KEIDS3-1702
Understanding secondary electron interactions and stochastic limits in EUV photoresists
Topic
IDS3. Photoresist: 50 Years of Innovation and the Next 50 Years (Sponsored by DONGJIN SEMICHEM)
When and Where
Sep 29, 2026
15:30 - 16:30
Room 201
Session Chairs
Hyun-Dam JEONG
Presenter(s)
Greg Denbeaux (University at Albany)
Co-Author(s)
Abstract
Extreme ultraviolet (EUV) lithography has enabled continued pattern scaling for advanced semiconductor manufacturing; however, resist performance is increasingly limited by stochastic effects and fundamental reaction-transport processes. In EUV exposures, 92 eV photons generate approximately 80 eV primary photoelectrons and cascades of low-energy secondary electrons that drive most of the chemically amplified resist reactions. Therefore, understanding the spatial distribution of electron-induced reactions and subsequent acid diffusion is critical for future high-resolution lithography.
To investigate these effects, photoresists were exposed using controlled low-energy electron sources (5 to 80 eV), allowing measurement of reaction distributions independent of optical imaging effects. Reaction blur was quantified through post-development thickness-loss measurements and ellipsometric analysis. For PMMA exposed to 80 eV electrons, the secondary-electron reaction blur (SEB) was determined to be approximately 1.8 nm, representing the depth within which 67% of reactions occur. In chemically amplified resists, acid diffusion further broadens the effective reaction volume, increasing the total blur and degrading image fidelity.
Measurements of deprotection kinetics further indicate that most post-exposure chemistry occurs within only a few seconds at typical bake temperatures, suggesting opportunities to optimize post-exposure bake conditions. Additionally, stochastic limitations become increasingly significant as feature sizes shrink. Representative EUV resist volumes contain only a few absorbed photons, acids, PAG molecules, and quencher molecules, making shot noise, molecular inhomogeneity, component segregation, and diffusion processes major contributors to line-edge roughness and defect formation.
These findings show that future EUV and High-NA EUV photoresists will require simultaneous control of secondary-electron transport, acid diffusion, quencher kinetics, and resist compositional uniformity. As feature dimensions continue to decrease, these chemically driven stochastic effects are expected to become the dominant factors governing resolution, line-edge roughness, and pattern fidelity.
To investigate these effects, photoresists were exposed using controlled low-energy electron sources (5 to 80 eV), allowing measurement of reaction distributions independent of optical imaging effects. Reaction blur was quantified through post-development thickness-loss measurements and ellipsometric analysis. For PMMA exposed to 80 eV electrons, the secondary-electron reaction blur (SEB) was determined to be approximately 1.8 nm, representing the depth within which 67% of reactions occur. In chemically amplified resists, acid diffusion further broadens the effective reaction volume, increasing the total blur and degrading image fidelity.
Measurements of deprotection kinetics further indicate that most post-exposure chemistry occurs within only a few seconds at typical bake temperatures, suggesting opportunities to optimize post-exposure bake conditions. Additionally, stochastic limitations become increasingly significant as feature sizes shrink. Representative EUV resist volumes contain only a few absorbed photons, acids, PAG molecules, and quencher molecules, making shot noise, molecular inhomogeneity, component segregation, and diffusion processes major contributors to line-edge roughness and defect formation.
These findings show that future EUV and High-NA EUV photoresists will require simultaneous control of secondary-electron transport, acid diffusion, quencher kinetics, and resist compositional uniformity. As feature dimensions continue to decrease, these chemically driven stochastic effects are expected to become the dominant factors governing resolution, line-edge roughness, and pattern fidelity.
