How we turned “nothing” into one of the tightest limits on dark matter
A PPC spotlight on Phys. Rev. D 105, 092007 (2022) — CMS, “Search for invisible decays of the Higgs boson produced via vector boson fusion in proton–proton collisions at √s = 13 TeV” (arXiv:2201.11585, analysis HIG-20-003)
Most Higgs analyses live or die by what the detector sees — two photons here, four leptons there. This one is different. It asks the opposite question: what if the Higgs boson we discovered in 2012 sometimes decays into something CMS cannot see at all?
In the Standard Model, the Higgs almost never does this — the invisible branching fraction from H → ZZ → 4ν is a minuscule 0.1%. But if dark matter is light enough, the Higgs could act as a portal to it, decaying straight into invisible dark-matter particles that sail out of the detector without a whisper. Measure how often the Higgs vanishes, and you are measuring how strongly it talks to the dark universe. That is the prize this paper chased — and it is exactly the kind of “seek what you cannot see” physics the PPC group has made its trademark, with Guillelmo Gomez-Ceballos deep in the machinery of the analysis.
The trick: tag the production, not the decay
If the decay is invisible, you have to catch the Higgs by the way it was born. The analysis targets vector boson fusion (VBF): two quarks each radiate a weak boson that fuse into a Higgs, leaving behind two forward jets flying off in opposite directions with a huge rapidity gap and a large dijet mass. That distinctive “two jets, wide apart, and a pile of missing momentum in between” signature is the fingerprint.
Using 101 fb⁻¹ of 13 TeV data from 2017–2018, the team built two complementary categories — one driven by the missing transverse momentum, one by the VBF dijet topology — and, crucially, tamed the dominant irreducible background (Z → νν produced with jets, which is genuinely invisible too) using a high-statistics photon + jets control region to pin down that “real nothing” from data rather than trusting simulation alone.
“The signal is literally nothing — missing energy between two jets. So the whole art is knowing your backgrounds better than you know your signal. You measure the invisible Z from a photon sample you can see, and you let the data tell you how much ‘nothing’ is normal.”
// Guillelmo Gomez-Ceballos, PPC / MIT
The result: a boson kept on a short leash
Combining the new data with every previous CMS VBF invisible search — 2012 (8 TeV), 2015, and 2016 — the analysis sets an observed (expected) upper limit on the Higgs invisible branching fraction of:
B(H → inv) < 0.18 (0.10 expected) at 95% confidence level, assuming the Standard Model production rate.
No sign of invisible decays — but a powerful constraint. And when recast as a Higgs-portal dark-matter model, this collider result out-reaches the great underground direct-detection experiments (XENON1T, LUX, PandaX-4T, DarkSide-50, and friends) precisely in the low-mass regime — for dark-matter masses below about 12 GeV (fermion) or 6 GeV (scalar) — where those experiments run out of steam. The LHC, it turns out, is one of the best dark-matter detectors on Earth for light dark matter.
The most significant plot
Of all the figures in the paper, the one that captures the whole enterprise is Figure 11 — the headline result:
Figure 11 (Phys. Rev. D 105, 092007): Observed and expected 95% CL upper limits on (σ_H/σ_H^SM)·B(H→inv) for each data-taking period and for their combination, assuming a 125.38 GeV Standard Model Higgs boson. The combined bar lands at 0.18 observed / 0.10 expected.
Why this one? Because it tells the entire story in a single frame: every year of data, each contributing its own limit, marching down the page and collapsing into one combined number that represents the best VBF invisible-Higgs limit CMS had ever produced. It is the plot you point to when someone asks, “So how dark is the Higgs allowed to be?” — and the answer is “less than 18% of the time, and we’re still squeezing.”
For the dark-matter aficionados, Figure 13: the spin-independent DM-nucleon cross section versus DM mass, laid over the direct-detection limits, is the crowd favorite, showing exactly where the LHC ‘beats’ the underground detectors. But Figure 11 it is the measurement; check out Figure 13 for its interpretation.
The PPC angle
This is PPC physics to the core: a rare, background-dominated signature; a data-driven strategy that turns a control region into the real analysis engine; and a result that speaks directly to the dark-sector question the group has been circling for years. It sits right alongside the Mono-X and dark-shower program –another way of casting the net for dark matter, this time using the Higgs itself as the bait.
“We spent a decade measuring everything the Higgs does that we can see. This is us measuring what it does when it thinks no one is watching.”
// Guillelmo Gomez-Ceballos, PPC / MIT
Reference: CMS Collaboration, Phys. Rev. D 105, 092007 (2022), arXiv:2201.11585; public figures at CMS HIG-20-003.