⚛️ phenomenology

Semi-visible higgs decay as a probe for new invisible particles

This paper investigates the High-Luminosity LHC's sensitivity to new invisible scalars and fermions with masses below 50 GeV via semi-visible Higgs decays within the dark-SMEFT framework, utilizing both cut-based and multivariate analysis techniques to distinguish operator structures and benchmark results against invisible Z-width constraints and perturbative unitarity.

Original authors: Sally Dawson, Arnab Roy, German Valencia

Published 2026-07-13

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sally Dawson, Arnab Roy, German Valencia

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the Large Hadron Collider (LHC) as a giant, high-speed particle slingshot, smashing protons together to see what tiny pieces fly out. Usually, scientists look for "invisible" particles—ghostly things that zip right through detectors without leaving a trace. But what if these ghosts leave a little bit of a footprint? That's the idea behind this paper: looking for "semi-visible" Higgs decays.

The authors, a team of physicists, are playing a game of cosmic detective. They are hunting for new, invisible particles (which they call "Dark Matter" or DM) that might be hiding inside the decay of a Higgs boson. Specifically, they are looking at a scenario where a Higgs boson splits into two visible things (like a pair of electrons or two jets of particles) and a pair of these new, invisible ghosts. The "missing" energy from the ghosts creates a gap in the energy balance, which is the clue the scientists are chasing.

The Main Discovery: A New Way to See the Invisible
The paper suggests that by focusing on a specific production mode—where a Higgs boson is created alongside a Z boson (another heavy particle)—scientists at the High-Luminosity LHC (HL-LHC) could spot these new particles even if they are quite light, weighing less than 50 GeV.

Think of the Higgs boson as a magician. Usually, when it disappears, it might vanish completely into a "black box" (invisible decay) or turn into standard particles. This paper proposes a trick where the magician pulls out a visible rabbit (leptons or jets) and a ghost. The ghost is invisible, but because the rabbit is there, we know the trick happened. The authors ran detailed computer simulations (using tools like MadGraph and Pythia) to see if the HL-LHC could catch this trick. They found that with enough data (3000 fb⁻¹ of collisions), they could potentially see these events with a statistical confidence of 3-sigma (which means it's a strong hint, though not yet a confirmed discovery).

What They Ruled Out and What They Argue Against
The paper is very careful about what it doesn't think will work.

No Spin-1 Dark States: The authors explicitly argue against including "spin-1" dark particles (like dark photons) in their analysis. They explain that if these particles were massless or very light, the math would break down and give nonsensical results. So, they are only looking for spin-0 (scalars) or spin-1/2 (fermions) particles.
Not All Operators Are Equal: They rule out certain types of interactions called "pure Higgs-portal" operators. Why? Because other experiments have already looked for these specific interactions and found them to be extremely rare or non-existent. The paper says, "Let's not waste time looking for ghosts that we already know aren't there."
The "Higgs-Neutrino Floor" isn't a Wall: In the past, scientists worried that a background noise called the "Higgs-neutrino floor" (where the Higgs decays into standard neutrinos) would drown out any new signals. This paper argues that for their specific setup, this background is actually "reducible." Imagine trying to hear a whisper in a noisy room; usually, the noise wins. But here, the authors show that by using a specific filter (looking at the mass of the two visible particles), they can silence the noise and hear the whisper. This means they can potentially see signals below the level that was previously thought to be the limit.

How Sure Are They?
It is important to note that these results are based on simulations, not actual measurements from the LHC yet. The authors have not "found" these particles; they have calculated that the HL-LHC should be sensitive enough to find them if they exist.

The Confidence: They are confident in their method. They used two different ways to analyze the data: a simple "cut-based" approach (like setting strict rules for what counts as a signal) and a more advanced "multivariate" approach using a Boosted Decision Tree (a type of AI that learns to spot patterns). Both methods agreed that the signal is detectable.
The Limits: They found that for some types of invisible particles (those interacting via a specific "derivative current"), the existing limits on the Z boson's invisible width are actually stronger than what the Higgs decay can see. However, for other types of particles (those interacting via "Yukawa-type" couplings), the semi-visible Higgs decay is the most sensitive tool available.
The "Unitarity" Check: They also checked their math against a rule called "perturbative unitarity," which ensures the laws of physics don't break at high energies. They found that for their proposed signals to make sense, the new physics scale must be around a few hundred GeV to 1 TeV. If the new physics were much heavier, the math would suggest the signals they are looking for shouldn't exist.

The Takeaway
In short, this paper is a roadmap for future experiments. It suggests that if new, light, invisible particles exist and interact with the Higgs boson in specific ways, the HL-LHC has a good chance of spotting them by looking for a "semi-visible" decay. It's like saying, "If there are ghosts, they might be holding hands with a rabbit. If we watch the rabbit closely, we can prove the ghost is there." The authors have simulated this scenario and shown that with the right tools and enough data, the rabbit's dance could reveal the ghost's existence, provided the ghost isn't too heavy and isn't the kind that breaks the rules of physics.

Technical Summary: Semi-visible Higgs Decay as a Probe for New Invisible Particles

Problem Statement
The paper investigates the sensitivity of the High-Luminosity Large Hadron Collider (HL-LHC) to detect new invisible particles (scalars $\phi$ or fermions $\chi$ ) with masses $m \lesssim 50$ GeV. These particles are hypothesized to be produced in semi-visible Higgs decays within the process $pp \to ZH$ , where the $Z$ boson decays to jets ($jj$) or leptons ( $\ell^+\ell^-$ ), and the Higgs boson decays to a pair of Standard Model particles plus missing transverse energy ( $E_T$ ). The study addresses the challenge of distinguishing these signals from the irreducible Standard Model (SM) background, specifically the "Higgs-neutrino floor" arising from $H \to ZZ^* \to \ell^+\ell^-\nu\bar{\nu}$ .

Methodology
The analysis is conducted within the "dark-SMEFT" (dSMEFT) framework, utilizing dimension-6 operators that respect a discrete $Z_2$ symmetry (where new particles are odd and SM particles are even). The study excludes operators involving multiple Higgs fields without other SM fields, as they are tightly constrained by invisible Higgs width measurements.

Theoretical Framework: The authors catalog relevant operators for scalar and fermion dark matter (DM) candidates. These include derivative currents involving $(H^\dagger \overleftrightarrow{D}_\mu H)$ and Yukawa-type interactions. Theoretical constraints are derived from perturbative unitarity bounds on scattering amplitudes (e.g., $e^+e^- \to \bar{\phi}\phi$ and $HZ \to \bar{\phi}\phi$ ) and limits on the invisible $Z$ -width ( $\Gamma_{inv} < 2.04$ MeV at 95% CL).
Simulation and Event Generation: Signal and background events for $pp \to ZH$ at $\sqrt{s} = 14$ TeV are generated using MadGraph5_aMC@NLO with the UFO model files derived from FeynRules. Parton showering and hadronization are simulated with Pythia8, and detector effects are modeled using Delphes3 with CMS configuration.
Analysis Strategies:
- Cut-based Analysis: Events are selected based on kinematic variables such as dilepton invariant mass ( $M_{\ell\ell} < 60$ GeV), transverse mass ( $M_T < 125$ GeV), and $E_T$ thresholds. A $b$ -jet veto is applied to suppress top-quark backgrounds.
- Multivariate Analysis (MVA): A Boosted Decision Tree (BDT) is employed using the ROOT TMVA framework. Input variables include kinematic masses, transverse momenta, and azimuthal angular separations ( $\Delta\phi$ ). The BDT is trained to distinguish signal from backgrounds (including $ZH$, $ZZ$, $t\bar{t}Z$ , etc.) and to discriminate between different operator structures.
Kinematic Discrimination: The study specifically examines whether kinematic distributions can distinguish between derivative current operators (e.g., $O_{DH\partial\phi}$ ) and Yukawa-type operators (e.g., $O_{e\phi}$ ).

Key Contributions

Reduction of the Higgs-Neutrino Floor: The paper demonstrates that for operators involving derivative currents (which induce diagrams similar to the SM $H \to ZZ^*$ ), the "Higgs-neutrino floor" is a reducible background. A cut on the dilepton invariant mass ( $M_{\ell\ell}$ ) effectively removes the SM $H \to ZZ^*$ background because the signal topology involves a partial cancellation that suppresses the $Z$ -peak in the dilepton mass distribution for these specific operators.
Operator Discrimination: The authors show that kinematic variables, particularly the azimuthal angle separation between leptons ( $\Delta\phi(\ell_1, \ell_2)$ ), provide strong discrimination between derivative operators and Yukawa-type operators.
Comprehensive Sensitivity Comparison: The study provides a direct comparison of HL-LHC sensitivity for semi-visible Higgs decays against constraints from the invisible $Z$ -width and perturbative unitarity limits.

Results

Sensitivity Reach: With an integrated luminosity of $3000 \text{ fb}^{-1}$ , the semi-visible Higgs decay analysis can probe Wilson coefficients down to $C/\Lambda^2 \sim 40 \text{ TeV}^{-2}$ for derivative operators and $\sim 150 \text{ TeV}^{-2}$ for Yukawa-type operators (at $3\sigma$ significance).
Dominant Constraints:
- For derivative operators ( $O_{DH\partial\phi}, O_{DH\chi\chi}, O_{DH\chi\chi2}$ ), the invisible $Z$ -width provides the strongest constraints for DM masses up to $\sim 45$ GeV. Semi-visible Higgs decays become competitive only for higher masses.
- For Yukawa-type operators ( $O_{e\phi}, O_{u\phi}, O_{d\phi}$ ), the semi-visible Higgs decay analysis offers the most sensitive probe, surpassing both the $Z$ -width limits and unitarity bounds in the relevant mass range.
Unitarity Consistency: Most of the parameter space probed by the semi-visible Higgs analysis is consistent with perturbative unitarity up to $\sqrt{s} \approx 0.3$ TeV. However, for quark-coupled operators ( $O_{u\phi}, O_{d\phi}$ ), the probed region is largely excluded by unitarity at $\sqrt{s} = 1$ TeV and is also strongly constrained by monojet searches.
Branching Fractions: The analysis can probe Higgs branching fractions into semi-visible channels ( $BR(H \to \text{semi-visible})$ ) that lie below the standard Higgs-neutrino floor ( $BR \approx 1.06 \times 10^{-2}$ ), specifically for leptonic operators where the background is reducible.

Significance
The paper claims that semi-visible Higgs decays offer a complementary and, in specific cases (Yukawa-type operators), superior probe for light invisible particles compared to traditional invisible width measurements. A primary significance is the demonstration that the "Higgs-neutrino floor" is not an absolute limit for these specific BSM scenarios; by exploiting the distinct kinematic topologies of dSMEFT operators, the SM background can be significantly reduced. Furthermore, the ability to kinematically distinguish between operator structures (derivative vs. Yukawa) suggests that if a signal is observed, the HL-LHC could provide insights into the underlying UV completion of the new physics. The authors note that while these sensitivities are generally weaker than direct detection limits for thermal dark matter scenarios, they remain crucial for non-thermal scenarios (e.g., asymmetric or co-annihilating dark matter) where other detection methods may be suppressed.

More like this