Geometrically asymmetric electrodes for probing electrochemical reaction kinetics: a case study of hydrogen at the Pt-CsH2PO4 interface.

ABSTRACT

Electrochemical reactions can exhibit considerable asymmetry, with the polarization behavior of oxidation at a given metal|electrolyte interface differing substantially from that of reduction. The reference-less, microcontact electrode geometry, in which the electrode overpotentials are geometrically constrained to the working electrode (by limiting its area) is experimentally convenient, particularly for fuel cell studies, because the results do not rely on accurate placement of a reference electrode nor must oxidant and reductant gases be sealed off from one another. Here, the conditions under which the critical assumption of this geometry applies-that the overpotential at the large-area counter electrode can be ignored-is numerically assessed. It is found that, for cells of sufficiently large area, the effective radius of the counter electrode (which defines the area through which the majority of the current passes) can be expressed directly as a function of electrolyte thickness and the materials properties, sigma, the conductivity of the electrolyte, and k, the reaction rate constant for the electrochemical reaction at zero-bias. From this effective radius and the true radius of the working electrode, the fraction of electrode overpotential at the latter, defined as the extent of isolation, can be readily computed. Experimental studies of hydrogen electro-oxidation/proton electro-reduction at the Pt|CsH(2)PO(4) interface using two cells of differing dimensions both validate the computational results and demonstrate that asymmetry in such reactions are readily revealed in the micro-electrode, reference-less geometry. The study furthermore confirms the insensitivity of the results to the precise placement of the working electrode, while indicating the importance of very high isolation values (>99%) to ensure that overpotential contributions of the counter electrode do not influence the measurements, particularly as bias is increased.