Access Official Technical PDF

Share Summary:LinkedInEmail

Silicon photomultiplier sensors (SiPMs) can detect light at the single-photon level while offering the practical benefits of a solid-state device. Their performance, however, depends on more than a headline sensitivity figure. Engineers comparing sensors need to consider the light source, required signal quality, operating conditions, readout design, package and production needs as parts of the same system.

A SiPM combines an array of closely packed single-photon avalanche diodes (SPADs), each with a quench resistor. The outputs of these microcells are summed to produce an analog signal. This structure makes SiPMs useful in applications that need to detect very low light levels, but it also means that parameters such as operating bias, temperature and microcell design affect how the sensor behaves.

How to compare silicon photomultiplier sensors

Start by translating the application requirements into measurable sensor requirements. Identify the wavelengths to detect, the expected light level, the needed signal-to-noise ratio, timing or count-rate demands, and the environmental and assembly constraints. Then compare candidate sensors against those needs rather than choosing by one specification alone.

Match sensitivity to the light source

Photon detection efficiency (PDE) describes the proportion of incident photons that produce a detectable, internally amplified event. It varies with wavelength and bias voltage, so a peak PDE value does not tell the whole story. Check the sensor’s wavelength response at the wavelengths used in the application, and consider how much bias is required to reach the stated performance.

Responsivity is another useful measure of response to light across wavelength. A SiPM’s high internal gain can support useful responsivity beyond the wavelength at which PDE peaks, but the actual response curve should still be compared with the application’s source and optical path.

Evaluate noise and signal quality

Dark count rate (DCR) is the rate of single-photon-level events generated by thermal electrons in the silicon rather than by incoming light. It contributes to sensor noise and can reduce signal-to-noise ratio. DCR is particularly relevant in low-light measurements and when signals are integrated over longer periods. It also changes with temperature and overbias, and scales with sensor area.

When comparing DCR specifications, confirm whether a value is stated per unit area or for the whole device. Also check the measurement conditions and whether the figure represents selected or binned devices rather than the broader product line.

Crosstalk and afterpulsing can also create unwanted events. Their importance depends on the measurement: the acceptable level for one application may not suit another. Both are influenced by overbias and microcell size. Package or device features such as trenches may reduce crosstalk, but their presence alone does not establish how a sensor will perform; compare measured specifications.

Consider timing, bias and temperature together

For timing-sensitive systems, distinguish the microcell’s switching behavior from the output pulse observed by the readout electronics. Sensor impedance affects the measured rise time, while recovery time depends mainly on microcell capacitance and quench resistance. Sensor size and series resistance can also influence pulse recovery.

Some SiPM designs provide a dedicated fast output in addition to the standard anode and cathode connections. This can support faster timing measurements and may present lower capacitance to the readout circuit. Whether that feature is valuable depends on the required timing resolution and count rate; for less demanding applications, device and microcell size may be sufficient considerations.

Breakdown voltage is the bias point at which the electric field supports a Geiger discharge. The operating bias is set above breakdown, and this overvoltage affects sensitivity as well as noise-related parameters. Account for how breakdown voltage varies between devices, whether multiple sensors can share a bias source, and what the supply and control circuitry require.

Temperature affects both breakdown voltage and DCR. A temperature change can alter overbias if the applied voltage remains fixed, while thermal generation increases DCR as temperature rises. Bias compensation can address changes in overbias; it does not eliminate the temperature-driven increase in DCR. Review the expected temperature range and decide whether compensation or cooling is required by the application.

Choose a package for optics and production

The package affects the sensor’s physical integration as well as its optical and electrical behavior. Compare package size, edge dead space, terminal style and compatibility with the intended assembly process. Pin-based parts may suit manual assembly, while pads that support reflow soldering are relevant to automated, higher-volume production.

Optical transmission through the package material matters when the application uses wavelengths that the encapsulant could absorb. Package materials can also matter in specialized environments: ferrous components may be a concern in strong magnetic fields, and trace radioactive elements may matter in low-background research. These are application-specific constraints, not universal selection criteria.

For production planning, check delivery format, moisture sensitivity level (MSL), storage requirements and soldering compatibility. These details can affect assembly workflows and the package options available to a contract manufacturer.

Assess consistency, reliability and support

When a system uses many sensors, variation in breakdown voltage and optical response can complicate biasing, calibration and system-level consistency. Ask how the manufacturer tests devices, controls process variation and reports uniformity. Determine whether any stated tight parameter range comes from routine production control or from selecting devices into narrower bins.

Reliability testing and documentation are also part of the evaluation. Because a common reliability assessment program does not exist specifically for SiPMs, ask what test approach is used and whether it follows established integrated-circuit practices. Confirm that datasheets and application information address the needs of the design team, and that a support channel is available for technical questions.

Let application metrics guide the final choice

Translate system goals into the sensor characteristics most likely to affect them. For example, positron emission tomography (PET) timing performance is influenced by PDE and single-photon timing resolution. Gamma-ray spectroscopy depends on energy resolution, with sensitivity important and DCR or crosstalk potentially significant at low energies. LIDAR and gesture recognition require accurate timing, while low returned-photon levels make wavelength-specific PDE and crosstalk relevant.

The practical lesson is to evaluate silicon photomultiplier sensors as components in a complete measurement chain. Compare sensitivity, noise and timing under relevant operating conditions, then check that package, manufacturing consistency, reliability and support fit the intended system and production scale. This application-led approach helps engineers avoid optimizing a single specification at the expense of overall system performance.

Access Official Technical PDF

About The Author