HH · two at once

When does a Higgs pair show up?

Two Higgs bosons out of one collision. Nobody has seen it yet. Here is how far away it is.

Today: 1.1σ

1.1σobserved significance (1.3σ expected)

< 2.5×HH rate limit vs the Standard Model, 95% CL (1.7× expected)

0.8^+0.9_−0.7measured HH signal strength

−0.71 … 6.1κ_λ (the Higgs self-coupling, 1 in the Standard Model) allowed, 95% CL (−1.3 … 6.7 expected)

ATLAS and CMS together, full Run 2 at 13 TeV, 126–140 fb⁻¹ of data. Discovery means 5σ, so this is a whisper. Run 3 isn’t in these numbers yet.

Slide the luminosity

140 fb⁻¹per experiment

rough √L line (not an official curve) published HL-LHC, S2 published, S3 Run 2, expected

1.6σon the rough √L line

1.3σnearest published number

The dashed line is just significance ∝ √L, pinned to the published 7.2σ at 3000 fb⁻¹. It is a guide between the published points, not an official curve. It starts above today’s 1.3σ because the projection assumes 14 TeV and the newer analysis techniques it describes.

What the projection says

each	ATLAS	CMS	both	κ_λ 68% CL, both
2 ab⁻¹, S2	3.7σ	3.5σ	6.0σ	−32% +37%
3 ab⁻¹, S2	4.3σ	4.2σ	7.2σ	−27% +31%
3 ab⁻¹, S3	4.5σ	4.5σ	7.6σ	−26% +29%

Expected, 14 TeV, Standard Model pairs assumed; κ_λ intervals are 68% CL. S2: reduced systematic uncertainties, theory ones halved where feasible. S3: S2 plus recent reconstruction gains. The previous European Strategy projection was 4σ. One experiment alone: unlikely at 2 ab⁻¹, possibly in reach at 3 ab⁻¹.

The full 3 ab⁻¹ is also a ruler for the hat: κ_λ to better than 30% (−26% / +29%, S3). Drag it in the Lab → Back to the hat →

The calendar

2015–2018 · Run 2≈140 fb⁻¹ per experiment. The 1.1σ above.
2022–2026 · Run 3more than 170 fb⁻¹ per experiment expected (as of March 2025).
June 2026 · LHC offSwitched off for Long Shutdown 3.
June 2030 · HL-LHC startsRun 4 start-up planned (CERN, October 2024; still “2030” in June 2026).
by 2041 · 3,000 fb⁻¹total delivered, including about 520 fb⁻¹ recorded in Runs 1–3 (delivered, not the same as the projection’s 3 ab⁻¹ of physics-quality data). Approved to run 2030–2041.

Schedules move; each date is quoted from its source with the source’s date.

Decay the pair

Each Higgs picks its own decay, with Standard Model odds. Tap.

Higgs 1H

Higgs 2H

two Higgs, not decayed yet

your pairsStandard Model odds0 pairs

Decay fractions only. What an experiment keeps after triggers, selection and efficiency is a much smaller slice of each. bbWW and bbZZ count every W/Z decay; “other” is everything not listed (mostly bb + gluons or charm).