Forum: BrightSpots Lithography F... |
We talked a fair bit during the live panel about whether there really needs to be a “battle” between EUV and direct-write e-beam, for example, or whether they are two complementary technologies that will together to help the industry progress in their own ways. We heard great responses from Franklin Kalk and Aki Fujimura in support of peaceful co-existence.
But while EUV for much of the industry seems increasingly a foregone conclusion, there still seems to be a fair bit of concern from some camps about whether EUV will ever really be a viable solution — in the time it’s needed.
Does the industry need to shift more of its attention to alternatives — whether that’s e-beam, nanoimprint or continued 193 nm immersion (triple, quadruple patterning)? Or would that be diluting the efforts that still need to go into EUV?
If I can put my two cents in here, I think it makes sense to increase the amount of investment in alternatives to EUV. Both nanoimprint and e-beam lithography have received a fraction of the funding of EUV, and yet both have made major strides along the development curve for semiconductor manufacturing. It’s true that there’s still a ways for both to go, but several applications are already benefiting from the work done to date with both technologies. Nanoimprint is being utilized for bit patterned media development by several of the top manufacturers in the data storage industry including Toshiba (http://www.theregister.co.uk/2010/09/03/seagate_hamr_blow/) and is also being investigated for use in photonic crystal production in LEDs. Developments in e-beam technology are enabling additional apps beyond direct-write e-beam, such as design for e-beam mask technology to extend optical lithography, which has the support of dozens of companies across the semiconductor food chain. So, like with many other new technologies, development efforts are yielding benefits in multiple applications—not just one.
The idea of putting all eggs into one basket to me seems a bit too risky also. EUV will happen eventually, but when and at what cost? Will it be economically viable for all leading-edge applications, or just very high-volume ones to justify the CoO?
Thanks for your input, David. It seems that nanoimprint lithography is making good strides in HDD, LED and photonic applications. But when I ask mainstream chip manufacturing folks about nanoimprint's chances for the future, most people don't seem to have much faith.
What do you think is holding it back? Is it simply funding, or would it ever really have the support of the industry?
I think funding is certainly an issue. From what I understand, the biggest barriers to nanoimprint for mainstream semiconductor use is overlay, defectivity and throughput. In each of these areas, nanoimprint technology has made a lot of progress over the past couple of years. So I don't see any permanent technical barriers to the technology that "no amount of money" would be able to solve. I do believe that additional funding will only accelerate the progress that nanoimprint has made. Unfortunately, I think the semiconductor industry has a tendency at times to be extremely slow to change course (take 157nm litho for example), which is ironic given the incredible amount of innovation that goes on here.
The items that inhibit nanoimprint's introduction into semiconductor manufacturing (e.g., template defect management and manufacturing turn time) are not big issues for bit-patterned media. NIL seems ideally suited for BPM. The integration issues with semiconductor manufacturing are considerable, however. The infrastructure to make, say, 15nm defect-free templates in short turn times (the current IC mask mfg strategy) would require new writer technology, new inspection technology, and possibly new repair technology.