China's ambition to build a massive chip factory powered by a particle accelerator could potentially sidestep US sanctions and establish the country as a leader in the semiconductor chip industry. Researchers are utilizing innovative technology to overcome restrictions on lithography machines that are crucial for microchip production. By using particle accelerators to create a new laser source, they aim to revolutionize semiconductor fabrication.

A storage ring design for steady-state microbunching to generate EUV radiation Photo: Tsinghua University

The plan currently underway involves constructing a particle accelerator with a circumference of 100-150 meters, approximately the size of two basketball courts. The electron beam generated by the accelerator will serve as a high-quality light source for on-site chip manufacturing and scientific exploration. The team at Tsinghua University is actively coordinating with authorities in Xiongan New Area to identify a suitable location for this groundbreaking project.

Unlike commercial players such as Advanced Semiconductor Materials Lithography (ASML), which focuses on shrinking chip-making machines for export, the Chinese project seeks to localize manufacturing by building a colossal factory housing multiple lithography machines around a single accelerator. This unique approach has the potential to facilitate high-volume, cost-effective chip production and position China as an industrial leader in advanced chip manufacturing, including the production of 2nm chips and beyond.

Lithography systems are among the most advanced pieces of machinery ever created. Currently, extreme ultraviolet (EUV) technology with ultra-short wavelength is widely used for chip production with 7nm nodes and below. ASML holds the exclusive technology rights in this field, making it the dominant player in the market. As of the end of 2022, ASML had delivered 180 EUV systems, with plans to ship an additional 60 units this year.

While numerous researchers are pursuing similar technology, the Chinese scientists have taken a different path. This project has been in progress since 2017 but recently gained attention due to Huawei's advancements in chip manufacturing.

The research conducted by Professor Tang Chuanxiang's team at Tsinghua University is based on a new luminescence mechanism called steady-state microbunching (SSMB). SSMB theory, proposed by Professor Zhao Wu and his student Daniel Ratner from Stanford University in 2010, utilizes the energy released by accelerated charged particles to generate a narrow bandwidth, low angle scattering, and continuous extreme ultraviolet (EUV) light.

Charged particles emit light during acceleration, and accelerators that leverage this phenomenon produce some of the brightest artificial light sources available. The main challenge lies in guiding the distribution of electrons within the accelerator's storage ring to achieve collective synchronous radiation. The SSMB source has the potential to produce high-quality radiation ranging from terahertz waves with a wavelength of 0.3mm to EUV waves with wavelengths of 13.5nm. Unlike free-electron lasers that produce pulsed lasers with high peak power, SSMB sources generate continuous light with high average power.

Compared to ASML's EUV technology, SSMB offers a more ideal light source with higher average power and increased chip production output at a lower unit cost. ASML's EUV source is created by projecting strong laser pulses onto liquid microdroplets of tin, resulting in EUV pulse light after complex filtering and focusing. However, the EUV beam loses energy with each reflection from the 11 mirrors it encounters, leading to a power reduction to less than 5W at the wafer. This limitation becomes more pronounced as manufacturing shifts to 3nm or 2nm nodes. SSMB technology overcomes these challenges by achieving a higher output power of 1000W and requiring fewer reflecting mirrors due to its narrow bandwidth, resulting in higher terminal power.

The team at Tsinghua University conducted the initial verification phase of SSMB technology at the Metrological Light Source (MLS) in Berlin, Germany, in 2019. The experiment was successful, and they published a paper on the phenomenon in the peer-reviewed journal Nature in 2021. In 2022, the team designed another prototype at Tsinghua University, and the SSMB-EUV project was presented by Professor Pan Zhilong at an academic workshop in January. The team has visited Xiongan to select a construction site and discuss the project's plan.

While the establishment of a chip factory involves securing funding and addressing engineering details, it is crucial to cultivate new ideas that can bring about new technical approaches. Tang emphasizes the need for further effort from the team and the industry to advance SSMB technology. The experimental verification of the technology has been carried out, but the next step is to build a functional SSMB light source operating in the EUV band. This will enable the cultivation of scientific and industrial users and the refinement of SSMB technology. Tang believes that SSMB-based EUV light sources offer an alternative to sanctioned technology, potentially helping China overcome future sanctions. However, he acknowledges that independent development of EUV lithography machines still requires significant progress.

In addition to its potential impact on chip production, the realization of SSMB-EUV light sources could offer new tools for frontier research in areas such as materials science, basic physics, and biochemistry.