Russia outlines EUV litho chipmaking tool roadmap through 2037 — country eyes replacing DUV with EUV

News
By last updated

Because why not?

Micron
(Image credit: Micron)

The Institute of Microstructure Physics of the Russian Academy of Sciences (via Dmitrii Kuznetsov) has laid out a long-term roadmap for domestic extreme ultraviolet (EUV) lithography tools operating at a wavelength of 11.2 nm, extending the information the organization shared last December. The new project spans from 2026, utilizing 40nm manufacturing technology, and extends to 2037, incorporating sub-10nm fabrication processes. The latest roadmap looks more realistic than some previous ones, yet it still has to prove itself to be executable. Furthermore, if executable, it may not be used for commercial purposes.

The very first thing that strikes the eye is that the proposed EUV* systems avoid replicating the architecture of ASML's tools. Instead, the plan is to use an entirely different set of technologies: hybrid solid-state lasers, xenon plasma-based light sources, and mirrors made from ruthenium and beryllium (Ru/Be) that reflect light at a 11.2nm wavelength. The choice of xenon instead of tin droplets in ASML's EUV tools eliminates debris that harms photomasks, which dramatically reduces maintenance. Meanwhile, when compared to ASML's DUV tools, the lower complexity is meant to avoid high-pressure immersion fluids and multi-patterning steps for advanced nodes.

  • The first system, scheduled for 2026–2028, is a 40nm-capable lithography machine with a two-mirror objective, 10nm overlay accuracy, an exposure field of up to 3 x 3 mm, and throughput exceeding five wafers per hour.
  • The second stage (2029–2032) introduces a 28 nm (with potential for 14 nm) scanner utilizing a four-mirror optical system. It offers 5 nm overlay precision, a 26 x 0.5 mm exposure field, and output exceeding 50 wafers per hour.
  • The final system (2033–2036) targets sub-10 nm production with a six-mirror configuration, 2 nm overlay alignment, and field sizes of up to 26 x 2 mm. It is designed to deliver throughput above 100 wafers per hour.

In terms of resolution, these tools are expected to support a range from 65 nm down to 9 nm, matching requirements for many of today's and tomorrow's critical layers in 2025 - 2027. Each generation advances optical precision and scan efficiency, while presumably maintaining a significantly lower per-unit cost structure compared to ASML's Twinscan NXE and EXE platforms.

Notably, the developers claim several unexpected benefits from using EUV for trailing nodes. However, they never mention complexities arising from using a laser with an 11.2nm length (different mirrors, different tools to polish the mirrors, different optics, different light source, different power supply units, resist, just to mention a few), which is a non-industry standard wavelength when it comes to EUV lithography. Click the tweet below to expand the roadmap.

In general, this roadmap may outline Russia's plan to achieve self-reliance in chip production by circumventing traditional EUV limitations. However, it is unclear how executable this plan is, as it would leapfrog the entire industry, something that has yet to be proven.

Rather than targeting maximum throughput for hyperscale fabs, the tools are aimed at cost-effective adoption by smaller foundries. By offering a clean, efficient, and scalable lithography system that doesn’t require immersion or tin-based plasma, the Russian platform may also appeal to international customers who are currently excluded from ASML’s ecosystem. If fully realized, the project could enable advanced chip manufacturing for both domestic and export use at significantly lower capital and operational costs.

*EUV systems were once called Soft X-Ray, hence the slide in Russian mentioning 'röntgen' lithography.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button!

Anton Shilov
Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

12 Comments Comment from the forums
  • bit_user
    I'll bet they just end up buying Chinese EUV equipment. The cost differential between that vs. developing it on their own isn't one I think they'll be able to justify, especially when you consider the limited market for the chips they'll be producing along the way and when they get there. Plus, buying it from China will surely accelerate their roadmap quite substantially, even if China doesn't sell them its latest tech.

    The article said:
    Instead, the plan is to use an entirely different set of technologies: hybrid solid-state lasers, xenon plasma-based light sources, and mirrors made from ruthenium and beryllium (Ru/Be) that reflect light at a 11.2nm wavelength.
    If these choices made sense, wouldn't ASML be going back and retrofitting its equipment to help customers further reduce costs of older nodes? Is revising older equipment something they ever do?
    Reply
  • Stomx
    Russians always plan to beat everyone in the world. Gigantism is a part of mentality. Like the wavelength of this low debris 11.2nm EUV source is clearly more advanced than ASML 13.5nm one, not much, but still. ASML design is kind of traditional one, Russians took more advanced laser system design irradiating nanoclusters.

    But the realization of schedules and the final cost for all that in reality will be very different subjects and the whole thing may never happen. There even exists the proverb: "The severity of Russian laws is countermanded by the lack of obligation to implement them"
    Reply
  • IBM296
    bit_user said:
    I'll bet they just end up buying Chinese EUV equipment. The cost differential between that vs. developing it on their own isn't one I think they'll be able to justify, especially when you consider the limited market for the chips they'll be producing along the way and when they get there. Plus, buying it from China will surely accelerate their roadmap quite substantially, even if China doesn't sell them its latest tech.


    If these choices made sense, wouldn't ASML be going back and retrofitting its equipment to help customers further reduce costs of older nodes? Is revising older equipment something they ever do?
    Yeah they'll probably buy Chinese equipment. The Russian plan (in theory) is to reach 9nm in 2037. China is already well on its way to reach 5nm by 2030.

    Seems a lot more practical (and cheaper) to just buy the equipment from China.
    Reply
  • Elusive Ruse
    Yesterday I outlined a roadmap to purchasing a video card that would enable me to be emancipated from upscaling technology by 2030.
    Reply
  • Electro-man
    Ok, you will have the printer(let say) but what will you print? :ROFLMAO: Having a technologically relevant IP is more important than lithography capacity.
    Reply
  • shady28
    Stomx said:
    Russians always plan to beat everyone in the world. Gigantism is a part of mentality. Like the wavelength of this low debris 11.2nm EUV source is clearly more advanced than ASML 13.5nm one, not much, but still. ASML design is kind of traditional one, Russians took more advanced laser system design irradiating nanoclusters.

    But the realization of schedules and the final cost for all that in reality will be very different subjects and the whole thing may never happen. There even exists the proverb: "The severity of Russian laws is countermanded by the lack of obligation to implement them"

    I think there is a big underestimation of Russia, largely linked to political ideological bias'.

    54% of Russians have a bachelors or higher degree, compared to just 18% of Chinese. They are #2 globally in the 25-64 age group, and 28% of Russians hold a masters or higher. Think about that - more than one in 4 Russians have a masters degree.

    Think about the internal competition those rates engender.

    For a relatively small population (~140M), they do pretty significant things. It wasn't until about 10 years ago for example, that China could build a jet engine for their SU-27 knockoffs that was comparable to the Russian made one.

    It was just a few days ago that Germany was complaining about Russian satellites 'shadowing' and even 'chasing' their satellites. Dare you to compare the space program of Russia to say, Japan, or the EU, or China.

    I mean, they have their own GPS system, completed in 2011. China's wasn't complete until 2020.

    Their technology strengths are more in physics and materials than digital though. Still I wouldn't underestimate their ability to make up ground in this space. They aren't shooting for 2nm next year, they're shooting for something more like 5nm in 10-12 years. I suspect that is very doable for them.
    Reply
  • bit_user
    shady28 said:
    I think there is a big underestimation of Russia,
    I don't see that. Nobody here said they weren't smart enough. I think the skepticism comes from a willingness to commit the necessary resources, especially when the potential for return on those investments seems quite dubious.

    On that latter point, keep in mind that they'll be trailing China for pretty much the entire time. So, anyone wanting chips fabbed outside the sphere of Western influence will just go to China. They won't need Russia's fabs, and there's no way Russia can compete with China on price.

    shady28 said:
    they have their own GPS system, completed in 2011. China's wasn't complete until 2020.
    In all of your comparisons vs. China, you should keep in mind that Russia had a gigantic head start. China is passing them in just about everything and only started working on this stuff a lot more recently.

    And, speaking of space programs, it's notable that Russian & China had some joint missions but the Russian part kept resulting in mission failure. So, now China just goes it alone. See also Luna 25.
    Reply
  • shady28
    bit_user said:

    On that latter point, keep in mind that they'll be trailing China for pretty much the entire time. So, anyone wanting chips fabbed outside the sphere of Western influence will just go to China. They won't need Russia's fabs, and there's no way Russia can compete with China on price.


    I don't think they are trying to compete with China. They are looking for 'good enough' tech in a model that is both more efficient, and distributable for small fabs (mentioned in the article). They are not seeking high production numbers or mega fabs like the US / China.

    Their approach, if it's successful, could be extremely disruptive - even if it is 10Y behind whatever exists in 2035 in terms of node size. Small production facilities that could be setup without pouring tens of billions into them...

    bit_user said:
    In all of your comparisons vs. China, you should keep in mind that Russia had a gigantic head start. China is passing them in just about everything and only started working on this stuff a lot more recently.

    And, speaking of space programs, it's notable that Russian & China had some joint missions but the Russian part kept resulting in mission failure. So, now China just goes it alone. See also Luna 25.

    I think Russia still has a big lead on China in terms of capability. I mean this in terms of aeronautical engineering and marine engineering. Chip technology isn't everything.

    China spends too much time copying, and not enough time innovating. Look at the list of combat aircraft that China has.

    All but two are either Russian designs, or in one case a low end export knock-off of the F-16. They actually bought 24 SU-35S, an aircraft that was introduced in 2014. They were delivered in 2019.
    In the second case, the J-35 stealth fighter, that one India said they could see on their radar from hundreds of miles away.

    So no, China is not yet in the mode of innovating.
    Reply
  • igorcd
    Electro-man said:
    Ok, you will have the printer(let say) but what will you print? :ROFLMAO: Having a technologically relevant IP is more important than lithography capacity.
    Russia doesn't need advanced nodes to fulfill its government and military needs.

    As for IP, I hope you know or at least have heard of RISC-V.

    You can laugh as much as you want, but Russia is one of the few countries developing its own independent projects.
    Reply
  • igorcd
    bit_user said:
    They won't need Russia's fabs, and there's no way Russia can compete with China on price.
    It's not about price to begin with.
    Reply