AmericanChemicalSociety.com

Photosensitized charge separation in Ni²⁺:ZnO and Ni²⁺:InTaO₄ mesostructured photoanodes

With ACS-PRF support, we have been studying sensitized photoinduced charge separation in transition-metal-doped oxides that exhibit water splitting under visible light irradiation. By photocurrent action, absorption, and magnetic circular dichroism (MCD) spectroscopies, we were able to identify the electronic transitions leading to charge separation in Ni²⁺:ZnO and Ni²⁺:InTaO₄, and to determine that sub-bandgap photocurrent activity is mainly due to charge transfer transitions between the nickel dopant and oxide semiconductor.

Doping has been shown to improve visible light absorption in wide bandgap semiconductors for more efficient water splitting, but the process by which dopants facilitate the formation of free carriers suitable for water cleavage is largely unknown. One of the few oxide materials that has been shown to efficiently and robustly catalyze overall water splitting is Ni²⁺-doped InTaO₄, which can do so under visible light irradiation. Undoped InTaO₄, with a bandgap of 3.7-4.3 eV, exhibits little activity in the visible region. It has been suggested that an additional d-d absorption band arising from nickel doping is responsible for the increased photoactivity at >420 nm, but such d-d bands are usually highly localized and hence ineffective for photoredox reactivity. To assist our investigation, we have also been studying Ni²⁺:ZnO, a well known doped oxide semiconductor that can be used as a model for other Ni²⁺ doped oxides.

Ni:InTaO₄ is typically prepared by solid-state synthesis, a method that results in powders that cannot be used for many optical studies. For optical quality Ni:InTaO₄ films, a sol gel method was developed, in which a sol solution of tantalum ethoxide, nickel acetate, and indium acetate in acid/ethanol was deposited onto a substrate, followed by calcination. Ni²⁺:ZnO films were prepared by spin-coating colloidal suspensions of Ni²⁺:ZnO nanoparticles onto transparent conducting glass. Photovoltaic cells were constructed from the oxide photoanodes, an I^‐/ I₃^‐ electrolyte, and a platinum counter electrode.

For Ni:InTaO₄, we observed an indirect bandgap of 3.8 eV in InTaO₄, which narrowed to 3.2 eV upon nickel doping. A charge transfer transition could be detected below the bandgap. MCD spectroscopy indicated a broad charge transfer band between 22,000-29000 cm^-1, as seen in the absorption spectrum. This charge transfer region overlaps with the pseudo-octahedral ⁴A₂à⁴T₁(P) ligand field transition around 22,000cm^-1, predicted from ligand field calculations. Using photocurrent action spectroscopy in conjunction with the above measurements, we observed internal quantum efficiencies (IQEs) corresponding to bandgap, charge transfer, and ligand field excitations. With these data, we could successfully identify the electronic transitions giving rise to the observed photocurrent.

Direct bandgap excitation gives the highest IQEs, followed by the charge transfer excitation, and then the ligand field excitation. Ni²⁺:ZnO films showed a similar trend, except that the ⁴T₁à⁴T₁(P) ligand field transition in this material does not give rise to any detectable photocurrent. Despite suggestions in the literature that d-d transitions are responsible for the visible light photoresponse of this class of materials, therefore, these transitions are concluded to be localized. Instead, charge transfer transitions that also result from doping are concluded to be responsible for the new visible light photoresponse. These charge transfer excitations are less efficient at generating free charge carriers than direct bandgap excitation is. The extremely small quantum efficiencies observed with some d-d transitions were shown to arise from sensitization of the charge transfer states by d-d exciation.

Solar water oxidation with oxide photoanodes

Storage of solar energy as hydrogen fuel via photocatalytic water splitting is a central challenge in solar energy conversion research. With ACS-PRF support, we have been studying electrocatalysis by oxide photoelectrodes. Water electrolysis is a multielectron process involving challenging catalysis that often requires high overpotentials to occur. We have shown that, by applying a competent cobalt-based electrocatalyst onto a mesostructured Si-doped α-Fe₂O₃ photoelectrode, solar water splitting is now accessible at an external bias ≥350 mV below what is typically required for Si-doped iron oxide alone. The results demonstrate that separating the tasks of photoabsorption, charge separation, and redox catalysis can improve the efficiency of solar water splitting.

Mesostructured oxide films were synthesized by atmospheric pressure chemical vapor deposition (APCVD). Si doping improved performance by reducing feature sizes to within the hole diffusion length (2-4 nm). For the electrocatalyst, we used “Co-Pi”, an amorphous electrocatalyst consisting of Co:K:P (2.7:1:1) shown to catalyze water oxidation at overpotentials as low as 0.28 V. Deposition of “Co-Pi” onto the oxide photoanode following literature procedures resulted in a >350 mV cathodic shift in the onset potential for photoelectrochemical water oxidation. SEM images showed that Co-Pi electrodeposition onto Si-doped α-Fe₂O₃ is conformal. Absorption and incident photon to current conversion efficiencies (IPCE) of α-Fe₂O₃ and Co-Pi/α-Fe₂O₃ photoanodes indicated that photocurrent generated from the composite photoanode still derived from α-Fe₂O₃ excitation. The IPCE curve of Co-Pi/α-Fe₂O₃ exhibited the same features as Fe₂O₃, indicating iron oxide is the photoactive species. Co-Pi alone is not photoactive.

This was the first successful demonstration of such a composite photoelectrode. Instead of using α-Fe₂O₃ for both photon absorption and catalysis, a cobalt-based electrocatalyst is used as the redox catalyst. By electrodepositing Co-Pi onto α-Fe₂O₃, photon absorption (α-Fe₂O₃) and redox catalysis (Co-Pi) can be individually optimized. Our work demonstrates that partitioning photoabsorption, charge separation, and redox catalysis is an effective strategy for improving solar water splitting with photoelectrochemical cells. Further optimization of photocurrent densities will require a better understanding of the Co-Pi/α-Fe₂O₃ interface and hole transfer across this interface. Additional work will seek to enhance Co-Pi/α-Fe₂O₃ performance by optimization for front-side illumination and by understanding the Co-Pi/α-Fe₂O₃ interface using various electrochemical techniques, including time-resolved and impedance spectroscopies.