Probing compressed mass spectra in the type-II seesaw model at the LHC

This paper proposes that a compressed mass spectrum scenario in the type-II seesaw model, which has likely evaded previous LHC searches, can be effectively probed using multivariate analysis of same-sign dilepton events with low invariant mass on existing Run 2 and future HL-LHC data.

The Gist

Imagine the Standard Model of physics as a giant, mostly complete puzzle. We know most of the pieces, but we are missing a few crucial ones that explain why neutrinos (tiny, ghost-like particles) have mass. One popular theory to fill this gap is the Type-II Seesaw Model.

Think of this model as adding a new, special "triplet" piece to the puzzle. This piece comes in three flavors: a doubly charged particle, a singly charged particle, and a neutral one. Scientists have been hunting for these particles at the Large Hadron Collider (LHC)—the world's biggest particle smasher in Switzerland—for years.

The Problem: The "Invisible" Search

So far, the LHC has found nothing. It's like looking for a specific type of car in a parking lot, but every time you look, the car is either too fast to see or hiding behind a truck.

In physics terms, previous searches looked for these particles when they were heavy and far apart in mass. If the particles were heavy, they would decay (break apart) into high-energy, easy-to-spot particles. The LHC experiments (ATLAS and CMS) have ruled out these "heavy and obvious" scenarios.

However, the authors of this paper suggest we might be looking in the wrong place. They propose a scenario called "Compressed Mass Spectra."

The Analogy: The Tightly Packed Suitcase

Imagine the new particles are like a set of nesting dolls or a tightly packed suitcase.

• The Old Search: Looked for a suitcase where the dolls were huge and far apart. Easy to spot.
• The New Scenario: The dolls are all the same size and packed so tightly together that they barely have any room to move.

In this "compressed" scenario, when the heavy particle breaks apart, the resulting pieces (leptons) are very soft (low energy) and move slowly. They are like whispers in a crowded, noisy room.

• The Noise: The LHC is a chaotic place filled with "background noise" from standard processes (like jets and fake particles).
• The Whisper: The signal from our compressed particles is so quiet and low-energy that the standard search filters (which look for loud, energetic signals) simply ignore it. It's like trying to hear a whisper while someone is playing a drum solo; the whisper gets lost.

The Solution: Listening for the Whisper

The authors realized that if you try to filter out the noise by demanding loud signals, you accidentally filter out the whisper too. So, they decided to change the strategy.

Instead of looking for loud, high-energy explosions, they decided to look for a specific pattern of two same-sign leptons (like two electrons with the same charge) that are low energy and close together.

To do this, they used a Multivariate Analysis (MVA).

• The Metaphor: Imagine a bouncer at a club. The old bouncer only let in people wearing flashy, expensive suits (high energy). The new bouncer is a super-smart AI (the BDT classifier). This AI doesn't just look at one thing; it looks at the whole picture: the shape of the crowd, the timing, the specific way people are standing, and the subtle differences between a real VIP and a lookalike.
• The AI was trained to distinguish the "whisper" signal from the "noise" of fake particles and misidentified charges (like mistaking a positive electron for a negative one).

The Results: Finding the Hidden Gems

After running their new "AI bouncer" through the data collected by the LHC so far (Run 2) and simulating future data (High-Luminosity LHC), they found:

1. A Hidden Safe Zone: There is a huge chunk of the "parameter space" (the possible settings for these particles) that the LHC has completely ignored because the particles were too "soft."
2. We Can Find Them: With the data already collected, they could potentially find these particles if they weigh up to 260 GeV. With future data (HL-LHC), they could find them up to 360 GeV.

Why This Matters

This paper is like telling the search team, "Stop looking for the loud drums! The answer might be in the quiet whispers we've been ignoring."

It suggests that the Type-II Seesaw model isn't dead; it's just hiding in a "compressed" corner that previous searches were too aggressive to see. By relaxing their rules and using smarter data analysis, we might finally find the missing pieces of the neutrino mass puzzle right under our noses.

In short: The authors found a new way to look for hidden particles that are too quiet for old methods to hear, proving that a significant part of the universe's secrets might still be waiting to be discovered in the LHC's existing data.

Technical Summary

1. Problem Statement

The Type-II Seesaw model extends the Standard Model (SM) by introducing an $SU(2)_L$ triplet scalar field ($\Delta$) with hypercharge $Y=1$. This model naturally explains neutrino masses and predicts physical states including doubly charged ($H^{\pm\pm}$), singly charged ($H^{\pm}$), and neutral ($H^0, A^0$) scalars.

The Core Issue:

• Experimental Limits: Extensive searches at the Large Hadron Collider (LHC) have placed stringent lower limits on the mass of $H^{\pm\pm}$ (up to $\sim 1020$ GeV for decays to $\ell^\pm \ell^\pm$ and $\sim 420$ GeV for $W^\pm W^\pm$) in regions where the mass splitting ($\Delta m = m_{H^{\pm\pm}} - m_{H^{\pm}}$) is large.
• The "Blind Spot": A significant portion of the parameter space remains unconstrained. This occurs when the mass splitting is small ($\Delta m \sim \mathcal{O}(10)$ GeV) and the triplet vacuum expectation value (VEV) is small ($v_t \lesssim 10^{-4}$).
• Kinematic Challenge: In this "compressed" regime, $H^{\pm\pm}$ and $H^{\pm}$ decay via off-shell $W$ bosons and neutral scalars ($H^0/A^0$). Since $H^0/A^0$ decays invisibly to neutrinos ($\nu\nu$) in this regime, the final state consists of soft leptons and jets with low missing transverse momentum ($p_T^{miss}$).
• Background Overwhelm: Standard search strategies relying on high-$p_T$ jets (from Initial State Radiation) or large $p_T^{miss}$ drastically reduce signal acceptance. Furthermore, the low-$p_T$ signal is swamped by SM backgrounds (QCD multijets, Drell-Yan, top production) and specific instrumental effects like fake leptons and electron charge misidentification.

2. Methodology

The authors propose a specialized search strategy tailored to the compressed spectrum, utilizing a Multivariate Analysis (MVA) to distinguish signal from background.

A. Signal Model & Production

• Process: Pair production of triplet-like Higgses via Drell-Yan mechanisms ($q\bar{q} \to \gamma^*/Z^* \to H^{++}H^{--}$ and $qq' \to W^* \to H^{\pm\pm}H^{\pm}$).
• Benchmark Point (BP1): $m_{H^{\pm\pm}} = 200$ GeV, $\Delta m = 30$ GeV.
• Decay Chain: $H^{\pm\pm} \to W^{\pm*} H^{\pm} \to W^{\pm*} W^{\pm*} H^0/A^0 \to \ell^\pm \nu \ell^\pm \nu \nu\nu$.
• Final State Signature: A pair of same-sign leptons ($\ell^\pm \ell^\pm$) with low invariant mass, accompanied by soft jets and low $p_T^{miss}$.

B. Simulation & Backgrounds

• Tools: Events generated with MadGraph5_aMC@NLO, showered with PYTHIA 8.2, and detector effects simulated via Delphes 3.4.2 (CMS card).
• Background Categories:
  1. Prompt: Diboson ($WZ, ZZ, WW$), Triboson, Top-pair ($t\bar{t}$).
  2. Non-prompt/Fake: Jets misidentified as leptons. Modeled conservatively with a $p_T$-dependent probability (0.1–0.3%).
  3. Charge Misidentification: Opposite-sign leptons (mainly from Drell-Yan and $t\bar{t}$) misidentified as same-sign due to bremsstrahlung. Modeled with a $p_T$- and $\eta$-dependent probability.

C. Event Selection & MVA

• Pre-selection:
  • Two same-sign isolated leptons ($e$ or $\mu$) with $p_T > 15$ GeV.
  • Invariant mass cut: $1 \text{ GeV} < m_{\ell\ell} < 60 \text{ GeV}$ (to suppress $Z$-boson background and $J/\psi$ resonances).
  • Angular separation $\Delta R_{\ell\ell} > 0.05$ and $\Delta R_{\ell,j} > 0.4$.
• Multivariate Analysis (BDT):
  • Classifier: Boosted Decision Tree (BDT) using the TMVA toolkit (AdaBoost algorithm).
  • Input Features: $p_T(\ell_{1,2})$, $p_T^{miss}$, $m_{\ell\ell}$, $\Delta R_{\ell\ell}$, $p_T(\ell\ell)$, scalar sum of $p_T$ ($L_T$), and azimuthal separation $\Delta\phi(\ell\ell, p_T^{miss})$.
  • Optimization: The BDT response cut was set to > 0.4, where the discovery significance peaks.
  • Key Finding: The variable $\Delta R_{\ell\ell}$ provided the best separation power, while individual lepton $p_T$ was least important.

3. Key Contributions

1. Identification of a Viable Search Channel: The paper demonstrates that the "compressed" region of the Type-II Seesaw model, previously considered inaccessible due to soft kinematics, can be probed by relaxing high-$p_T$ requirements and focusing on low-invariant-mass same-sign dileptons.
2. Robust Background Handling: The analysis explicitly accounts for electron charge misidentification and fake leptons, which are the dominant backgrounds in low-$p_T$ same-sign dilepton searches.
3. MVA Optimization: The use of a BDT classifier significantly enhances the signal-to-background ratio compared to simple cut-based analyses, allowing for the retention of signal events that would otherwise be lost.

4. Results

The authors projected the sensitivity of the search using existing Run 2 data and future High-Luminosity LHC (HL-LHC) data.

• Sensitivity with Run 2 Data ($\sim 139 \text{ fb}^{-1}$):
  • The model can be probed up to $m_{H^{\pm\pm}} \approx 260$ GeV for a $5\sigma$ discovery.
  • Exclusion limits (95% CL) can reach $\approx 330$ GeV.
• Sensitivity with HL-LHC Data ($3000 \text{ fb}^{-1}$):
  • Discovery reach extends to $m_{H^{\pm\pm}} \approx 360$ GeV.
  • Exclusion reach extends to $\approx 420$ GeV.
• Significance: These results cover a parameter space ($\Delta m \sim 30$ GeV, $v_t \lesssim 10^{-4}$) that is currently unconstrained by ATLAS and CMS searches, which typically assume larger mass splittings or higher $p_T$ thresholds.

5. Significance

• Theoretical Impact: This work revitalizes the Type-II Seesaw model as a viable candidate for new physics at the electroweak scale. It shows that the lack of discovery so far does not rule out the model but rather indicates that the mass spectrum may be compressed.
• Experimental Strategy: It provides a concrete roadmap for LHC experiments to search for "stealth" signals characterized by soft objects and low missing energy, a strategy applicable not only to Type-II Seesaw but also to other BSM scenarios with compressed spectra (e.g., Supersymmetry).
• Future Prospects: The study confirms that the full Run 2 dataset and the upcoming HL-LHC run have the statistical power to either discover or definitively exclude this specific compressed region of the Type-II Seesaw parameter space.

In conclusion, the paper successfully argues that by employing advanced multivariate techniques and relaxing traditional kinematic cuts, the LHC can effectively probe the elusive, compressed mass spectra of the Type-II Seesaw model, potentially uncovering the mechanism behind neutrino masses.