EUV lithography has been at the core of winning the chip war. This technology is indispensable for producing high-end microchips in sharpening the edge of civilian products and military weapons. Unlike nuclear bombs, EUV lithography is benign; but it’s a vital tool for keeping opponents behind, both on civilian and military fronts. Hence, both the USA and China have been desperate to master it. Furthermore, the USA has deployed regulatory measures to ensure that China cannot access it—causing a chip war.
As we succeed to keep increasing microchip density, the edge of Innovation continues sharpening. Consequentially, missiles succeed in tracking and hitting targets with increasing accuracy, and smartphones have become more powerful and energy efficient. Hence, finding profitable means to keep increasing chip density has been at the core of gaining an innovation edge. The core capability of increasing chip density is to make the lithography tool keep reducing the dimension of transistors. Consequently, the microchip performance increases, and the cost per Transistor falls. The race for profitable exploitation of this potential gave birth to Moore’s law in 1965. Since then, lithography tool makers have been competing intensely to find ways to enhance resolution. Hence, keeping Moore’s alive has been the underlying driving force for lithography innovation—reaching EUV exploitation.
But the journey of turning the faint beginning of EUV into a vital competitive tool has been long and, most importantly, risky. The 20+ years-long journey of active persuasion, costing more than $25 billion in public and private funds, had always been fraught with uncertainty—would it succeed to be a better alternative? Even two years before delivering the first commercial machine in 2019, experts raised doubts. Hence, this article reviews the necessity, pervasive risk, and management approach to turn EUV lithography possibility into a vital business innovation.
The urgency of EUV lithography—profiting from keeping Moore’s law alive
The transistor dimension, determining microchip density, largely depends on the wavelength of the light source used in semiconductor lithography. In the 1970s, the semiconductor industry investigated X-ray lithography for the shortest wavelength. But X-ray lithography, using a giant synchrotron source, was too expensive and ultimately failed in the 1980s. Hence, investigations on several wavelengths from 4nm to 40nm started. Consequentially, the industry found 13.5 nm wavelength Extreme ultraviolet (EUV) or soft X-ray as the next sweet spot. The concept originated in Japan in the mid-1980s when Hiroo Kinoshita projected the first EUV image. It was based on research on multilayer mirrors completed in Russia in the 70s.
Due to the belief that optical lithography would hit the wall at 65nm or 45nm, in the 1990s, EUV became a candidate for the next-generation lithography (NGL) technology. Hence, Laboratories in the United States and the Netherlands soon began to explore this potential in developing NGL.
Can you do anything useful with EUV?
By the way, EUV was not the only candidate technology researchers are exploring for keeping Moore’s alive. Among other viable options, Electron beam and ion beam lithography kept surfacing. However, Dutch company ASML made an “educated bet” on EUV lithography as the technology core for NGL.
By the way, unlike ASML, research groups in Japan and the USA were not attracted. They kept raising doubt and shying away. For example, at an SPIE conference in 2020, Hiroo Kinoshita, a researcher at NTT in the 1980s, described his frustration with convincing his fellow scientists that EUV lithography had a chance. Like him, Andrew Hawryluk, semiconductor industry veteran and then a Lawrence Livermore National Laboratory researcher, recalled similar frustrations. He mentioned a professor’s remark in 1987 on knowing groundbreaking research in EUV: “But can you really do anything useful with these things?”
How EUV lithography works, and why is it so difficult?
Like optical lithography, EUV lithography also produces light of the target wavelength and guides it through masks to project the chip design patterns on photoresist-coated wafers. But the challenge is lies in generating enough intensity light and collecting and guiding it.
First, a power source converts plasma into light at 13.5nm wavelengths. Ejected tin droplets need to be hit by high lasers in two stages to turn them pancake-shaped first, followed by conversion into plasma. The challenge is to increase the droplet hitting rate with precision to produce enough intensity. A key performance factor is how much of the laser power gets turned into EUV light, referred to as conversion efficiency (CE). Hence, a computer vision system should keep tracking and guiding laser beams to hit them at the right altitude.
Producing enough intensity light for meeting industry throughput requirements, like patterning 125 wafers per hour, faced an extreme challenge. The next one is about the reality that everything, including the air and lenses, absorbs EUV. Hence, the light source had to operate in a vacuumed chamber; novel optics should collect and guide the light to pass through the mask. For collecting EUV, a multilayer mirror became the target. Once the EUV light is generated, the photons hit a multi-layer collector mirror. Upon bouncing off the collector, the EUV travels through an intermediate focus unit into the scanner. But tin splatters, and the material accumulates on the collector, negatively affecting CE. Hence, the challenge of the time-consuming and expensive process of collector replacement surfaced.
Managing EUV technology possibility for business innovation
Experimenting with EUV and demonstrating the concept of high-resolution patterning is an issue of exploiting science to create technology possibilities. It must meet some performance requirements to turn it into a business innovation. The first one is the power of the EUV light source. Hence, improving CE became a challenge. The next one is the efficiency of reflecting mirrors. In 2016, each multi-layer mirror reflected about 70% of the light, resulting in a transmission rate of only 4% of the EUV lithography scanner. Among others, the sensitivity of the photoresist to EUV and the defectivity of masks were notable.
For example, although the minimum requirement of the light source power was 250w to succeed as a business innovation, in 2014, ASML’s scanner was still stuck somewhere between 10W to 15W. By this time already, $15-$20 billion was spent. Hence, industry experts kept raising the question of whether, at last, EUV would be able to deliver the solution that remained to be seen.
The availability issue was also hindering the growth of EUV lithography to succeed as a business innovation. For example, in 2016, EUV tool availability was 70% to 80%, below the industry’s target of 90% or higher. The next challenge was sensitivity to photoresists. In 2016, EUV resists had sensitivities around 31 millijoules per cm², which was far below than minimum target of 20 millijoules per centimeter square for getting close to cost parity with immersion triple patterning. Furthermore, EUV arises the issue of a stochastic phenomenon affecting defectivity. It causes line-edge and contact-hole roughness caused by photon shot noise and other sources. Hence, EUV lithography faced a Line-edge roughness (LER) problem, causing linewidth variation.
EUV lithography technology risk—from generation, projection to photoresist
As mentioned above, EUV lithography faced multiple limitations in meeting performance requirements to succeed as a business innovation. And nobody knew whether it was feasible to address them. Hence, experts kept raising doubt all the way. Although NGL was supposed to disrupt the landscape, many of the predictions kept proving to be wrong. As EUV lithography as NGL was not ready as late as 2016, conventional optical lithography kept remaining the workhorse technology in the fab.
Due to the uncertainty of NGL, the industry kept developing techniques to push the limit of 193nm technology. One of the notable developments was the immersion-based approach and resolution enhancement techniques (RETs). As single exposure sometimes doesn’t provide enough resolution for the critical layers, Chipmakers pursued multiple patterning plus a simple two-step process. Furthermore, the progress of competing technologies like nanoimprint lithography followed by Canon and Toshiba also kept casting a shadow on the future of EUV lithography.
Besides, it was beyond the scope of a single company to address all of the challenges. Hence, ASML took the lead in sourcing technology capabilities from an array of providers, forming partnerships to refine them, and finally integrating them precisely.
Public-private cooperation for technology development, transfer, and R&D financing
By 2014, industry and government agencies had already spent some $15-$25 billion on EUV development. One of the notable investments made by the US government was DOE’s funding to Lawrence Livermore and Sandia Labs’ national program for EUV lithography research. Later on, it got the name of “Virtual National Laboratory.” The leadership of this laboratory was also instrumental in recruiting companies, including Intel, to adopt EUV. To reduce the risk associated with the industrialization of EUV lithography, participating U.S. chipmakers formed “EUV LLC” to collaborate with the Virtual National Laboratory. As reported by Michael Borrus in 1998, to accelerate the development and commercialization of EUV technology, a collaboration of the National Labs with Intel, Motorola and AMD formed the EUV LLC project. It targeted to enhance and license technologies in light-source, optics, thin-film coatings, and metrology.
Among other major EUV R&D projects, Japan’s initiatives are notable. In the late 1990s, the University of Hyogo, in partnership with Hitachi and Nikon, worked on imaging EUV optics Design and fabrication. For studying optics technology, mask process, and wafer resists process, the collaborative national project ASET EUVL started in October 1998. In Europe, in 2006, France, Germany, Italy, and The Netherlands launched EUV advanced generation lithography in Europe (EAGLE) project. This project focused on advancing the results of earlier MEDEA+ and EU projects that had involved leading European companies, research institutes, and universities. Among notable European developments, Carl Zeiss shipped EUV optical system to ASML in 2009. It was the outcome of 15 years long research. German and European public institutions funded the development with more than 20 million euros (about $29 million).
Among the academics, EUV optics professor Fred Bijkerk projected the first EUV image in the Netherlands in 1990. American SUNY and Belgian IMEC played a vital role in testing and fine-tuning the prototype EUV lithography tool being developed by ASML. For example, in the spring of 2008, SUNY produced the world’s first full-field EUV test chip using its demo tool.
Partnership with component suppliers for sourcing specialization
Upon taking the leadership role in exploiting the commercial prospect of EUV lithography, to complement in-house capability, ASML started sourcing competence from outside. Its extended supply chain consists of as high as 5,000 suppliers. For example, Carl Zeiss, a German optics firm, has been relentlessly working on polishing mirrors to reflect EUV and fine-tuning other optical pieces. A Dutch company VDL makes robotic arms that feed wafers into the machine. And American Cymer is engrossed in developing the light source.
To access a specialized lens technology, as early as 2000, ASML acquired Silicon Valley Group Inc (SVG). To cement partnerships to keep improving critical building blocks, ASML has taken the approach of buying stakes. Notable ones include the acquisition of American Cymer (2012, for $3.7 billion) for sourcing lithography light source specialization. To access optics for collecting and guiding EUV light, ASML formed a long-term partnership with CARL Zeiss, culminating in acquiring Zeiss’s 24.9% stake in 2016(for 1 billion euros). To advance optical fabrication and high-precision manufacturing further, ASML acquired Berlin Glass in 2020. Partnership with Berlin glass is critical for wafer tables and clamps, reticle chucks, and mirror blocks.
Pattern verification is a critical task in advanced lithography. Hence, in 2016, AML acquired Taiwanese Hermes Microvision Inc. (HMI), a supplier of pattern verification systems for advanced semiconductor devices (for $3.1 billion). For enhancing Verification, Reticle Enhancement Technology (RET), and Optical Proximity Correction (OPC), ASML acquired Brion Technologies in 2006 (for $270 million).
Making lead users development partners for ensuring financing and sharing risk
To address the increasing complexity of EUV technology and financial hurdles to turn possibility into business innovation, ASML focused on risk sharing among customers and suppliers. Hence, to mobilize finance and ensure demand for its EUV lithography solution, in 2012, ASML management announced a customer co-investment program. It was a strategy of engaging lead users like Intel, TSMC, and Samsung through minority equity investments in ASML.
Responding to a customer co-investment program, Intel committed to funding €829 million of R&D activities over five years. In July 2012, Intel Corp announced its intention to take a 15 percent stake in ASML for around $3.07 billion. Soon after, Samsung followed Intel. Samsung announced, in August 2012, to contribute EUR 276 million to ASML’s research and development of next-generation lithography technologies over five years. The announcement also included Samsung’s decision to invest EUR 503 million in a 3% ASML equity stake.
Subsequently, in November 2012, TSMC got into an agreement to acquire five percent of the equity of ASML for US$1.04 billion. TSMC also committed €276 million over five years to ASML’s research and development program.
Rise of late-entrant ASML as a monopoly in the EUV lithography market
In 1984, ASML was born as a joint venture between Dutch Philips and ASMI. In a wooden hut, 31 employees of ASML started the journey. Soon after, the company got into a financial struggle, reaching the point of death in infancy. But due to strong management, science, and technology capability, ASML reported a 5.58 billion Euro net profit from 18.61 billion Euro revenue in 2021. Through the process, a late entrant, suffering from weak health in unfancy, has grown as the most valuable technology company in Europe. Besides, ASML is the sole supplier of the most sophisticated microchip machine—costing around $250 million apiece.
With the continued progress in expanding innovation capabilities, ASML’s share (in unit shipment) in the lithography market segment kept expanding. It grew from less than 40% in 2005 to over 60% in 2018. But for EUV lithography, ASML emerged as the sole supplier in 2019. In device count, although only 26 of the 229 lithography machines ASML sold in 2019 are EUV machines, they made up a third of sales by revenue. As neither Canon nor Nikon pursued EUV technology, investors concluded that ASML would enjoy its nanoscopic monopoly for a while. Hence, market capitalization has kept growing since 2010. Between 2010 and 2020, its market capitalization grew tenfold to around €114bn. By Sept 2021, due to accelerated share price appreciation (reaching USD858.87 on Sept 10, 2021), ASML became the 32nd most valuable company in the world.
EUV lithography gets caught in the Chip War
By 2022, by leveraging ASML’s EUV lithography, TSMC succeeded in growing as a monopoly in high-end chip making. As high as 90% of chips processed with EUV lithography at 5nm and 4nm process nodes come from TSMC’s foundry. Consequentially, both America and China depend on TSMC’s foundries operating with ASML’s EUV machines. Hence, both find such dependence as a national security issue. Therefore, China has embarked on a semiconductor independence mission by creating a flow of more than 100 billion-dollar funds. Besides, China has been using its military and political machinery to increase its influence on Taiwan to ensure the supply of TSMC foundry service and steady transfer of EUV lithography-operated foundry expertise.
America finds growing dependence on Taiwan for high-end chips and the ambition of China an agenda deserving of state interventions. Therefore, the USA has come up with Chip acts, pouring $52 billion in public funds and $23 billion in tax incentives for facilitating investment in the foundry, R&D, and human capital. Furthermore, the USA has unleashed regulatory measures to ban the export of ASML’s EUV lithography machines to China. Consequentially, the success of EUV lithography to keep Moore’s law alive got caught in the Chip War between the USA and China.
...welcome to join us. We are on a mission to develop an enlightened community by sharing the insights of Wealth creation out of technology possibilities as reoccuring patters. If you like the article, you may encourage us by sharing it through social media to enlighten others.
Related articles:
- Chip War
- Semiconductor Lithography Economics–fuelling Moore’s law and market power
- Microchips–invention, evolution, transformation and chip war
- Moore’s Law–fuelling monopolization, Reinvention, and migration
- Moore’s Law is Dead–Chiplet redefines semiconductor industry
- Semiconductor–illusive technology core changing world order
- Winning Chip War–fuelling reinvention waves for changing world order
- Chiplet Technology — a weak reinvention core?
- Semiconductor Economics–will Chiplet era slow down the growth?
- ASML–growing pearl gets caught in Chip War
- China’s Semiconductor Independence–prematurely caught?
- India’s Semiconductor Dream–pushed in the slow lane?
- Semiconductor Value Chain–globally distributed ecosystem
- Semiconductor IndustryWaves
- Intel Falling Due to PC and Mobile Waves
- ASML Lithography Monopoly from Sustaining Innovation
- Taiwan’s Semiconductor Monopoly – How did it arise?
- ASML TSMC Nexus Fuels Semiconductor Monopoly
- ASML Monopoly in Semiconductor — where is magic?
- SEMICONDUCTOR MONOPOLY DUE TO WINNING RACE OF IDEAS
- Semiconductor Industry Growth–personalities, new waves, and specialization underpin
- Transistor–technology core shaping global trade and power
- Loss of America’s Inventions–blame semiconductor economics?
- Microprocessor Evolution Fueling Disruptive Innovations
