Even if we give the complete set of drawings, Chinese people still cannot build high-end lithography machines?
[Copy link]
As a number of chip companies rush to list on the Science and Technology Innovation Board, the basic tools for chip production - domestic lithography machines - have also been successfully pushed to the hot search list in the technology circle.
First, the “Chinese Academy of Sciences releases 5nm laser lithography technology” was misinterpreted as “you don’t need to buy the EUV high-end lithography machine from ASML, the Dutch lithography machine giant, to produce chips”; then the “post-95 college student UP host made a lithography machine” was quickly debunked; then there was the rumor that “20,000 people were in seclusion for two years, and Huawei developed its own lithography machine”…
Public sentiment is like the midsummer sun, hot and passionate. Any disturbance related to domestic lithography machines is quickly magnified and fermented.
But if we can calm down and think about it, we will find that the people's desire for advanced lithography machines comes from the hope that domestic chips will no longer be "choked". However, this mentality towards domestic lithography machines is always accompanied by a sense of anxiety, restlessness and even impetuousness.
Obviously, this mentality is not desirable when pursuing the goal of catching up with advanced countries.
This article is reprinted from the WeChat public account "Economic National Weekly" (ID: ENNWEEKLY). The original article was first published on July 29, 2020, with the original title "Give us a full set of drawings, and the Chinese can't build high-end lithography machines? !"
1. The crown jewel of the semiconductor industry
What is a lithography machine?
Simply put, a photolithography machine is a machine that uses light as a blade to engrave chip drawings on a wafer. It is the most basic and critical link in chip production and processing.
The entire photolithography process can be understood as printing the designed chip pattern on the mask, and then using the laser to pass through the mask and objective lens to expose the chip pattern on the photoresist coating. This is somewhat similar to the exposure of camera film, but more precise.
If you ask industry insiders to define a photolithography machine, it is probably best described as “the jewel in the crown of the semiconductor industry.” A photolithography machine is made up of tens of thousands of parts and is a product that integrates top technologies in mathematics, optics, fluid mechanics, polymer physics and chemistry, surface physics and chemistry, precision instruments, machinery, automation, software, and image recognition.
On a cutting-edge lithography machine, you can see the gathering of top technologies from all over the world: Germany's Zeiss lens, Japan's special composite materials, Sweden's industrial precision machine tools, the United States' advanced control software and power supply, etc.
The "origin" of the lithography machine can be traced back to more than 70 years ago.
In 1947, Bell Labs invented the first point-contact transistor, and photolithography technology began to develop. In 1959, the world's first transistor computer was born, photolithography technology was proposed, and Fairchild Semiconductor developed the world's first single-structure silicon chip.
Although photolithography machines are so advanced now, before 1970, photolithography was not considered advanced high-tech, and most semiconductor companies could design their own tools and tools. Even Japan's Nikon and Canon entered this field in the late 1960s. After all, photolithography machines at that time were no more complicated than cameras.
However, in the following 10 years, lithography machines gradually emerged from the subsidiary departments of semiconductor companies and formed a real market. In particular, in 1978, the American GCA company launched the truly modern automated stepper lithography machine, which officially opened the era of the great development of lithography machines with many competitors.
Riding on the momentum of domestic large companies and chaebols entering the semiconductor industry, Nikon, Canon and other companies began to compete with American companies such as GCA, Ultratech, and Eaton. By 1984, Nikon and GCA each had 30% of the market, Ultratech had about 10%, and the remaining companies such as Eaton, Canon, and Hitachi each had less than 5% of the market share.
Since then, due to the large-scale overcapacity of dynamic memory chips in Japan, the price has fallen by nearly 80%, and the US semiconductor industry has suffered a huge blow. A number of lithography machine manufacturers such as GCA and Ultratech have encountered serious financial crises and have been marginalized ever since.
On the contrary, ASML, a small company far away in Europe, did not suffer much loss due to the lack of attention from its investor Philips and the small size of the company, so it continued to develop new products as planned.
Later, the United States began to counter the Japanese semiconductor industry, which had been prosperous for more than a decade, and when it was studying the use of EUV (extreme ultraviolet) to solve the 193nm wavelength problem that DUV (deep ultraviolet) could not cross, it blocked Nikon from entering the "EUV LLC (Extreme Ultraviolet Limited Liability Company)" organization, which brought together major technology giants. ASML successfully "picked up a leak" by promising to set up a research center, build a factory, and purchase 55% of raw materials in the United States. It experienced the whole process of the feasibility of EUV lithography machines being fully verified by the "big guys", laying a key foundation for its future "ascension to the gods".
ASML EUV lithography machine, each system costs hundreds of millions of euros
Today, looking at the lithography machine orders of global chip manufacturers, ASML, which was once the "little brother", has become the "big brother" in the industry. Taking 2019 as an example, ASML shipped a total of 229 lithography machines, with net sales of 11.82 billion euros and net profit of 2.52 billion euros; Nikon shipped 46 units and Canon shipped 84 units.
Moreover, among the 229 lithography machines shipped by ASML, 26 are EUV lithography machines, which represent the highest level in the industry today. EUV lithography machines use a 13.5nm wavelength light source and are an indispensable tool for breaking through the 10nm chip process node. Without it, TSMC, Samsung, and Intel's 5nm chip production lines cannot be put into production.
Some people say that if the lithography machine is the jewel in the crown of the human semiconductor industry, then ASML is the person holding the brightest jewel.
2. “Even if I give you a complete set of drawings, you still can’t make it.”
The story of the rise of China's lithography machines goes back 20 years.
On February 27, 2001, Wang Yangyuan, academician of the Chinese Academy of Sciences and director of the Institute of Microelectronics at Peking University, gave a report on "Microelectronics Science and Technology and Integrated Circuit Industry", analyzing the necessity, urgency, measures and suggestions for the development of my country's microelectronics and integrated circuit industries.
After listening to the report, the State Council leader who presided over the meeting immediately proposed that integrated circuits are the "heart" of electronic products and the foundation of the information industry, and must be given high attention.
Soon, photolithography machines became the primary research project in microelectronics equipment during the "15th Five-Year Plan"; the following year, photolithography machines were included in the National 863 Major Scientific and Technological Research Program. Under the joint promotion of the Ministry of Science and Technology and the Shanghai Municipal Government, Shanghai Microelectronics Equipment Co., Ltd. (hereinafter referred to as "Shanghai Microelectronics"), which undertook the task of the key project, was officially established.
Soon after the company was founded, the ambitious general manager He Rongming took his technical team to Europe, the United States and other places to develop technical cooperation. But when He Rongming proposed that "Chinese people also want to make lithography machines", the foreign experts were surprised and more or less disdained and looked down on Chinese scientific and technological personnel. One senior European expert even told him: "Even if I give you a full set of drawings, you still can't make it."
This remark irritated He Rongming's self-esteem, and he was very unconvinced. Even after more than a decade, He Rongming recalled that time and admitted that for a long time after returning from abroad, he led the team to fight with a grudge.
But the words of foreign experts are not entirely false.
A photolithography machine is made up of tens of thousands of parts, and the requirements for technology, precision, and speed are so high that it is unimaginable. Temperature, humidity, light, etc. will affect the final success or failure. Even He Rongming had to admit: "It is equivalent to two large airplanes moving side by side from takeoff to landing. A knife extends from one plane and carves words on the rice grains of another plane. There can be no mistakes."
Such resentment was not dissipated until many years later when Shanghai Microelectronics designed and delivered the first domestically produced lithography machine in the field of advanced packaging to a Taiwanese customer. The customer told He Rongming: "We really never thought that the Chinese could make lithography machines. As a Chinese, I am proud of this."
Perhaps it was the affirmation from the customers that finally allowed He Rongming to find an outlet for his emotions. At a later company cadre meeting, He Rongming finally expressed a sentence that had been suppressed in his heart for a long time: "If the existence and efforts of our group of people can gain the respect of our peers, then everything we do will be worth it!"
Today, Shanghai Micro Electronics has become the most advanced lithography equipment manufacturer in China, with a domestic market share of 80% and a global market share of 40% for packaging lithography products. Its LED/MEMS/power device lithography machines have leading performance indicators, and it ranks first in the LED lithography machine market share.
However, it should be noted that lithography machines are divided into front-end lithography machines and back-end lithography machines. Front-end lithography machines are used for chip manufacturing, while back-end lithography machines are mainly used for chip packaging. ASML's lithography machine is a front-end lithography machine.
Currently, the global front-end lithography equipment is completely monopolized by ASML, Nikon, and Canon, and the most advanced front-end lithography equipment that Shanghai Microelectronics can mass produce is at the 90nm process node. Even though Shanghai Microelectronics disclosed that it will deliver a 28nm process front-end lithography equipment in 2021, it is still far behind ASML's 5nm process lithography equipment.
What consequences will this gap lead to?
To give a typical example, as soon as Shanghai Microelectronics announced a breakthrough in its 28nm lithography machine, ASML immediately cut prices and sold off the same and previous generation lithography machines in an effort to seize the Chinese market and squeeze Shanghai Microelectronics' market space.
Some people may say that EUV lithography machines are not necessary at present. The case often used to support this view is that SMIC's N+1 and N+2 generation processes will not use EUV processes, and TSMC only began to introduce EUV in the third-generation 7nm process.
But in fact, the smaller the process, the more advanced the technology, the smaller the line width in the integrated circuit, the more advanced the lithography machine is needed. "Although EUV technology is not necessary for the 7nm process, the injection of EUV technology can improve the yield rate and has a good effect," said Zhang Guobin, CEO of Electronic Innovation Network.
A widely circulated saying in the industry is that even if all imported lithography machines stopped working instantly, China would still have chips available, but the chips produced by 90nm lithography machines are only equivalent to the level of the Pentium 4 processor in 2004.
In other words, after more than a decade of hard work, although domestic lithography machines have made considerable progress, they are still a long way behind the leading level and have never been able to open the door to the high-end market. If you can't enter the high-end field, you will have to bear the slap in the face from others.
3. Why are high-end lithography machines difficult to buy and manufacture?
There is an old Chinese saying: If you want to do your work well, you must first sharpen your tools.
Why is it that the chip manufacturing "weapon" we want is still difficult to buy and manufacture after so many years? Was China's foundation so poor from the beginning?
Here we have to mention a well-known historical factor in the semiconductor industry - the "Wassenaar Arrangement". Western countries generally follow the "N-2" principle for the approval of China's semiconductor technology and products, which means two generations behind the most advanced technology. If there is a proper delay in the approval process, the technology and equipment that China can obtain will basically be three generations or more behind the most advanced in developed countries.
For example, Intel, Samsung and other companies were able to buy ASML's 10nm lithography machine in 2015, while SMIC could only buy ASML's 32nm lithography machine produced in 2010. Even if SMIC could buy second-hand lithography equipment through cooperation with the Belgian Microelectronics Research Center (IMEC), it would have to wait for IMEC to use it for five years before it could get it.
Five years is enough time for the semiconductor industry to iterate several times.
In addition to being unable to buy the latest equipment, the Wassenaar Arrangement also restricts Chinese engineers from entering the core departments of well-known semiconductor companies in Europe and the United States to prevent technology leaks. However, the existing high-end lithography machines are the result of cooperation among many countries and involve multiple scientific fields. It can be said that the difficulty of China's semiconductor industry, which is under all-round suppression from production equipment to technical talents, to build high-end lithography machines is no less difficult than building an atomic bomb back then.
In fact, China’s lithography technology was not lagging behind others from the beginning.
In 1952, soon after the founding of the People's Republic of China, the computer industry was launched and an electronic computer research group was established, led by Hua Luogeng, director of the Institute of Mathematics. In the following decade, China successfully trial-produced the first crystal triode, the first germanium transistor, the first generation of silicon planar transistors, and the first integrated circuit.
Since my country's first integrated circuit was born in 1965, a widely accepted view in the industry is that my country began to use photolithography technology to manufacture integrated circuit chips around 1965.
In 1977, my country's first lithography machine, the GK-3 semi-automatic lithography machine, was born; in 1978, 1445 developed the GK-4 based on the GK-3, increasing the processing wafer diameter from 50 mm to 75 mm and improving the degree of automation. At this time, ASML had not yet appeared.
By 1985, the 45th Institute of the Ministry of Machinery and Electronics had developed a step-by-step projection lithography machine, which was technically appraised by the Ministry of Electronics to have reached the level of the 4800DSW lithography machine launched by the American GCA Company in 1978.
If we take a brief look back, we can summarize the development of the entire semiconductor industry in mainland China as follows: it got off to a good start in the 1950s, moved forward all the way through the 1960s and 1970s, and in the first half of the 1980s it was second only to the United States, even on par with Japan, and ahead of South Korea and Taiwan.
But the turning point happened to occur in the second half of the 1980s. At that time, Japan's semiconductor production capacity was in excess, and it was engaged in a price war with the United States. The idea of "it is better to buy than to make" began to rise in China. A large number of companies abandoned the principle of independence and self-reliance, blindly introduced and opened up, and embarked on a development path guided by "trade, industry and technology".
At that time, China's integrated circuits were also out of touch with scientific research, education, and industry. In scientific research, they followed the technical standards and technical systems formulated by foreign countries. In education, all majors related to finance and trade became popular, and software engineers became an unpopular profession. In industry, some companies were keen on assembling for foreign companies, exchanging a large amount of cheap labor for economic benefits.
How can the eggs remain intact when the nest is overturned? China’s lithography machine industry has also been affected by the overall environment and has not been able to leave a name in the field of high-end lithography machines.
It was not until the late 1990s and early 2000s that the Chinese chip industry began to see the rise of overseas returnees and private enterprises, and the domestic lithography industry reawakened. But by then, the gap between domestic lithography and the most advanced technology had stretched from 7 years to more than 20 years.
4. A glimmer of hope has emerged amid difficulties
Compared with ASML, which has the support of advanced research results in Europe and the United States and has huge R&D funds in exchange for equity investments from Intel, Samsung, TSMC and other companies, the domestic lithography machine industry, which is under the "Wassenaar Arrangement", is clearly at a real disadvantage in terms of its growth foundation.
But being a step behind others makes every breakthrough on the road to growth exciting.
In order to speed up the development of domestic lithography machines, in 2008, the state established the "Extremely Large Scale Integrated Circuit Manufacturing Equipment and Complete Process Special Project" (referred to as the 02 Special Project) to establish independent high-end lithography technology and industrial development capabilities as a major and core strategic goal.
The lithography machine industry chain can be divided into two parts. One is the core components of the lithography machine, including key subsystems such as light source, lens, dual worktable, immersion system, etc. The other is the lithography supporting facilities, including photoresist, photomask, coating and developing equipment, etc.
Therefore, Project 02 also organized multiple departments to participate, each responsible for different sub-items, and has gradually achieved gratifying results:
On June 21, 2017, the "key technology of extreme ultraviolet lithography" developed by the Changchun Institute of Optics, Fine Mechanics and Physics of the Chinese Academy of Sciences (now Beijing Guowang Optics) passed the acceptance;
Tsinghua University and Beijing Huazhuo Precision Technology are responsible for the dual worktable system. They completed the construction of the R&D and pilot production base in 2019 and are the second in the world to master this technology (the first is ASML's Twinscan technology);
The State Key Laboratory of Fluid Power and Electromechanical Systems of Zhejiang University and Zhejiang Qier Electromechanical are responsible for the immersion system of immersion lithography machines. Currently, their level ranks third in the world, with ASML and Nikon ranking the top two respectively.
The Institute of Optoelectronics of the Chinese Academy of Sciences is responsible for the excimer laser light source system, and Beijing Keyi Hongyuan is responsible for industrial transformation. The domestically produced 40W 4kHz ArF light source has been delivered, which is the third in the world after the American Cymer Company (acquired by ASML in 2013) and the Japanese Gigaphoton Company.
On November 29, 2018, the "super-resolution lithography equipment" developed by the Chinese Academy of Sciences passed the acceptance. The lithography resolution reaches 22nm, and combined with double exposure technology, it can be used to manufacture 10nm-level chips in the future;
The 28nm lithography machine that Shanghai Micro Electronics will deliver in 2021 will be designed and integrated by Shanghai Micro Electronics, with Beijing Keyi Hongyuan providing the light source system, Beijing Guowang Optics providing the objective lens system, Guoke Precision providing the exposure optical system, Huazhuo Jingke providing the double workbench, and Zhejiang Qier Electromechanical providing the immersion system.
"Now in our ultra-clean factory, each lithography machine covering dozens of square meters stands there like a giant of thought." He Rongming often gazes through the glass at these complex systems composed of tens of thousands of ultra-precision parts and millions of codes, and seems to be able to see the glimmer of hope flashing in the midst of difficulties.
5. Don’t learn blindly
Having come this far, what should be the next step for domestic lithography machines?
If we summarize ASML's success, it is nothing more than a few key factors, such as the special era environment, a leader with strong decisiveness, the determination to continue to increase R&D investment, and the courage to integrate industrial chain resources. However, as far as the current development of domestic lithography machines is concerned, it is not enough to learn from the experience of only one giant company.
A British economist once pointed out that long-term and strategic guidance and investment by the government and enterprises are an absolute prerequisite for breakthrough innovation. He Rongming has also repeatedly expressed his agreement with this view. In the field of lithography machines, the country's strategic investment and guidance of risk sharing between the government and enterprises have always been the biggest pillar of Chinese enterprises.
In this sense, if we want to plan the development prospects of domestic lithography machines, we might as well look at the key development nodes of global semiconductors and find referenceable experiences that are jointly guided and invested by the government and enterprises.
In the mid-1980s, the U.S. semiconductor industry experienced a decline, but then quickly recovered and achieved revival. One of the keys to turning the tide was the establishment and development of the Semiconductor Manufacturing Technology Alliance (Sematech).
The Sematech Alliance was initiated by Charlie Sporck, CEO of National Semiconductor, and brought together 14 companies that accounted for 80% of the output value of the U.S. semiconductor industry at the time, such as Intel, IBM, HP, Micron, Motorola, etc. It aimed to rebuild the international competitiveness of the U.S. semiconductor industry through research and development and industrialization of semiconductor manufacturing-related technologies in the short and medium term.
Of course, it is difficult to control the situation of many heroes gathering together. At the beginning of its establishment, Sematech was in a state of chaos due to the competition and mutual vigilance among member companies. After noticing this problem, the US government quickly sent the Defense Advanced Research Projects Agency (DARPA) to participate in the management and operation of the organization on behalf of the US government.
It is also with the joint cooperation of government and industrial forces that the US semiconductor industry chain was sorted out in the shortest time, reshaping the up-down and down-to-the-left and right relationships, while promoting cooperation among members and avoiding competition, and improving the resilience of the entire industry chain.
Can this experience be directly copied to the development of domestic lithography machines? Of course not. China is facing a different industrial situation. We can learn from it, but not blindly.
One of the characteristics that Sematech is praised for is that it is an alliance organization with corporate attributes, led by a CEO who understands technology, management, has an industry background and connections, and an entrepreneurial spirit, thus ensuring the efficient connection between production, learning, and research to the greatest extent possible.
Although China also has projects such as the 02 Special Project to sort out and shape the domestic lithography machine industry chain, the production, education, and research sectors as a whole are still in a "separate management" situation. It is necessary to establish an alliance organization like Sematech that is jointly managed and operated by the government and enterprises to unify the planning and lead the development process of the entire domestic lithography machine industry.
However, details such as the selection of specific leaders, the government's role in the alliance, the formulation of industry norms, and regulatory intervention need to be adjusted and implemented in light of China's specific situation.
Secondly, since most American companies are not very interested in basic research and applied research, the financial subsidies provided by the government to Sematech are mainly used for medium- and short-term technology development.
Considering that what enterprises are really interested in are innovative products that can bring actual profits, only when enterprises are interested can technological innovation be motivated, so Sematech's move is reasonable.
However, we must clearly realize that the reason why the US government can allow companies to be uninterested in basic research and focus on medium- and long-term technology development that can quickly generate innovative value is that the US has a solid reserve of intelligence and talent in the field of basic science. At this stage, China does not have enough confidence in this regard.
Many industry insiders believe that when encouraging the domestic lithography machine industry to increase technological research and development, the idea of absorbing multiple parties and guiding in layers may be more desirable.
For example, establish an industry alliance similar to Sematech to attract leading companies and small and medium-sized enterprises with ideas and potential to join, encourage large companies with stable revenue to continue investing in cutting-edge technology research and development, and guide small and medium-sized enterprises to carry out medium- and short-term technology development.
This can not only help small and medium-sized enterprises increase their ability to quickly generate revenue, but also enable them to perceive and be exposed to the latest technological trends and industrial signals in an alliance atmosphere, and gradually establish their own appropriate development position in the industrial chain. Thirdly, it will help accelerate the establishment, maturity and consolidation of the entire upstream and downstream of the industrial chain.
It is important to note that supporting real technology development is both the original intention of government subsidies and the bottom line of subsidies. Even for short- and medium-term technologies, there must be innovative technology development, not just supporting product production.
Because there are no products with technological development, even if the product belongs to an industry that sounds very new, it only supports the expansion of low-end production capacity and is limited to low-end duplication and oversupply.
6. Protecting the Technological Airspace
Money can buy blueprints, but not core technology; the market can bring investment, but never innovation. Technology and innovation, in essence, come from talent.
How can the water be so clear? Because it comes from a source of fresh water. In addition to retaining existing talents, it is also a commonplace to continuously cultivate new talent reserves, which is the top priority .
At the same time, we should always keep in mind that we should formulate the correct development strategy from the perspective of the global system. In the face of relatively weak industrial conditions and talent reserves, it is not advisable to advocate the idea of "complete localization" at every turn.
Not to mention, the number of lithography machine parts and the complexity of technology sources make it difficult to achieve complete localization in the short term. Even if we insist on localizing all of them to ASML's current 5nm level, ASML will be far ahead. What's more, in today's globalized world, the internationalization of the supply chain has become an unstoppable development trend.
Rather than "making everything ourselves if others don't allow us to buy it", "making others have no choice but to sell it to us" may be a more practical and effective solution. Rather than simply emphasizing "complete localization", it is better to look at the long term, increase the accumulation of patent technology, strengthen international cooperation, and make it impossible for others to avoid your patent scope.
At a work report, He Rongming made a special sharing. During the period, he mentioned that after countless technical confrontations, they found that many foreign companies applied for patent protection scopes in Chinese airspace that were very large. For this reason, Shanghai Microelectronics launched a very difficult patent airspace defense war, using its own patents to block many patents with too large patent protection scopes in China, reducing the latter's protection scope by half, thereby winning innovation space for its own patent technology layout.
Innovation is a battle, and patents are a battlefield without the smoke of gunpowder. International cooperation in the field of high-end technology is often fought for, and begging for cooperation is powerless and useless. Only by tackling technical problems and becoming a competitor, and through constant confrontation, can both parties have the intention to cooperate and have a more equal attitude.
Driven by Moore's Law, the technology of international lithography machines is developing rapidly. As domestic lithography machines strive to catch up with the future, the argument of "breaking through the technological blockade" is often advocated. This will not only easily breed blind self-confidence and trigger stronger external suppression, but will also be of no benefit to the development of my country's scientific and technological airspace.
In 2017, at the Shanghai Science and Technology Innovation Pioneer Advanced Deeds Report Conference, the gray-haired He Rongming took the stage to speak as a "Science and Technology Innovation Pioneer."
He said: "I deeply feel and understand that in the future world, the size of a country is no longer measured by the area of its territory, but by the size of its technological airspace. If our generation does not take the responsibility of protecting and expanding our country's technological airspace, who will take the responsibility?"
There is still a long way to go before the rise of domestic lithography machines.
|