The Need
The search for sustainable, high-performance battery materials is hindered by reliance on finite metal-based resources and slow, trial-and-error development cycles. Organic electrode materials (OEMs), composed of earth-abundant elements, offer a more sustainable path but present challenges in identifying cost-effective candidates that balance energy density, solubility, and ease of synthesis. Existing discovery approaches are too slow and inefficient to keep pace with the demands of emerging grid and portable energy storage markets.

The Technology
This technology introduces SPARKLE, a machine learning–driven framework that enables zero-shot prediction and discovery of optimal OEMs by jointly considering performance (specific energy), stability (solubility), and cost (synthesizability). SPARKLE uses symbolic regression, generative molecular design, and a new symmetry-adapted synthetic accessibility score to evaluate a massive design space of over 670,000 compounds. It identified ~5000 viable candidates, 27 of which were synthesized and tested. SPARKLE-discovered materials outperformed state-of-the-art OEMs in energy density, cost-efficiency, and cycling stability.

Benefits/Advantages
• 3× Higher Discovery Success Rate than human expert selection (62.9% vs. 20.8%)
• Zero-shot discovery: no prior electrochemical data required
• Superior energy performance: Specific energy >250 Wh/kg in top candidates
• Lower material cost through symmetry-enhanced synthetic scoring
• Generalizability to new OEM classes, including unexplored p-type materials like benzidines