They Made Ammonia From Air, Seawater, Iron Salts, and a UV Lamp

Subtitle: Université de Bordeaux chemists report in Nature Chemistry an artificial photosynthetic route that couples nitrogen reduction to water oxidation — quantum yield 5% at room temperature and pressure, 10.8% under pressurized N₂, and it still works in unpurified seawater or ambient air.

Modern civilization runs on a reaction most people never think about. Roughly half of the nitrogen in a human body, on average, has passed through an industrial ammonia plant. The Haber–Bosch process — high temperature, high pressure, fossil-derived hydrogen — feeds the world. It also burns a large share of global industrial energy and, by the paper’s own framing, accounts for about 1.8% of greenhouse-gas emissions. For a century the trade-off was accepted because the alternative was hunger.

On August 25, 2026, Nature Chemistry published a different kind of answer. Doha Zidouhia, Thierry Tassaing, Nicolas Penin, Etienne Grau, and Dario M. Bassani at Université de Bordeaux / CNRS report light-driven production of ammonia from dinitrogen in water at room temperature and pressure, using inexpensive iron(II) salts, ultraviolet-C light, and a molybdenum catalyst — and they show the same chemistry still runs in unpurified seawater or under ambient air (DOI 10.1038/s41557-026-02218-2). A News & Views piece by Jens Uhlig followed on September 17, 2026, framing the result as a promising new pathway for one of chemistry’s most important reactions.

This is not another “hydrogen from seawater” leaf story. The molecule that matters here is NH₃ — fertilizer feedstock, chemical building block, and a candidate hydrogen carrier — made by fixing atmospheric nitrogen while oxidizing water.

What “photosynthetic” means in this paper

Many lab recipes for photochemical ammonia are photocatalytic in a thermodynamic sense: they dump energetic sacrificial reductants (amines, alcohols, exotic metal reagents) downhill into ammonia and call it a day. The Bordeaux team’s claim, sharpened in their related Coordination Chemistry Reviews essay, is that this system is photosynthetic: the net chemistry is endergonic by 329.2 kJ mol⁻¹. Light does real thermodynamic work. Water is the terminal electron donor. Nitrogen is reduced. Oxygen is the stoichiometric counterpart of water oxidation.

The measured headline efficiencies:

Photochemical quantum yield 5.0% for ammonia formation in water at normal temperature and pressure.
10.8% quantum yield once they push dinitrogen pressure to overcome nitrogen’s miserable solubility in water.
7.6% calculated light-to-chemical energy conversion efficiency for the endergonic N₂ + H₂O → NH₃ + O₂ stoichiometry.

Those are not Haber–Bosch plant numbers. They are laboratory quantum yields that put a homogeneous, aqueous, water-coupled nitrogen-fixation photochemistry into a range the field has chased for years without needing sacrificial SmI₂ or sodium amalgams as the electron bank.

The iron trick

The architecture is almost provocatively simple. Under 254-nm UVC, aqueous iron(II) salts generate solvated electrons. Those electrons reduce a molybdenum catalyst of the Nishibayashi family (a Mo–PCP pincer complex), which then does the hard work of activating N₂ and stepping it toward ammonia. The oxidized iron(III) does not sit as waste: it also absorbs ultraviolet light, returns to Fe(II), and in the process drives water oxidation. Iron is both the photoelectron source and the shuttle that couples the two half-reactions.

Isotope experiments with ¹⁵N₂ and H₂¹⁸O, plus transient spectroscopy, support the net picture: water is oxidized; nitrogen is reduced; selectivity for N₂ over proton reduction is unexpectedly good because solvated electrons are quenched by the Mo catalyst on a diffusion-controlled timescale. First-author communication around publication noted that under ambient-air conditions the setup could reach on the order of 0.28 g L⁻¹ ammonia after 19 hours — a lab concentration, not a warehouse, but a concrete “room air + lamp” number.

Salt and oxygen do not kill the reaction. That is why natural seawater and ambient air appear in the abstract rather than as afterthoughts. Haber–Bosch plants demand purified feeds and megawatt thermal envelopes. This flask asks for light, iron, a Mo catalyst, and wet nitrogen.

What this is not

It is not a drop-in replacement for Haber–Bosch. Global ammonia production is measured in hundreds of millions of tonnes per year; this is a peer-reviewed demonstration of a photosynthetic mechanism and quantum yields under controlled lamps. It is not sunlight-ready in the form reported: the photochemistry leans on UVC (254 nm), not the visible spectrum that hits a farmer’s field. Scaling will need either UVC engineering (already common in water treatment) or a redesign that moves the iron photochemistry toward longer wavelengths. It is not catalyst-free — molybdenum still sits at the nitrogen-activation heart — and pressurized N₂ helps because chemistry cannot invent dissolved nitrogen that water refuses to hold.

Hold those caveats in the same breath as the shareable line. The shareable line is still extraordinary: room-temperature, aqueous, water-coupled artificial photosynthetic nitrogen fixation that tolerates seawater and air, with a 5% quantum yield at 1 bar and 10.8% when solubility is forced.

Why Sunday readers should care

Fertilizer chemistry is planetary infrastructure. Every percentage point of fossil ammonia that can eventually be displaced by light-driven, water-coupled routes is a climate and food-security story at once. Prior homogeneous Mo systems often leaned on harsh sacrificial reductants; semiconductor photocatalysts often posted tiny apparent quantum yields. Bordeaux’s Fe/Mo photochemistry occupies a middle lane that looks chemically honest: iron for electrons, molybdenum for N₂, water for both protons and the oxidative half, light for the thermodynamic lift.

Uhlig’s News & Views puts the aspiration cleanly — iron salts and a molybdenum catalyst driving photosynthetic reduction of atmospheric nitrogen to ammonia with water as the terminal electron donor. The next experiments write themselves: longer-wavelength iron photosensitizers, continuous-flow reactors that recycle Mo, independent labs checking seawater and air runs, and an honest life-cycle comparison against green-hydrogen Haber–Bosch. None of that is promised by an August paper. All of it is unlocked by one.

A century ago, Haber and Bosch taught humanity to squeeze nitrogen with heat and pressure. Zidouhia, Grau, Bassani, and colleagues just showed that under a UVC lamp, iron and molybdenum can ask nitrogen the same question more gently — in water, even seawater, even from the air in the room — and get ammonia back with a quantum yield you can put in a headline without inventing a factory.

No fossil hydrogen cylinder. No 400 °C pressure vessel in the story’s core. Air or seawater. Iron salts. A Mo catalyst. A UV lamp. Five percent quantum yield at ambient pressure; ten-point-eight when they squeeze more N₂ into solution. Not the end of Haber–Bosch. A peer-reviewed beginning for photosynthetic ammonia that finally couples the hard half-reaction to water oxidation in a flask Sunday science can explain.

Sources

Zidouhia, D., Tassaing, T., Penin, N., Grau, E. & Bassani, D. M. (2026). Artificial photosynthetic reduction of nitrogen to ammonia at room temperature using natural seawater. Nature Chemistry. Published online 25 August 2026. https://doi.org/10.1038/s41557-026-02218-2
Uhlig, J. (2026, September 17). Forging ammonia with iron and light from water and nitrogen. Nature Chemistry (News & Views). https://doi.org/10.1038/s41557-026-02253-z
Zidouhia, D., Grau, E. & Bassani, D. M. (2026). Photocatalytic vs. photosynthetic reduction of nitrogen into ammonia: The importance of the sacrificial reducing agent. Coordination Chemistry Reviews, 562, 217875. https://doi.org/10.1016/j.ccr.2026.217875
Zidouhia, D. et al. (2026). Dataset for artificial photosynthetic reduction of nitrogen to ammonia at room-temperature using natural seawater. Zenodo. https://doi.org/10.5281/zenodo.20489299