Back Electron–Hole Recombination in Hematite Photoanodes for Water Splitting

Abstract

The kinetic competition between electron-hole recombination and water oxidation is a key consideration for the development of efficient photoanodes for solar driven water splitting. In this study, we employed three complementary techniques, transient absorption spectroscopy (TAS), transient photocurrent spectroscopy (TPC), and electrochemical impedance spectroscopy (EIS), to address this issue for one of the most widely studied photoanode systems: nanostructured hematite thin films. For the first time, we show a quantitative agreement between all three techniques. In particular, all three methods show the presence of a recombination process on the 10 ms to 1 s time scale, with the time scale and yield of this loss process being dependent upon applied bias. From comparison of data between these techniques, we are able to assign this recombination phase to recombination of bulk hematite electrons with long-lived holes accumulated at the semiconductor/electrolyte interface. The data from all three techniques are shown to be consistent with a simple kinetic model based on competition between this, bias dependent, recombination pathway and water oxidation by these long-lived holes. Contrary to most existing models, this simple model does not require the consideration of surface states located energetically inside the band gap. These data suggest two distinct roles for the space charge layer developed at the semiconductor/electrolyte interface under anodic bias. Under modest anodic bias (just anodic of flatband), this space charge layer enables the spatial separation of initially generated electrons and holes following photon absorption, generating relatively long-lived holes (milliseconds) at the semiconductor surface. However, under such modest bias conditions, the energetic barrier generated by the space charge layer field is insufficient to prevent the subsequent recombination of these holes with electrons in the semiconductor bulk on a time scale faster than water oxidation. Preventing this back electron-hole recombination requires the application of stronger anodic bias, and is a key reason why the onset potential for photocurrent generation in hematite photoanodes is typically ~500 mV anodic of flat band and therefore needs to be accounted for in electrode design for PEC water splitting.

Achieving Highly Efficient Photoelectrochemical Water Oxidation with a TiCl4 Treated 3D Antimony‐Doped SnO2 Macropore/Branched α‐Fe2O3 Nanorod Heterojunction Photoanode

“…The TiO 2 blocking layer fabricated by hydrothermal decomposition of TiCl 4 or electrodeposition has been capable to block the solution‐mediated recombination at the highly conductive substrates 23. Hence, the exposed conductive mpATO and FTO glass substrate would be sufficiently covered by the TiO 2 nanoparticles formed through the current TiCl 4 treatment, which leads to a declined back reaction and an enhanced photocurrent response.…”

Electron trapping induced electrostatic adsorption of cations: a general factor leading to photoactivity decay of nanostructured TiO2

“…[15][16][17][18][19] They may cause Fermi level pinning or partial pinning in some nanostructured photoelectrodes, 15 which is responsible of increasing the overpotential needed for PEC water oxidation. [16][17] On the other hand, during PEC processes these sites may act as recombination sites 18 or active sites for hole interface transfer (water oxidation) reactions. 19 In this work, we will demonstrate for the first time that surface defect states may exert another important influence on the dynamics of electrons in TiO 2 electrode which leads to the general photoactivity decay.…”

How Should Iron and Titanium be Combined in Oxides to Improve Photoelectrochemical Properties?

“…8,10 However, the reported water oxidation efficiencies for hematite to date are notoriously lower than the predicted maximum value despite intense efforts. 18 This is caused by two principal limiting physical properties: the short lifetime of the photogenerated charge carriers (≤ 10 ps), and the low mobility of the minority carriers (0.2 cm 2 ·V -1 ·sec -1 ). 18 The combined effect is that the minority carrier (hole) diffusion length is only 2-4 nm, whereas full absorption of the incident light requires much thicker films due to the indirect nature of the bandgap (the penetration depth of light is ~118 nm at a wavelength of 550 nm).…”

“…18 The combined effect is that the minority carrier (hole) diffusion length is only 2-4 nm, whereas full absorption of the incident light requires much thicker films due to the indirect nature of the bandgap (the penetration depth of light is ~118 nm at a wavelength of 550 nm). 18 …”