Search for resonances in four top quark events in the 2 lepton final state

Using 138 fb⁻¹ of 13 TeV and 35 fb⁻¹ of 13.6 TeV proton-proton collision data, this study presents the first search for beyond-the-Standard-Model resonances in four-top-quark production via the two-lepton channel, finding no significant excess and setting exclusion limits on various mediator models, including Z' bosons up to 850 GeV.

The Gist

Imagine the Large Hadron Collider (LHC) as the world's most powerful particle smasher. Scientists fire protons at each other at nearly the speed of light to see what happens when they crash. Usually, these crashes create a chaotic mess of particles, but sometimes, they create something rare and special: four top quarks at once.

Think of top quarks as the "heavyweights" of the particle world. Creating four of them in a single crash is like trying to hit four bowling balls with a single ping-pong ball. It's incredibly rare and difficult to spot.

The Big Question: Is the Standard Model Broken?

For a long time, the "Standard Model" (the rulebook of physics) predicted exactly how often these four-top-quark crashes should happen. But recently, experiments found that they happen more often than the rulebook says they should.

This is exciting! It's like if you were baking cookies and the recipe said you should get 10 cookies, but you kept getting 12. You might wonder: "Is there a secret ingredient I'm missing?" In physics, that "secret ingredient" could be New Physics—particles or forces we don't know about yet.

The Detective Work: Hunting for the "Mediator"

The scientists in this paper (from the CMS collaboration) decided to play detective. They hypothesized that maybe a new, invisible "messenger" particle (called a mediator) is crashing into the protons, breaking apart, and then turning into those four top quarks.

They looked for three types of potential messengers:

1. Z' (Z-prime): A heavy, electrically neutral cousin of the Z boson.
2. Scalars/Pseudoscalars: New types of Higgs-like particles.
3. ALPs (Axion-Like Particles): Ghostly particles that might explain dark matter.

The Strategy: Finding a Needle in a Haystack

The team analyzed a massive amount of data (173 "inverse femtobarns," which is a fancy way of saying they looked at trillions of collisions). They focused on a specific "signature":

• The 2-Lepton Channel: They looked for crashes that produced exactly two "leptons" (like electrons or muons).
• The Analogy: Imagine a crime scene. Most crimes leave a messy trail (QCD background). But this specific type of crime leaves two very distinct footprints (the two leptons) and a pile of debris (the other particles). By looking for those two specific footprints, they could ignore the messy noise and focus on the interesting stuff.

The New Tool:
To catch the heavy top quarks, they used a new, smarter "net" called a HOTVR jet.

• Old Net: A fixed-size net that only caught very fast, heavy fish.
• New Net: A smart net that changes its size based on how fast the fish is swimming. If the fish is slow, the net gets bigger to catch it. This allowed them to catch more of the "four-top" events they were looking for.

They also used a super-smart computer brain (a Boosted Decision Tree) to distinguish between a real "four-top" event and a fake one that just looked like one.

The Results: No New Physics... Yet

After sifting through all the data, the scientists found no significant evidence of these new messenger particles.

• The Verdict: The number of four-top-quark events they saw matched the "Standard Model" predictions (within the margin of error).
• The "Exclusion": While they didn't find the new particles, they did set a "No Trespassing" sign. They proved that if these new particles do exist, they cannot be lighter than a certain weight.
  • For the Z' mediator, they ruled out anything lighter than 850 GeV (about 900 times the mass of a proton).
  • They also set limits on the other types of particles (scalars and ALPs).

Why Does This Matter?

Think of this like searching for a specific type of alien in a forest.

• The Search: They scanned the whole forest (the data) with high-tech binoculars (the new detectors and algorithms).
• The Outcome: They didn't find the alien.
• The Value: Even though they didn't find the alien, they now know for sure that the alien isn't hiding in the bottom 850 meters of the forest. This tells future explorers exactly where not to look, so they can focus their search on the higher, more interesting branches.

The Future

This was just the first search using the very latest data from 2022 (when the LHC was running at a slightly higher energy). The scientists are optimistic that with even more data from the future (Run 3), they will have a bigger net and a sharper eye, potentially catching that "secret ingredient" that explains why the universe is the way it is.

In short: They looked for new physics in a very rare particle crash, used clever new tools to find it, didn't find it this time, but successfully narrowed down the search area for the next generation of physicists.

Technical Summary

1. Problem and Motivation

The paper addresses the observation of the four-top quark ($4t$) production process, which was confirmed by ATLAS and CMS in 2023. A persistent anomaly exists where the measured cross-section for $4t$ production is higher than the Standard Model (SM) prediction. This discrepancy suggests the potential contribution of Beyond-the-Standard-Model (BSM) physics.

The primary goal of this analysis is to search for BSM resonances that decay into four top quarks ($t\bar{t}t\bar{t}$). Specific theoretical models motivating this search include:

• Top-philic $Z'$ bosons: Vector mediators coupling predominantly to top quarks.
• Scalar and Pseudoscalar bosons: Predicted in Two-Higgs-Doublet Models (2HDM).
• Axion-Like Particles (ALPs): Pseudoscalars coupling to top quarks.

2. Methodology and Analysis Strategy

Dataset and Kinematics

• Data: The analysis utilizes the full CMS Run 2 dataset ($\sqrt{s} = 13$ TeV, $138 \text{ fb}^{-1}$) combined with early Run 3 data ($\sqrt{s} = 13.6$ TeV, $35 \text{ fb}^{-1}$). Total integrated luminosity is $173 \text{ fb}^{-1}$.
• Signal Topology: The search targets a specific decay chain where a heavy mediator ($X$) decays into a top-antitop pair ($t\bar{t}$), which then produces two additional top quarks via SM processes or associated production. The specific final state requires:
  • Two top quarks decaying leptonically ($t \to Wb \to \ell\nu b$), yielding 2 leptons.
  • Two top quarks decaying hadronically ($t \to Wb \to q\bar{q}'b$), reconstructed as large-radius jets.
• Advantage: This "2-lepton" channel suppresses Quantum Chromodynamics (QCD) backgrounds significantly compared to 1-lepton or all-hadronic channels, while the hadronic tops allow for full reconstruction of the resonance mass.

Object Reconstruction and Selection

• Leptons: Two isolated leptons (electrons or muons) are required.
• Small-Radius Jets: Two small-radius jets are required, with at least one $b$-tagged. Signal regions are categorized by the number of $b$-tagged jets (1 or $\ge$ 2).
• Large-Radius Jets (Top Reconstruction):
  • Standard anti-$k_T$ jets ($R=0.8$) are inefficient for top quarks with $p_T < 800$ GeV.
  • Solution: The analysis employs HOTVR jets, a variable-radius algorithm where the jet radius scales inversely with $p_T$. This optimizes the capture of hadronic top decay products across a wider $p_T$ spectrum.
• Top Tagging: A new Boosted Decision Tree (BDT) based top tagger was developed specifically for HOTVR jets.
  • Performance: It achieves a 32% signal efficiency for low-$p_T$ tops ($200 < p_T < 400$ GeV) compared to 20% for previous cut-based taggers.
  • Working Point: An operating point with 84% signal efficiency and a 22% mistag rate was selected to maximize sensitivity given the low signal rates.

Background Estimation

• Dominant Background: $t\bar{t}$ production where tops decay leptonically, and additional jets are mistagged as hadronic tops.
• Estimation Method: A transfer factor derived from simulation is applied to control regions (requiring 0 top-tagged HOTVR jets) to estimate the background in the Signal Regions (SR).
• Secondary Background: Events where at least one HOTVR jet originates from a real top quark are modeled directly via simulation.

Statistical Analysis

• A binned maximum likelihood fit is performed on the invariant mass of the two leading HOTVR jets ($m_{JJ}$).
• The fit is split by lepton flavor, $b$-jet category, and center-of-mass energy ($\sqrt{s}$).
• Limits are set on the production cross-section times branching fraction ($\sigma \times \text{BR}$) for various mediator masses and widths.

3. Key Contributions

1. First 2-Lepton Search: This is the first search for BSM resonances in the $4t$ process specifically targeting the 2-lepton final state.
2. Inclusion of 13.6 TeV Data: It is the first $4t$ search to incorporate data from the 13.6 TeV run, leveraging the ~30% higher signal cross-section at this energy despite the lower luminosity.
3. Novel Tagger Development: The introduction of a BDT-based top tagger for variable-radius (HOTVR) jets, significantly improving signal acceptance for lower-$p_T$ top quarks compared to previous methods.
4. Comprehensive Model Testing: The analysis sets limits not only on vector $Z'$ bosons but also on scalar, pseudoscalar, and ALP mediators.

4. Results

• Observation: No significant excess of events over the SM background prediction was observed.
• Data Fluctuation: An upward fluctuation was noted in the first bins of the $m_{JJ}$ distribution, resulting in observed limits that are weaker than expected, particularly for lower mediator masses. The local significance of this fluctuation is 2.2 standard deviations for a 500 GeV mediator with a 4% width.
• Exclusion Limits:
  • $Z'$ Mediators: For a $Z'$ with a 50% decay width, masses up to 850 GeV are excluded at 95% Confidence Level (CL) (Expected: 1000 GeV).
  • Scalar/Pseudoscalar: Limits were set for $t\bar{t}\phi$ and $t\bar{t}a$ processes, showing similar exclusion power to the $Z'$ search due to comparable kinematic properties.
  • ALPs: Limits on the ALP-top coupling ($c_t/f_A$) were derived and compared with existing CMS results from $t\bar{t}$ resonance searches and missing transverse momentum ($p_T^{miss}$) searches.

5. Significance and Conclusion

This analysis represents a significant step in probing the anomalous $4t$ cross-section. By utilizing the 2-lepton channel and advanced jet reconstruction techniques (HOTVR + BDT), the collaboration has expanded the sensitivity to BSM physics in a challenging final state.

While no new physics was discovered, the exclusion of $Z'$ mediators up to 850 GeV (and 1000 GeV expected) constrains top-philic BSM models. The authors note that the analysis is currently statistically limited. Future improvements are anticipated with the full Run 3 dataset, which will provide higher statistics to reduce uncertainties and potentially uncover the source of the $4t$ cross-section excess.