Better Together: Ilmenite/Hematite Junctions for Photoelectrochemical Water Oxidation

Hematite (α-Fe2O3) is an earth-abundant indirect n-type semiconductor displaying a band gap of about 2.2 eV, useful for collecting a large fraction of visible photons, with frontier energy levels suitably aligned for carrying out the photoelectrochemical water oxidation reaction under basic conditions. The modification of hematite mesoporous thin-film photoanodes with Ti(IV), as well as their functionalization with an oxygen-evolving catalyst, leads to a 6-fold increase in photocurrent density with respect to the unmodified electrode. In order to provide a detailed understanding of this behavior, we report a study of Ti-containing phases within the mesoporous film structure. Using X-ray absorption fine structure and high-resolution transmission electron microscopy coupled with electron energy loss spectroscopy, we find that Ti(IV) ions are incorporated within ilmenite (FeTiO3) near-surface layers, thus modifying the semiconductor–electrolyte interface. To the best of our knowledge, this is the first time that an FeTiO3/α-Fe2O3 composite is used in a photoelectrochemical setup for water oxidation. In fact, previous studies of Ti(IV)-modified hematite photoanodes reported the formation of pseudobrookite (Fe2TiO5) at the surface. By means of transient absorption spectroscopy, transient photocurrent experiments, and electrochemical impedance spectroscopy, we show that the formation of the Fe2O3/FeTiO3 interface passivates deep traps at the surface and induces a large density of donor levels, resulting in a strong depletion field that separates electron and holes, favoring hole injection in the electrolyte. Our results provide the identification of a phase coexistence with enhanced photoelectrochemical performance, allowing for the rational design of new photoanodes with improved kinetics.

The electrophoretic deposition of mesoporous hematite (MPH) films on FTO was adapted from previous reports. [1] Briefly, in a Teflon beaker FeCl 3 •6H 2 O (0.55 g, >99%, Sigma Aldrich) was dissolved in a mixed solution of ethanol (20 mL) and water (5 mL) containing sodium acetate (0.8 g, ≥98%, Sigma Aldrich), and kept in a steel autoclave at 180 °C for 12 h. The resulting red powder (iron oxide nanoparticles) was washed several times with water and acetone and then suspended in acetone (50 mL). An aliquot (5 mL) of this colloidal dispersion was then mixed with a solution containing iodine (20 mg, Sigma Aldrich, ≥99.8%) in acetone (45 mL), and sonicated for 10 minutes. Two FTO slides (2 x 3 cm) were mounted at a distance of 0.8 cm, immersed in the colloidal solution and polarized at 10 V for 35 s in two electrode configuration using a ECO Chemie Autolab PGSTAT 302/N potentiostat. A homogeneous coating of the iron oxide nanoparticles was obtained on the negative electrode, which was then washed with acetone and annealed at 550°C for 1 h in air, then ramped up to 800°C for 20 min, yielding the MPH photoanodes.
As regards the preparation of Ti-modified samples, the same procedure was used, but the proper amounts of a 10 mM titanium(IV) butoxide (Sigma Aldrich, 97%) solution in ethanol were added to the deposition dispersion just before the application of the 10 V potential. A nominal 5 and 10% concentration of Ti(IV) with respect to the molar concentration of iron

Electrochemical impedance spectroscopy
The photoelectrochemical water oxidation efficiency ( WO ) of the different photoanodes can be calculated using the resistance and capacitance values obtained from the fit of EIS data, according to literature reports. [2,3] Being related to the interfacial performances of the photoelectrodes under illumination, the  WO depends on: (i) the kinetics S4 of water oxidation (i.e. the reaction of the surface trapped holes with the surface bound water molecules), to which a rate constant accounting for the charge transfer to the electrolyte (k CT ) is associated; (ii) the e -/h + recombination rate (k REC ) at the surface of the hematite electrode, which depends on the applied bias, and (as far as the hole concentration is concerned) on the excitation intensity; (iii) the charge recombination in the bulk of the film (since this aspect does not significantly change upon surface modifications, it will be neglected in this method).
Both k CT and k REC can be extracted from the fit of EIS data. In particular, the former can be calculated as follows: where R CT,SS is the resistance of the charge transfer from the surface states to the electrolyte and C CT is the space charge capacitance of hematite, obtained from the corresponding constant phase element (CPE CT ) admittance (extracted from the fit), according to the equation: being ω the angular frequency corresponding to the largest imaginary component of the charge transfer arc and n is the CPE exponent (0.7 ≤ n ≤ 1).
k REC can instead be obtained by the following simplified equation, valid at high applied bias: [4] , ≅ • , where R SC is the charge transport resistance through the space charge.