Chapter 112 111: Mass Production
Li Qingsong had originally thought that after successfully breaking through the 7-nanometer, 5-nanometer, and 3-nanometer nodes in quick succession, the 2-nanometer chip technology would also be conquered with ease.
But he hadn't expected that at this moment, as they approached the very limits of physics, a multitude of technical challenges, each one impossibly complex and difficult, would all surface at once.
'Come to think of it, this is actually quite normal.'
Li Qingsong sighed softly. "To precisely etch over forty billion transistors onto a silicon wafer just a few square centimeters in size, while also ensuring they maintain their designed connections and work in concert... How could that possibly be simple?"
But no matter how difficult, it had to be done.
Li Qingsong reviewed the existing technical hurdles and once again began the task of overcoming them.
He discovered that a major constraint preventing him from mastering 2-nanometer chip technology was the short-channel effect, which had become far more severe than in previous nodes.
At such a minuscule scale, leakage current was nearly impossible to control. This prevented the transistors from functioning correctly; as soon as power was applied, vast numbers of them would burn out instantly.
Li Qingsong mobilized the brainpower of tens of thousands of Clones, simultaneously thinking, running simulations on a supercomputer, and conducting a wide range of experiments.
Ultimately, he devised a solution.
'Perhaps... I can change the transistor's structure? By wrapping the channel with the gate, I might be able to improve control and reduce the leakage current.'
With this idea in mind, Li Qingsong immediately began his attempts.
Using the lab's high-precision lithography suite, Li Qingsong attempted to fabricate a brand-new chip based on the gate-all-around field-effect transistor model, then powered it up for testing.
After a series of tests, Li Qingsong looked at the current data on the screen and let out a long sigh of relief.
He knew he had conquered the technical problem of leakage current.
But this was merely the first hurdle in the process of developing the 2-nanometer chip.
Soon after, Li Qingsong ran into the second technical challenge.
How could he further increase the light source's power to achieve an even shorter wavelength?
Obviously, only with a shorter wavelength of light could smaller transistors be etched.
But the lithography machine Li Qingsong was currently using was already what one would call an "extreme ultraviolet" (EUV) light source machine.
And the "extreme" in its name was no exaggeration.
But now, even "extreme" was not enough; he had to find a way to push it even further.
Li Qingsong once again marshaled immense brainpower and resources, beginning a new round of attempts, sparing no expense.
Several days later, a new idea emerged from the mind of one of the Clones and was synchronized to Li Qingsong's own.
'Perhaps we can try a different material for the Laser-produced plasma source.'
After a series of experiments, Li Qingsong finally discovered that bombarding tin droplets with a high-power carbon dioxide Laser produced a light source with a wavelength that met his requirements.
This problem, too, was conquered.
But, with the light source problem solved, the issue of photoresist floated to the surface.
The existing photoresist couldn't effectively absorb and react to light of this wavelength, making it impossible to properly etch transistors onto the silicon wafer.
Furthermore, its sensitivity was too low. While its associated defects had been tolerable for the previous chip generation, they were now magnified to an unacceptable degree in the manufacturing of 2-nanometer chips.
'I have to develop a new material.'
Li Qingsong set his mind to it, once again launching a series of material science experiments. He directed hundreds of thousands of Clones to conduct a total of over ten million trials, and in the end, they found a material that met the requirements.
It was a new type of photoresist based on metal oxides and organic-inorganic hybrid materials, which perfectly solved all the problems of the previous generation.
After this, yet another problem emerged.
The reflectivity of the mask used in the lithography process was insufficient, causing the light source to lose energy and preventing the etching of high-quality transistors.
Li Qingsong tackled the problem again, managing to solve it by using multi-layer molybdenum-silicon reflective mirrors.
After that, problems still appeared, one after another.
Issues with the optical system, stability, precision, thermal management, heat dissipation systems—the list went on and on, an almost endless stream of problems.
Often, he would solve one problem, thinking he could finally move forward, only for hundreds or even thousands of new issues to appear in the next moment.
He would solve those hundreds or thousands of problems, only for tens of thousands more to emerge.
Sometimes, a single, minuscule problem could stall the millions of chip-research Clones under Li Qingsong's command for over ten days. But even when he mobilized immense brainpower and resources to finally resolve it, he would find he had only advanced by one tiny step.
Time slowly slipped by, consumed by the endless flood of new problems and the millions of Clones' round-the-clock trials, deliberations, and research.
Finally, one day, after resolving yet another issue, Li Qingsong subconsciously glanced at the problem list. He was stunned to find that the list... was empty.
Empty?!
A jolt of exhilaration shot through Li Qingsong.
'After all this time, all this struggle... have I finally mastered 2-nanometer chip manufacturing technology?'
'Granted, what I've mastered is only the laboratory fabrication technique. It's still a long way from mass production.'
'Once we enter the mass production stage, countless more problems will surely emerge, and I'll have to solve each one.'
'But at least, from the standpoint of fundamental principles and processes, I have truly and completely mastered this technology! All future issues will be mere engineering problems, not matters of fundamental science!'
Energized, Li Qingsong immediately began upgrading one of the factories that had been producing 3-nanometer chips. He had more precise equipment manufactured to replace the old machinery and began a trial production run.
Sure enough, as soon as trial production began, a host of new problems appeared. But it didn't matter. Li Qingsong had a million Clones on standby.
When a problem appeared, he simply solved it.
And so, the chip factory began a period of rapid optimization and iteration. Finally, after six months, the production line stabilized.
The first batch of 2-nanometer process chips was finally produced.
Although the yield was currently less than 10%—meaning that of the factory's annual capacity of 10 million wafers, fewer than 1 million chips met specifications—it didn't matter.
'At least I can build the supercomputer first! The yield can be improved later.'
Thus, Li Qingsong took the factory's entire first-year output of usable chips—approximately 1 million of them—and used them all in the construction of a single supercomputer.
At the same time, a full 100 newly built 2-nanometer process chip factories had already begun production.