Performance Limiting Factors in Co-Pi Catalyzed, Spray-Deposited BiVO4 Photoanodes
Fatwa F. Abdi, Nienke Firet and Roel van de Krol Department of Chemical Engineering/Materials for Energy Conversion and Storage, Delft University of
Technology
P.O. Box 5045, 2600 GA Delft, The Netherlands BiVO4 is considered to be a promising photoanode material for solar water splitting applications. The monoclinic phase has a bandgap of ~2.4 eV, which corresponds to an optical absorption edge at ~520 nm. However, the quantum efficiencies reported so far do not exceed 55% [1-3], and the main performance-limiting factors are still not clear. In this work,
photo-electrochemical characterization of spray-deposited, dense films of BiVO4 results in an internal quantum efficiency exceeding 95% at low light intensities of few μW/cm2 [4]. This indicates that BiVO4 performance is not affected by interface traps, bulk defect states, or the presence of a Schottky barrier at the back contact. By comparing front- and back-side illumination, electron transport was found to be the main rate-limiting factor under these conditions [2]. This is further confirmed by a time-resolved microwave conductivity (TRMC) study that we performed, which reveals a carrier mobility of ~3 x 10 -3 cm2/Vs. This is 2 – 3 orders of magnitude lower than is typically observed for metal oxide photoanodes.
Intriguingly, Mott-Schottky analysis of BiVO4 films consistently indicates very high donor densities (>1019 cm-3), which implies that the width of the space charge region is just a few nm. To reconcile this with the high quantum efficiencies observed, the minority carrier diffusion length (Lp) would have to be rather large, in the order of the film thickness. This is indeed consistent with the recent findings of Zhong et al. [7], who reported Lp to
be ~100 – 200 nm. Such a high Lp value, together with the low carrier mobility from the TRMC experiments, suggests that the carrier lifetimes in BiVO4 must be unusually high. We will present the results of a recent study in which electrochemical impedance spectroscopy was combined with TRMC measurements and transient absorption spectroscopy in order to elucidate the interplay between donor density, carrier mobility and carrier lifetimes in BiVO4.
Integrating the quantum efficiencies over the solar spectrum leads to a predicted AM1.5 current density of 3.6 mA/cm2. However, the observed photocurrent density under simulated AM1.5 illumination is much lower than the predicted value (~0.6 mA/cm2) due to extensive carrier recombination. Under these high light intensity conditions, poor water oxidation kinetics (hole transfer) become rate-limiting, as evidenced by a significant increase of photocurrent density when hydrogen peroxide is added as a hole scavenger [4]. Based on these insights, we modified the sample with cobalt-phosphate as a water oxidation catalyst [5] to address the poor hole transfer, and tungsten as a donor-type dopant to enhance electron transport. This resulted in an AM1.5 photocurrent density of 2.3 mA/cm2 at 1.23 VRHE, as shown in Figure 1. While this is lower than the state-of-the-art photocurrent of 2.8 mA/cm2 recently reported by Luo et al. (see Table 1), our results have been obtained for a five-times thinner compact film of only 200 nm and a low cost cobalt phosphate as the co-catalyst. In addition, the amount of light scattering in these spray-deposited films is very small. This is an important advantage since these films would ultimately have to be incorporated in a tandem device structure to achieve overall water splitting.
0.5 1.0 1.5 2.0 0 2 4 1%W:BiVO4 + Co-Pi
j (mA/cm
2)
V vs. RHE (V)
BiVO4 BiVO4 + Co-Pi dark 1.23 VRHEFigure 1. AM1.5 photocurrent vs. voltage for BiVO4
photoelectrodes. The effects of the electrodeposition of Co-Pi catalyst and W-doping are shown by the red and blue curves, respectively.
Table 1. Performance summary of different BiVO4-based
photoanodes reported in the literatures.
References
[1] K. Sayama et al., J. Phys. Chem. B, 110 (2006) 11352 [2] Y. Liang et al., J. Phys. Chem. C, 115 (2011) 17594 [3] W. Luo et al., Energy Environ. Sci., 4 (2011) 4046 [4] F. F. Abdi and R. van de Krol, J. Phys. Chem. C
(accepted)
[5] M. Kanan and D. Nocera, Science, 321 (2008) 1072 [6] K. Sayama et al., Chem. Lett., 39 (2010) 17 [7] D. K. Zhong et al., J. Am. Chem. Soc., 133 (2011) 18370
[8] S. Pilli et al., Energy Environ. Sci., 4 (2011) 5028 [9] J. A. Seabold and K. S. Choi, J. Am. Chem. Soc., 134 (2012) 2186