Designing mechanosensitive molecules from molecular building blocks: A genetic algorithm-based approach

Abstract

Single molecules can be used as miniaturized functional electronic components, when contacted by macroscopic electrodes. Mechanosensitivity describes a change in conductance for a certain change in electrode separation and is a desirable feature for applications such as ultrasensitive stress sensors. We combine methods of artificial intelligence with high-level simulations based on electronic structure theory to construct optimized mechanosensitive molecules from predefined, modular molecular building blocks. In this way, we overcome time-consuming, inefficient trial-and-error cycles in molecular design. We unveil the black box machinery usually connected to methods of artificial intelligence by presenting all-important evolutionary processes. We identify the general features that characterize well-performing molecules and point out the crucial role of spacer groups for increased mechanosensitivity. Our genetic algorithm provides a powerful way to search chemical space and to identify the most promising molecular candidates.