Gold(I) Complexes with Ferrocenylphosphino Sulfonate Ligands: Synthesis and Application in the Catalytic Addition of Carboxylic Acids to Internal Alkynes in Water

: The synthesis and characterization of novel gold(I) complexes containing hydrophilic ferrocenylphosphino sulfonate ligands, i.e., compounds [AuCl{( η 5 -C 5 H 3 PR 2 (SO 3i Pr))Fe( η 5 -C 5 H 5 )}] (R = Ph ( 2a ), p -Tol ( 2b ), Cy ( 2c )), are presented, including a single-crystal X-ray di ﬀ raction study on 2a . Complexes 2a – c were checked as catalysts for the intermolecular addition of carboxylic acids to nonactivated internal alkynes using water as a green reaction medium. The best results in terms of activity were obtained with 2a in combination with AgOAc, which was able to promote the selective anti addition of a variety of aromatic, aliphatic, and α , β -unsaturated carboxylic acids to both symmetrical and unsymmetrical internal alkynes at 60 ◦ C, employing metal loadings of only 2 mol %.


Introduction
Enol esters represent an important class of synthons in organic chemistry, commonly employed as intermediates, among others, in cross-coupling [1][2][3], asymmetric hydrogenation [4][5][6], and cyclization reactions [7][8][9], as well as monomers in polymerization and oligomerization processes [10][11][12]. Among the different methods of synthesis of these valuable molecules, the intermolecular addition of carboxylic acids to alkynes catalyzed by transition metals (hydro-oxycarbonylation reaction) is probably the most straightforward and atom-economical one because the starting materials are widely available and no byproducts are generated in the process. A large number of catalytic systems, mainly involving Groups 8-11 metals, have been reported in the literature, with those based on ruthenium being probably the most popular [13][14][15][16][17]. However, the vast majority of studies have focused on the hydro-oxycarbonylation of terminal alkynes. Examples of the intermolecular addition of carboxylic acids to internal alkynes still remain scarce [18], and in most of the cases, only activated substrates, such as trifluoromethylated alkynes [19], acetylenic esters [20], ynol ethers [21], ynamides [22], Given the tremendous effect that ligands can exert in homogeneous Au catalysis [31], we reasoned that a change in the nature of the coordinated phosphine ligand could improve these previous results. In particular, we turned our attention to the ferrocenylphosphino sulfonates 1a-c recently synthesized by us (Scheme 2), because preliminary studies showed their utility as auxiliary P-donor ligands in ruthenium-catalyzed C-H activation processes in aqueous media [32]. In addition, compounds 1a-c represent rare examples of hydrophilic ferrocenylphosphines, a particular class of ligands whose potential in aqueous catalysis remains almost unexplored. Thus, in this work, the preparation of gold(I) complexes containing the ferrocenylphosphines 1a-c is presented, as well as an evaluation of their catalytic activity in hydro-oxycarbonylation reactions of nonactivated internal alkynes in water. Scheme 2. Synthetic route employed for the preparation of the ferrocenyl sulfonate ligands 1a-c.

Results and Discussion
The treatment of dichloromethane solutions of the ferrocenylphosphino sulfonates 1a-c with one equivalent of [AuCl(SMe 2 )] at room temperature led to the clean and fast formation of the novel gold(I) complexes 2a-c through the expected displacement of the labile dimethylsufide ligand (Scheme 3). The characterization of complexes 2a-c, which were isolated as air-stable yellow solids in 79-94% yield, was straightforward following their analytical and spectroscopic data (details are given in Materials and Methods). In particular, the 31 P{ 1 H} NMR spectra were very informative, showing in all the cases a singlet resonance at δ P 26.2-49.8 ppm, strongly deshielded with respect to that of the free ferrocenylphosphino sulfonates 1a-c (δ P from −22.2 to −12.0 ppm) [32]. The 1 H and 13 C{ 1 H} NMR spectra were also fully consistent with the proposed formulations, featuring two sets of signals for the R substituents of the phosphino groups, as well as for the diastereotopic methyl units of the O-i Pr moieties (i.e., two doublet signals at δ H 0.91-1.46 ppm ( 3 J HH = 6.0-6.3 Hz) and two singlets at δ C 22.6-23.6 ppm), as a consequence of the planar chirality of the ferrocenyl fragments due to the 1,2-disubstitution of one of their Cp rings. In addition, the structure of complex 2a was unambiguously confirmed by means of a single-crystal X-ray diffraction study. X-ray quality crystals were obtained by slow diffusion of hexane into a saturated solution of 2a in dichloromethane. An ORTEP-type view of the molecule, along with selected structural parameters, is shown in Figure 1. 2.2893(9) Å, respectively. These bonding parameters compare well with those previously found in the solid-state crystal structures of related ferrocenylphosphine-gold(I) complexes, like [Au 2 Cl 2 (µ-dppf)] (dppf = 1,1´-bis(diphenylphosphino)ferrocene; Au-P = 2.2262(13) Å, Au-Cl = 2.2815(13) Å, and P-Au-Cl = 179.59(5) • ) [33], [Au 2 Cl 2 {µ-(η 5 -1,3-C 5 H 3 PPh 2 (OSi i Pr 3 )) 2 Fe}] (Au-P = 2.2282(9) Å, Au-Cl = 2.2814(10) Å, and P-Au-Cl = 177.35(4) • ) [34] or [AuCl{(η 5 -1,2-C 5 H 3 PPh 2 (1-naphthyl))Fe(η 5 -C 5 H 5 )}] (Au-P = 2.2278(11) Å, Au-Cl = 2.2845(11) Å, and P-Au-Cl = 174.16(5) • ) [35]. On the other hand, the observed bond distances within the ferrocenylphosphine sulfonate skeleton were very similar to those found in the structure of the free ligand 1a (± 0.04 Å) [32]. The same can be said about the bond angles, where the most noticeable differences were the increase in ca. 5-6 • of the C(1)-P(1)-C(11) and C(1)-P(1)-C(17) angles when passing from 1a to 2a. All these observations indicate a negligible influence of the gold coordination on the geometry of the ligand.
With the complexes [AuCl{(η 5 -C 5 H 3 PR 2 (SO 3 i Pr))Fe(η 5 -C 5 H 5 )}] (2a-c) in hand, we next explored their catalytic potential in hydro-oxycarbonylation reactions of internal alkynes in water. In particular, a first set of experiments was performed with complex 2a and the model substrates hex-3-yne (3a) and benzoic acid (4a) (see Table 1). [AuCl{(η 5 -C 5 H 3 P(p-Tol) 2 (SO 3 i Pr))Fe(η 5 -C 5 H 5 )}] (2b) AgOAc 89 9 [AuCl{(η 5 -C 5 H 3 PCy 2 (SO 3 i Pr))Fe(η 5 -C 5 H 5 )}] (2c) AgOAc 77 10 ---- AgOAc 4 system [AuCl(PPh 3 )]/AgOAc (see Scheme 1) [30], all the reactions collected in Table 1 proceeded cleanly without any side-reaction associated with the hydration of the C≡C bond or oligomerization processes, and with an exquisite anti selectivity (only the Z isomer of 5aa is formed). A blank experiment with AgOAc alone confirmed that gold is responsible for the catalytic activity observed (entry 10). We would also like to highlight in this point that the effectiveness shown by the 2a/AgOAc system in this reaction compares favorably with that reported for [AuCl(PPh 3 )]/AgOAc (5 mol % of both reagents were needed to generate 5aa in a comparable yield under identical experimental conditions) [30], which could be explained in terms of the higher solubility in the water of complex 2a (ca. 10 mg/mL at 60 • C). However, the different reactivities observed between complexes 2a, 2b, and 2c do not seem to be related to their water solubilities, since the most soluble one 2c (ca. 13 mg/mL at 60 • C) is the least effective (the solubility of 2b is identical to that of 2a).
The scope of Au(I) complex 2a was subsequently explored by varying firstly the carboxylic acid reagent. Thus, as shown in Scheme 4, different benzoic acids 4b-f could be successfully added to hex-3-yne (3a), regardless of the electronic nature and substitution pattern of the aryl ring. As for 5aa, the resulting enol esters 5ab-af were exclusively obtained as the corresponding Z isomers, as assessed by comparison of their NMR data with literature values [30]. The use of 4-hydroxybenzoic acid (4f) led to an interesting result because, in addition to the expected hydro-oxycarbonylation reaction, the hydroalkoxylation of 3a also occurred, leading to the previously unknown enol ester 5af (to obtain this compound in good yield, a 3a/4f ratio of 2.4:1 was employed). We would like to remark here that, although gold-catalyzed intermolecular hydroalkoxylation reactions of alkynes have been extensively studied [36], no previous examples in water can be found in the literature. As exemplified with compounds 5ag-an, the addition process could also be extended to benzylic, homobenzylic, aliphatic, and α,β-unsaturated carboxylic acids, thus confirming the wide scope of 2a towards the carboxylate partner (Scheme 2). The enol esters 5ab-an were isolated, after extraction of the reaction mixture with diethyl ether and subsequent chromatographic purification, in 61-92% yield (conversions ≥ 85% were in all cases observed by GC), with reaction times ranging from 5 to 24 h.
Additional studies varying the nature of the alkyne were also performed. Thus, as observed for hex-3-yne (3a), the addition of benzoic acid to oct-4-yne (3b) and but-2-yne (3c) also proceeded cleanly under the standard conditions, leading to the corresponding enol esters 5ba and 5ca, which were isolated in 82-92% yield (see Scheme 4). In contrast, a very poor result was obtained when diphenylacetylene (3d) was employed as substrate, the reaction leading to a maximum 20% GC yield of the desired enol ester product 5da after 24 h (Scheme 4). The higher steric constraints of this particular alkyne could be behind this negative result.
To complete the study, the catalytic addition of benzoic acid (4a) to a nonsymmetrically substituted alkyne, i.e., 1-phenyl-1-propyne (3e), was finally explored. As shown in Scheme 5, the reaction proceeded in high yield after 20 h, but with a relatively low regioselectivity. Thus, a nonseparable mixture of the regioisomeric enol esters 5ea and 5ea´was formed in ca. 2.2:1 ratio. Scheme 4. Addition of different carboxylic acids to symmetrically substituted internal alkynes (isolated yields are given).

Materials and Methods
All the synthetic procedures described in this article were performed under inert atmosphere (dry argon) using vacuum-line and Schlenk or sealed-tube techniques. Organic solvents were purified by standard methods and distilled under inert atmosphere before use [37]. All reagents employed in this work were obtained from commercial suppliers and used as received, with the exception of the ferrocenylphosphino sulfonate ligands (η 5 -C 5 H 3 PR 2 (SO 3 i Pr))Fe(η 5 -C 5 H 5 ) (1a-c) [32] and the gold(I) complex [AuCl(SMe 2 )] [38], which were synthesized as previously described in the literature. NMR spectra were recorded at room temperature making use of Bruker DPX-300 or AV400 instruments (Billerica, MA, USA). The residual signal of the deuterated solvent was employed as reference for 13

General Procedure for the Addition of Carboxylic Acids to Internal Alkynes Catalyzed by Complex 2a
Under an argon atmosphere, the corresponding internal alkyne 3a-e (1.2 mmol) and carboxylic acid 4a-n (1 mmol), the gold complex [AuCl{(η 5 -C 5 H 3 PPh 2 (SO 3 i Pr))Fe(η 5 -C 5 H 5 )}] (2a) (0.014 g; 0.02 mmol), AgOAc (0.003 g; 0.02 mmol), and water (3.0 mL) were introduced into a Teflon-capped sealed tube, and the reaction mixture was stirred at 60 • C for the indicated time (see Schemes 4 and 5). The course of the reaction was monitored regularly, taking samples of ca. 5 µL, which, after extraction with CH 2 Cl 2 , were analyzed by GC. Once the reaction finished, the mixture was extracted with diethyl ether (3 x 5 mL), and the organic phase was dried over MgSO 4 and evaporated to dryness. The resulting oily residue was purified by flash column chromatography over silica gel using diethyl ether/hexane (1:10) as eluent. The identity of the enol ester products was assessed by comparison of their NMR spectroscopic data with those previously reported by us [30] or, in the case of 5ag, by Tsukada and coworkers [39]. Copies of the NMR spectra are included in the Supplementary Materials.

X-ray Crystal Structure Determination of Compound 2a
Crystals suitable for X-ray diffraction analysis were obtained by slow diffusion of hexane into a saturated solution of compound 2a in dichloromethane. The most relevant crystal and refinement data are collected in Table 2. Diffraction data were recorded on an Oxford Diffraction Xcalibur Nova single-crystal diffractometer, using Cu-Kα radiation (λ = 1.5418 Å), with a crystal-to-detector distance fixed at 62 mm and using the oscillation method, with 1 • oscillation and variable exposure time per frame of 2.5-6.5 s. The data collection strategy was calculated with the program CrysAlis Pro CCD [40]. Data reduction and cell refinement were performed with the program CrysAlis Pro RED [40]. Empirical absorption correction was applied by means of a SCALE3 ABSPACK algorithm as implemented in the program CrysAlis Pro RED [40]. The software package WINGX was used for space group determination, structure solution, and refinement [41]. The structure was solved by direct methods using SHELXL97 [42]. Isotropic least-squares refinement on F 2 using SHELXL97 was performed [42]. During the final stages of the refinement, all the positional parameters and the anisotropic temperature factors of all the non-H atoms were refined. The coordinates of the H atoms were found from different Fourier maps and included in the refinement with isotropic parameters. The function minimized was [Σw(  International Tables for X-ray Crystallography [43]. Geometrical calculations were made with PARST [44]. The crystallographic plots were made with DIAMOND [45]. Supplementary crystallographic data for the structural analysis have been deposited with the Cambridge Crystallographic Data Centre, CCDC-1961685.