Japan on edge of EUV lithography chip-making revolution

The Okinawa Institute of Science and Technology (OIST) has designed a new type of extreme ultraviolet (EUV) lithography equipment that could significantly reduce the cost to produce 7nm and smaller semiconductors, and thus revolutionize the chip manufacturing supply chain.

According to reports, the EUV equipment’s optical system is greatly simplified while power consumption is reduced by a factor of ten, raising the prospect of much cheaper advanced chip-making machines.

If so, it could mark the end of ASML’s monopoly on EUV lithography, which would have serious implications for semiconductor manufacturers, investors and governments.

Recall that US sanctions forbid the sale of EUV lithography equipment to China, making it much more difficult and expensive for Chinese companies to make semiconductors at 7nm and 5nm, and impossible at the 3nm node now in production at Taiwan’s TSMC and the 2nm and smaller nodes currently under development.

AI processors, the low-power semiconductor devices used in smartphones, and the latest high-density memory chips are made using highly complex EUV lithography equipment that is expensive to maintain and consumes enormous amounts of power.

OIST Professor Tsumoru Shintake says the invention is a breakthrough technology that can nearly completely solve these problems.

To paraphrase OIST, this is how it works: In traditional optical systems such as cameras, telescopes and older lithography tools, the aperture and lenses are symmetric to the central axis – i.e., arranged in a straight line. This configuration enables high optical performance with minimal aberrations, resulting in high-quality images.

However, this is not possible with extremely short-wave EUV light, which is absorbed by most materials and cannot travel through transparent lenses. For this reason, in EUV lithography systems, the light is directed using crescent-shaped mirrors that reflect the rays in an asymmetrical zigzag pattern.

This method “sacrifices important optical properties and reduces the overall performance of the system,” according to OIST.

To eliminate this problem, Professor Shintake aligned two axis-symmetric mirrors in a straight line and used a total of only four mirrors instead of ten.

Because highly absorbent EUV light weakens by 40% with each reflection, only about 1% of the energy from the light source reaches the wafer when bounced off ten mirrors while more than 10% does when only four mirrors are used.

This makes it possible to use a smaller EUV light source with only one-tenth the power.

Over 20 years ago, Phil Ware, an American engineer working for Japan’s Canon, told a technology seminar at the Semicon West industry exhibition in San Francisco that the problem with EUV lithography was that its power consumption was measured in “HDEs – Hoover Dam Equivalents.”

If Professor Shintake’s design works as intended, this problem may finally be solved. “Like the egg of Columbus, [it] may seem impossible at first glance, but once solved, it becomes very simple,” Shintake said about the problem of EUV power consumption.

As for the projector that transfers the circuit pattern on the photomask to the silicon wafer, OIST’s design consists of only two reflective mirrors, like an astronomical telescope.

“This configuration is unimaginably simple,” Shintake says, “given that conventional projectors require at least six reflective mirrors. This was made possible by carefully rethinking the aberration correction theory of optics.”

Furthermore, “the performance has been verified using optical simulation software and it is guaranteed to be sufficient for the production of advanced semiconductors.”

OIST has filed a patent application for the technology, which it first plans to demonstrate with a half-scale model. After proving the concept, it will be used to build a working EUV lithography system in cooperation with one or more Japanese corporate partners in 2026.

If that all goes to plan, Japan’s global position in the geopolitically important semiconductor industry will be greatly enhanced.

The most likely partner would be Nikon, which gave up on EUV lithography some 15 years ago due to technical difficulties and high costs but which still makes previous-generation deep ultraviolet (DUV) lithography systems.

Canon is also a potential partner but it is busy trying to commercialize nanoimprint lithography, a completely different technology that uses circuit pattern molds instead of optics.

Follow this writer on X: @ScottFo83517667