Ligand Enhanced Activity of In Situ Formed Nanoparticles for Photocatalytic Hydrogen Evolution
Zoe C. Simon, Eric M. Lopato, Maya Bhat, Paige J. Moncure, Sarah M. Bernhard, John R. Kitchin, Stefan Bernhard and Jill E. Millstone, “Ligand Enhanced Activity of In Situ Formed Nanoparticles for Photocatalytic Hydrogen Evolution”, ChemCatChem 14(2) (2022; published online December 2021). https://doi.org/10.1002/cctc.202101551
This summary was written by a language model from the paper and has not been checked line by line. Rely on the published paper for details.
Question
High-throughput screens of metal salts that form nanoparticle (NP) catalysts in situ during photocatalytic hydrogen evolution can test hundreds of compositions, but when a composition is inactive, or less active than computation predicts, it is unclear why. The authors hypothesized that some of these discrepancies come not from intrinsic inactivity of the metals but from failure to form, or to keep, discrete colloidal particles. They asked whether NP capping ligands, usually thought to poison catalysts by blocking sites, could instead be needed to sustain activity (Simon et al. 2021).
Approach
The experiments used the group’s parallel photoreactor (up to 108 wells) in which H2 is detected by chemosensitive tape photographed every 10 minutes and converted to H2 amounts from RGB values with a calibration curve. Each 440 µL well contained an iridium photosensitizer, [Ir(Fmppy)2dtbbpy]PF6, triethanolamine (TEOA) as sacrificial reductant, DMSO and water, plus 1 mM total metal. Wells were irradiated with 440 to 450 nm blue LEDs for three rounds of about 15 hours (1000 minutes), with photosensitizer and TEOA replenished between rounds. Because the active catalyst area was unknown, no figure of merit was computed; a well was simply called active if it produced more than 2 µmol H2. Three catalyst preparations were compared for 100% Au, 75/25 Au/Cu and 50/50 Au/Cu: in situ NPs from metal salts with no ligand, pre-synthesized NPs capped with 1 kDa thiolated poly(ethylene glycol) (PEGSH), and in situ NPs formed with PEGSH added at a 1:1 metal to ligand ratio. Particles were characterized by TEM, UV-vis-NIR extinction spectroscopy and ICP-OES.
Key findings
Ligand-free in situ NPs of all three Au and Au/Cu compositions produced some H2 in the first round but were inactive by the third, and TEM showed aggregated particles. Pre-synthesized PEGSH-capped NPs (1.8 ± 0.2 nm diameter before reaction) stayed colloidally stable and active over the full 45 hour experiment, although they grew in size, with the solutions changing from brown to red within minutes of irradiation. Adding PEGSH to the in situ reactions gave discrete NPs and activity in all three rounds for every composition.
For Au, the ligand identity mattered little: oligoPEGSH (about 271 Da, five ethylene glycol units), 1, 2 and 5 kDa PEGSH, and 8-mercaptooctanoic acid all gave discrete, active NPs, both in situ and pre-synthesized (the latter prepared by ligand exchange from 4-diphenylphosphinobenzoic acid capped Au to keep size constant). The weaker binding phosphine ligand bis(p-sulfonatophenyl)phenylphosphine (BSPP) stabilized pure Au, which stayed active, but not the Cu-containing compositions, which aggregated and were inactive. Mercury poisoning after one round deactivated the PEGSH-stabilized in situ catalysts in every case, supporting the NPs as the active species. Finally, Ni, Co and Pd, which the group had previously found inactive at 1 mM, formed no discrete NPs and no H2 without ligand, but formed discrete NPs and evolved H2 when PEGSH was added (the Co case also showed large sheet-like structures).
Significance
The study shows that strongly binding thiolated ligands can be beneficial and even necessary for sustained activity of in situ formed photocatalytic HER cocatalysts, contrary to the usual view of ligands as site-blocking poisons. For screening, it means that apparent inactivity can reflect colloidal instability rather than intrinsic chemistry, so compositions dismissed earlier may deserve retesting. For computational prediction of such catalysts, the authors argue that particle formation, stability and possibly surface ligand chemistry must be considered both in making predictions and in judging them against experiment. The authors note that H2 amounts cannot be compared quantitatively across preparations because the active surface area is unknown.