Our experienced and innovative engineering team has a thorough knowledge of image sensor technology. In this page, you will find a non-exhaustive list of technologies and parameters which are important for image sensor developments.
Needless to say that nobody likes noise in an image. The lower it is, the better the image. The IMA300 family achieves noise performance close to the single electron. Our Single Photon Avalanche Detectors will generate a digital signal for every electron they detect. As each visible or infra-red photon generates a single electron in silicon, this corresponds to the detection of a single photon.
High full well
In some applications, it is important to be able to detect millions of photons in a single frame, without saturating the sensor. Our family of sensors for intraoral dental imaging achieves linear full of 4 millions of electrons
High dynamic range
Special designs are needed to achieve high dynamic range. Our patented technology is used in the family of sensor for intraoral dental, achieving a dynamic range of 94.3 dB, corresponding to 15.7 of dynamic range. The sensors in the IMA300 family also achieves high-dynamic range.
Charge domain binning
Our patented pixel technology (US 11445129) achieves noiseless charge binning for a 2×2 group of pixel. By lumping charges together right in the pixel, no extra noise is added to the binned signal, thus achieving ideal performance from binning pixels together
Do you need a detector which is as big as possible? We are experienced in designing large sensors, up to the full size of a CMOS wafer, and whether 200 or 300 mm in diameter. We know how to handle the challenge of having high imaging performance as well as a great yield. See here for examples of wafer-scale design, including 3-side buttable
Photons are precious and we want to see them all. With back-side illumination BSI quantum efficiency values close to 100% can be achieved. When the light wavelength goes into the infrared, silicon is fairly inefficient but thick layer will help recovering the quantum efficiency. With fully depleted, thick layers, the high quantum efficiency is achieved without losing spatial resolution. DUV-optimised technology is particularly suitable for sensors for the semiconductor industry.
Detecting the invisible
Going beyond the visible range opens up many applications. On the longer wavelength side, efficient detection of near infrared can be achieved with thick, fully depleted silicon layer. To go beyond the silicon cut-off at 1.1 µm, readout integrated circuits are coupled with different materials. On the shorter wavelength side, efficient detection of ultraviolet light requires efficient back-thinning. This type of back-thinning would also help detecting low energy X-rays, with increasingly thicker silicon is needed there. And for very high energy X-rays, converting materials like scintillators are used.
Detecting charged particles
CMOS image sensors are now the detector of choice for detecting charged particles. Whether in scientific applications like particle or nuclear physics, or in lifetime or material science with electron microscopes, CMOS image sensors achieve ideal efficiency for the detection of charged particles, thus greatly simplifying the detectors for these applications. See here for our latest developments for transmission electron microscopy (TEM).
High speed imaging
Since their early days, it was recognised that CMOS image sensors would be the detector of choice for imaging sensors working at high-speed. IMASENIC is continuously pushing the frontier of high-speed imaging, bringing it also into the realm of large area sensors. We also have options for the proximity electronics of these fast sensors, generating hundreds of gigabit per second, like for the Sagara1212 sensor. Also for super-fast phenomena, at the nanosecond scale, we can offer solutions to your imaging problems.
Single Photon detection
Detecting photons is what imaging is about. We want to see them all, and take care of them one by one. Single photon detection is the ultimate goal of imaging. Single Photon Avalanche Detector (SPAD) generate a large digital signal each time a photon is detected. Very low noise sensors can also achieve photon counting capability. This is for visible light detection. For higher energy photos, more than one electron-hole pair can be generated by a single photon, so noise performance is eased, although it then becomes interesting to be able to say more about the photons, for example measure their energy. This is at the base of so-called “colour X-ray” imaging, for analogy with colour visible light imaging where colours correspond to photons of different energy.
Once there were passive pixels, then active pixels came and revolutionised the imaging world. Is now the time for the hyperactive pixels? The term was first used by CMOS imager’s inventor Eric Fossum in 2005, For many years, the word disappeared until it was then used by our CEO in 2018, see presentations. It was used to highlight the growing importance of new sensor types whose pixels contain more than a diode and a few transistors, as it occurs in active pixels. In these group can be found:
- self-triggering pixels
- event-driven pixels
- single photon avalanche detectors
- pixels with in-pixel ADCs
- artificial intelligence (AI) pixels with some local processing
Stacked technologies, where each sensor is comprised of two or three silicon layers, free the design constraints of conventional CMOS imaging technology, allowing separate optimisation of the sensing and processing units in a pixel, as well as in the periphery of the sensor. This also opens the way to having more complexity in the pixel without affecting the imaging performance.
In many applications there is a need to measure the distance of the object in a frame. Different technologies exist. Single Photon Avalanche Detectors (SPAD) can give the precise arrival time of photons, and the distance of an object can then be calculated from the time of flight of the light that it reflects. In IMASENIC we are also developing a different way of measuring the distance, based on a proprietary architecture of image sensors, the Depth Scanning Image Sensor (DSIS). In this way, it is possible to have a sensor that works like a standard imager when it comes to measure conventional 2D images, but that it can also detect high-resolution depth maps.