One Anion Swap Turns Copper Oxide Into a Better Solar Water-Splitting Co-Catalyst

Swap nitrate for sulfate in the electroplating bath — nothing else changes — and a solar water-splitting photoanode’s photocurrent jumps 80%.

Photoelectrochemical (PEC) water splitting is one of the more direct routes to green hydrogen: sunlight hits a semiconductor photoanode, generates charge carriers, and drives the oxygen evolution reaction (OER) that splits water. Bismuth vanadate (BiVO₄) is a favorite light absorber for this job — good bandgap, reasonably stable — but on its own it suffers from heavy electron-hole recombination and sluggish OER kinetics. It needs a co-catalyst to actually perform.

Copper oxide (CuO) is an attractive candidate: earth-abundant, non-toxic, cheap. But making CuO in the nanostructure that actually helps — high surface area, well-defined porosity — has usually meant a trade-off. Solvothermal synthesis in toxic DMF can take 20–40 hours. Thermal evaporation of copper foil needs temperatures up to 900 °C. Template-assisted routes need surfactants that have to be removed later. None of that is friendly to scale-up.

The Obstacle

Electrodeposition has always been the economically sensible alternative — fast, low-energy, room temperature. The problem is that most prior work on electrodeposited CuO focused on tuning the metal cation and deposition voltage, while largely overlooking a second lever sitting in the same bath: the anion. Nobody had systematically shown that the anion alone — same copper source concentration, same deposition conditions, same everything else — could be the difference between a high-surface-area porous nanostructure and a dense, poorly crystalline one.

The Innovation

Truong-Giang Vo, Shu-Ju Chang, and Chia-Ying Chiang (Department of Chemical Engineering, National Taiwan University of Science and Technology) closed that gap in Dalton Transactions (Vo et al., 2020, DOI: 10.1039/c9dt04626g).

They electrodeposited copper layered hydroxide salt (Cu-LHS) precursors directly onto a BiVO₄ photoanode from two otherwise-identical baths — one using Cu(NO₃)₂/NaNO₃, the other CuSO₄/Na₂SO₄ — then calcined both at 400 °C for 1 hour to convert the hydroxide precursor into CuO. The measured ionic strength of the two baths (0.16 mol/L for nitrate vs. 0.38 mol/L for sulfate) drives genuinely different nucleation kinetics: the nitrate bath produces plentiful small nuclei that grow into thin nanosheets (~1 µm long, 10–20 nm wide), while the sulfate bath produces fewer, larger nuclei that grow into thicker, irregular nanoplates. On calcination, that difference in the precursor carries straight through to the final oxide.

Nitrate produces thin nanosheets; sulfate produces irregular, less homogeneous nanoplates as seen in the SEM images above: Cu-LHS/N (a), Cu-LHS/S (b), CuO-N (c), and CuO-S (d).

Key Results

The anion used in the plating bath alone decides whether the resulting CuO cocatalyst turns out porous and well-crystallized or dense and poorly ordered, and that structural difference cascades through every measurement that follows: cleaner Raman-confirmed phase conversion, a photocurrent that very nearly doubles, faster charge separation timed directly by MRID, and higher charge-injection and Faradaic efficiencies.

Only the Nitrate Route Crystallizes into a Well-Ordered, High-Surface-Area CuO

BVO/CuO-N forms interconnected 20–30 nm nanoparticles with clean (002)-plane lattice fringes at a 2.52 Å spacing — well crystallized. BVO/CuO-S keeps its sheet-like shape with only a few tiny pores and no resolvable lattice fringes at all. That gap shows up in surface area too: double-layer capacitance puts CuO-N’s electrochemically active surface area at 10 mF/cm², versus 7 mF/cm² for CuO-S.

MRID Raman (532 nm) Confirms Clean, Phase-Pure Conversion

Raman spectra (150–1000 cm^-1) of both BVO/CuO-S and BVO/CuO-N show the expected scheelite-monoclinic BiVO₄ bands (211.8, 322.9, 367.8, 710.4, 825 cm^-1) plus one new peak at 285.9 cm^-1 — the CuO vibrational mode — absent from bare BVO. Neither sample shows any residual Cu-LHS or Cu(OH)₂ bands, confirming complete, clean conversion to CuO on both routes, consistent with the XRD pattern in the same figure.

The Porous CuO-N Cocatalyst Nearly Doubles Photocurrent

At 1.23 V vs. RHE, BVO/CuO-N reached 0.9 mA/cm² (an 80% increase over bare BiVO₄) with a 280 mV cathodic shift in onset potential, versus 0.7 mA/cm² (a 40% increase) and a 180 mV shift for BVO/CuO-S — porosity, not just the presence of CuO, is doing the work.

That faster charge separation is quantitative too: MRID’s time-resolved PL measured the average excited-state lifetime shortening from 0.21 ns (bare BVO) to 0.17 ns (BVO/CuO-N).

Charge Injection Nearly Doubles, Faradaic Efficiency Climbs to 95%

Hole injection efficiency for BVO/CuO-N reached 65% at 1.23 V vs. RHE — almost double the bare BiVO₄ electrode. Overall O₂-evolution Faradaic efficiency rose from 84% (bare BVO) to 95% (BVO/CuO-N), confirmed by gas chromatography.

Every one of these measurements traces back to the same root cause: the nitrate route’s open, porous CuO nanostructure gives photogenerated charges more surface area to reach and more direct pathways to get there, while the sulfate route’s denser sheets fall short on nearly every metric that matters for solar water splitting.

Why ProTrusTech Is Featuring This Research

The Experimental Section names MRID twice, for two different jobs, and that’s exactly the case study worth highlighting. First: “Raman spectroscopy was performed on a confocal Raman microscope system (MRID, Protrustech, Taiwan). A solid-state laser (λ = 532 nm) was used as the excitation source.” Second: “All photoluminescence (PL) and time-resolved photoluminescence (TR-PL) lifetime measurements were performed using an integrated spectrometer (MRID, ProTrusTech Co., Taiwan) with a 405 nm laser as the excitation source.”

That’s the two-laser, dual-modality workflow MRID is built around, running on real catalyst-screening science rather than a showcase 2D material:

– Raman answers “did the conversion actually happen cleanly?” (structure),

– and PL/TRPL answers “does it actually change how charge carriers behave?” (function).

Neither question alone tells the full story — a phase-pure CuO that doesn’t improve charge separation would be a synthesis footnote, not a co-catalyst. Together, on one platform, they explain why the photocurrent and Faradaic efficiency numbers moved the way they did.

PL emission (a) shows CuO-N slightly quenches BVO’s characteristic 530/575 nm emission, indicating suppressed electron-hole recombination.

Time-resolved PL (b) confirms it quantitatively: the average excited-state lifetime shortens from 0.21 ns (bare BVO) to 0.17 ns (BVO/CuO-N), measured via time-correlated single-photon counting at the 530 nm peak.

Worth noting for anyone measuring beam-sensitive catalyst films: the authors kept Raman exposure to just 3 seconds with 10 accumulations specifically “to prevent degradation” of the sample. Fast acquisition on the same platform that also runs the PL/TRPL side matters when the material itself can’t tolerate a long laser dwell.

For the academic researcher, this is a template for connecting a synthesis variable (anion identity) all the way through to charge-carrier dynamics without switching instruments or samples between measurements.

For the industrial engineer evaluating earth-abundant co-catalysts for solar fuels, it’s a demonstration that Raman-verified phase purity and PL-verified charge dynamics can both be checked, on the same bench, before committing to a scaled-up electrode.

Engineering the CuO Co-Catalyst Solution

The manufacturability story here is arguably as strong as the science. This isn’t a lab curiosity that only works with exotic reagents — it’s the same electroplating bath, the same voltage, the same temperature, with one salt swapped for another. Cathodic electrodeposition at −0.1 V vs. Ag/AgCl, room temperature, then a single 400 °C calcination step. No surfactants, no toxic solvents, no week-long solvothermal runs, no vacuum thermal evaporation. That’s a single, cheap, easily-controlled knob — anion identity — for tuning a co-catalyst’s porosity and performance, which is exactly the kind of variable a production process can standardize around.

Featured Instruments:

Why This Solar Water-Splitting Co-Catalyst Strategy Matters

The generalizable principle isn’t really about CuO specifically — it’s that anion chemistry, not just cation chemistry, is an underused lever in electrodeposited nanostructure design. The paper’s own introduction notes that the layered hydroxide salt structure this strategy relies on isn’t unique to copper — it’s a general M²⁺(OH)_2-x(Aᵐ^–)_x/_m·nH₂O framework that also forms with Mg²⁺, Ni²⁺, Cd²⁺, and Co²⁺. That suggests the same anion-controlled morphology strategy could extend to other earth-abundant transition-metal oxide co-catalysts for solar fuel applications, well beyond this one CuO/BiVO₄ system.

What’s Next

• Hole injection still has headroom. The authors’ own data shows injection efficiency reaching 65% for BVO/CuO-N — a large improvement over bare BVO, but the paper explicitly notes “large room for further enhancing hole injection efficiency” remains.
• Extension to other layered hydroxide salt systems. The same anion-controlled electrodeposition-then-calcination strategy is a plausible fit for Ni-, Co-, or Cd-based layered hydroxide salts used as co-catalysts elsewhere in the PEC and OER literature.
• Mechanistic confirmation of the Cu(II)/Cu(III)-oxo cycle. The proposed OER mechanism (surface Cu(II) oxidized to a highly valent Cu(III)-O active species) draws on in-operando Raman work from other groups — a natural follow-up would be in-operando Raman on this specific CuO-N system to confirm the same active species forms during operation.

Key Takeaways

• Swapping only the anion in an otherwise identical electroplating bath (nitrate vs. sulfate) controlled whether the calcined CuO product was a porous, high-surface-area nanosheet or a denser, poorly crystalline nanoflake.
• ProTrusTech’s MRID ran both halves of the characterization: 532 nm Raman confirmed phase-pure conversion of the hydroxide precursor to CuO, and 405 nm PL/TRPL quantified the resulting charge-carrier lifetime shortening (209.8 ps → 170.9 ps).
• The porous, nitrate-derived CuO co-catalyst delivered an 80% photocurrent increase, a 280 mV cathodic onset-potential shift, and lifted O₂ Faradaic efficiency from 84% to 95% versus bare BiVO₄.
• The entire fabrication route — cathodic electrodeposition plus one calcination step — uses no toxic solvents, no exotic temperatures, and no templates, making it a genuinely scalable co-catalyst manufacturing strategy.

Featured Publications: