Papers
Topics
Authors
Recent
2000 character limit reached

CLD & IDEA Detectors for FCC-ee

Updated 28 November 2025
  • CLD and IDEA Detectors are tracker-based solutions that integrate all-silicon and helium-based drift chamber technologies to meet FCC-ee’s stringent precision and flavor physics benchmarks.
  • They leverage innovative PID methodologies including Time-of-Flight, dE/dx, and cluster counting to achieve high spatial resolution and efficient particle discrimination across varying momentum ranges.
  • Quantitative benchmarks indicate that while CLD excels in low-momentum tracking with silicon-only approaches, IDEA’s combination of cluster counting and ToF delivers superior performance for extended momentum regimes.

The CLD and IDEA detector concepts constitute the principal tracker-based solutions developed for the precision and flavor physics program at the Future Circular Collider (FCC-ee). These detectors are distinguished by their approach to charged particle tracking, particle identification (PID), and their overall material and geometric configuration. CLD is an all-silicon architecture optimized for spatial precision and robust hit timing, whereas IDEA employs a very light helium-based drift chamber with cluster-counting readout, supplemented by silicon vertex and timing layers. Both concepts aim to satisfy the demanding FCC-ee flavor, electroweak, and Higgs physics benchmarks, with PID strategies that rely exclusively on tracker information, as neither include dedicated Cherenkov or RICH systems in their baseline configuration (Beck et al., 21 Nov 2025).

1. Detector Architectures and Tracking Subsystems

CLD (CLIC-like Detector):

  • All-silicon tracking system with an innermost layer of pixel vertex detectors near the interaction point, surrounded by silicon strip or pixel disks at larger radii.
  • Pixel single-point resolution of a few microns; per-pixel hit timing in the range O(10)–100 ps.
  • Material budget per layer: ∼0.3% X0X_0 (pixels), ∼1% X0X_0 (strips); outermost tracking radius \sim1.8 m.
  • Enclosed by a 2 T solenoidal field at the Z pole, upgradeable to 3 T at higher s\sqrt{s}.
  • No gas-based tracking layers (Azzi et al., 2021).

IDEA (Innovative Detector for e⁺e⁻ Accelerators):

  • Combines a silicon MAPS vertex detector with a large, ultra-low mass, full-stereo drift chamber (He-based, 90/10 He/iC₄H₁₀; thickness 1.6–5% X0X_0).
  • Cluster-counting readout in the drift chamber enables high-precision dNN/dx measurements for PID.
  • Outer silicon “wrapper” provides additional tracking anchors and fast timing (LGAD: \sim100 ps).
  • Tracking lever arm extended to r = 2 m; drift chamber single-hit resolution \sim100 μm.
  • Designed for high-precision momentum measurement, large tracking efficiency (>>99% for pTp_T>$0.2 GeV), and low multiple scattering (Elmetenawee et al., 2022, Ilg, 30 Oct 2025).

Summary Table: Principal Tracking Features

CLD (Silicon) IDEA (Drift + Si)
Tracking type Pixel+strip Si tracker Cluster-count drift chamber + Si
Max radius ∼1.8 m 2 m (drift chamber), 2.08 m (Si)
Point res. 3–7 μm (pixels), 7 μm (strips) 100 μm (DCH), 3–14 μm (Si)
Material 0.3–1.0% $X_0perlayer</td><td>1.65 per layer</td> <td>1.6–5% X_0(DCH),0.4 (DCH), 0.4% X_0(Si)</td></tr><tr><td>Timing</td><td>O(10100)ps(pixels)</td><td>1ns(Si),100200ps(wrapper)</td></tr></tbody></table></div><h2class=paperheadingid=trackerbasedparticleidentificationmethodologies>2.TrackerBasedParticleIdentificationMethodologies</h2><p><strong>TimeofFlight(ToF):</strong></p><ul><li>BothCLDandIDEAexploitToFinsiliconlayers:measure (Si)</td> </tr> <tr> <td>Timing</td> <td>O(10–100) ps (pixels)</td> <td>1 ns (Si), 100–200 ps (wrapper)</td> </tr> </tbody></table></div><h2 class='paper-heading' id='tracker-based-particle-identification-methodologies'>2. Tracker-Based Particle Identification Methodologies</h2> <p><strong>Time-of-Flight (ToF):</strong></p> <ul> <li>Both CLD and IDEA exploit ToF in silicon layers: measure \Delta tbetweenfirstandlastSihitsalongknown between first and last Si hits along known L;; \beta=v/c=L/(c \Delta t).</li><li>ToFseparationbetweenspecies:.</li> <li>ToF separation between species: \Delta t_{12} = (L/c)(1/\beta_1 - 1/\beta_2) \approx L/c \cdot (m_1^2 - m_2^2)/(2p^2)athigh at high p.</li><li>ToFresolutionpertrack.</li> <li>ToF resolution per track \sigma_\mathrm{ToF}resultsfromperlayertimingandtrackfit;O(30ps)isbenchmarked(<ahref="/papers/2511.17447"title=""rel="nofollow"dataturbo="false"class="assistantlink"xdataxtooltip.raw="">Becketal.,21Nov2025</a>).</li></ul><p><strong>SpecificIonization(dE/dx)inSilicon:</strong></p><ul><li>Sumofperhitenergydepositions results from per-layer timing and track fit; O(30 ps) is benchmarked (<a href="/papers/2511.17447" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">Beck et al., 21 Nov 2025</a>).</li> </ul> <p><strong>Specific Ionization (dE/dx) in Silicon:</strong></p> <ul> <li>Sum of per-hit energy depositions E_ialongthetrack;PIDusestruncatedmeanorharmonicmeanestimators.</li><li>Maximumionizationdifference( along the track; PID uses truncated-mean or harmonic-mean estimators.</li> <li>Maximum ionization difference (\mathrm{\pi,\,K,\ p})atintermediate) at intermediate p( (\beta\gamma\approx15).</li><li>dE/dxresolution:–5).</li> <li>dE/dx resolution: \sigma(\mathrm{d}E/\mathrm{d}x)\simeq k/\sqrt{N_\text{hits}},with, with k\sim1520</ul><p><strong>ClusterCounting(dN/dx,IDEAonly):</strong></p><ul><li>Countsindividualprimaryionizationclusters–20% per hit (<a href="/papers/2511.17447" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">Beck et al., 21 Nov 2025</a>).</li> </ul> <p><strong>Cluster Counting (dN/dx, IDEA only):</strong></p> <ul> <li>Counts individual primary ionization clusters N_iperdriftcell,summedoverthetrack: per drift cell, summed over the track: dN/dx = (\sum_i N_i)/L_\text{track}.</li><li>ClusterPoissonstatisticsyieldnarrowerPIDpeaksthanLandaudominateddE/dx.</li><li>Clustercountingefficiency.</li> <li>Cluster Poisson statistics yield narrower PID peaks than Landau-dominated dE/dx.</li> <li>Cluster-counting efficiency \epsilon_{cc}entersresolution: enters resolution: \sigma(dN/dx)/\langle dN/dx\rangle\simeq1/\sqrt{\epsilon_{cc}\cdot N_\text{clusters}}(<ahref="/papers/2511.17447"title=""rel="nofollow"dataturbo="false"class="assistantlink"xdataxtooltip.raw="">Becketal.,21Nov2025</a>,<ahref="/papers/2211.12568"title=""rel="nofollow"dataturbo="false"class="assistantlink"xdataxtooltip.raw="">Elmetenaweeetal.,2022</a>).</li></ul><h2class=paperheadingid=quantitativepidperformancebenchmarks>3.QuantitativePIDPerformanceBenchmarks</h2><p>BenchmarkingagainstrepresentativeFCCeeflavor,raredecay,andjetflavortaggingtasksyieldsthefollowingresults(<ahref="/papers/2511.17447"title=""rel="nofollow"dataturbo="false"class="assistantlink"xdataxtooltip.raw="">Becketal.,21Nov2025</a>):</p><ul><li><strong>Low (<a href="/papers/2511.17447" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">Beck et al., 21 Nov 2025</a>, <a href="/papers/2211.12568" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">Elmetenawee et al., 2022</a>).</li> </ul> <h2 class='paper-heading' id='quantitative-pid-performance-benchmarks'>3. Quantitative PID Performance Benchmarks</h2> <p>Benchmarking against representative FCC-ee flavor, rare decay, and jet-flavor tagging tasks yields the following results (<a href="/papers/2511.17447" title="" rel="nofollow" data-turbo="false" class="assistant-link" x-data x-tooltip.raw="">Beck et al., 21 Nov 2025</a>):</p> <ul> <li><strong>Low-p( (0.31GeV) GeV) B_sflavortagging(samesideKaon):</strong><ul><li><em>CLD</em>(-flavor tagging (same-side Kaon):</strong> <ul> <li><em>CLD</em> (\sigma_\mathrm{ToF}=30ps,nodE/dx): ps, no dE/dx): \epsilon_\text{sig}\sim75%, ϵbkg\epsilon_\text{bkg}\sim$5%
  • IDEA (cluster counting, $\epsilon_{cc}=80\%):): \epsilon_\text{sig}\sim85%, ϵbkg\epsilon_\text{bkg}\sim$3%; with ToF=30 ps: $\epsilon_\text{bkg}\sim2%
  • Medium-pp ($1$–$5$ GeV) rare bsb\to s\ell\ell decays:
    • CLD: ToF 30 ps reduces contamination an order of magnitude below kinematics-only (\sim0.02%)
    • IDEA: Cluster counting (ϵcc=80%\epsilon_{cc}=80\%) yields \sim0.005%; ToF adds modest further gain
  • High-pp (10–50 GeV) Higgs ss-jet tagging:
    • CLD: Even with σToF=10\sigma_{\mathrm{ToF}}=10 ps, separation weak (ϵsig80%\epsilon_\text{sig}\sim80\%, ϵbkg50%\epsilon_\text{bkg}\sim50\%)
    • IDEA: Cluster counting (ϵcc=80%\epsilon_{cc}=80\%): ϵsig80%\epsilon_\text{sig}\sim80\%, ϵbkg25%\epsilon_\text{bkg}\sim25\%; with ToF 50 ps: ϵbkg\epsilon_\text{bkg} approaches 20%
  • Timing resolution below 30 ps improves suppression in background-limited rare decays, while cluster counting shows strong suppression across pp regimes, with only minor dependence on ϵcc\epsilon_{cc} down to 50%.

    4. Cluster Counting Technique in IDEA

    The drift chamber for IDEA is a 4 m long, 2 m outer radius device, filled with 90% He, 10% iso-butane, precisely engineered for minimal material (\sim1.6% X0X_0 in the barrel). Signal yields correspond to a cluster density Nc12.5N_c \simeq 12.5/cm (He–iC₄H₁₀), with single-hit spatial resolution of \sim100 μ\mum. Cluster counting directly timestamps individual primary ionization clusters, improving both dE/dx\mathrm{d}E/\mathrm{d}x and spatial resolution—empirically observed to yield %%%%70\sim71%%%% better dE/dx\mathrm{d}E/\mathrm{d}x resolution than the truncated-mean approach. For practical implementation, a cluster-counting efficiency ϵcc\epsilon_{cc} is included in simulations and learning, reflecting realistic digitization and noise (Elmetenawee et al., 2022, Ilg, 30 Oct 2025).

    Key equations:

    • S/N=Qc/σnS/N=Q_c/\sigma_n (cluster charge over noise)
    • σdE/dxΔE/NcL\sigma_{dE/dx}\simeq \Delta E/\sqrt{N_c L}
    • σx=vdσt\sigma_x=v^d \sigma_t (spatial from timing)
    • Occupancy is limited to <<2% per cell at Z-pole luminosity (Elmetenawee et al., 2022)

    5. Impact on Flavor and Precision Physics

    Tracker-based PID in both CLD and IDEA directly impacts the reach of flavor-physics measurements at FCC-ee across multiple observables (Beck et al., 21 Nov 2025, Azzi et al., 2021):

    • bb-flavor and ss-jet tagging benefit from suppression of misidentified hadrons, critical for same-sign Kaon counting and ss-jet enrichment.
    • Rare decay suppression in bsb\to s FCNC decays (notably Bs0K+Kμ+μB_s^0\rightarrow K^+K^-\mu^+\mu^-), where mass resolution and PID combine to limit background to sub-percent levels, especially when leveraging IDEA’s cluster counting.
    • Jet tagging at high-pp remains the principal limitation for tracker-based PID: both ToF and dE/dx in silicon lose discrimination power for π/K\pi/K beyond 5–10 GeV; the cluster-count drift chamber of IDEA outperforms these, but none reach the efficacy of dedicated Cherenkov approaches for this regime.

    Both detector concepts achieve per-mil–level systematic control on tracking and PID for heavy-flavor and rare-decay channels, enabling significant improvement over LEP and prior e+ee^+e^- collider results.

    6. Limitations and Prospects for Dedicated PID

    A critical limitation—shared by both CLD and IDEA baselines—is the inability to match the π/K\pi/K separation at high momenta that is provided by RICH or DIRC systems. For ultimate suppression (sub-percent backgrounds and systematic uncertainties on mis-ID) in jet-flavor and extreme low-pp domains, a dedicated PID such as a time-of-propagation or RICH detector would be required. Tracker-based PID as implemented in CLD and IDEA is however found to suffice for a large fraction of the FCC-ee flavor and rare-decay program, potentially allowing deferral of dedicated systems to later upgrades (“should be the subject of future paper”) (Beck et al., 21 Nov 2025).

    IDEA’s approach—ultra-light cluster-counting drift chamber plus ToF wrapper—provides robust PID over a wide pp-range, with cluster counting out-performing ToF alone in many scenarios and being relatively insensitive to realistic reductions in cluster-count efficiency (Beck et al., 21 Nov 2025, Elmetenawee et al., 2022). CLD’s silicon-only ToF+dE/dx is optimal at low-pp, moderate at medium pp, and ineffective at the multi-GeV scale relevant for hadronic Higgs decays.

    7. Comparative Evaluation and Quantitative Performance

    The following table summarizes key performance features for CLD and IDEA as evaluated in simulation benchmarks (Azzi et al., 2021):

    Feature CLD (All-Si) IDEA (Drift+Si)
    PID method ToF (30 ps), dE/dx Cluster-counting dE/dx, ToF (100 ps)
    π/K\pi/K sep. (p<5p<5 GeV) ToF/dE/dx (%%%%93s\sqrt{s}94%%%% low pp) %%%%96s\sqrt{s}97%%%% up to p30p\sim30 GeV
    π/K\pi/K sep. (p>10p>10 GeV) Low; ToF ineffective Moderate (dN/dx), no match to RICH
    Jet energy res. 25%/E25\%/\sqrt{E} 28%/E28\%/\sqrt{E}
    Mom. res. @50 GeV 0.20%0.20\% 0.15%0.15\%
    Tracking eff. >98%>98\% >98%>98\%

    Both CLD and IDEA deliver impact-parameter resolution σd0=ab/(psin3/2θ)\sigma_{d_0}=a\oplus b/(p\sin^{3/2}\theta) with a4a\sim46μ6\,\mum, approaching the FCC-ee benchmarks for flavor-tagging and rare-decay vertexing (Azzi et al., 2021).

    References

    Whiteboard

    Follow Topic

    Get notified by email when new papers are published related to CLD and IDEA Detectors.