Matters Arising in -- "Plasmon-driven carbon-fluorine (C(sp^$3$)-F) bond activation with mechanistic insights into hot-carrier-mediated pathways".

In a recent paper~[Nature Catalysis 3, 573 (2020)], Robatjazi {\em et al.} demonstrate hydrodefluorination on Al nanocrystals decorated by Pd islands under illumination and under external heating. They conclude that photocatalysis accomplishes the desired transformation \ce{CH3F + D2 -> CH3D + DF} efficiently and selectively due to "hot" electrons, as evidenced by an illumination-induced reduction of the activation energy. Although some of the problems identified in prior work by the same group have been addressed, scrutiny of the data in~[Nature Catalysis 3, 573 (2020)] raises doubts about both the methodology and the central conclusions. First, we show that the thermal control experiments in~[Nature Catalysis 3, 573 (2020)] do not separate thermal from "hot electron" contributions, and therefore any conclusions drawn from these experiments are invalid. We then show that an improved thermal control implies that the activation energy of the reaction does not change, and that an independent purely thermal calculation (based solely on the sample parameters provided in the original manuscript) explains the measured data perfectly. For the sake of completeness, we also address technical problems in the calibration of the thermal camera, an unjustifiable disqualification of some of the measured data, as well as concerning aspects of the rest of the main results, including the mass spectrometry approach used to investigate the selectivity of the reaction, and claims about the stoichiometry and reaction order. All this shows that the burden of proof for involvement of hot electrons has not been met.