Performance of the LHCb muon detector in Run 3

This paper presents the operation, calibration, and performance evaluation of the upgraded LHCb muon detector during Run 3, demonstrating that despite a fivefold increase in instantaneous luminosity, the system achieves a muon identification efficiency above 90% with sub-percent hadron misidentification probability.

The Gist

Imagine the Large Hadron Collider (LHC) at CERN as the world's most intense, high-speed particle racetrack. For years, the LHCb experiment has been a specialized camera trying to take photos of very rare, fleeting moments in this race—specifically, the decay of heavy particles containing "beauty" and "charm" quarks.

This paper is a report card on how well the Muon Detector (a specific part of the camera) performed during Run 3, a period where the racetrack got five times more crowded and chaotic than before.

Here is the story of how they upgraded the system to keep taking perfect photos in the chaos.

1. The Challenge: From a Busy Street to a Mosh Pit

In the previous era (Run 2), the racetrack had a certain amount of traffic. The LHCb detector was like a security guard at a busy airport, good at spotting specific people (muons) in the crowd.

But for Run 3, the organizers decided to increase the traffic by five times. Suddenly, the airport wasn't just busy; it was a mosh pit.

• The Problem: With five times more particles flying around, the old security guard's system would get overwhelmed. It would start confusing random people in the crowd (hadrons) with the VIPs (muons), or it would get so tired it would miss the VIPs entirely.
• The Goal: They needed to upgrade the security system to handle the crush without losing accuracy.

2. The Hardware Upgrade: Rebuilding the Security Gates

The muon detector is essentially a series of four giant "gates" (stations M2–M5) placed behind thick iron walls. Only the fastest, most penetrating particles (muons) can punch through the iron and hit the gates.

To handle the new traffic, they didn't just tweak the software; they did a complete overhaul of the electronics:

• The Old System: Think of the old electronics as a single, slow typist trying to write down every person who passed.
• The New System: They installed a high-speed digital scanner (called nODE) that can process data at 40 million times per second. It's like upgrading from a typewriter to a supercomputer that can read a library in a second.
• The Shielding: They also added extra "bouncers" (tungsten shields) near the entrance to block low-energy particles that were just causing noise and confusion.

3. The Software: The "Smart Filter"

Even with the new high-speed scanner, the data is overwhelming. The team developed a new muon identification algorithm (a smart filter) to sort the VIPs from the crowd.

They used a two-step process, like a bouncer at an exclusive club:

1. Step 1: The "IsMuon" Check (The Rough Scan): This is a quick, loose check. "Did you hit a few gates in a row?" If yes, you get a "maybe."
2. Step 2: The "Chi-Square" Check (The Deep Dive): This is the smart part. It looks at the pattern of your hits.
   • The Analogy: Imagine a muon is a bullet passing through a wall; it leaves a straight, clean line of holes. A random particle (a hadron) might bounce off things, creating a messy, scattered pattern of holes.
   • The new algorithm calculates the "messiness" of the pattern. If the pattern is too messy, it's likely a fake muon. If it's clean, it's a real one.

4. The Calibration: Tuning the Instruments

Before they could trust the results, they had to tune the detector, much like tuning a musical instrument before a concert.

• Time Alignment: The particles arrive in tiny 25-nanosecond bursts. The electronics had to be synchronized perfectly so that a hit at 12:00:00.000000001 is recorded at the exact right moment. They used special "calibration events" (isolated proton bunches) to adjust the clocks on every single wire.
• Spatial Alignment: They had to make sure the physical position of the detector chambers hadn't shifted. They used the tracks of known particles (like J/ψ mesons) to measure if the chambers had moved even a few millimeters and corrected for it.

5. The Results: A Perfect Performance

The paper reports on data collected in 2024. The results are excellent:

• Efficiency: The detector successfully spotted 90% or more of the real muons. It didn't miss the VIPs.
• Accuracy: It confused real muons with random particles (hadrons) less than 1 time in 1,000.
• Resilience: Even when the "mosh pit" got incredibly crowded (high occupancy), the system didn't break down. It kept performing just as well as it did in the quieter days of Run 2.

The Bottom Line

The LHCb muon detector was like an old, reliable car that was suddenly asked to drive a Formula 1 race. Instead of just driving faster, the team completely rebuilt the engine, installed a new GPS, and trained the driver on a new racing line.

The result? The car didn't just survive the race; it performed exactly as well as it did on the old track, proving that the Upgrade I was a massive success. They successfully turned a chaotic, high-speed environment into a precise laboratory for discovering new physics.

Technical Summary

1. Problem Statement

The LHCb experiment entered Run 3 (2022–2026) with a significant increase in instantaneous luminosity, rising from $4 \times 10^{32} \text{ cm}^{-2}\text{s}^{-1}$ in Run 2 to $2 \times 10^{33} \text{ cm}^{-2}\text{s}^{-1}$ (a factor of five). This increase posed two primary challenges for the muon detector system:

1. High Particle Rates: The increased flux threatened to overwhelm the readout electronics, causing dead-time and reducing detection efficiency.
2. Trigger Performance: The new fully software-based trigger (High Level Trigger 1 and 2) required robust muon identification algorithms capable of maintaining Run 2 performance levels (high efficiency, low hadron misidentification) despite the higher background and combinatorial hits.

The core objective was to verify that the upgraded muon system could sustain a single-hit efficiency above 99% and achieve muon identification efficiencies above 90% with sub-percent hadron misidentification probabilities under these extreme conditions.

2. Methodology and Upgrades

The paper details the hardware interventions, operational strategies, and calibration methods employed to meet these challenges.

A. Hardware Interventions (Upgrade I)

• Shielding: New tungsten shielding and improved beam plugs were installed downstream of the calorimeters to reduce the low-energy particle rate in the innermost regions (R1) of the M2 station by approximately 25%.
• Readout Granularity: The logical layer grouping pads was removed in regions R2–R4 of station M2 and R4 of station M5. This increased granularity to mitigate dead-time effects from the front-end electronics under high rates.
• Electronics Overhaul:
  • Front-End Boards (FEBs): Retained from Run 1/2 (radiation-hardened up to 100 kGy).
  • Off-Detector Electronics (nODE): A major upgrade replacing the previous readout. The nODE boards utilize custom nSYNC ASICs for clock synchronization, bunch crossing alignment, and time measurement. They enable full event readout at 40 MHz via GBTx optical links.
  • Trigger: The system feeds a fully software trigger (HLT1 at 30 MHz, HLT2 at 1 MHz) capable of real-time reconstruction.

B. Operation and Monitoring

• Gas System: Operates with an Ar/CO2/CF4 mixture (38:57:5) in closed mode with >95% recirculation efficiency.
• High Voltage (HV): Set between 2.53–2.62 kV (gain $4-7 \times 10^4$). A novel method was developed to monitor luminosity online using MWPC currents, corrected for gas pressure and beam-gas collisions (SMOG2), providing an independent safety measure against luminosity spikes.
• DAQ Stability: Initial desynchronization issues (14 links/hour) were resolved by adjusting GBTx power supplies and firmware, reducing DAQ inefficiency to negligible levels.

C. Calibration

• Time Alignment: A revised procedure aligns individual nODE channels (previously grouped) using "Time Alignment Events" (isolated bunches). Corrections are applied in 1.6 ns steps to ensure hits fall within the 25 ns bunch crossing window.
• Spatial Alignment: Performed annually using optical references and Kalman filters on $J/\psi \to \mu^+\mu^-$ tracks. In 2024, displacements up to 10 mm in the x-direction were corrected via the conditions database.

3. Key Contributions and Algorithms

The paper introduces and validates two critical components for Run 3 muon identification:

1. IsMuon (Loose Selection): A boolean selection requiring hits in consecutive stations within a Field Of Interest (FOI).
   • The FOI size is momentum-dependent: $FOI = a + b \times \exp(-cp)$.
   • In Run 3, the FOI size was reduced by 20% to lower misidentification probability, with a negligible (<1%) loss in efficiency.
2. $\chi^2_{corr}$ Algorithm: A new algorithm developed specifically for Run 3 to handle combinatorial hits.
   • It calculates a $\chi^2$ based on the distance between the track extrapolation and the closest muon hit, accounting for correlations induced by multiple scattering through the ECAL, HCAL, and iron absorbers.
   • The $\chi^2$ is converted into a likelihood and combined with RICH and calorimeter information for global particle identification.

4. Results

Performance was evaluated using calibration samples collected in 2024 (integrated luminosity of 1.2 fb$^{-1}$ for ID studies).

• Hit Efficiency:
  • Measured using detached $J/\psi \to \mu^+\mu^-$ decays (tag-and-probe method).
  • Result: All detector regions achieved a hit efficiency >99%, meeting the design requirement.
• Muon Identification Performance:
  • Evaluated using samples of muons ($J/\psi$), pions ($K_S^0$), kaons ($\phi$), and protons ($\Lambda$).
  • Efficiency vs. Misidentification: For tracks with $p > 10$ GeV/c, a working point was found with >90% muon efficiency.
  • Hadron Rejection:
    • Pions and Kaons: Misidentification probability at the per mille level (few $\times 10^{-3}$).
    • Protons: Misidentification probability is an order of magnitude lower ($\sim 10^{-4}$).
  • Occupancy Impact: In high-occupancy events (average 120 long tracks vs. 50 in Run 2), muon efficiency remained stable. Proton misidentification varied by a factor of two but remained within acceptable limits.
  • Momentum Dependence: Misidentification is highest at low momentum (where multiple scattering dominates) but drops significantly for $p > 10$ GeV/c.

5. Significance

This paper confirms that the LHCb Upgrade I muon detector successfully met its primary physics goals despite the fivefold increase in luminosity.

• Robustness: The combination of hardware shielding, increased readout granularity, and the new nODE electronics effectively managed the high particle rates without significant efficiency loss.
• Algorithmic Success: The new $\chi^2_{corr}$ algorithm successfully mitigated the increased background of combinatorial hits, allowing the software trigger to maintain Run 2-level particle identification performance.
• Future Readiness: The integrated charge analysis indicates that while the innermost chambers (M2R1) will approach tested limits by the end of Run 4, the outer regions (R3, R4) are sufficiently robust to operate through the future High-Luminosity Upgrade II (up to 2041).

In summary, the LHCb muon system has transitioned successfully to the high-luminosity era, providing the high-purity muon samples necessary for precision measurements of charm and beauty quark physics.