I would like to suggest the following:

- better clarify (also in Abstract and Conclusions) the range of concentrations of H2 for which the proposed system is valid. I see a range of 50-250 ppmv in the tests;

- indicate related applications for that range (I assume not for atmospheric measurements).

- be more rigorous in the use of the terms "sensitivity" (which is voltage vs concentration) and "accuracy" (measured concentration vs real concentration), and to report (Abstract and Conclusion) the accuracy of the standard that can be obtained.

Dear authors, thank for precedent clarification.

Based on the SI, the long term test confirms the repeatability of the method. Concerning stability over time, the current yield stay close to 1 during 14 hours. Anyway hydrogen production seems stable during the first two hours after the initial run-in phase, then a regular drift is observed (from 2 to 14 h, with an increase of +20 ppm). We notice “peaks” at regular intervals as if compensation has been made.

Could you explain what experimentally leads to this ?

Best regards.