A New Method to Link Metal Ions via Cycloaddition of Metal-Azides to Metal-Acetylides

Veige, Adam

The goal of this grant was to develop a deeper understanding of the 1,3-dipolar cycloaddition reaction between metal-azides and metal-acetylides, a transformation we have labeled as iClick. Our preliminary results (Scheme 1) have shown that gold(I) is amenable to this type of reaction, as demonstrated by the reaction of triphenylphosphinegold(I) azide (1-Ph) and triphenylphosphinegold(I) phenylacetylide (2), which undergo a 1,3-dipolar cycloaddition reaction to yield the triazole bridged product, 4-phenyl-1,5-triphenylphosphinegold(I) 1,2,3-triazolate (3). As outlined in the proposal, we planned to probe the scope of the iClick reaction, in terms of metal compatibility, ancillary ligand effects, and acetylene substituents.

A series of kinetic experiments were designed to probe the mechanism of the iClick reaction between a variety of gold(I) azides and gold(I) acetylides. The reaction proved to be first order in both reactants, or second order overall, with respect to gold(I) concentration. This was not surprising, as the generally accepted mechanism of Copper(I)-catalyzed Azide Alkyne Cycloaddition reactions (CuAAC) is also second order with respect to the copper concentration, a first row congener of gold. The proposed mechanism is depicted in Scheme 2, alongside the proposed CuAAC mechanism. The necessity for ligand dissociation was confirmed by a significant rate inhibition when free triphenylphosphine was systematically added to the reaction. This finding is further corroborated by results indicating N-heterocyclic (NHC) supported gold(I) azides and acetylides to not proceed under ambient conditions; as NHC ligands are stronger σ–donors than phosphine ligands, it is harder for this crucial ligand dissociation to occur. Lastly, when the para-proton of the phenyl acetylene was systematically substituted with both electron withdrawing (NO₂ and F) and electron donating (OMe) groups, the gold(I)/gold(I) iClick reaction rates followed a linear Hammet plot, with a slope of 0.976. This slope indicated there is a significant buildup of negative charge in one of the intermediates, which is stabilized by the more electron withdrawing substituents, leading to faster reaction rates. A nucleophilic attack of the terminal nitrogen of the azide by the b-carbon of the acetylide, could indicate a reasonable location of the partial negative charge in the intermediate.

When the triphenylphosphine ligands are substituted with the more basic, yet conically smaller triethylphosphine ligands (1-Et and 4-Et), the reaction is effected notably. While the reaction proceeds to a similar triazolate-bridged, dinuclear intermediate (5), strong intermolecular aurophilic interactions cause a dimerization (6) to occur (Scheme 3). This dimerization is permitted by the less sterically demanding triethylphosphine ancillary ligands. The crystallographically characterized product consists of a pseudo C₂-symmertic distorted tetrahedral geometry of gold(I) ions, and can be seen in Figure 1.

To demonstrate the compatibility of other metals in an iClick type reaction, a series of phosphine supported platinum(II) di-azide complexes were synthesized. Triphenylphosphine supported platinum(II) diazide 7-Ph readily undergoes two subsequent iClick reactions with two equivalents of gold(I) acetylide 4-Ph. When this reaction is carried out in CDCl₃, the stepwise iClick reactions can be monitored over a 12 hour period, but when run in C₆D₆, at slightly elevated temperatures, the desired, trinuclear product (8-Ph) crystalizes out in 92% isolated yield.

The effect of the ancillary ligand was monitored, and once again, triphenylphosphine was substituted with triethylphosphine. While the desired trinuclear product 8-Et was isolated, the yield of the reaction suffered from deleterious side-reactions, which were the result of anionic ligand exchange of the reactants. While anionic ligand exchange was not observed when triphenylphosphine was utilized, the more strongly σ–donating triethylphosphine located trans to the azido functionalities leads to increased labilization. The other two isolated products, were the gold(I)/gold(I) iClick product, 6, and 9, the result of a single iClick reaction between one of the azides on the platinum(II), and the acetylide of the gold(I), followed by anionic ligand exchange of the other azide with an acetylide from another gold(I) (Scheme 5). Multinuclear and 2-D NMR spectroscopy data provides reasonable evidence for the identity of 9, though its isolation from the product mixture was not possible.

To create a complex in which an gold(I)-gold(I) interaction could be manipulated, we sought to forcibly constrain the gold(I) ions in close proximity. Tethering the gold(I) acetylide units across a bis(diphenylphosphino)methane (dppm) bridge (10) provides the necessary structural constraint, and allows the synthesis of the trimetallic Pt^II/Au^I₂ complex 11 in 36% yield (Scheme 6). Figure 2 shows the solid state structure of 8-Ph, 8-Et, and 11, with hydrogen atoms and disordered atoms removed for clarity. Both pseudo-C₂-symmetric, square planar complexes, 8-Ph and 8-Et, the gold(I) ions orient anti, relative to each other, but the C₁-symmetric 11, features gold ions oriented in a syn fashion, with an aurophilic interaction between them (Au-Au bond distance of 3.1878(10) Å).

We also attempted to employ the second row palladium congener of 7, however the reaction results in an intractable mixture of products. Both third and second row iridium(I) and rhodium(I) azides do participate in iClick with gold(I) acetylides, as depicted in Scheme 7. These products have been characterized by NMR spectroscopic techniques.

In conclusion, this grant has supported the research to better understand the scope of the iClick reaction. We now have a firm understanding of the reaction mechanism, but more importantly, we now understand the limitations of the iClick reaction. This project, as is its intention, is to nurture new research directions. The results from this project were used as preliminary result to obtain additional support from the DOE-BES materials division. The project was initiated by one undergraduate student (Trevor Del Castillo) who went on to win the prestigious NSF graduate fellowship and is now pursuing his PhD at Cal Tech under the guidance of Prof. Jonas Peters. Currently three graduate students are working on this project within the Veige laboratories. Andrew Powers is scheduled to graduate with a Ph.D. in the spring of 2015. Xi Yang (Ph.D., 2016) and Chris Beto (Ph.D., 2018) will continue the project