Grand unification (GUTs)

The Standard Model is a triumph built on an apology. Its gauge group, $SU(3)\times SU(2)\times U(1)$ , is a product of three factors, and a product group carries one coupling constant per factor — the strong $g_s$ , the weak $g$ , and the hypercharge $g'$ . Three numbers, put in by hand, plus a scattering of fractional electric charges and a menagerie of fermion multiplets that look unrelated: this is the price of admission. Grand unification is the conjecture that the apology is unnecessary — that the three factors are the low-energy splinters of a single simple group $G$ , with one coupling, broken at some enormous energy $M_{\text{GUT}}$ down to the group we see. If so, the three couplings are not three numbers but one number viewed through three windows, and they must converge when extrapolated to high enough energy. The astonishing thing is that they almost do. The couplings are not constants: quantum corrections make them run logarithmically with energy, and the strong coupling weakens (asymptotic freedom) while the electromagnetic-hypercharge coupling grows. Run them up from the $Z$ mass and the three curves sweep toward one another, crossing in a tight knot near $10^{15}$ – $10^{16}$ GeV — a scale no accelerator will ever reach, yet one the couplings point to unambiguously. This single observation, made by Georgi, Quinn, and Weinberg in 1974, turned unification from philosophy into a quantitative, falsifiable program: it predicts the weak mixing angle, quantizes electric charge, and — fatally or triumphantly — makes the proton decay.

Key takeaways

The Standard Model's three gauge couplings are independent only because its group is a product; a single simple group would give one coupling, forcing the three to unify at a high scale $M_{\text{GUT}}$ .
Couplings run logarithmically: $\alpha_i^{-1}(\mu)=\alpha_i^{-1}(M_Z)-\dfrac{b_i}{2\pi}\ln\dfrac{\mu}{M_Z}$ , and extrapolated from $M_Z$ they converge near $10^{15}$ – $10^{16}$ GeV.
The minimal group is $SU(5)$ (Georgi–Glashow, 1974): one generation's 15 left-handed fermions fill $\bar{\mathbf 5}\oplus\mathbf{10}$ , and 12 new $X,Y$ leptoquark bosons join the 12 Standard-Model gauge bosons in the adjoint $\mathbf{24}$ .
Because electric charge is a traceless generator of $SU(5)$ , summing $Q$ over a multiplet gives zero — which forces $Q_d=-\tfrac13 Q_e$ : charge is quantized in thirds because there are three colours.
Unification predicts $\sin^2\theta_W=\tfrac38$ at $M_{\text{GUT}}$ , running down to the observed $\approx0.231$ ; and it predicts proton decay $p\to e^+\pi^0$ with $\tau\sim M_X^4/(\alpha_{\text{GUT}}^2 m_p^5)$ .
Key idea: the three Standard-Model couplings are independent only because the group is a product; if that product is the low-energy shadow of one simple group, the couplings converge at a high scale, electric charge is quantized, sin²θ_W is predicted, and the proton must decay.

The core insight

The three Standard-Model gauge couplings are independent only because the symmetry group is a product; if that product is the low-energy shadow of one larger simple group with a single coupling, the three couplings must converge at a high scale, the fractional quark charges are forced to be quantized, the weak mixing angle is predicted, and baryon number is no longer exactly conserved — so the proton decays.

Where this lives today

Grand unification is untestable at accelerators — no machine will reach $10^{16}$ GeV — so it is probed indirectly, and three fronts are active. The first is proton decay: Super-Kamiokande's 50-kiloton water tank has watched $\sim10^{34}$ proton-years and set the world's best limits, and its successor Hyper-Kamiokande (starting later this decade, roughly eight times the mass) together with the liquid-argon DUNE detector will push the reach on $p\to K^+\bar\nu$ — the golden mode of supersymmetric GUTs — into the $10^{34}$ – $10^{35}$ -year range where those models actually live. A single confirmed event would be among the most important discoveries in the history of physics. The second front is neutrino mass: the $SO(10)$ spinor demands a heavy right-handed neutrino, and the seesaw mechanism it enables, $m_\nu\sim m_D^2/M_R$ with $M_R\sim M_{\text{GUT}}$ , naturally explains why neutrinos are light — tying the sub-eV neutrino masses now measured in oscillation experiments directly to the unification scale. The third is precision coupling running: the LHC's measurements of $\alpha_3$ and the electroweak couplings, extrapolated with the renormalization group, continue to test whether the curves truly converge, and the quality of the meeting remains the sharpest low-energy argument for supersymmetry or other new thresholds. Grand unification thus sits at the crossroads of the deepest underground detectors, the lightest known particles, and the highest-energy colliders — a scale we cannot reach, inferred from clues we can.

The physical picture

The whole case for grand unification rests on one graph. The gauge couplings are not fixed numbers but slowly varying functions of the energy at which you probe: vacuum polarization dresses each charge differently, and the three run at different rates set by the particle content that carries each charge. The strong coupling is asymptotically free and weakens with energy; the weak coupling weakens more gently; the hypercharge coupling, alone among them, grows. Plot the inverse couplings against the logarithm of energy and you get three straight lines with different slopes — and, extrapolated across a dozen orders of magnitude, they aim at very nearly the same point.

Figure

The three inverse couplings run toward each other. Starting from the measured values α₁⁻¹ = 59.0, α₂⁻¹ = 29.6, α₃⁻¹ = 8.5 at the Z mass and running with one-loop slopes fixed by b = 41/10, −19/6, −7, the Standard-Model lines nearly meet near 10¹⁵ GeV but miss: the three pairwise crossings spread from 10¹³ to 10¹⁷ GeV. Adding supersymmetry sharpens the meeting to a single point at 2×10¹⁶ GeV with α_GUT⁻¹ ≈ 24

If the couplings truly meet, the meeting point $M_{\text{GUT}}$ is where the larger symmetry is restored. Below it, the unified group is broken and the fermions we know are shattered fragments of its multiplets; above it, they reassemble. The minimal candidate group is $SU(5)$ : it is the smallest simple group with room for $SU(3)\times SU(2)\times U(1)$ , and one full generation of fifteen left-handed Weyl fermions fits — with nothing left over and nothing missing — into two of its representations, the $\bar{\mathbf 5}$ and the $\mathbf{10}$ . The second figure is where the arithmetic becomes magic: once the fermions sit in these multiplets, the demand that electric charge be a traceless $SU(5)$ generator forces the down quark's charge to be exactly $-\tfrac13$ the electron's, and fixes the weak mixing angle at unification to the pure number $\tfrac38$ .

Figure

One generation fits into two SU(5) multiplets. The anti-fundamental 5̄ holds the three colours of the down antiquark d^c (Q = +1/3 each) plus the lepton doublet (ν, e⁻); the antisymmetric 10 holds the quark doublet, u^c, and e⁺. Because electric charge Q is a traceless generator, its sum over each multiplet is zero: 3(+1/3) + 0 + (−1) = 0. That single constraint quantizes charge — Q_d = −1/3 Q_e — and the ratio Tr T₃² / Tr Q² = (1/2)/(4/3) fixes sin²θ_W = 3/8 at the GUT scale

The core equations

The engine of unification is the renormalization-group running of the couplings. Writing $\alpha_i=g_i^2/4\pi$ for the three factors ( $i=1,2,3$ for $U(1)_Y$ , $SU(2)_L$ , $SU(3)_C$ ), each inverse coupling runs linearly in the logarithm of energy at one loop, with a slope fixed entirely by the particle content.

\frac{1}{\alpha_i(\mu)}=\frac{1}{\alpha_i(M_Z)}-\frac{b_i}{2\pi}\,\ln\frac{\mu}{M_Z},\qquad (b_1,b_2,b_3)=\left(\tfrac{41}{10},\,-\tfrac{19}{6},\,-7\right) \qquad (1)

where: $\alpha_i$ — the running coupling of factor $i$ ; $\mu$ — energy scale; $M_Z=91.19$ GeV — the reference scale; $b_i$ — one-loop beta coefficients for the Standard Model with three generations and one Higgs doublet. Read the signs physically: $b_3=-7<0$ makes $\alpha_3^{-1}$ grow with energy (the strong coupling weakens — asymptotic freedom), while $b_1=+\tfrac{41}{10}>0$ makes $\alpha_1^{-1}$ shrink (hypercharge strengthens). The couplings run toward each other. To compare the three on an equal footing the hypercharge coupling must be normalized so that all generators are measured with the same yardstick inside the unified group. This is the origin of the famous factor $\tfrac53$ .

\alpha_1=\frac{5}{3}\,\alpha_Y,\qquad \alpha_Y=\frac{\alpha}{\cos^2\theta_W},\qquad g'=\sqrt{\tfrac35}\,g_1 \qquad (2)

where: $\alpha_Y$ — the "bare" hypercharge coupling as it appears in the Standard Model; $\alpha_1$ — its GUT-normalized version; $\alpha=e^2/4\pi$ — the fine-structure constant. The factor $\tfrac53$ comes from requiring $\text{Tr}\,T^2$ be equal for every generator of the simple group; without it the couplings would never meet. Unification is the statement that all three normalized couplings coincide at one scale, where the single coupling of the parent group takes over.

\alpha_1(M_{\text{GUT}})=\alpha_2(M_{\text{GUT}})=\alpha_3(M_{\text{GUT}})\equiv\alpha_{\text{GUT}} \qquad (3)

where: $M_{\text{GUT}}$ — the grand-unification scale; $\alpha_{\text{GUT}}$ — the common coupling there. Two equations (three curves through one point is two conditions) turn three measured inputs into two predictions: the value of $M_{\text{GUT}}$ and one relation among the low-energy couplings — historically, a prediction of $\sin^2\theta_W$ . At the unification scale the weak mixing angle is not a free parameter but a pure group-theory number, obtained as the ratio of the trace of $T_3^2$ to the trace of $Q^2$ over any complete multiplet.

\sin^2\theta_W(M_{\text{GUT}})=\frac{\text{Tr}\,T_3^2}{\text{Tr}\,Q^2}=\frac{1/2}{4/3}=\frac{3}{8} \qquad (4)

where: $T_3$ — the third weak-isospin generator; $Q$ — the electric charge operator. The value $\tfrac38=0.375$ is fixed once and for all by how the fermions sit in the multiplets; running it down to $M_Z$ with equation (1) yields the observed $\sin^2\theta_W\approx0.231$ . Finally, the unified group's extra gauge bosons — the $X$ and $Y$ leptoquarks of mass $M_X\sim M_{\text{GUT}}$ — connect quarks to leptons and violate baryon number, so the proton becomes unstable. Dimensional analysis fixes its lifetime.

\Gamma_p\sim \alpha_{\text{GUT}}^2\,\frac{m_p^5}{M_X^4},\qquad \tau_p\sim\frac{1}{\alpha_{\text{GUT}}^2}\,\frac{M_X^4}{m_p^5} \qquad (5)

where: $m_p$ — proton mass; $M_X$ — leptoquark mass. The decay proceeds through a four-fermion contact interaction of strength $g^2/M_X^2$ , exactly like Fermi theory with $M_X$ in place of $M_W$ ; the enormous $M_X^4$ in the denominator of $\tau_p$ is what stretches the lifetime to $\sim10^{30}$ years and beyond, making it the sharpest testable prediction of the whole framework.

Derivation AWhere the couplings meet, and what the meeting predicts

The goal is to show that the three measured low-energy couplings, run upward with the one-loop equations, converge near $10^{15}$ – $10^{16}$ GeV, and that the pattern of their convergence turns three inputs into a prediction of the unification scale and the weak mixing angle. We take the values at $M_Z$ , extrapolate each along its own slope, locate the crossings, and read off both what works and what the plain Standard Model gets wrong.

Step 1. Fix the starting points. From the electroweak measurements $\alpha^{-1}(M_Z)=127.9$ and $\sin^2\theta_W=0.231$ , together with $\alpha_3(M_Z)=0.118$ , the three normalized inverse couplings are $\alpha_1^{-1}=\tfrac35\cos^2\theta_W\,\alpha^{-1}=59.0$ , $\alpha_2^{-1}=\sin^2\theta_W\,\alpha^{-1}=29.6$ , and $\alpha_3^{-1}=8.5$ . Why: these three numbers are the entire experimental input — everything downstream is pure extrapolation, so the prediction stands or falls on them alone.

Step 2. Run each line. Equation (1) gives slopes per decade of energy of $-1.50$ , $+1.16$ , and $+2.57$ for $i=1,2,3$ ; the strong coupling, starting lowest, climbs fastest, while the hypercharge coupling, starting highest, descends. Why: the different slopes are what let three curves that are far apart at $M_Z$ come together at all — unification is a race whose finish line exists only because the runners move at different speeds.

Step 3. Find the crossings. Setting the lines equal pairwise gives $\alpha_1=\alpha_2$ at $10^{13.0}$ GeV, $\alpha_1=\alpha_3$ at $10^{14.4}$ GeV, and $\alpha_2=\alpha_3$ at $10^{17.0}$ GeV. In the plain Standard Model the three do not pass through a single point — the crossings straddle roughly four orders of magnitude. Why: this near-miss is the honest experimental fact; it says unification is suggested but not achieved by the minimal field content, and something must be added.

Step 4. Restore exact meeting, and extract the payoff. Adding the superpartners of the Standard Model (the MSSM) changes the coefficients to $(b_1,b_2,b_3)=(\tfrac{33}{5},1,-3)$ ; the lines then cross at a single point $M_{\text{GUT}}\approx2\times10^{16}$ GeV with $\alpha_{\text{GUT}}^{-1}\approx24$ . Turned around, demanding unification predicts $\sin^2\theta_W\approx0.231$ from $\tfrac38$ run down — a number known to permille precision. Why: that a two-parameter fit reproduces the measured mixing angle, and that supersymmetry sharpens a near-miss into a bullseye, is the strongest quantitative hint for both grand unification and low-energy supersymmetry.

Go deeper · Derivation B · 1 minCharge quantization and sin²θ_W = 3/8 from the SU(5) embedding

Derivation BCharge quantization and sin²θ_W = 3/8 from the SU(5) embedding

The aim is to show that placing one generation into the $\bar{\mathbf 5}\oplus\mathbf{10}$ of $SU(5)$ does two things no Standard-Model argument can: it forces the fractional quark charges to be exactly what they are, and it predicts the weak mixing angle at the unification scale to be the pure number $\tfrac38$ . Both follow from a single fact — that inside a simple group the electric charge operator is a generator, and generators are traceless.

Step 1. List the $\bar{\mathbf 5}$ . In left-handed convention it contains the colour anti-triplet $d^c$ (three states) and the lepton doublet $(\nu_e,e^-)$ : five states in all. Their electric charges are $Q_{d^c}$ (three copies), $Q_\nu=0$ , and $Q_{e^-}=-1$ . Why: the multiplet mixes a quark and leptons, so any statement about the multiplet as a whole ties quark and lepton charges together — which is exactly what quantization needs.

Step 2. Impose tracelessness. Since $Q$ is a generator of $SU(5)$ , $\text{Tr}\,Q=0$ over the multiplet: $3Q_{d^c}+Q_\nu+Q_{e^-}=0$ , i.e. $3Q_{d^c}=+1$ , so $Q_{d^c}=+\tfrac13$ and hence $Q_d=-\tfrac13$ . Why: the down quark's charge is not fitted to data — it is forced to be one-third of the electron's, and the "one-third" is literally the reciprocal of the three colours summed over. Charge quantization becomes a theorem, not an accident.

Step 3. Compute the normalization ratio. Over the same multiplet, the isospin generator contributes only through the lepton doublet: $\text{Tr}\,T_3^2=(+\tfrac12)^2+(-\tfrac12)^2=\tfrac12$ . The charge-squared trace is $\text{Tr}\,Q^2=3(\tfrac13)^2+0+(-1)^2=\tfrac13+1=\tfrac43$ . Why: equal normalization of all generators inside the simple group means $g^2\,\text{Tr}\,T_3^2=g'^2\,\text{Tr}\,(Y/2)^2$ style relations hold, and the ratio of these traces is precisely what sets the mixing angle.

Step 4. Assemble the angle. At unification $\sin^2\theta_W=\dfrac{g'^2}{g^2+g'^2}=\dfrac{\text{Tr}\,T_3^2}{\text{Tr}\,Q^2}=\dfrac{1/2}{4/3}=\dfrac38$ . Equivalently, using $g'^2=\tfrac35 g^2$ at $M_{\text{GUT}}$ from equation (2): $\sin^2\theta_W=\dfrac{3/5}{1+3/5}=\dfrac38$ . Why: the two routes agree because the factor $\tfrac53$ and the trace ratio are the same statement of equal generator normalization; $\tfrac38$ is the unavoidable prediction, and its agreement with the run-down measured value is the second pillar of the SU(5) case.

Go deeper · Derivation C · 1 minThe proton lifetime from dimensional analysis

Derivation CThe proton lifetime from dimensional analysis

The goal is to build, from nothing but the leptoquark mass and the unified coupling, the order-of-magnitude lifetime of the proton, and to see why that lifetime is astronomically long yet still within experimental reach. The logic is identical to Fermi's treatment of the weak interaction, with the $X,Y$ bosons of mass $M_X$ playing the role the $W$ plays in beta decay — only vastly heavier.

Step 1. Write the effective interaction. Exchanging a heavy $X$ boson between two quarks and a quark–lepton pair generates a four-fermion contact operator of strength $G_X\sim g^2/M_X^2\sim 4\pi\alpha_{\text{GUT}}/M_X^2$ — a baryon-number-violating analogue of the Fermi constant. Why: below $M_X$ the heavy boson cannot propagate, so its exchange shrinks to a point interaction whose strength carries two inverse powers of the huge mass — the origin of the suppression.

Step 2. Restore dimensions. A decay rate has units of energy; the amplitude carries $G_X\sim M_X^{-2}$ , so the rate must be $\Gamma\sim G_X^2\times(\text{energy})^5$ , and the only energy in the problem is the proton mass. Hence $\Gamma_p\sim\alpha_{\text{GUT}}^2\,m_p^5/M_X^4$ . Why: the fifth power of $m_p$ is the same phase-space counting that makes the muon and beta-decay rates scale as the fifth power of the released energy — dimensional analysis, not dynamics, fixes it.

Step 3. Put in numbers. With $\alpha_{\text{GUT}}\approx\tfrac{1}{40}$ , $M_X\approx10^{15}$ GeV, and $m_p=0.938$ GeV: $\Gamma_p\sim(1/40)^2(0.938)^5/(10^{15})^4\approx4.5\times10^{-64}$ GeV, so $\tau_p=\hbar/\Gamma_p\approx1.5\times10^{39}$ s $\approx5\times10^{31}$ years. Why: a lifetime $10^{22}$ times the age of the universe sounds untestable, but with $10^{33}$ protons in a tank of water you expect tens of decays per year — which is exactly why kiloton detectors were built.

Step 4. Confront experiment. Super-Kamiokande sees no $p\to e^+\pi^0$ and sets $\tau_p>2.4\times10^{34}$ years — some 500–1000 times longer than minimal $SU(5)$ predicts. Why: this gap kills the simplest model outright; supersymmetric GUTs, with $M_X\approx2\times10^{16}$ GeV, push the lifetime up by $(20)^4\sim10^5$ and shift the dominant mode to $p\to K^+\bar\nu$ , landing right at the current frontier.

Limiting cases

Case	Result	Meaning
$\mu\to M_{\text{GUT}}$	$\alpha_1=\alpha_2=\alpha_3=\alpha_{\text{GUT}}$	one coupling; the product group is restored to a single simple group
$\mu\ll M_{\text{GUT}}$	three separate couplings	the symmetry is broken; we see only the fragments $SU(3)\times SU(2)\times U(1)$
Plain Standard Model content	crossings spread over $10^{13}$ – $10^{17}$ GeV	couplings nearly meet but miss — unification hinted, not achieved
MSSM content	single meeting at $2\times10^{16}$ GeV, $\alpha_{\text{GUT}}^{-1}\approx24$	superpartners bend the slopes into an exact intersection
$M_X\to\infty$	$\tau_p\to\infty$	proton becomes stable; baryon number effectively exact at low energy
$SU(5)\to SO(10)$	$\bar{\mathbf 5}\oplus\mathbf{10}\oplus\mathbf 1\to\mathbf{16}$	one generation in a single spinor, with a built-in $\nu_R$ for the seesaw

Common trapThe Standard-Model couplings meet exactly at a single point

They do not — and pretending they do oversells the case. Run the three measured couplings up with the plain Standard-Model field content and the pairwise crossings land at $10^{13.0}$ , $10^{14.4}$ , and $10^{17.0}$ GeV: a triangle nearly four orders of magnitude wide, not a point. The couplings come tantalizingly close, which is the real hint, but minimal non-supersymmetric $SU(5)$ fails this test at the several-percent level — and independently fails the proton-decay test by a factor of hundreds. What does meet at a single point is the MSSM: adding superpartners at the TeV scale changes the slopes just enough to sharpen the near-miss into a genuine intersection at $2\times10^{16}$ GeV. So the honest statement is precise: the couplings unify only with new physics between $M_Z$ and $M_{\text{GUT}}$ , and their approach is evidence for such physics, not proof that bare $SU(5)$ is correct. Treating the near-meeting as an exact meeting hides the very clue — supersymmetry, or some other threshold — that the graph is actually pointing at.

Worked examples

1 · Where do the couplings meet, and how big is the miss? (fully worked)

Given $\alpha_1^{-1}(M_Z)=59.0$ , $\alpha_2^{-1}(M_Z)=29.6$ , $\alpha_3^{-1}(M_Z)=8.5$ and one-loop slopes per decade $-1.50$ , $+1.16$ , $+2.57$ , locate the $\alpha_1=\alpha_2$ crossing and quantify how badly the third curve misses it.

Step 1. Set up the lines in $x=\log_{10}(\mu/\text{GeV})$ , anchored at $x_0=\log_{10}M_Z=1.96$ : $\alpha_1^{-1}=59.0-1.50(x-1.96)$ , $\alpha_2^{-1}=29.6+1.16(x-1.96)$ , $\alpha_3^{-1}=8.5+2.57(x-1.96)$ .

Step 2. Cross $\alpha_1$ and $\alpha_2$ : $59.0-1.50\,\delta=29.6+1.16\,\delta$ with $\delta=x-1.96$ , so $29.4=2.66\,\delta$ , giving $\delta=11.05$ and $x=\mathbf{13.0}$ — i.e. $\mu\approx\mathbf{10^{13}\ \text{GeV}}$ . The common value is $\alpha^{-1}=29.6+1.16(11.05)=\mathbf{42.4}$ .

Step 3. Evaluate $\alpha_3^{-1}$ at that same scale: $\alpha_3^{-1}=8.5+2.57(11.05)=8.5+28.4=\mathbf{36.9}$ . The strong curve sits at $36.9$ where the other two sit at $42.4$ — it has already overshot the crossing, because with its steep slope $\alpha_3$ meets $\alpha_2$ only much later, at $x\approx17$ .

Step 4. Quantify the miss. The three pairwise crossings are at $x=13.0$ , $14.4$ , $17.0$ : a spread of four decades in energy, and the couplings differ by $\Delta\alpha^{-1}\approx5$ at any single scale in the knot. In the plain Standard Model the couplings do not unify — they miss by far more than the experimental error. Repeating the exercise with MSSM slopes closes the triangle to a point at $x\approx16.3$ ( $2\times10^{16}$ GeV): the single most-cited quantitative argument for weak-scale supersymmetry.

2 · The proton lifetime in minimal SU(5), against Super-Kamiokande

Estimate $\tau_p$ for $M_X=10^{15}$ GeV and $\alpha_{\text{GUT}}=1/40$ , and compare with the experimental bound $\tau(p\to e^+\pi^0)>2.4\times10^{34}$ years.

Step 1. Rate from equation (5): $\Gamma_p\sim\alpha_{\text{GUT}}^2\,m_p^5/M_X^4$ . With $\alpha_{\text{GUT}}^2=(1/40)^2=6.25\times10^{-4}$ , $m_p^5=(0.938)^5=0.727\ \text{GeV}^5$ , and $M_X^4=(10^{15})^4=10^{60}\ \text{GeV}^4$ : $\Gamma_p\sim6.25\times10^{-4}\times0.727/10^{60}=\mathbf{4.5\times10^{-64}\ \text{GeV}}$ .

Step 2. Convert to a lifetime using $\hbar=6.58\times10^{-25}$ GeV·s: $\tau_p=\hbar/\Gamma_p=6.58\times10^{-25}/4.5\times10^{-64}=\mathbf{1.5\times10^{39}\ \text{s}}$ .

Step 3. In years: $\tau_p=1.5\times10^{39}/(3.16\times10^7)\approx\mathbf{5\times10^{31}\ \text{years}}$ — about $10^{22}$ times the age of the universe, yet the proton is not truly "stable".

Step 4. Compare. The Super-Kamiokande limit $\tau>2.4\times10^{34}$ years is roughly 500 times longer than this estimate, so minimal non-supersymmetric SU(5) is experimentally excluded. Because $\tau_p\propto M_X^4$ , restoring unification with supersymmetric $M_X\approx2\times10^{16}$ GeV multiplies the lifetime by $(20)^4\approx1.6\times10^5$ , to $\sim10^{36}$ – $10^{37}$ years for this mode — safely above the bound, while the SUSY-dominant channel $p\to K^+\bar\nu$ lands near $10^{34}$ years, exactly where the next generation of detectors will look.

Physics of the portrait

Grand unification has no single discoverer; it is a body of ideas, and its defining artefact is a model — the $SU(5)$ theory that Georgi and Glashow wrote down in a single 1974 paper, the simplest way to sew the three forces into one.

Origins · 2 minthe Georgi–Glashow SU(5) model and how this was discovered

the Georgi–Glashow SU(5) model

1974, Howard Georgi and Sheldon Glashow's "Unity of All Elementary-Particle Forces" · grand unification of the strong, weak, and electromagnetic interactions

In February 1974 Howard Georgi and Sheldon Glashow published a two-page paper proposing that the entire gauge structure of particle physics is $SU(5)$ , broken at a high scale to the Standard Model. Georgi has recounted that he worked out the fermion assignment — one generation into $\bar{\mathbf 5}\oplus\mathbf{10}$ — in an evening, and was at first so dismayed by the prediction that the proton must decay that he went to bed convinced the idea was wrong. It was not. The model achieved three things at a stroke that no theory before it could: it explained why electric charge is quantized (the down quark's $-\tfrac13$ is forced by the tracelessness of charge in a simple group), it predicted the weak mixing angle as $\tfrac38$ run down to low energy, and it unified the couplings — a program made quantitative the same year by Georgi, Helen Quinn, and Steven Weinberg, who showed the couplings converge near $10^{15}$ GeV and that this scale is fixed by their measured low-energy values. The price was the proton's mortality, through new $X,Y$ leptoquark bosons that turn quarks into leptons. That prediction, $\tau_p\sim10^{30}$ – $10^{31}$ years for minimal $SU(5)$ , drove the construction of the great underground detectors of the 1980s — IMB, Kamiokande — and when they saw no proton decay, minimal $SU(5)$ was falsified. But the framework survived and deepened: $SO(10)$ swept a whole generation plus a right-handed neutrino into one spinor, supersymmetric GUTs restored exact coupling unification, and the seesaw mechanism tied neutrino masses to the GUT scale. Georgi and Glashow's model is wrong in its minimal form and foundational in its ideas — the template for every attempt to find the one symmetry behind the three.

History

1974

— Georgi and Glashow propose SU(5), the first grand unified theory

1974

— Georgi, Quinn, and Weinberg show the couplings run together near 10¹⁵ GeV and predict sin²θ_W

1975

— Fritzsch, Minkowski, and Georgi introduce SO(10), fitting a generation plus a right-handed neutrino in a 16

1981

— Dimopoulos, Georgi, and Sakai construct supersymmetric SU(5), restoring exact unification

1983

— the IMB and Kamiokande detectors begin dedicated searches for proton decay

2020

— Super-Kamiokande sets τ(p → e⁺π⁰) greater than 2.4×10³⁴ years, excluding minimal SU(5)

Going further

Texts. Peskin & Schroeder, An Introduction to Quantum Field Theory, §20.3 (grand unification) and the running-coupling analysis of Ch. 12 · Griffiths, Introduction to Elementary Particles, 2nd ed., §12.2 (grand unified theories) · Halzen & Martin, Quarks and Leptons, §15.5 (the SU(5) model and proton decay) · Thomson, Modern Particle Physics, §17.5 (unification and running couplings) · Schwartz, Quantum Field Theory and the Standard Model, Ch. 29–30 · Langacker, "Grand Unified Theories and Proton Decay," Phys. Rep. 72 (1981). The key anchor numbers used throughout this article are: the GUT-normalized inverse couplings $\alpha_i^{-1}(M_Z)=(59.0,29.6,8.5)$ ; SM pairwise crossings at $10^{13.0}$ , $10^{14.4}$ , $10^{17.0}$ GeV and the MSSM single meeting at $2\times10^{16}$ GeV with $\alpha_{\text{GUT}}^{-1}\approx24$ ; $Q_d=-\tfrac13$ and $\sin^2\theta_W=3/8$ from the $\bar{\mathbf 5}$ traces; and $\tau_p\approx5\times10^{31}$ yr for $M_X=10^{15}$ GeV, below the Super-K bound $2.4\times10^{34}$ yr.

Prerequisites. The Standard Model · The Local Gauge Principle · The Electroweak Gauge Group $SU(2)\times U(1)$ · Colour and QCD $SU(3)$ . Leads to. Proton Decay · Supersymmetry · The Hierarchy Problem · Open Questions of the Standard Model.

Ask the tutor: Ask me to run the three couplings together line by line and read off the unification scale, or to work out why the down quark's charge is forced to be exactly one-third of the electron's.

Does grand unification mean there is really only one force?

At energies above $M_{\text{GUT}}\sim10^{16}$ GeV, yes — in a grand unified theory the strong, weak, and electromagnetic interactions are a single force with one coupling, mediated by the gauge bosons of one simple group. What we call three forces are the low-energy fragments left after that symmetry is spontaneously broken, in the same way electromagnetism and the weak force are two faces of the broken electroweak symmetry. But note the qualifier: this unifies the three gauge interactions only. Gravity is not included — that is the further ambition of "theories of everything" like string theory — and the Higgs and Yukawa sectors that give particles mass are not tidied up by gauge unification either. So "one force" is true, but only for three of the four, and only above an energy no experiment can reach.

If the proton can decay, why is ordinary matter stable?

Because the predicted lifetime is fantastically long — at least $2.4\times10^{34}$ years, more than $10^{24}$ times the age of the universe. "Unstable" here does not mean fragile; it means that baryon number is not an exact symmetry, so a proton has a vanishingly small probability per unit time of decaying. In a single human body, with roughly $10^{28}$ protons, you would expect fewer than one decay in a lifetime. Stability is a matter of degree: the proton is the longest-lived unstable particle we know of, and the only reason we can hope to see its decay at all is that a large enough tank of water contains so many protons ( $\sim10^{33}$ in a kiloton) that even a $10^{34}$ -year lifetime yields a handful of events per year.

What exactly are the X and Y bosons, and why haven't we seen them?

They are the twelve extra gauge bosons of $SU(5)$ — beyond the eight gluons, three weak bosons, and one hypercharge boson of the Standard Model — that fill out the 24-dimensional adjoint representation. Unlike any known force carrier, they are "leptoquarks": they carry both colour and electric charge ( $\pm\tfrac43$ for $X$ , $\pm\tfrac13$ for $Y$ ) and couple a quark directly to a lepton, or two quarks to each other, violating baryon and lepton number. We have not seen them because their mass is $\sim M_{\text{GUT}}\approx10^{16}$ GeV — roughly a trillion times the reach of the LHC. Their only observable footprint is the rare, slow process they mediate at low energy: proton decay.

Why does supersymmetry make the unification so much better?

Because superpartners change the slopes of the running couplings. Each new particle that carries a given charge contributes to that coupling's beta coefficient, so doubling the spectrum with superpartners at around the TeV scale shifts all three slopes — and it happens to shift them by exactly the amounts needed to turn the Standard Model's four-decade-wide near-miss into a single clean intersection at $2\times10^{16}$ GeV. This is not something supersymmetry was designed to do; it falls out of the same superpartners introduced to solve the hierarchy problem. That two independent motivations point to the same new physics at the same scale is why supersymmetric grand unification was, for decades, the leading paradigm — though the LHC's failure so far to find superpartners has put that picture under real pressure.

How is SO(10) an improvement on SU(5)?

$SU(5)$ needs two representations to hold one generation ( $\bar{\mathbf 5}\oplus\mathbf{10}$ ) and has no place for a right-handed neutrino. $SO(10)$ fixes both: its 16-dimensional spinor representation holds all fifteen Standard-Model fermions of a generation plus a sixteenth state — a right-handed neutrino — in one irreducible multiplet. That is aesthetically cleaner (one generation, one representation) and physically consequential: the right-handed neutrino is exactly what the seesaw mechanism needs to explain why the known neutrinos are so light. Because $SO(10)$ contains $SU(5)$ as a subgroup, it inherits all of $SU(5)$ 's successes — charge quantization, $\sin^2\theta_W=\tfrac38$ , coupling unification — while curing its two blemishes, which is why it remains the most popular grand-unified group today.

Is grand unification an established part of physics or a hypothesis?

It is a well-motivated hypothesis, not an established fact. The evidence for it is circumstantial but genuine: the three couplings really do nearly meet, electric charge really is quantized in the pattern $SU(5)$ predicts, and $\sin^2\theta_W$ comes out close to right. But no direct prediction has been confirmed — proton decay has never been observed, the $X,Y$ bosons are far out of reach, and the minimal model is already excluded. So grand unification stands where electroweak unification stood before 1973: a compelling theoretical structure with strong indirect support, awaiting a decisive experimental signature. That signature would be a single proton-decay event, and the detectors now being built are designed to find it.