Complementary Metal-Oxide-Semiconductor (CMOS) Image Sensor (⚠️WORK IN PROGRESS⚠️)

A CMOS sensor (Complementary Metal-Oxide-Semiconductor sensor) is an electronic image sensor that converts light into digital signals using the photodetector principle. Photons are captured and converted into electrons by the photodiode, which are then converted to voltage. An analog-to-digital converter (ADC) then converts this analog information to digital information, which is further processed to create the final image. This technology is used in most modern digital cameras, smartphones, webcams and many scientific imaging devices¹⁾.

A microlens array is a layer of tiny lenses positioned above the pixels of an image sensor. Each microlens is aligned with a single pixel and concentrates the incoming light onto the photodiode, which is the light-sensitive region of the pixel.

There is three types of vertical photodiodes suitable for fabrication in standard N-well CMOS processes : photodiodes using the N-well/P-sub, P⁺/N-well, and N⁺/P-sub junctions.

The best type of junction depends on the CMOS structure ⁶⁾ :

• for 0.7 $\mu$m CMOS process : N⁺/P-sub
• for 0.35 $\mu$m CMOS process : N-well/P-sub
• for 0.18 $\mu$m CMOS process : P⁺/N-well

The most-used type of photodiode for CMOS is the pinned photodiiode. ⁷⁾ The pinned photodiode has a p⁺/n-well/p-sub architecture, with a 4T pixel structure.

The potential well in a CMOS image sensor is the electron storage region formed inside the N-type pinned photodiode. The combination of the N photodiode and the shallow P+ pinning layer creates a controlled electrostatic minimum that stores photoelectrons with low noise and allows efficient transfer during readout. This structure is the foundation of essentially all modern CMOS image sensors.

The potential well will determine key sensor of the CMOS image sensor such as the Full Well Capacity (FWC), the Charge Transfer Efficiency (CTE), the dark current, the dynamic range, etc.⁸⁾

This image shows how a photodector works. The photon penetrates the silicon and create a electron-hole pair connection. The photonic penetration depth on silicon, i.e., the light absorption, depends on the wavelength. This means that the longer the wavelength λ [m] of a photon, the loweris its energy, and the further it can delve into silicon. ⁹⁾

Silicon photodetectors can create photogenerated currents for impinging light with wavelengths across the complete visible range. The produced photocurrent is proportional to the intensity of the incident light and is given by : $$I_{ph} = \frac{e \times QE \times \lambda \times P_i}{hc} = \frac{e \times QE \times P_i}{hv} $$

where I_ph [A] is the photocurrent, e [C] is the elementary charge, λ [m] is the wavelength of the incident light, QE [%] is the quantum efficiency, P_i[W] is the incident optical power, h [J.Hz⁻¹] is Plank’s constant, and c [m.s⁻¹] is the velocity of light in a vacuum. The quantity $E = \frac{hc}{\lambda} = h\nu$[eV] is the energy of a photon and ν [Hz] is the frequency of the photons. Formally, the quantum efficiency is determined by the ratio of the generated electrons N_e to the incident photons N_ph within the photodetector :

$$QE = \frac{N_e}{N_{ph}} = \frac{\left(\frac{I_{ph}}{e}\right)}{\left(\frac{P_i}{h\nu}\right)} = R \times \left(\frac{h\nu}{e}\right)$$

where R [A/W] is the responsivity of the photodetector. Responsivity is very important because it relates the generated photocurrent Iph [A] with the impinged optical power P_i [W], e.g., R = I_ph/P_i [A/W]. Hence, after acquiring the photocurrents, the primary physical quantity that was obtained is the responsivity. Subsequently, the quantum efficiency is derived from Equation (2)

3T-APS, 4T-APS ¹²⁾

More for early sensors ¹³⁾ Since its invention in 1993, CMOS image technology has evolved. The first architectures were front-illuminated, meaning the microlens and color filter were on top, followed by metal wiring for interconnects and the photodiode on the bottom. Since light enters the image sensor through the metal layers, some light information is reflected and lost before reaching the photodiode. This affected the performance of front-illuminated sensors, but Sony Corporation solved the problem by moving the photodiode to the top, next to the color filter. This architecture is known as the back-illuminated (BI) CMOS image sensor, which greatly improved the sensor’s performance.

Current designs ¹⁵⁾ Following the back illuminated sensor, the idea to stack the pixel and the logic circuit sections was proposed to reduce the size of the sensor in the X and Y directions. The pixel section containing the photodiodes was placed on the top and the logic circuitry was moved to the bottom of the architecture on the supporting substrate. This is called the Stacked Back-Illuminated CMOS Image Sensor, which was proposed by Sony Corporation in the year 2012.

In the popular Bayer color pattern (RGGB), each color “pixel” consists of a 2-by-2 array of photo-detectors where two are designated to green channel and each of the other two are for the blue and red channels, as shown in figure 4a.