The chill of early spring had not yet fully receded from the North China Plain, yet within the lithography‑machine R&D base, it already felt like a scorching battlefield. Unlike the futuristic, physics‑challenging magnificent poetry of the EUV prototype, on another crucially important production line, an even more arduous and grueling battle‑decisive for reality and efficiency—was underway: **the final breakthrough of immersion deep‑ultraviolet lithography mass‑production technology**. EUV is the key to the future, but immersion DUV is the true industrial backbone supporting over seventy percent of global chip manufacturing—deserving its name. The goal of Xiuxiu's team was not only to master that key, but also to temper that backbone to its utmost, enabling it to support the nation's urgent demand for more advanced process nodes and higher‑performance chips.
Xiuxiu stood before the observation window of the upgraded immersion‑DUV lithography machine, brow tightly furrowed, like the gathering clouds outside refusing to disperse. Inside the machine, another full‑speed operation test was in progress, aiming to achieve stable, high‑speed, near‑zero‑defect mass production of 28‑nanometer‑node‑and‑below process chips. Yet on the monitoring screen, the constantly fluctuating defect‑density data and occasional red alarms representing pattern failure resembled cold needles pricking her taut nerves.
What they faced was precisely the last, most stubborn fortress that must be conquered when pushing immersion technology from laboratory proof‑of‑principle into the harsh environment of large‑scale mass production. The core of this **immersion revolution**, seemingly a simple physical principle, was in engineering‑implementation difficulty comparable to harnessing a never‑ending, perfect miniature tsunami at the nanometer scale.
Its principle originated from an ingenious optical concept. In traditional dry lithography, 193‑nanometer‑wavelength deep‑ultraviolet light propagates through air (refractive index n≈1.0), its theoretical resolution limited by the classic **Rayleigh criterion**: CD = k₁ * λ / NA. Where CD is critical dimension, λ is wavelength, NA is numerical aperture, k₁ is process factor. To obtain smaller CD, either shorten wavelength (the arduous path leading to EUV), or increase numerical aperture NA.
Immersion technology chose a more ingenious path—**changing the refractive index of the light‑propagation medium**. It injects a layer of ultrapure water between the final lens element of the projection objective and the silicon wafer coated with photoresist. Water's refractive index is approximately 1.44, far higher than air's 1.0. When 193‑nm light enters water from the lens (glass with higher refractive index), its **effective wavelength** shortens to λ_water = λ_air / n_water ≈ 193 nm / 1.44 ≈ **134 nanometers**!
This, without changing actual light source or lens design, essentially "created" a shorter‑wavelength light source out of thin air, thus breaking through dry‑lithography's resolution limit in one stroke, making it possible to continuously push chip processes toward smaller nanometer nodes using relatively mature DUV technology.
The principle was clear and elegant, but the devil lay in the details. Transforming this principle into stable, reliable, high‑speed mass‑production technology required overcoming a series of daunting engineering challenges, the core being absolute control over that thin, continuously flowing water film.
**Extreme Purity of Ultrapure Water:** This wasn't the "pure water" of daily concept. It needed to achieve near‑theoretical absolute purity; any tiny ions, organic particles, even bacterial residues would become "boulders" on nanometer‑scale circuit patterns, causing fatal defects. The team adopted over a dozen stages of filtration, ion‑exchange, ultraviolet sterilization, and degassing systems, ensuring water resistivity reached the extreme 18.2 MΩ·cm, total organic‑carbon content below parts‑per‑billion levels. Any tiny contamination fluctuation would leave clear, frustrating imprints on final defect‑distribution maps.
**Constant‑Temperature Water‑Flow Control—Nanometer‑Scale Thermal Expansion/Contraction:** Lithography is an extremely temperature‑sensitive process. Lenses, silicon wafers, even photoresist undergo nanometer‑scale deformation due to minute temperature changes, causing pattern distortion. Flowing water itself is both heat‑transfer medium and thermal‑capacity body. The temperature of injected water flow must be controlled within astonishing stability ranges—usually requiring fluctuation less than ±0.01 degrees Celsius. Any tiny temperature fluctuation would cause minute changes in water's refractive index (dn/dT effect), altering effective focal length, causing imaging blur. More frighteningly, non‑uniform temperature fields could induce unpredictable convection within the water film, directly interfering with stable light‑wave propagation. For this, they designed complex stratified constant‑temperature systems; from water‑source reservoir, pipeline delivery, to final injection head, every link was wrapped in precision temperature‑control sleeves, as if nurturing a life form with extreme temperature demands—an exquisitely delicate organism.
**Perpetual War Against Bubbles:** This was the most stubborn, hardest‑to‑eradicate "chronic illness" in immersion‑process technology. Gases dissolved in water, under pressure or temperature changes, precipitate forming micron‑even nanometer‑scale bubbles. These bubbles, for 134‑nm effective‑wavelength light, were like tiny, opaque "black holes" or strong scattering centers. A tiny bubble lingering in the exposure region could ruin local patterns across an entire chip.
* **Bubble Generation:** Originates from cavitation effects as water flows through valves, pump bodies, microscopic roughness on pipeline inner walls, even precipitation of dissolved gases from water itself under pressure release.
* **Bubble Transport and Capture:** Tiny bubbles move with water flow, possibly captured by microscopic structures on wafer surfaces or vortices generated by high‑speed wafer‑stage motion, adhering to critical regions.
* **Bubble Elimination:** The team adopted multiple strategies. Powerful online degassing modules continuously strip dissolved gases from water; flow‑channel design underwent countless computational‑fluid‑dynamics simulation optimizations, avoiding low‑pressure zones and vortices; injection‑head nozzle shapes were meticulously designed, ensuring water flows with stable, laminar‑flow state cover wafer surface, forming uniform, gapless "liquid curtains." Yet bubbles were like phantoms, always appearing at unexpected moments, in unexpected ways, challenging engineers' patience and wisdom.
Xiuxiu was under immense pressure. National‑special‑project support meant higher expectations and more urgent timelines. Downstream chip‑design companies in the industrial chain awaited domestic manufacturing platforms with higher performance and higher yields. Every test run carried high costs and enormous team‑energy investment. Those lingering defect alarms on screens, and electron‑microscope‑observed pattern bridging or fracture caused by minute water‑flow fluctuations or stubborn bubbles, all weighed like heavy stones upon her heart.
She demanded near‑harsh standards from the team, and herself. Every process parameter was repeatedly scrutinized, every component required to reach performance limits, every failure required root‑cause analysis, traced back to the minutest details. She stayed long hours at the workshop, analyzing data with engineers, adjusting parameters, even personally participating in key‑component disassembly/inspection. Dark circles beneath her eyes grew heavier, voice huskier from constant discussion and instruction.
"Reduce water‑flow speed another 0.1 meters per second, observe boundary‑layer stability!"
"This batch of filter‑cartridge ultrapure‑water indicator shows 0.01% fluctuation, immediately investigate upstream supply system!"
"Optimize wafer‑stage acceleration curve, reduce turbulence excitation in exposure region!"
Her directives were clear, decisive, leaving no room for doubt; yet team members could see that deep, hard‑to‑conceal anxiety and weariness in her eyes. They understood this strictness, for all knew they were assaulting another resplendent, hard pearl on the semiconductor‑manufacturing domain's crown. Any compromise might mean total failure.
After thirty‑six consecutive hours of struggle, attempting yet again failing to solve a periodically occurring imaging‑drift problem related to water‑temperature fluctuations, Xiuxiu alone returned to her office. She closed the door, shutting out all external sound, exhausted, collapsing into her chair. Outside, night was deep, the base's lights in the cold wind appearing especially lonely.
She felt a bone‑deep helplessness. The joy brought by the EUV‑prototype success had long been diluted by this seemingly more "conventional" yet equally insurmountable process mountain. She knew the principle, knew the path, even knew most answers, yet couldn't perfectly, stably integrate them, transform them into the machine's steady‑running pulse. This feeling—seeing the finish line, yet always tripped in the final steps by invisible obstacles—felt more frustrating than facing entirely unknown EUV.
Subconsciously, she retrieved the encrypted communicator, finger hovering above that familiar group. She wanted to hear Mozi's and Yue'er's voices, even just a few irrelevant greetings. But she hesitated. Mozi faced financial‑sniper pressure; Yue'er struggled in academic‑debate whirlpools. She couldn't, shouldn't, disturb them with her own predicament.
Just as she prepared to pocket the communicator, the screen lit up on its own. A message from Yue'er, without text, only a simple mathematical symbol ∞ representing "thinking" and "focus," plus a picture of a star twinkling faint light in the night sky.
Simultaneously, Mozi's message also arrived, equally concise: "Steady. Only with solid foundation can one bear weight."
No queries, no comfort, just these two seemingly ordinary yet directly striking her deepest‑heart messages at this moment. Yue'er told her the exploration process itself was infinite, hardship the norm; Mozi reminded her of the importance of solidifying foundations, and his presence as a backstop.
Xiuxiu stared blankly at the screen, nose suddenly stung; days‑long pressure, grievance, and exhaustion seemed to find an outlet, tears welling. But she forcefully held back. She couldn't fall.
She took a deep breath, wiped away moisture from her eye corners, replied two words: "Received."
Then she stood, walked back toward the workshop. Lights cast her shadow long, yet her steps were firmer than before. The immersion revolution wasn't yet successful; the war for water purity, constant temperature, and bubble‑free operation continued. She knew she must, and could, lead her team, harness this nanometer‑scale miniature tsunami, transform this 134‑nm effective wavelength into the sharpest engraving blade carving China's chip future. The darkness before dawn is deepest, yet heralds the imminent light.