When you walk into conference room 220A, it will be under the watchful eyes of the “lithography luminaries”. These life-size posters of people who have made out-sized contributions to lithography are part of SPIE’s celebration of the International Year of Light. Some are a sampling of the great scientists who contributed to optical science and technology: Lord Rayleigh, Ernst Abbe, Frits Zernike. The rest are a sampling (of course, incomplete) of folks who have contributed directly to semiconductor lithography. I’m proud (and a tad embarrassed) to be among them. And then someone pointed out the hyphen. I guess one should always be mindful of one’s mortality (Chris Mack, 1960 – …).
I spent the morning of day three learning about progress on directed self-assembly (DSA). Defectivity levels are below 1/cm2, a major milestone but still too high. DSA for contact hole shrinking and uniformity enhancement seems very close to mature enough for manufacturing. Is it ready? I wish that more manufactures were here giving papers talking about their progress in this and other areas. Samsung is conspicuous in its relative absence. On the other hand, Intel, a company that perennially listens more than it talks, is much better represented this year than in the past.
DSA for lines and spaces still needs work. It is interesting to see that line-edge roughness is one of the big problems for DSA, since we all thought that low LER would be one of its major advantages. It doesn’t help that its competitor is SAQP (self-aligned quadruple patterning), a technique with about the lowest LER possible. Comparisons of DSA to SAQP will determine if DSA is ready for manufacturing, both in terms of performance and cost. It is getting close. Dan Millward of Micron reported on Monday that it is only 10% away on the LER metrics.
I also heard some competing exercises comparing the impact on design of choosing SAQP or EUV for the 7nm node. Lars Liebmann showed some standard cells designed making assumptions about the restrictions that SAQP and EUV would impose, then found that routing would be impacted the most, with EUV having an area advantage. Julien Ryckaert of Imec should that a clever introduction of an extra middle of the line (MOL) metal layer, as well some other optimizations, could eliminate the area penalty of using SAQP. The Liebmann and Ryckaert studies were complimentary, and more of these exercises are certainly necessary to understand the impact of lithography on 7nm-node design.
In the afternoon I heard about the progress of multi-electron beam tools. IMS Nanofabrication talked about the most promising application of multibeam writers: mask making. Their tool has made it through the alpha stage and they are hoping to have a production instrument in 2016 (despite having “Ready for Use” in the title of their talk). Mapper seems to continue its trend of pushing out their schedule by one year every year. Still, if they can demonstrate one wafer per hour in the next year or two, that will be a major milestone worth celebrating. The REBL program was dropped by KLA-Tencor in the past year, but it seems that TSMC has not given up on it. It will be interesting to see if TSMC can find another supplier to pick up that technology.
At the poster session I was very pleased to see the posters much more spread out than in years past – thank you SPIE! It was very pleasant to mingle and talk to authors with an adequate amount of room.
After a sampling of wonderful hospitality suites (thanks to the vendors who let me in despite that I will never buy their products), I ended the evening at the PROLITH party. It was 30 years ago that I gave my first talk at this conference (“PROLITH: a comprehensive optical lithography model”). Of course, I had no idea that this paper would have such a momentous impact on my life. It has been a great 30 years, and I am grateful for all the friends I have made in this community. It was also 30 years ago that I held the first bathtub party. (If you don’t know what that means, you can read some of my past blogs from this conference here). I was very glad to end the evening carrying my 30th anniversary PROLITH beer glass.