Abstract: This study evaluates silicon and gallium arsenide (GaAs) pad sensors in a 5 GeV electron beam at the DESY-II facility. The sensors, featuring 500 µm thickness and $5 \times 5 \text{ mm}^2$ pads, are designed for compact electromagnetic sampling calorimeters. Innovative connectivity solutions—copper traces on Kapton for silicon and embedded aluminum traces for GaAs—allow for sensor planes thinner than 1 mm. Results demonstrate high signal-to-noise ratios, excellent response homogeneity, and manageable cross-talk, supporting their use in high-precision detectors like those for the LUXE experiment.
Introduction
In high-energy physics, electromagnetic calorimeters measure particle energy by initiating showers in dense absorber materials. A key performance metric is the Molière radius, which describes the transverse spread of these showers. To maintain a small Molière radius and thus high spatial resolution, the gaps between absorber plates must be minimized.
Traditional sensor mounting techniques, such as wire bonding, often require significant vertical space, which "dilutes" the calorimeter and increases the effective Molière radius. This paper investigates two novel sensor technologies and connectivity schemes designed to keep the total sensor plane thickness below 1 mm, ensuring maximum calorimeter compactness.
Sensor Technologies and Connectivity
The researchers compared two sensor types, both 500 µm thick with $5 \times 5 \text{ mm}^2$ pads.
Silicon (Si) Sensors
The silicon sensors utilize a standard $p+$ on $n$ substrate. The primary innovation is the readout connectivity: instead of bulky wire bonds, copper traces on a flexible Kapton foil are attached to the pads using conductive glue. This approach significantly reduces the vertical profile of the sensor assembly.
Gallium Arsenide (GaAs) Sensors
GaAs sensors, often preferred for their radiation hardness, were fabricated from high-resistivity single crystals. In this design, the readout traces are embedded directly on the sensor substrate within the narrow gaps between pads, further reducing the need for external components and maintaining a slim profile.
Experimental Methodology
The sensors were tested using 5 GeV electrons at the DESY-II test-beam facility. A pixel telescope provided precise tracking for every electron, while the readout was performed by the FLAME ASIC, a custom 32-channel chip designed in 130 nm CMOS technology. The entire system was orchestrated by a Trigger Logic Unit. The team used a modified Hough transform to align the sensors with the telescope tracks with micron-level precision. Performance was benchmarked against Geant4 Monte Carlo simulations to validate energy deposition and sensor geometry models.
Key Results and Insights
- Signal-to-Noise Ratio (SNR): Both sensors performed excellently, with an SNR of approximately 15 for Silicon and 17 for GaAs.
- Homogeneity: The Most Probable Value (MPV) of the signal was stable across the sensor surface, varying by only about 1% for Silicon.
- Edge Effects: GaAs showed a more pronounced signal drop (up to 40%) in the gaps between pads compared to Silicon, primarily because the inter-pad gaps in the GaAs prototype were significantly larger (300 µm vs 10 µm).
- The "Trace Effect": In the GaAs sensors, because the readout traces run between the pads, electrons hitting the trace area can induce small "fake" signals. The study quantified this, finding it affects only about 1% of electrons and is manageable for calorimetry.
- Molière Radius Preservation: By keeping the sensor planes thin, the effective Molière radius remains close to that of pure tungsten (~9.3 mm), which is critical for the LUXE experiment's goal of measuring high-density particle showers.
Limitations and Future Work
While the prototypes were successful, future iterations will aim to reduce the inter-pad gaps in GaAs sensors to improve spatial uniformity. Additionally, future versions of the FLAME ASIC will likely include dynamic gain switching to handle the massive energy depositions expected in the core of high-energy electromagnetic showers. The embedded traces in GaAs, while space-efficient, require detailed simulation and calibration to ensure energy measurements remain unbiased in full-scale calorimeter applications.
Summary updated