Abstract: CO2 electrocatalytic reduction to value-added products stands out as one of the most favorable utilization pathways for captured CO2 to reduce emissions and mitigate climate change. However, current electrode fabrication techniques limit the implementation of the process at an industrial scale, as they are neither reproducible nor automatized. In this work, an automatic spray pyrolysis technique is studied by assessing the effect of different fabrication variables, such as (i) spraying nozzle height, (ii) distance between steps, and (iii) ink flowrate, from a statistical point of view, developing a linear regression model and a neural network-based predictive model that can forecast the behavior of the electrodes fabricated under different conditions. These statistical models are developed to advance from a rudimentary to an automatized electrode fabrication method, considering this as a first step towards establishing an electrode manufacturing protocol based on machine learning.

Authorship

Download reference