Catalyst Development Using Soft Computing and High Throughput Screening (9) Optimization of Copper Catalyst for Methanol Synthesis Using Activity Map Envelope from Artifical Neural Network

摘要

We developed combinatorial tools consisting of high-throughput screening reactor using 96-well microplate and activity mapping by artificial neural network. The tool was applied for optimization of Cu-Zn-Al-Cr-B-Zr-Ga oxide catalyst for methanol synthesis. As optimization method, instead of genetic algorithm, more straightforward method was applied: all the 230, 230 activities of possible combination of catalyst component with 5 percent resolution were predicted by neural network (NN). These activities were visualized by mapping using two parameters to find the global optimum.