Alternative technologies and news to replace EUV have emerged.

Intel, Samsung Foundry, and TSMC use ASML's EUV lithography machines, which can "print" integrated circuits with a resolution of 13nm, building chips on their latest manufacturing nodes. However, the laser-produced plasma (LPP) EUV light source (i.e., the CO2 laser applied to small tin droplets) is not the only way to generate 13.5nm EUV radiation to "print" chips. According to Spectrum.IEEE.org, Japanese researchers are exploring the use of free-electron lasers (FELL) from particle accelerators to manufacture chips with cutting-edge feature sizes.

The High Energy Accelerator Research Organization (KEK) in Tsukuba, Japan, is exploring the use of free-electron lasers (FEL) generated by energy recovery linear (ERL) accelerators for chip manufacturing. They say that energy recovery linear accelerators can economically and efficiently produce tens of kilowatts of EUV power while powering multiple lithography machines. In contrast, ASML has produced a 500W EUV light source for its Twinscan NXE:5800E and is considering eventually increasing the power of its EUV light source to 1000W.

The ERL works completely differently from EUV lithography tools. First, an electron gun injects electrons into a tube cooled at low temperature, and superconducting radio frequency cavities accelerate them. Electrons pass through an undulator, emitting light, which is amplified through a process called self-amplified spontaneous emission (SASE). After emitting light, the used electrons loop back through the radio frequency accelerator in the opposite phase, transferring their remaining energy to newly injected electrons, which are then processed in the beam. This energy recovery allows ERL to accelerate more electrons with the same amount of electricity, making it an efficient and cost-effective method for generating ultra-high power EUV light.

"The extremely high power of the FEL beam, its narrow spectral width, and other characteristics make it suitable for applications in future lithography," KEK Senior Light Source Researcher Norio Nakamura told Spectrum.IEEE.org in an interview.

In 2021, before the onset of severe global inflation, the KEK team estimated the construction cost of the new ERL system at $260 million, $50 to 60 million higher than the price of the Twinscan NXE:3800E—although the latter is a fully integrated tool, while the former is essentially just a light source. The system will provide 10 kW of EUV power, powering multiple lithography machines (KEK did not specify how many), and its annual operating cost is expected to be about $25.675 million.

"Compared to the estimated cost of today's laser-produced plasma sources, the estimated cost per exposure tool in our setup is still relatively low," Nakamura told Spectrum.IEEE.org.There are some minor issues with FEL and ERL. First, like any particle accelerator, energy recovery linear accelerators are huge. Second, a set of extremely complex mirrors is required to guide 10kW of EUV radiation to multiple lithography tools without significant power loss, and this set of mirrors has not yet been invented (the cost is still unknown).

Finally, assuming there is a set of complex mirrors that can guide 10kW of EUV radiation to 10 lithography tools, there are no photoresists and thin films compatible with such a powerful light source. However, this will not be a problem, because the current experimental ERL can produce 20-micron infrared light, which is far from 13.5nm EUV light.

Nakamura acknowledges that many technical challenges must be addressed before ERL-based lithography tools can achieve the high performance and operational stability required for commercial chip manufacturing.

Three Alternatives to EUV Lithography

In the current chip manufacturing process, lithography is one of the essential and important steps.

Lithography requires the use of lithography equipment, and when manufacturing 7nm and below chips, EUV lithography equipment is needed. EUV lithography equipment is very complex, and only one company, ASML, can produce it globally.

At the same time, the cost of EUV lithography equipment is very high, with a price of over 1 billion yuan for one EUV lithography machine, and the latest 0.55 numerical aperture EUV lithography machine costs more than 300 million US dollars (2 billion yuan).

This is an unbearable cost for ordinary wafer companies, so the industry has been exploring alternatives to EUV lithography equipment.

There are currently three confirmed alternatives to EUV lithography equipment, which are 1. Nanoimprint Lithography (NIL); 2. Direct Self-Assembly (DSA) Lithography; 3. Electron Beam Lithography (EBL).Electron Beam Lithography (EBL) machines are likely familiar to many. The American company Zyvex has developed an EBL machine capable of creating features as small as 768 picometers (0.768 nm). However, this lithography method is slow and not suitable for large-scale chip manufacturing.

Nanoimprint lithography technology is currently being heavily invested in by Canon. Canon is already able to provide nanoimprint lithography (NIL) machines. For example, Toshiba has been interested in using NIL machines to produce NAND flash memory. In terms of NIL lithography patents, Canon is far ahead.

Of course, NIL lithography machines cannot currently compare with EUV (Extreme Ultraviolet) lithography machines in terms of precision, but the future is uncertain.

Directed self-assembly (DSA) lithography currently does not have a specific manufacturer investing heavily in it, similar to Canon's approach to NIL technology. However, as shown in the figure below, the major players have already laid out their strategies for NIL, EUV, and DSA lithography machines.

Canon is primarily focused on NIL and holds an absolute advantage in NIL patents. As for EUV, they are not very concerned, and they have essentially given up on DSA lithography.

TSMC and Samsung, the two major wafer manufacturers, have prepared for all three lithography technologies: NIL, EUV, and DSA, with EUV patents being relatively more numerous, followed by NIL, and then DSA.

ASML is, of course, focusing on EUV but also pays attention to NIL. In terms of patents, there are 345 for EUV and 187 for NIL, while there are only 16 for DSA.

Zeiss, as an important part of the EUV lithography machine supply chain, mainly focuses on EUV, with 353 patents, 38 for NIL, and 0 for DSA.