In recent times, there has been an increase in news related to lithography machines in China, with headlines such as “Significant Breakthrough in Domestic EUV Lithography Machine!” and “Substantial Breakthrough in EUV Lithography Machine!” grabbing attention. However, upon closer analysis, these news articles tend to exaggerate things that are still in the theoretical or laboratory stages.
Experts argue that the development of EUV lithography machines not only requires a stable institutional environment but also a craftsman spirit. Under the current Chinese communist regime, the focus is more on generating widespread corruption rather than fostering genuine security and retaining top-tier talents.
A lithography machine mainly consists of three major components: a light source, a lens, and a platform, with the light source being particularly crucial.
Currently, the mainstream laser light source in lithography machines is the LPP scheme adopted by the Dutch company ASML, which has been extensively verified, and all existing EUV light sources are based on LPP.
The mainland’s LPP scheme is primarily led by the Shanghai Institute of Optics and Fine Mechanics, where Lin Nan, the former head of the light source technology department at ASML, spearheads the project. Based on the current research progress, Lin Nan’s team has achieved a conversion efficiency of up to 3.42% from 1 μm solid-state laser to 13.5 nm EUV on their established laser-driven plasma extreme ultraviolet (LPP-EUV) light source experimental platform, with the main focus being single-pulse energy exceeding 20 mJ at 13.5 nm.
This result has been considered to be at the forefront internationally and leading domestically.
Apart from the LPP scheme, China’s research and development on EUV light source technology also explore other technical pathways, such as the discharge plasma extreme ultraviolet light source (DPP) from Harbin Institute of Technology. This approach is believed to have advantages in high energy conversion efficiency, lower cost, smaller size, and lower technical complexity, providing extreme ultraviolet light with a central wavelength of 13.5 nanometers.
Despite the simplicity of the DPP scheme compared to LPP, industry experts note that Cymer (later acquired by ASML) was unable to increase power and had to abandon DPP in favor of LPP.
In addition, Guangdong Intelligent Machine Research Institute and Huazhong University of Science and Technology are experimenting with using high-power fiber laser to shoot liquid tin targets, bypassing the route of ASML’s ultra-high power and high repetition frequency carbon dioxide laser.
The last innovative and disruptive approach is the SSMB-EUV accelerator light source, a steady-state microbeam extreme ultraviolet light source route led by Tsinghua University.
The various light source routes developed in China have been hailed by domestic media as “significant breakthroughs” and “substantial breakthroughs,” creating an impression that domestically produced EUV lithography machines are imminent.
However, according to ASML’s development experience, it took three years just to demonstrate the feasibility of EUV light source and another 15 years to achieve practical engineering of the 13.5nm EUV light source.
ASML’s CEO Christophe Fouquet, at a conference in April this year, stated, “Generating some EUV light is possible, but China still needs many, many years to manufacture EUV machines.”
Former purchasing manager of TSMC, Xue Zongzhi, remarked that whenever someone asks if mainland news is true or not, he always replies that if you see such news boasting about achievements, it’s likely false because China loves to exaggerate its capabilities. If they truly achieve something, they wouldn’t want the world to know because revealing it could lead to control from others, which is self-destructive.
Xue Zongzhi mentioned that China will need at least 20 years to reach the level of ASML. What they are producing now is controlled in laboratories and purely for verification purposes, using large equipment without cost constraints, aiming for one successful trial. Those working in engineering understand that transitioning from a laboratory setting to mass production can take up to 20 years and many professors may write impressive papers but fail to translate them into products for mass production. Mass production ensures consistent quality output every second.
Taiwan University professor Lin Zongnan pointed out that the tendency to exaggerate achievements is primarily due to the inability to deliver. If they could, the product would be directly launched.
In terms of power output, ASML’s top-of-the-line light source has 600 watts, while Harbin Institute of Technology’s power is approximately less than 100 watts, indicating a significant disparity and inability to be used in actual factory manufacturing.
Director of the Policy and Regional Research Division at Taiwan Industrial Technology Research Institute, Li Guanhua, stated that current news in Chinese media mainly revolves around laboratory inventions. Labs focus on publishing papers, but commercialization is a long way off.
He emphasized that EUV light sources are critical, but the enormous machine involves more than just the light source, including lenses, and harmonizing the entire system, especially for advanced processes like 3 nanometers and 2 nanometers. Even a tiny tolerance difference in a component or zero component would lead to very poor results.
“Advanced lithography machines are not just about research and development; they require cooperation with advanced process manufacturers. Everyone should collaborate for mutual development. China faces a disadvantage as it lacks manufacturers with advanced process capabilities, hindering the absorption of manufacturing experiences to optimize their lithography machines,” Li added.
Industry insiders believe that a lithography machine constitutes a complete nanoscale industrial set-up, where achieving the nanoscale precision in every component and process of the entire machine is pivotal.
Some domestic Chinese experts have admitted that throwing money at lithography machines won’t bridge the technological gap. Science requires creativity, and creativity thrives in an environment that fosters interest and allows trial and error. Treating science as a mission or even a command will likely yield counterproductive outcomes. Not everything can have a knock-off version; products like aircraft engines, chips, and lithography machines are challenging to replicate just by observing samples.
Director of Strategic and Resource Department at Taiwan Institute for National Defense and Security Studies, Su Ziyun, mentioned that the micro-physical world involves proprietary knowledge and technical thresholds, coupled with machine manufacturing integration. Unlike massive weapons systems like aircraft carriers that can be reversely engineered, micro-physical equipment cannot be reverse-engineered.
Experts underscore that lithography machines, as intricate and massive engineering endeavors, demand high standards from individuals and the societal system.
Lin Zongnan explained that the semiconductor industry’s value chain is extensive, spanning various fields that require cumulative experiences. Without a stable, free, and conducive social environment, it becomes challenging to accumulate technological expertise, and regimes like the CCP may not generate high-quality talents or accumulate such talents.
Li Guanhua highlighted that a lithography machine is a precise system requiring a traditional craftsmanship approach akin to that in Japan. It demands extreme focus, precision, and high-quality standards in every step, necessitating repeated simulations to form the final product meticulously.
He noted that although China has such talents, they struggle to thrive in China due to the rapid market entry philosophy prevalent. Many actions that should be taken meticulously and step by step are often skipped in the rush to achieve results. By neglecting the process and solely focusing on the end product, the outcome may not meet expectations.
Li Guanhua pointed out that individuals with experience in this industry know that achieving foreign achievements within a short span of two to three years is impossible. However, taking longer to succeed may result in insufficient funding, leading many to deceive and quickly achieve results in the market, as evident in the chip corruption scandal.
Su Ziyun suggested that the calculated corruption within the CCP, under the guise of the semiconductor industry, leverages government funds. Various local governments collude with unscrupulous individuals to establish multiple fake semiconductor companies that collapse within two years.
He added that under the CCP’s one-party rule, no individual truly feels secure because the party could label anyone for reasons like leaking secrets or other security concerns, preventing the retention of top-tier talents.
To date, ASML’s EUV lithography machine stands as a unique marvel in global manufacturing. Even if ASML has some publicly available data, no other country has managed to replicate its success.
Xue Zongzhi pointed out that while ASML’s ideas are widespread, the reason China fails to produce similar results is simple. Just as China purchasing Mercedes-Benz does not enable them to immediately match the quality of Mercedes-Benz vehicles, the experience of Mercedes-Benz, BMW, and similar firms accumulated over centuries provides a significant advantage.
He reflected, “I was previously responsible for that technology, and it’s highly challenging. I must respect technology and talent.” He affirmed that the United States, as well as Taiwan and Japan, have struggled to reproduce ASML’s success, making ASML the unparalleled leader worldwide.
Xue Zongzhi emphasized the significance of long-term data accumulation within enterprises and the crucial role of personal experiences and legacies.
“I always say, humans are the most potent AI, especially some skilled craftspeople. In the past, when we went to repair vehicles, a skilled craftsman could discern the issue with just a listen. However, AI lacks such capabilities since it requires massive data. Even with AI, without data, it’s ineffective. Thus, an understanding of optical technology from 4 inches to 12 inches, from micrometers to nanometers, with the necessary data, is solely possessed by ASML worldwide,” he emphasized.
He highlighted how the AI capabilities are reliant on the data and could be executed efficiently using AI’s support to process and integrate the vast amount of data but without proper data accesse, technology cannot be perfected.
He elaborated that even in the United States, to establish TSMC’s Arizona factory, the entire TSMC team and big data had to be relocated to ensure the successful operation of the factory.
“This is termed soft power, China has excelled in areas like solar panels and LED lighting due to their lower thresholds, but they’ve hit a roadblock in semiconductors,” Xue added.
He also noted that another essential factor is ASML’s EUV’s collaborative development with customers, emphasizing the importance of customer feedback in product development. Just like designing a toaster-oven, relying solely on personal ideas often leads to subpar outcomes compared to collaborating with experienced bakers who understand how it should be used, resulting in significant improvement.
Xue Zongzhi pointed out that even if China could produce an EUV machine today, whether developed by Tsinghua University or Peking University, it would struggle to find willing customers to test it. Currently, China lacks any fabricators that can produce 3-nanometer and 5-nanometer wafers. SMIC only operates at 10 nanometers, and claims of 7 nanometers are erroneous, with low yields and short lifespans.
According to the latest data from the Center for Security and Emerging Technology (CSET), an independent think tank in Washington, D.C., China gained a negligible market share in the lithography tools sector from 2019 to 2024. The Netherlands continues to be the major supplier of lithography tools (79%), followed by Japan (17%), underscoring the crucial bottleneck lithography technology poses to China’s semiconductor technological ambitions, with foreign established companies maintaining a robust leading position.
Shanghai Microelectronics Equipment Co., Ltd. (SMEE) from China holds only a 4% market share in conventional chip manufacturing I-line (365 nm) lithography machines.
Guanhua Li mentioned that while China’s progress in advanced processes is slow under the tight restrictions of the United States, and significant advancement is challenging. If countries with advanced semiconductor capabilities continue progressing while China remains constrained, the widening gap between them is inevitable.
Su Ziyun estimated the gap with ASML to be around 30 years currently. While China may have some breakthroughs narrowing the gap to 20 years, democratic nations are also advancing technologically. Therefore, the gap remains manageable.