Organic Electrode Battery Materials : A Journey from Quantum Mechanics to Artificial Intelligence

Batteries have become an irreplaceable technology in human life as society becomes progressively more dependent on electricity. The demand for novel battery technologies has increased fast, especially with the popularisation of different portable devices. However, the current battery industry relies heavily on non-renewable resources that are also prone to provoke environmental harm. Among the possible candidates for the next generation of batteries, organic electroactive materials (OEMs) have become attractive due to a series of advantages: vastly accessible from renewable raw materials; highly versatile due to the possible functionalisation mechanisms; possibly lower production costs; reduced environmental impacts; etc. Nevertheless, some drawbacks need to be overcome before OEMs become competitive. Issues with energy density, rate capability and cycling stability hinder their final technological application. This thesis thereby discusses fundamental aspects of OEMs and proposes novel techniques to accelerate the materials discovery process. The first part of this thesis presents a pathway to systematically investigate organic materials by combining quantum mechanics calculations and crystal structure predictions. An evolutionary algorithm predicts the crystal structure of several OEMs, enabling an initial assessment of the electronic structure and the thermodynamics of the ionic insertion mechanism in these compounds. Furthermore, this first part also suggests an approach to tailor OEMs, identifying their charge storage limits and the possible occurrence of metastable phases during the ion insertion process. However, the presented strategy, while accurate, is seriously limited by its high computational demands, which are unrealistic for high-throughput screening of novel materials. Since organic materials represent a possibly limitless universe of compounds, alternative techniques are needed. Thus, the second part of this thesis combines quantum mechanics and artificial intelligence (AI), rendering a powerful platform to aid this task. An “AI-\textit{kernel}” was employed to analyse millions of organic compounds, discovering novel possible organic battery materials. Moreover, the AI accurately identified common functional groups associated with higher-voltage electrodes and suggested features that may aid future materials design. Furthermore, the kernel can also identify materials suitable for Na- and K-ion batteries and anticipate their redox stability. In conclusion, this thesis has focused on investigating fundamental properties of organic electroactive materials, particularly the ionic insertion process in batteries. Furthermore, AI-driven methodologies have also been proposed, accurately evaluating OEMs and enabling fast access to the gigantic organic realm when searching for novel battery electrode materials