CMOS sensor radiation tolerance is a key consideration wherever monolithic pixel detectors must operate in a radiation environment. CMOS Monolithic Active Pixel Sensors (CPS), also known as MAPS, combine sensing elements and signal-processing electronics on one chip. This integration supports thin sensors with fine pixel granularity, but it also means radiation can affect both the charge-collection system and the circuitry that reads it out.
Early CPS used standard CMOS processes that were not designed specifically for particle detection. Their sensing regions were generally non-depleted or only partly depleted, and the sensors collected much of their signal charge by diffusion. Research and process improvements have extended the radiation tolerance of these devices, while newer CMOS technologies have enabled a move toward fully depleted designs.
Why radiation affects both the sensor and its electronics
Radiation damage is commonly divided into two categories: ionizing and non-ionizing. A particle or photon can produce both, so their effects may be present together in an irradiated device. Separating the mechanisms is useful when assessing performance and considering design changes.
Ionizing damage in insulating structures
Ionizing radiation creates electron-hole pairs in materials such as silicon dioxide (SiO2), used for insulation in CMOS devices. Some of the resulting holes become trapped near the silicon–oxide interface. The accumulated charge and interface defects can alter local electric fields and transistor behavior.
Possible consequences include shifts in transistor threshold voltage and unwanted leakage paths. The scale of the effect depends on features such as oxide thickness, electric field, device geometry, and the manufacturing process. Thin transistor gate oxides can be less susceptible to some forms of charge buildup, but thicker isolation structures, including shallow trench isolation, may remain vulnerable.
Enclosed-layout transistors are one approach to reducing the impact of radiation-sensitive isolation beside a transistor channel. Their geometry can limit exposure to such structures, but the design also involves trade-offs, including increased gate capacitance and transistor size.
Non-ionizing damage in the silicon lattice
Non-ionizing radiation can displace silicon atoms from their crystal-lattice positions, creating defects within the material. These defects can increase the leakage current of charge-collection diodes and reduce the effective lifetime of signal charge carriers through trapping or recombination.
Non-ionizing damage is often described using the Non-Ionizing Energy Loss (NIEL) framework, with exposure expressed as an equivalent fluence of 1 MeV neutrons. This provides a useful basis for comparing radiation exposures, although it does not capture every microscopic difference between radiation sources or materials.
Radiation-induced defects can also change the effective doping of silicon. In p-type material, acceptor removal and the creation of additional defects can alter the active region’s electrical properties. The outcome depends on factors including the initial material and radiation type, and can evolve through annealing.
How radiation damage changes CPS performance
Classical CPS often use a relatively thin, low-resistivity epitaxial layer as their active medium. Because this region is not fully depleted, much of the signal charge reaches collection diodes by thermal diffusion rather than being swept directly by an electric field. This helps explain why radiation-induced trapping and changes in charge collection are important concerns: a longer or less reliable collection process can reduce the signal available to the pixel electronics.
The signal in these sensors can be small, making noise and leakage current especially relevant. An increase in diode leakage contributes additional shot noise and can affect the pixel’s ability to distinguish a particle signal. Damage to transistor characteristics can also influence the amplification and readout chain, which is integrated into the same chip as the sensing elements.
Radiation effects may vary across pixel architectures. For example, a traditional three-transistor pixel uses a reset cycle and correlated double sampling to manage its signal. Self-bias pixels instead compensate for leakage continuously, while clamping pixels add circuitry for signal sampling and discrimination. These designs have different operating behaviors, so radiation assessment needs to consider the complete pixel architecture rather than the sensor material alone.
Assessing radiation tolerance in monolithic sensors
Because CPS integrate sensing and readout, the response of the complete device is central to testing. Relevant measurements described for these sensors include:
- Temperature and leakage current: Track operating conditions and radiation-related changes in diode current.
- Pixel noise: Evaluate the noise of the analogue readout and its effect on signal detection.
- Charge-collection measurements: Use radioactive sources to examine signal amplitude, gain, charge-collection efficiency, and signal-to-noise ratio.
- Pixel clearing behavior: Characterize how quickly a pixel’s signal returns toward its baseline in architectures with continuous leakage compensation.
Testing can use different radiation sources to investigate ionizing and non-ionizing effects. Results should be interpreted with attention to device temperature and annealing history, because these conditions can influence measured leakage and the evolution of radiation damage.
Design strategies and the move to depleted CPS
Radiation hardening has involved adapting device layouts and pixel designs to reduce sensitivity to identified damage mechanisms. Examples include enclosed-layout transistors to limit the influence of isolation structures and circuit architectures developed to manage leakage current or readout constraints. The effectiveness of a strategy depends on the process and device under evaluation; no single design choice addresses every radiation effect.
Improved CMOS processes have also opened a path toward depleted monolithic active pixel sensors, often called DMAPS. A depleted sensing region can collect charge through drift, unlike the diffusion-dominated collection in classical CPS. Newer process options can also ease earlier design constraints, including the separation of transistor wells from the sensitive volume. These developments support research into sensors with improved radiation tolerance and rate capability, although performance must be established for each implementation.
Practical implications
Evaluating CMOS sensor radiation tolerance requires attention to the whole device: silicon defects, oxide and interface damage, leakage current, pixel noise, charge collection, and readout architecture. Engineers and technical decision-makers should match irradiation tests to the expected damage mechanisms, document temperature and annealing conditions, and assess signal performance alongside electrical measurements.
For applications with demanding radiation conditions, process choice and pixel design are both important. Hardening measures can address specific weaknesses in established CPS, while depleted CMOS sensors offer a different approach to charge collection. The practical lesson is to base sensor selection on measured performance under relevant conditions rather than assuming that a CMOS process or pixel architecture is inherently radiation tolerant.