Simulate Single-Photon Avalanche Diode (SPAD) microcell arrays, 106 Geiger gain, sub-100 ps timing resolution, 100% PET/MRI magnetic immunity, and 20-year cleanroom CMOS wafer fab economics.
A Silicon Photomultiplier (SiPM) solves the classical dilemma of semiconductor photodetectors: PIN photodiodes lack internal amplification (gain = 1), while linear Avalanche Photodiodes (APDs) only provide gains of 50 to 200 with severe temperature sensitivity. The SiPM bypasses this limitation by operating an integrated two-dimensional array of microcells in Geiger mode, biased several volts above their breakdown voltage (\(V_{\text{bias}} = V_{\text{bd}} + V_{\text{ov}}\)).
When an optical photon from a scintillation crystal generates an electron-hole pair within the depletion zone, it initiates an explosive, self-propagating impact ionization cascade. Each microcell contains an integrated polysilicon quench resistor (\(R_q \approx 300\text{ k}\Omega\)) wired in series. As the avalanche current flows, the voltage drop across \(R_q\) instantly lowers the junction bias below breakdown, quenching the discharge in less than one nanosecond. This creates a quantized, highly reproducible charge packet:
Because all microcells are connected in parallel to a single readout line, the total output current is simply the sum of all concurrently fired microcells. For scintillation pulses within the microcell density limit, the output charge is strictly proportional to the number of incident photons.
For six decades, Positron Emission Tomography (PET) relied exclusively on vacuum Photomultiplier Tubes (PMTs). However, vacuum PMTs cannot operate inside Magnetic Resonance Imaging (MRI) scanners: even modest magnetic fields (\(>0.05\text{ Tesla}\)) exert Lorentz forces (\(\mathbf{F} = q(\mathbf{v} \times \mathbf{B})\)) that violently deflect secondary electrons away from dynodes, completely extinguishing tube gain.
In contrast, SiPMs are solid-state silicon chips. The high internal electric fields in the avalanche zone exceed \(3 \times 10^5\text{ V/cm}\), producing carrier drift velocities of \(10^7\text{ cm/s}\) over sub-micron distances. The magnetic deflection over these nanometer mean-free paths is utterly negligible. SiPMs maintain 100% gain and timing stability inside 1.5T, 3T, and even 7T ultra-high-field MRI systems, making simultaneous PET/MRI clinical imaging a commercial reality.
Furthermore, SiPMs exhibit ultra-fast single photon timing resolution (SPTR < 100 ps), allowing paired scintillation detectors with LSO or LYSO crystals to achieve Coincidence Timing Resolution (CTR) below 200 ps. This enables Time-of-Flight (TOF) PET reconstruction, drastically improving signal-to-noise ratio (SNR) and lesion detectability while minimizing patient radiotracer dose.
| Year | Revenue ($) | Free Cash Flow ($) | Discounted FCF (10%) | Cumulative Cash Flow ($) |
|---|
High-density lutetium oxyorthosilicate crystal for 511 keV PET coincidence and 40 ns decay.
Peltier-cooled sub-0.1 pF SDD for handheld XRF metal sorting and SEM-EDS elemental mapping.
High-Purity Germanium 77K semiconductor gamma spectrometry with 1.8 keV FWHM resolution.
Microchannel plate continuous dynodes, chevron ion trapping, and sub-25 ps fast timing.
Microcrystalline phosphor screen for multi-MeV alpha Bragg peak capture and Lucas radon cells.
An SiPM operates each of its thousands of constituent SPAD microcells slightly above the breakdown voltage in Geiger mode. When an absorbed photon creates an electron-hole pair, it triggers a self-sustaining avalanche breakdown multiplying carriers until the cell's integrated polysilicon quench resistor drops the bias below breakdown. This passive discharge produces a standardized charge packet of 10^5 to 10^6 electrons from a bias of only 30V to 60V, matching vacuum PMT gain without high-voltage power supplies.
Traditional photomultiplier tubes accelerate electrons through free vacuum paths across dynodes where Lorentz forces from external magnetic fields deflect the beam, rendering PMTs unusable inside MRI bores. In an SiPM, charge carriers travel across microscopic sub-micron silicon depletion regions governed by intense internal electric fields exceeding 300,000 V/cm. The Lorentz deflection over these nanometer mean-free paths is negligible, allowing SiPMs to operate stably in magnetic fields up to 7 Tesla and higher.
During an avalanche breakdown, hot-carrier recombination emits secondary photons (approximately 3 photons per 10^5 avalanche electrons). These isotropic near-infrared photons can travel through the silicon substrate into neighboring microcells, triggering false coincident breakdowns called optical crosstalk. Fabricating deep silicon isolation trenches filled with tungsten or dielectric materials between microcells blocks these photons, cutting crosstalk from over 20% down to under 4%.
The dark count rate (DCR) in high-purity silicon SPADs is primarily driven by Shockley-Read-Hall (SRH) thermal generation of electron-hole pairs within the depleted avalanche junction. According to Boltzmann statistics, thermal excitation over trap states is exponential with temperature. Cooling the silicon sensor by approximately 8°C cuts thermal carrier generation in half, allowing thermoelectric Peltier cooling to reduce ambient noise by orders of magnitude for ultra-low-light applications.
Smaller microcell pitches (e.g. 15 to 25 micrometers) provide vastly higher microcell densities per square millimeter, dramatically expanding dynamic range and delaying non-linear saturation when exposed to bright scintillation flashes like LSO. However, smaller cells have a lower active geometric fill factor due to guard rings and quench resistors, which slightly reduces overall Photon Detection Efficiency (PDE). Larger pitches (35 to 50 micrometers) achieve fill factors over 80% for peak PDE but saturate at lower photon fluxes.