ASML은 일론 머스크가 지원한 입자 가속기 기반 칩 제조 기술을 거부하고 EUV 스캐너용 1,000W 레이저 생성 플라즈마 시스템에 대한 투자를 두 배로 늘렸다.
One of the most significant challenges in developing extreme ultraviolet (EUV) lithography scanners is creating a powerful and reliable light source. ASML, the sole manufacturer of EUV lithography tools, relies on complex laser-produced plasma (LPP) technology to generate EUV radiation. In contrast, numerous companies are proposing free-electron lasers (FELs), which utilize particle accelerators, for EUV generation. While FEL technology offers distinct advantages and has even garnered endorsement from Elon Musk, ASML is unlikely to adopt it, according to JPMorgan. “Given laser advances, ASML sees no reason to try new 'FEL' light source favored by Musk,” reports Semi Doped, citing a JPMorgan note for clients.
Modern EUV lithography systems employ LPP light sources that fire high-power CO₂ laser pulses at microscopic droplets of molten tin, approximately 30 microns in diameter. This vaporizes the droplets into ionized plasma with electron temperatures reaching several tens of electron volts, emitting 13.5-nm EUV radiation. The emitted light is captured by a roughly 0.5-meter elliptical collector mirror coated with alternating layers of molybdenum and silicon, which selectively reflects maximum 13.5-nm radiation toward the scanner’s intermediate focus. Because virtually all materials absorb EUV radiation—even specialized multilayer mirrors lose a substantial portion—the entire optical path must operate in a vacuum using reflective rather than conventional refractive optics. This fundamental constraint makes generating sufficient EUV source power exceptionally difficult.
(Image credit: ASML)
Despite these engineering hurdles, ASML has steadily increased its LPP light source output, scaling from approximately 250W to around 500W. The company plans to double this capacity to 1,000W in the coming years while nearly doubling its droplet generation rate to 100,000 tin droplets per second.
(Image credit: ASML)
A free-electron laser generates EUV light by accelerating electrons to nearly the speed of light and directing the beam through an undulator, a series of alternating magnets that forces electrons to oscillate and emit radiation. The interaction between the electrons and their own radiation causes them to form microscopic bunches, producing coherent light at the target 13.5-nm wavelength. This method eliminates tin droplets and the associated debris that currently necessitate protective pellicles on photomasks. It also promises substantially higher EUV power than LPP systems. Furthermore, a single FEL paired with a large-scale EUV beam-distribution network could potentially replace multiple LPP sources.
However, the approach carries major trade-offs. Unlike the relatively compact LPP architecture, an FEL requires a highly complex particle accelerator, a dedicated electron source, a long undulator, precise electron-beam control, extensive radiation shielding, and an extremely intricate distribution system featuring mirrors capable of handling and routing very high EUV power without significant loss along the way. The entire apparatus must meet semiconductor fabrication-level standards for availability, efficiency, and cost—benchmarks that took ASML and the broader industry years to achieve.
(Image credit: xLight)
Consequently, despite strong enthusiasm for FEL technology across China, the U.S., and Japan, it will likely take a decade, if not more, before FEL can realistically compete with LPP in commercial semiconductor production facilities. Developing the technology will consume billions of dollars in the interim, meaning not every entity currently pursuing FEL will survive the timeline.