|Home | About | Journals | Submit | Contact Us | Français|
The resting state of the gold(I)-catalyzed hydroarylation of 1 changes in the presence of Ag+, with silver free catalysts resting at the dinuclear gold structure 5 and Ag+ containing solutions resting at a heteronuclear species like 6. Adventitious Ag+ (typically from LAuCl activation) can therefore intercept key organogold intermediates and effect the catalysis even when it does not effect the reaction in Au free control experiments.
Silver salts are commonly used to activate metal halides for catalytic transformations, and this strategy is especially useful in homogeneous gold catalysis.1 After control experiments demonstrating that the AgX (X = BF4, OTf, NTf2, etc.) activator does not catalyze the reaction of interest, the silver salts are often used in excess to ensure quantitative activation of the precatalyst.2 Despite proving that the silver salts themselves are unreactive, there are examples of gold(I)-catalyzed reactions where Ag+ effected either activity or selectivity,3 but models for accommodating such observations are not well appreciated in the gold catalysis community. We report herein observations that point to Ag+ ions intercepting Au(I) catalytic intermediates and subsequently effecting catalyst speciation and reaction kinetics; a structural model is also suggested.
In a recent investigation of the gold(I)-catalyzed hydroarylation of 1 to 2,4 we noted that the catalyst rested at the digold vinyl intermediate 5,5 which could additionally be synthesized by the combination of Ph3PAuNTf2 (4) and the monoaurated vinylgold(I)-complex 3 (Scheme 1). Based on the doubly aurated aryl compounds of Schmidbaur, 5 was formulated as an out of plane Au–C–Au 3-center-2-electron bond with a stabilizing d10 closed shell aurophilic interaction.6 Motivated by the observation of Ag+-effects in gold(I)-catalysis and the known examples of auro-argentophilic closed shell interactions in polymetallic complexes, we examined the effect of Ag+ on the catalyst speciation.7
As expected, the catalyst generated in situ from Ph3PAuCl/AgNTf2 (5 mol %/25 mol %) was found to be effective for the conversion of 1 to 2, while the AgNTf2 activator was much less reactive at RT.8 Monitoring the reactions by NMR, however, revealed a significantly different behavior from Ph3PAuNTf2 (4) alone.9 Instead of 5 being observed using 4, the Ag+-containing catalyst formulations rested at a new structure with a singlet in the 31P NMR at δ = 41 ppm (Figure 1a) and regenerated 4 at the end of the reaction. This compound, 6, could be independently synthesized by reacting a mixture of 4 and AgNTf2 with 1 in the presence of 2,6-di-tert-butylpyridine, which acts to arrest protodemetalation.10 The resonances of this new species (Figure 1b) exactly matched those observed in the 1H and 31P NMR of the in situ monitored catalytic reactions. Figure 1c and d additionally display the corresponding spectra for the isolated intermediates 5 and 3, respectivley.
Analysis of the 1H NMR of 6 indicated a complex with a PPh3 to carbocyclic framework ratio of one. Both vinyl protons additionally were found to couple to one phosphorus nuclei (Hsyn: 4JPH = 5 Hz; Hanti: 4JPH = 14 Hz), and like 5 and previously reported multimetalated compounds, the quaternary bimetalated sp2 carbon showed no signal in the 13C NMR, perhaps a result of the quadrupole moment of gold.6a High resolution mass spectrometry showed an m/z of 899.1046 (calcd 899.0966) for the [M–NTf2]+ ion with the isotope pattern expected for a Au–Ag containing compound.
Inspection of X-ray data for numerous simple Au–Ag complexes, e.g. A,12b suggests that Ag+ likely binds to the vinyl carbon and the Au ion without significantly perturbing the gold-vinyl structure.12 In other words, the 3-center-2-electron interaction is not equal and the stronger C–Au bond dominates the structure. Since only a single diastereomer is observed, one of the diastereofaces must be preferentially populated (i.e. 6a or 6b).13
Since the putative role of 6 is to react with acid to give 2 and regenerate 4 and AgNTf2, the acid stability of 6 was tested. Compound 6 was unchanged in the presence of 5 equivalents of AcOH,5 but did provide traces of 2 with 35 equivalents over 48 h. One equivalent of the stronger acid HNTf2 (generated in situ during catalysis) resulted in a rapid protodeauration to form 2. When treated with Br−, precipitation of AgBr drives the conversion of 6 to the monoaurated goldvinyl 3 (Scheme 2).
The rapid and clean reaction of 3, a putative reaction intermediate, with 2 equivalents of AgNTf2 to form 6, suggests that this process may be viable under catalytic conditions, and thus act to trap intermediate Au(I)-compounds. Similarly, the dinuclear gold intermediate 5 reacted with 3 equivalents of AgNTf2 to provide the mixed metal species 6 showing that even preformed digold structures can be driven to the heterometallic structure.14
Such a silver-for-gold substitution seems only slightly downhill as it was found in catalytic experiments that the resting state monitonically shifted from 6 to 5 as the Au/Ag ratio was increased. The change in speciation was especially dramatic when gold was added as Ph3PAuCl, which simultaneously decreased the Ag+ concentration by precipitation of AgCl (Figure 2 and and3).3). Under suitably balanced conditions it is therefore possible to detect both digold and mixed gold/silver resting states.15
The effect of excess AgNTf2 on the kinetics of the cyclization was also investigated. At a Au(I) loading of 5 mol %, the rate with a 3-fold excess of Ag+ was significantly retarded relative to the gold only experiment (Figure 4). Unexpected was the increased rate of 1 consumption as the reaction progressed.
We show herein that even if the reactants are untouched by Ag+, silver ions can still influence gold-catalyzed reactions by intercepting key organogold intermediates to form dinuclear intermediates with their own unique reactivity. A structural model based on existing Au/Ag cluster chemistry literature is proposed, which may rationalize known Ag+ effects.
We thank J. Jacob Byrne (University of North Carolina at Chapel Hill) for samples of 1. Financial support is gratefully acknowledged from the Fulbright Foreign Student Program (for D.W.) and the National Institute of General Medicine (GM-60578).