Towards a Hydrobromic Acid Splitting Device Using Earth-Abundant Materials

Citation

Roske, Christopher William (2017) Towards a Hydrobromic Acid Splitting Device Using Earth-Abundant Materials. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z9NS0RWS. https://resolver.caltech.edu/CaltechTHESIS:09062016-162323180

Abstract

This thesis disembarks from the traditional approach of tailoring a system to the water splitting reaction. As detailed in Chapter 2, this thesis predicts that two silicon photoelectrons connected in parallel are ideally suited to electricity storage in an integrated light collector and chemical storage device driving the splitting of hydrobromic acid (2HBr -> H₂ + Br₂). The predicted dual photoelectrode system could potentially obtain high solar-to-hydrogen conversion efficiencies of up to an η_{STH, HBr} of 12 %, whereas an equivalent water splitting system is not possible due to the small band gap of silicon. Unfortunately, silicon possesses low catalytic activity for both the hydrogen evolution half-reaction and the bromide oxidation half-reaction. In the past, the electrocatalysis of silicon has been aided by using Pt/Ir alloys to act as both a protective and electrocatalytic layer. Herein, efforts are detailed to replace these precious metals, where possible, by using only earth-abundant materials to decrease the cost of a module. Our hope is that efforts along this path will aid the field of artificial photosynthesis as a whole.

We begin by further testing a chemical insight previously noted within our group and discover a surprisingly high activity electrocatalyst for the hydrogen evolution reaction by cobalt phosphide (CoP) nanoparticles, detailed in Chapter 3. Falling on a traditional technique of increasing the surface area of particular facets, we nanostructured our crystalline CoP to increase its surface area of exposed (111) facets and hoped it would increase our catalytic activity; however, we found that simple structuring resulted in poor adhesion of nanostructures and poorer activity than our multi-faceted CoP nanocrystals (see the appendix to find out more). Our original catalysis efforts spurred a flurry of activity in the literature, and consequently, alternative devices that are more scalable arose. We detail the developments occurring since our work in the last appendix.

Now, with a potential catalyst in hand, comes the difficulty of balancing the delicate interplay between light absorption and catalysis, as detailed in Chapter 4. While CoP is active for HER, our particles possess a relatively low turnover frequency compared to hydrogenase or platinum, and thus require high mass loadings of material (2 mg/cm²) to obtain competitive extrinsic performance. Planar electrodes are incompatible with our particles because of substantial light absorption by the thick catalyst overlayer. By structuring our photoelectrode, we abnegate our catalyst limitations by exploiting the properties of microwires. High-aspect ratio microwires have shown promise as potentially low-cost materials for future photovoltaic applications as well as photocathodes functioning as part of an energy storage device. We discuss how to integrate our materials with silicon microwires (the wires were grown by an unscalable process to serve in place of functional CVD wires with radial emitters) to prototype a candidate photocathode. While a parasitic resistance limited the overall efficiency of the photocathode candidate, it still had promising stability. The parasitic resistance was addressed by electrodepositing the cobalt phosphide, thereby giving us a promising efficiency limited by the quality of the p-n junction.

While high-catalytic activity for the HER in acidic solutions using earth-abundant materials represents a significant advance, the photocathode is just one component of what is necessary for a complex system of splitting hydrobromic acid. Silicon, by its virtue of being a small band gap material, is easily passivated in aqueous solutions by the formation of a silicon oxide. In the past, our colleagues had shown that a monolayer of graphene could occasionally provide protection in a test solution, but batch-to-batch variability provided a considerable challenge. The putative hypothesis offered for the degradation argued defects in the crystalline graphene at grain-boundaries were the culprit. In Chapter 5 we present a method to passivate defects in the graphene crystal by light fluorination and observe a considerable enhancement in stability relative to typical graphene-protected silicon photoanodes. We had hoped that catalysis for bromide oxidation would be aided by the near-perfect graphene liquid junction, but electrodeposited Pt was required to effect photoxidation. A cursory stability test shows promising stability for one-half of an hour, but we would like to avoid using Pt. Finally, we also turned our attention to protecting silicon surfaces from oxidation by exploiting covalent silicon surface chemistry, accessible via a two-step chlorination/alkylation procedure, and explored the deposition of potentially protective thin-film metal oxides (see the appendix).

Thesis Files