Mitigation methods for errors in oxygen measurement with redox cycling of materials for hydrogen and syngas production

In the search for new and improved materials for hydrogen and syngas production by solar thermochemical looping, test-reactors are employed which include a temperature controlled sample chamber and adjustable gas flows through or past the sample. The experiments performed in these devices enable researchers to find limiting factors like mass transfer, heat transfer, kinetics, and material durability in a time and cost efficient manner. The devices have proven their utility by their near universal employment by groups seeking and studying new materials. A review of past studies has revealed that the measurement of oxygen partial pressure during the reduction state is key to the evaluation of material productivity, yet the methods for this measurement are varied across different publications and are often given little focus. The majority of O2 sensing is achieved using a mass spectrometer or gas chromatograph, inferring behavior at the sample from measurements of gas that has traveled for some distance and time. In this paper, we investigate the potential errors which may be introduced by taking a single measurement of oxygen production at the system outlet to infer O2 production curves, and demonstrate some methods to correct this. We also investigate some of the issues related to including an oxygen sensor near the sample. Issues discussed include temporal delays between sensors, oxygen leakage, sensing an incompletely mixed flow, diffusion, and mixing downstream from the sample. Oxygen entering the system through inlet gas or leakage accounted for the largest source of error, but these errors can be corrected by straightforward methods. Numerical simulations are employed to investigate the mixing of the flow, while diffusion is estimated with an analytical model. During an example experiment, the applied correction methods reduced differences between two sensors' data from 20% to 7%, while the corrections led to a 36% change in calculated total oxygen production from raw to corrected data.