What if the key to making urea selectively is not a new molecule, but a cavity barely two nanometers wide? This study uses heat to sculpt those tiny spaces around single copper atoms. Urea remains the cornerstone nitrogen fertilizer, sustaining nearly half of the world’s population.

But current urea production using the Haber-Bosch and Bosch-Meiser processes accounts for 2% of global energy consumption and 1.5% of total carbon dioxide emissions. Electrosynthesis from carbon dioxide and nitrate offers a carbon-neutral alternative under ambient conditions, potentially transforming environmental pollutants into high-value chemicals.

Using renewable electricity could bypass the high energy barriers of traditional manufacturing and enable decentralized urea production directly from wastewater and flue gas. The central problem is an intrinsic kinetic mismatch among carbon dioxide reduction, nitrate reduction, and the hydrogen evolution reaction, which disrupts synchronized carbon and nitrogen intermediate generation and hinders carbon-nitrogen coupling.

Coordination-engineered single-atom catalysts reach only 28% urea Faradaic efficiency because intermediates have short residence times on unconstrained surfaces. Bimetallic copper systems raise Faradaic efficiency to 70.1–89.4%, but exposed interfaces still let intermediates diffuse away or enter competing pathways.

Existing nanoconfined catalysts improve retention, yet their polydisperse nanoclusters in 4–11 nanometer mesopores lack atomic site homogeneity and molecular-scale confinement-electronic synergy. The challenge is to combine a defined coordination environment with a size-matched nanocavity that regulates intermediate generation, enrichment, and coupling.

The strategy uses programmable zinc-node gasification in zeolitic imidazolate frameworks to create size-controlled nanoconfined cavities hosting optimized copper-nitrogen-four single-atom sites. Unlike conventional carbonization, the multi-stage programmable pyrolysis protocol forms 1.5–1.8 nanometer cavities intended to minimize the energy barrier of adsorbed molecules and suppress side reactions.

A confinement descriptor quantitatively links cavity geometry, local electric field, and electronic structure, defining a generalizable design principle for nanoconfined electrocatalysis. In panel b, the blue cavity shading changes across the temperature series, while the microscopy row shows the dodecahedral particles remaining recognizable through the first three conditions before becoming visibly collapsed in the final column.

In panel d, the rightmost AC HAADF-STEM image contains many individually resolved bright spots, distributed across the carbon framework rather than gathered into larger particles. The thermally sculpted ZIF-Cu catalysts were fabricated through a designed multi-stage pyrolysis protocol optimized using thermogravimetric analysis and differential scanning calorimetry of the ZIF-Cu precursor.

The protocol used tailored temperature ramps and isothermal holds for each target pyrolysis temperature. Below 600 degrees Celsius, solvent removal and organic-linker condensation transform the flexible framework into a rigid, nitrogen-doped carbon scaffold.

At 800 degrees Celsius, the material retains the ZIF precursor’s intrinsic 1.5 nanometer pore size, but its pore volume remains underdeveloped. Near zinc’s 905 degree Celsius boiling point, zinc species undergo reduction and gasification, acting as an internal sacrificial template and thermal chisel, while copper and nitrogen stabilize the architecture against collapse.

For tZIF-Cu900, the proportion of 1.8 nanometer micropores increases substantially as new cavities form through ligand polycondensation and graphitization. At 1000 degrees Celsius, 2.6 nanometer pores develop and disperse, while tZIF-Cu1100 undergoes excessive graphitization and framework collapse.

Electrochemical characterization found that tZIF-Cu900 achieved the highest carbon-nitrogen co-reduction current density and the most positive nitrate-reduction onset potential, with moderate hydrogen-evolution and carbon-dioxide-reduction activity. This combination suppresses parasitic reactions and achieves kinetic matching among nitrate reduction, carbon dioxide reduction, and hydrogen evolution in alkaline electrolyte.

tZIF-Cu900 also delivered the smallest Tafel slope, 47.6 millivolts per decade, the largest electrochemical surface area, 128 square centimeters per gram, and the lowest charge-transfer impedance. By contrast, tZIF-Cu800 had sluggish reaction kinetics, while tZIF-Cu1000 and tZIF-Cu1100 showed severe loss of active area.

The catalysts were evaluated in both an H-type cell and a gas-diffusion-electrode cell, with carbon dioxide and zero point one molar potassium nitrate and potassium bicarbonate. The H-type cell screened intrinsic activity, while the gas-diffusion-electrode cell enhanced mass transfer through a three-phase interface for higher product yield under industrially relevant validation.

Across zero point zero one to zero point thirty molar potassium nitrate, the 1.5 nanometer cavity directed the largest fraction of converted nitrate toward urea, reaching 90.2% at zero point ten molar. The 1.0 nanometer cavity preferentially produced ammonium, while nitrite accumulation was greater in the 1.8 and 2.6 nanometer cavities.

In the gas-diffusion-electrode cell, tZIF-Cu900 delivered a urea yield of up to 71.32 grams per gram of catalyst per hour, with 98.9% Faradaic efficiency and a turnover frequency of 9430 per hour at minus zero point seven volts versus the reversible hydrogen electrode. In panel a, the tZIF-Cu900 curve extends to the most negative current densities, while panel b shows the corresponding Tafel slopes labeled at forty-seven point six, eighty-three point one, one hundred thirty-one point six, and sixty-three point seven millivolts per decade.

Panel c also shows distinct capacitance slopes, spanning zero point six five to two point seven nine millifarads per square centimeter. Molecular-probe inhibition and reactant-feed experiments revealed a pore-size-dependent confinement effect. Small potassium ferricyanide and 2-mercaptoethanol could access multiple cavities, while progressively larger probes were used for size-dependent spatial blocking.

The figure block shows plots for the k-three-weighted copper K-edge extended X-ray absorption fine-structure signals and a schematic structural model of the nanoconfined copper-nitrogen-four active site. In panel c, the tZIF-Cu900 trace shows its main radial feature near one point five angstroms, aligning with the Cu–N reference rather than the Cu–Cu feature of copper foil.

Panel d reinforces this assignment: the strongest intensity is concentrated around roughly one point four angstroms in R-space, with no comparable Cu–Cu-like maximum at larger distance. Methylene blue most strongly inhibited urea formation in the 1.5 nanometer cavity, cysteamine preferentially inhibited the 1.8 nanometer cavity, and larger probes produced the greatest inhibition in the 2.6 nanometer cavity.

In contrast, 2-mercaptoethanol caused strong but non-selective inhibition through copper-sulfur interactions, while potassium ferricyanide moderately inhibited all catalysts. These size-selective inhibition profiles link cavity accessibility directly to urea-forming activity and support the role of 1.5–1.8 nanometer confinement in promoting carbon-nitrogen coupling.

Molecular-dynamics simulations tested the geometrical confinement hypothesis using four representative copper-nitrogen-four-embedded carbon cavities with diameters of 1.0, 1.5, 1.8, and 2.6 nanometers. The models compared pore-size-dependent diffusion of reactants, including carbon dioxide and nitrate, and key intermediates, including carbon monoxide, nitric oxide, and CONO.

The 1.0 nanometer cavity restricted hydrated nitrate entry, with a mean intracavity population of only 1.00, compared with 9.33 and 8.63 in the 1.5 and 1.8 nanometer cavities. In co-confined carbon monoxide and nitric oxide systems, the mean populations were 4.36 and 2.04 at 1.5 nanometers and 4.83 and 4.49 at 1.8 nanometers, suggesting relatively balanced co-occupancy at 1.8 nanometers.

The diffusion coefficients of carbon monoxide and nitric oxide increased in the 2.6 nanometer cavity, suggesting greater molecular mobility in the largest cavity. The 1.5–1.8 nanometer cavities may combine sufficient reactant access with reduced carbon-monoxide and nitric-oxide separation, providing a favorable confined environment for carbon-nitrogen coupling.

Finite-volume-method simulations, parameterized with the molecular-dynamics diffusivities, predicted pore-dependent steady-state concentration fields. Modeled urea concentration was highest at 1.5 nanometers and remained high at 1.8 nanometers; transport was restricted at 1.0 nanometer, while the 2.6 nanometer channel favored competing formation of formic acid and ammonium.

Beyond geometric confinement, thermal sculpting induces electronic modulation and electronic confinement at the active sites. The pyridinic-nitrogen-rich concave carbon cavity generates a strong localized electric field of 3.05 to 4.09 volts per angstrom, exerting compressive strain on the central copper atom and weakening hybridization between copper three-d and nitrogen two-p orbitals.

Density functional theory verified that this strain upshifts the copper d-band center from minus 3.36 electron volts to minus 1.85 electron volts. In panel a, the nitrate trajectories become visibly more concentrated within the copper-containing cavity as the pore narrows, whereas the two-point-six-nanometer model shows much broader excursions.

In panel b, the labeled d-band centers shift across the four pore sizes, from minus three point two six electron volts to minus one point eight five electron volts. Control experiments over tZIF-Cu900 identified key carbon-nitrogen coupling intermediates.

Carbon monoxide was the dominant carbon intermediate, while nitric oxide, hydroxylamine, and CONH2 were critical nitrogen intermediates. No urea was detected when nitrogen or ammonium was the sole nitrogen source, verifying the core carbon-monoxide and nitric-oxide coupling pathway.

After co-generation of carbon monoxide and nitric oxide at nanoconfined copper-nitrogen-four sites, the preferred urea pathway begins with direct carbon-monoxide and nitric-oxide coupling to form CONO. Sequential proton-coupled electron transfer produces CONH2, which couples with nitrite to form the second carbon-nitrogen bond and is then hydrogenated to urea.

The maximum uphill free-energy changes are 1.39, 0.96, 0.72, and 1.41 electron volts for the 1.0, 1.5, 1.8, and 2.6 nanometer cavities, respectively, making the 1.5–1.8 nanometer regime most favorable. Direct CONO formation has reaction free energies of minus 0.33, minus 0.62, minus 0.37, and 1.39 electron volts, suggesting a thermodynamic preference within the 1.0–1.8 nanometer cavities.

The 1.5 nanometer cavity showed the strongest CONO adsorption, while hydrogen adsorption was most favorable in the 2.6 nanometer cavity, consistent with enhanced carbon-nitrogen coupling and hydrogen-evolution competition, respectively. For pilot implementation, tZIF-Cu900 was integrated into a modular zero-gap proton-exchange-membrane electrolyzer stack paired with gas-diffusion electrodes at a catalyst loading of zero point zero two milligrams per square centimeter.

Operating continuously for 300 hours on industrial waste flue gas and nitrate-rich wastewater, the catalyst delivered a stable urea production rate exceeding 31.52 grams per gram of catalyst per hour with 90% Faradaic efficiency. After operation, the pristine rhombic morphology and atomic copper dispersion were largely preserved, with no obvious copper leaching or active-site agglomeration.

In panel a, the pathway proceeds through direct coupling of adsorbed carbon monoxide and nitric oxide, followed by proton-assisted steps and a second coupling near the lower sequence. Panel b maps these intermediates onto free-energy profiles, with highlighted barriers of two point zero two, two point one nine, and zero point six zero electronvolts marking key steps.

The isolated pilot-scale product closely matched pure urea in solid-state carbon-thirteen nuclear magnetic resonance and X-ray diffraction, with no discernible additional carbon resonances. A techno-economic assessment estimated a gross production cost of 472.44 U.S.

dollars per tonne of urea at a total energy demand of 16,440 kilowatt-hours per tonne. Including direct carbon-dioxide utilization, avoided nitrate-wastewater treatment, water recycling, and waste-derived feedstocks reduced the net production cost to 241.2 U.S.

dollars per tonne, below the comparison range of 460–700 U.S. dollars per tonne for conventional urea production. Under the stated resource-credit and carbon-price assumptions, the scenario-adjusted net cost was 175.16 U.S. dollars per tonne of urea.

In summary, programmable thermal sculpting regulates nanoconfinement around copper-nitrogen-four sites through temperature-controlled zinc volatilization. The size-matched cavities of tZIF-Cu900 balance molecular transport and local reactant enrichment while modulating the electronic structure of copper-nitrogen-four sites, promoting carbon-nitrogen coupling.

The confinement descriptor integrates cavity dimensions, molecular transport, and active-site electronic properties, providing a quantitative basis for confined catalyst design. Under laboratory conditions, tZIF-Cu900 achieved 98.9% urea Faradaic efficiency and a yield rate of 71.32 grams per gram of catalyst per hour.

A pilot-scale electrolyzer using industrial flue gas and nitrate-containing wastewater operated continuously for 300 hours, delivering 90% Faradaic efficiency and a yield rate of 31.52 grams per gram of catalyst per hour. The lower pilot-scale performance probably arises from feed-composition fluctuations, residual impurities, and less uniform mass transport.

Although transferability beyond the tZIF-Cu system requires further validation, the techno-economic analysis supports its potential for urea production from carbon dioxide and nitrate-containing waste streams. The size-matched cavities balance reactant access, intermediate retention, and copper electronic structure, producing near-unity urea selectivity in the lab and stable operation for 300 hours at pilot scale.