Dinuclear thiazolylidene copper complex as highly active catalyst for azid–alkyne cycloadditions

A dinuclear N-heterocyclic carbene (NHC) copper complex efficiently catalyzes azide–alkyne cycloaddition (CuAAC) “click” reactions. The ancillary ligand comprises two 4,5-dimethyl-1,3-thiazol-2-ylidene units and an ethylene linker. The three-step preparation of the complex from commercially available starting compounds is more straightforward and cost-efficient than that of the previously described 1,2,4-triazol-5-ylidene derivatives. Kinetic experiments revealed its high catalytic CuAAC activity in organic solvents at room temperature. The activity increases upon addition of acetic acid, particularly for more acidic alkyne substrates. The modular catalyst design renders possible the exchange of N-heterocyclic carbene, linker, sacrificial ligand, and counter ion.

For standard CuAAC reactions, copper(I) carboxylates [5,38], mononuclear copper(I) phosphine carboxylate complexes [39] or copper(I) salts are an ideal compromise of low catalyst cost and high catalyst activity. However, CuAAC reactions of some substrates are not compatible with heterogeneous catalysis at the surface of insoluble copper(I) compounds. Instead, they depend on highly active molecular catalysts under homogeneous reaction conditions. Our research group has already established molecularly defined dicopper catalysts with unprecedented activity under diluted conditions with low catalyst loading [37,40]. Thus, we aimed at an even more facile synthesis of dicopper complexes with bis-N-heterocyclic carbene ancillary ligands.

Results and Discussion
We herein report the synthesis of an ethylene-linked bisthiazol-2-ylidene dicopper(I) complex 2 that features high catalytic activity in CuAAC reactions. The advantage of this new complex 2 in comparison to dicopper complexes previously described by our research group [40] is its less time-consuming and more cost-efficient synthesis. Commercially available, inexpensive 4,5-dimethylthiazole is used as azole starting material instead of 4-aryl-1,2,4-triazoles. The precursor 1a for the NHC ancillary ligand is synthesized via a double S N 2-reaction of two equivalents of thiazole derivative with 1,2-dibromoethane. In order to avoid the presence of halide ions as inhibitory ligands for copper(I) [2,41], bisthiazolium hexafluorophosphate 1b was obtained by a salt metathesis from bromide salt 1a with aqueous hexafluorophosphoric acid ( Figure 1).

Scheme 2:
Synthesis of dinuclear copper complex 2. The final step is the reaction with copper(I) acetate and sodium acetate as additional base in order to deprotonate the thiazolium salt 1b and to form the bisthiazolylidene copper(I) complex 2.
Due to the relatively high acidity of the thiazolium precursor (pK a ≈ 18 [44]), a weak base such as sodium acetate yields small equilibrium concentrations of thiazol-2-ylidene. The latter then irreversibly binds to copper(I) ions (Scheme 2). The structure of the resulting complex 2 is presumably similar to the previously reported dinuclear bis(1,2,4-triazol-5-ylidene)copper(I) complexes that had been synthesized, characterized, and structurally characterized in our group [40]. The similarity of NMRspectroscopic data of the 1,2,4-triazol-5-ylidene and the 1,3thiazol-2-ylidene dicopper complex indicate that the complexes of both NHC ligand types consist of a bis-NHC ligand, two copper(I) ions and a labile acetate ligand that bridges the metal centers. The thiazolylidene complex 2 is air-stable in the solid state for at least several days. Stability tests in solution were taken under an atmosphere of nitrogen. Small amounts of a brown precipitate were formed in solution after one day. However, the NMR spectra showed no changes even after one week and in the presence of acetic acid. Therefore, we assume that complex 2 is quite robust against oxidization.
To test the catalytic performance of complex 2 with the help of continuous NMR spectroscopy, the reaction of benzyl azide with either phenylacetylene or ethyl propiolate in deuterated dichloromethane at room temperature was used (Table 1 and Figure 2). Due to the highly exothermic nature of the triazole formation, a high dilution of the reaction mixture and low catalyst loadings are necessary to prevent a thermal runaway. In order to compare the catalytic activity of complex 2 with conventional catalysts a kinetic study with copper(I) acetate was performed. All kinetic experiments were carried out under an atmosphere of nitrogen because of the air-sensitivity of complex 2 in solution (see Supporting Information File 1 for the detailed procedure). The reaction with phenylacetylene and 1.8 mol % copper complex 2 ( Table 1, entries 1 and 2) reaches 50% conversion within 37 min (without acetic acid, green triangles in Figure 2) and is slightly accelerated by addition of acetic acid (half conversion within 22 min, blue dots). In contrast, the half conversion time for this reaction catalyzed by a saturated homogeneous solution of copper(I) acetate in deuterated dichloromethane is about 3 h (Table 1, entry 3, black dots in Figure 2). Therefore, the reaction with complex 2 is about 4.5 times (without HOAc) to 7.5 times (with HOAc) more effective compared to the homogeneous reaction with copper acetate. Under heterogeneous catalytic conditions, however, larger amounts of commercially available CuOAc powder with vivid stirring or shaking of the reaction mixture give rise to rapid CuAAC conversion. Thus, the molecular NHC dicopper catalyst complexes excel in homogeneous CuAAC reactions [40], while CuOAc excels in heterogeneous catalysis and in cost-effectiveness [38].
The reaction with ethyl propiolate in the presence of 0.9 mol % catalyst is very slow with a half conversion time of more than nine hours (Table 1, entry 4, yellow diamonds). We attribute this poor catalytic activity to the formation of copper acetylide clusters or even coordination polymers. Analogous dicopper complexes of more sterically demanding bis-1,2,4-triazolylidene ancillary ligands are quantitatively converted to octacopper hexaacetylide clusters under the same conditions [37].
To date, we have not been able to characterize thiazolylidene copper acetylides. Addition of acetic acid greatly increases the rate of the CuAAC reaction with ethyl propiolate, so that half-conversion is reached after 6 min (Table 1, entry 5, red squares). These observations are again consistent with the formation of a thermodynamically stable copper acetylide species [37], which are regenerated in the presence of acid to catalytically active dicopper acetylide intermediates [36]. The reaction catalyzed by copper acetate proceeds very slow. The extrapolated half-conversion is reached within approximately one day (Table 1, entry 6, grey triangles).

Conclusion
In summary, we have presented a molecularly defined bisthiazolylidene dicopper(I) complex that features high catalytic activity in CuAAC reactions. Its three-step synthesis is straightforward and cost-efficient. The modular design of this class of catalysts renders possible the tuning of the complex's properties and its features according to specific demands. Dicopper complexes with thiazolylidene ancillary ligands provide for improved availability, air-stability and convenience for the growing community of CuAAC users.

Experimental
General methodology . For processing, analysis and interpretation of NMR spectra, the program TopSpin 3.2 by Bruker was used. All observed signals are singlets. Elemental analyses were carried out in the Department of Chemistry at the University of Heidelberg on the instruments vario EL and vario MICRO cube by Elementar Analysensysteme GmbH. Infrared spectra were recorded on a Bruker Lumos instrument with a Germanium ATR-crystal. The positions of the peaks are indicated in wavenumbers ν in cm -1 . The following abbreviations were used to describe both the intensity and profile of the signals: w (weak), m (medium), s (strong), br (broad). Mass spectra were recorded by the Mass Spectrometry Service Facility of the Organic-Chemical Department of the University of Heidelberg using the following instruments: Vacuum Generators ZAB-2F, Finnigan MAT TSQ 700, JEOL JMS-700, Bruker ICR Apex-Qe hybrid 9.4 T FT-ICR. In general the ionization method was specified. Apart from the method of ionization and the peak of the molecular ion, the base peak and characteristic fragmentation peaks with their relative intensities are reported.
After cooling to room temperature, the resulting solid was suspended in ethanol (2 mL) and the mixture was filtered. The solid residue was washed with ethanol (3 × 2 mL) and diethyl ether (3 × 3 mL) and dried in vacuo to give the beige product 1a (1.05 g, 2.53 mmol, 86%). 1  and dichloromethane or tetrahydrofuran (3 mL) were added. The suspension was stirred at room temperature for 5 d.
a) Procedure for the reaction in dichloromethane: The suspension was filtered over a frit and the solution was concentrated by reducing the solvent in vacuo to 2 mL. Diethyl ether (4 mL) was added and the formed precipitate was filtered, washed with diethyl ether (3 × 2 mL) and dried in vacuo to give the light beige product 2 (0.06 g, 0.10 mmol, 56%).
b) Procedure for reaction in tetrahydrofuran: The solvent was removed under reduced pressure and dichloromethane (4 mL) was added in the glove box. The suspension was filtered over a frit and the solution was concentrated by reducing the solvent in vacuo to 2 mL. Diethyl ether (4 mL) was added and the formed precipitate was filtered, washed with diethyl ether (3 × 2 mL) and dried in vacuo to give the light beige product 2 (0.09 g, 0.16 mmol, 86%). 1

Supporting Information
Supporting Information File 1 Author contributions, details of the procedures for the kinetic measurements, and figures of NMR spectra.