The nitrogen element is a necessary ingredient for building proteins and nucleotides, making it essential for all living organisms. Although the atmosphere contains 78 vol% N₂, N₂ molecules cannot be directly used by most living organisms because of the inert reaction activity caused by the extremely difficult bond cleavage. To overcome this problem, the Haber-Bosch process was invented. It has a harsh reaction condition requirement of over 300 °C and 200 atm and is currently the most applied artificial nitrogen fixation method. The Haber-Bosch process provides over 50% of nitrogen in human bodies. However, to acquire and maintain the rigorous reaction conditions, ~2% of the total world energy supply is consumed and ~1% of the total world CO₂ emissions is produced. The realization of energy-efficient, environmentally friendly artificial nitrogen fixation under mild conditions is thus an important goal around the world.

Photocatalytic nitrogen fixation has attracted blossoming interests for its ability of achieving nitrogen fixation under mild conditions with solar energy as the sole energy source. Since TiO₂ was first reported for water splitting under solar light illumination, semiconductor photocatalysts have received enormous attention for different chemical reactions. As for photocatalytic nitrogen fixation, semiconductor photocatalysts can provide electrons and holes for the redox reactions upon proper light illumination, and defects on semiconductors or additionally introduced species can function as adsorption and activation sites for N₂ molecules, which can reduce the cleavage energy of N₂ molecules and allow the reaction to be conducted under mild conditions. However, most semiconductors whose conduction band minima are sufficiently negative for the nitrogen reduction reaction have wide bandgaps. They therefore only absorb light in the ultraviolet or near-ultraviolet region, leading to rather low solar-to-chemical conversion efficiencies (SCCEs). In this regard, hybrid nanostructure photocatalysts made of plasmonic metals and semiconductors emerge due to their wonderful characteristics in solar energy harvesting and catalysis. Localized surface plasmon resonance (LSPR) originates from the collective oscillations of free charge carriers under photon excitation. It allows light to be spatially confined to nanoscale, it can generate hot charge carriers for redox reactions, and it endows photocatalysts with the light absorption ability in the visible and near-infrared regions. The integration of plasmonic nanoparticles with semiconductors greatly improves the photocatalytic nitrogen fixation efficiency and has pushed this field into a new era. However, this design also introduces a Schottky barrier at the metal–semiconductor interface, limiting the injection of the hot electrons from the plasmonic metal into the semiconductor and further transfer to the nitrogen-active sites, which places a limit on the nitrogen fixation performance. So far, the highest SCCE in photocatalytic nitrogen fixation has been 0.1%, which is too low for practical applications.

Herein we report oxygen-deficient molybdenum(VI) oxide nanospheres as a promising photocatalyst to achieve a higher SCCE. The oxygen vacancies are created under an oxygen-deficient environment and can be well controlled by annealing in air for different periods of time. These oxygen vacancies not only introduce abundant free electrons, which form the basis for the occurrence of strong localized plasmon resonance in the visible to near-infrared region, but also effectively chemisorb and activate N₂ molecules through considerable electron transfer between the N₂ molecule and the oxygen vacancy. The recombination of the hot charge carriers is effectively suppressed and their lifetime is prolonged in a large extent due to the trapping effect of the introduced defect electronic states. The Schottky-barrier-free nature guarantees the free movement of the hot charge carriers, which results in an efficient utilization of the hot charge carriers. As a result, the Schottky-barrier-free plasmonic MoO_3-x nanoparticles give record-high apparent quantum efficiencies of over 1% in the spectral region from 700 nm to 1064 nm, and a SCCE of 0.057% without use of any hole scavenger. Through the optimization of the reaction conditions, the SCCE has recently been further improved to an even higher value. We believe that this new concept of the Schottky-barrier-free design offers a great opportunity for the rational design of new efficient photocatalysts with ever higher SCCEs toward practical applications.