In June this year, the US Department of Commerce notified the Dutch semiconductor equipment manufacturer ASML of the suspicion that one of its Extreme Ultraviolet (EUV) lithography machines had flowed into China. ASML denied this claim, but has not yet provided any evidence to support their denial.
ASML is the only company in the world capable of manufacturing EUV lithography machines. Whether China has acquired an entire EUV machine, only key subsystems, or components, the question arises: what benefits can China gain from obtaining these technological parts?
According to the latest analysis from the reputable website specializing in US national security, defense policies, and foreign affairs, War on the Rocks, the biggest concern for the US is that once China obtains these machines, they are likely to engage in reverse engineering to steal ASML’s design and production technology.
It was revealed by Western media as early as last December that some engineers who had previously worked at ASML were hired by a Chinese laboratory. Through reverse engineering, they managed to help create a prototype of an EUV lithography machine.
While the US concerns are valid, the technology behind EUV lithography machines is extremely advanced and complex. It may not be easy for China to replicate the entire system through physical reverse engineering. Even if China can quickly grasp the architecture of components and sub-systems by obtaining a machine, it may not be able to transfer all the system integration knowledge, testing, process accumulation experience, and relationships with suppliers, essential for replicating the technology and maintaining operational efficiency.
ASML, founded in 1984 as a spinoff of the Dutch electronics company Philips, manufactures EUV lithography machines capable of producing advanced semiconductor chips at 7 nanometers and below, essential for smartphones, data centers, and AI accelerators. It plays a pivotal role in the development of artificial intelligence and is at the core of the US-China technological competition and related export control regimes.
The advantage of EUV lithography machines lies in their outstanding performance. Printing highly intricate patterns on semiconductor wafers (at the nanometer scale) with tolerances smaller than atoms involves thousands of interconnected processes. Even minor deviations in alignment, contamination, positioning, synchronization, speed, temperature, vibration, or other parameters can reduce yield and affect the entire process, making EUV lithography machines one of the most complex products in the world.
The exceptional performance of ASML’s EUV lithography machine components and subsystems includes precise features such as the laser hitting tin droplets at about 50,000 times per second to create plasma 40 times hotter than the sun’s surface; mirror accuracy at the picometer-level, allowing deviations to be only one-tenth of a millimeter when scaled to the size of Germany; and the mask (photomask) that carries patterns with acceleration equivalent to a car reaching 100 kilometers per hour in an instant.
The manufacturing of machine components and subsystems themselves poses a significant technical challenge and industrial puzzle, but the core challenge is to ensure that each component and subsystem operates within the tightly coupled system, while managing the heat, vibration, contamination, electromagnetic interference, and other disturbances from other components to achieve the desired performance. Developing components that can operate under stringent environmental conditions without affecting the performance of surrounding components is crucial.
The complexity of producing a single EUV system is staggering, with over 100,000 components and over 700,000 components including subsystem internal parts, sourced from 5000 suppliers. ASML collaborates closely with these specialized suppliers in the design and development processes, providing technical specifications that each component must meet and continuously refining them until reaching an optimal balance.
It is this immense complexity that makes the machine’s cost exorbitant, with prices of an EUV lithography machine reaching nearly $180 million, and the latest high numerical aperture EUV lithography machine priced at nearly $380 million.
The article emphasizes that China’s ambitious goal of successfully reverse engineering such a high-precision machine by acquiring a prototype and some blueprints, and then recruiting former engineers with high salaries, is unrealistic. While China may engage in industrial espionage, reverse engineering, and talent poaching, it may achieve limited success in architecture selection or accelerated development of specific subsystems, but cannot guarantee the commercial success regarding the required process knowledge, yield, reliability, and capacity.
To date, China is still struggling even in replicating the previous generation of Deep Ultraviolet (DUV) lithography machines from the late 2000s, let alone copying EUV lithography machines. ASML has already introduced a new generation of high numerical aperture equipment, placing China on an outdated target while striving to catch up with the constantly evolving goals.
Even if China manages to produce fully functional machines, the problem of the supply chain remains unresolved. An EUV lithography machine cannot independently produce cutting-edge chips but rather relies on a highly concentrated material and equipment ecosystem. For instance, chemicals responsible for forming circuit patterns – photoresists – are mainly supplied by a few Japanese companies, such as JSR, Tokyo Ohka Kogyo (TOK), and Shin-Etsu Chemical. As the photoresists must satisfy strict and sometimes conflicting requirements for resolution, sensitivity, and line edge roughness, changing suppliers is challenging.
This situation extends throughout the entire supply chain, whether it’s photoresist coating and developing machines, defect-free reticle substrates used for writing circuit patterns, or EUV-grade metrology and inspection equipment – all are either monopolized by one company or highly concentrated among a few companies.
China has been striving to establish its own supplier ecosystem; however, these input materials are developed jointly with ASML specifically for the machine. For example, the formulation of photoresists tailored to specific exposure doses, wavelengths, and outgassing properties; reticle substrates designed for specific flatness and defect densities; and specific purity requirements for tin. The machines and consumables have undergone meticulous adjustments for years, honed to atomic-level precision. Even if China can reproduce the scanner and supply photoresists, reticles, and chemicals, they will face the second integration challenge involving the entire industry chain.
Even after solving the production issues, maintaining normal machine operations remains a significant challenge. Continuous calibration and adjustments are required for EUV systems, relying on uninterrupted spare parts supply, software updates, and on-site maintenance from ASML’s global customer support center consisting of 10,000 customer support engineers working 12-hour shifts day and night. This specialized knowledge cannot be entirely transferred through hardware alone.
ASML’s customer support engineers undergo nearly a year of training, rotating through the company’s centers in Taiwan, the US, and Germany to obtain qualifications for independent machine operations. Fault diagnosis often depends on the engineer’s personal experience and intuition rather than standardized processes. For instance, an engineer may detect potential faults by detecting almost imperceptible deformations on cooling hoses, preventing hose ruptures that could lead to machine shutdown.
The article emphasizes that ASML’s machine, machine components, and the supplier ecosystem are highly complex, significantly limiting China’s ability to take shortcuts. The US and its allies can leverage this to formulate targeted policies to prevent China from copying and stealing.
Firstly, talent is the key bottleneck, and the US and its allies should establish talent retention and counter-recruitment measures to prevent China from poaching at the component and subsystem level.
Secondly, the US and its allies should regulate and protect critical nodes in the system, such as optical mask inspection tools and photoresists, to counter Chinese espionage activities, research, and attempts to exploit vulnerabilities.
Thirdly, China’s strategies of industrial espionage, talent poaching, and government acquisitions to acquire foreign technology can be countered by encouraging technologists to develop more technology that is beyond replication, making it difficult for China to accumulate the necessary experience to surpass opponents.
Lastly, although China continuously explores alternative solutions, its efforts falter due to internal integration and ecosystem barriers, undermining its ability to replace ASML and its subsidiary Cymer’s meticulously refined tin droplet plasma sources developed over several years. Therefore, Western countries should consider cutting-edge alternative solutions (not just existing bottlenecks) as a competitive area and exclude China from it as soon as possible.