Source: laserfocusworld.

com (CMOS detectors: TCAD simulation of image sensors

catches light in silicon 05/01/2006) - IJU JACOB, ALP GENCER, GERGO LETAY,
AND PAUL PFAEFFLI
Section: TCAD-driven optimization
The challenge is to extract the relevant data from the wealth of information produced by simulation to give
statistically meaningful results. As a first step, a multidimensional response surface model is fitted to the data,
providing correlation information for process variations, design parameters, and observation parameters-
collectively called a process compact model. While TCAD simulations typically take hours, the PCM can be
evaluated immediately.

Source: stemmer-imaging.com/en/knowledge-base/vorder-oder-rueckseitig-beleuchtete-sensoren/

Front side or back side illuminated sensors

Conventional sensors are designed to convert the light falling on the front side of the sensor to electrons.
To maximise sensitivity we are seeing additional optical lightguides added to microlenses to channel as much
light as possible to the limited photosensitive area, however a new approach is now being seen in some of the
latest small pixel sensors that overturn the pixel design to illuminate the pixel from the back side (Back side
illuminated or BSI), so the control electronics do not limit the light sensitive area. Due to the structure of the
silicon pixel, light in a BSI sensor has to travel further into the silicon to be converted into electrons and is
therefore less efficient at converting the light (lower quantum efficiency or QE).

Frontside illuminated sensor with microlenses Frontside illuminated sensor with light-guides

Backside illuminated sensor

Source: A. Crocherie, P. Boulenc, J. Vaillant, F. Hirigoyen, D. Hérault and C.

Tavernier, “From photons to electrons: a complete 3D simulation flow
for CMOS image sensor,” in IEEE 2009 Int. Image Sensor Workshop
(IISW).
We couple Optical simulations with TCAD process and device simulations to combine both approaches benefits.
comparison is made between our coupled Optical/TCAD simulations results and real Image Sensors optical
measurements.

As pixel size shrinks in CMOS image sensors markets, the crosstalk effects become more and more critical.

Optical part

For small pixels that are used nowadays in CMOS image sensors, diffraction effects can substantially affect light
propagation and photon collection. Thus, ray-tracing description is not accurate anymore. We chose to adopt an
electromagnetic simulation tool based on Finite Difference Time Domain (FDTD). FDTD is a fully vectorial
method the response of the system over a wide range of wavelengths could be obtained in a single simulation.
wavelength of interest (450nm, 532nm, and
633nm).

Materials used, like oxides, nitrides, silicon,

color filters, are based on
STMicroelectronics experimental database.

Accuracy of simulation: mesh step

N~10

photogenerating rate of electron/hole pairs given from optical to electrical tool.

TCAD simulations
Process simulations that deal with modelling the frontend- of-line (FEOL) steps in the sensor manufacturing.
Then,
device simulations are the electro-optical behaviour. Back-end-of-line (BEOL)= interconnect design Front-end-of-
line (FEOL)=process design. To accurately consider electrical crosstalk i.e. electrons that may flow from one
pixel to all its neighbors.
Experimental protocol consists of the three following steps:
pixels Reset, Illumination, Readout.
The Reset is performed by initializing electron Fermi level to an arbitrary high value in the pixels in order to
empty all the photodiodes. Then, thermal generation of electron-hole pairs (Shockley- Read-Hall model with
Scharfetter doping dependence
fills the photodiodes by dark current.
Light is then turned on with the Lumerical Optical Generation map in the 3D Bayer and integration is performed
for tint=66.7ms. Finally, the 3D electron density is integrated in each pixel 1ms after light is stopped to obtain the
number of collected electrons. Another simulation without light is performed to obtain the number of dark current
electrons to be subtracted in the final pixel quantum efficiency (QE) calculation :

where λ is the wavelength, n is the electron density, tint is the

integration time and Ф is the incident photon flux.
COUPLED OPTICAL/TCAD SIMULATIONS RESULTS
Bayer filter
From Wikipedia, the free encyclopedia

The Bayer arrangement of color filters on the pixel array of an image sensor

Profile/cross-section of sensor

A Bayer filter mosaic is a color filter array (CFA) for arranging RGB color filters on a square grid of photosensors.
Its particular arrangement of color filters is used in most single-chip digital image sensors used in digital
cameras, camcorders, and scanners to create a color image. The filter pattern is 50% green, 25% red and 25%
blue, hence is also called BGGR, RGBG, GRGB, or RGGB. [1][2] [3] [4]

At each wavelength, a scaling factor is applied on simulated Blue, Green and Red QE(λ). It is calculated as the
ratio between measured Bayer average QE(λ) and simulated Bayer raw average QE(λ). This factor reflects
imperfections that are not considered in optical simulations (like interface roughness) and is consequently
wavelength dependent.

CONCLUSION
In very small pixels, crosstalk becomes one of the main performance limiting factors, with strong 3D and
neighboring effects. Moreover crosstalk arises from two distinct phenomena (optical and electrical) that are in
strong interaction. Therefore, an approach that couples dedicated simulations tools is mandatory. Although
some adjustments are required, like global optical stack transmission scaling and 3D corrections to 2D doping
distributions, electro-optical performances of four different sensors can be reproduced within a few percents
error.

SCMOS
Source: Wikipedia
sCMOS (scientific Complementary metal—oxide—semiconductor) is a technology based on next-
generation CMOS Image Sensor (CIS) design and fabrication techniques. sCMOS image sensors offer
extremely low noise, rapid frame rates, wide dynamic range, high quantum efficiency, high resolution, and a
large field of View .
While back-illuminated electron-multiplying CCD (EMCCD) cameras are optimum for certain uses that
require the lowest noise and dark currents, sCMOS technology, with its higher pixel count and lower cost, can be
the choice for a wide range of high- precision applications. sCMOS devices can capture data in a global-shutter
“snapshot” mode over all the pixels or rectangular subsets of pixels, and can also operate in a rolling-shutter
models]
The sCMOS sensor's low read noise and larger area provides a low-noise, large field-of—View
(FOV) image that enables researchers to scan across a sample and capture high-quality images. The New York
University School of Medicine uses sCMOS cameras for their research. With that sCMOS-technology, they have
a new optical method to study biological molecules and processes in real-time at nanometer scale.

Source: sCMOS Technology- ANDOR

sCMOS is poised for widespread recognition as a true
scientific grade CIS, capable of out-performing most scientific
imaging devices on the market today. overcoming the 'trade-
offs' that are inherent to current scientific imaging technology
standards, and eradicating the performance drawbacks that
have traditionally been associated with conventional CMOS
imagers.
The fact that the sCMOS device can achieve down to 1
electron rms read noise while reading out 5.5 megapixels at
30 frames/sec renders it truly extraordinary in the market.
By way of comparison, the lowest noise Interline CCD, reading out only 1.4 megapixels at 16 frames/sec would
do so with ~ 10 electrons read noise.
The low noise readout is complemented by up to 30,000:1 dynamic range. Usually, for CCDs or EMCCDs
(~1,800:1) to reach their highest dynamic range values, there needs to be a significant compromise in readout
speed, yet sCMOS can achieve this value while delivering high frame rates. The unique dual amplifier
architecture of sCMOS allows for high dynamic range by offering a large well depth, despite the relatively small
6.5 μm pixel size, alongside lowest noise.

The sensor features a split readout scheme in which the top

and bottom halves of the sensor are read out independently.
Each column within each half of the sensor is equipped with
dual column level amplifiers and dual analog-to- digital
converters (ADC), represented as a block diagram in below. This architecture was designed to minimize read
noise and maximize dynamic range simultaneously.
The dual column level amplifier/ADC pairs have independent gain settings, and the final image is reconstructed
by combining pixel readings from both the high gain and low gain readout channels to achieve a wide intra-
scene dynamic range from such a small pixel pitch.

Each pinned-photodiode pixel has 5 transistors (‘5T’ design), enabling the novel
‘global shutter’ mode (The time to transfer charge after the exposure is complete is less than 1μs, rendering the
sensor useful for fast electronic shuttering and ‘double exposure’ techniques such as Particle Imaging
Velocimetry (PIV)) and also facilitating correlated double sampling (CDS) and a lateral anti-blooming drain. The
sensor has anti-blooming of >10,000:1, meaning that the pixels can be significantly oversaturated without charge
spilling into neighboring pixels. It is also possible to use the anti-blooming capability to hold all or parts of the
sensor in a
state of ‘reset’, even while light is falling on these pixels. The sensor is integrated with a microlens array that
serves to focus much of the incident light per pixel away from the transistors and onto the exposed silicon,
enhancing the QE.

New Scientific CMOS Cameras with Back-Illuminated Technology- Princeton Instruments

(Low-Light, Time-Resolved Imaging and Spectroscopy Applications Get New Boost)

For more than five decades, CCD cameras have provided the single-photon sensitivity and moderate frame rates required for
scientific imaging and spectroscopy applications. More recently, scientific CMOS (sCMOS) cameras that are capable of achieving
low read noise and higher frame rates have become an alternative to CCD cameras in several applications. However, the first
generations of these sCMOS devices fall short on sensitivity owing to their front-illuminated architecture, which imposes a
fundamental limit on their quantum efficiency (i.e., the fraction of incident photons detected in each pixel). Aided by the latest
CMOS fabrication technology, sCMOS devices can finally be created with a back- lluminated sensor architecture. As a result,
sCMOS sensors are now capable of CCD like quantum efficiency (>95%) and dynamic range without compromising the low read
noise and high frame rates for which they are known. Back-illuminated sCMOS camera technology is a serious contender as an
optical detector for myriad applications, including hyperspectral imaging, astronomy, cold-atom imaging, quantum imaging,
fluorescence spectroscopy, and high-speed spectroscopy.

A Biologically Inspired CMOS Image Sensor (mukul Sarkar)-Springer (2013)

CMOS image sensors on the other hand use generic fabrication process, and in principle work very similarly to
CCD sensors. In addition to being less expensive when compared to CCDs, they also allow for the integration
of processing circuitry with the photoreceptor. CMOS image sensors also allow for random accessibility,
enhancing the output bandwidth. The image quality of a CMOS image sensor is however not at par with that of a
CCD as the in-pixel and in-column circuitry often produces a “fixed pattern noise” in the image. The CMOS
image sensors need lower voltage than the CCD sensors, and thus lower power consumption. The CMOS
image sensors also allows for random access and thus region based readout is possible, allowing the CMOS
image sensors to operate at very high speeds. The CMOS image sensors are easily capable of running at over
500 frames per second at megapixel resolution.
CMOS image sensor have yet not been able to replace CCDs in the applications where very low noise operation
is desired. The CMOS image sensors with their many amplifiers at each pixel, suffers from the fixed pattern
noise, which arises due to the process variations while fabricating the amplifiers. CCDs are less susceptible to
noise compared to CMOS and thus create higher quality, lower noise images. The CMOS image sensors are
also less light sensitive compared to the CCDs. In CCDs, each individual units are only collecting light while the
amplifiers inside the pixel in the CMOS image sensor reduces the light collecting area and thus are less
sensitive to light. However the CMOS image sensors are being actively used as they are much less expensive
and allow for the integration of processing circuitry with the photoreceptor, allowing for random accessibility,
enhancing output bandwidth.