When the battle for EUV vacuum environment had just achieved a phased victory—stabilizing chamber pressure at the ultra‑high‑vacuum level of 10^‑7 Pa—Xiuxiu's team immediately turned its gaze to the next, and the most core, most formidable fortress of EUV lithography: **the extreme‑ultraviolet light source itself**. Without light, the most perfect vacuum and optical systems were merely useless shells. The technology path they chose to conquer aligned with that of the world's leading ASML company: **laser‑produced plasma** (**LPP**) technology.
Deep inside the laboratory, a new, more complex experimental area was isolated. This was no longer merely a silent world pursuing ultimate vacuum, but a physical site soon to be filled with violent energy conversion. The central equipment was a massive device termed "light‑source prototype," integrating a high‑power pulsed laser, a tin‑droplet generator, and a collector‑mirror system for gathering extreme‑ultraviolet light.
Xiuxiu stood before the observation window, gazing through thick lead‑glass at the device interior. Her mood mingled excitement of venturing into the unknown with sobriety facing extreme difficulty. She knew they were about to attempt taming an immensely violent, hard‑to‑control physical process.
During a technical discussion session, she clearly explained the basic principle of **LPP** to the team.
"Comrades, the 13.5 nm extreme‑ultraviolet light we need cannot be generated directly by exciting gas like traditional lasers, because almost all materials strongly absorb light at this wavelength." Xiuxiu sketched a schematic on the whiteboard. "**The core idea of LPP technology is to 'tear apart' a specific target material instantaneously with 'brute force,' turning it into a high‑temperature, high‑density plasma; during this plasma's cooling process, it radiates the EUV photons we need.**"
She pointed to a key part of the schematic: "Our chosen target material is **liquid tin**. Why tin? Because when tin atoms transform into plasma, internal‑electron energy‑level transitions happen to very efficiently radiate photons whose wavelengths cluster near 13.5 nm—this is called 'window radiation.'"
"Then, how do we turn tin into plasma?" Xiuxiu's gaze swept the team members. "We use a **high‑power carbon‑dioxide pulsed laser** to bombard **micron‑sized liquid‑tin droplets**."
She detailed the process:
**Tin‑droplet generation**: An extremely precise tin‑droplet generator, inside the vacuum chamber, would produce perfectly spherical liquid‑tin droplets roughly 20–30 µm in diameter, at high frequency (tens of thousands per second) with high stability. These droplets needed uniform size, stable flight trajectory—like an invisible, ultra‑fine necklace of tin beads. **Laser bombardment**: When a droplet reached the pre‑determined focal spot, a precisely tuned, extremely high‑power‑density CO₂ pulsed laser (power typically reaching tens of kilowatts) would strike the tiny droplet with nanosecond‑level timing accuracy. **Plasma formation and radiation**: The huge laser energy would be absorbed instantaneously by the tin droplet, raising its temperature dramatically to hundreds of thousands of degrees Celsius. The droplet would instantly vaporize and ionize, forming a very‑high‑temperature **tin plasma** emitting a blinding white glow. As this plasma expanded and cooled, tin ions inside would undergo energy‑level transitions, radiating extreme‑ultraviolet light including 13.5 nm. **Light collection**: Because EUV light is absorbed by almost any material, transmissive lenses cannot be used; instead, special **multilayer‑coated mirrors** based on Bragg‑reflection principles are needed to collect EUV light radiated in specific directions from the plasma and guide it into the lithography illumination system.
The principle sounded clear and direct, but engineering‑implementation difficulty resembled a bottomless chasm.
The first integrated test began in an atmosphere close to tragic solemnity. All systems readied, vacuum met specs, laser stood by, droplet generator started.
"Tin‑droplet stream stable."
"Laser‑sync signal normal."
"Three, two, one… trigger!"
The instant the high‑power laser pulse fired, a brief, piercing flash erupted inside the observation window, accompanied by a dull thud. But immediately after, the detector screen monitoring EUV spectrum displayed only a faint, almost negligible 13.5 nm signal peak, which swiftly vanished.
"EUV‑pulse energy… below detection threshold." The engineer responsible for data monitoring reported softly, voice filled with disappointment.
Failure. Utter failure.
They began dissecting the failure like anatomizing a giant beast. High‑speed‑camera footage revealed the laser pulse hadn't acted completely, uniformly on the tin droplet; the droplet appeared to have undergone uncontrollable splashing and deformation at the instant of impact, rather than forming a neat, efficiently radiating plasma sphere.
The problem lay in **tin‑droplet control** and **laser‑droplet interaction**.
**Tin‑droplet control technology** itself was a major challenge. Generating micron‑sized droplets with uniform size, extremely stable position and speed required precise nozzles, stable pressure‑control systems, and overcoming various fluid‑dynamic instabilities. Their prototype nozzle produced droplets with percent‑level size fluctuations and micrometer‑level trajectory jitter—fatal for laser targeting demanding nanometer‑level accuracy.
Moreover, even if the laser struck accurately, optimizing laser‑droplet interaction to maximize EUV conversion efficiency was another bottomless subject. Direct bombardment often caused massive tin‑liquid splashing; these splashing droplets would contaminate expensive collector mirrors, greatly lowering their reflectivity, while this "explosive"‑type interaction yielded poor‑shaped, low‑efficiency‑radiating plasma.
Test after test, failure after failure. A flavor of anxiety and frustration began permeating the lab. Piercing flashes lit up again and again, but the hoped‑for stable, powerful EUV light remained absent. Collector mirrors already showed signs of tin‑residual contamination—meaning they not only failed to produce sufficient light, but were damaging extremely precious, hard‑to‑repair optical components.
Team members grew exhausted; some began doubting this technology path's feasibility, privately discussing whether to explore other more "obscure" but perhaps easier EUV‑generation approaches.
Xiuxiu bore enormous pressure. She almost lived in the lab; eyes bloodshot from prolonged high‑speed‑video review and spectral‑data analysis. She personally adjusted laser parameters, disassembled and cleaned contaminated collector mirrors with team members, analyzed tin‑debris morphology after each failure.
During a crucial failure‑analysis meeting, facing members' doubts and low morale, Xiuxiu stood up. Her voice was hoarse with fatigue, yet carried undeniable firmness:
"I know everyone's tired, frustrated. The LPP path is acknowledged as the hardest road, but currently the only one proven capable of meeting lithography power and reliability demands. ASML succeeded—why can't we?"
She walked to the whiteboard, drew the tin‑droplet‑deformation image recorded by high‑speed camera during that failed experiment.
"The problem isn't the principle; it's in the details! Our control isn't precise enough, our understanding not deep enough! Droplet uniformity, laser waveform, interaction timing… each parameter needs us to refine it like embroidery!"
"Every failure, the high‑speed camera tells us how the droplet responded; the spectrometer tells us how energy converted; the residues tell us how the process ran out of control. These aren't failures; they're data! Signposts toward success!"
Her eyes burned, sweeping each weary face: "What we're touching now is the hardest bone of EUV light‑source. Gnaw through it, and open terrain lies ahead. Fail to gnaw it, and all our previous efforts—vacuum, optics… become meaningless! We cannot retreat, have no retreat path!"
Her resilience and clear thinking acted like a strong stimulant, stabilizing the teetering morale. The team rallied, plunging into more meticulous, more tedious parameter optimization and experiments.
Yet one key bottleneck remained unbroken: domestic tin‑droplet nozzles, under prolonged high‑frequency operation, saw their micro‑holes deform easily due to tin‑liquid erosion and thermal load, causing droplet uniformity to degrade. They urgently needed a nozzle made of special material that was more heat‑resistant, more erosion‑resistant, with higher machining precision. Such cutting‑edge components were also on embargo lists.
Late one night, in the lab's temporary rest‑room, Xiuxiu stared at microscopic‑structure photos of nozzle wear on her laptop, feeling near despair. Unconsciously, she sent a short message in a small group chat that had only three people—herself, Yue'er, and Mozi. No specifics, just a deeply weary sigh: "Another bottleneck jammed; feels like bashing my head against steel plate."
Mozi replied swiftly, no superfluous comfort, just one sentence: "Send me the bottleneck's concrete technical parameters and requirements; maybe there's a way."
Xiuxiu hesitated a moment, but trusting Mozi's mysterious capability and past help, she compiled nozzle material‑performance needs, working environment, and facing predicament into a concise technical‑demand document and sent it over.
She held little hope—this was highly specialized, precision componentry.
Yet, only three days later, Mozi contacted her again, giving her contact information and a liaison name for a renowned German university's affiliated precision‑engineering laboratory. "They have a team researching a new silicon‑carbide composite micro‑machining technology that might meet your requirements. I've made preliminary contact; they're willing to discuss technology exchange and sample testing."
Xiuxiu could hardly believe it. She immediately organized the team to connect with the other side. Communication went smoothly; their research direction precisely addressed her pain point. More surprisingly, they indicated they could prioritize providing a batch of experimental nozzle samples for testing, with very reasonable terms.
Later, Xiuxiu incidentally learned that Mozi, through a complex overseas technology‑investment‑fund channel, had provided that laboratory with substantial R&D funding under the guise of "sponsoring frontier‑material research," specifying the research direction, thereby rapidly advancing the matter. He didn't directly purchase embargoed components, but in a more subtle, longer‑term way, opened an alternative path for her to obtain key technology.
When the first batch of high‑precision silicon‑carbide‑composite nozzle samples, gleaming with ceramic‑characteristic luster, arrived at the lab, Xiuxiu stroked their smooth, hard surface, emotions surging. She again sent a message in that three‑person group chat—just two words, yet weighty as a thousand pounds: "Thank you."
Mozi's reply remained terse: "Looking forward to your good news."
Holding the new nozzles, Xiuxiu and her team reignited fighting spirit. They knew this wasn't the finish line; countless hardships and dangers remained on the LPP road. Yet at this darkest moment, this support from afar—silent yet unwavering—was like a sturdy walking‑stick handed to them while trudging through mud, enabling her and her team to steady themselves, gather strength, prepare to launch a new round, firmer assault toward that violent "outburst of tin droplets." Failures might still come one after another, but the spark of hope, guarded carefully by this timely assistance, flickered stubbornly amid the ruins.