Oxygen deficient α-Fe2O3 photoelectrodes: a balance between enhanced electrical properties and trap-mediated losses

Intrinsic doping of hematite through the inclusion of oxygen vacancies (VO) is being increasingly explored as a simple, low temperature route to preparing active water splitting α-Fe2O3–x photoelectrodes.

was deposited by thermal ALD using trimethyl-aluminium (TMA) and water in an Oxford Instruments OpAL reactor. The reactor was modified by the addition of a process controlled 'hold' valve between the pump and the process chamber, enabling precursor 'soak' steps. Soak steps were used in the process to ensure good film density and to promote conformal coating of the highly textured hematite surface. Coating was carried out at a low process temperature of 120 o C, selected to avoid accidental modification of the hematite samples. Films were deposited using 10 cycles of ALD targeting ~1 nm of Al2O3, using the following sequence: {(50 ms TMA Dose)(10 s TMA hold)(10 s purge)(30 ms H 2 O)(10 s H 2 O hold)(10 s purge)} The growth rate of the Al 2 O 3 was estimated using a Rudulph Auto EL IV ellipsometer operating at 633 nm using silicon 'witness' samples. This confirmed that 10 cycles gave close to 1 nm of Al 2 O 3 . In line with previous reports, 1 the effect of Al 2 O 3 layer is found to be retained for at least 20 minutes in 1 M NaOH indicating a reasonable level of stability and we are able to obtain kinetic data at up to two wavelengths.

Electrochemical measurements
A custom designed H-photoelectrochemical cell containing an α-Fe 2 O 3 photoanode, illuminated (ca. 0.5 cm -2 illuminated area) from the electrolyte-electrode (EE) side, a platinum gauze counter electrode and a 3 M KCl Ag/AgCl reference electrode (SSE, Bioanalytical Systems Ltd.) protected from the electrolyte solution by a KCl double junction was used in all experiments. Potentials quoted in the text are converted to RHE using equation 1.
Where E 0 Ag/AgCl = 0.205 V at room temperature. The electrolyte used throughout was 1 M NaOH (Aldrich, pH 13.6) prepared with Milli-Q water (Millipore Corp, 18.2 MΩ cm at 25 o C) and prior to all experiments it was thoroughly degassed with a stream of argon for at least 20 minutes. Hole scavenging experiments were carried out in the presence of 0.5 M H 2 O 2 in the electrolyte. Due to the low stability of Al 2 O 3 at high pH values the electrolyte was purged thoroughly prior to addition to the electrochemical cell which was held under an argon atmosphere. Linear sweep voltammograms were recorded both in the dark and under illumination with a 75 W Xe lamp (ca.10 mW cm -2 ) using either a PalmSens 3 (PalmSensBV) or a Ministat (Thomson Scientific) potentiostat typically at 5 mV s -1 unless otherwise stated.

Transient measurements
The TA apparatus has been described elsewhere. 2 Briefly, the third harmonic of a Nd:YAG laser (Continuum, Surelite I-10, 355 nm, 4-6 ns pulse width) operating at 0.33 Hz is the UV excitation source. The repetition rate of the laser was chosen to ensure that all photogenerated charge carriers had fully decayed prior to the next excitation event. A laser intensity of ca.200 J cm -2 at 355 nm was incident on the photoelectrochemical cell, leading to an intensity of ca.100 J cm -2 at the sample. We note that high pump laser intensities can lead to excessively high electron-hole recombination rates and in mind of this we use the relatively low excitation power of ca.100 J cm -2 throughout, in order to limit/eliminate this effect. There are a number of past studies on hematite regarding these laser intensities which give us confidence that the parameters used here are relevant in providing useful information. 3 A 75 W Xe lamp (OBB Corp.) coupled to monochromator (OBB Corp., set to 4 nm resolution) acts as the probe light and the change in optical density of the sample is calculated by measuring the transmitted light using a Si Photodiode (Hamamatsu) and a homemade amplification system coupled to both an oscilloscope (Textronix TDS 220) and data acquisition card (National Instruments NI-6221). In order to generate the TA spectra at a range of biases, 200 laser shots per wavelength were recorded. Accurate kinetic traces were recorded in a separate experiment using a minimum of 500 shots per wavelength. All TA experiments were carried out on the PEC cell used for electrochemical measurements, under potentiostatic control at the potential indicated in the text. Transient photocurrent measurements were recorded using the same apparatus with the light from the Xe probe lamp blocked. The current was obtained by measuring the voltage drop across a 47 ohm resistor in series with the working and counter electrode. From the absorption spectrum at 355 nm and the known pump light intensity of ca.100 J cm -2 we can estimate the photo-generated carrier density and compare it to the N d . We calculate the number of incoming photons to be on the order of 1.78x10 14 cm -2 demonstrating that expected oxygen vacancy concentration (assumed to be the difference in N d ) is much greater than the photo-generated charge carrier density.

Linear sweep voltammograms in the presence of H 2 O 2 hole scavenger
In the presence of 0.5 MH 2 O 2 we observe a photocurrent at potentials as low as 0.7 VRHE on both α-Fe 2 O 3 and α-Fe 2 O 3-x confirming that (i) initial charge separation is occurring to some degree in both materials at potentials well below the water oxidation photocurrent onset potential and (ii) that in both samples the photoholes are able to reach the surface to participate in oxidation reactions. The significantly higher photocurrent for the oxygen deficient sample in the presence of H 2 O 2 (ca. x2) is in line with expectations that the greater degree of band bending in Fe 2 O 3-x enables higher initial charge separation efficiencies. It is significant to note however that although this measurement indicates that the initial charge separation efficiency in Fe 2 O 3-x is approximately x2 that of Fe 2 O 3 this is intself is not sufficient to explain the complete lack of water splitting activity on Fe 2 O 3 . This is in good agreement with Figure 3 of the main text which also shows that a significant population of photoholes remain in Fe 2 O 3 on the micro-to millisecond timescale in NaOH at 1.4 V RHE.   Unlike in Fe 2 O 3-x , the addition of an Al 2 O 3 ALD layer onto an air annealed α-Fe 2 O 3 sample showed no improvement in activity confirming that sub-surface oxygen vacancies are required for photoelectrochemical activity in these otherwise un-doped samples.

at (a) 700 nm (photoholes) in α-Fe 2 O 3-x in the presence of hydrogen peroxide hole scavenger at 1.4 V RHE following UV excitation (355 nm) in 1 M NaOH (red line on green curve).
The TA trace measured at 700 nm (holes) in the presence of hydrogen peroxide hole scavenger was found to be well fitted to a bi-exponential decay of the form. In contrast to experiments solely in 1 M NaOH no slow hole injection term is identified. This is in agreement with past studies which suggested that scavenging of surface holes by H 2 O 2 is an ultrafast process, occurring on timescales faster than studied here (2 s -1 s). 8 We are therefore able to assign that the a significant population of holes transfers to the SCLJ on the sub 2 microsecond timescale, in agreement with a recent fs-TAS study. 9 We note that the low concentration of surface trapped holes remaining here likely to be due to the limited availability of H 2 O 2 at the surface proposed to be due the finite concentration of H 2 O 2 present at the electrode surface under the conditions examined (500 mM H 2 O 2 in the bulk solution).