Photocatalytic Hydrogen Evolution
Photocatalytic hydrogen evolution produces H2 by driving proton or water reduction directly with light rather than with electricity. In common homogeneous-heterogeneous systems a molecular or particulate photosensitizer absorbs light, is reductively quenched by a sacrificial reductant, and passes the electron to a hydrogen evolution cocatalyst, often a metal nanoparticle; see also Hydrogen evolution reaction.
Screening bimetallic colloidal cocatalysts (2020)
Lopato and coworkers screened more than 700 bimetallic water reduction cocatalysts formed in situ from metal chloride salts (Pd/Sn, Pd/Mo, Pd/Ru, Pd/Pb, Pd/Ni, Ni/Sn, Mo/Sn and Pt/Sn) with an iridium photosensitizer and triethanolamine, run in air under blue LEDs (Lopato et al. 2020). Measuring synergy as the bimetallic rate minus the sum of the single-metal rates, they found clear enhancement for Pd/Sn, Pd/Mo, Pd/Ru and Pt/Sn; Pd/Sn showed its strongest synergy at only 7.7% Pd, and its highest absolute rates near a 1:1 ratio (0.2 mM each), where 4.7 nm particles containing both Pd and Sn formed. Summary: lopato2020.
Colloidal stability of in situ formed cocatalysts (2022)
Simon et al. used a 108-well parallel photoreactor with an iridium photosensitizer and triethanolamine to compare Au and Au/Cu nanoparticle cocatalysts formed in situ from metal salts, pre-synthesized with a thiolated PEG ligand, or formed in situ with that ligand added (Simon et al. 2021). Ligand-free in situ particles aggregated and stopped producing H2 by the third 15 hour irradiation round, while both ligand-stabilized preparations stayed discrete and active; Hg poisoning supported the particles as the active species. Ni, Co and Pd, previously found inactive, became active once a thiol ligand let discrete particles form, showing that screening results can reflect colloidal stability rather than intrinsic composition activity. Summary: simon2022.
Optimizing photosensitizer, metal and ligand concentrations (2022)
Bhat and coworkers optimized the concentrations of an iridium photosensitizer, metal salt and PEGSH ligand for in situ formed Au, Cu, Ni and Fe cocatalysts in a 108-well parallel photoreactor (Bhat et al. 2022). They found that activity was limited by the photosensitizer at the 0.25 mM used previously, rose steeply up to about 0.4 mM, and gained little beyond about 0.5 mM, so 0.5 mM suffices for metal loadings up to about 1.2 mM. Optimal compositions clustered near 1.2 to 1.6 mM metal and 0.8 to 1.1 mM PEGSH, with best measured yields of 17.9 (Cu) to 24.3 (Ni) µmol H2 per well. XPS showed metallic Au and Cu, mixed metallic Ni and NiO, and iron sulfide for Fe, the last being about as active as Au and stable for more than 45 hours. Summary: bhat2022.
Synergy in Cu-Ru-Fe colloidal cocatalysts (2023)
Bhat et al. screened in situ formed, PEGSH-stabilized Cu, Ru and Fe nanoparticles and their binary and ternary mixtures as cocatalysts with an iridium photosensitizer and triethanolamine (Bhat et al. 2023). Defining synergy as the H2 from a mixture minus the summed H2 of its pure metals at the same concentrations, they found that CuRu was strongly synergistic although Cu and Ru alone were weak or inactive, RuFe gave positive synergy, and CuFe gave both positive and negative synergy. At 0.8 mM total metal the best ternary well (0.1/0.2/0.5 mM Cu/Ru/Fe) outperformed the binaries. STEM-EDS showed that active wells held mixtures of particle types (for example CuRu-core/Cu-shell particles, or Cu plus CuFe particles), so the active composition could not be pinned down. Summary: bhat2023.
Machine-learned screening of in situ bimetallic cocatalysts (2023)
Broderick and coworkers compared about 16 million GemNet-predicted hydrogen adsorption energies with 5300 parallel photoreactor experiments across 84 monometallic and bimetallic systems, in which metal salts are reduced in situ to nanoparticle cocatalysts by an iridium photosensitizer with triethanolamine in DMSO (Broderick et al. 2023). The most active systems all exposed surfaces near thermoneutral hydrogen binding, but many systems with favorable predicted surfaces were inactive, and noble-metal-containing systems were about an order of magnitude more active than base-metal systems with similar binding energies. The authors point to particle formation, surface expression, deactivation (oxides and sulfides on Fe and Ni) and strain in Au and Ag nanoparticles as factors beyond adsorption thermodynamics. Summary: broderick2023.