The joyful atmosphere from the successful mass‑production of "String‑Light One," like a carefully maintained greenhouse, had not yet allowed every plant within the park to fully stretch out when a cold front from across the ocean, carrying the biting breath of technological change, pierced all soundproof glass and strategic buffers, blowing directly onto Xiuxiu's desk. The large LCD screen was live‑streaming a global semiconductor technology summit in Silicon Valley. The chief technology officer of overseas semiconductor giant "Cylinx"—a scholar‑entrepreneur known for radical technical roadmaps and grand‑narrative ability—stood under the spotlight, the title on the screen behind him strikingly alarming: "Beyond Nanometers: Opening a New Era for Transistors."
Xiuxiu sat alone in her office; outside the window lay the orderly scenery of the String‑Light Research Institute park, yet her entire mind was focused on the screen, on what was about to be announced. She sensed this would not be a routine technical update, but a strategic declaration, long‑plotted, aimed at reshuffling the industry landscape. Sure enough, after reviewing the severe challenges faced by current FinFET (Fin Field‑Effect Transistor) technology below the 3‑nanometer node, the speaker pivoted and dropped a bombshell.
"Ladies and gentlemen, we stand at a historic inflection point." The Cylinx CTO's voice came through the high‑quality sound system, clear and compelling. "Traditional planar scaling is approaching physical limits; even FinFET encounters nearly insurmountable obstacles in electrostatic control and drive‑current capability. To extend Moore's Law's vitality, we must fundamentally innovate the transistor's architecture!"
The large screen behind him switched; a complex, exquisite 3D animation began to demonstrate. This was no mere patch‑up of existing structures, but a wholly new, disruptive design.
"Today, I am honored to introduce to the world that our Cylinx laboratory has achieved decisive breakthroughs in the research and development of **CFET**—**Complementary Field‑Effect Transistor**—and has successfully verified it with tape‑out!" His voice carried undisguised pride.
Xiuxiu's pupils contracted slightly. CFET! This new architecture, discussed in industry‑academia circles for years, regarded as one of the most promising candidates for the post‑Moore era, had been the first to achieve a critical breakthrough by the opponent!
On screen, the animation parsed the CFET principle in detail. It was radically different from today's mainstream FinFET and earlier planar transistors. **FinFET** can be visualized as fin‑shaped semiconductor channels standing upright on the silicon substrate, with the gate wrapping around three sides of the "fin," providing better electrostatic control than planar transistors, thus allowing transistor dimensions to shrink further. However, when the transistor's gate length shrinks to a dozen nanometers or less, even the most optimized FinFET cannot indefinitely shrink the spacing between "fins," hindering integration‑density improvement, and the fin height is limited, affecting drive current.
**CFET**, in contrast, adopts a radical **three‑dimensional stacking** approach. The animation showed it stacks the **N‑type field‑effect transistor and P‑type field‑effect transistor**—the two fundamental units constituting CMOS logic circuits—not side‑by‑side on the silicon‑wafer plane, but like building blocks, vertically! The N‑type transistor at the bottom, P‑type transistor on top (or vice versa), sharing the same gate structure.
"The CFET architecture brings revolutionary advantages!" the Cylinx CTO expounded passionately. "First, it dramatically increases **transistor density**. Through vertical stacking, it can essentially double the effective number of transistors without increasing chip area, or achieve higher functional integration in the same area—key to extending Moore's Law!"
Xiuxiu stared tightly at the animation, her mind racing. Yes, vertical stacking—an excellent idea to bypass the bottleneck of 2D planar scaling. But the process complexity behind this would rise exponentially.
"Second, it significantly optimizes **interconnect length and performance**." The animation displayed that because N‑ and P‑type transistors stacked vertically, interconnects between them could be made extremely short, reducing signal‑delay and power consumption—crucial for boosting overall chip speed and energy efficiency.
"Third, it provides superior **electrostatic control** and **design flexibility**." The speaker continued introducing. "Optional designs of shared‑gate or independent‑gate provide circuit designers new tools to optimize performance‑power balance."
The principle was clear, the advantages enticing. Yet Xiuxiu, as an expert deeply immersed in manufacturing processes, saw more of the terrifying engineering challenges hidden behind that dazzling animation.
How to fabricate a bottom‑layer transistor on a silicon wafer, then nearly perfectly grow another layer of single‑crystal semiconductor material on top, and fabricate a top‑layer transistor thereon, ensuring atomic‑level precision in interface quality, electrical isolation, and stress control between the two layers? This involved extremely complex **epitaxial‑growth technology**, **wafer‑bonding technology**, or more radical **bottom‑up nanowire/nanosheet growth technology**.
CFET fabrication placed unprecedented new demands on **lithography technology**. It was no longer merely carving finer lines, but required high‑precision **multi‑layer pattern alignment** and **deep etching** in three‑dimensional space. Sidewall roughness, etch‑selectivity ratios, stress matching between layer materials… any tiny deviation could cause the entire stacked structure to fail. High‑NA EUV lithography machines offered higher resolution, but how to utilize that resolution to achieve the complex 3D structures CFET required needed entirely new **computational‑lithography models**, **multi‑patterning strategies**, and matching **etch** and **thin‑film‑deposition** processes. Lithography‑technology progress must evolve synergistically with device innovation, even needing to intervene early, tailoring lithography and process solutions for new device architectures.
The Cylinx CTO did not divulge many specifics of how they overcame these technical difficulties—that was core confidential. But the preliminary test data he displayed were sufficient to shock: test chips based on CFET structures showed, at equal power, a forty‑percent performance increase, or, at equal performance, a nearly fifty‑percent power reduction, with transistor density reaching unprecedented levels.
"…This is not merely a technology iteration; this is a paradigm revolution!" the speaker concluded with a stirring statement. "We believe CFET will lead the semiconductor industry's development trend for the next decade! Cylinx already holds the key to the future!"
The live‑stream ended. The screen darkened, reflecting Xiuxiu's grave countenance.
The office fell silent, only the low hum of the air‑conditioning system. Leaning back in her chair, Xiuxiu closed her eyes. Pressure, like substantial seawater, surged from all sides, heavy and cold. She could almost hear the commotion in capital markets triggered by this news, see the renewed worry in domestic industrial‑chain partners' eyes, feel the expectations and inquiries from higher levels.
"String‑Light One" had just achieved catching‑up, even neck‑and‑neck in some metrics, yet the opponent had already fired the starting gun on the next‑generation line. This was the cruelty of technological competition—unending, like sailing against the current.
Yet beneath this immense pressure, a familiar, almost instinctual emotion began to surge in Xiuxiu's heart—not fear, not dejection, but a challenge‑ignited, intense fighting spirit and innovative impulse.
She opened her eyes; her gaze again turned sharp and clear. Picking up the internal communicator, she dialed the core R&D team.
"Ten minutes, conference room one, emergency technical discussion." Her voice was calm, yet carried unquestionable decisiveness.
In the conference room, the atmosphere was somewhat oppressive. Team members had evidently also learned of the Cylinx announcement; each face bore gravity.
Xiuxiu did not waste time soothing moods; she went directly to the electronic whiteboard and picked up the stylus.
"Everyone, Cylinx's CFET—you've all seen it." Her gaze swept over every technical lead present. "This is indeed a powerful challenge, a technological 'second wave.' It tells us that merely mastering High‑NA EUV lithography technology is far from enough. Device innovation and lithography advancement must intertwine like DNA's double helix, evolving synergistically."
She drew a simple schematic on the whiteboard: lithography machine on one side, transistor structure on the other, connected by double arrows.
"CFET's emergence is not our doom, but points out the next strategic high ground we must conquer!" Xiuxiu's voice gradually rose, filled with strength. "The challenges it brings are immense, but the opportunities are equally immense! This forces us to leap out of the original thinking framework, not content with merely optimizing process parameters within existing architectures."
She began guiding the team through brainstorming.
"First, we need to immediately launch a **CFET Special‑Research Group**." Xiuxiu wrote "CFET" on the whiteboard, circling it. "Concentrate efforts on conquering CFET's key process modules: high‑quality selective epitaxial growth, low‑damage etching technology, high‑aspect‑ratio contact‑hole filling, and optimization of inter‑layer dielectric materials… We must, in the shortest time, thoroughly digest its technological principles and manufacturing difficulties."
"Second, the **Computational‑Lithography Team** needs early involvement." She looked at the computational‑lithography expert. "Develop new optical‑proximity‑correction models, inverse‑lithography‑technology algorithms targeting CFET's complex 3D structures. We need to simulate light propagation and reflection in multi‑layer stacked structures, predict and compensate possible pattern distortion. This requires more powerful computing power and more advanced algorithms."
"Third, the **Materials Team**'s task is even more arduous." Xiuxiu's gaze turned to the materials‑department head. "Find new channel materials more suitable for CFET stacking? How to control stress between stacked layers, avoid defect generation? How to ensure perfect electrical isolation between two transistor layers? These need source innovation in materials science."
"Finally, and most importantly," Xiuxiu put down the pen, placed both hands on the conference table, leaning forward slightly, her eyes burning at everyone, "we cannot be content with merely following and imitating. Cylinx gave a direction, but paths to that direction may be more than one. While deeply understanding CFET, we must think: are there our own, more competitive transistor‑architecture innovations? For instance, combining our prior technological accumulation in nanowire transistors, exploring **gate‑all‑around nanowire CFET**? Or exploring possibilities using 2D materials as channels?"
Her words, like a stone thrown into a calm lake, stirred sparks in the team members' eyes. Pressure began transforming into drive; gravity gradually replaced by the excitement of meeting challenges.
"Dr. Xiu, you mean… we not only follow up, but think about surpassing?" a young engineer could not help asking, voice tinged with excitement.
"Why not?" Xiuxiu countered, even a confident curve touching her lips. "We could build 'String‑Light One'; can't we conquer this CFET fortress? Moreover, what we must think is not simply conquering it, but how to conquer it via a better path, lower cost, higher performance! This second wave—we not only need to stand firm, but strive to ride its crest!"
The meeting lasted several hours, from breaking down technical difficulties to allocating resources, from setting short‑term goals to planning long‑term technology roadmaps. When Xiuxiu finally declared adjournment, though weariness marked each face, that familiar glint—of overcoming difficulties—rekindled in their eyes.
Xiuxiu was the last to leave the conference room. She walked to the corridor window, gazing outside. Sunset's afterglow gilded the park gold. What she felt was no longer the heavy pressure right after the announcement, but a weighty sense of responsibility and a burgeoning innovative passion.
The opponent's strength precisely proved this path's value. This second wave struck swiftly, yet thoroughly ignited her and the team's next‑round huge innovative drive. She knew the road ahead remained treacherous, yet she and her team were ready to meet the wave head‑on. The lithography‑machine battle temporarily concluded, but the more magnificent innovative long‑march belonging to China's chip industry had just entered new deep waters.