Single-Photon Detection Process

• Single-photon detection (SPD) is a process that converts an individual photon into a time-stamped electrical signal using avalanche mechanisms in semiconductor devices.
• Thermal engineering and optimized device structures minimize dark counts and afterpulsing, ensuring high detection efficiency and reliable performance.
• Integrated readout electronics and active quenching techniques enable high-speed, low-noise operation essential for quantum communications, imaging, and metrology.

Single-photon detection (SPD) encompasses the physical, electronic, and statistical processes by which the arrival of an individual photon is transduced into a macroscopic, time-stamped electrical event (“click”). High-performance SPD is foundational to quantum communications, quantum imaging, quantum metrology, and photon-starved sensing, demanding precise engineering of device structure, thermal management, noise suppression, and system integration to maximize detection efficiency, minimize noise (dark counts, afterpulses), and ensure reliability.

1. Fundamental Device Physics and Avalanche Mechanisms

Contemporary SPDs exploit discrete physical mechanisms matched to photon energy. For 0.3–1.7 μm (visible to near-infrared), semiconductor single-photon avalanche diodes (SPADs)—modeled here as InGaAs/InP or Si pn-junctions—are reverse-biased above breakdown (Geiger mode). An absorbed photon generates a carrier in a field-engineered depletion region. If the excess bias $V_{\rm ex} = V_{\rm bias} - V_{\rm br}$ is sufficient, this carrier triggers a self-sustaining impact-ionization avalanche:

$$M(V_{\rm ex}) \approx \frac{1}{1 - \int_0^w \alpha[E(x)]\,dx}$$

where $M$ is gain, $\alpha[E(x)]$ is the ionization coefficient, and $w$ is multiplication width (Xu et al., 2023). Optimization of the absorption (e.g., InGaAs layer for near-IR, Si for visible, 4H-SiC for UV (Zhao et al., 2024)), grading/charge layers, and high-field multiplication regions maximizes quantum efficiency and uniform breakdown, with design features including floating guard rings and stepped diffusions for breakdown localization.

Avalanche quenching is implemented either passively—through an integrated series resistor $R_{\rm q}$ (e.g., $200\,\mathrm{k}\Omega$ for InGaAs/InP NFADs: $\tau_q \approx 0.8\,\mathrm{ns}$)—or actively with fast electronic circuits or ASICs, determining dead time $\tau_{\rm dead}$ and maximum count rate. Shorter quenching enables higher throughput and reduced afterpulsing (Xu et al., 2023, Zhao et al., 2024).

2. Thermal Engineering and Temperature Dependence

Thermal noise, primarily via thermal carrier generation and trap-assisted tunneling, is the dominant source of dark counts:

$$\mathrm{DCR}(T)\propto \exp\left(-\frac{E_a}{k_{\rm B}T}\right)$$

where $E_a$ is material-specific (e.g., $0.5$–$0.7\,\mathrm{eV}$ for InP). To suppress DCR, InGaAs/InP SPDs are routinely cooled by compact cryocoolers or thermoelectric coolers. For example, a thermoacoustic cooler allows setpoints from 173–273 K (with $\Delta T < 0.04\,\mathrm{K}$) (Xu et al., 2023). Excessively low $T$ can increase afterpulsing due to longer trap lifetimes and compromise absorption efficiency.

Materials with wide bandgaps (SiC, GaN) or van der Waals heterostructures can operate closer to room temperature while maintaining lower DCR due to reduced carrier generation rates (Zhao et al., 2024, Abraham et al., 5 Sep 2025). Likewise, integrating device structures for localized cooling further enhances DCR suppression.

3. Readout Electronics, Quenching, and Suppression of Noise Artifacts

Efficient SPD requires extraction of sub-nanosecond avalanche pulses against large capacitive and RF transients, especially in high-frequency gated systems:

• Monolithic integrated readout circuits (MIRC), often in low-temperature co-fired ceramics, embed low-pass filtering and low-noise amplification, substantially reducing parasitic capacitance ($C_{\rm par}<0.1\,\mathrm{pF}$) and supporting GHz-gated operation (Fang et al., 2020, Zhengyu et al., 2024, Jiang et al., 2017, Jiang et al., 2018).
• Negative-feedback or thin-film resistors provide passive quenching at sub-ns scales, while active quenching/active reset ICs (e.g., AQAR) allow programmable hold-off and fine control of dead time and quench voltage swings, achieving timing jitter <50–500 ps and maximizing maximum count rate (Fang et al., 2020, Zhao et al., 2024, Xu et al., 2023).

Afterpulse probability $P_{\rm AP}$, originating from trapped carriers released after an avalanche, is addressed by:

• Adequate hold-off (dead) times, typically $2–20\,\mu$s, scaling $P_{\rm AP} \propto \tau_{\rm hold}^{-\alpha}$; faster, timed quenching reduces charge and trap population, lowering $P_{\rm AP}$ to $<$3% in state-of-the-art devices (Xu et al., 2023, Zhao et al., 2024).
• Minimizing total avalanche charge via short gates in GHz sine-wave-gated designs—e.g., $t_{\rm gate} <$ 150 ps—suppresses afterpulsing and enables high-repetition operation (Fang et al., 2020, Jiang et al., 2017, Jiang et al., 2018).

4. Fundamental Performance Metrics and Calibration

Key SPD metrics (all experimentally determined):

Metric	Definition	Range in State-of-the-Art Devices
Photon Detection Efficiency (PDE)	$$PDE(V_{\rm ex}) = \eta_{\rm abs}\cdot P_{\rm trigger}(V_{\rm ex})$$	60–84% (state-of-art Si, InGaAs, SiC) (Fang et al., 2020, An et al., 24 Jul 2025)
Dark Count Rate (DCR)	Counts/sec with no illumination	2 kcps (InGaAs), 260 cps (Si), 138 kcps (SiC UV) (Xu et al., 2023, An et al., 24 Jul 2025, Zhao et al., 2024)
Afterpulse Probability ($P_{\rm ap}$)	Fraction of counts from afterpulses	2.4–8% (InGaAs), <3% (SiC) (Xu et al., 2023, Zhao et al., 2024)
Timing Jitter	FWHM of time distribution of photon events	49 ps (InGaAs), 360 ps (Si), 400 ns (passive Si) (Xu et al., 2023, Govdeli et al., 2023)
Dead time ($\tau_{\rm dead}$)	Minimum time required between events	$<$1–10 $\mu$s (passive/active), down to 60 ns (SiC) (Zhao et al., 2024, Xu et al., 2023)
Maximum Count Rate (MCR)	$MCR = 1/\tau_{\rm dead}$	$220$ MHz (gated InGaAs), $13$ MHz (SiC) (Zhengyu et al., 2024, Zhao et al., 2024)

Evaluation methods include Poisson statistics on pulsed illumination, time-tagging for afterpulse/jitter extraction, and dark-rate measurements with laser off. Calibration can be performed absolutely via heralded-photon (Klyshko) methods with portable bi-photon sources (Pani et al., 2024), crucial for system metric traceability.

5. Variants: Materials, Integration, and Advanced Detection Modes

The choice of detector architecture reflects operational wavelength, application, and integration constraints:

• Si SPADs (visible, $<1.1\,\mu\mathrm{m}$): thick-junction, backside-illuminated devices engineered for $>$84% PDE, low DCR, and multi-mode operation (free-running, gated, hybrid) (An et al., 24 Jul 2025).
• InGaAs/InP SPADs (telecom, $1.3–1.7\,\mu\mathrm{m}$): advanced separate-absorption/multiplication, guard-ring structures, with high-frequency gating and monolithic readouts for compact, low-noise modules (Xu et al., 2023, Zhengyu et al., 2024, Jiang et al., 2018, Jiang et al., 2017, Fang et al., 2020).
• 4H–SiC SPADs (UV): mesa-terminated, thick p–n structures achieving high UV PDE, fast active quenching, and afterpulse suppression, competitive with PMTs for lidar (Zhao et al., 2024).
• Van der Waals Heterostructures (e.g., BP/MoS$_2$/Graphene): enable room-temperature SPD at 1550 nm, trading avalanche gain for single-electron sensitivity (EQE up to 21% with $<$1 kHz DCR) (Abraham et al., 5 Sep 2025).
• SNSPDs and Calorimetric Detectors: superconducting nanowires detect photons by creating resistive hotspots; calorimetric graphene devices register photon-induced electronic heating with Josephson-junction readout for near-unity quantum efficiency and sub-kHz dark rates at $\sim$1 K (Huang et al., 2024, Schroeder et al., 2017).

Quantum network models formalize the pre-amplification interaction as a passive filter characterized by a transmission amplitude $T(\omega)$, with detection efficiency $\eta(\omega)=|T(\omega)|^2$. Networks of coupled levels, balanced decay channels, and critical coupling achieve near-unit spectral efficiency, bandwidth tailoring, and group delay shaping (Propp et al., 2019).

6. Statistical Inference, Tomography, and System-level Enhancement

SPADs, lacking intrinsic photon-number resolution, support statistical protocols for number-state reconstruction. Maximum-likelihood estimators, incorporating measured PDE, DCR, dead time, and afterpulsing models, recover photon distributions from only single-photon-resolving data (Banner et al., 2023). In QKD and metrology, advanced encoding (copy-and-detect, ESPD) architectures using quantum logic (e.g., C-NOT gates plus threshold detection across SPD arrays) can asymptotically suppress dark-count-induced errors and boost effective system detection efficiency to near unity, bypassing physical SPD limitations (Shu, 2022, Shu, 1 Dec 2025).

Absolute calibration utilizes correlated photon pairs (SPDC/Klyshko) and squeezes vacuum models, accounting for multi-photon and loss effects to achieve sub-percent accuracy in SPD efficiency calibration across sites (Pani et al., 2024).

7. Practical Integration and Application-driven Optimization

Contemporary SPDs are deployed as compact, thermally managed modules, integrating SPADs, TECs, readout ASICs, FPGA-based processing, and advanced filtering (e.g., UNIC, LTCC MIRC) in footprints of $<$100 cm$^3$ (Zhengyu et al., 2024, Jiang et al., 2018). Application-driven trade-offs balance PDE, DCR, afterpulsing, jitter, and count-rate for quantum key distribution, lidar, time-of-flight ranging, time-resolved spectroscopy, and fundamental quantum optical experiments (Xu et al., 2023, Yu et al., 2017, Banerjee et al., 31 Oct 2025, Zhou et al., 2017).

Optimization strategies include:

• Field engineering for maximal absorption and carrier injection (e.g., backside illumination, optical stacks) (An et al., 24 Jul 2025, Fang et al., 2020).
• Active quenching to reduce afterpulsing and enable fast reset (Fang et al., 2020, Zhao et al., 2024).
• Operating point selection to achieve desired SNR or rate, via tuning $V_{\rm ex}$, hold-off time, and temperature (Xu et al., 2023, Zhao et al., 2024).
• Integration of temperature compensation and real-time monitoring for stability (Xu et al., 2023, Zhengyu et al., 2024).

Through tight synergy of heterostructure design, high-speed electronics, thermal engineering, and system-level calibration, single-photon detection achieves the performance envelope necessary for modern quantum technologies and high-sensitivity measurements (Xu et al., 2023, Shu, 1 Dec 2025, An et al., 24 Jul 2025, Zhao et al., 2024, Pani et al., 2024).