Single-photon detection efficiency evaluation is the core metrology process for quantifying how effectively photodetectors such as SPADs, SiPMs, SNSPDs, and PMTs convert individual incident photons into registered count events. Accurate photon detection efficiency (PDE) characterization is essential for the performance, calibration, and reliability of quantum key distribution systems, time-of-flight LiDAR, quantum optics experiments, and ultra-low-light imaging platforms. This article explains the technical principles, sample types, key testing parameters, measurement methods, standards references, and practical applications of single-photon detection efficiency evaluation performed by accredited third-party laboratories.
Technical Overview
Single-photon detection efficiency evaluation measures the probability that a photon incident on a single-photon detector produces a measurable output count. Because single photons represent optical power at the attowatt level, this parameter cannot be assessed with conventional power meters; it requires photon-counting statistics, calibrated attenuation chains traceable to national metrology institutes, and rigorous uncertainty analysis. Several related efficiency quantities must be clearly distinguished during evaluation:
- Photon Detection Efficiency (PDE): the intrinsic device-level probability that a photon arriving at the photosensitive region triggers a registered avalanche or switching event, determined by quantum efficiency, fill factor, and triggering probability.
- System Detection Efficiency (SDE): the counts registered per photon delivered to the detector module input, typically a fiber connector or defined optical aperture, including all internal coupling and packaging losses.
- Coupling efficiency: the ratio linking SDE and PDE, covering alignment, reflection, and mode-matching losses between the optical interface and the active area.
Detection efficiency directly determines the secure key rate in quantum key distribution, the maximum ranging distance of photon-counting LiDAR, and the signal-to-noise ratio of fluorescence and Raman instrumentation. Evaluations are therefore performed under controlled conditions with traceable reference standards, uncertainty evaluated in accordance with ISO/IEC Guide 98-3 (the Guide to the Expression of Uncertainty in Measurement, GUM), and testing conducted by ISO/IEC 17025-accredited third-party laboratories.
Testing Samples
Single-Photon Avalanche Diodes (SPADs)
Silicon SPADs operating from 400 nm to 1100 nm and InGaAs/InP SPADs covering 900 nm to 1700 nm are the most frequently submitted samples. Devices may be discrete chips, wire-bonded dies, or packaged modules, tested in free-running or gated modes depending on the operational protocol.
Silicon Photomultipliers (SiPMs)
SiPMs consist of thousands of Geiger-mode microcells with integrated quenching resistors. Their PDE is strongly dependent on overvoltage and microcell fill factor, so evaluation typically includes efficiency-versus-overvoltage sweeps alongside the standard wavelength characterization.
Superconducting Nanowire Single-Photon Detectors (SNSPDs)
Niobium nitride or titanium nitride nanowire devices operated at 0.8 K to 4 K in closed-cycle cryostats offer the highest system efficiencies (frequently above 90%) and lowest dark count rates. Evaluation requires cryogenic integration, fiber-to-chip coupling verification, and bias-current-dependent efficiency mapping.
Photomultiplier Tubes (PMTs) and Hybrid Detectors
Vacuum-tube detectors are evaluated by converting cathode radiant sensitivity into quantum efficiency and by photon-counting measurements under standardized illumination geometry across the specified spectral range.
Fiber-Coupled Modules and Complete Receiver Assemblies
Integrated modules containing detector, quenching circuitry, discriminator, and thermoelectric or cryogenic cooling are evaluated at the system level, commonly at 850 nm, 1064 nm, 1310 nm, and 1550 nm telecom wavelengths.
Testing Items
A comprehensive single-photon detection efficiency evaluation typically covers the following parameters:
- System detection efficiency and photon detection efficiency at specified operating wavelengths
- Spectral PDE scan across the full operating band (for example, 400–1100 nm for silicon devices, 1260–1650 nm for InGaAs devices)
- PDE as a function of excess bias voltage and breakdown voltage characterization
- Dark count rate in counts per second, and dark count density normalized to active area
- Afterpulsing probability, particularly critical for InGaAs/InP SPADs
- Dead time, recovery time, and count-rate saturation behavior
- Timing jitter (single-shot resolution, FWHM) using pulsed laser excitation
- Spatial uniformity of the photosensitive area under scanned illumination
- Temperature dependence of detection efficiency and dark counts
- Count-rate linearity over the intended photon flux range and maximum linear count rate
Testing Methods
Calibrated Attenuation Method
A stable continuous-wave or pulsed laser is attenuated through a chain of calibrated optical attenuators, reducing the flux to approximately 10⁴ to 10⁶ photons per second at the device under test. The photon flux is calculated from a reference power meter reading and the attenuation values, both traceable to SI optical power standards maintained by national metrology institutes.
Count-Rate Measurement and Statistical Correction
The detector count rate is recorded with the signal applied and with the source blocked, and the dark background is subtracted. Because avalanche detectors exhibit dead time, the measured rate is corrected using an appropriate statistical model; for a non-paralyzable detector the true rate equals the measured rate divided by (1 minus the product of the measured rate and the dead time). Photon flux is deliberately kept low so that corrections remain small and well-conditioned.
Substitution Method with a Reference Detector
A calibrated reference radiometer of the type described in IEC 61040 is alternated with the device under test at the same optical plane. This method minimizes errors from beam geometry and source drift and provides an independent cross-check of the attenuation-based flux determination.
Correlated-Photon (Klyshko) Method
For highest-accuracy absolute calibration, a spontaneous parametric down-conversion source generates photon pairs. The detection efficiency is derived from coincidence-to-singles ratios, eliminating dependence on conventional optical power standards and their associated uncertainties.
Operating Condition Control
Evaluation requires precise control of bias voltage, temperature (thermoelectric stabilization for SPADs and SiPMs, cryogenic temperatures for SNSPDs), polarization state, and fiber-coupling reproducibility. All measurements are performed in light-tight enclosures, and laser operation follows the safety requirements of IEC 60825-1.
Uncertainty Evaluation and Reporting
The uncertainty budget accounts for reference detector calibration, attenuator calibration and linearity, wavelength accuracy, coupling reproducibility, dead-time model assumptions, and counting statistics. Expanded uncertainties (coverage factor k = 2) of approximately 2% to 5% relative are typical, with all results reported in accordance with ISO/IEC Guide 98-3 and issued by ISO/IEC 17025-accredited laboratories.
Applications
Verified single-photon detection efficiency is a decisive performance parameter across advanced photonics and quantum technology sectors:
- Quantum key distribution (QKD), where secure key rate and link budget depend directly on detector efficiency at 1310 nm and 1550 nm
- Photonic quantum computing and boson sampling, which require high-efficiency, low-jitter photon counting at the output stage
- Quantum networks, heralded photon sources, and quantum repeater research
- Single-photon time-of-flight LiDAR and long-range three-dimensional imaging
- Fluorescence lifetime imaging microscopy (FLIM) and single-molecule spectroscopy
- Ultra-weak Raman and luminescence spectroscopy in materials science
- Deep-space and free-space optical communication receivers
- Astronomy, including adaptive optics wavefront sensing and transit photometry
- Positron emission tomography and other biomedical imaging modalities using SiPM readout
- Optical time-domain reflectometry, distributed fiber sensing, and semiconductor photon-emission failure analysis
FAQ
What is the difference between PDE and SDE?
PDE describes the intrinsic probability that a photon reaching the photosensitive region generates a count, while SDE includes all packaging and coupling losses between the module input (such as a fiber connector) and the active area. SDE is always equal to or lower than PDE, and the ratio between them defines the coupling efficiency.
Which wavelengths can be evaluated?
Typical coverage includes 400 nm to 1100 nm for silicon SPADs and SiPMs, and 900 nm to 1700 nm for InGaAs/InP devices, with standard test points at 850 nm, 1064 nm, 1310 nm, and 1550 nm. SNSPDs are commonly characterized from 980 nm to 1650 nm, and spectral scans can be performed on request.
How is the photon flux calibrated?
The flux is established by measuring laser power with a reference detector and applying calibrated attenuation values, both traceable to national optical power standards. Alternatively, the correlated-photon Klyshko method provides an absolute calibration without relying on conventional power standards.
Why is dead-time correction necessary?
Single-photon detectors are blind for a finite recovery period after each event, so the measured count rate systematically underestimates the true event rate. Applying the appropriate paralyzable or non-paralyzable statistical model removes this bias and prevents detection efficiency from being understated at moderate count rates.
What measurement uncertainty can be expected?
With traceable attenuation chains and careful coupling control, expanded uncertainties of approximately 2% to 5% relative (k = 2) are achievable. The complete uncertainty budget is documented in the test report in accordance with ISO/IEC Guide 98-3.
How are cryogenic SNSPDs evaluated?
SNSPD samples are mounted in closed-cycle cryostats operating below 4 K, with fiber feedthroughs, bias-tee circuitry, and calibrated RF readout. Efficiency is mapped against bias current, and dark count rate and timing jitter are recorded simultaneously at the cryogenic operating point.
Which standards govern this evaluation?
Testing is performed under the laboratory competence requirements of ISO/IEC 17025, with uncertainty evaluation per ISO/IEC Guide 98-3, reference detectors consistent with IEC 61040, and laser safety per IEC 60825-1, together with applicable national photodetector test specifications such as relevant GB/T standards for photon-counting device characterization.
