Dawei Wang, Rong Cai, Sripadh Sharma, James Jirak, Sravan K. Thummanapelli, Novruz G. Akhmedov, Hui Zhang, Xingbo Liu, Jeffrey L. Petersen, and Xiaodong Shi
J. Am. Chem. Soc.
Occasionally the topic of trace palladium catalysis in processes reported to be iron catalysed appears in the literature and, more often, in questions asked in seminars on iron catalysis. The trace palladium in these cases can come from reaction flasks than haven’t been properly cleaned or from impurities in the metal that’s added. In gold catalysis, a silver salt with a non-coordinating counter ion (eg. SbF6-) is often added to abstract a halogen from the gold species, leaving the homogenous gold catalytic system and notoriously insoluble silver halides, which can be removed by filtration. However, often the precipitate is simply left in the reaction flask in the assumption that it won’t participate in the reaction. But does it?
A study from the Shi group in West Virginia begins to answer this question in an impressively thorough manner. Their initial interest began when a report in the literature contradicted work going on in the group. Further investigation showed that if the gold catalyst was filtered through Celite to remove silver salts, then no activity was observed. But, if it was not filtered or filtered using only filter paper, then the catalytic system retained activity. X-ray photoelectron spectroscopy was employed to determine the metal composition of the two filtrates and showed that filter paper wasn’t sufficient to remove all the silver, suggesting a cooperative catalytic system. Phosphorus NMR studies also revealed that the gold complexes in solution were influenced by silver, promoting catalytic activity.
Having solved the original issue, the group commendably set about investigating a wide range of gold catalysed reactions reported in the literature and the “silver effect” on the reactivity of the catalytic systems. They divided the reactions into 3 categories;
- Type 1 – True [L-Au]+ Catalysis
- Type 2 – Bi-Metallic Catalysis – Requires Au/Ag mixture
- Type 3 – Silver Assisted Catalysis – Variable reactivity with Au, Ag or Au/Ag mixture
Of the 14 literature reactions tested only 2 were found to be true gold catalysis with the silver content of the system having no effect on activity. Five examples were shown to be Type 2 – requiring both gold and silver to be present for reaction. Indeed, if the catalyst was activated and then filtered through Celite to remove any silver traces, the catalysts were inactive, but the addition of “insoluble” silver chloride, which cannot activate gold by halide abstraction, formed a catalytically active system. The majority of the remaining reactions were Type 3; pure gold systems could catalyse the reactions but the presence of silver increased catalytic activity. Two reactions which were broadly classed as type 3 were equally as good with gold, silver or a mixture.
This work potentially opens up new areas in gold-silver bimetallic catalysis and certainly highlights the importance of thorough investigation in the development of gold-catalysed reactions.