The LHCb Beam Conditions Monitor in LHC Run 3
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment |
| Published | 2026-08-04 |
| Authors | J. Albrecht, F. Alessio, Claus Maximilian Bäcker, Christian Bäumer, M. S. Bieker |
| DOI | 10.1016/j.nima.2026.171913 |
The upgraded LHCb Beam Conditions Monitor triggered six beam dumps from 2022 to 2025 while operating with beams at the LHC. Each of its two stations contains eight polycrystalline chemical vapor deposited, or pCVD, diamond sensors. LHCb operated at an instantaneous luminosity five times that of Run 2. The system is integrated with the LHC Beam Interlock System and the LHCb Vertex Locator safety chain.
Upgraded stations and mechanical support
Section titled “Upgraded stations and mechanical support”The BCM monitors particle flux upstream and downstream of the interaction point. These positions cover beam trajectories that could strike the VELO, the smallest aperture in LHCb. The upstream station, BCM-U, uses two half-rings on a sliding support divided into vertical halves. The split design gives maintenance access even though the PLUME luminometer stands directly in front of the station. Its sensors were rotated 15 degrees around the beam pipe from their earlier positions.
The downstream station, BCM-D, kept its general design and remains inside the LHCb dipole magnet. Its support uses two half-rings milled from polyaryletheretherketone, or PEEK, to reduce the detector material budget. New sensors and revised mechanical support were installed in 2021. The new readout electronics followed in 2023.
Diamond sensor quality assurance
Section titled “Diamond sensor quality assurance”Thirty-two pCVD diamond tiles were procured from Element Six Technologies US Corporation for the upgrade. Each tile measured 10 mm by 10 mm by 0.5 mm. Electrodes covered 8 mm by 8 mm on both faces and combined 50 nm titanium with 50 nm gold. The authors identify two material-related concerns. Charge trapping at grain boundaries lowers the ratio of detected charge to released charge. Irregular dark currents can cause unwarranted beam dumps and require sensor rejection.
Each sensor was biased at 200 V and exposed to a constant electron flux from a Strontium-90 source. A picoamperemeter recorded its current for 90 minutes. Sensors with erratic current behavior were discarded during this first test stage. That decision removed 25% of the procured sensors.
The second stage measured relative charge collection efficiency, CCE. Absolute CCE could not be determined because the source’s spatial electron flux distribution was unknown. The authors assumed that the flux was constant and uniform across all angles. They calculated relative CCE from the average current over the final 10 seconds of each 90-minute measurement, after the current reached a stable plateau. The results showed little variation among the accepted sensors. The authors distributed them evenly between the two stations and a set of spares.
The team validated the test at the West German Proton Therapy Center Essen. Sensors used in Runs 1 and 2 received the same electron exposure and were irradiated with a 100 MeV proton beam 12.3 mm in diameter. Total collected proton-beam charge came from integrating sensor current over irradiation. For an independent set of four sensors, that charge correlated linearly with the stable current measured during electron exposure. The authors therefore treated the QA setup as a reliable measure of sensor performance.
FPGA readout and interlock logic
Section titled “FPGA readout and interlock logic”Each station feeds a Current to Frequency Converter card, which measures current every 40 microseconds. Two redundant optical links per card carry the digitized data to the Machine Interface and Beam Abort Decision board. Its FPGA checks for missing or malformed data packages, then selects one redundant input from each station for analysis.
The board applies two dump criteria with time scales of 80 microseconds and 1,280 microseconds. A decision uses the measured absolute current and the prevalence of a specific spatial distribution across a station. Unsafe conditions trigger a beam abort through a direct hardware connection to the LHC Beam Interlock System. The BCM also sends its operating status to the VELO Safety System through a hardware interlock.
For each dump, the board stores the preceding 200 seconds of data for review. It also records operating parameters in real time for display in the LHCb control system. Injection thresholds are higher than collision thresholds because the VELO is moved farther from the interaction point during injection. If the BCM malfunctions, the VELO moves automatically to a safe position. Beam interruptions do not begin until the BCM fault is resolved.
Thresholds and VELO coverage
Section titled “Thresholds and VELO coverage”Geant4 simulations supplied the basis for the Run 3 dump thresholds. The authors calculated energy deposition in the sensors over a 40 microsecond interval under nominal operating conditions. The resulting induced currents were four times the currents measured during Run 2. The dump thresholds were therefore increased by a factor of four.
The simulation also checked whether the BCM still covered trajectories that could damage the VELO. The distance from the VELO sensor edges to the interaction point fell to 5.1 mm. The distance between the VELO aluminum foil and the beam fell from 5.5 mm to 3.5 mm. Adverse trajectories were defined through a beam-center offset and a momentum inclination relative to nominal conditions. Nominal values were x equals 0 mm and theta equals 0 mrad. Across the simulated adverse conditions, energy deposition in the VELO sensors remained proportional to energy deposition in the BCM sensors. The authors conclude that the BCM remains able to protect the VELO.
Commissioning and beam dump record
Section titled “Commissioning and beam dump record”Commissioning placed the new MIBAD readout in parallel with the legacy TELL1 board. Both systems received identical data from the front-end cards, so engineers could test MIBAD in situ while TELL1 controlled the interlock. Full transfer to MIBAD occurred during the LHC technical stop in June 2024.
The first MIBAD-triggered dump occurred during injection kicker magnet tests for LHC beam 2. The magnets sit about 160 m from the LHCb interaction point. Collimators behind them intercept misdirected beams and produce particle showers. Both BCM stations registered the increased particle flux. MIBAD withdrew beam permission and triggered the dump.
From 2022 to 2025, the BCM operated continuously in the presence of beams and triggered six dumps in total. The operating record covers upgraded stations, revised mechanical support, and the transition from the legacy readout to MIBAD.
Material selection and unresolved limits
Section titled “Material selection and unresolved limits”The results define several practical selection criteria for diamond beam-loss sensors. Stable dark current is essential because erratic behavior can cause unwarranted beam dumps. Such sensors made up 25% of the new tiles rejected during QA. Charge collection must also be consistent across tiles. The accepted set showed little relative CCE variation and could then be distributed between stations.
Two quantitative limits remain in the reported data. It reports relative CCE because absolute CCE could not be measured. It also gives a fourfold threshold change without publishing the final absolute dump-current values. The threshold decision rests on Geant4 predictions of nominal operation, while the trajectory study checks coverage of the VELO.
These results support pCVD diamond for this protection role. Sensor stability was tested before installation, redundant readout was verified in parallel, and the station geometry was checked against VELO exposure.
Materials and support
Section titled “Materials and support”The BCM uses 0.5 mm thick pCVD tiles with Ti and Au electrodes for beam-loss monitoring. 6CCVD offers polycrystalline diamond grown by MPCVD in thicknesses from 0.1 to 500 micrometers, which includes the 0.5 mm tile, with metallization options including Ti and Au. The diamond materials catalog covers that material family, while engineering support provides a route for custom fabrication and engineering discussions.
View Original Abstract
At the Large Hadron Collider (LHC), the sensitive detector components of the experiments need to be protected against adverse beam conditions. Thus, the beam conditions have to be closely monitored at multiple locations along the collider. At the LHCb experiment, the Beam Conditions Monitor (BCM) is mounted around the beam pipe at two stations, upstream and downstream of the interaction point. It is built of 8 polycrystalline diamond sensors per station and integrated into the LHC Beam Interlock System to trigger beam dumps if unsafe beam conditions are detected. The BCM system has operated reliably during LHC Run 1 and 2. During the Long Shutdown 2, the LHCb detector underwent a major upgrade to maintain its performance in Run 3 at a fivefold instantaneous luminosity with respect to Run 2. Within this upgrade, the BCM was equipped with new diamond sensors and a new FPGA-based readout system. Additionally, the mechanical support of the BCM upstream station was adjusted to allow for the installation of the PLUME luminometer. After installation of the new sensors and mechanical support, the new readout system was commissioned in 2023. In total, the BCM induced six beam dumps from 2022 to 2025. This proceeding gives an overview of the upgraded BCM detector and its performance during Run 3 operation until 2025.
Products
View Specs →
Advanced Thermal Simulation Service Diamond Heat Spreader Performance Analysis
View Specs →