⚛️ high-energy experiments

Search for long-lived particles decaying into muons in proton-proton collisions at $\sqrt{s}$ = 13.6 TeV using data scouting

Using 62.4 fb $^{-1}$ of 13.6 TeV proton-proton collision data collected via a high-rate data scouting stream, the CMS experiment performed a search for long-lived particles decaying into muons, finding no significant excess over the Standard Model and setting improved upper limits on branching fractions for various beyond-the-Standard-Model scenarios.

Original authors: CMS Collaboration

Published 2026-06-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: CMS Collaboration

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 massive, ultra-fast particle smasher. It fires two beams of protons at each other, creating a chaotic explosion of debris. Usually, scientists are looking for the "big bang" moments—huge, heavy particles that appear and vanish instantly.

But this paper is about looking for the "ghosts" in the machine.

The Mystery: Invisible Travelers

The scientists are hunting for Long-Lived Particles (LLPs). Think of these as cosmic ghosts. In the Standard Model (our current rulebook for physics), particles usually decay (break apart) almost instantly. But these "ghosts" are different. They are created in the collision, travel a noticeable distance through the detector, and then decay into something we can see: a pair of muons (heavy cousins of electrons).

Why do we care? Because these ghosts might be Dark Matter or particles from a "Hidden Sector" that we don't understand yet. They are the missing pieces of the puzzle that explain why the universe has mass but we can't see it.

The Detective's Tool: The "Scouting" Camera

Here is the problem: Standard cameras in the LHC are set to take photos of only the biggest, brightest explosions. If a ghost particle travels a few centimeters and then breaks into two tiny, slow-moving muons, the standard cameras ignore it. They think it's just background noise.

To catch these ghosts, the CMS team used a special technique called "Data Scouting."

The Analogy: Imagine a security guard at a stadium. The standard guard only stops people carrying large bags or weapons (high-energy particles). They let everyone else through.
The Scouting Guard: This new guard is different. They are told to stop everyone who looks even slightly suspicious, even if they are carrying a tiny, slow-moving item. To do this, they don't take a full 4K video of every person (which would clog the system); instead, they take a quick, low-resolution snapshot of just the important details.
The Result: This allowed the team to collect a massive amount of data (62.4 "snapshot" units) specifically looking for these slow, displaced particles that the standard cameras would have missed.

The Hunt: Looking for the "Displaced"

The team looked for a specific signature:

Two Muons: A pair of particles appearing together.
Displaced Vertex: Instead of appearing right where the collision happened (the "primary vertex"), these muons appeared a few centimeters to a few meters away.
The "Ghost" Path: If you trace the muons backward, they don't point to the center of the explosion. They point to a spot where a ghost particle traveled and then popped.

They checked every possible distance the ghost could travel, from just inside the detector's inner layers all the way out to the outer edges (up to 70 cm). They also looked for cases where four muons appeared, which would happen if two ghosts were created at once.

The Models: What They Were Looking For

The scientists tested three main theories about what these ghosts could be:

The Dark Photon: A hidden version of the photon (light particle) that mixes with our world. It's like a secret messenger that can turn into muons.
The Dark Shower: Imagine a particle that doesn't just break into two pieces, but shatters into a whole "shower" of hidden particles, which then decay into muons. It's like a firecracker that explodes into smaller firecrackers, which then explode again.
The Scalar from a Bottom Quark: A heavy particle (a "bottom quark") decaying into a new, long-lived scalar particle that then turns into muons.

The Verdict: No Ghosts Found (Yet)

After analyzing millions of events, the team found no evidence of these long-lived particles. The data matched the "Standard Model" prediction perfectly. There were no unexpected spikes or "ghostly" peaks in the data.

What does this mean?

Ruling Out Suspects: Even though they didn't find the ghosts, they didn't fail. By not finding them, they proved that these particles cannot exist with certain properties.
New Boundaries: They drew a new, stricter line in the sand. If these particles exist, they must be even more elusive (heavier, lighter, or longer-lived) than the team previously thought.
The "Dark" Zone: They successfully explored a region of physics that was previously invisible to the LHC, specifically looking for particles that travel further and have lower energy than ever before.

Summary

The CMS team used a special "scouting" method to take a closer look at the quiet, slow-moving debris of particle collisions. They were hunting for "ghost" particles that travel a short distance before turning into muons. They didn't find any ghosts, but by proving they aren't there in the places they looked, they have narrowed down the search for the secrets of Dark Matter and the hidden universe. They have effectively told the universe: "If you are hiding there, you are hiding very well indeed."

Technical Summary: Search for Long-Lived Particles Decaying into Muons in Proton-Proton Collisions at $\sqrt{s} = 13.6$ TeV Using Data Scouting

Problem and Motivation
The origin of dark matter remains a primary unsolved problem in particle physics, motivating searches for a hidden (dark) sector that interacts weakly with Standard Model (SM) particles. This paper addresses the search for long-lived particles (LLPs) within several Beyond the Standard Model (BSM) frameworks, specifically targeting scenarios where the Higgs boson decays into dark sector particles or where long-lived scalars arise from $b$ -hadron decays. Key models include the Hidden Abelian Higgs Model (HAHM), where the Higgs decays into pairs of long-lived dark photons ( $Z_D$ ); dark-shower models involving confining dark gauge groups producing "dark showers" of dark pions and photons; and minimal extensions involving a singlet scalar ( $\phi$ ) mixing with the Higgs.

A significant challenge in detecting these signatures, particularly for low-mass resonances ( $m < 10$ GeV) with macroscopic decay lengths, is the limitation of standard trigger streams. Standard triggers typically require high transverse momentum ( $p_T$ ) particles, rendering low-mass, displaced muon pairs inaccessible. Furthermore, previous analyses using CMS data scouting (a high-rate trigger stream retaining reduced event information) were limited to decay displacements ( $l_{xy}$ ) less than 11 cm due to pixel hit multiplicity requirements.

Methodology
The analysis utilizes proton-proton collision data collected by the CMS experiment at $\sqrt{s} = 13.6$ TeV during 2022 and 2023, corresponding to an integrated luminosity of 62.4 fb $^{-1}$ .

Trigger Strategy: The search employs a dedicated "data scouting" dimuon trigger stream. This stream retains only reduced information (four-momenta, hit counts, and track details) at the High-Level Trigger (HLT), allowing for a high-rate selection of events with low transverse momentum thresholds ( $p_T \gtrsim 3-4$ GeV). Crucially, updates to the scouting triggers in 2022 removed the pixel hit multiplicity requirement, enabling the reconstruction of muon pairs with decay displacements up to 70 cm ( $l_{xy} > 11$ cm), a phase space previously inaccessible to CMS scouting analyses.
Event Selection: Events are selected based on the presence of at least one pair of oppositely charged (OS) muons associated with a secondary vertex (SV). Selection criteria include:
- SV displacement $l_{xy} < 70$ cm.
- Vertex fit quality ( $\chi^2/\text{dof} < 3$ ) and coordinate uncertainties.
- Suppression of material interactions by excluding SVs within tracker modules.
- Topological requirements ensuring the dimuon momentum vector is collinear with the SV vector (pointing) to reject random combinatorial backgrounds.
- Isolation criteria and impact parameter significance ( $|d_{xy}|/\sigma_{d_{xy}} > 2$ ) to distinguish displaced muons from prompt backgrounds.
- Specific handling of "excess hits" ( $N_{\mu\mu}^{\text{excess}}$ ) to reject prompt muons that accidentally form a displaced vertex.
Categorization: Selected events are categorized into:
1. Dimuon events: Single OS muon pairs, further split by $l_{xy}$ (8 bins from 0 to 70 cm), pointing vs. non-pointing topology, isolation status, and $p_T^{\mu\mu}$ (< 25 GeV vs. > 25 GeV).
2. Four-muon events: Events with two OS muon pairs, categorized as "multivertex" (spatially resolved SVs) or "overlapping vertex" (SVs overlapping within uncertainties).
Background Modeling: The SM background is estimated directly from data using smooth analytical functions (Bernstein polynomials, power laws, exponentials) fitted to the multimuon invariant mass spectrum in sliding windows. Known SM resonances (e.g., $J/\psi$ , $\Upsilon$ ) are masked.
Signal Interpretation: Signal Monte Carlo (MC) samples are generated for HAHM ( $H \to Z_D Z_D$ ), dark-shower scenarios (Scenarios A and B1), and $b$ -hadron decays ( $h_b \to \phi X$ ). Upper limits on branching fractions are set using a profile likelihood ratio with the CL $_s$ criterion.

Key Contributions

Extended Phase Space: The analysis significantly extends the accessible phase space for displaced muon searches by utilizing updated scouting triggers to probe decay lengths up to 70 cm, overcoming the 11 cm limitation of previous scouting analyses.
Low-Mass Sensitivity: By employing low- $p_T$ thresholds, the search targets narrow resonances with masses below 10 GeV, a region often excluded by standard high- $p_T$ triggers.
Comprehensive Categorization: The inclusion of four-muon topologies (both resolved and overlapping vertices) and the detailed binning of dimuon events by displacement and isolation allow for sensitivity to a wide variety of BSM production and decay mechanisms.
Data-Driven Background: The reliance on data-driven background modeling ensures robustness against simulation uncertainties in the low-mass, high-multiplicity regime.

Results
No significant excess of events above the Standard Model prediction was observed in the dimuon or four-muon invariant mass spectra across all analyzed categories. Consequently, 95% confidence level (CL) upper limits were set on the branching fractions for the considered BSM models:

Hidden Abelian Higgs Model: Limits on $\mathcal{B}(H \to Z_D Z_D)$ were set for dark photon masses ( $m_{Z_D}$ ) ranging from 0.5 to 50 GeV and proper decay lengths ( $c\tau_{Z_D}^0$ ) of 0.1, 1, 10, and 100 cm. The results improve upon previous constraints, particularly for $m_{Z_D} < 5$ GeV and $c\tau_{Z_D}^0 > 1$ cm.
Dark-Shower Models: Upper limits on $\mathcal{B}(H \to \Psi\Psi)$ were established for Scenarios A and B1. The analysis extends the excluded range of proper decay lengths for dark photons ( $A'$ ) and dark pions ( $\tilde{\pi}_3$ ) from 10 cm up to 100 cm, surpassing previous limits.
Long-Lived Scalar from $b$ -Hadron: Limits on the product $\mathcal{B}(h_b \to \phi X)\mathcal{B}(\phi \to \mu\mu)$ were set for scalar masses $m_\phi$ up to 4 GeV and lifetimes of 0.1, 1, and 10 cm. These constraints are found to be comparable to previous results, with the improved trigger efficiency compensating for the smaller dataset relative to earlier runs.

Significance
The paper claims that the resulting constraints represent the most stringent limits to date for a wide range of parameter space in the Hidden Abelian Higgs model (specifically for $m_{Z_D} < 5$ GeV and $c\tau_{Z_D}^0 > 1$ cm) and for dark-shower models (for proper decay lengths $> 10$ cm). The work demonstrates the efficacy of the data scouting technique in exploring "blind spots" in the detector acceptance, specifically low-mass, long-lived particles that are inaccessible to standard trigger streams. By extending the reach to larger decay lengths and lower masses, the analysis significantly narrows the viable parameter space for these specific BSM hypotheses.