Dependence of the estimated diffusion coefficient of LixWO3 films on the scan rate of cyclic voltammetry experiments

In order to assess the effect of the scan rate on the Li+ ion diffusion coefficient as calculated from cyclic voltammetry (CV), a series of CV experiments at various rates (namely 2, 5, 10, 20 and 50 mV/s) has been carried out for the same WO3 film prepared by electron gun deposition at room temperature. The diffusion coefficient (D) for lithium ion intercalation into the WO3 film, calculated from the Randles–Sevcik equation has been compared to that calculated by the Galvanostatic Intermittent Titration Technique (GITT), carried out on the same WO3 film. It has been found that at high scan rates (above 2 mV/s) the Randles–Sevcik method tends to grossly overestimate diffusion coefficient, possibly due to the occurrence of large potential gradients within the WO3 film. Compared to the results deduced from GITT, the voltammetric diffusion coefficient was found to be up to 67% higher. It can therefore be concluded that diffusion coefficient values deduced by CV peaks with use of the Randles–Sevcik method must be treated with caution and that low scan rates (2 mV/s or below) are required for credible results.