betaFeedback

WorksPhysics and Astronomy

Wire-by-Wire Tracking Efficiency Plots: A New Diagnostic for the Belle~II Central Drift Chamber

Abstract

Large detectors are often monitored at the channel level (drift time, collected charge, and hit maps), which validates hardware but not tracking performance. A wire-by-wire tracking efficiency diagnostic is presented for the Belle~II Central Drift Chamber~(CDC). The method is directly analogous to the extrapolation-based efficiency measurement standard in resistive-plate-chamber~(RPC) stacks developed for the India-based Neutrino Observatory~(INO). A reference track (helix) is extrapolated to each wire layer; the fraction of crossings that contain an associated hit defines the per-wire efficiency. Implemented in the Belle~II Data Quality Monitoring~(DQM) framework and validated on Monte Carlo simulation with controlled dead-wire conditions, the method reveals localised tracking failures that are invisible to conventional channel-level diagnostics. The resulting plots provide direct feedback for run selection, operations, and long-term ageing studies.

Highlights

  • Wire-by-wire tracking efficiency for the Belle II CDC, computed by extrapolating helices to each layer.

  • Method adapted from the standard RPC diagnostic developed for the India-based Neutrino Observatory.

  • Reveals localised failures invisible to channel-level monitoring.

  • Integrated into Belle II DQM; supports run selection and long-term ageing studies.

Introduction

Large detector systems are routinely treated as black boxes: if aggregate performance metrics look reasonable, the internals are assumed to be fine. This assumption breaks down in two related ways. First, a localised hardware failure can be absorbed into the global acceptance, leaving no clear signature in standard monitoring plots. Second, in multi-detector systems one sub-detector can silently compensate for another, creating false confidence in overall performance while leaving the root cause undetected.

Motivation for per-component diagnostics comes from the India-based Neutrino Observatory (INO), whose proposed Iron Calorimeter (ICAL) called for approximately 28,800 glass Resistive Plate Chambers of 2 m × 2 m area [3]. At a target lifetime of roughly twenty years, even a small fraction of failed modules would carry measurable physics impact. This motivated systematic quality assurance at every level, from individual gas-gap integrity tests on prototypes [5] to layer-by-layer efficiency monitoring on the prototype stacks.

The Belle II experiment at the SuperKEKB e+e−e^{+}e^{-} collider applies the same per-component approach. The Central Drift Chamber (CDC) is its primary charged-particle tracking detector. We show that the extrapolation-based efficiency method, standard in RPC stacks, transfers directly to CDC wires and reveals failures invisible to conventional channel-level monitoring.

Extrapolation-Based Efficiency in RPC Stacks

An RPC detects ionising radiation through the avalanche it produces in the gas gap between two resistive plates (Figure 1). The standard diagnostic in a multi-layer RPC stack is extrapolation-based efficiency measurement. A reference track is reconstructed from all layers except the one under test. That track is extrapolated to the test layer; the strip it crosses is the expected active element, and the layer efficiency is ε=Nobserved/Nexpected\varepsilon= N_{\mathrm{observed}}/N_{\mathrm{expected}}, where NobservedN_{\mathrm{observed}} counts crossings for which that strip recorded a hit and NexpectedN_{\mathrm{expected}} is the total number of crossings. The efficiency, together with the strip hit counts and the cluster multiplicity (Figure 2), forms a set of three plots that answer three distinct questions:

Image not extracted

Figure 1. Schematic of a resistive plate chamber traversed by ionising radiation, illustrating the avalanche in the gas gap and signal pickup on the readout strips.

Image not extracted

Figure 2. Per-layer efficiency (a), strip hit count (b), and cluster multiplicity (c) for a representative INO-ICAL RPC stack layer. Each plot answers a specific diagnostic question.

  1. whether the layer detected the particle (efficiency)

  2. whether individual strips are alive (hit count map)

  3. whether the gain is at the right operating point (cluster multiplicity)

The Belle II Central Drift Chamber

The same extrapolation principle applies to the Belle II CDC, where each sense wire plays the role of an RPC strip.

Detector overview

The Belle II CDC [1] is a cylindrical drift chamber filling the radial range from ∼16\sim16 cm to ∼113\sim113 cm. It consists of 56 wire layers organised in 9 superlayers (one innermost axial superlayer followed by alternating stereo and axial superlayers), providing up to 56 hits per track. The CDC is supplemented by the four-layer Silicon Vertex Detector (SVD) and the two-layer Pixel Detector (PXD). Together they provide the primary vertex measurement and supply seed tracks to the CDC pattern recognition algorithms (Figure 3).

Image not extracted

Figure 3. Left: XY cross-section of the Belle II CDC showing the 56 concentric wire layers grouped into superlayers, with circular boundaries marking the transition between axial and stereo superlayers. Right: drift-cell field map illustrating the hexagonal cell geometry and the field-wire arrangement around each sense wire.

Conventional wire health monitoring

Standard CDC DQM monitors each wire through:

  • Hit maps: the number of times each wire fires per unit luminosity (Figure 4);

  • Drift time distributions: sensitive to gas conditions and high-voltage stability;

  • Collected charge: sensitive to gas gain and ageing.

Image not extracted

Figure 4. CDC wire hit map appearing globally healthy despite a localised tracking failure, illustrating why channel-level monitoring alone is insufficient.

Each quantity confirms that the hardware is functioning within limits, but none directly addresses whether the detector is finding the tracks it should.

Why channel-level monitoring is insufficient

The Belle II tracking chain is designed for robustness. If the CDC does not find a track, the SVD attempts independent reconstruction; the SVDtoCDCCKF Combinatorial Kalman Filter [2] then tries to attach CDC hits to the SVD seed. This resilience is a strength for physics, but a weakness for diagnostics. A localised CDC inefficiency is partially absorbed by SVD, suppressing any efficiency drop and leaving channel-level plots uninformative.

Wire-by-Wire Efficiency Method

The efficiency is computed from track helices provided by the standard Belle II reconstruction. As a charged particle traverses a CDC wire layer, the helix is extrapolated to identify the sense wire nearest to the crossing point. That wire is the expected hit: it should have recorded a signal if functioning. If the reconstructed track has an associated hit on that wire, the crossing is counted as observed. Unlike the strict RPC procedure, the layer under inspection is not excluded from the fit: with up to 56 layers per track, the gain in fit independence is marginal and the benefit does not justify the added complexity. The result is accumulated in a two-dimensional histogram of wire index vs. layer index (or equivalently vs. ϕ\phi), giving a geometric view of the full CDC acceptance. Because the efficiency is defined relative to the track, the diagnostic needs no external reference detector or baseline run, making it suitable for early failure detection. The DQM module runs both online and offline, accumulating statistics across all trigger types.

Simulation Setup

Software and track selection

Tracks are selected from the standard Belle II reconstruction output via the publicly available basf2 framework [4]. The results presented here use a Monte Carlo sample of e+e−→B0Bˉ0e^{+}e^{-} \to B^{0}\bar{B}^{0} events without beam background. Beam background is omitted because the method requires only reconstructed tracks; its validation is independent of the underlying event conditions.

Dead-wire conditions

The simulation payload includes scattered genuine dead wires together with two injected full-depth board failures, one in superlayer 2 (inner, low radius) and one in superlayer 4 (outer, high radius), each spanning all layers of the respective superlayer. This two-radius design quantifies how severely the track quality estimator penalises a gap closer to the interaction point.

Results and Discussion

Efficiency map and localised failure

Figure 5 shows the wire-by-wire efficiency canvas for the simulation described in § 6. The efficiency map (c) shows drops at two azimuthal locations: a pronounced drop near ϕ≈+135∘\phi\approx+135^\circ from the disabled SL2 board (inner, closer to the interaction point), and a less severe drop near ϕ≈+42∘\phi\approx+42^\circ from the disabled SL4 board (outer). The one-dimensional distribution (d) reveals that 83.29% of wires have efficiency above 0.72, while 14.11% fall in the intermediate range (0.08<ε<0.720.08 < \varepsilon< 0.72), and 2.59% have ε<0.08\varepsilon< 0.08 (effectively dead); the mean wire efficiency is 77.9%. The 2.59% dead-wire fraction corresponds to 371 wires out of 14 336, consistent with the injected dead-wire conditions. The per-wire efficiency values allow run-by-run selection: runs where specific ϕ\phi regions fall below threshold can be excluded or down-weighted rather than rejected wholesale. The thresholds are indicative and can be refined with operational experience.

Image not extracted

Figure 5. Wire-by-wire tracking efficiency canvas for a Monte Carlo simulation of e+e−→B0Bˉ0e^{+}e^{-}\to B^{0}\bar{B}^{0} events (Belle II simulation, no beam background). (a) Observed wire hits. (b) Expected wire hits from track extrapolation. (c) Wire efficiency map showing clear drops at the locations of the two simulated dead-board regions. (d) One-dimensional efficiency distribution cleanly separating active from masked wires.

Investigation

Both drops also appear in the ϕ\phi distribution of the standard CDC DQM module, but that plot alone provides no further explanation. When a full axial superlayer is missing, CDC tracking produces incomplete tracks; the MVA-based track quality estimator removes these as poorly formed. SVD standalone tracking then recovers the particle, and the SVDtoCDCCKF algorithm [2] attaches CDC hits only up to the gap, unable to bridge the missing layer. Because the track does not reach the outer CDC layers, it fails to extrapolate to the Electromagnetic Calorimeter (ECL), preventing calorimeter association; this is particularly serious for electron and photon-related analyses. The efficiency map alone is therefore sufficient to localise the failure and reveal its downstream consequences. This makes it a direct input to operational decisions.

Operational impact

The efficiency histograms are now part of the Belle II CDC DQM. They provide three classes of information:

  • Immediate failure detection: a disabled board or dead high-voltage channel produces a drop over a contiguous region of wires, visible within a single run.

  • Run selection: analysts can exclude or down-weight runs where specific ϕ\phi regions show reduced efficiency, rather than applying a global run quality flag.

  • Ageing and long-term studies: the per-wire efficiency trend accumulated over many runs provides a sensitive probe of wire ageing, gas gain evolution, and detector degradation.

Summary

The extrapolation-based efficiency method, established in the INO RPC prototype programme [3, 5], transfers directly to drift-wire detectors. Applied wire by wire to the Belle II CDC, it exposes localised hardware failures invisible to hit-map monitoring, including their downstream effect on track quality rejection and calorimeter association. The method is implemented in the Belle II DQM framework [4] and provides immediate, wire-resolved feedback for operations and physics analysis. The approach generalises to any tracking detector with individually instrumented channels, enabling reconstruction-level quality monitoring beyond channel-level plots. The failure mode also motivates retraining the MVA-based CDC track quality estimator to handle partially missing axial superlayers, improving track retention under realistic detector conditions.

CRediT authorship contribution statement

Suryanarayan Mondal: Conceptualization, Methodology, Formal analysis, Visualization, Writing – original draft.

References

  1. [1]Abe, T., et al., 2010. Belle II technical design report. arXiv arXiv:1011.0352.arxiv.org/abs/1011.0352
  2. [2]Bertacchi, V., et al., 2021. Track finding at Belle II. Computer Physics Communications 259, 107610. doi:10.1016/j.cpc.2020.107610, arXiv:2003.12466.DOI
  3. [3]ICAL Collaboration, 2017. Physics potential of the ICAL detector at the India-based Neutrino Observatory: a comprehensive study. Pramana 88, 79. doi:10.1007/s12043-017-1373-4.
  4. [4]Kuhn, T., et al., 2019. The Belle II core software. Computing and Software for Big Science 3, 1. doi:10.1007/s41781-018-0017-9.
  5. [5]Mondal, S., et al., 2019. Leak test of Resistive Plate Chamber gap by monitoring absolute pressure. Journal of Instrumentation 14, P04009. doi:10.1088/1748-0221/14/04/P04009, arXiv:1812.00277.DOI

Paper details

vlnc:W_wgup8sxs
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents