ALICE FIT: Fast Interaction Trigger

• ALICE FIT is a forward detector system that uses hybrid Cherenkov and scintillator technologies to achieve precise timing with sub-50ps resolution and robust trigger capabilities.
• It delivers real-time collision time, online luminosity, and event-plane measurements using advanced dual-gain front-end electronics and scalable readout systems.
• The design supports high operational rates with >90% efficiency in pp and near-unity in Pb–Pb collisions while maintaining minimal dead-time and reliable background suppression.

The ALICE Fast Interaction Trigger (FIT) is the primary forward detector system in the ALICE experiment at the CERN Large Hadron Collider (LHC) for Run 3 and Run 4. FIT provides minimum-bias and centrality triggers, precision collision-time reference, online luminosity monitoring, charged-particle multiplicity measurement, reaction-plane orientation, event-plane reconstruction, and background rejection across proton-proton and heavy-ion collisions. FIT's hybrid design integrates Cherenkov and scintillator detector technologies, high-speed front-end electronics, and scalable readout and control systems, sustaining rates up to 1 MHz (pp) and 50 kHz (Pb–Pb) with sub-50 ps timing precision and flexible trigger logic (Maevskaya, 2018, Melikyan, 14 Oct 2024, Maevskaya, 2020, Roslon, 7 Mar 2025, Roslon, 8 Jan 2025, Roslon, 18 Mar 2025, Ferretti, 2022, Mermer et al., 21 Nov 2025).

1. Physics Motivation and Functional Requirements

FIT was developed to address the stringent physics and operational requirements imposed by increased instantaneous luminosity and interaction rates of the LHC in Run 3/4. Primary scientific drivers include:

• Minimum-bias trigger efficiency greater than 90% for pp collisions and near-unity for Pb–Pb, enabling unbiased event selection across luminosity and multiplicity ranges (Maevskaya, 2018, Melikyan, 14 Oct 2024).
• Collision time determination with $\sigma_t < 50\,\mathrm{ps}$, required for preserving Time-Of-Flight (TOF) PID performance and providing a global clock for the ALICE experiment (Maevskaya, 2018, Maevskaya, 2020).
• Online luminosity measurement with a few-percent precision for both pp and Pb–Pb systems, supporting real-time accelerator feedback and normalization of physical yields (Melikyan, 14 Oct 2024, Mermer et al., 21 Nov 2025).
• Charged-particle multiplicity, centrality, and event-plane measurement in the forward region, underpinning centrality percentile assignment, reaction-plane orientation for flow analyses, and cross-section normalization (Maevskaya, 2018, Maevskaya, 2020).
• Sustained operation at high rates: up to $1\,\mathrm{MHz}$ in pp and $50\,\mathrm{kHz}$ in Pb–Pb, with a total trigger latency below $425\,\mathrm{ns}$, matching the upgraded ALICE readout chain (Melikyan, 14 Oct 2024, Maevskaya, 2020, Ferretti, 2022).

These requirements drove the transition from separate Run 2 forward detectors (T0, V0, FMD, AD) to the fully integrated FIT system in Run 3 (Maevskaya, 2020).

2. System Architecture and Detector Technologies

FIT employs a composite architecture involving three functionally distinct but synergistic subsystems arrayed around the interaction point (IP):

Subsystem	Sensor Type	z-Position (m)	$\eta$ Coverage	Primary Roles	Timing Resolution (per module)
FT0-A/C	Quartz Cherenkov radiators + MCP-PMTs	+3.5 (A), –0.8 (C)	4.2–5.0	Precision $t_0$, MB/centrality triggers	$\sim$17 ps (pp); 4.4 ps (Pb–Pb)
FV0	Plastic scintillator disk + clear fibers	–3.5 (C)	2.2–5.1	Multiplicity, centrality, event plane, MB	$\sim$200 ps
FDD-A/C	Scintillator tiles + WLS bars + PMTs	+17 (A), –19.5 (C)	up to 6.9	Diffractive/UPC tag, background veto	$\sim$150 ps

• FT0: Each array comprises fused-silica radiator bars, mirror-coated and directly optically coupled to multi-anode MCP-PMTs (Planacon XP85012/FIT-Q), optimized for high rate, high gain, low transit time spread, and radiation tolerance (Maevskaya, 2018, Melikyan, 14 Oct 2024). FT0 operates as the primary $t_0$ source for TOF PID.
• FV0: Large-area EJ-204 plastic scintillator segmented into concentric rings and azimuthal sectors, using $\sim$50,000 clear PMMA fibers for direct fast light transfer to fine-mesh PMTs. This avoids wavelength-shifting losses and preserves dynamic range and timing for centrality/multiplicity (Melikyan, 14 Oct 2024, Maevskaya, 2020).
• FDD: Very-forward double-layer arrays of fast scintillator tiles targeting single- and double-diffractive processes, beam–gas rejection, and background vetoes up to $|\eta|\sim7$ (Melikyan, 14 Oct 2024, Maevskaya, 2020, Roslon, 7 Mar 2025).

All subsystems integrate custom front-end boards providing amplification, discrimination, and digitization (TDC/ADC), with low jitter clocking and data serialization via GBT or IPbus to the Common Readout Unit (CRU) and the Detector Control System (DCS) (Roslon, 8 Jan 2025, Roslon, 7 Mar 2025).

3. Electronics, Trigger Logic, and Readout Integration

• Front-End Electronics (FEE): Each FEE module executes pre-amplification, constant-fraction discrimination (CFD), digitization, and FPGA-based trigger logic. Advanced dual-gain amplification is employed on FV0/FDD to extend dynamic range up to $\sim$1500 MIPs (previously $\sim$100 MIPs), enabling high-fidelity centrality and event-plane measurements in extreme-central Pb–Pb (Roslon, 7 Mar 2025, Maevskaya, 2018).
• Trigger and Clock Module (TCM): Collects timing and amplitude primitives, applies hardware-based trigger algorithms (minimum-bias, vertex, centrality, background veto), and distributes the low-jitter global clock. All FIT triggers are delivered to the ALICE Central Trigger Processor (CTP) within a global latency budget of $<425\,\mathrm{ns}$ (Maevskaya, 2020, Melikyan, 14 Oct 2024).
• Data Flow: Raw and processed data are serialized over GBT or IPbus, aggregated via CRU into the O2 framework, and supplied to the Event Processing Nodes (EPNs) for both online calibration and offline asynchronous reconstruction (Roslon, 8 Jan 2025, Maevskaya, 2020).
• Detector Control System (DCS): The DCS stack employs WinCC OA for supervisory control, using the ALFRED/FRED framework for standardized SCADA integration. The FIT-specific “IPbus-ALF” approach provides atomic, high-reliability register access across firmware generations without resynthesis, resulting in reduced operator intervention and improved remote control (Roslon, 8 Jan 2025, Roslon, 7 Mar 2025, Mermer et al., 21 Nov 2025).

4. Trigger Algorithms, Online Functionality, and Data Processing

• Minimum-Bias (MB) Trigger: Defined by logical coincident signals from FT0-A and FT0-C above dynamically set thresholds within a sub-nanosecond window; offers $>$90% efficiency in pp and nearly full coverage in central Pb–Pb (Melikyan, 14 Oct 2024, Maevskaya, 2020, Maevskaya, 2018).
• Centrality and Vertex Triggers: Built from multiplicity and amplitude sums in FV0/FT0, mapped via pre-fitted Glauber and Negative Binomial Distribution (NBD) models to centrality percentiles. The centroid is calculated from the mean times on the two FT0 sides (Maevskaya, 2020, Roslon, 7 Mar 2025).
• Event-Plane Reconstruction: Utilizes the azimuthal segmentation of FV0 (eight segments/ring) and FT0 modules, forming event flow vectors $Q_n = \Sigma_i w_i\,e^{in\phi_i}$ and resolving event-plane angles $\Psi_n = \arg Q_n / n$ (Maevskaya, 2020).
• Background Suppression: FDD anti-coincidences and tight timing cuts are applied to reject beam–gas, late background, and ultra-peripheral events (Melikyan, 14 Oct 2024, Maevskaya, 2020).
• Real-Time Data Flow: All FIT hit and trigger primitives are timestamped, clustered, and flagged by EPNs and are part of the synchronous and asynchronous O2 data-processing pipeline, supporting GPU-accelerated calibration and full reconstruction (Maevskaya, 2020).

5. Measured Performance and Operational Experience

Extensive simulation, beam-test, and operational data demonstrate that FIT achieves and exceeds the Run 3/4 requirements:

Metric	Value (pp)	Value (Pb–Pb)	Comments
Minimum-bias trigger efficiency	$>90\%$	$\sim99.8\%$	$$\varepsilon_\mathrm{MB}=N_\mathrm{triggered}/N_\mathrm{minBias}$$ (Maevskaya, 2018, Melikyan, 14 Oct 2024)
Collision time resolution	$17\,\mathrm{ps}$	$4.4\,\mathrm{ps}$	FT0 (measured, Run 3/4) (Melikyan, 14 Oct 2024, Roslon, 7 Mar 2025)
Latency	$<425\,\mathrm{ns}$	$<425\,\mathrm{ns}$	End-to-end hardware, meets CTP/O2 integration (Maevskaya, 2020)
Centrality resolution (0–10%)	$<1\%$	$<1\%$	FV0 + FT0, Glauber-constrained (Maevskaya, 2018, Maevskaya, 2020)
Dynamic range (scintillator chains)	$\sim64$ dB		Dual-gain FEE (Roslon, 7 Mar 2025)
FEE rate capability	$1.2$ MHz	$50$ kHz	No dead-time at design rates (Roslon, 18 Mar 2025, Maevskaya, 2018)

• Dead time: Effectively zero in between bunch crossings; operation demonstrated at bunch-by-bunch (40 MHz) level (Melikyan, 14 Oct 2024).
• Aging: MCP-PMTs have absorbed $>1\,\mathrm{C}/\mathrm{cm}^2$ anode charge, with gain recovery via HV adjustment; radiation tolerance verified to $0.5\,\mathrm{MRad}$ (Melikyan, 14 Oct 2024).
• Long-term uptime: FIT delivered stable Level-0 trigger and luminosity/timing reference information with $>99\%$ uptime during 2022–2024 ALICE operation (Roslon, 18 Mar 2025).

6. Control, Diagnostics, and Operator Training

• Detector Control System (DCS): ALFRED/FRED-based topology with atomic command sequencing, DIM-based telemetry, and finite-state machine (FSM) support across the FIT FEE and SCADA layers, yielding improved operability and error recovery (Roslon, 8 Jan 2025, Roslon, 7 Mar 2025, Mermer et al., 21 Nov 2025).
• Human-in-the-Loop Integration: A dedicated FIT coaching station with hardware and software replica of production DCS trains on-call shifters, reducing error rates and mean alarm response time by 30–40%, and enlarging the expert pool (Roslon, 18 Mar 2025).
• AI Support: Integration of AI-based support assistant employing LLMs, Retrieval-Augmented Generation pipelines, and policy guardrails, offering alarm diagnosis, corrective-action suggestions, and accelerated troubleshooting directly within SCADA operator GUIs (Mermer et al., 21 Nov 2025).
• Quantitative impacts: Operator training period reduced from $\sim$2 months to $\sim$2 weeks; subsystem downtime due to operator error reduced by $\sim$30% (Roslon, 18 Mar 2025).

7. Upgrades, Novel Features, and Outlook

• Front-End Enhancement: Adoption of dual-gain amplifiers, improved CFD modules, and pile-up tagging via pulse-shape analysis are expanding the dynamic range and time-walk performance, facilitating robust operation under extreme rates and multiplicity (Roslon, 7 Mar 2025).
• DCS Evolution: Migration to ALFRED-based control for IPbus-only FEE, enabling fully atomic operations, stateful error recovery, and plug-and-play extensibility (Roslon, 8 Jan 2025, Roslon, 7 Mar 2025).
• Environmental Innovations: High-rate, radiation-hard ALD-coated MCP-PMTs; direct optical coupling and novel mechanical tolerancing under the constraints of high stray fields ($0.45\,\mathrm{T}$), high dose, and severe envelope restrictions (Melikyan, 14 Oct 2024).
• Future Developments: Proposals for Run 5 forward detector arrays targeting $|\eta|\sim7$, with latency $<25\,\mathrm{ns}$, time resolution $<20\,\mathrm{ps}$, and dynamic range $>2000\,\mathrm{MIPs}$ (Roslon, 7 Mar 2025). Continuous firmware/software upgrade cycles are planned, including integration of direct CRU IPbus gateways and advanced online calibration.

FIT's integration of hybrid detection and readout, flexible trigger logic, low-latency synchronization, and modern control and diagnostics positions it as the front-door system for ALICE’s high-rate, high-precision heavy-ion and rare-probe physics in the LHC’s high-luminosity era (Maevskaya, 2020, Melikyan, 14 Oct 2024, Maevskaya, 2018, Roslon, 7 Mar 2025, Roslon, 8 Jan 2025, Roslon, 18 Mar 2025, Mermer et al., 21 Nov 2025).