Predictive Catalysis is based on combining Principal Component Analysis (PCA) with DoE. A collaboration with the University of Bristol provided high quality properties (descriptors) for ligands from DFT calculations. The resulting PCA maps of these ligands are then used to investigate a small but diverse set of catalyst systems representative of the total chemical space of the reaction of interest. Note that this approach is completely unbiased towards any aspects of ligand or pre-catalyst.
Any of the reaction parameters can be investigated to meet the ultimate design requirements of the client. This may include:
- Commercial availability of catalysts.
- Cost considerations, including activity and loading.
- Catalyst performance, such as selectivity and robustness.
- Freedom from patent constraints.
- Manufacturability and operability.
- Environmental impact and sustainability.
- Process safety.
For examples of combining DoE and PCA in catalysis see the following case studies: