Barrel Timing Layer (BTL) Overview

• Barrel Timing Layer (BTL) is a precision, barrel-shaped timing detector that provides sub-100 ps measurements for vertexing and particle identification in high-luminosity colliders.
• It employs advanced sensor technologies such as LYSO:Ce/SiPM for CMS and fast CMOS MAPS for Belle II, achieving time resolutions as low as 25 ps under high radiation.
• The design mitigates event pileup and optimizes vertex association, significantly enhancing physics analyses through improved particle ID and efficient signal recovery.

The Barrel Timing Layer (BTL) designates large-area precision timing detectors implemented as barrel-shaped layers around collider interaction regions to provide sub-100 ps time-of-arrival measurements for minimum-ionizing particles (MIPs). BTLs are central to both event pileup mitigation and time-of-flight-based particle identification (PID) in modern high-luminosity collider experiments. The leading implementations are the CMS BTL—part of the MIP Timing Detector (MTD) upgrade for the HL-LHC—and the STOPGAP BTL, a proposed upgrade to the Belle II experiment. While their mechanical contexts and system-level roles differ, both exemplify the frontier of precision timing in high-rate, high-radiation particle physics environments.

1. Functional Role and System Integration

The BTL in CMS, planned for HL-LHC operation, is a 38–40 m² cylindrical layer mounted at a radial position of approximately 1.15–1.17 m, between the outer silicon tracker and the electromagnetic calorimeter, covering $|\eta| < 1.48$ (Dutta, 2018, Palluotto, 18 Jan 2026, Addesa et al., 15 Apr 2025). Its primary function is to deliver a per-track time resolution of 30–60 ps throughout the HL-LHC data taking, thereby reducing the $O(200\,\mathrm{ps})$ spread of the collision time and enabling 4D vertexing, pileup mitigation, and new time-of-flight (TOF) searches for slow or long-lived particles. The BTL modules do not significantly increase the tracking material budget and are supported on a cold volume, sharing cooling infrastructure with the tracker (Palluotto, 18 Jan 2026, Dutta, 2018).

In Belle II, the Barrel Timing Layer ("STOPGAP") is proposed as a supplementary layer placed in the nominal $\sim 2\,\mathrm{cm}$ gaps between the existing fused-silica Time-of-Propagation (TOP) bars. It is specifically designed to recover $\sim 6\,\%$ of nominally uncovered tracks, as well as reduce the additional $\sim 3\%$ degraded by edge effects, thus enhancing the overall barrel PID acceptance by $\sim 10\%$ (Hartbrich et al., 2022). The STOPGAP BTL is a compact, thin, modular system situated between the CDC outer shell and the inner face of the TOP enclosure, constrained to a radial space of about 45 mm.

2. Detector Module Architecture and Sensor Technology

CMS BTL:

The fundamental sensing element is a Lutetium-Yttrium Orthosilicate (LYSO:Ce) scintillator bar, typically $3.75 \times 3.12 \times 54.7\,\mathrm{mm}^3$, polished, optically isolated, and read out at both ends by SiPM arrays ("double-ended" configuration). Each module contains 16 bars, arranged into sensor modules (SMs), which are encapsulated with front-end electronics in a copper support housing ("Detector Module," DM). High-density arrays of SiPMs (microcell pitch 25 $\mu$m) are matched to the bar ends and maintained at $T \approx -45\,^\circ$C with integrated TECs (Addesa et al., 2024, Addesa et al., 15 Apr 2025, Palluotto, 18 Jan 2026).

Parameter	CMS BTL Value	Belle II STOPGAP BTL Value
Geometry	38–40 m², $|\eta|<1.48$	$\sim$1–3 m², 16 azimuthal sectors
Sensing	LYSO:Ce/SiPM	Fast CMOS MAPS (baseline)
Active Area	331,776 channels (CMS)	$2 \times 10^3$–$6 \times 10^3$ channels (STOPGAP)

STOPGAP (Belle II BTL):

Each STOPGAP module employs two silicon-sensor layers (targeting monolithic CMOS MAPS in 65 nm HV-CMOS or BiCMOS processes). Pixel sizes range from $200 \times 200\,\mu\mathrm{m}^2$ up to $1\,\mathrm{cm^2}$ depending on granularity requirements. The mechanical support is ultrathin (goal: $\leq 2$ mm active + $\leq 1$ mm support per layer), constructed with carbon-fiber frames and integrated microchannel cooling (Hartbrich et al., 2022).

Alternate sensor options evaluated for STOPGAP include:

• LGADs (Low-Gain Avalanche Diodes): Time resolution 20–30 ps, but requiring double layers for efficiency.
• LYSO+SiPM: Excellent timing (30 ps) but excessive material budget for this application.

3. Timing Performance, Readout, and Calibration

The BTL's core design objective is single-track, single-hit timing at or below 30 ps initially, and maintaining sub-60 ps through lifetime (bearing end-of-life radiation and dark-count rates).

CMS BTL:

• Test-beam results for the final module geometry demonstrate 25 ps (unirradiated) to 55 ps ($2 \times 10^{14}\,n_\mathrm{eq}/\mathrm{cm}^2$ fluence, V$_\mathrm{OV}$ = 1 V) per-bar time resolution (Addesa et al., 15 Apr 2025, Abbott et al., 2021, Addesa et al., 2024).
• The timing error is modeled as:

$$\sigma_t^\mathrm{BTL} \approx \sqrt{\sigma_\mathrm{elec}^2 + \sigma_\mathrm{photo-stat}^2 + \sigma_\mathrm{DCR}^2 + \sigma_\mathrm{clock}^2}$$

with main contributions from electronics noise, photon counting statistics, SiPM dark-count-induced jitter, and clock distribution (Addesa et al., 2024).

• SiPM dark-count rates rise up to 20 GHz after $2 \times 10^{14}\,n_\mathrm{eq}/\mathrm{cm}^2$; power and noise are kept within system budget using bias reduction and in-situ annealing at $+60\,^\circ$C (Addesa et al., 15 Apr 2025, Addesa et al., 2024).

Electronics (CMS):

• The TOFHIR2 ASIC provides per-channel preamplification, differential leading-edge discrimination (DLED), a three-threshold trigger, dual-ended analog buffering, and multi-level digitization (TAC, QAC, 10-bit ADC, on-chip TDC with 11 ps binning) (Albuquerque et al., 2024).
• Operational at up to 2.5 MHz/channel with no rate-dependent time resolution loss.

STOPGAP (Belle II):

• Targeting $\sigma_\mathrm{sensor} \lesssim 50$ ps, achieved with monolithic CMOS MAPS using in-pixel amplifier/discriminator/TDC (bin $\lesssim$ 20–30 ps), power consumption $O(100)\,\mathrm{mW/cm}^2$ (Hartbrich et al., 2022).

Calibration protocols in both design lines include per-channel time and amplitude correction (for time-walk), reference clock loop-back, and dedicated light-injection or minimum-ionizing-track scans. Quality assurance achieves $\Delta\sigma_t < 2$ ps across module positions (Addesa et al., 15 Apr 2025).

4. Radiation Tolerance and Thermal Management

CMS BTL:

• LYSO:Ce remains effectively radiation hard up to $2 \times 10^{14}\,n_\mathrm{eq}/\mathrm{cm}^2$ at $T < -30\,^\circ$C (Dutta, 2018, Addesa et al., 2024).
• SiPMs exhibit increasing dark count and degraded gain after irradiation, managed via cooling to $-45\,^\circ$C, operational overvoltage adjustments, and periodic annealing (Addesa et al., 2024, Addesa et al., 15 Apr 2025).
• The TOFHIR2 ASIC demonstrates resilience to total ionizing dose up to 7 Mrad, with all performance parameters returning to nominal after annealing cycles; single event upset rates at HL-LHC are mitigated with triple-modular redundancy (Albuquerque et al., 2024).

STOPGAP (Belle II):

• The expected non-ionizing energy loss is three orders of magnitude below MAPS and ASIC technology failure thresholds; radiation levels ($$\mathrm{NIEL} \sim 1 \times 10^{9}\,n_\mathrm{eq}/\mathrm{cm}^2$$, TID $\sim$ 0.16 kRad) are negligible compared to the MAPS qualification range ($>1 \times 10^{13}\,n_\mathrm{eq}/\mathrm{cm}^2$, $>1$ Mrad) (Hartbrich et al., 2022).

Thermal control in both systems employs microchannel CO₂ or ethanol circulation, with finite-element analysis (FEA) confirming sensor temperature variations $\Delta T < 10\,\mathrm{K}$ (Hartbrich et al., 2022, Palluotto, 18 Jan 2026, Addesa et al., 15 Apr 2025).

5. Performance Impact and Physics Reach

CMS BTL:

• 30–60 ps timing reduces vertex merging rates from $15\% \rightarrow 1\%$ under 200-pileup conditions, cuts track-to-vertex misassociation by a factor of two, and substantially improves lepton/photon isolation, $b$-tagging, and missing $E_T$ performance (Dutta, 2018, Palluotto, 18 Jan 2026).
• Enables time-of-flight separation of heavy states and sensitivity to long-lived particle signatures.
• Uniformity across the entire active volume is within 2 ps; spatial resolutions of a few millimeters along bars and sub-millimeter across bars are realized (Abbott et al., 2021).
• For the decay $\phi \rightarrow K^+ K^-$, the STOPGAP S/N=0.79 compared to S/N=0.37 for the original TOP system, more than doubling statistical power for hadronic-physics analyses at Belle II (Hartbrich et al., 2022).

STOPGAP (Belle II):

• Simulation of $\Upsilon(4S)\rightarrow B\bar{B}$ events shows $\epsilon_K > 90\%$ up to $p\lesssim 2$ GeV/c for kaons, with $\pi\rightarrow K$ misID below 5% and full recovery of the problematic acceptance gaps (Hartbrich et al., 2022).

6. Optimization and Production Status

Broad R&D benchmarks SiPM cell sizes (15–30 $\mu$m), crystal bar thicknesses (2.4–3.75 mm), and irradiation campaigns up to $2 \times 10^{14}\,n_\mathrm{eq}/\mathrm{cm}^2$ (Addesa et al., 2024). The CMS design has converged on 25 $\mu$m pitch SiPMs with T1 (3.75 mm) bars as optimal, balancing photon detection efficiency, gain, and dark-count/power at both startup and end-of-life (Addesa et al., 2024). BTL modules are in full production, with over 50% assembled and system-level QA in progress (Palluotto, 18 Jan 2026).

Recommendations for further R&D on the STOPGAP BTL include a 3–5 year fast timing MAPS program in 65 nm HV-CMOS, system integration prototyping, and the development of on-chip TDC and sparsification logic (Hartbrich et al., 2022).

7. Future Prospects and Extensions

The BTL architecture is applicable beyond HL-LHC and Belle II. Emerging monolithic CMOS MAPS and hybrid SiPM/crystal solutions offer pathways to extend 4D tracking and trigger-level timing into future collider environments and heavy ion physics. STOPGAP's demonstration of ultra-thin, high-granularity timing modules within severe space constraints provides a blueprint for covering PID acceptance losses or adding TOF-based trigger legs in other systems (Hartbrich et al., 2022).

Advanced on-chip digital logic, multi-level analytical methods for timing, and continual irradiation resilience studies remain focal areas for future large-area fast-timing detectors in high-energy physics.