KAHIBARO
Table of Contents
Basic idea
Silicon photomultipliers, often called SiPMs, are light sensors that detect very small flashes of light. They are widely used in radiation detection because many detectors, especially scintillation detectors, produce light when radiation deposits energy. The role of the SiPM is to convert that faint light into an electrical signal that can be measured.
A silicon photomultiplier is a semiconductor device made of many tiny avalanche photodiodes connected together on one chip. Each tiny element is often called a microcell or pixel. Every microcell works in a digital-like way, it either fires or it does not fire. When many microcells are combined, the total output becomes proportional to the number of detected photons, as long as not too many cells fire at once.
Structure and operation
A SiPM consists of a large array of microcells. Each microcell is a reverse biased photodiode operated above its breakdown voltage. This operating region is called Geiger mode. In this mode, a single charge carrier created by an incoming photon can trigger a large avalanche of carriers inside the microcell.
Because the avalanche would continue uncontrollably if nothing stopped it, each microcell includes a quenching resistor. This resistor reduces the voltage across that microcell after avalanche starts, so the avalanche stops. After that, the microcell slowly recovers and becomes ready to detect another photon.
The total SiPM signal is the sum of the signals from all fired microcells. If one photon triggers one cell, the output is one standard pulse. If several photons trigger several different cells at nearly the same time, the pulses add.
Important operating rule:
A SiPM microcell is biased above breakdown voltage:
$$V_{\text{bias}} > V_{\text{bd}}$$
The excess amount
$$V_{\text{ov}} = V_{\text{bias}} - V_{\text{bd}}$$
is called the overvoltage.
The overvoltage strongly affects gain, photon detection efficiency, noise, and timing.
Why it is called a photomultiplier
The name comes from its ability to produce a large electrical signal from very little light, similar in purpose to a traditional photomultiplier tube. However, the physical principle is different. A photomultiplier tube uses vacuum electronics and dynodes, while a SiPM uses semiconductor avalanches in silicon.
Each fired microcell releases a nearly fixed charge. This charge is approximately
$$Q_{\text{cell}} = C_{\text{cell}} \, V_{\text{ov}}$$
where $C_{\text{cell}}$ is the capacitance of one microcell.
The gain is the number of electrons produced per detected photon event, and is approximately
$$G = \frac{Q_{\text{cell}}}{e} = \frac{C_{\text{cell}} V_{\text{ov}}}{e}$$
where $e$ is the elementary charge.
Because this gain can be very large, even a very weak light pulse can be measured.
Microcell array behavior
The microcells are connected in parallel, so their currents add together. If the detector has $N_{\text{cells}}$ microcells, then at most that many can fire at exactly one instant before saturation appears. This gives the SiPM a finite dynamic range.
For small light levels, the number of fired cells is roughly proportional to the number of incident photons. For high light levels, many cells are already occupied, so the response becomes nonlinear.
A common idealized relation is
$$N_{\text{fired}} = N_{\text{cells}} \left(1 - e^{- \mu / N_{\text{cells}}}\right)$$
where $\mu$ is the average number of effective triggering photons. This shows that the output approaches a maximum as more and more light arrives.
Important limitation:
A SiPM is linear only when the number of fired cells is much smaller than the total number of cells.
When many cells fire, saturation occurs and the output no longer increases proportionally with light.
Photon detection efficiency
Not every incoming photon is detected. The probability that an incident photon produces a fired microcell is called the photon detection efficiency, or PDE.
The PDE depends on several factors. A useful conceptual expression is
$$\text{PDE} = \text{fill factor} \times \text{quantum efficiency} \times P_{\text{trigger}}$$
The fill factor is the fraction of chip area that is photosensitive. The quantum efficiency describes how often a photon creates a charge carrier in silicon. The trigger probability is the chance that the carrier starts an avalanche.
The PDE depends on wavelength. Silicon is especially sensitive in parts of the visible and near infrared range, which makes SiPMs very suitable for many scintillators.
Pulse shape and recovery
When a microcell fires, it produces a pulse with a fast rise and a decay determined mainly by the cell capacitance and quenching resistor. The recovery time is approximately
$$\tau \approx R_q C_{\text{cell}}$$
where $R_q$ is the quenching resistance.
A short recovery time allows the device to respond to rapidly repeated light pulses. A long recovery time reduces the maximum count rate and affects pulse shape.
Noise and non-ideal effects
Real SiPMs are not perfect. Several effects can produce extra pulses that are not caused directly by the desired light signal.
Dark counts are pulses produced by thermally generated carriers inside the silicon. Even in complete darkness, some microcells may fire. The dark count rate usually increases with temperature.
Afterpulsing happens when carriers become trapped during an avalanche and are released later, causing delayed extra avalanches.
Optical crosstalk occurs when an avalanche in one microcell emits secondary photons inside the silicon, and these photons trigger neighboring cells. This makes one detected event look like several.
These effects worsen energy resolution and counting accuracy if not controlled.
Important practical fact:
SiPM noise increases with temperature, and breakdown voltage also changes with temperature.
For stable operation, the bias voltage often must be adjusted as temperature changes.
Output and single photon sensitivity
One important advantage of SiPMs is that the pulse from one fired microcell is very uniform. Because of this, the output charge spectrum often shows separated peaks corresponding to 0, 1, 2, 3, and more fired cells. This means SiPMs can have excellent single photon counting capability.
If the number of fired microcells is measured, then the light signal can be estimated. This is very useful in low light applications such as scintillation timing, medical imaging, and particle detectors.
Comparison with photomultiplier tubes
SiPMs and photomultiplier tubes often serve similar purposes, but their properties differ.
| Property | SiPM | Photomultiplier Tube |
|---|---|---|
| Technology | Semiconductor | Vacuum tube |
| Operating voltage | Low to moderate | High |
| Size | Compact | Larger |
| Magnetic field sensitivity | Very low | Can be strong |
| Mechanical robustness | High | More fragile |
| Single photon detection | Good | Good |
| Dark noise | Often higher | Often lower in some cases |
| Active area | Usually smaller per device | Often larger |
SiPMs are especially attractive where compact size, low voltage, ruggedness, and operation in magnetic fields are important.
Use with scintillation detectors
In a scintillation detector, radiation deposits energy in a scintillator and produces a flash of light. The SiPM is optically coupled to the scintillator and detects part of that light. The electrical pulse amplitude is then related to the number of detected scintillation photons, and therefore to the deposited energy.
Because SiPMs are small, they are often matched to small crystal scintillators, plastic scintillators, or arrays of detector elements. Their fast timing performance also makes them useful in coincidence measurements and time of flight systems.
Advantages
SiPMs have several important strengths. They are compact and lightweight. They operate at much lower voltage than traditional photomultiplier tubes. They are insensitive to magnetic fields, which is valuable in systems that also contain magnets. They have high gain, fast response, and can resolve very small light signals. They are also mechanically robust and easy to integrate into electronic systems.
Limitations
SiPMs also have limitations. Their active area is often relatively small, so large detectors may need arrays of many SiPMs. They suffer from dark counts, afterpulsing, and optical crosstalk. Their response saturates when too many microcells fire. Their properties depend on temperature, so bias control and thermal stabilization are important.
Practical operating considerations
To use a SiPM well, the bias voltage must be chosen carefully. Increasing overvoltage usually increases gain and PDE, but it also increases noise and crosstalk. There is therefore a tradeoff between sensitivity and signal quality.
Optical coupling to the scintillator must also be good. Poor coupling causes light loss and reduces signal size. Electronics connected to the SiPM must be fast and low noise enough to preserve the information in the pulse.
Key practical tradeoff:
Higher overvoltage gives larger gain and usually higher PDE, but also increases dark counts, afterpulsing, and optical crosstalk.
Good SiPM operation requires balancing these effects.
Summary
A silicon photomultiplier is a semiconductor light detector made of many Geiger mode avalanche microcells connected in parallel. Each detected photon can trigger a standard avalanche pulse, and the sum of many such pulses gives the output signal. SiPMs are highly sensitive, compact, fast, and well suited for reading scintillation light in radiation detectors. Their main challenges are noise, saturation, and temperature dependence, but their advantages have made them very important in modern detector systems.