Substitution Vs. Precipitation*

Hematite is predicted to have a high water-splitting efficiency but is limited by a low carrier conductivity. This limitation can be overcome by increasing the carrier concentration through doping; however, the formation of dopant precipitates hinders the free polaron conductivity. Recent theoretical calculations predict that a critical concentration of dopants can be incorporated into the hematite structure at a given annealing temperature, above which a simple dopant-oxide precipitate begins to form. This saturation point varies with the type of dopant and for Sn and Ge, synthesized at 800^oC, they are 0.22% and 0.44% respectively. Using EXAFS data analysis, we test these predictions by determining the fraction of dopants going into hematite and the fraction going into a dopant-precipitate for various dopant concentrations. We find that at low concentrations the dopants disperse evenly into the hematite lattice and use these results to model the dopant in hematite, while the dopant in a dopant precipitate is modeled using results from diffraction data. Our results suggest the saturation point is between 0.23-0.41% and 0.20-0.53% for Sn and Ge respectively, above which the fraction of dopants going into the precipitate increases logarithmically with total dopant concentration. We find that a simple dopant-oxide SnO₂ forms in the Sn doped system, whereas a more complex structure Fe₈Ge₃O₁₈ forms in the case of Ge. Our results for the saturation point are consistent with theoretical predictions.