This Artificial Leaf Makes Clean Hydrogen From Dirty Seawater

Subtitle: An NTU Singapore “artificial leaf” turns sunlight into hydrogen from seawater while destroying hydrazine — a dual-duty photoelectrochemical trick published in Nature Communications and amplified in mid-September 2026.

Most green-hydrogen pitches start with a confession: you need clean water, or you need a lot of electricity, or you need both. Seawater is free and endless, but chloride ions wreck electrodes and spawn corrosive chlorine chemistry. Freshwater electrolysis works — and competes with drinking water. Solar panels plus grid electrolyzers work — and pile on capital cost.

On September 16, 2026, Nanyang Technological University (NTU) Singapore highlighted a different bet: a leaf-inspired device that harvests sunlight, splits contaminated seawater into hydrogen, and chemically dismantles hydrazine — a toxic industrial pollutant — in the same light-driven circuit. No external power pack. Hydrogen shows up at both electrodes. The peer-reviewed core appeared earlier in Nature Communications (May 2026); the institutional and trade press wave hit mid-September.

Lead researcher Lydia H. Wong of NTU’s School of Materials Science and Engineering frames it as solving two problems at once. Independent commenter James Durrant of the University of Oxford, who was not on the paper, called the dual-purpose design a practical route for solar-fuels systems where economic value and environmental impact have to travel together.

Why seawater usually fights back

Classic water electrolysis needs a voltage high enough to run the oxygen-evolution reaction at the anode while hydrogen forms at the cathode. In seawater, chloride ions muscle in. Efficiency drops. Toxic and corrosive chlorine compounds form. Electrodes die young.

Hydrazine (N₂H₄) is a nasty industrial contaminant — rocket-fuel chemistry, polymer and pharmaceutical waste streams — and a known carcinogen at tiny doses. The U.S. EPA’s drinking-water health advisory sits around 10 parts per billion. It is also, chemically, a gift: oxidizing hydrazine to nitrogen (and producing extra hydrogen) takes far less voltage than oxygen evolution. Swap the hard anode reaction for hydrazine oxidation, and the energy budget of a solar-driven cell suddenly looks friendlier — while the pollutant disappears.

That is the NTU design logic: treat hydrazine not as an inconvenience to remove first, but as the anode fuel that unlocks bias-free operation in salty water.

Anatomy of the artificial leaf

The self-powered “artificial leaf” couples two halves:

Photocathode: a lead-halide perovskite that absorbs sunlight and generates electrons for the hydrogen-evolution reaction. NTU coated it with a conductive epoxy loaded with silver and copper particles, plus titanium foil, to keep the moisture-sensitive semiconductor alive in alkaline seawater.
Anode: a noble-metal-free spinel oxide, Fe₀.₃Co₀.₉Cr₁.₈O₄ (iron–cobalt–chromium), on a conductive support. It catalyzes hydrazine oxidation to nitrogen and hydrogen, resists corrosion, and — crucially — suppresses chlorine chemistry even when salt is present.

First author Mahmoud G. Ahmed and colleagues report a photocurrent density of 25 mA cm⁻² under one-sun illumination — among the highest figures cited for lead-based perovskite photocathodes in this class. NTU’s release adds a hydrogen generation rate of 466 μmol cm⁻² h⁻¹, in the same ballpark as comparable solar-driven devices, with stable operation for more than 72 hours (the paper abstract says stability for three days under 1-sun).

When the team scaled the leaf, residual hydrazine fell from a brutal starting concentration of 0.5 M (about 1.6% by weight) to below 1 ppb in roughly 30 hours of zero-bias operation. NTU’s public numbers put the floor at 0.5 ppb — more than twenty times under the EPA’s 10 ppb advisory. Colorimetric assays tracked the cleanup; gas chromatography confirmed hydrogen and nitrogen products. Hydrogen forms on both sides of the circuit when light hits the leaf.

Collaborators named in NTU’s September feature include Stener Lie, Anupam Sadhu, and Hao Zhe Chun. NTU has filed a Singapore provisional patent covering the approach.

What this is — and what it is not

It is not a shipping container of green hydrogen, a municipal water plant, or proof that every coastal city can plug seawater into a leaf and retire fossil fuels tomorrow. The demonstration still leans on hydrazine-spiked seawater as the working electrolyte. Lead-halide perovskites remain a stability and toxicity engineering problem at scale. Upscaling beyond lab areas, handling real industrial waste matrices (oils, metals, mixed organics), and integrating gas separation and safety systems are open work. The authors themselves say they are already hunting catalysts that attack other pollutants and turn waste into fuels or chemicals.

It is peer-reviewed evidence that a bias-free perovskite photocathode plus a cheap oxide anode can co-produce solar hydrogen and detoxify an extreme hydrazine load in alkaline seawater without external electricity — while dodging the chloride oxidation trap that usually kills seawater electrolysis demos. That combination is rarer than another incremental efficiency bump on pure freshwater cells.

From metaphor to multifunctional hardware

“Artificial leaf” is not a marketing invention of 2026. Daniel Nocera and others spent years pushing photoelectrochemical water-splitting toward something that looks, functionally, like a leaf: light in, fuels out, no black-box grid in between. Most descendants still optimize a single job — make hydrogen, or make oxygen, or hit a higher photocurrent on idealized electrolytes.

The NTU leaf’s twist is multifunction under constraint. Alkaline seawater. Chloride present. A pollutant concentrated enough to be a hazmat story. Zero applied bias. Noble-metal-free anode. Hydrogen stoichiometry on both sides. That is a different product thesis than shaving another half-percent off a freshwater tandem cell in a glovebox.

Trade coverage in pv magazine’s Hydrogen Stream (September 18, 2026) slotted the result next to commercial electrolyzer deployments and airport refueling trials — a reminder that lab leaves and port hardware now share the same news beat, even when readiness levels are years apart. The honest gap remains: this device proves a chemistry and a device architecture, not a plant. The useful claim for public readers is narrower and stronger — sunlight can pay for detox and fuel in one seawater circuit when the anode reaction is chosen on purpose.

Why the dual duty matters for climate tech

Green hydrogen’s sales pitch is clean molecules for steel, shipping, chemicals, and long-duration storage. Its quiet tax is water and power. Devices that earn a second paycheck — destroying a regulated toxin while making fuel — change the economics of where you site solar-fuels hardware. Wastewater lagoons, industrial outfalls, and coastal process plants suddenly look less like liabilities and more like feedstock.

Durrant’s Oxford line lands the strategic point: multifunctional photoelectrochemical systems, not single-task leaves, are likely to matter when deployment math has to clear both carbon and cash. Wong’s team puts the same idea in plainer language — energy problem and pollution problem, one sunlight-powered circuit.

None of that ships as a product on September 20, 2026. What shipped is a credible lab leaf with hard numbers: 25 mA cm⁻², >72 hours, ~30 hours to sub-ppb hydrazine, hydrogen on both electrodes, seawater that does not have to be desalinated first.

An artificial leaf that drinks dirty water and exhales clean fuel is still a prototype. It is also a sharper story than “we made slightly better hydrogen again.”

Sources

Ahmed, M. G., Wong, L. H., et al. (2026). Self-powered artificial leaf using perovskite photocathode for solar hydrogen production and hydrazine degradation. Nature Communications. https://doi.org/10.1038/s41467-026-73326-3
Nanyang Technological University Singapore. (2026, September 16). “Artificial leaf” generates clean hydrogen from contaminated seawater. https://www.ntu.edu.sg/research/research-hub/news/detail/artificial-leaf--generates-clean-hydrogen-from-contaminated-seawater
Tech Xplore / NTU. (2026, September 17). 'Artificial leaf' generates clean hydrogen from contaminated seawater. https://techxplore.com/news/2026-09-artificial-leaf-generates-hydrogen-contaminated.html
Matalucci, S. / pv magazine. (2026, September 18). The Hydrogen Stream: Singapore researchers present artificial leaf for H2 production from contaminated seawater. https://www.pv-magazine.com/2026/09/18/the-hydrogen-stream-singapore-researchers-present-artificial-leaf-for-h2-production-from-contaminated-seawater/
PubMed abstract (PMID 42161985) for DOI 10.1038/s41467-026-73326-3. https://pubmed.ncbi.nlm.nih.gov/42161985/