Executive Summary: Material Selection via Data Mining

The goal of my project was to identify potential Li-ion fuel cell cathode materials that could be synthesized via the organic steric entrapment synthesis method used by Prof. Kriven’s research group. To accomplish this, I used information available on the Materials Project and results from queries of the Materials Project’s Web application programming interface (API), through the use of pymatgen, a Python library for materials analysis.

The Materials Project API was used due to the wide range of potential candidate materials it indexes (40,000+) and ease of availability of useful material properties (eg; volumetric charge capacity, unit cell expansion, delivered voltage, etc.). However, the organic steric entrapment synthesis method used by my research group (the polyvinyl alcohol, or PVA method), imposes limitations of its own upon possibly synthesized powders, including the relative size of cations, maximum cation valency, and geometric constraints.

After using a literature search to devise a weighted system for candidate selection criteria, an initial candidate pool of potential Li-ion fuel cell cathode materials of ~500 was reduced to a pool of 14 prospective strong candidates for synthesis. These candidates were not novel - they have already been synthesized through other synthesis methods and in some cases studied extensively - but as a result of applying selection criteria weights imposed by both the PVA method and desirable properties for Li-cell cathode materials, it's clear that these materials can be synthesized through the PVA method. The five most promising candidate materials, in descending order, are LiMnO2, LiCoO2, LiCoPO4, LiMn2O4, and LiTiO2.


Please feel free to contact Max McKittrick if you have any questions!