mathrm{MeV}$ . The spacings agree to a few MeV, confirming the equal-spacing rule; take the mean, $\overline{\Delta m}\approx150\ \mathrm{MeV}$ .

Step 2. Extrapolate one rung. The apex sits one unit of hypercharge below $\Xi^*$ : $m_{\Omega}\approx m_{\Xi^*}+\overline{\Delta m}=1533+150=\mathbf{1683\ \mathrm{MeV}}$ (Gell-Mann's $1962$ estimate was $\approx1685\ \mathrm{MeV}$ ).

Step 3. Read off the quantum numbers from the geometry. The apex lies at $I_3=0$ (so $I=0$ , an isosinglet), $Y=-2$ , hence $S=Y-B=-2-1=-3$ : three strange quarks, $sss$ . Its charge is $Q=I_3+\tfrac12 Y=0+\tfrac12(-2)=-1$ , and being in the spin- $\tfrac32$ decuplet it has $J^P=\tfrac32^+$ .

Step 4. Compare with experiment. The $\Omega^-$ was found at Brookhaven in $1964$ with mass $1672\ \mathrm{MeV}$ , $S=-3$ , $Q=-1$ , $J^P=\tfrac32^+$ — matching every predicted quantum number and the mass to under $1\%$ . Prediction $\approx1685$ , observation $1672\ \mathrm{MeV}$ : the discovery that established flavour $SU(3)$ .

3 · The pseudoscalar octet and η–η′ mixing

Use the quadratic Gell-Mann–Okubo relation $4m_K^2=m_\pi^2+3m_{\eta_8}^2$ with $m_\pi=138$ , $m_K=495.6\ \mathrm{MeV}$ to predict the octet mass $m_{\eta_8}$ , and interpret the gap to the physical $\eta$ and $\eta'$ .

Step 1. Solve for the octet mass: $m_{\eta_8}^2=\tfrac13(4m_K^2-m_\pi^2)=\tfrac13(4\times495.6^2-138^2)=\tfrac13(982\,400-19\,044)\ \mathrm{MeV}^2=321\,000\ \mathrm{MeV}^2$ , so $m_{\eta_8}=\mathbf{567\ \mathrm{MeV}}$ .

Step 2. Interpret. The physical $\eta$ sits at $548\ \mathrm{MeV}$ — close to the predicted $\eta_8$ — while the $\eta'$ at $958\ \mathrm{MeV}$ lies far above any octet member. The gap signals that the physical states are mixtures of the octet $\eta_8$ and the singlet $\eta_1$ , with a mixing angle $\theta_P\approx-11^\circ$ to $-24^\circ$ . Mesons obey GMO in mass-squared, and the $\eta'$ anomaly is the singlet refusing to stay put — a preview of the axial $U(1)$ problem solved by QCD instantons.

Physics of the portrait

Flavour $SU(3)$ is the work of one physicist who saw a periodic table where others saw a zoo, and who trusted the pattern enough to bet a new particle on an empty seat.

Origins · 2 minMurray Gell-Mann and how this was discovered

Murray Gell-Mann

1929–2019 · the Eightfold Way and the quark model

Murray Gell-Mann was, by common consent, the dominant theorist of hadron physics in its formative decades. A prodigy who entered Yale at fifteen, he made his first mark in 1953 by introducing strangeness — a new additive quantum number, conserved by the strong force but violated by the weak — and, with Kazuhiko Nishijima, the charge relation $Q=I_3+\tfrac12 Y$ that still bears both their names. By the late 1950s the proliferation of "strange" particles had become an embarrassment of riches, a zoo without an organising principle. In 1961 Gell-Mann (and, independently, Yuval Ne'eman) found the principle: the flavour group $SU(3)$ , whose eight generators he playfully christened the Eightfold Way after the Buddhist path. Hadrons, he argued, must fill $SU(3)$ multiplets — octets and decuplets — and their quantum numbers should tile the $(I_3,Y)$ plane. The theory's boldest moment came in 1962 at the CERN conference, where the spin- $\tfrac32$ baryons formed a decuplet with nine of ten seats occupied. Gell-Mann stood up and predicted the tenth: a particle he named the $\Omega^-$ , with strangeness $-3$ , charge $-1$ , and a mass he could read off the equal-spacing rule, near $1685\ \mathrm{MeV}$ . Its discovery at Brookhaven in 1964 was a sensation. That same year, chasing the deeper meaning of the $\mathbf 3$ that generated everything, Gell-Mann proposed quarks — fractionally charged constituents, three flavours filling the fundamental representation (George Zweig independently proposed the same, calling them "aces"). Gell-Mann received the 1969 Nobel Prize in Physics "for his contributions and discoveries concerning the classification of elementary particles and their interactions." His fingerprints are everywhere in the modern Standard Model: strangeness, the Eightfold Way, quarks, current algebra, and — with Harald Fritzsch and Heinrich Leutwyler — the naming and shaping of quantum chromodynamics itself.

History

1953

— Gell-Mann and Nishijima introduce strangeness and the relation Q = I₃ + Y/2

1961

— Gell-Mann and, independently, Ne'eman propose the Eightfold Way, flavour SU(3)

1962

— Gell-Mann and Okubo derive the mass formula; Gell-Mann predicts the Ω⁻ at CERN

1964

— the Ω⁻ is discovered at Brookhaven at 1672 MeV; Gell-Mann and Zweig propose quarks

1969

— Gell-Mann receives the Nobel Prize in Physics

Going further

Texts. Thomson, Modern Particle Physics, §9.6 (SU(3) flavour symmetry) and §9.6.1–9.6.2 · Griffiths, Introduction to Elementary Particles, 2nd ed., Ch. 5 (Bound States), §5.1–5.5 (the Eightfold Way and the quark model) · Halzen & Martin, Quarks and Leptons, Ch. 2–2 (SU(3) and the quark model) · Peskin & Schroeder, An Introduction to Quantum Field Theory, §15.4 and §A (Lie-algebra conventions) · Georgi, Lie Algebras in Particle Physics, 2nd ed., Ch. 6–7 (SU(3) roots, weights, and tensor methods) · Particle Data Group, Review of Particle Physics (baryon and meson mass tables). Key numerical checks: (1) $9-1=8$ generators, rank $2$ ; (2) $Q=I_3+\tfrac12 Y$ returns $+\tfrac23,-\tfrac13,-\tfrac13$ for $u,d,s$ ; (3) baryon GMO $(m_N+m_\Xi)/2=1128.6$ vs $(3m_\Lambda+m_\Sigma)/4=1135.1\ \mathrm{MeV}$ , agreement to $0.6\%$ ; (4) decuplet spacings $153,148,139\ \mathrm{MeV}$ and $\Omega^-$ predicted $\approx1685$ vs observed $1672\ \mathrm{MeV}$ ; (5) $m_{\eta_8}=567\ \mathrm{MeV}$ from $4m_K^2=m_\pi^2+3m_{\eta_8}^2$ .

Prerequisites. Isospin · Symmetries in Quantum Mechanics · Flavour Symmetry. Leads to. Meson Multiplets · Combining Quarks into Baryons · Colour and QCD.

Ask the tutor: ask me to walk you through building the pseudoscalar meson nonet seat by seat with the ladder operators, or to re-derive the $\Omega^-$ prediction from the decuplet spacing.

Is flavour SU(3) the same as the colour SU(3) of QCD?

No — and confusing them is the single most common error. They are two different $SU(3)$ groups acting on two different labels. Flavour $SU(3)$ rotates the three light quark flavours $(u,d,s)$ ; it is approximate, broken by the $\sim100\ \mathrm{MeV}$ spread in their masses, and it is not a gauge symmetry — no force couples to it. Colour $SU(3)$ rotates the three colour charges $(r,g,b)$ that every quark carries; it is exact, unbroken, and gauged, with the eight gluons as its force carriers. Flavour $SU(3)$ organises the hadron spectrum into multiplets; colour $SU(3)$ is the dynamics that binds quarks in the first place. The two share only the abstract group and its Gell-Mann matrices; physically they could hardly be more different.

What exactly is hypercharge, and how does it relate to strangeness?

Hypercharge $Y$ is one of the two additive quantum numbers that label $SU(3)$ states — the eigenvalue of the second diagonal generator, $Y=\tfrac{2}{\sqrt3}\hat T_8$ . Physically it packages baryon number and strangeness together: $Y=B+S$ (for the light quarks; with charm and beyond one adds those flavour numbers too). For an ordinary baryon $B=1$ , so $Y=1+S$ ; the nucleon ( $S=0$ ) has $Y=1$ , and the $\Omega^-$ ( $S=-3$ ) has $Y=-2$ . Its usefulness is that, together with $I_3$ , it gives every hadron a unique address on the weight plane, and through Gell-Mann–Nishijima $Q=I_3+\tfrac12 Y$ it fixes the electric charge. The vertical axis of every multiplet diagram is $Y$ .

Why do the mesons obey Gell-Mann–Okubo in mass-squared but the baryons in mass?

It is an empirical fact with a field-theory rationale. For baryons — fermions — the natural mass scale in the Dirac equation is linear in $m$ , and the linear Gell-Mann–Okubo relation $(m_N+m_\Xi)/2=(3m_\Lambda+m_\Sigma)/4$ works to $0.6\%$ . For mesons — bosons — the Klein-Gordon and Bethe–Salpeter descriptions make $m^2$ the natural variable, so the octet relation reads $4m_K^2=m_\pi^2+3m_{\eta_8}^2$ . In practice using mass-squared for mesons and mass for baryons is what fits the data best; the underlying reason is that the symmetry-breaking operator enters the two-point function as a mass shift, which is linear in the pole mass for fermions and appears as $m^2$ in the propagator for scalars.

How did Gell-Mann know the Ω⁻ would have those exact quantum numbers?

Because they are forced by the geometry of the decuplet, not guessed. The spin- $\tfrac32$ baryons form a triangular multiplet with four rows of decreasing hypercharge. Nine seats were filled — the four $\Delta$ 's, three $\Sigma^*$ 's, two $\Xi^*$ 's — leaving exactly one empty seat at the apex. Its location on the weight plane is its quantum numbers: $I_3=0$ (so isospin $I=0$ ), $Y=-2$ (so strangeness $S=Y-B=-3$ ), charge $Q=I_3+\tfrac12 Y=-1$ , and spin-parity $\tfrac32^+$ from the multiplet it belongs to. The equal-spacing rule then supplied the mass. Everything about the particle except its lifetime was fixed before anyone saw it — which is why the 1964 discovery was so persuasive.

Could the scheme be extended to a fourth quark, an SU(4) flavour symmetry?

Formally yes, physically barely. Adding the charm quark gives $SU(4)$ flavour, and the multiplets become three-dimensional polyhedra (the mesons fill a $\mathbf 4\otimes\bar{\mathbf 4}=\mathbf{15}\oplus\mathbf 1$ ). But a symmetry is only useful to the extent it is a good approximation, and the charm quark at $\sim1.3\ \mathrm{GeV}$ is more than ten times heavier than the up and down. The breaking is so large that $SU(4)$ makes essentially no reliable predictions — the "multiplet" members have wildly different masses. This is why particle physics uses flavour $SU(3)$ but treats charm, bottom, and top with heavy-quark symmetries of a completely different kind, exploiting that those quarks are heavy rather than that they are light and interchangeable.

Where does the flavour singlet go, and why is the η′ so heavy?

The flavour singlet $\eta_1=\tfrac1{\sqrt3}(u\bar u+d\bar d+s\bar s)$ is the combination invariant under every $SU(3)$ rotation. Naively it should sit near the octet, and if $SU(3)$ and the classical symmetries held exactly it would be light. Instead the physical $\eta'$ — mostly singlet — weighs $958\ \mathrm{MeV}$ , far above the octet. The resolution lies outside flavour $SU(3)$ : the singlet couples to the axial- $U(1)$ anomaly of QCD, and gluonic effects (instantons) give it a large extra mass that the octet does not receive. So the $\eta'$ mass is a window onto non-perturbative QCD, and the small residual $\eta$ – $\eta'$ mixing ( $\theta_P\approx-11^\circ$ to $-24^\circ$ ) measures how the octet and singlet leak into each other once the symmetry is broken.