Measuring self-assembled micelle topologies of functionalised rylenes to build a predictive machine learning model

This project centres around the creation of machine learning models which will ultimately allow for the prediction of the morphology and ultimately the physical properties of self-assembled aggregates. Due to the mechanisms which lead to molecular self-assembly being poorly understood, the design of new materials for devices is often unachievable. We therefore are developing a model that would predict the morphology of the aggregate from chemical structure alone. In order to generate models capable of predicting the properties of self-assembled aggregates, high-resolution data across a range of compounds needs to be acquired to achieve a prediction with sufficient confidence in order to be useful. This could pave the way for the design of responsive organic materials with the capability of replacing metals in high-value mechanoresponsive devices, amongst other applications.本项目以构建机器学习模型为核心，最终实现对自组装聚集体（self-assembled aggregates）的形貌及物理性质的预测。由于当前对分子自组装的作用机制尚未形成充分认知，面向器件应用的新型材料设计往往难以实现。为此，我们正开发一种仅基于化学结构即可预测聚集体形貌的模型。为构建可准确预测自组装聚集体性质的机器学习模型，需获取覆盖多种化合物的高分辨率数据，以得到具备足够置信度、具备实用价值的预测结果。该研究可为响应型有机材料的设计开辟路径，这类材料有望在高价值力响应（mechanoresponsive）器件中替代金属，并可应用于诸多其他场景。