**IEEE Transactions on Electron Devices on**

**Solid-State Image Sensors**

In recent years there have been enormous advances in solid-state image sensors in CMOS (CIS). Improvements in pixel density, quantum efficiency, power dissipation, temporal noise, fixed-pattern noise are just some of the advancements that are permitting the widespread adoption of image acquisition in consumer appliances such as personal digital assistants, digital still cameras, camcorders, cell phone handsets as well as in automotive, industrial, medical and scientific applications. This special issue will provide a focal point for reporting these advancements in an archival journal and serve as an educational tool for the solid-state image sensor community. Previous special issues on solid-state image sensors were published in 1976, 1985, 1991, 1997, 2003 and 2009.

Areas of interest include, but are not limited to:

- Pixel device physics (New devices and structures, Advanced materials, Improved models and scaling, Advanced pixel circuits, Performance enhancement for QE, Dark current, Noise, Charge Multiplication Devices, etc.)
- Image sensor design and performance (New architectures, Small pixels and Large format arrays, High dynamic range, 3D range capture, Low voltage, Low power, High frame rate readout, Scientific-grade, Single-Photon Sensitivity)
- Image-sensor-specific peripheral circuits (ADCs and readout electronics, Color and image processing, Smart sensors and computational sensors, System on a chip)
- Non-visible “image” sensors (Enhanced spectral response e.g., UV, NIR, High energy photon and particle detectors e.g., electrons, X-rays, Ions, Hybrid detectors, THz imagers)
- Fabrication, packaging and manufacturing (stacked image sensors, back-side illuminated devices)
- Miscellaneous topics related to image sensor technology
**Submission Deadline: February 28**^{th}, 2015**Targeted Publication Date: January 2016**Guest Editor-in-Chief:

Prof. dr. Albert Theuwissen, Harvest Imaging, Bree, Belgium, and Delft University of Technology, Delft, the Netherlands.

Guest Editors:

Prof. dr. Eric Fossum, Thayer School of Engineering, Dartmouth, NH, USA,

Dr. Boyd Fowler, Google, Mountain View, CA, USA,

Prof. dr. Shoji Kawahito, Shizuoka University, Shizuoka, Japan,

Prof. dr. Pierre Magnan, ISAE, Toulouse, France,

Dr. Junichi Nakamura, Brillnics, Tokyo, Japan,

Dr. Johannes Solhusvik, Omnivision, Oslo, Norway,

Prof. Nobukazu Teranishi, University of Hyogo, Japan, and Shizuoka Univeristy, Japan,

Dr. John Tower, SRI International, Princeton, NJ, USA.

**Please submit papers by using the website : http:/mc.manuscriptcentral.com/ted , be sure to mention the special issue within the cover letter.**

For those who are interested, you can find all information at www.harvestimaging.com/forum.php

Albert, 30-08-2014.

]]> *Figure 1 : Effect of the slanted edge angle on the accuracy of the evaluation technique to characterize the MTF.*

As a message from the simulation results shown in Figure 1 : angles of the slanted edge between 2 deg. and 10 deg. are very well suited for the MTF analysis. This can be seen by checking their comparison to the ideal curve (see black dots in Figure 1). Once the angle is larger than 10 deg., the slanted edge method starts loosing accuracy. The simulation results obtained here are fully in line with the advices of the ISO standard, which suggests also to use an angle of the slanted edge between 2 deg. and 10 deg. Just to highlight the deviations, an extra insert is included in Figure 1 to shown the MTF behaviour around the Nyquist frequency.

One of the reasons why angles larger than 10 deg. give deviating results is the fact that the number of measurement points in a particular sensor column is getting smaller and smaller if the angle is getting larger. This can indeed result in nasty effects on the MTF values and can lead to aliasing effects in the sampling of the data. To avoid these type of issues, it is possible to take the data of multiple columns into account instead of using the data of just a single column. To show the overall working principle of oversampling, the spatial frequency responses (SFR) of 4 neighbouring columns are shown before (Figure 2) and after (Figure 3) multiplexing the data into a single SFR.

*Figure 2 : Spatial Frequency Response of 4 neighbouring columns.*

Figure 2 shows the data collected in 4 neighbouring columns, respectively columns 94, 93, 92 and 91, running vertically and crossing the slanted edge (which has an angle of 15 deg.). Four SFRs are obtained, all with a relative low amount of data points. As illustrated in Figure 3, the amount of data points can be increased by means of multiplexing the 4 curves of Figure 2 into one single SFR with 4 times the amount of samples.

*Figure 3 : Spatial Frequency Response of the 4-times oversampled data, after multiplexing the 4 curves shown in Figure 2.*

Based on this method of 4-times oversampling (also recommended by ISO 12233), the accuracy of the MTF measurement as a function of the angle of the slanted edge is checked again. The results of this exercise are shown in Figure 4.

*Figure 4 : Effect of the slanted edge angle on the accuracy of the evaluation technique to characterize the MTF, based on a 4-times oversampling of the SFR data.*

The result is quite remarkable : up to an angle of 20 deg. no deviation between the various obtained MTF curves can be seen, even not in the insert showing the MTF around the Nyquist frequency.

So oversampling when obtaining the SFR is absolutely recommendable, because it results in more data points to generate the SFR. And in the simulation result described in this and previous blogs, the “measured” modulation transfer function does not deviate from the theoretical one.

Albert, 01-08-2014.

]]>More information can be found on www.harvestimaging.com/forum.php

Albert, 25-7-2014.

]]>All organizational and logistics details are settled, so I can open the registration for the second Solid-State Imaging Forum. All information of the forum is shown on the website : www.harvestimaging.com/forum.php

Please notice, like last year, I will limit the number of seats to maximum 24. This will enhance the learning experience. Only in the case that we get substantially more registrants than this upper limit of 24, a second session can/will be considered. If you are interested to attend, early registration is recommended for two reasons :

1) for you : to make sure you get a seat (first come, first serve),

2) for me : to get as early as possible an idea whether a second session is needed or not. (A scond session can not be organized just a few weeks before the event will take place,)

Thanks, and looking forward to see you at the forum,

Albert.

15-07-2014.

]]>The result is shown in Figure 1.

*Figure 1 : Effect of the slanted edge angle on the accuracy of the evaluation technique to characterize the MTF.*

Shown are the MTF results obtained for a simulation of the angle being equal to 2 deg., 4 deg., 6 deg., 8 deg., 10 deg. and 12 deg. The ideal curve, obtained by the calculation of the sinc-function, is included as well. As can be seen from the curve :

- All evaluations based on an angle between 2 deg. and 10 deg. seem to fit very well to the ideal curve,
- The simulation result for an angle of 12 deg. shows some minor deviations from the ideal curve.

As a message from this simulation : angles of the slanted edge between 2 deg. and 10 deg. are very well suited for the MTF analysis. Once the angle is larger than 10 deg., the slanted edge method starts loosing its accuracy. The simulation results obtained here are fully in line with the advice of the ISO standard, which suggests also to use an angle of the slanted edge between 2 deg. and 10 deg.

Next time : how to implement oversampling and how to avoid aliasing effects during the measurements.

Albert, 04-07-2014.

]]>There are several good references describing the slanted-edge method, e.g. :

- M. Estribeau and P. Magnan., in SPIE Proceedings, Vol. 5251, Sept. 2003,
- T. Dutton et al. in SPIE Proceedings, Vol. 4486, 2002, pp. 219-246,
- P.B. Burns, in Proceedings IS&T, 2000, pp 135-138,
- S.E. Reichenback et al. In Optical Engineering, pp. 170-177, 1991.

This slanted edge method became an ISO standard, namely ISO 12233. This is one of the very few ISO standards for image sensor and/or camera measurements.

The technique of the slanted edge can be described as follows :

- Image a vertically oriented edge (or a horizontal one for the MTF measurement in the other direction) onto the detector. The vertical edge needs to be slightly tilted with respect to the columns of the sensor. The exact tilting is of no importance, it is advisable to have a tilt of minimum 2
^{o}and maximum 10^{o}w.r.t. the column direction. A tilt within these limits gives the best and most reliable results for the MTF characterization. - Each row of the detector gives a different Edge Spread Function (ESF), and the Spatial Frequency Response (SFR) of the slanted edge can be “created” by checking the pixel values in one particular column that is crossing the imaged slanted edge.
- Based on the obtained SFR, the Line Spread Function (LSF) can be calculated, the LSF is simply the first derivative of the SFR.
- Next and final step is calculating the Fourier transform of the LSF. This results in the Modulation Transfer Function, because the MTF is equal to the magnitude of the optical transfer function, being the Fourier transform of the LSF. Plotting the MTF as a function of the spatial frequency can be done after normalizing the MTF to its DC component and normalizing the spatial frequency to the sampling frequency.

(In one of the coming blogs more info will be given on further improvement and/or sophistication of this procedure.)

A very helpful strategy in understanding how this MTF measurement method works and to check the algorithms, is to run a simulation and create an artificial image with a slanted edge that is sampled by an artificial sensor (e.g. with a pixel fillfactor of 100%). Next the theoretical, geometric MTF can be calculated as a sinc-function of the spatial frequency, while the synthetic image can be used as the input image to evaluate the MTF by means of the technique explained above (ESF, SFR, LSF, MTF). Such a simple simulation tool can also be used to check the influence of the various system parameters on the measurement technique. An example of such a simulation is shown in the following figures.

First of all a synthetic image is generated that results in a slanted edged of 4 deg. w.r.t. the column direction. A region-of-interest (ROI) of 200 (H) x 300 (V) pixels is created around the black-white transition of the slanted edge. This synthetic image is shown in Figure 1.

*Figure 1 : ROI containing the slanted edge or black-white transition.*

A particular column is selected (in this example column number 96), and all pixel values in this column are recorded to generate the SFR or Spatial Frequency Response. The result of this operation is shown in Figure 2, with reference to the left vertical axis.

*Figure 2 : Spatial Frequency Response, being the values of the pixels present in column 96 of the image shown in Figure 1, and Line Spread Function, being the first derivative of the SFR.*

Next the LSF or Line Spread Function is generated, simply by numerically calculating the first derivative of the SFR. The LSF is shown in Figure 2 as well, with reference to the right vertical axis.

Once the LSF is known, the magnitude of the FFT of this LSF is calculated. Plotting the FFT magnitude versus spatial frequency results in the MTF of the sensor, as shown in Figure 3. Notice that the MTF is normalized with its value a zero input frequency (= DC), while the spatial frequency is normalized to the spatial sampling frequency of the sensor. In this simulation example, the pixel pitch is equal to 6.5 µm.

*Figure 3 : MTF of the simulated pixel (6.5 µm, 100 % FF), as well as the theoretical, geometric MTF of the same pixel.*

In Figure 3 and next to the outcome of the MTF simulation, also the theoretical geometric MTF of the pixel is shown (6.5 µm, 100 % FF), for comparison reasons. This geometrical MTF is calculated by means of the well-known sinc-function. As can be seen, both curves coincide very nicely, indicating that the slanted edge method and the algorithms used in the calculation seem to do the job that they were developed for !

Before showing real measurements, in the next blog(s) a few additional improvements of the slanted edge method will be highlighted.

Albert, 18-06-2014.

]]>In Figure 1, the results of the MTF measurement are shown.

*Figure 1 : Modulation Transfer Function of a monochrome device for various wavelengths of the incoming light.*

It is quite nice to see the influence of the wavelength of the incoming light :

- The absorption coefficient of silicon for “red” photons, or photons with a lower energy, is relatively low. The absorption depth for light with a wavelength of 630 nm can reach a few microns. So part of the electrons generated in the silicon will be generated below the depletion region of the photodiodes, and before these electrons can get collected by the photodiodes, they need to “travel” in the neutral bulk/epi-layer. Because there is no electrical field present in these regions to guide the electrons to the right photodiodes, the chance that these electrons finally land in a neighbouring pixel is relatively large. In this way the contrast in the image is reduced, so is the MTF,
- The absorption coefficient of silicon for “blue” photons, or photons with a higher energy, is relatively large. The absorption depth for light with a wavelength of 470 nm is just a few tens of a micron. So most of the electrons generated in the silicon will be generated within the depletion region and the chance of diffusion of these electrons to neighbouring pixels is limited. The contrast in the image is not reduced by the effect described above for the “red” photons, neither is the MTF,
- The green light, with a wavelength of 525 nm, has an absorption coefficient situated between the red light and the blue light. So not surprising that the MTF for the green light is lying between the blue and red results.

The effect explained here by means of the MTF measurements is also known as electrical cross-talk. The loss in contrast or loss in MTF is due to the diffusion of electrons. The effect is also illustrated in Figure 2.

*Figure 2 : Illustration of the electrical cross-talk.*

Figure 2 shows a cross section of a hypothetical image sensor with an RGB filter. Illustrated is the fact that the “red” photons can penetrated much deeper into the silicon than the “blue” ones. This is the origin of the larger electrical cross-talk for the light having a longer wavelength.

To conclude a few numbers :

- Absorption coefficient for a “red” photon (630 nm) = 4000/cm, resulting in an absorption depth of 2.5 um,
- Absorption coefficient for a “green” photon (525 nm) = 10,000/cm, resulting in an absorption depth of 1 um,
- Absorption coefficient for a “blue” photon (470 nm) = 20,000/cm, resulting in an absorption depth of 0.5 um.

Albert, 19-04-2014.

]]>Mark now already your agenda for the second solid-state imaging forum, scheduled for Dec. 11-12, 2014.

After the succesful first forum in 2013, I am happy to announce a second one. Also this second solid-state imaging forum will be a high-level, technical, short course focusing on one particular hot topic in the field of solid-state imaging. The audience will be strictly limited to 28 people, just to stimulate as much as possible the interaction between the participants and speaker(s). The subject of the second forum will be : “Advanced Digital Image Processing”.

More information about the speaker and the agenda of the second forum will follow in the coming weeks, but I wanted to share this announcement with you as early as possible to make sure you can keep your agenda free on these days.

Albert,

April 6th, 2014.

In Figure 1 the result of the MTF measurement is shown.

*Figure 1 : Modulation Transfer Function for two settings of the lens F-number.*

It is quite nice to see the influence of the F-number :

- a low F-number is referring to a large lens opening (= a lot of light goes to the sensor, a short exposure time is needed), and in that case the incoming light is reaching the sensor with a large chief-ray angle (= deviation from the normal),
- a large F-number is referring to a small lens opening (= much less light goes to the sensor, a long exposure time is needed), and in that case the incoming light is reaching the sensor with a small chief-ray angle (= almost perpendicular to the sensor).

Light that is perpendicularly reaching the sensor will suffer less from optical cross-talk in comparison to light that is reaching the sensor under a certain angle and is deviating more from the normal. More (optical) cross-talk does result in less contrast between neighbouring pixels, thus lowering the MTF for larger spatial frequencies. And this effect is observed in Figure 1 !

Next time something about colour and MTF.

Albert, 25-03-2014.

]]>