Finding new wide-bandgap light absorbers that are stable in aqueous solutions is a long-standing challenge in photoelectrochemical water splitting research. Two papers in this issue describe recent advances in high throughput experimentation that may accelerate the discovery of suitable materials. Finding new wide-bandgap light absorbers that are stable in aqueous solutions is a long-standing challenge in photoelectrochemical water splitting research. Two papers in this issue describe recent advances in high throughput experimentation that may accelerate the discovery of suitable materials. Semiconducting materials that are able to absorb sunlight and use this energy to directly drive an electrochemical reaction at their surface have fascinated researchers for decades. Most photoelectrochemical (PEC) studies have focused on water splitting, a thermodynamically up-hill reaction that could offer an attractive pathway for the long-term capture and storage of solar energy by producing ‘green’ hydrogen. Other useful PEC reactions are the oxidation of harmful pollutants or even the reduction of CO2 to carbon monoxide or hydrocarbons. To drive the electrochemical reaction of interest, photovoltages of at least 1.5 V are needed. Such high voltages require tandem systems in which at least two absorbers are connected in series. While silicon is usually considered as the optimal bottom absorber, finding an efficient and chemically stable top absorber with a bandgap of around 1.5 – 2.0 eV has been a long-standing and largely unsolved challenge.1van de Krol R. Parkinson B.A. Perspectives on the photoelectrochemical storage of solar energy.MRS Energy & Sustainability. 2017; 4: E13Crossref Google Scholar The search for suitable photoanode materials has mostly concentrated on metal oxides, one the few materials classes that shows good chemical stability under the highly oxidizing conditions that enable water splitting. Metal oxides are easy to synthesize and process compared to classic semiconductors like Si and GaAs, but those that absorb visible light tend to suffer from poor charge transport properties due to the formation of polarons, excitons, or defects. The good news is that only a tiny fraction of the ~19,000 possible ternary oxides and > 200,000 quaternary oxides has been studied thus far,2Woodhouse M. Parkinson B.A. Combinatorial approaches for the identification and optimization of oxide semiconductors for efficient solar photoelectrolysis.Chem. Soc. Rev. 2009; 38: 197-210Crossref PubMed Google Scholar which makes it likely that the best materials are still waiting to be discovered. The bad news is that there are currently no robust and proven strategies to identify promising candidates. The calculation of polaronic and midgap defect states is still too computationally expensive for in silico materials design with high throughput electronic structure calculations. As a result, the search for novel photoelectrode materials has thus far mostly relied on chemical intuition and serendipitous discoveries. Two papers in this issue report advances in the discovery and optimization of metal oxide photoanodes using high throughput experimentation. Both papers are from the high throughput team at the Joint Center for Artificial Photosynthesis (JCAP) in the USA. In the first paper, Zhou et al. have studied a family of copper vanadate-based quaternary compositions, Cu-V-M-O, where M = Mg, Ca, Sr, or Fe.3Zhou L. et al.Quaternary Oxide Photoanode Discovery Improves the Spectral Response and Photovoltage of Copper Vanadates.Matter. 2020; (this issue): 1614-1630Abstract Full Text Full Text PDF Scopus (5) Google Scholar Copper vanadates (CVOs) have attracted interest because they exhibit narrow bandgaps as well as good photochemical stability, a combination that is relatively rare for n-type semiconductors. Thin film gradient composition libraries were made with reactive co-sputtering and the structural, optical, and photoelectrochemical properties of many hundreds of distinct compositions were measured. Especially noteworthy are the extensive photoelectrochemical measurements; using a scanning droplet cell, no less than four chopped-light voltammograms, each with a different illumination wavelength, were measured for each of the 840 samples in the library. This dataset is one of the largest ever made in PEC studies and allowed the authors to identify several general trends. One is that alloying with divalent alkaline earth elements on the Cu sites improves the PEC properties of CVOs, and another is that V-rich compositions showed relatively high photocurrents at low photon energies. Furthermore, a large spread in the generated photovoltage was found for samples that show the highest photocurrents. This is important because while most PEC studies tend to emphasize photocurrent as a performance metric, it is the internally generated photovoltage that determines the available driving force for the electrochemical reactions. Sr-alloyed Cu5V2O10 showed the highest performance in this study, and six promising new quaternary compositions were identified that warrant follow-up studies. In the second paper, the JCAP team switched from high throughput photoanode discovery to high throughput modification of a single photoanode material, β-Cu2V2O7.4Newhouse P.F. et al.Enhanced Bulk Transport in Copper Vanadate Photoanodes Identified by Combinatorial Alloying.Matter. 2020; (this issue): 1601-1613Abstract Full Text Full Text PDF Scopus (4) Google Scholar Cu2V2O7 shows good visible light absorption starting at photon energies of 2 eV, but this does not result in good photoactivity at lower energies due to the formation of excitons.5Wiktor J. Reshetnyak I. Strach M. Scarongella M. Buonsanti R. Pasquarello A. Sizable Excitonic Effects Undermining the Photocatalytic Efficiency of β-Cu2V2O7.J. Phys. Chem. Lett. 2018; 9: 5698-5703Crossref PubMed Scopus (20) Google Scholar Newhouse et al. attempted to disrupt the formation of excitons by combinatorial (co)doping (or rather alloying) of β-Cu2V2O7. Using inkjet printing, they fabricated combinatorial libraries of Cu-rich, Cu-poor, and stoichiometric Cu2V2O7 that were co-alloyed with six elements (Ca, Gd, Hf, La, Lu, Zr) and their pairwise combinations. With concentrations ranging between 0 and 7.3%, more than 1800 different compositions were synthesized. They found that Ca appears to enhance to oxygen evolution reaction and that the combination of Ca with either Hf, Zr, or La leads to a 2.7-fold increase in photoactivity. One subtle, yet intriguing finding is that the optimal ratio of co-alloying elements, i.e., z in Cu1-xVxOδ:(Ca1-zBz)y, where B is La, Hf, or Zr, shifts to higher values if the base composition becomes increasingly Cu-poor (i.e., at higher values of x). This trend would have been nearly impossible to identify in a regular PEC study and illustrates the power of the high throughput approach. The latter is even more clearly illustrated by Figure 1, which reveals that while it is indeed possible to improve the transport properties (evidenced by the increase in generated power, Pmax), it comes at the cost of reduced optical absorption at lower photon energies. The first combinatorial studies on photoelectrode materials date back to 2002 with the work of McFarland et al.,6Baeck S.H. Jaramillo T.F. Brändli C. McFarland E.W. Combinatorial electrochemical synthesis and characterization of tungsten-based mixed-metal oxides.J. Comb. Chem. 2002; 4: 563-568Crossref PubMed Scopus (104) Google Scholar followed a few years later by Parkinson (2Woodhouse M. Parkinson B.A. Combinatorial approaches for the identification and optimization of oxide semiconductors for efficient solar photoelectrolysis.Chem. Soc. Rev. 2009; 38: 197-210Crossref PubMed Google Scholar and refs. therein) and several other groups. Most of these early studies report modifications of established photoelectrode materials and focus mainly on establishing high throughput methods based on electrodeposition or variations of ink jet printing. Later efforts by e.g., Ludwig et al.7Vidyarthi V.S. et al.Enhanced photoelectrochemical properties of WO3 thin films fabricated by reactive magnetron sputtering.Int. J. Hydrogen Energy. 2011; 36: 4724-4731Crossref Scopus (81) Google Scholar extended this to vacuum-based deposition techniques, such as sputtering. One of the first examples of a truly new photoelectrode material was p-type CuBi2O4, a 1.8 eV bandgap material reported by the group of Sayama.8Arai T. Konishi Y. Iwasaki Y. Sugihara H. Sayama K. High-throughput screening using porous photoelectrode for the development of visible-light-responsive semiconductors.J. Comb. Chem. 2007; 9: 574-581Crossref PubMed Scopus (118) Google Scholar The work of the high-throughput team at JCAP, headed by John Gregoire, probably represents the largest concerted effort in the field. Over the past decade, they have built up an impressive infrastructure for high throughput synthesis and characterization of photoelectrodes as well as electrocatalysts. High throughput studies are not only demanding in terms of the required research infrastructure, the design of such studies and the analysis of the data also comes with many challenges. Every high throughput experiment needs a good starting point, and the JCAP team have established an effective research pipeline in which promising starting points are provided by high throughput theory.9Yan Q. Yu J. Suram S.K. Zhou L. Shinde A. Newhouse P.F. Chen W. Li G. Persson K.A. Gregoire J.M. Neaton J.B. Solar fuels photoanode materials discovery by integrating high-throughput theory and experiment.Proc. Natl. Acad. Sci. USA. 2017; 114: 3040-3043Crossref PubMed Scopus (116) Google Scholar To convert large datasets into useful insights, a judicious choice of performance indicators is needed. Takeuchi already warned that combinatorial techniques are not suitable for every application,10Koinuma H. Takeuchi I. Combinatorial solid-state chemistry of inorganic materials.Nat. Mater. 2004; 3: 429-438Crossref PubMed Scopus (355) Google Scholar and it should be realized that photoelectrodes are not the easiest candidates due to the strong influence of processing conditions on the performance. Incomplete crystallization, small amounts of defects, and the presence of grain boundaries can easily lead to false negatives. Conversely, the relatively low photoactivity of most oxides can lead to false positives, since it often easy to improve a poorly performing material by making it slightly less poor. The chance of a false positive will of course decrease as the baseline performance of the library improves. The large amount of resources (time, money, personnel) that have to be invested in an effective high throughput infrastructure means that such studies are unlikely to be verified or reproduced in their entirety by other groups, as is common for other types of PEC studies. Instead, high throughput studies contribute by offering concrete suggestions for promising new compositions, such as the Sr-alloyed Cu5V2O10 and the additional six new quaternary compositions proposed by Zhou et al.3Zhou L. et al.Quaternary Oxide Photoanode Discovery Improves the Spectral Response and Photovoltage of Copper Vanadates.Matter. 2020; (this issue): 1614-1630Abstract Full Text Full Text PDF Scopus (5) Google Scholar They also contribute by revealing subtle trends that are normally obscured by the relatively large sample-to-sample variations encountered in typical PEC studies. Such trends may offer interesting avenues for further exploration, or even provide direct support for theoretical predictions (like the role of Cu 3d states in CVO3Zhou L. et al.Quaternary Oxide Photoanode Discovery Improves the Spectral Response and Photovoltage of Copper Vanadates.Matter. 2020; (this issue): 1614-1630Abstract Full Text Full Text PDF Scopus (5) Google Scholar). Last but not least, high throughput studies can tell us what does not work. An excellent example of this is Figure 1, which shows that β-Cu2V2O7 is unlikely to show high energy conversion efficiencies at low photon energies.5Wiktor J. Reshetnyak I. Strach M. Scarongella M. Buonsanti R. Pasquarello A. Sizable Excitonic Effects Undermining the Photocatalytic Efficiency of β-Cu2V2O7.J. Phys. Chem. Lett. 2018; 9: 5698-5703Crossref PubMed Scopus (20) Google Scholar The importance of reporting negative results cannot be overstated; the PEC community can no longer afford to spend decades on a single material, like we did with TiO2. With so many candidate materials and so little time left to transform our energy infrastructure, it is important to ‘fail quickly’1van de Krol R. Parkinson B.A. Perspectives on the photoelectrochemical storage of solar energy.MRS Energy & Sustainability. 2017; 4: E13Crossref Google Scholar. The recent developments in high throughput experimentation, exemplified by the two papers from the JCAP team, offer a promising path to also succeed quickly. Enhanced Bulk Transport in Copper Vanadate Photoanodes Identified by Combinatorial AlloyingNewhouse et al.MatterOctober 1, 2020In BriefPhotoanodes are materials that require tuning of many properties, and copper vanadates have shown initial promise. Alloying is a primary materials research strategy to co-optimize multi-functional performance, although predictive models for performance of complex metal oxide alloys are generally lacking. High-throughput methods can effectively explore the vast search space, as demonstrated by our exploration of 15 co-alloy composition spaces that reveal new avenues for improved visible light photoactivity in copper vanadate-based photoanodes. Full-Text PDF Open ArchiveQuaternary Oxide Photoanode Discovery Improves the Spectral Response and Photovoltage of Copper VanadatesZhou et al.MatterOctober 1, 2020In BriefWe present an accelerated photoanodes discovery in Cu-V-X-O (where X is Mg, Ca, Sr, Fe) oxide composition spaces by combing combinatorial synthesis and high-throughput photoelectrochemical screening. A set of six quaternary oxide phases is identified as photoanodes, with several exhibiting photoresponse at sub-2.4 eV illumination. Several phase mixtures in the Cu-V-X-O composition spaces exhibit simultaneous improvement against both the photon energy and the applied bias performance limitations of Cu-V-O photoanodes, demonstrating a path forward for the development of efficient metal oxide photoanodes. Full-Text PDF Open Archive