At a time when the global semiconductor industry is facing the challenge of slowing down Moore's Law, the breakthrough achievement brought by the Fudan University team - the world's first 32-bit RISC-V architecture microprocessor based on two-dimensional semiconductor materials "Wuji" has caused an industry sensation. This chip not only broke through the bottleneck of large-scale integration of 2D semiconductor materials, but also set a scale record of 5,900 transistors, far exceeding the previous record of 115 transistors at the highest international level, and achieved a 99.77% inverter yield in the experiment. This achievement is not only a milestone breakthrough for China's semiconductor industry, but also rekindles the industry's confidence in surpassing Moore's Law.
Ⅰ Moore's Law Anxiety: The Semiconductor Industry's Extreme Dilemma
Moore's Law has been a cornerstone of the semiconductor industry since the '60s, with the number of transistors doubling every 18 to 24 months. However, as silicon-based processes approach their physical limits, the cost and difficulty of chip manufacturing have risen dramatically, making it difficult for traditional silicon-based semiconductors to continue to grow exponentially. For example, the current state-of-the-art 3nm process chips are already facing serious mass production challenges, and the R&D focus of leading companies such as TSMC has begun to shift from simple transistor scaling to new architecture design and material innovation to meet the challenge of slowing Moore's Law.
Against this backdrop, there are high hopes for two-dimensional semiconductor materials. Due to their atomic-level thickness and excellent electronic properties, 2D materials are expected to break through the physical limits of silicon-based chips and push the semiconductor industry into a new stage of development. However, the controllable growth, device integration, stability and large-scale manufacturing process of 2D materials still face many technical bottlenecks, resulting in the highest 2D semiconductor digital circuit integration in the world was only 115 transistors. The "Wuji" chip of the Fudan University team directly increased this number to 5,900 transistors, which undoubtedly injected a shot in the arm for the development of global two-dimensional semiconductor chips.
Ⅱ Technological Leap: How to Achieve the 5900 Transistor Breakthrough
1. Advanced 2D semiconductor process
The breakthrough of the Fudan team first comes from the innovative use of two-dimensional semiconductor materials. Traditional silicon-based chips rely on the high-quality growth of monocrystalline silicon, while two-dimensional semiconductor materials (such as transition metal chalcogenides such as MoS₂ and WS₂) must rely on processes such as chemical vapor deposition (CVD) for large-area growth. However, the CVD process has the challenges of many material defects and low uniformity, which greatly limits the application of 2D materials.
To this end, the Fudan team adopted flexible plasma treatment technology, combined with precise control of atomic-level interfaces, to achieve high-quality growth of 2D semiconductor materials. This not only reduces the material defect rate, but also improves device consistency, laying the foundation for large-scale chip integration.
Figure: Fudan team develops the world's first 2D semiconductor RISC-V processor "Wuji"
2. "Dual-engine" optimization: AI-driven precise control of the whole process
In terms of manufacturing process, the Fudan team uses full-process AI algorithm optimization, that is, machine learning analyzes thousands of sets of experimental data to automatically optimize key process parameters. This "AI + semiconductor manufacturing" model enables precise control of all links from material growth, device manufacturing to integrated processes, and ultimately ensures the high-quality output of chips.
Through these innovations, the Fudan team successfully broke through the stability problem of 2D semiconductor devices in high-integration scenarios, and finally achieved a breakthrough in the integration of 5,900 transistors. This achievement not only set a new international record, but also laid the technical foundation for more complex and high-performance 2D chips in the future.
Ⅲ 99.77% Yield: How to Guarantee Reliability?
For any commercial chip, yield (i.e., the proportion of chips that are successfully manufactured to specification during the production process) is a core measure of its commercial viability. The inverter yield of the "Wuji" chip is as high as 99.77%, which means that its manufacturing process has reached a very high degree of maturity.
The yield improvement of 2D materials has always been a pain point in the industry, mainly due to material defects, electrical stability and process consistency. The breakthrough points of the Fudan team include:
Optimized device interface engineering: The use of a low defect density dielectric layer reduces electronic traps and improves switching ratio and signal stability.
Reduced leakage current: Through material doping optimization, the closed-state leakage current is reduced to the level of silicon-based chips, which improves the overall reliability of the device.
Single-stage high-gain amplification design: Increases the gain of the inverter, making the switching speed of the logic circuit faster and the signal clearer.
It is these engineering breakthroughs that make the "Wuji" chip not only have advantages in theory, but also show excellent reliability in actual manufacturing.
Ⅳ RISC-V architecture: the key layout of an autonomous and controllable ecosystem
In terms of chip architecture, "Wuji" adopts RISC-V architecture, which is of great strategic significance. As an open-source instruction set architecture, RISC-V not only circumvents foreign IP licensing restrictions, but also enables domestic manufacturers to freely customize chip architectures to adapt to different application scenarios.
For China's semiconductor industry, the wide application of RISC-V architecture means that there is no need to be subject to ARM or x86 ecology in the future, and it can completely independently build a complete ecosystem from underlying hardware to software adaptation. The Fudan team's choice of RISC-V architecture not only improves the scalability of the chip, but also provides more possibilities for subsequent commercial implementation.
Ⅴ Industrialization prospects: When will 2D chips enter the market?
At present, the Fudan team has applied for more than 20 core process patents and proposed a set of manufacturing solutions that can be compatible with existing silicon-based production lines. They said that about 70% of the 2D semiconductor integration process can continue to use the silicon-based mature process, which means that the industrialization cost of 2D chips is expected to be greatly reduced in the future.
In terms of application, the "Wuji" chip has the characteristics of low power consumption and high computing efficiency, and is suitable for scenarios such as the Internet of Things, edge computing, and AI inference. The next goal of the Fudan team is to further improve the integration on the basis of existing technologies, and carry out industrial cooperation with domestic and foreign enterprises to promote 2D semiconductors from the laboratory to the market.
Ⅵ Conclusion: How will 2D chips change the future?
Fudan University's "Wuji" chip is not only a major technological breakthrough, but also an important step for China's semiconductor industry to move towards an independent and controllable computing architecture. With the continuous progress of 2D materials and the maturity of the RISC-V ecosystem, we may see 2D chips gradually move towards the mainstream market in the future and become a new computing paradigm in the post-Moore era.
At a time when the pattern of the global semiconductor industry is undergoing profound changes, the success of the "Wuji" chip has not only injected new impetus into China's scientific and technological innovation, but also provided a new development direction for the entire industry. We have reason to believe that in the near future, China will play a more important role in the global semiconductor field and lead the industry change with innovation.
