Discovering an Invisibly Decaying Higgs at Hadron Colliders

Worth your time, but read Table II with the abstract in the other hand. The new content is real: the WBF invisible-Higgs analysis at the Tevatron (Sec. III, Tables VI–VII) has not been done before, the combination of WBF with Z+h_inv giving 3σ at 7 fb⁻¹ per detector is a concrete and useful target, and the ratio σ(Zh)/σ(WBF) as a model-independent mass handle (Table V) is a genuinely nice idea — custodial symmetry kills the coupling dependence, so the ratio depends on m_h alone across a wide class of models. The h_inv+j null result is a clean negative finding, correctly attributed to the Z→νν+j background being three to four orders of magnitude larger. The cuts are all written as equations, the perturbative backgrounds come from Madgraph, and the tables are checkable against the text. That is more than most papers in this genre give you. The authors are also candid about what they did not simulate, which I credit them for. Now the soft spot, and it is the load-bearing one. Table II entries are S/√B, pure statistics. The paper itself derives a 7.1% background-normalization uncertainty at 10 fb⁻¹ (Sec. II.2) and uses it in Table III, where it becomes a 33% uncertainty on the signal at m_h=120 GeV. Apply that same 7.1% to the discovery significance and the 75 GeV entry goes from 5.3σ to roughly 2.6σ; the 65 GeV entry from 4.9σ to about 2.1σ. The paper never makes this combination — systematics appear in the mass-extraction tables but not in the discovery claim. Note also that even on its own statistical-only terms, the 65 GeV entry with the Z+jets estimate is 4.9σ, already below the abstract's 'above 5σ'. The fix is easy: the 7.1% is statistics on the Z→ℓℓ control sample, so it falls to 2.2% at 100 fb⁻¹, where the significance recovers to ~8σ. The claim should be stated at 100 fb⁻¹, or stated as statistics-only at 10. The imported Z+jets and CDF fake-MET backgrounds are acknowledged and argued conservative; those are minor by comparison. This is for people building invisible-Higgs search programs at ATLAS/CMS and for anyone writing the Tevatron endgame projections. Send it out; ask for the significance table redone with the normalization uncertainty included.

The paper's headline claim (abstract:s2) rests on Table II, whose entries are S/√B — a purely statistical significance. The paper itself estimates the background normalization uncertainty for Z+h_inv at 7.1% for 10 fb⁻¹ (Sec. II.2, paragraph beginning 'We estimate the total background normalization uncertainty'), and uses that number in Table III, where it produces a 33% uncertainty on the signal cross section at mh=120 GeV. That 33% is simply (B/S)×7.1% with S/B≈0.22 at the optimal pT/ cut. If the same 7.1% is applied to the discovery significance — S/√(B+(0.071B)²) — the 75 GeV entry falls from 5.3σ to about 2.6σ, and the 65 GeV entry from 4.9σ to about 2.1σ. The paper never performs this combination: systematics are included for mass extraction (Table III) but omitted from the discovery claim (Table II). This is not the reader's concern. The imported Z+jets background is acknowledged, argued to be conservative, and at the optimal 75 GeV cut only moves 5.7σ to 5.3σ; even doubling it leaves the entry near 4.9σ. The statistical-only significance is unacknowledged and far more load-bearing for the abstract's 'above 5σ' sentence.

The paper is a competent projection study, and its central claim is the Z+h_inv discovery reach at the LHC (abstract:s2). For that claim to hold, the signal significance must survive both the irreducible backgrounds and the uncertainty on their normalization. The reader correctly identifies the imported Z+jets fake-pT/ background as a soft spot, and the authors do acknowledge it (Sec. II.1, 'quite likely a conservative one'; 'beyond the scope of our analysis'). But at the optimal 75 GeV cut this background is only 9 fb out of 59 fb total, so even a factor-2 increase leaves the significance near 4.9σ; the reader's concern is real but not the most load-bearing one. The more load-bearing gap is that Table II quotes S/√B only, while the paper's own Sec. II.2 estimate of the background normalization uncertainty (7.1% at 10 fb⁻¹) implies a 32% systematic on the signal yield — larger than the 8.8% statistical error. Combining them gives ~2.6σ at 75 GeV, well below the abstract's 'above 5σ'. This is a correctness/overstatement risk, not a soundness failure: the physics message that an invisible Higgs is discoverable at the LHC survives via the WBF channel (Ref. [13]) and the mass-extraction analysis, and the Tevatron combination arithmetic (1.9σ ⊕ 1.6σ over two detectors → 7 fb⁻¹ for 3σ) checks out. The verdict therefore remains CONDITIONAL, but for the statistical-only significance rather than the Z+jets import. A concrete recomputation of Table II with the paper's own 7.1% would settle whether the headline sentence needs restating.

The paper is a careful, well-documented simulation study of a well-motivated scenario, and several of its elements are genuinely useful: the cuts are stated as equations so they can be reproduced or reweighted; the signal and the main perturbative backgrounds come from a public generator; the rescaling formulas of Eqs. (13) and (14) make the results reusable for non-SM production rates and partial invisible branching; and the negative result on h_inv+j, while less quotable, is a real service to the community because it closes off the largest-rate channel with an explicit background estimate. The observation that the missing transverse energy in ℓ+ℓ−+MET is controlled by the well-measured dilepton system, and therefore carries smaller systematic uncertainty than the forward-jet-based WBF channel, is a sound and useful point. The mass-extraction discussion, particularly the ratio method of Table V, is a legitimate and falsifiable proposal. If the headline significance survives a systematic-uncertainty accounting, the result is of real phenomenological interest for the pre-LHC and early-LHC program.

The paper asks whether a Higgs boson lighter than twice the W mass that decays almost entirely into particles the detector cannot see could still be discovered at a hadron collider, and argues that it can. The strategy is to produce the Higgs together with a Z boson that decays into an electron or muon pair: the pair is reconstructed, the Higgs leaves only missing transverse energy, and the two together form a signature whose missing energy is fixed mostly by the well-measured lepton momenta rather than by jet energy resolution. Using parton-level simulation with a dilepton mass window around the Z, a cut on the azimuthal separation of the two leptons, and a missing-energy cut of 65 to 100 GeV, the authors find the LHC reaches $5$ standard deviations at $m_h = 120$ GeV with 10 fb$^{-1}$ of integrated luminosity, and extends to 160 GeV with 30 fb$^{-1}$. They further show that taking the ratio of the Z-plus-invisible-Higgs rate to the weak-boson-fusion rate removes the production couplings, so the Higgs mass can be extracted in a comparatively model-independent way. At the Tevatron neither channel reaches $3$ standard deviations on its own at plausible luminosities, but combining weak-boson fusion with the previously studied Z channel and both detectors gets there with 7 fb$^{-1}$ per detector.

What carries the argument. Two devices carry the argument. The first is the selection itself: a Z boson reconstructed in $ell^+ell^-$ gives a trigger and a mass constraint, so that the invisible Higgs appears as missing transverse energy whose distribution is largely determined by the measured lepton momenta, and whose shape stiffens with increasing Higgs mass. The second is the ratio method. Writing $r = sigma(Z + h_{rm inv}) / sigma({rm WBF} to h_{rm inv})$, the production couplings cancel between numerator and denominator under custodial symmetry, and the logarithmic slope $(dr/dm_h)/r$ of about $-0.011$ to $-0.013$ per GeV converts a measured rate ratio into a mass, with the luminosity uncertainty dropping out entirely. The paper also defines the analogous single-channel slope $rho = (dsigma_S/dm_h)/sigma_S$, which is roughly five times larger for Z-plus-invisible-Higgs than for weak-boson fusion, so that the Z channel is the more mass-sensitive of the two even though it has fewer events.

Core claim. The central claim is that the process $pp to Z( to ell^+ell^-) + h_{rm inv}$ is a discovery channel for an invisibly decaying Higgs at the LHC, not merely a marginal one. Requiring the dilepton invariant mass to sit within 10 GeV of $m_Z$, requiring the two leptons not to be back-to-back ($Delta phi_{ellell} < 2.5$), and cutting on missing transverse energy above 65 to 100 GeV suppresses the $ZZ$, $WW$ and $ZW$ backgrounds enough that a 120 GeV Higgs decaying entirely invisibly reaches above $5sigma$ with 10 fb$^{-1}$, combining electron and muon pairs, and remains above $5sigma$ up to 160 GeV with 30 fb$^{-1}$. The paper also makes a second, distinct claim: because the $hZZ$ and $hWW$ couplings are tied by custodial $SU(2)$ in any model built only from Higgs doublets and singlets, the ratio of the Z-plus-invisible-Higgs and weak-boson-fusion cross sections depends on the Higgs mass alone, giving $Delta m_h$ of 35 to 50 GeV with 10 fb$^{-1}$ and 15 to 20 GeV with 100 fb$^{-1}$. Finally, it reports a negative result: the gluon-fusion channel $h_{rm inv} + j$ is drowned by $Z to nu barnu$ plus jets, roughly $10^3$ to $10^4$ times the signal, with no kinematic handle to separate them.

The premise it rests on. The Z-plus-jets background that fakes missing energy is never simulated here; it is imported from an earlier study and rescaled by hand, by about 1.6 for the wider lepton acceptance and by about 2.5 in going from a 65 GeV to a 75 GeV missing-energy cut, and the authors acknowledge that this estimate is uncertain and that simulating it is beyond their scope, while calling the result conservative.

A Higgs between the LEP limit of 114 GeV and roughly 160 GeV that decays invisibly is discoverable at the LHC in a channel with clean leptonic systematics, independent of weak-boson-fusion jet tagging.

The mass of an invisibly decaying Higgs can be bounded without knowing its production coupling strength, by comparing the Z-associated and weak-boson-fusion rates in models with only doublets and singlets.

Observing the Higgs in weak-boson fusion but not in Z-plus-invisible-Higgs would set a lower limit on the Higgs mass in that same class of models.

At the Tevatron, combining weak-boson fusion with the Z channel and both detectors brings a $3sigma$ observation of a 120 GeV invisible Higgs within reach at 7 fb$^{-1}$ per detector, provided the background normalisation is controlled to better than about 10 percent.

Gluon-fusion production with one jet should be abandoned as an invisible-Higgs search channel at both machines, despite having the largest rate.

Because the missing-energy spectrum in the Z channel is set by the dilepton system, the same data set that measures the signal also calibrates the dominant $ZZ$ background in the $ell^+ell^-$ + invisible control region, which is a stronger handle than the paper's error budget makes explicit.

The ratio method's model independence rests on custodial symmetry, so a measured mismatch between the mass extracted from the Z channel and from the ratio would itself be a signal of Higgs-sector structure beyond doublets and singlets, such as a triplet or a mixing with a heavier state.

The paper's negative conclusion on $h_{rm inv} + j$ is a statement about a mono-jet trigger at hadron colliders generally: any search relying on a single jet plus missing energy inherits the same irreducible $Z to nu barnu$ plus jet floor unless a photon or an initial-state tag is added.

The quoted Tevatron combination at 7 fb$^{-1}$ per detector is the kind of target that a later reanalysis with a central-jet veto could reach with roughly half the luminosity, since the paper estimates a factor of three improvement in signal-to-background from that veto alone.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review