A research team from the Institute of Metal Research (IMR) of the Chinese Academy of Sciences, together with collaborators, has revealed a new mechanism by which spatially ordered heteroatoms induce bulk polarization in nonpolar semiconductor photocatalysts, offering a promising pathway to overcoming the long-standing challenge of efficient charge separation for solar water splitting.

Their findings were published in Advanced Materials on August 27.

Solar water splitting, often regarded as the "holy grail" of artificial photosynthesis, aims to directly convert solar energy into hydrogen fuel using semiconductor photocatalysts. However, in nonpolar semiconductors, photogenerated electrons and holes rapidly recombine within the bulk, severely limiting efficiency. Introducing a built-in electric field that penetrates the entire particle and provides a strong driving force for charge separation is an ideal strategy for suppressing bulk recombination.

Inspired by polar semiconductors, such as piezoelectric and ferroelectric materials, the researchers reasoned that achieving non-centrosymmetric atomic arrangements in nonpolar semiconductor particles could induce bulk polarization, creating a built-in electric field throughout the particle. Although heteroatom doping induces local lattice polarization, heteroatoms are randomly distributed at low concentrations, resulting in randomly oriented local polarizations that cancel out at the macroscopic scale.

Using interstitial Li-doped TiO₂ as a model system, the team combined density functional theory calculations with machine learning and revealed a "spatial ordering" mechanism driven by "local symmetry reduction → orbital hybridization → electron localization." 

Specifically, the local symmetry of the interstitial lithium (Li) ion intrinsically forbids hybridization between its s orbital and neighboring titanium (Ti) d orbitals. However, introducing additional interstitial lithium at specific sites reduces this local symmetry, enabling the previously forbidden hybridization to occur and forming new lithium-s/titanium-d hybridized states. The excess electrons introduced by the interstitial lithium then occupy these hybridized states, lowering the system's energy and making this specific, spatially ordered doping configuration thermodynamically favorable. 

Further introduction of N^3- for charge compensation suppresses recombination centers while maintaining the spatial ordering of Li/N heteroatoms.

Theoretical calculations predicted bulk polarization intensities of 2.36 and 3.47 μC/cm² along the [100] and [001] directions, respectively, for Li/N co-doped TiO₂ at a 5 at.% doping concentration. Moreover, the team successfully synthesized the predicted material, and structural characterization and performance tests confirmed the theoretical predictions.

Importantly, this mechanism is not unique to TiO₂. The team generalized it to two narrow-bandgap nonpolar visible-light responsive photocatalysts: Y₂Ti₂O₅S₂ and SrTaO₂N, demonstrating its potential generalizability to other nonpolar semiconductor functional materials.

This work provides a new strategy for promoting bulk charge separation in nonpolar semiconductors, with broad implications for photocatalytic energy conversion.

Schematic illustration of bulk polarization induced by an ordered spatial distribution of interstitial dopants. (Image by IMR)

Bulk polarization induced by interstitial Li and charge-compensated N-codoping. (Image by IMR)

Inducing bulk polarization with interstitial Li and charge-compensated codoping in Y₂Ti₂O₅S₂ and SrTaO₂N. (Image by IMR)