Machine-Guided Design of Synthetically-Accessible, Processable, Cost- Effective, High-Performance, High-Temperature Polymer Dielectrics
Abstract
Biaxially oriented polypropylene (BOPP) continues to be the industry standard for electrostatic capacitive energy storage devices and pulsed power applications. Innovations in the last decade have indeed led to a few alternative dielectric polymer options. Our own recent ONR funded work involving iterative computations, machine learning, synthesis, processing and testing has enabled the identification of new materials that may surpass BOPP on many performance measures, including high temperature performance. However, given the practically infinite polymer chemical space, it is extremely likely that a number of new and much better dielectric polymers are awaiting discovery. Significant barriers exist, however, to identifying promising new useful polymers. Chief among them is the staggering size of the chemical space to be sifted through, and a lack of rapid prediction models for all (or most) relevant properties that span a large enough polymer chemical, synthesizable and processable spaces. In this proposed work, we seek to take important critical steps towards overcoming such barriers using emerging computational and data mining methodologies. Our primary objective here will be to add critical new polymer informatics capabilities to (1) assist synthetic chemists in their decision-making process, and (2) perform an unprecedented and exhaustive search of the polymer chemical space to identify polymer dielectric candidates that are high performing at high temperatures, while also being cost effective and accessible to synthesis and processing. All planned work will be done collaboratively with Profs. Sotzing, Cao and Cakmak, and with other ONR Dielectric Materials and Film Research Program researchers, when appropriate and possible. The planned tasks are: A. Reassess and redraft property-based screening criteria for polymer dielectrics (Table 1).B. Extend informatics capability from homopolymers to copolymers and polymer blends, such that predictions of relevant properties may be made across this entire space. C. Create a large candidate set of (billions of) polymers using #virtual# synthesis, starting from known polymerization reactant templates and commercially available monomers. D. Create a synthetic-feasibility score for polymers (p-score) that estimates the ease with which a polymer in the candidate set (or any new polymer, for that matter) may be synthesized. E. Use the properties and desired target values of Table 1 to screen the candidate set of billions of polymers to identify a set of recommended polymers meeting target property criteria. F. Validate the recommended polymers through a combination of physics-based computations and lab synthesis (Sotzing) and characterization (Cao). The ultimate outcomes of this effort are anticipated to be a set of design criteria for capacitor dielectrics, a knowledge framework for the rapid prediction of the properties of polymers, and validated designs of high temperature high energy density polymers that are cost effective, amenable to easy synthesis, processing, and manufacturability at-scale. Approved For Public Release.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 03, 2023
- Source ID
- N000142312279
Entities
People
- Ramamurthy Ramprasad
Organizations
- Georgia Tech Research Corporation
- Office of Naval Research
- United States Navy