What do you make of this paper?
Non-American doesn't mean Chinese. It could also mean Dutch or Japanese and since litho (ASML) and the photoresist processing (Tokyo Electon) are not American but also not-Chinese, that 30% non-American is a much smaller Chinese proportion. Add two points to SME to 5% and that is global.
www.sinodefenceforum.com
It seems enormously significant to me as a layman, but I'd like to hear from someone with experience in the industry like yours
Those are the expected results from a size perspective.
LW is the line width, which is in the tens of nm (22-68 nm)
LER is line error roughness. that's the deviation from straightness of a feature, not the feature size itself, and note how they're limited at around 2-3 nm. So errors get bigger relative to feature size as you move down in process.
What they're doing in that paper is developing an EUV photoresist with high resolution and testing a measurement platform. many photoresists in use now for ArF processes are polymeric resists which are easy to coat (spin coating the polymer solution) but have a resolution problem.
Let's take a positive photoresist - where exposed, it becomes more soluble. These contain a photoacid generator (PAG) which is an organic molecule that turns into a water soluble acid when exposed to UV light. This acid either directly dissolves away as part of the photoresist, or reacts with the photoresist.
Here's the problem - many organic molecules are soluble in polymers due to their loose pack molecular structure and similar chemical properties. So they'll move over time. Even motion of a few nm means that the edge of the exposed feature is no longer sharp. So for EUV, they need to develop a photoresist specifically for high resolution EUV processes.
Indeed. This paper is on the development of EUV resist. From the images provided, I would say this resist is not very impressive.
It's easy to achieve good imaging performance in terms of CD uniformity and Line Edge roughness with very high dose. As you can see at the smallest line width, the dose is is 290mJ. EUV resist in use in production today is roughly between 20 to 50mJ. Scanner throughput is inversely proportional to dose. For exmaple, running at 40mJ means the throughput would be half of that if run with 20mJ.
Currently the industry is working toward higher contrast resist that could provide good imaging performance at low dose. This paper showing results with dose from 80mJ and up is hardly anything to gloat about.
Furthermore, the SEM images show linewidth down to 22nm...but the problem is this is not 1:1 pitch (e.g. 44nm pitch). The pitch size appears to be >100nm. All of these are worse than existing EUV resist in production.
Lastly, there's a big difference between demonstrating performance of imaging via SEM image than actual production capability. I would chalk this one up as a mere research paper on their progress...and nothing to be too excited about.
