Photocatalytic water reduction is a promising alternative to traditional hydrogen production pathways, because it does not rely on fossil fuel feedstocks or electrical power sources and instead requires only visible light and water. However, photocatalytic pathways of water reduction require the discovery and development of new catalysts to be a viable alternative hydrogen production route. Current and promising catalysts are primarily nanoparticulate precious metals, and a major barrier to discovering new and better alternatives is catalyst screening. Our previously reported high-throughput method addresses this screening challenge in that hundreds of metal combinations and concentrations can be tested efficiently to determine promising candidates. However, other variables, such as the formation of nanoparticles in situ and the stability of those nanoparticles, complicates interpretation of successful metal stoichiometries, and in particular the ability to isolate impactful properties like metal stoichiometries and surface composition. Without a clear understanding of how to control and stabilize nanoparticle formation in these in situ systems, it is not only difficult to determine how to optimize the nanoparticles themselves but also to develop computational models to predict high performing catalysts. Here, we compare the catalytic performance of in situ formed Au and AuCu nanoparticles to nanoparticles of the same composition synthesized prior to being used as catalysts, as well as compare how these two suites of nanoparticles are stabilizing via exogeneous ligand addition. While strongly binding particle ligands (such as those containing thiols in the case of Au and Cu) are typically considered detrimental to catalytic activity, we show that they are necessary for hydrogen production by comparing the effects of adding a strongly binding thiolated ligand to that of a weaker binding phosphinated ligand to both in situ and pre-synthesized nanoparticles. We then test the generality of this result by expanding our suite of metals used in situ to include Co, Ni and Pd, and observe again that a strongly binding ligand leads to enhanced catalytic activity. Taken together, these results suggest nanoparticle formation and stability are critical for activity, and importantly, we show that ligands can be beneficial, and even necessary, to produce active catalysts in situ.

John Kitchin

Carnegie Mellon University

Appearances