In the cleanroom of the lithography‑machine R&D base, time seemed stretched to infinity, each second saturated with a near‑solidified anxiety and anticipation. What unfolded here was no longer the magnificent epic of EUV challenging physical limits, but a silent, brutal trench warfare on the nanometer scale—against fluid dynamics, surface physics, and extreme cleanliness requirements. The final, and most stubborn, barrier of **immersion‑DUV lithography mass‑production technology** loomed before Xiuxiu and her team: **"water‑droplet control,"** or rather, the stable formation and maintenance of that **ultrapure‑water film** which served as the optical medium, bearing the miraculous mission of shortening 193‑nm wavelength to 134 nm.
The principle had long been crystal clear. Yet transforming that principle into a stably maintained, flowing water film—uniform in thickness, bubble‑free, vibration‑free, contamination‑free, and non‑contact with both—above a silicon wafer moving several meters per second with nanometer‑precision stepping, and below the ultimate lens element worth tens of millions of dollars and exquisitely delicate, was a difficulty comparable to maintaining absolute straightness and stability of a spider‑silk thread in a hurricane.
Xiuxiu stood before the observation window of the upgraded immersion lithography machine, her gaze penetrating the specialty glass, locked onto the key component called the **"immersion head."** Positioned between the projection objective's final lens and the high‑speed silicon wafer, its core mission was to execute this exquisitely intricate yet extremely fragile **"dance of water droplets."**
The challenges were multiple and coupled:
**The paradox of non‑contact:** The water film must never contact the lens lower surface nor the photoresist on the wafer. Any physical contact meant potential contamination, particle residue, even damage to the optical element. Yet water must be brought infinitely close to both surfaces to form a complete optical channel. This required the immersion head to "conjure" a layer of water out of thin air, suspending it within the tiny gap—like a magician. **Flow stability:** Water must flow continuously to carry away possible minute heat and any potential contaminants. But this flow must be absolutely **laminar**; the slightest turbulence would introduce unpredictable refractive‑index variations and microbubbles into the water film, causing imaging distortion and defects. Maintaining such laminar flow during the wafer‑stage's high‑speed scanning, abrupt stops, and directional changes was like maintaining an absolutely calm water surface behind a speedboat. **Edge control and recovery:** The water film must be "cut off" cleanly and efficiently at the wafer edge, with no droplet splashing or residue on the wafer backside or inside the machine—otherwise it would become a catastrophic contamination source.
To tackle these problems, the team focused on two core areas: **micro‑environment control technology** and **hydrophilic/hydrophobic material surface engineering**.
**Micro‑environment control technology** aimed to create an absolutely controlled "micro‑climate" within the tiny gap between lens and wafer. Xiuxiu's team designed the immersion head itself as an extremely complex micro‑fluidics system.
* **Ultrapure‑water injection system:** Water flow, purified through over a dozen stages and precisely temperature‑controlled, was injected into the gap via multiple micron‑scale‑aperture nozzles at extremely low speed and uniform pressure. Nozzle arrangement and angle were optimized through countless computational‑fluid‑dynamics simulations, ensuring the flow began as steady laminar.
* **Gas‑curtain protection system:** Around the injected water flow, an extremely thin but stable **inert‑gas (e.g., nitrogen) curtain** was simultaneously injected. This gas curtain acted like a loyal guard, on one hand isolating the water flow from external possible particles, on the other helping shape the flow's form and assisting clean separation of the water film at the wafer edge. Gas flow rate, pressure, and water‑flow matching needed to achieve microsecond‑level synchronization and balance.
* **Negative‑pressure recovery system:** At the wafer edge and water‑film outflow region, precision negative‑pressure suction holes were designed, like efficient miniature vacuum cleaners, instantly "sucking" the outflowing water film into recovery piping, which after re‑purification cycled it back, ensuring no liquid retention or splash.
Yet relying solely on precise fluid‑dynamics control was far from enough. Water, as a liquid with surface tension, behaves when contacting solid surfaces depending on the material surface's chemical nature. This was where **hydrophilic/hydrophobic material surface engineering** shone.
* **Hydrophobic lens protection:** The final lens's lower surface was coated with an extremely robust, smooth **super‑hydrophobic coating**. This coating's microstructure mimicked the lotus‑leaf effect, making the contact angle between water molecules and the lens surface extremely large (usually >150°). When water droplets or film attempted to approach the lens surface, they were strongly repelled by this "water‑repelling" coating, forming an air‑cushion‑like effect, thereby physically preventing contact—achieving a key link of "non‑contact" immersion. Maintaining this super‑hydrophobic coating's stability under long‑term high‑speed water‑flow impact and high‑energy UV irradiation was itself a huge materials‑science challenge.
* **Hydrophilic wafer guidance:** On specific regions of the wafer carrier stage (not the wafer itself, which has photoresist), **hydrophilic materials or coatings** might be employed. Such hydrophilic surfaces "attracted" water molecules, helping guide the water film to spread smoothly and flow toward the recovery port at the wafer edge, avoiding random film rupture or irregular retraction, ensuring water‑film shape stability and recovery efficiency.
Theory was complete, design exquisite. But practice was cold, brutal failure, again and again. In the cleanroom, the machine started over and over, attempting to form that ideal perfect water film. The high‑speed‑camera footage, however, was always disheartening:
Sometimes the water film trembled violently during initial formation, like a soap bubble shattered by wind;
Sometimes, at the instant of wafer‑stage acceleration, laminar flow broke, invisible micro‑vortices arose in the film, leading to subsequent exposure patterns developing hard‑to‑trace, randomly distributed defects;
Sometimes the edge‑recovery system showed millisecond‑level delay or fluctuation, causing extremely tiny droplets to be flung out like lethal stray bullets, landing where they shouldn't;
More often, those ghost‑like, nano‑scale bubbles of unknown origin stubbornly appeared in the water film, casting fatal shadows during exposure.
Xiuxiu bore immense pressure. Her face grew increasingly haggard, yet her eyes, due to extreme concentration, appeared unusually bright—even somewhat alarming. She practically lived in the workshop, analyzing high‑speed camera data with team members, adjusting fluid parameters, checking every valve, every sensor, every pipeline's cleanliness. Her demands were near‑harsh, pursuing any tiny anomaly relentlessly.
"Gas pressure down another 0.001 Pa!"
"Zone B recovery negative‑pressure curve insufficiently coupled with wafer‑stage acceleration curve—rematch!"
"This batch of super‑hydrophobic coating contact‑angle test data shows 0.5‑degree fluctuation—suspend entire batch of lenses, send for inspection!"
Driven by her nearly burning will, team members wound up like clockwork, struggling forward in cycles of hope and disappointment. The air was thick with a mixture of exhaustion, tenacity, and a faint trace of despair.
During a large‑scale systematic test targeting water‑film stability under wafer‑stage high‑speed scanning mode, the team had been fighting continuously over forty hours. Failed data piled up like a mountain; everyone's physical and mental energy neared the limit. Xiuxiu felt her temples hammered, dizzy spells washing over; she nearly ordered the test paused for rest.
Just then, the young engineer monitoring the main screen let out an almost distorted, incredulous whisper: "St‑stabilized?!"
All eyes instantly focused on the primary screen displaying real‑time water‑film morphology and key stability parameters. That thin water film—visible under high‑speed camera, suspended between lens and wafer—after the wafer‑stage completed an entire cycle of violent acceleration‑constant‑speed‑deceleration‑reverse scanning, actually… remained nearly perfectly flat and stable! Edge recovery was clean, without any visible tremor or bubble generation.
Seconds ticked by.
Thirty seconds… one minute… three minutes…
Key parameters—water‑film thickness uniformity, refractive‑index stability, edge‑contact‑line fluctuation—all stayed within the stringent green qualification zone.
Five minutes… eight minutes…
The cleanroom was utterly silent, only the faint hum of equipment running. Everyone held their breath, eyes glued to the screen, as if fearing the slightest sound would disturb this fragile, precious balance.
**Ten minutes!**
When the timer steadily crossed the ten‑minute threshold, with all parameters still perfect, the silence broke.
Not cheers, not shouts. First came suppressed, choked sobs. A gray‑haired senior engineer removed his antistatic glasses, wiping his eyes forcefully with the back of his hand. Then, like a chain reaction, more eyes reddened, tears sliding down tired faces over many days, dripping onto cleanroom suits, spreading tiny wet marks.
Xiuxiu stood in place, body trembling slightly; she clenched her lower lip, trying to hold back the surging sourness and excitement. But in the end, scalding tears broke through the dam, blurring her vision. She gazed at the steadily flowing "dance of water droplets" on the screen—like a most perfect artwork—and looked at these comrades who had stood shoulder‑to‑shoulder with her, endured countless failures yet never gave up. An immense sense of achievement and relief, almost suffocating, flooded her, washing away all exhaustion and pressure.
They had done it. On this nanometer‑scale battlefield, they had finally preliminarily tamed that unruly water flow, making this exquisite dance last so long, so stable, for the first time.
Those ten minutes were not merely a time datum. They were a milestone, proving the correctness and feasibility of their chosen path of micro‑environment control and surface engineering; they were a shot of adrenaline, injecting courage and hope into a team on the verge of collapse; they were the first truly solid light illuminating the road to mass production.
Xiuxiu let the tears flow; she didn't wipe them, just reached out and tightly, firmly shook the hand of the nearest team member—also eyes brimming with hot tears. No words were needed; everything was said without saying.
The dance of water droplets had finally found its rhythm. On the road to immersion mass production, the hardest boulder had finally been pried open—a crack. Hope, like that steadily flowing ultrapure water, began trickling once more through this space once filled with frustration.