(Invited) Halide Vapor Phase Epitaxy of III-V Alloys for Unassisted Water Photoelectrolysis

Abstract

Photoelectrochemical water splitting devices comprising III-V semiconductors alloys hold record solar-to-hydrogen efficiencies, achievable only by complex architectures involving tandem cells i.e an inverted metamorphic GaInAsP/GaInAs solar cell with an efficiency of 32.6% under 1-sun and the GaInAs/p-GaInP PV/PEC with a solar to hydrogen conversion efficicncy close to 17% under 1-sun as well; nanostructures and buried junctions, which leads to high manufacturing complexity. These approaches have been necessary to circumvent the shortcomings towards unassisted water splitting of state-of-the art III-V alloy-based single absorber devices, particularly of InGaP2, InGaAs, GaPN, and GaAsN, such as a) inadequate band edge alignment with respect to the HER and OER redox potentials, b) insufficient photovoltage to favor charge separation at the solid electrolyte liquid junction, and/or iii) charge recombination. An alternative path to overcome these limitations while maintaining device simplicity and competitive solar fuels costs is to develop novel III-V photoabsorbers with adequate band energetics.First-principles DFT+U calculations incorporating the local density approximation and generalized gradient approximation have shown that incorporation of Sb into GaN or GaP forms GaSbxN1-x and GaSbyP1-y alloys with narrowed band gap in the former case and changes the electronic band structure from indirect to direct in the latter. Theoretical computations predict that with band gaps in the order of 2 eV, these materials could be able to straddle the electrochemical potentials for water oxidation and proton reduction in acidic electrolytes.Metal organic chemical vapor deposition (MOCVD) and hydride vapor phase epitaxy (HVPE) have been used to synthesize GaSbxN1-x and GaSbyP1-y alloys, respectively, achieving single crystalline quality in a wide range of homogeneous Sb incorporation. Experimental measurements confirm the significant band gap reduction from 3.4 eV in pure GaN to 1.5 eV in GaSbxN1-x as determined by Tauc plot analysis of diffuse reflectance data and low temperature photoluminescence spectroscopy; and suggest reduction of the direct transition of GaP at 2.68 eV in the case of GaSbyP1-y alloys.The photoelectrochemical performance towards water splitting of GaSbxN1-x and GaSbyP1-y alloys has been assessed in 2- and 3-electrode standard methods to determine their characteristic attributes, e.g. flat-band and onset potentials, photovoltage, zero-bias photocurrent density, fill factors, carrier concentrations, and most importantly, their ability for gas evolution by in situ fluorescence probing. Furthermore, photoelectrochemical impedance measurements have shown the absence of charge transfer resistance at the interface between electrode and electrolyte in the case of GaSbyP1-y alloys that is clearly observed in pure GaP, suggesting antimony can also modify the charge transport properties of the material.These promising results have also motivated our efforts for better understanding of the synthesis processes, particularly for the HVPE technique given its fast growth rate is both an advantage and a challenge for precise thickness and composition control. A computational model using Chemkin Prowas used to validate the previous experimental results for the growth of Gallium Antimonide Phosphide (GaSbyP1-y) films. The model was developed for a reactor having one heating zone for precursor formation and a second zone for film deposition. The first zone was assumed to be in thermodynamic equilibrium. The second (deposition) zone was modeled using a chemistry set describing various gas phase and gas-solid kinetics. The substrate temperature and system pressure were successfully correlated with the observed experimental values for the growth rate and other important features such as the crystalline quality and degree of Sb incorporation. The results also point out the major gas phase chemical species involved in the growth, such as GaCl, SbCl, P2 and P4, which opens the door for future optimization of the synthesis process oriented towards a better precursor utilization and higher antimony incorporation at elevated substrate temperatures, which are required for single crystalline material quality. Acknowledgements: Financial support from Conn Center for Renewable Energy Research and NSF EPSCoR Program (1355438)