At the 2026 SPIE Advanced Lithography and Patterning Conference, the future development of advanced lithography technology will be one of the hot topics of discussion. Participants generally agreed that existing technologies will not be able to meet the needs in ten years, so what is the future direction of development? 

One approach is to increase the numerical aperture (NA) while shortening the wavelength.

In a presentation Tuesday afternoon titled "Ultra-High Resolution: A Necessary Step to Further Enhance Lithography Resolution," Jos Benschop, Executive Vice President of Technology at lithography equipment manufacturer ASML, explained the necessity of taking this first step. He pointed out that changes in wavelength require corresponding adjustments to the light source, optics, masks, photoresist, and other aspects of the lithography process.

In the past, the industry has continuously pushed the limits of numerical aperture. "We expect this to happen again," Benshop said when discussing extreme ultraviolet (EUV) lithography.

Extreme ultraviolet lithography

The most advanced technology currently available—Extreme Ultraviolet (EUV) lithography—uses light with a wavelength of 13.5 nanometers to create feature patterns as small as 8 nanometers. The minimum feature size is determined by resolution, which is proportional to the wavelength divided by the numerical aperture. ASML's latest equipment has a numerical aperture of 0.55 and is known as high numerical aperture EUV. 

In his presentation, Benschop discussed the next generation of equipment under development. These will have a numerical aperture of 0.75, reducing the minimum printable size by 36% to approximately 5 nanometers, all other things being equal. Even higher numerical apertures are called ultra-high numerical aperture (hyper NA). In contrast, past lithography technologies have all eventually achieved numerical apertures greater than 1.0, although the initial numerical apertures were less than 1.0.

According to Benschop, ASML has already begun advancing this technology. For example, optics manufacturer Zeiss has designed a lens assembly that is only slightly larger than the lens used in a 0.55 high numerical aperture (NA) tool. ASML plans to manufacture lithography tools capable of accommodating ultra-high NA or high NA lens assemblies. Therefore, ultra-high NA lenses can directly replace existing lenses.

Said by Benschop

Benschop acknowledged that a major challenge of increasing chip size in North America is the reduced depth of field. This means that images on the chip are more prone to blurring, impacting chip performance and yield. Benschop stated that the solution lies in improving focus control. This requires improvements to sensors and other scanners, or the use of flatter wafers.

Benschop predicts that the problems with ultra-high-throughput nucleic acid detection technology will be resolved. "When ultra-high-throughput nucleic acid detection technology is needed, it will be readily available," he said.

IBM's Lithography Roadmap

In a presentation titled "IBM Lithography Roadmap: The Needs for Future Lithography Tools and Masks and the Requirements for Photoresists to Avoid Random Defects," delivered Tuesday by Allen Gabor, Principal Graphics Engineer at IBM Semiconductors, an alternative was explored. He looked to the next wavelength, focusing on 3.1 nanometers. 

Using a wavelength four times shorter than current wavelengths offers several advantages. First, the numerical aperture required to achieve a specific resolution is smaller. Therefore, the depth of field is greater than when using longer wavelengths.

Gabor points out that importantly, the new wavelength allows for adjustments to optical designs and brings other improvements, reducing line edge roughness by 20% compared to using 13.5-nanometer wavelength light. Reduced line edge roughness brings the printed lines on the wafer closer to their ideal state. This is crucial because each layer of pattern on the wafer must be aligned with the pattern below it. If the alignment is too large, or if the lines shift at specific locations, the holes etched through the thin film may deviate from the intended structure, leading to short circuits or open circuits. Edge misalignment can also cause device failure in other ways.

Gabor states that minimizing edge positioning errors is a key factor in determining the wavelength. In fact, reducing edge positioning errors while continuously shrinking feature sizes is one of the main goals of IBM's lithography technology roadmap for the next 15 years.

Gabor notes that there is still much work to be done to improve the lithography infrastructure. Furthermore, several issues must be addressed. For example, current mirrors capable of reflecting 3.1-nanometer wavelength light have a reflectivity of approximately 35% to 40%. 

It is unclear how many mirrors a lithography device might need to focus or guide light. During the Q&A session following his presentation, Gabor stated that the number of mirrors ranges from 2 to 10. If a lithography machine uses 5 mirrors with a reflectivity of 35%, then 99.5% of the light emitted from the light source will be lost.

The solution is to increase the number of reflectors and decrease the number of mirrors. This problem, along with several others, must be resolved before the technology can be truly put into use.

Source: Compiled from SPIE

View more at EASELINK