Li Qingsong originally thought that under the premise that the 7nm, 5nm, and 3nm technologies had been successfully broken through one after another, the 2nm process chip technology would also be successfully broken through. But he did not expect that at this moment when he was closest to the limit, many extremely complex and difficult technical problems emerged. “If you think about it, this is actually quite normal.” Li Qingsong sighed softly: "To accurately carve more than 40 billion transistors on a silicon wafer within a few square centimeters, and to make them maintain the connection as shown in the blueprint and work together to perform their functions, how can this be easy..." But, no matter how difficult it is, we have to do it. Li Qingsong sorted out the existing technical obstacles and started the task of tackling them again. He found that a major obstacle restricting his mastery of 2-nanometer chip technology was the more serious short-channel effect than before. Leakage current is almost uncontrollable at such a tiny size. This causes the transistors to not work properly at all, and once power is turned on, a large number of transistors will burn out instantly. Li Qingsong mobilized the brainpower of tens of thousands of clones, thinking while using supercomputers for simulations and conducting various experiments. Finally, he came up with a plan.
"Perhaps... I can change the structure of the transistor? By wrapping the channel with a gate, I might be able to improve control and reduce leakage current." Thinking of this, Li Qingsong immediately started trying. Using a set of high-precision lithography equipment in the laboratory, Li Qingsong tried to build a new chip based on the surround gate field effect model and tested it by powering it on. After a series of tests, looking at the current data on the screen, Li Qingsong let out a long sigh. He knew that he had solved the technical problem of leakage current. But this is just the first difficulty encountered in the process of developing 2-nanometer chips. After this, Li Qingsong immediately encountered the second technical difficulty. How to further increase the light source power and achieve a shorter wavelength? Obviously, only by making the wavelength of light shorter can smaller transistors be carved out. But now, the lithography machine used by Li Qingsong can be called an "extreme ultraviolet light source" lithography machine. The "极" here means extreme. But now, even "extreme" is not enough, and we have to find a way to improve it further. Li Qingsong once again mobilized a lot of brainpower and resources and began an attempt regardless of cost. A few days later, a new idea emerged from the mind of one of the clones and was synchronized to Li Qingsong's head. "Perhaps we can try other materials in laser plasma light sources." After some attempts, Li Qingsong finally found that the wavelength of the light source produced by bombarding tin droplets with a high-power carbon dioxide laser generator could meet his requirements. This problem was also overcome. However, after the light source problem was solved, the problem of photoresist surfaced again. Existing photoresists cannot effectively absorb and react to light sources of this wavelength, and therefore cannot effectively engrave transistors on silicon wafers.
Moreover, its sensitivity is not high enough. Although its related defects were acceptable in the previous generation of chips, they have been magnified to an unacceptable level in the manufacturing of 2-nanometer chips. “We need to develop new materials.” Li Qingsong thought about it and started the materials science experiment again, manipulating hundreds of thousands of clones to conduct tens of millions of experiments in total. Finally, he found a material that met the existing needs. This is a new type of photoresist based on metal oxides and organic-inorganic hybrid materials, which perfectly solves the various problems of the previous generation of photoresists. After this, new problems emerged. The masks used in the photolithography process have insufficient reflectivity, resulting in energy loss from the light source and an inability to carve out transistors of sufficient quality. Li Qingsong once again started to tackle the problem and tried to use multi-layer molybdenum and silicon mirrors, and successfully solved the problem. After this, one problem after another still arises. There are issues related to the optical system, stability, precision, thermal management, cooling systems, and so on. The problems are almost endless. Often, when one problem is solved, we think the project can be implemented smoothly, but the next moment hundreds or thousands of problems arise. After solving these hundreds of problems, tens of thousands more problems will arise. Sometimes a very small problem could delay Li Qingsong's millions of chip research clones for more than ten days. But even if Li Qingsong mobilized huge brainpower and material resources to finally solve the problem, he found that he had only made a small step forward. Time passed slowly amidst the numerous emerging problems and the endless attempts, thinking and research of millions of clones day and night. Finally, one day, Li Qingsong solved a problem again. When he subconsciously looked at the problem list, he suddenly found that the problem list was... empty. Empty?! Li Qingsong suddenly felt excited. After such a long struggle, have I finally mastered the 2-nanometer chip manufacturing technology? Although what he has mastered at the moment is only the manufacturing technology in the laboratory, it is still some distance away from large-scale mass production. After entering the mass production stage, more problems will inevitably emerge and need to be solved one by one. But at least, from the basic principle and process, I have truly mastered this technology! All subsequent problems will only be technical problems, not problems of principle! Excited, Li Qingsong immediately began to upgrade one of the factories that originally produced 3-nanometer chips, producing more sophisticated equipment, replacing the original equipment, and trying trial production. As expected, many problems emerged again when the trial production just started. But it didn't matter, Li Qingsong had prepared one million clones ready at any time. When a problem arises, it is solved. So the chip factory began to optimize and iterate rapidly. Finally, after half a year, the production line was finally stable. The first batch of 2nm process chips have finally been produced. Although the yield rate is less than 10% at the moment, that is, out of the annual production capacity of 10 million chips of this chip factory, only less than 1 million chips meet production expectations, it doesn't matter. At least the supercomputer can be built first! The yield rate can be gradually improved later. As a result, all the good products of the chip factory in the first year, about 1 million chips, were taken away by Li Qingsong and used in the construction of just one supercomputer.
At the same time, 100 newly built 2-nanometer process chip factories have been put into production.