Tuesday, May 18, 2010

Aptina Explains its BSI-FSI Strategy, Announces New Sensors

Business Wire: Aptina made few big announcements today explaining its A-Pixel technology, FSI and BSI pixel plans and also launched two new sensors. I was given a media kit accompanying these announcements. The pictures below are from this kit.

The corporate presentation says that Aptina has achieved $547M sales in 2009 and the company has about 700 employees. Aptina's historical milestones are shown on the slide below:

The most interesting part here is an acquisition of GG&C in 2007. I do not recall any official announcements about that. A quick search on Linked-In shows that GG&C developed a telescope with extended depth of field.

Market share slide is rarely seen chart based on revenue, rather than units. It shows quite different picture from the unit market shares:

Next slide shows conflicting requirements of stills photography and video, with video mostly demanding low light:

This results in Aptina FSI vs BSI approach:

Here comes the explanation of Aptina A-Pix advantages - combination of deep photodiode with lightguide results in improved light sensitivity:

Next slide says that BSI is not really necessary at 1.4um pixel node, if one has 65nm process instead. First BSI products will be announced in the second half on this year at 1.1um node:

3rd generation A-Pix features better fill factor dues to thinner metals in 65nm process:

Aptina FSI vs BSI whitepaper says about BSI problems:

"For 1.4 micron BSI pixels, QE is typically in the 50-60 percent range with crosstalk in the 15-20 percent range. The combination of BSI’s high QE and somewhat degraded crosstalk at 1.4 micron results in a net overall image quality that is comparable to FSI for 1.4 micron pixels.

Today, 1.1 micron BSI pixels are still in the early stages of development, but when they are production-ready, they would be expected to have a QE approaching 50-60 percent with crosstalk in the 10-30 percent range. These 1.1 micron BSI pixels should be outperforming 1.1 micron FSI pixels at that point due to the fabrication challenges in shrinking FSI pixels to 1.1 micron.

A tipping point for BSI will be the 1.1 micron pixel node where FSI will likely be unable to achieve the market-required performance – necessitating a transition to BSI for applications that require this smaller pixel.

Additionally, Aptina announced two new sensors based on 3rd generation 1.4um A-Pix technology: the 3MP SOC MT9T113 and the 5MP MT9P017. Both sensors have small die size suited for integration into 6.5 x 6.5 mm modules, parallel and MIPI interfaces and integrated AF VCM driver.

The 1/5-inch MT9T113 SOC features JPEG thumbnail support, Scalado SpeedTags and 720p/30fps HD video. The 1/4-inch MT9P017 has adaptive noise reduction, 2D dynamic defect correction, and supports 1080p/30fps or 720p/30fps HD video.

Update: There is Youtube video with A-Pix explanations by Gennadiy Agranov, Aptina's VP of Imaging Technology:

Update #2: EETimes published an article on the new Aptina announcements.


  1. Albert TheuwissenMay 18, 2010 at 10:26 PM

    I am happy to see some real images and a (tiny) bit of technical details. Much better than the "blind" marketing stuff of others.
    Wishing you success and a lot of fun in pursuing pixels of 0.9 um. What a world of difference compared to our CCD work at the time I started at Philips in 1983. We were using design rules (!) of 0.9 um but also that work was a lot of fun !
    "There's more to the picture than meet the eye" (Neil Young 1978)

  2. Nice movie also.

    Does anyone know which is the large sensor that they show at timestamp 0:31 in the movie? A wafer with 36 dies only. The die must be larger than an APS size 24x16 mm sensors for SLR camera's. This obviously isn't gone have 1.4 micron pixels...

    The largest Aptina sensor I know are some high speed sensors, but they are smaller than the one shown in the movie.

  3. TSMC has shown such "light-guide" at a conference, maybe 5-6 years ago at Hawai. Do they have such kind of process option too ??

    thanks !

  4. Lightguide is a natural extension of copper-based process, while TSMC FSI offerings are alumina-based. While it's still possible to integrate lightguide into alumina process, I do not think that TSMC offers that option.

  5. it looks like aptina mainains the title of innovation leader.

  6. Just wanted to mention that we (Aptina) are doing light guide both on Aluminum process and Cooper process

  7. Nice work Gennadiy! I also enjoyed the animation.

    Hey, wouldn't the pixel work alot better if those photoelectrons were collected by drift instead of allowed to just dance around? :)

  8. Hopefully this would kill some of OVT's hot air

  9. Thank you Eric. You mean Tango rather than Charleston? :)

  10. (cont)... Talking seriously, you are right. We are continuously working on Electric Field Engineering, so that “good electrons” (the ones that increase quantum efficiency) moves toward the accumulation site, and the “bad electrons” (the ones that create crosstalk) are pushed away. When we are talking about deep PD, it doesn’t necessary mean that metallurgical p-n junction is very deep in the Si. But rather electric field is engineered such way that electrons generated in the deep portion of the Si drift toward the PD.


All comments are moderated to avoid spam.