at constant potential. It is critical for the stabilization strategy that the OER activity of the modified
catalyst does not decrease significantly. The activity of the catalysts supported on the gold-coated
quartz crystal microbalances is plotted on Figure 6a. The TiO2-modified samples exhibit an 11% lower
current density at 1.9 VRHE. This is less than expected from the 20 % TiO2 content. Over time, the loss
of Mn could lead to an enrichment of Ti in the surface, which would be detrimental to the activity. We
ruled out this possibility, by measuring the Mn:Ti ratio of a tested sample with XPS, which again
showed about 20 % of Ti at the surface, see Figure S2d. Since the ratio between Mn and Ti, for the 1
nm depth probed by XPS, is constant throughout the test it is likely that the role of the Ti is to block
dissolution only when located at the undercoordinated sites. Figure 6b shows the corresponding mass
losses of manganese evaluated with ICP-MS; at 1.9 VRHE the TiO2 addition leads to 40 % lower losses.
This value is four times as high as the loss in OER current, indicating that a better compromise
between activity and stability can be achieved by modifying MnO2 with TiO2. It should also be noted
that RuO2 thin films prepared and characterized with the same methods exhibit mass losses that are
roughly six times higher than MnO2 at 1.8 VRHE.32 The findings for EQCM substrates were further
validated by RDE tests, combined with ICP analysis, of the catalyst films supported on polycrystalline
gold disks, see Supporting Information.
Figure 6. a) OER current after 1-hour of the chronoamperometry tests at 1.8 and 1.9 VRHE. b)
Mass loss for MnO2 and TiO2-modified MnO2 based on ICP-MS analysis. The amount of Mn
measured in the electrolyte is then corrected for volume of electrolyte and converted into
corresponding MnO2 weight. Blue bars show the MnO2 while the red bars show TiO2-modified