89
Part Two · Quantum fields and particles

The Higgs discovery and what it measured

On 4 July 2012, in a packed CERN auditorium with Peter Higgs and François Englert in the audience, the spokespeople of CMS and ATLAS each showed a small bump: an excess of photon pairs, and of four-lepton events, near 125 GeV. Each experiment alone had five standard deviations. Higgs wiped his eyes. Forty-eight years after the mechanism of Chapter 77 was proposed, its particle had been found. This chapter rebuilds the discovery from the data and shows what has been measured since.

On the plate: the photon-pair mass spectrum. Add data and watch the bump rise out of the background.

How a Higgs is made

The Higgs couples to mass, but the proton’s light quarks hardly couple to it at all. Most Higgs bosons at the LHC are made by gluon fusion, about 87%: two gluons join through a loop of top quarks, the heaviest particle, which couples most strongly. Smaller shares come from two quarks radiating W or Z bosons that fuse (vector-boson fusion, about 7%), from a W or Z radiating a Higgs (about 4%), and from production with a top-quark pair (about 1%). At 13.6 TeV the total cross section is about 60 pb, roughly one Higgs per second at full luminosity (Chapter 65). Established

How a Higgs decays

A 125 GeV Higgs decays mostly to bottom quarks (58%), then to W pairs (21%, one of them off-shell), gluons, taus, charm and Z pairs. Photon pairs, through loops, are only 0.23%. Yet the two “golden” channels in 2012 were H→γγH\to\gamma\gamma and H→ZZ→4ℓH\to ZZ\to4\ell: rare, but measured with excellent mass resolution and modest background, while bbˉb\bar b is buried under ordinary QCD jets. The plate shows the branching fractions.

The bump

Accumulate data: 10 fb⁻¹. The background falls smoothly with mass; the signal is a narrow peak, its width set by the detector’s photon energy resolution. Significance grows as the square root of the data. By mid-2012, with about 10 fb⁻¹ per experiment, the combination of channels passed 5σ in each. The full Run 2 data, 140 fb⁻¹, show the peak unmistakably. The mass is now known to be 125.20±0.11125.20 \pm 0.11 GeV. Established

Couplings line up with mass

In the Standard Model the Higgs couples to each fermion in proportion to its mass, and to W and Z in proportion to their masses squared. Plot the measured coupling strength, suitably scaled, against particle mass, and the points should lie on one straight line. They do, from the muon at 0.1 GeV to the top at 173 GeV, three orders of magnitude, each to within 5–20%. Deviations are expressed as coupling modifiers κ\kappa, all consistent with 1. Couplings to the light quarks and the electron remain unmeasured.

The shape of the potential

The Mexican hat of Chapter 76 has a definite shape fixed by mhm_h and vv: the self-coupling λ=mh2/2v2≈0.13\lambda = m_h^2/2v^2 \approx 0.13. It is measured by making two Higgs bosons at once, a process about a thousand times rarer than single production. ATLAS’s 2025 analysis of the full Run 2 and early Run 3 data in the bbˉγγb\bar b\gamma\gamma channel constrains the self-coupling modifier to −1.6<κλ<6.6-1.6 < \kappa_\lambda < 6.6 at 95% confidence. Observing it, and checking that the vacuum is the shape the Standard Model says, is a central goal of the High-Luminosity LHC. Established, as a bound only.

What you now have

Higgs production modes, branching fractions and the golden channels, how the 2012 discovery emerged from a growing bump, the mass, the coupling–mass line and κ modifiers, and the self-coupling still to be observed.

· · ·

Three questions to carry with you.

Why is gluon fusion the main production mode?tap
Gluons are plentiful in protons, and the loop runs through the top quark, which couples strongly to the Higgs.
Why was H → γγ a discovery channel despite 0.23% branching?tap
Excellent mass resolution and a smooth, manageable background.
What would measuring the self-coupling test?tap
The shape of the Higgs potential, and so the nature of the electroweak vacuum.

And one to work out. With σ=55\sigma = 55 pb =55,000= 55{,}000 fb, BR(γγ)=0.00227\text{BR}(\gamma\gamma) = 0.00227 and 140 fb⁻¹, how many H→γγH\to\gamma\gamma decays were produced?

N =

Sources: ATLAS Collaboration, Phys. Lett. B 716, 1 (2012); CMS Collaboration, Phys. Lett. B 716, 30 (2012). ATLAS, Nature 607, 52 (2022) and CMS, Nature 607, 60 (2022) (ten-year coupling summaries). LHC Higgs Working Group branching ratios. Particle Data Group (m_H = 125.20 ± 0.11 GeV). ATLAS Collaboration, ATLAS-CONF-2025-005 (HH → bbγγ, 308 fb⁻¹, Run 2 + early Run 3). Checked by code: branching fractions sum to 1.00; produced H → γγ in 140 fb⁻¹ ≈ 17,500.

Next · Chapter 90
Beyond the Standard Model: SUSY, extra dimensions, composite Higgs

Why is the Higgs so light? Every answer predicted new particles at the LHC. None have appeared yet.

Turn the page