The biennial IISW, this year on its North American leg, was held in mid-June in Utah at the Snowbird Resort. As per usual, the attendance was capped at 150. The volume of research presented was impressive. Three invited talks, 46 regular paper presentations, and 46 short poster presentations were presented in a single track over four days. The major topics of discussion included small pixel development, high-speed imaging, SPADs, global shutter, high dynamic range, binary pixels, and medical imaging. The usual suspects included key inventors of this technology sector; this year we were lucky enough to hear from the pioneers of digital imaging, including Mike Tompsett and Gene Weckler.
Chipworks’ extensive library of image sensor benchmarking reports enabled me to produce a trends paper highlighting innovative technology elements in use in high-volume consumer electronics. The paper and presentation (which will be made publicly available at a later date) include device highlights of recent DSLRs, MILCs, compact cameras, smartphones, and tablets. Some emerging technologies we’re seeing enter production include on-chip phase detection pixel arrays, 3DICs (stacked chip CIS), vertical gate transfer transistors, and several developments in back illuminated CIS.
Note that we are all living off our paper copy of the proceedings until September, when the IISS will electronically publish the conference papers. It is not practical to regurgitate the full scope of the workshop, although Albert Theuwissen posted daily summaries on his blog. To give a preview of the September release, I’ll comment on two papers related to small pixel back illuminated device research. As back illuminated CIS technology matures, technologists are continually improving the performance of back anti-reflective layers and optical stacks. Samsung presented its work on a “lensed color filter” and gapless microlens design for 1.1 µm generation pixels. The focus of this work was on lowering cross talk and increasing SNR for small pixels. Essentially, the microlens etch back extends down into the color filter so the microlens photoresist and a portion of the color filter contribute to the microlens dome. While the color filter thickness was increased, the overall optical stack was reduced by 15%. The paper showed an SNR10 improvement, and perhaps most importantly, enabled a performance gain without a materials change. Devices using this approach were said to be in production “soon.”
Nearly everyone who wants to has made it to the 1.1 µm generation. The challenge now is to optimize that generation and continue developing sub-micron pixels. Aptina presented its work on 1.1 µm and 0.9 µm devices, including an advanced color filter array to combat cross talk. The approach is somewhat reminiscent of Panasonic’s SmartFSI, in that a grid material is used to form sidewalls in between neighboring color filters, essentially serving as light pipes. Of course, Aptina has recently published its Clarity+ whitepaper, so I expect we might see this new CFA structure deployed with 50% panchromatic pixels. The paper also discussed a new multi-conversion gain pixel architecture, including distributed source follower, to extend dynamic range and enable pixel level binning.
The work presented in the Samsung and Aptina papers will likely be utilized in the near future for smartphones or tablets. As for what’s next, there are enough ideas on the table that one of them is bound to be the next big thing. Whether that is a true paradigm shift, or a variation on the same old theme remains to be seen. It’s not too risky to suggest there isn’t even an agreed upon path forward within the imaging community. This was my third workshop, and every time I attend, I’m reminded that if imaging scientists and technologists can build something, they probably will. The invited speakers reminded us that sometimes when technologists “can’t” do something, they find a way to do it anyway. After all, our smartphones all contain cameras that were supposed to have been impossible to make.
Innovative Technology Elements for Large and Small Pixel CIS Devices by Ray Fontaine