Saturday, June 04, 2011

Invisage Holds Promise for Machine Vision

Test & Measurement World published an article about Invisage QuantumFilm with QuantumShutter sensor. The article seems to be based on an interview with Michael Hepp, InVisage’s director of marketing. Few advantages for machine vision market are presented:

  1. Adjustable bandgap of QuantumFilm tunes the sensitivity for specific wavelength range
  2. QuantumShutter (a nice word for global shutter) eliminates rolling shutter artifacts
  3. Improved QE in range of 80 to 90%
  4. Process is simpler than BSI

Hepp said that the QuantumFilm sensors designed for camera phones will have a 1.4um pixel size, and next-generation devices are planned at 1.1um. Invisage's 1.1um pixels can use a 110nm process. "In a 1.4-micron pixel, we can achieve a 12,000-electron well," says Hepp.

TSMC is expected to sample chips made with the process this summer, said Hepp.


  1. Quantum film, Black Silicon, etc. All these things are promising and exciting, but we didn't see any concret any realizations. I think that these guys should show the world something, it's a so long time....

  2. Have these technical journalists seen the real stuff? Or they just transcribe what Invisage guys said?

  3. I once saw a demo from Invisage. I saw a b/w low light image that was grainier than the image I saw from my color iphone that was used to photograph the same target under the same illumination. Took a snap side by side in other words.

    It works but they have more work to do from my vantage point.

  4. Personally, I hope the quantum film photodetector can deliver as promised. It would be great to get away from silicon and the non-scalability of QE.

    Obviously as previously discussed, dark current QE, lag etc all need to perform as advertised.

    I also want to know the post CDS read noise for the pixel circuit and photodetector. This is where the rubber meets the road!

    I don't understand why TSMC will be sampling the chips. One would think it would be Invisage.

  5. @ "I don't understand why TSMC will be sampling the chips. One would think it would be Invisage."

    I noticed this too. I'd guess this is just a misinterpretation of T&M World. On the other hand, it quotes Hepp on that. May be the intention was like this:

    TSMC is expected to [ship] sample chips...

  6. I saw an image from SiOnyx during SPIE DSS at Orlando. Even the image is rough and can give no interpretation at all, but we have seen an image from a test chip under probing machine. The marketing guy said that they would like to get some incoming now. This is very positive. Personaly I'm waiting it for long time.

    During the ISE at London this year, A Simens guy presented also a polymer based photosensitive material with quantum dots added. This material is capable to sense in the SWIR band.

    If InVisage guys read this blog, please understand that a lot of technical guys from imaging field waiting for some demonstration of your technology!

  7. Eric, if it is non-silicon why does Invisage need TSMC? How capable is TSMC in making this kind of device?

  8. I'm concerned about pixel bleed. In one of Bob Street's (Xerox) papers he was discussing topside photodiodes for TFT and he noted that you have to be very careful in how you set up the boundaries of a pixel or you get significant bleed from non-seperate photodiodes.

    The other question I have is really how big is the advantage of larger 2d area capacitors. My understanding was that deep trenches in DRAM had pretty much solved that problem. If well depth is really a desirable feature, it seems like we will just see deeper wells in silicon.

    Conceptually I really like the technique, though I've sat through enough talks on nanodot preparation to be skeptical about thier process control. I remember a talk from 3 or so years ago where someone was sharing the challenges of upscaling nanodots for lighting. It is a major challenge to step up from a few grams in the laboratory to kgs. Evidently one of the real chalenges was agitation and temperature control since these structures are pretty much only metastable.


  9. Also to respond to the TMSC question, this is still a CMOS device under the QDs, it still needs masking and doping and everything thats normally associated with making a chip. From my reading they are taking a 'standard-esque' cmos pixel and ripping out the photodiode/detector. Then they apply thier film as a post process. So they still need to make an array of pixels with A-D and amplification.

  10. @Sloan. You must be new to this field. It takes more than reading a few papers and sitting through a few talks before you jump in. Deep trenches? No.

  11. alright, no problem there. Everyone has to start somewhere.

  12. @ "It takes more than reading a few papers and sitting through a few talks before you jump in."

    I think that reading some papers and attending some talks is jumping in, and also that the more people who do so the better.

    My son who is just finishing the second-grade is the youngest kid in a class that is rife with red-shirting and superficial competitive metrics. He's had people who are older and more experienced treat him with disdain and apathy all year long, and I can tell you both are interest-killers. He doesn't deserve that, and neither do constructive blog participants.

  13. @ Anonymous 7:37AM

    I must be new to this field as well, as I have only been doing CIS since the mid 90's.

    I never had the privilege of working with deep trenches, but always thought it would be a good idea for increasing the full well, provided one could control the dark current of the etched sidewall.

    You say a categorical "NO" to deep trenches. Can this newbie as "Why"?


  14. What I'm curious about is the longevity of the QD film. From my understanding, all Invisage does is asking TSMC to deposit their film over gates, and that sounds like a fine business model. QD films are nothing new either. The question is about the process and stability.

    Specifically, if the lifetime of the film turns out to be <2years, it might be Ok for cell phones, but not suitable for industrial applications. Any knowledge on that to share?


All comments are moderated to avoid spam and personal attacks.