Recently, the Shanghai Artificial Intelligence Laboratory, in collaboration with multiple research institutions, successfully solved the stability preparation problem of high-end KrF photoresist resins using the "Shu Shen" scientific large model and automated R&D platform. This breakthrough marks a major progress in China's core material field for chips, with key industrial indicators of related products already meeting expectations.

The development of photoresist resins has long relied on manual experience, with researchers needing to repeatedly test and error through thousands of formulations and reaction conditions. This traditional R&D model is not only inefficient but also prone to operational errors, making it difficult to ensure the high stability required for mass-produced chips.

AI-Driven Closed-Loop R&D

Through the construction of a "AI decision-making + automated synthesis" closed-loop system, the research team achieved full-process operation from experimental design to automated post-processing. The platform uses precision control technology to stably keep the metal impurity content of the finished resin at an extremely low level, significantly improving the purity and consistency of the material.

By leveraging the self-evolution capabilities of AI models, the research team successfully transitioned from "experience-driven" to "data-driven" R&D. Key data generated during experiments automatically returns to the large model, driving continuous algorithm optimization for the next round of plans, thereby greatly shortening the development cycle of high-end materials.

Reducing Reliance on Overseas Supply

This technological breakthrough enables the production of high-end photoresist resins without relying on the "black box" capabilities of a few foreign suppliers, offering a standardized path for global chip material development. Currently, the relevant achievements have entered client validation stages, laying a solid foundation for enhancing the independence of China's semiconductor industry chain.

This research result demonstrates the great potential of artificial intelligence in fundamental scientific discovery and indicates that AI-native workflows will deeply reshape the field of materials science. In the future, this highly modularized intelligent synthesis platform will continue to drive more core semiconductor materials to achieve technological breakthroughs.