Gold nanoclusters have been known to have a high catalytic reactivity. However, the difficulty in the synthesis of monodisperse nanoclusters (typically consisting of fewer than 100 atoms) in large amounts has been a considerable hurdle to catalytic research. Recently, bidentate ligands of the general formula P(Ph)2-(CH2)M-P(Ph)2 {labeled LM}, where Ph = phosphine and M is the length of the aliphatic chain separating the P atoms, is found to remarkably have the size selecting capacity for gold nanoclusters depending on M. To investigate the origins of size selectivity of the diphosphine ligands towards small-sized Au clusters, we performed scalar relativistic density functional theory (DFT) calculations using the projector augmented wave scheme (PAW). We find that a diphosphine with long spacer such as L5 can relieve a strain induced by the spacer more easily than can one with short spacer such as L3. Hence, while L5 can interact effectively with a broad range of gold clusters of various sizes, L3 can interact only with a narrow range of gold clusters demonstrating its size-selecting power towards small gold clusters such as Au11+3. Based on these results, we propose a two-body ligand system for an ideal, highly-selective ligand, in which one part of the ideal ligand provides a high reactivity towards the broad range of gold clusters and the other part of the ideal ligand provides the control over the reactivity, which could be the form of a short length of spacer, as in diphosphine, but not necessarily limited to, or any type of controllable, reactivity-poisoning component. The controllable competition between the two components of an ideal, highly-selective ligand system will produce a desirable selectivity for the generation of monodisperse nanoclusters of interest through tailoring process.

This work is supported in part by US-DOE under Grant No.DE-FG02-07ER46354.