Case Study: Showa Denko accelerates solvent blend discovery with Citrine Informatics AI Platform

Showa Denko discovers 50+ solvent blends in 5 months with Citrine Informatics

The customer, Showa Denko, faced the challenge of identifying high-performance solvent blends from over 850 million possible combinations. Their initial testing data contained no blends that met all their downstream customer constraints, making it costly and time-consuming to find viable candidates. They partnered with vendor Citrine Informatics to leverage the Citrine AI Platform and target their testing efforts on the most promising blends.

Using its AI platform, Citrine Informatics implemented a sequential learning process to model and predict blend properties. This approach, combined with quantum chemistry simulations, efficiently explored over 100 million candidates. The solution delivered significant results: within five months, Showa Denko discovered more than 150 improved solvent blends, over 50 of which satisfied all project constraints, dramatically accelerating their innovation cycle.

Showa Denko