Simple Rinsing Step Balances Light Direction in Bismuth Vanadate Photoanodes

Solar hydrogen production depends on photoelectrochemical devices that can split water using nothing but sunlight and carefully engineered semiconductors. Among the materials competing to serve as the light-absorbing workhorse of such devices, bismuth vanadate, or BiVO4, has earned a reputation as one of the most promising photoanode candidates available today. Its band gap of roughly 2.4 electronvolts allows it to harvest a substantial portion of visible light, its band edge positions align well with the water oxidation reaction, and it remains chemically stable under neutral aqueous conditions. Theoretical calculations suggest that an ideal BiVO4 photoanode could generate a photocurrent density of about 7.5 milliamperes per square centimeter under standard AM 1.5G illumination, a figure that has motivated more than a decade of intensive research into doping, nanostructuring, and cocatalyst loading.

Yet a stubborn problem has persisted across nearly every BiVO4 electrode ever fabricated: the material performs very differently depending on which side the light strikes. Because the hole diffusion length in BiVO4 is shorter than 100 nanometers, photogenerated holes created far from the electrolyte interface tend to recombine before they can oxidize water. When light enters through the transparent conductive oxide substrate, a geometry known as back-side illumination, most charge carriers are born close to the substrate where electrons can be collected efficiently, and the measured photocurrent is comparatively high. When light enters directly through the semiconductor surface, front-side illumination, holes generated near the electrolyte interface must travel only a short distance, but electrons generated deeper in the film face a long, lossy journey to the substrate. The result is a pronounced asymmetry, with back illumination routinely outperforming front illumination by a wide margin, a disparity that complicates device design and wastes a portion of the absorbed solar energy.

A research team led by Sarwar Saad, Wonjung Choi, and Yiseul Park at Pukyong National University in Busan, Republic of Korea, has now demonstrated that this direction-dependent behavior can be dramatically reduced with a processing step so simple it borders on the mundane: rinsing. In work published in Advances in Industrial and Engineering Chemistry, the group showed that selectively washing away weakly adhered bismuth particles after electrodeposition produces BiVO4 thin films whose photocurrents under front and back illumination are nearly identical. The finding, published as open access on 9 December 2025, suggests that interface engineering through precursor control may be as important as the more elaborate strategies the field has traditionally pursued.

The fabrication route began with fluorine-doped tin oxide, or FTO, glass substrates that were ultrasonically cleaned in acetone, ethanol, and deionized water. Bismuth was then electrodeposited from a solution of bismuth nitrate and sodium perchlorate dissolved in dimethyl sulfoxide at 80 degrees Celsius, using a pulsed potential of minus 1.4 volts in a three-electrode configuration. The researchers prepared two sets of electrodes. In the first, the bismuth films were simply dried at 100 degrees Celsius and carried forward as deposited. In the second, the films were rinsed sequentially with ethanol and deionized water before drying, a step designed to strip away the loosely bound, overgrown bismuth branches that electrodeposition tends to leave behind. Vanadium was introduced by drop-casting a vanadyl acetylacetonate solution onto the bismuth-coated substrates, and a slow anneal in air at 450 degrees Celsius for three hours converted the stacked precursors into the characteristic yellow BiVO4 phase. A final soak in sodium hydroxide removed unreacted vanadium species.

Scanning electron microscopy revealed just how consequential the rinsing step was. The non-rinsed bismuth precursors displayed the classic signatures of diffusion-limited electrodeposition: dendritic structures at low deposition charges that coalesced into large, irregular aggregates as the passed charge increased to 300 and 500 millicoulombs. Cross-sectional imaging exposed voids at the bismuth-FTO interface, evidence of weak adhesion and uncontrolled crystal growth. After thermal conversion, these defects propagated directly into the BiVO4 layer, which emerged as a coarse, porous film riddled with interparticle voids and separated from the substrate by distinct gaps, a configuration prone to delamination during operation. The rinsed precursors told a different story. With the weakly attached branches removed, the remaining bismuth formed a compact, densely packed, and uniformly distributed layer that remained void-free at the interface even at the highest deposition charge. The converted BiVO4 films inherited this discipline, showing fine grains, tight packing, and a well-defined, gap-free boundary with the FTO substrate.

Optical measurements reinforced the structural picture. Ultraviolet-visible spectrophotometry showed that absorbance in the non-rinsed films climbed steadily with deposition charge, as thicker precursors yielded thicker, more light-hungry BiVO4 layers with a strong absorption edge near 520 nanometers. The rinsed films, by contrast, absorbed less light across the visible spectrum, reflecting their reduced thickness. Crucially, the researchers interpret this lower optical density not as a degradation but as a refinement: by thinning the film, rinsing brought its dimensions closer to the carrier diffusion length intrinsic to BiVO4, optimizing the trade-off between light harvesting and charge collection. The improved transparency of the rinsed films also carries practical appeal for tandem cell architectures, where a semi-transparent top photoanode must allow unused photons to reach an underlying absorber.

The photoelectrochemical tests delivered the headline result. Under back-side illumination in a sulfite electrolyte at pH 7, the non-rinsed BiVO4 electrodes generated photocurrents of roughly 1.5 milliamperes per square centimeter, but their front-side response was far weaker, reproducing the familiar asymmetry that has long plagued the material. The rinsed electrodes collapsed this gap to less than 1 milliampere per square centimeter, delivering nearly identical photocurrents from either direction. Moreover, at equivalent deposition charges, the rinsed films consistently outperformed their non-rinsed counterparts overall, despite absorbing less light. The team attributes this to the compact film-substrate interface formed during rinsing, which facilitates efficient electron extraction from BiVO4 into the FTO, minimizes series resistance, and suppresses the recombination pathways that porous, poorly adhered films inevitably introduce. In effect, the optimized electron transport compensates for the reduced optical thickness, and the electrode becomes genuinely direction-independent.

Durability testing added a further layer of significance. When a rinsed BiVO4 electrode was held at 1.25 volts versus the reversible hydrogen electrode under continuous AM 1.5G illumination, its photocurrent density remained essentially constant for more than five hours, with no noticeable degradation or fluctuation. The authors connect this stability directly to the mechanical and electrochemical robustness imparted by the rinsing step: strong bonding between the BiVO4 layer and the FTO substrate minimizes delamination and crack formation, while the compact, uniform microstructure suppresses photocorrosion and limits the defect sites that typically accelerate degradation during extended operation. For a technology whose commercial prospects hinge on devices surviving thousands of hours in real electrolytes, demonstrating that a simple fabrication modification can deliver both performance and longevity is a meaningful contribution.

What makes the result particularly attractive is what it does not require. Strategies to improve BiVO4 have historically involved dopant incorporation, elaborate nanostructuring, or the loading of oxygen evolution cocatalysts, each of which adds cost, complexity, and process variability. The rinsing strategy, by contrast, is a single additional washing step inserted into an existing electrodeposition workflow, requiring no exotic chemicals, no high-vacuum equipment, and no post-growth surface treatments. The authors position it as a scalable and cost-effective route to direction-independent BiVO4 photoanodes, and the underlying principle, that the morphology and adhesion of the metallic precursor dictate the quality of the final oxide film, is likely transferable to other electrodeposited photoelectrode systems. As solar fuels research matures from laboratory curiosities toward practical hydrogen production, results like this one serve as a reminder that sometimes the most powerful lever in materials engineering is not a new compound or a clever catalyst, but careful control of the interface where the material meets its substrate.

Subject of Research: Interface engineering of BiVO4 photoanodes for balanced front-back illumination performance in photoelectrochemical water splitting

Article Title: Interface-controlled BiVO4 thin films via selective rinsing for balanced front-back illumination performance

Article References: Saad, S., Choi, W., & Park, Y. (2025). Interface-controlled BiVO4 thin films via selective rinsing for balanced front-back illumination performance. Advances in Industrial and Engineering Chemistry, 1(1), Article 36. https://doi.org/10.1007/s44405-025-00034-9

Image Credits: AI Generated

DOI: 10.1007/s44405-025-00034-9

Keywords: BiVO4, photoanode, photoelectrochemical water splitting, solar hydrogen, electrodeposition, interface engineering, selective rinsing, FTO substrate, charge recombination, illumination symmetry, photostability, thin films