(Invited, Digital Presentation) The Low T Water Electrolyzer Current-Voltage Relationship: Electrocatalysis and More

Abstract

As described in a recent publication [1], mechanisms of electrocatalytic processes have a common pattern which provides the basis for a universal rate expression for electrode processes in fuel cells. The same universal rate expression is applicable for low T water electrolyzers, serving as an effective predictor of the form of the non-linear dependence of the overpotential h on log J. The general form of this equation is exemplified by equation (1): (1) J (E) = Cr g Fk0A* x f(E - E0 surf. redox) 10^(- D H#/2.3RT) 10^ (E- E0 cell)/b. It can be seen that this expression considers two effects of the electrode potential on the current: one is the effect on the activation energy through the term 10^(E- E0 cell)/b, and the other is the effect of the electrode potential on the fractional population of active surface sites (or the fractional surface coverage by adsorbed intermediates), described in Eq 1 as: f(E - E0 surf.redox) where E0 surf.redox is the standard potential of a surface redox couple which “mediates” the faradaic electrode process . Consequently, E0 surf.redox determines the value of the onset potentialfor that faradaic process.Overpotential-driven activation of catalyst surface sites is a key component of processes taking place in electrodes of water electrolyzers, much the same as in the case of the fuel cell. The reason is that, as a rule, catalyst surface sites are inactive under open circuit conditions. In the case of the anode of the water electrolyzer , the origin of the main component of the electrode kinetic losses, the reason is the relatively low oxidation state of the metal ions in the oxyhydroxide catalyst surface sites. Iron sites are typically Fe(III) and iridium sites are Ir(IV) under open circuit conditions in contact with the aqueous electrolyte and O2 , whereas a higher oxidation state is required to withdraw electrons from water to form an *OH adsorbed intermediate. Such higher oxidation states are first generated at an anode overpotential of the order 0.15-0.20V, i.e., at ~1.40V vs. RHE.Highly active OER catalyst are all characterized by an oxide redox couple M+nox/ M+(n+1)ox of E0 ~1.40V , as can be verified from relevant Pourbaix diagrams. This means that the loss associated with catalyst site activation is significantly lower than in the case of the ORR, where it is near 0.30V. It is interesting to examine the reasons for this apparent advantage of the oxide catalyst in the OER vs. the PGM catalyst in the ORR. As explained in detail in reference [1], each electrocatalytic process is characterized by a two-step sequence: a pre-RDS step (“pre-step”) which involves surface site activation, followed by the rate determining step. The pre-step determines the onset potential and the RDS determines the rate of the electrocatalytic process at a given potential beyond the onset [1]. To large degree, the electrocatalysts of highest activity are determined by the value of the onset potential, as is best set to achieve the optimized bond strength required by the “Sabatier Principle” for the catalytic processes, which all involve bond formation followed by bond scission. In the case of the ORR process, Pt and it’s alloys which exhibit highest activity are characterized by a M/Mox surface potential of ~085V vs. RHE This means that forming a surface population of 1% of active M sites for the ORR, requires a minimum overpotential of ~0.30V whereas metal catalysts of lower oxygen affinity than Pt , which exhibit “bare” metal sites at much lower overpotentials, fail to reactively bond the diatomic oxygen molecule in the RDS. In the case of the OER, the relatively high activity of oxide catalyst sites in forming *OH ,as reflected by the low onset overpotential, does not seem to result in an excessive scission energy for the *OOH intermediate which most likely forms last in the sequence of the OER.Finally and importantly, the typical form of the polarization curves observed for low T aqueous electrolyzers, is fully accounted for by Eq. 1 which predicts very well the form of the non-linear h vs. log J polarization characteristics.Reference:[1] Shimshon Gottesfeld , A Perspective of Polymer Electrolyte Fuel cell Science and Technology, highlighting a general mechanistic pattern and a general rate expression for electrocatalytic processes , JES, in press