Monday, May 16, 2011

CMOSIS Deals with BSI Passivation, Samsung Works on Separation between RGB and IR

CMOSIS filed BSI patent application US20110101482 proposing to create inversion layer on the backside by depositing a thin sapphire layer on the backside (also known as alumina or Aluminium Oxide):


The Al2O3 film contains a stable density of negative fixed charges. Typically, this density is in the region of 9E12/cm2. This creates an accumulation layer in the p-type substrate of the image sensor. This accumulation layer can effectively passivate the surface. It ensures that no photoelectric charges are trapped near the back surface and that thermally generated charges recombine. This reduces the dark current in backside illuminated image sensors. Other possible materials for the passivation layer include: Aluminiumfluoride (AlF3); Aluminiumoxynitride (AlON); Zirconiumoxide (ZrO2) under certain conditions; Calciumfluoride (CaF2). The Al2O3 film can serve as AR coating as well.

Samsung filed another application on RGB and depth sensor combination. The US20110102547 talks about RGB and IR color filter arrangement aimed to separate IR and RGB paths. Few different filter designs are proposed to achieve separation like this:

10 comments:

  1. Al2O3 has been used for BSI passivation in CCD for a long time. See the work done by Prof. Lesser at University of Arizona on scientific CCD devices and presented numerous times. What is the novelty of the patent?

    ReplyDelete
  2. Make Buzz to show they are imaginative...

    ReplyDelete
  3. @ "Make Buzz to show they are imaginative..."

    I suspect you will find no such thing. This is probably Vladimir sharing a couple of results he found interesting from what is likely a regular search for relevant recently-published patent applications.

    While I can't comment on the novelty of the idea since I haven't read the application, I think CMOSIS deserves the benefit of the doubt. I don't recall seeing any hyperbole-laden press releases from this company, and as a small player CMOSIS is constantly under pressure to deliver results or go under. There aren't other business divisions and economies of scale to keep it afloat, and that can be a really challenging situation to work in every day.

    ReplyDelete
  4. Indeed, the patent application was quietly filed with no buzz at all. The fact it's presented in the blog has nothing to deal with CMOSIS.

    As for the novelty, may be the fact that it's applied to CMOS sensors rather than CCD is enough for the patent office to grant it?

    ReplyDelete
  5. Well keep in mind that these are both just applications. First of all, nothing will happen for a long time due to the immense backlog at USPTO. Unless Congress relaxes its budgetary rules on PTO (e.g. letting PTO spend its fees on itself instead of Congress sweeping those funds to the general budget) the future just gets bleaker for turnaround time and fundamental change.
    Second, PTO examiners are demonstrably not-well-versed in image sensor technology so there is always a good chance a patent will be granted despite prior art. It is incumbent on CMOSIS to report all prior art that they are aware of so hopefully all will be mentioned. Failure to report relevant known prior art can invalidate a patent.
    Third, my guess is that some day in the future, if the patent is granted, some court case will decide if CMOSIS claims are truly valid or not. It would be ridiculous for CMOSIS not to try to patent technology, even if the claims are narrowed. Just imagine how one would feel if you decided not to file, and a competitor later filed and received a patent!

    ReplyDelete
  6. By the way, PTO is looking for image sensor specialists. CDM, maybe you can work for them from home?! Others too, you should consider this!

    ReplyDelete
  7. We recommande Eric as final patent application examinator, it will be surely better!

    ReplyDelete
  8. Hahahahahahahaha!

    That's a good one, Eric.

    ReplyDelete
  9. This is not a joke at all. I think that experienced people like Eric can make all the patents much more credible. No computer databases can replace the mental memory of an experimented guy!!!!

    ReplyDelete
  10. I don't doubt that the involvement of competent engineers in the examination process would improve the quality of issued patents. It's just that I've dealt with the US PTO directly as an inventor, and the amount of unnecessary frustration I endured leaves me less than enthusiastic about becoming an employee.

    While I'm not inclined to rule out any possibility, let's just say I'm much more likely to make the US team for the 2020 Olympic Games than to become a remote examiner.

    ReplyDelete

All comments are moderated to avoid spam.