Economic Observation: Where are China's opportunities in the post-Moore era?

China News Service, Beijing, August 3 (Reporter Liu Yuying) The so-called Moore's Law, that is, "under the condition of unchanged price, every 18-24 months, the number of transistors that can be accommodated in integrated circuits doubles and the performance doubles." Today, this golden rule of the industry has been challenged, and the future direction of the market is uncertain.

Has Moore's Law, proposed by Gordon Moore, one of the founders of Intel, reached its limit? In terms of technology, several major chip companies in the world are engaged in an "arms race". TSMC's 5nm process has been mass-produced, and the 3nm process will be mass-produced in 2022. Samsung is competing with TSMC in the 3nm competition and also plans to mass-produce the 3nm process in 2022. Intel, which has fallen behind in technology in recent years, suddenly announced that it will mass-produce 2nm chips in 2024.

It seems that Moore's Law is still moving forward. Han Xiaomin, general manager of Aiji Micro Consulting, which provides professional consulting services for the entire electronics industry chain, said in an exclusive interview with China News Service that the current chip manufacturing process is getting smaller and smaller, but it is already close to the limit of physical scale due to the effective transmission of internal signals in the chip. It is very difficult to improve the process level below 10 nanometers, and it requires higher costs to ensure the yield after mass production.

How to continue Moore's Law has become a question of the times. In May this year, the 18th meeting of the National Leading Group for Science and Technology System Reform and Innovation System Construction discussed "Potential Disruptive Technologies for Integrated Circuits in the Post-Moore Era".

The industry has already made some explorations. Han Xiaomin introduced that currently, major foundry companies are still constantly improving their processes and using other means to keep Moore's Law moving forward. One of the ways is to upgrade the structure of transistors from the previous plane to three-dimensional. Currently, Samsung is betting on this GAA (Gate All Around Transistor) process.

"GAA is a clear research and development route at present, which is mainly promoted by large companies and is basically commercially feasible. However, after GAA, there is no relatively unified route and direction," said Han Xiaomin.

Heterogeneous integration is another technical line. Intel believes that heterogeneous integration is to combine hardware with different process architectures, different instruction sets, and different functions into a computing system. At the same time, it is also the synergy of chips, packaging, systems, and software, rather than a single technical point.

At present, companies such as AMD, Intel, NVIDIA, Qualcomm, Unisoc, and Beijing Ingenic are actively conducting research and development on heterogeneous integration and launching corresponding solutions.

"Heterogeneous integration has improved the performance of chip systems by one generation through packaging methods. This technology still has potential to be tapped, such as what methods and materials can be used for interconnection and stacking, and it can be further optimized and improved," said Han Xiaomin.

The industry has high hopes for the replacement of silicon-based materials. If it can be industrialized, it will be a disruptive technology. As the chip manufacturing process approaches 2 nanometers, the potential of silicon-based chips has been exhausted, and the use of new materials is considered to be a fundamental solution to the problem. Countries have invested in carbon-based carbon nanotubes and graphene experiments, hoping to gain the technological commanding heights of the post-Moore era.

It is said that the performance and integration of carbon-based chips using 90-nanometer process are expected to be equivalent to those of silicon-based chips using 28-nanometer process.

"However, the research and development of these carbon-based materials are still in the laboratory and early stages of industrialization, and there is no possibility of replacing silicon-based materials. Even if they are made into devices in the future, they will still require a lot of process polishing," said Han Xiaomin. If carbon-based materials enter commercial use in the future, based on cost considerations, they will most likely only replace silicon-based materials in some areas, and are unlikely to replace all of them.

The third generation of semiconductors is also an iteration on chip materials. The so-called third generation semiconductors refer to the third generation of semiconductor materials represented by gallium nitride (GaN) and silicon carbide (SiC). 5G base stations, new energy vehicles and fast charging are all important application areas of third generation semiconductors.

China has made it clear that it will vigorously support the development of the third-generation semiconductor industry during the 14th Five-Year Plan period. Han Xiaomin said that driven by application upgrades and policies, the third-generation semiconductor will usher in a period of rapid development. However, it mainly solves power and radio frequency problems, and cannot solve the Moore's Law problem of CPU and GPU.

In addition, there are some new architectures and technical routes, such as spin, storage and computing, online computing, etc. Han Xiaomin believes that these can be regarded as bypassing Moore's Law, with the purpose of solving local problems such as computing.

Moore's Law is facing extreme challenges, and the industry has proposed a variety of technical directions. This is both a turning point and an opportunity, providing Chinese companies with a direction to catch up. China has identified heterogeneous integration, carbon-based technology, and third-generation semiconductors as key development directions, and has achieved some breakthroughs.

Han Xiaomin was the general manager of the Integrated Circuit Center of CCID Consulting of the Ministry of Industry and Information Technology of China. He graduated with a bachelor's and master's degree in microelectronics from Tsinghua University. He focuses on the field of semiconductor integrated circuits and has conducted research on the entire industry chain of integrated circuit design, manufacturing, packaging and testing, equipment and materials.

Han Xiaomin reminded that China's chip industry still has many shortcomings. Even some new technologies that have a clear direction are still far behind foreign giants. For example, in heterogeneous integration, Chinese companies are still at the low end, and the mid- and high-end are still in the hands of several international giants.

In addition, there are great uncertainties in some technologies that have not yet been industrialized. Even if some technologies seem to be one year away from industrialization, they may not be able to leave the laboratory and become commercial technologies. In addition, we must also consider that China's current chip industry chain is still blocked and restricted. Han Xiaomin believes that some technologies in the post-Moore era are being explored by global manufacturers, and have even been explored for a longer time. These challenges cannot be ignored.