Bond contribution model for the prediction of glass transition temperature in polyphenol molecular glass resists

Abstract

Molecular glass resists have shown potential as replacements for polymeric resists in next generation lithography, especially extreme ultraviolet lithography. One of the main concerns about molecular resists is their glass transition temperature (Tg) which can be very low in some cases due to their small molecular size and other factors. While most of the polymeric chemically amplified resist platforms used thus far have Tg’s above 100 °C, molecular resists investigated in the literature so far have shown a wide range of measured Tg’s from near room temperature to greater than 160 °C. This potential for low Tg values and the current lack of ability to easily predict their Tg is a concern when designing new compounds because a molecular resist may be synthesized with a Tg value that is too low for the required processing conditions (e.g., allowing for dewetting of the resist, flow of the resist features, or excessive photoacid diffusion). To enable rational molecular resist design and overcome these problems, a quantitative structure-property relation model based on bond additivity that allows for the prediction of the Tg of molecular resists based on their full chemical structure has been developed in this work. The model shows a good coefficient of determination (R²) of 0.84 with experimental data, and a standard deviation of only 12 °C for 57 compounds. It works well across multiple different levels of protection, different structural moieties, different molecular sizes, and different types of protecting groups. The model was also simplified to provide a simple heuristic for predicting Tg based on only two or three structural parameters, and this easy to use simplified model provides a similar level of quantitative agreement with experimental data to the full bond additivity model.