🔬 physics

Hunting for Long-Lived ALPs in Top Events via Displaced Dielectron Pairs at the LHC

This paper presents a generator-level sensitivity study proposing a search strategy at the LHC for long-lived axion-like particles (ALPs) produced in association with top-quark pairs that decay into displaced electron-positron pairs, demonstrating the potential to probe ALP masses between 17 MeV and 10 GeV with cross sections times branching fractions as low as $10^{-3}$ pb using 500 fb $^{-1}$ of data.

Original authors: Juliette Alimena, Freya Blekman, Cesare Cazzaniga, Xabier Cid Vidal, Jeremi Niedziela, Lovisa Rygaard, Rosalia Soto Rey

Published 2026-10-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Juliette Alimena, Freya Blekman, Cesare Cazzaniga, Xabier Cid Vidal, Jeremi Niedziela, Lovisa Rygaard, Rosalia Soto Rey

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

The universe we can see and touch is built from a known set of building blocks, but physicists have long suspected that this visible world is only a fraction of reality. Hidden within the gaps of our current understanding are particles that might exist for a brief moment before vanishing, or perhaps linger long enough to travel across a detector before decaying. These are known as long-lived particles, and they are a favorite target for scientists searching for physics beyond our standard models. Unlike the familiar particles that decay instantly upon creation, these elusive travelers can move a measurable distance away from their point of origin. Finding them requires looking for very specific, unusual patterns in the debris of high-energy collisions, particularly when the decay products are soft, faint, or difficult to spot directly. The challenge is that these particles often hide in plain sight, masked by the overwhelming noise of ordinary particle interactions, requiring researchers to devise clever ways to separate the signal from the background.

A team of researchers has now proposed a new strategy to hunt for one such hypothetical particle, called an axion-like particle, by looking at a very specific and energetic environment: the collision of protons that produces pairs of top quarks. Top quarks are the heaviest known elementary particles, and when they are created in the Large Hadron Collider, they decay almost immediately into a spray of other particles, including energetic jets and leptons. The researchers realized that these prompt, high-energy decay products could serve as a reliable beacon. Even if the new, long-lived particle they are hunting is invisible to standard triggers, the presence of the top quark pair provides a solid handle to select the event. This allows the search to focus on the long-lived particle's decay products, which might be too faint or slow to trigger the detectors on their own.

In this study, the scientists focused on a scenario where the long-lived particle decays into a pair of electrons and positrons. This specific channel is crucial because it opens the door to searching for particles with masses lower than what previous searches could reach. Earlier efforts looked for similar particles decaying into muons, but those searches hit a hard limit: if the particle is too light, it simply cannot produce two muons. By switching to electrons, the researchers could probe a much lighter mass range, extending down to just 17 millionths of a GeV, a scale that overlaps with intriguing anomalies seen in nuclear physics experiments. The team simulated millions of collision events to see if this strategy could work, using the geometry and capabilities of the CMS detector at the Large Hadron Collider as their guide. They did not build a physical detector or run a real experiment; instead, they created a detailed computer model of what would happen if these particles existed and how the detector would respond to them.

The simulation revealed that the strategy is highly effective at filtering out the background noise. The researchers set up a series of digital filters to identify events where a top quark pair was present alongside a pair of electrons that appeared to originate from a point slightly displaced from the main collision spot. They required these electrons to be separated from the main collision point by a distance between 0.1 millimeters and one meter, a range where the new particle would likely decay. They also applied strict rules to ensure the electrons were not just random debris from other processes. The results showed that this method could effectively isolate the signal, even for particles that are very light and decay very close to the collision point or far away. In the most sensitive regions of their search, the team found they could detect a signal if the production rate of these particles was as low as one-thousandth of a picobarn, a unit of probability for particle interactions, assuming the collider runs with a specific amount of data.

The study also explored how the sensitivity changes depending on the mass of the particle and how far it travels before decaying. They found that the method works best for particles with masses between 17 millionths of a GeV and 10 GeV, and for decay lengths ranging from a fraction of a millimeter to a meter. However, the researchers were careful to note the limitations of their work. Because they were working with a computer simulation, they assumed that the detector could perfectly reconstruct the paths of the electrons, even when they are very close together. In reality, when the particle is very light, the two electrons it produces are squeezed so tightly together that they might look like a single blob of energy rather than two distinct tracks. The simulation did not account for the difficulty of separating these tightly packed tracks or for the messy background effects that occur in a real detector, such as material interactions or misreadings. Therefore, the numbers they produced represent an idealized best-case scenario, showing the potential of the method rather than a guaranteed outcome.

Despite these simplifications, the work provides a clear roadmap for future experimental searches. It demonstrates that using the energetic debris of top quarks as a guide can successfully extend the search for long-lived particles into a low-mass region that has been largely unexplored. The researchers suggest that with the actual data from the Large Hadron Collider, and with more sophisticated tools to handle the tricky reconstruction of closely spaced electrons, this approach could reveal new physics. They also noted that for future upgrades of the detector, which will include more detailed calorimeters, the ability to resolve these tightly packed electron pairs might improve significantly. For now, the study stands as a proof of concept, showing that the top quark pair is a powerful ally in the hunt for the universe's most elusive, long-lived travelers.

Technical Summary: Hunting for Long-Lived ALPs in Top Events via Displaced Dielectron Pairs at the LHC

Problem and Motivation
Long-lived particles (LLPs) are a prominent feature of many Standard Model (SM) extensions, often characterized by macroscopic lifetimes resulting from small couplings or limited phase space. While displaced decays offer powerful discrimination against SM backgrounds, experimental searches are frequently hindered when decay products are soft, displaced, or difficult to trigger on directly. A strategy to overcome this involves producing LLPs in association with prompt particles that provide independent event selection handles. Top-quark pair ( $t\bar{t}$ ) production is particularly advantageous due to the energetic leptons, jets, and missing transverse momentum available from top decays.

While previous work (Ref. [4]) investigated long-lived axion-like particles (ALPs) produced with $t\bar{t}$ pairs decaying to displaced dimuons, this channel is kinematically closed for ALP masses below the dimuon threshold ( $m_a < 2m_\mu$ ). This paper addresses the gap by extending the search strategy to the dielectron channel ( $pp \to t\bar{t}a, a \to e^+e^-$ ). This extension is experimentally nontrivial because displaced electrons pose reconstruction challenges distinct from muons, and at low masses, the electron-positron pairs can be highly collimated and soft, falling outside standard triggering and reconstruction regimes. Furthermore, direct searches for long-lived ALPs decaying to electron pairs in association with $t\bar{t}$ have not yet been studied phenomenologically or experimentally.

Methodology
The study presents a generator-level sensitivity analysis using the CMS detector geometry and performance as a guide. The analysis covers ALP masses ( $m_a$ ) between 17 MeV and 10 GeV and proper decay lengths ( $c\tau$ ) between 0.1 mm and 1 m.

Simulation: Signal and background events were generated using MadGraph5_aMC@NLO at $\sqrt{s} = 13$ TeV. The background is modeled using inclusive SM $t\bar{t}$ production, normalized to the measured cross section. Signal samples were generated for various benchmark masses and decay lengths using an effective field theory implementation where the ALP couples to top quarks.
Event Selection: The strategy relies on the prompt $t\bar{t}$ $t \overset{ˉ}{t}$ system for triggering and initial selection, assuming 100% efficiency for this component to isolate the sensitivity of the displaced dielectron signature.
- Pre-selection: Requires at least one electron-positron pair with opposite charges, pseudorapidity $|\eta| \le 2.5$ , and a transverse displacement $l_{xy}$ between 500 $\mu$ m and 0.8 m.
- Signal Selection: Further suppresses background using kinematic and topological variables. Key discriminators include:
  - Invariant mass vetoes for $J/\psi$ , $\psi(2S)$ , and $Z$ bosons.
  - Electron isolation ( $I_e$ ) to reject electrons inside jets.
  - Distance of closest approach (DCA) between electron trajectories.
  - Transverse momentum of the softer electron ( $p_{T, \text{min}}$ ).
  - Collinearity angle ( $\phi_{\text{coll}}$ ) between the dielectron momentum and the displacement vector.
- Cut thresholds were optimized using the Punzi figure of merit for each benchmark.
Background Modeling: The study assumes inclusive $t\bar{t}$ events dominate the background. It explicitly excludes instrumental backgrounds (pileup, material interactions, misreconstruction) and does not apply detector-level reconstruction efficiencies for displaced electrons, treating the analysis as an idealized generator-level projection.

Key Contributions and Results
The primary output of the study is the computation of expected 95% confidence level (CL) upper limits on the production cross section times branching fraction, $\sigma(pp \to t\bar{t}a) \times \mathcal{B}(a \to e^+e^-)$ , as a function of $m_a$ and $c\tau$ .

Sensitivity: For an integrated luminosity of 500 fb $^{-1}$ , the study projects sensitivity to cross-section times branching fraction values of order $10^{-3}$ pb in the most sensitive regions of the parameter space.
Mass Reach: The strategy successfully extends the reach of top-associated LLP searches to masses below the dimuon threshold, covering the region relevant to the X17 anomaly (approx. 17 MeV).
Efficiency and Robustness:
- The total selection efficiency for signal events varies significantly with mass and decay length, primarily driven by the probability of the ALP decaying within the targeted transverse displacement range.
- Robustness tests indicate that the projected sensitivity is most strongly affected by the efficiency of reconstructing displaced dielectron pairs. Reducing the displaced-dielectron reconstruction efficiency to 10% or 1% weakens the limits by factors of 10 and 100, respectively. In contrast, reducing the prompt $t\bar{t}$ selection efficiency to 10% or increasing background yields by a factor of 10 only weakens limits by a factor of roughly 3.
Model Interpretation: The results are interpreted within a specific "top scenario" where the ALP couples only to top quarks at the cutoff scale. This allows for the translation of limits into constraints on the coupling $|c_{tt}|/f_a$ . The study demonstrates that the dielectron channel provides complementary sensitivity to the dimuon channel, specifically probing the low-mass region inaccessible to dimuon searches.

Significance and Claims
The paper claims to establish the generator-level potential of the top-associated strategy for displaced-dielectron signatures. It highlights that this approach allows for the probing of light long-lived particles decaying to electron pairs, a regime largely unexplored at the LHC.

The authors are explicit about the limitations of their work:

The results are idealized generator-level projections and do not include detector-level displaced-electron reconstruction efficiencies or instrumental backgrounds.
The study does not claim that the sensitivity to highly collimated, low-mass ALPs is fully achievable with current reconstruction algorithms; rather, it identifies the kinematic reach and the specific challenges (collimation, softness) that a future detector-level study must address.
The paper serves as a foundational step to motivate and guide future experimental searches in the Run 2, Run 3, and High-Luminosity LHC (HL-LHC) eras, noting that future upgrades like the CMS High-Granularity Calorimeter may aid in resolving collimated pairs.

In summary, the work quantifies the theoretical reach of a specific search channel, demonstrating that top-associated production is a promising avenue for discovering light, long-lived ALPs decaying to electrons, provided that the experimental challenges of reconstructing displaced, collimated electron pairs can be overcome.

Technical Summary: Hunting for Long-Lived ALPs in Top Events via Displaced Dielectron Pairs at the LHC

More like this