SiPM Readout Separates Cherenkov and Scintillation Light

Dual-readout calorimetry aims to measure scintillation and Cherenkov light separately, because the two signals carry different information about shower development. In dense scintillating crystals, however, both components reach the photosensor together, and separating them without sacrificing a compact detector layout is difficult. The authors report a beam-test demonstration in bismuth germanate (BGO) and bismuth silicate (BSO) crystals coupled to silicon photomultipliers (SiPMs), targeting an electromagnetic calorimeter concept for a future electron-positron Higgs factory.

Core Contribution

The paper’s central claim is an event-by-event separation scheme that combines spectral and timing information rather than relying on either alone. Cherenkov light and scintillation light differ in both wavelength distribution and temporal profile. The proposed readout uses an optical filter to bias one channel toward the Cherenkov component, then fits waveform templates to disentangle the remaining overlap. According to the authors, this is the first demonstration of Cherenkov–scintillation separation in BGO and BSO using SiPM readout.

That distinction matters for detector design. Earlier dual-readout work has established the value of separating crystal signals, but SiPMs offer a compact, magnetic-field-tolerant alternative to more traditional photodetectors. The contribution here is therefore not a new radiation process or crystal composition; it is a practical readout path that combines commercially available photosensors, optical filtering, and digitized pulse-shape information.

Technical Approach

The dual-readout layout assigns the signals to a Cherenkov-sensitive C channel and a scintillation-sensitive S channel. Figure 1 depicts that division: SiPMs observe the crystal through different optical selections, while subsequent waveform analysis exploits the distinct time structure of the two components. The method is designed to retain a measurement of both components for each event rather than extracting only an ensemble-average Cherenkov fraction.

The supplied figure inventory also documents the physical implementation: BGO and BSO crystals were wrapped with Mylar and instrumented with C- and S-channel SiPMs, an optical filter, and a crystal holder. The beam-test arrangement used the crystal under test together with three scintillators, T1 through T3, to form the trigger. These details make the work an experimental detector study rather than a simulation-only proposal.

The paper’s references situate the design within established dual-readout calorimetry, crystal-calorimeter, SiPM, optical-filter, waveform-digitizer, and GEANT4 work. They do not, by themselves, establish a released software package or dataset, so no artifact is listed here.

Results and Analysis

The reported beam measurements used high-energy muon and positron beams at the CERN SPS North Area. The clearest quantitative outcome is a Cherenkov yield of up to approximately 150 photoelectrons per GeV in electromagnetic showers. That is the relevant operational result: the filtered, SiPM-based chain retained enough Cherenkov signal for event-level separation rather than merely observing a small prompt-light excess.

The evidence supports the narrower claim that the two light components can be separated in these crystals with this readout architecture under test-beam conditions. It does not, from the supplied material, establish the full performance of an electromagnetic calorimeter: no energy-resolution comparison, large-array measurement, pileup study, radiation-tolerance result, or system-level detector integration metric is reported here. The 150-photoelectron-per-GeV figure is promising as a light-yield measurement, but it is not by itself a demonstration of improved calorimetric resolution.

For future-collider R&D, the result is useful because it turns a detector-design premise into a measured sensor-and-crystal signal chain. The next deciding evidence would be whether that separation remains stable when scaled from a single crystal arrangement to a calorimeter module, and whether the recovered C and S observables improve reconstructed electromagnetic-shower performance against a conventional single-readout baseline.