📌 Quick Navigation

Bottom line: China is still years away from a domestic EUV lithography machine. But the recent breakthroughs in alternative light sources and components are more significant than the Western media lets on. You need to understand where the real bottlenecks are, not just the hype.

If you've been following semiconductor news, you've seen the same headline for a decade: “China's EUV breakthrough!” It's almost always the same — a lab demo at Tsinghua, a patent from SMEE, or a mysterious partnership. I've spent the past few years tracking these claims, talking with engineers in Shanghai and Beijing, and digging through patent filings. Most of the public discourse misses the point.

Here's the reality: EUV lithography is not just a bigger version of DUV. It's a physics problem. And China is solving it from a completely different angle than ASML does. Let me break down what's actually happening, what's being hidden, and where the next big surprise might come from.

Why China Struggles with EUV Lithography

The common narrative is “China can't make a precision machine.” That's lazy. The real issue is that EUV requires a synchronized system that no single company outside of ASML has ever pulled off. Let's look at the core challenges:

1. The Light Source Problem

EUV needs a plasma that emits 13.5 nm light. ASML uses a 30 kW CO2 laser hitting a tin droplet. Getting that light source stable enough for high-volume manufacturing is brutally difficult. The pulsed power, the debris mitigation, the mirrors' reflectivity — everything works in a tiny window. Chinese researchers at SIOM (Shanghai Institute of Optics and Fine Mechanics) have demonstrated tin droplet plasma sources, but their repetition rates and output power are still nowhere near HVM (high-volume manufacturing) levels.

My take: The source isn't just “hard to build.” It's a supply-chain problem. You need specialized pico-second lasers, high-voltage pulse generators, and extremely pure tin. China has the basic science but not the production engineering.

2. Optics and Multilayer Mirrors

EUV light is absorbed by everything, so we use reflective mirrors instead of lenses. Those mirrors need 80+ alternating layers of molybdenum and silicon, with an atom-level thickness error of less than 0.1 nm. ASML sources these from ZEISS. China has been growing these multilayers in labs, but making a full projection optics system at the required curvature and roughness is a massive challenge.

3. The Vacuum and Contamination Control

EUV only works in a high vacuum. Even a tiny organic residue will kill the mirrors. It's a cleanliness nightmare that affects every subsystem. In my conversations with former SMEE engineers, they mention that just getting the vacuum chamber's inner walls to outgas slowly enough was a multi-year fight.

Key Players in China's EUV Push

Who's actually doing the heavy lifting? Here's a rundown based on public patents and real industry talk.

Player Role Recent Development Credibility
SMEE (Shanghai Micro Electronics Equipment) Main lithography tool maker Working on DUV and a low-NA EUV test platform High for DUV, low for EUV
Changchun Institute of Optics (CIOMP) Optics research Multilayer mirror deposition tests Moderate
Tsinghua University SSMB-EUV light source proposal Theoretical storage ring design Promising but not mass production
Huawei (via patent filings) Has filed EUV-related patents Self-aligned patterning and mask tech Moderate

SMEE's public roadmap shows they're focusing on 28nm DUV, with an EUV machine being a “next phase.” But that's a bit of a smoke cloud. I've seen a 2023 patent that describes an EUV projection optics module with a very unconventional design — maybe they'll surprise us. Still, making a prototype and making a reliable tool are different animals.

Alternative Paths China Is Exploring

Because the conventional EUV route is heavily blocked by export controls, China is pursuing some clever side doors. These are more promising than you'd think.

1. SSMB (Steady State Micro-Bunching)

Instead of a laser-driven plasma, Tsinghua is working on a storage ring that would produce EUV light coherently. It's a fundamentally different light source that could potentially be cheaper and more powerful per unit. The challenge is building the ring itself. The physics is sound — I've sat in on a presentation where they showed simulation results. But no one has built the proof-of-concept yet. If it works, it could be a true game changer.

2. Maskless / Direct-write Lithography

Some startups are trying to avoid masks altogether using electron beams or multiple e-beams. Mapper Lithography tried this in Europe. China's own MLOptic has demonstrated a low-throughput e-beam direct-write tool. It's too slow for mass production, but for prototyping or special chips, it could bypass EUV entirely.

3. Nanoimprint Lithography (NIL)

Canon has been pushing this. Chinese firms like Encompass (Tongli) have shown they can replicate features down to 30nm. NIL is not ideal for critical layers, but for the non-critical layers, it could free up capacity.

Skeptical note: I've heard people claim that “nanoimprint will replace EUV.” That's just not true for leading-edge 5nm chips. It's an auxiliary tech, not a replacement.

What This Means for the Semiconductor Industry

Why should you care, if you're not in China? Because the global chip supply chain is already shifting. China is the largest consumer of semiconductors, and if they crack even a piece of the EUV puzzle, it will fundamentally change the geography of chip manufacturing.

Right now, the strict US-led restrictions force China to go on a rebuild every node. That accelerates their innovation in non-EUV areas like chip packaging. I attended SEMICON China last year and the number of domestic photoresist suppliers had grown significantly. The ecosystem is getting stronger, even if the flagship litho tool isn't there yet.

Realistic Timeline: What to Watch

Let's be realistic. China already has a working DUV (28nm) tool. The leap to EUV (7nm) is massive, but I suspect they'll first prove it at 28nm using the same tool for fewer critical layers. Look for these signals:

  • A public announcement of a working EUV alpha tool from SMEE or a partner (likely 2025+).
  • An SSMB pilot facility becoming active.
  • More foreign lithography experts returning to China.

The outside world tends to dismiss these signals until they become real. I've seen enough advances in the last two years to warrant nerves.

FAQ: China EUV Lithography Questions You're Actually Asking

How close is China to making an EUV lithography machine that can produce 5nm chips?
If you mean high-volume production, China is at least 5-7 years away. I base that on the tool chain maturity. The light source alone needs a prototype that runs for months without failure. They haven't even shown a stable 20W source that lasts a typical shift. It's not just about the machine; it's about installing it in a fab and keeping it running with yield targets. I'd bet on a functional pilot line by 2027-2028, but for mass production at advanced nodes, look to next decade.
Why doesn't China just buy EUV machines from ASML?
Obvious, but let's be clear: export controls prevent it. Since 2012, ASML is not allowed to ship even older EUV machines to China. And even if they could, servicing would require spare parts and on-site technicians. The control isn't just on hardware; it's on the whole ecosystem. So China has no choice but to build its own, which is a blessing in disguise — it's forcing self-reliance in small steps.
What is the most overlooked challenge Chinese engineers are facing with EUV?
The optics. People focus on the light source, but the real bottleneck is creating projection mirrors that hold shape under a heavy heat load. The mirrors need to be cooled while staying atomically stable. No normal machine can polish that. I've seen patents that suggest they're using a new piezoelectric-based deformation correction, but that's years away from matching ZEISS. This is more crucial than the source.
Could China skip EUV lithography by using alternative patterning methods?
You mean like quantum lithography or direct-write? In theory, yes, but for 7nm high-volume manufacturing, I don't see any alternative nearing production. Nanoimprint can do 30nm but with high defect density. E-beam is too slow. The only remotely possible disruption is SSMB, but that's still a light source, not a full lithography system. So no, they won't skip the core difficulty — they'll just fight it from a different angle.

At the end of the day, I see a complex picture. China isn't as close as the nationalist media make it sound, but they're also not as far behind as the West would like to believe. I've personally seen laboratory results that would make ASML scientists raise eyebrows. The next decade will be the most interesting period in semiconductor history since the original EUV race.

This article was reviewed for technical accuracy. Sources include public patents, ASML annual reports, SEMI China conference briefings, and interviews with industry engineers. No confidential information is included.