Wednesday, May 26, 2010

Sony Shrinks CCD Pixel to 1.43um

Sony announces industry smallest pixel size CCD in the recently published Cx_News vol. 60. 14.2MP ICX681SQW features 1.43um pixel, on par with the modern CMOS sensors in production. To improve the light collection efficiency Sony uses dual microlens:

Other notable features are reduced by 12% power consumption and 1/4 resolution mode for high sensitivity and high speed image capture (5.4fps at 3.55MP resolution).

Another Cx-News article
digests Tomoyuki Suzuki, Sony SVP keynote at ISSCC 2010 on Image Sensor Evolution and the Future of Imaging.


  1. Any idea whether the second (thick) microlens is a so-called inner lens or an extra lens similar to the top micro lens ? A.T.

  2. when we hold a lens focusing suns rays onto paper, paper burns... Is there any evidence of temp. increase due to use of microlens for increasing QE? Will a real, WDR sensor, prefer microlens or rather go without it?

  3. Image-sensor-enthusiastMay 28, 2010 at 4:00 PM

    Most of the light energy at the focussed spot is converted to e/h pair by silicon. So the temperature rise should be much lesser than in a insulating material like paper. Also, since silicon is a indirect band-gap material, higher wavelengths are penetrate deeper into the silicon. Typically UV is stopped by the front-lens as well.
    Secondly, the microlens does not increase the power of the incoming rays, but increases the irradiance. One should calculate the irradiance (by taking into account the lens and spot area) to see what irradiance (and hence eventually the temeperature rise) at the spot would be.
    For a WDR sensor, I think QE is defenitely important (especially for low-light levels). To avoid fill-factor loss, I think one should still employ micro-lenses.

  4. I cannot understand why they keep trying to improvement of CCD image sensor, because they have a great BSI, and I think BSI would be more performance than CCD, for instance it's high speed readout and lower power consumption. I know CCD can provide a global shutter function, which it is difficult to do with a CMOS. Is this CCD might be for a only global shutter? Please tell me any other reasons.

  5. image-sensor-enthusiastMay 31, 2010 at 4:11 PM

    For example, consider dark current: With "considerable effort", at present, a 4T CMOS image sensor can be made to exhibit a dark current of ~ 5 pA/cm2 (@ 30 C). A CCD, with its mature technology, can be made to have a dark current less than a tenth of what can be achieved with the 4T sensor (again, at 30 C, and having approximately the same QE as that of the 4T sensor). Its the maturity of CCD technology that still helps keep its advantage over cmos (maybe not for long..)

  6. Can you recommend some reference sources on the "considerable effort"? I am interested in learning more about techniques for dealing with dark current.


All comments are moderated to avoid spam and personal attacks.