Design, fabrication and characterization of 8x9 n-type silicon pad array for sampling calorimetry

Mohamad Sawan, Ganesh Jagannath Tambave, J.-l. Bouly, O. Bourrion, T. Chujo, Amita Das, M. Inaba, Vipul Kashyap, C. Krug, R. Laha, C. Loizides, B. Mohanty, N. K. Mondal, N. Ponchant, S. Prasad, R. Singh, D. TourrèsJournal of Instrumentation 2025 · DOI · arXiv

TL;DR

Researchers have successfully designed and fabricated a new 8x9 n-type silicon pad array for the upcoming FoCal upgrade of the ALICE detector at CERN. While the detectors show excellent initial performance—with 90% of pads exhibiting leakage currents below 10 nA—radiation hardness tests suggest that these n-type sensors may face significant challenges in high-fluence environments compared to p-type alternatives.

What Problem Does This Paper Address?

The ALICE experiment at the Large Hadron Collider (LHC) is preparing for "Run 4" (2029–2034), which requires a new Forward Calorimeter (FoCal). To study the dynamics of the Quark-Gluon Plasma at extremely small Bjorken scaling factors ($x$), the FoCal needs high-granularity silicon detectors capable of precise tracking and calorimetry.

While silicon detectors are standard in high-energy physics, this paper addresses the specific challenge of developing large-area (8x9) n-type silicon pad arrays that are compatible with the HGCROC readout chip. Notably, this work represents the first time such large-area detectors have been fabricated domestically in India, providing a pathway for local production of high-tech experimental components.

Methods

The authors employed a full-lifecycle approach to detector development:

  • Simulation: They used Technology Computer-Aided Design (TCAD) software (Silvaco’s Athena and Atlas modules) to model the doping profiles and predict electrical behavior, such as breakdown voltage and full depletion voltage (FDV).
  • Fabrication: The detectors were manufactured on 6-inch, 325 $\mu$m thick, high-resistivity n-type silicon wafers at Bharat Electronics Limited (BEL) in Bangalore, India, using standard bipolar fabrication technology.
  • Characterization:
    • Electrical: Current-voltage (IV) and capacitance-voltage (CV) tests were performed using a Keithley 4200A-SCS parameter analyzer.
  • Functional: The arrays were integrated with the HGCROCv2 readout chip and tested using a blue LED (for channel verification) and a $^{90}$Sr beta source (to simulate particle response).
  • Radiation: Samples were exposed to a fast neutron beam at the RANS facility in RIKEN, Japan, to test their durability against the harsh conditions of the LHC.

Datasets

As an experimental physics paper, the "datasets" consist of the characterization measurements:

  • Electrical Maps: IV and CV data for 25 fabricated wafers.
  • Signal Profiles: ADC counts vs. time for all 72 pads under LED and beta source illumination.
  • Radiation Logs: Leakage current measurements recorded over 55 days post-irradiation at various neutron fluence levels (up to $5 \times 10^{13} \text{ 1 MeV n}_{eq}/\text{cm}^2$).

Key Insights and Ideas

  • Domestic Success: The successful fabrication at BEL proves that high-quality, large-area silicon sensors can be produced outside the traditional major hubs, meeting the stringent requirements for CERN upgrades.
  • Geometry Matching: The 8x9 array (72 pads) was specifically designed to match the 72-channel HGCROCv2 chip, allowing for a compact "flip-chip" style integration that reduces the overall detector footprint.
  • The N-Type Vulnerability: A critical "aha moment" came during radiation testing. The researchers observed that high neutron fluence likely caused "substrate-type inversion," where the n-type material begins to behave like p-type due to radiation-induced defects. This led to a three-order-of-magnitude increase in leakage current.
  • Temperature Sensitivity: The study confirmed that maintaining the detector below 20°C is vital for keeping leakage currents manageable, especially after the sensors have sustained radiation damage.

Key Results

  • Leakage Current: 90% of the pads showed leakage currents $< 10$ nA at the full depletion voltage, indicating high fabrication quality.
  • Voltage Tolerance: While simulations predicted a breakdown at 1000 V, physical tests confirmed stable operation up to at least 500 V.
  • Depletion: The full depletion voltage was found to be approximately 50 V, with a pad capacitance of 35 pF, aligning closely with TCAD simulations.
  • Particle Response: The $^{90}$Sr tests showed a clear separation between the "pedestal" (noise) and the electron signal, with a Minimum Ionizing Particle (MIP) peak around 25 ADC counts.
  • Radiation Impact: After a fluence of $5 \times 10^{13} \text{ n}_{eq}/\text{cm}^2$, the leakage current jumped from nanoamps to tens of microamps.

Limitations and Future Work

The primary limitation identified is the radiation hardness of the n-type substrate. The significant increase in leakage current and the potential for type inversion make these specific n-type detectors unsuitable for the highest-radiation zones of the FoCal.

Future Work: The authors explicitly recommend moving toward p-type substrate-based detectors for the final FoCal construction. P-type silicon is generally more resistant to the specific type of radiation damage (type inversion) observed here. Future studies will likely focus on characterizing p-type arrays using the same 8x9 geometry and fabrication pipeline established in this paper.

Summary updated

Abstract

Abstract This paper reports the development and testing of n-type silicon pad array detectors intended for the Forward Calorimeter (FoCal) detector, which is an upgrade of the ALICE detector at CERN, scheduled for data taking in Run 4 (2029–2034). The FoCal detector will consist of a hadronic and an electromagnetic calorimeter. The electromagnetic part is made up of alternating tungsten absorber and silicon detector layers, read out using the High Granularity Calorimeter Readout Chip (HGCROC). This paper covers the various processes involved in the design, fabrication and testing of the detector. The current vs voltage (IV) measurements showed that 90% of the pads in the detector had leakage current below 10 nA, and capacitance vs voltage (CV) measurements determined the full depletion voltage, consistent with the design simulations. The detector's performance was tested with an LED light source and a 90 Sr β - electron source. Furthermore, radiation harness studies were performed on 1 × 1 cm 2 pads with a neutron fluence of 5 × 10 13 1 MeV n eq /cm 2 .
BibTeX
@article{sawan2025design,
  author = {Mohamad Sawan and Ganesh Jagannath Tambave and J.-L. Bouly and O. Bourrion and T. Chujo and Amita Das and M. Inaba and Vipul Kashyap and C. Krug and R. Laha and C. Loizides and B. Mohanty and N. K. Mondal and N. Ponchant and S. Prasad and R. Singh and D. Tourrès},
  title = {{Design, fabrication and characterization of 8x9 n-type silicon pad array for sampling calorimetry}},
  journal = {Journal of Instrumentation},
  year = {2025},
  doi = {10.1088/1748-0221/20/05/p05007},
}

Loading related content...

Figure