Monday, July 19, 2010

Crosstalk Quantification, Analysis and Trends in CMOS Sensors

Applied Optics accepted for publication a paper on crosstalk:

Crosstalk quantification, analysis and trends in CMOS image sensors
Lior Blockstein, and Orly Yadid-Pecht, Fellow IEEE
The VLSI Systems Center/ Ben Gurion University, P.O.B. 653 Beer-Sheva, Israel.
Dept of Electrical and Computer Engineering, University of Calgary, Alberta, Canada.

Abstract
Pixel crosstalk (CTK) consists of three components, optical CTK (OCTK), electrical CTK (ECTK) and spectral CTK (SCTK). The CTK has been classified into two groups, pixel architecture dependant and pixel architecture independent. The pixel architecture dependant CTK (PADC) consists of the sum of two CTK components, i.e. the OCTK and the ECTK. This work presents a short summary of a large variety of methods for PADC reduction. Following, this work suggests a clear quantifiable definition of PADC. Three CMOS Image Sensors based on different technologies were empirically measured, using a unique scanning technology, the S-cube. The PADC is analyzed, and technology trends are shown.

The crosstalk is measured for 3.2, 2.2 and 1.75um pixels. There is also a great pixel crosstalk bibliography at the end of the article.

4 comments:

  1. cannot understand really what usage could be this paper. Any one can give more explanation ??

    ReplyDelete
  2. I think u can find the detailed info from pdf file, just click "paper on crosstalk" http://www.opticsinfobase.org/ao/upcoming_pdf.cfm?id=125402

    ReplyDelete
  3. Interesting paper.
    However it is somewhat impossible to separate optical from electrical crosstalk using this methodology, as spot is diffracted, particularly for small pixels.
    Moreover the Bayer pattern limits interpretation too.

    I suppose that Micron & Samsung are cited as they supplied test sensors... Other companies work on FDTD/coupling modeling.

    ReplyDelete
  4. Yes by using direct measurement using this method the optical and electrical processes happen simultaneously. Therefore measuring them separately is not possible.
    I am not sure what do you mean when you say "Moreover the Bayer pattern limits interpretation too."?

    ReplyDelete

All comments are moderated to avoid spam.