Ionic liquids (ILs) are salts, composed of asymmetric cations and anions, typically existing as liquids at ambient temperatures. They have found widespread applications in energy storage devices, dye-sensitized solar cells, and sensors because of their high ionic conductivity and inherent thermal stability. However, measuring the conductivity of ILs by physical methods is time-consuming and expensive, whereas the use of computational screening and testing methods can be rapid and effective. In this study, we used experimentally measured and published data to construct a deep neural network capable of making rapid and accurate predictions of the conductivity of ILs. The neural network is trained on 406 unique and chemically diverse ILs. This model is one of the most chemically diverse conductivity prediction models to date and improves on previous studies that are constrained by the availability of data, the environmental conditions, or the IL base. Feature engineering techniques were employed to identify key chemo-structural characteristics that correlate positively or negatively with the ionic conductivity. These features are capable of being used as guidelines to design and synthesize new highly conductive ILs. This work shows the potential for machine-learning models to accelerate the rate of identification and testing of tailored, high-conductivity ILs.
2 September 2022 in Molecular Systems Design & Engineering
Machine learning (ML) has emerged as a promising technology to accelerate materials discovery. While systematic screening of vast chemical spaces is computationally expensive, ML algorithms offer a directed approach to identifying and testing promising molecular candidates for specific applications. Two significant hurdles towards development of robust ML models are the quality and quantity of existing experimental and ab initio data for training ML models in new chemical spaces. Here we present a reliable, reproducible, and fully automated simulation pipeline that enables users with varied backgrounds to easily generate thermodynamic data. Our atomistic simulation pipeline is GPU accelerated and suitable for high-performance computing (HPC) environments. We validate our pipeline results against dedicated experimental work and existing literature data, then further demonstrate how ML may be employed via an active learning approach to further drive chemical exploration. First, ML models are trained to predict thermodynamic properties for a large set of small molecule candidates. Second, new molecules are picked and simulated based on the model uncertainty and expected model improvement. These additional simulations enhance the predictive capabilities of the model until we are satisfied with the overall prediction capability. We simulate 410 molecules using active learning within the automated simulation pipeline to enumerate properties of interest. Across our set of over 6000 small molecule candidates, our active learning procedure is able to predict monomer properties at error rates which are substantial improvements compared to a random selection baseline. We demonstrate that this approach is capable of reducing the number of completed simulations while simultaneously generating a reliable final model to predict thermodynamic properties for a large number of small molecules.
14 September 2026 in Computational Materials Science