Developing high-energy and efficient battery technologies is a crucial aspect of advancing the electrification of transportation and aviation. However, battery innovations can take years to deliver. In the case of non-aqueous battery electrolyte solutions, the many design variables in selecting multiple solvents, salts and their relative ratios make electrolyte optimization time-consuming and laborious. To overcome these issues, we propose in this work an experimental design that couples robotics (a custom-built automated experiment named "Clio”) to machine-learning (a Bayesian optimization-based experiment planner named "Dragonfly”). An autonomous optimization of the electrolyte conductivity over a single-salt and ternary solvent design space identifies six fast-charging non-aqueous electrolyte solutions in two work-days and forty-two experiments. This result represents a six-fold time acceleration compared to a random search performed by the same automated experiment. To validate the practical use of these electrolytes, we tested them in a 220 mAh graphite∣∣LiNi0.5Mn0.3Co0.2O2 pouch cell configuration. All the pouch cells containing the robot-developed electrolytes demonstrate improved fast-charging capability against a baseline experiment that uses a non-aqueous electrolyte solution selected a priori from the design space.
Electrolytes play a critical role in designing next-generation battery systems, by allowing efficient ion transfer, preventing charge transfer, and stabilizing electrode-electrolyte interfaces. In this work, we develop a differentiable geometric deep learning (GDL) model for chemical mixtures, DiffMix, which is applied in guiding robotic experimentation and optimization towards fast-charging battery electrolytes. In particular, we extend mixture thermodynamic and transport laws by creating GDL-learnable physical coefficients. We evaluate our model with mixture thermodynamics and ion transport properties, where we show improved prediction accuracy and model robustness of DiffMix than its purely data-driven variants. Furthermore, with a robotic experimentation setup, Clio, we improve ionic conductivity of electrolytes by over 18.8% within 10 experimental steps, via differentiable optimization built on DiffMix gradients. By combining GDL, mixture physics laws, and robotic experimentation, DiffMix expands the predictive modeling methods for chemical mixtures and enables efficient optimization in large chemical spaces.
We propose a low-cost automated salt solubility screening module. Using an excess-solvent workflow with a 3D-printed Archimedes-screw solid doser and computer-vision dissolution detection, it enables faster mapping of electrolyte solubility trends.