Stereoselective Gold(I)-Catalyzed Intermolecular Hydroalkoxlation of Alkynes

Abstract

We report the use of cationic gold complexes [Au(NHC)(CH₃CN)][BF₄] and [{Au(NHC)}₂(μ–OH)][BF₄] (NHC = N-heterocyclic carbene) as highly active catalysts in the solvent-free hydroalkoxylation of internal alkynes using primary and secondary alcohols. Using this simple protocol, a broad range of (Z)-vinyl ethers were obtained in excellent yields and high stereoselectivities. The methodology allows for the use of catalyst loadings as low as 200 ppm for the addition of primary alcohols to internal alkynes (TON = 35 000, TOF = 2188 h^–1).

The development of synthetic methods for the formation of C–O bonds is of great interest in synthetic organic chemistry. A very effective approach is the addition of alcohol O–H bonds across unsaturated C–C bonds in inter- or intramolecular fashion to provide ethers.¹ To avoid the use of harsh reaction conditions² and/or the need for strong bases,³ these hydroalkoxylation reactions are usually performed employing metal-catalyzed conditions. Numerous procedures have been developed that make use of complexes of Cu,⁴ Zn,⁵ Hg,⁶ Ru,⁷ Rh,⁸ Ir,⁹ Pd,¹⁰ Pt,¹¹ Au,¹² or Th¹³ as catalysts. These catalytic procedures are desirable over substitution reactions (that would form the same products) because the method avoids the generation of stoichiometric amounts of waste.¹⁴

The rapid growth of the field of homogeneous gold catalysis has resulted in the development of a vast number of gold-catalyzed organic transformations, most of which rely on the ability of cationic gold complexes to activate unsaturated C–C bonds.¹⁵ While gold-catalyzed intramolecular hydroalkoxylation reactions have been successfully employed in the synthesis of various heterocycles¹⁶ and natural products,¹⁷ reports on the more challenging intermolecular hydroalkoxylation reactions remain scarce.¹⁸ Teles and co-workers were the first to report intermolecular hydroalkoxylation of terminal alkynes.^18a The groups of Corma and Sahoo later achieved the addition of secondary and tertiary alcohols as well as phenols to internal alkynes.^18b,18c These reports have established that internal alkynes are more challenging substrates in hydroalkoxylation reactions compared to their terminal congeners. Although monoaddition to alkynes had already been demonstrated to be difficult, other challenges remain in terms of chemoselectivity, stereoselectivity, regioselectivity, and substrate scope (Scheme 1).

Scheme 1. Common Side-Products Formed in Hydroalkoxylation Reactions.

Cationic gold complexes [Au(NHC)(CH₃CN)][BF₄] and [{Au(NHC)}₂(μ–OH)][BF₄] (NHC = N-heterocyclic carbene) have been demonstrated to be highly active catalysts in various silver- and acid-free gold-catalyzed transformations.¹⁹ Herein, we showcase their efficiency to achieve good chemo- and stereoselectivity in hydroalkoxylation reactions of internal alkynes.

We first examined the addition of 1-phenylethanol (3a) to diphenylacetylene (2a), catalyzed by 1 mol % [{Au(IPr)}₂(μ–OH)][BF₄] (1a) in toluene at 80 °C (Table 1, entry 1; IPr = 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene). After 18 h, full conversion was observed (as monitored by GC) with high chemoselectivity, 90% of the desired vinyl ether 4a and only 10% of its corresponding hydrolysis product 5,^18c,18d,20 and stereoselectivity of 98% (Z)-4a. To the best of our knowledge, hydroalkoxylation reactions have not been reported previously with secondary benzylic alcohols.^18c Moreover, we were delighted to see that the corresponding acetal, resulting from the addition of two molecules of 3a to 2a, was not detected. Interestingly, the use of Gagosz-type monogold [Au(IPr)(NTf₂)],²¹ resulted in poor reactivity (Table 1, entry 2). Gratifyingly, better chemoselectivity and faster reactions were obtained under solvent-free conditions (Table 1, entry 3). Other NHC ligands were then tested using lower catalyst loading (0.3 mol %, Table 1, entries 4–6). Catalysts 1a and 1b bearing IPr and SIPr ligands gave very similar chemoselectivities. The catalyst bearing the least electron-donating NHC ligand, 4,5-dichloro-1,3-bis(2,6-diisopropylphenyl)imid-azol-2-ylidene (IPr^Cl) [{Au(IPr^Cl)}₂(μ–OH)][BF₄] (1c) enhanced the reactivity but reduced the chemoselectivity. The use of solvate monogold complexes [Au(NHC)(CH₃CN)][BF₄] 1d and 1e as catalysts led to a large decrease in chemoselectivity (Table 1, entries 7–8). We concluded that the hydroalkoxylation reaction could be performed most effectively using 0.3 mol % of [{Au(SIPr)}₂(μ–OH)][BF₄] (1b) or [{Au(IPr^Cl)}₂(μ–OH)][BF₄] (1c) at 80 °C under solvent-free conditions. Indeed, after carrying out these hydroalkoxylation reactions for 2 h (Table 1, entries 9–10), the desired vinyl ether 4a was formed with high chemoselectivity and stereoselectivity, 96% and 95% (Z)-4a, respectively.

Table 1. Catalyst Screening with NHC-Gold(I) Complexes^a.

entry	catalyst	t	conversion
	(loading in mol %)	(h)	(%)b (4a/5)c
1d	[{Au(IPr)}2(μ–OH)][BF4] 1a (1)	18	>99 (9/1)
2d	[Au(IPr)(NTf2)] (2)	18	7
3	[{Au(IPr)}2(μ–OH)][BF4] 1a (1)	1	>99
4	[{Au(IPr)}2(μ–OH)][BF4] 1a (0.3)	0.5	51 (32/1)
5	[{Au(SIPr)}2(μ–OH)][BF4] 1b (0.3)	0.5	63 (31/1)
6	[{Au(IPrCl)}2(μ–OH)][BF4] 1c (0.3)	0.5	86 (17/1)
7	[Au(IPr)(CH3CN)][BF4] 1d (0.6)	0.5	51 (2/1)
8	[Au(IPrCl)(CH3CN)][BF4] 1e (0.6)	0.5	74 (1/1)
9	[{Au(SIPr)}2(μ–OH)][BF4] 1b (0.3)	2	>99 (9/1)
10	[{Au(IPrCl)}2(μ–OH)][BF4] 1c (0.3)	2	>99 (8/1)

^aSIPr = 1,3-bis(2,6-diisopropylphenyl)imidazolin-2-ylidene. Reaction conditions: 2a (0.50 mmol), 3a (0.55 mmol, 1.1 equiv), neat, 80 °C, in air.

^bDetermined by GC analysis, with respect to 2a.

^cDetermined by ¹H NMR spectroscopy.

^dReaction in 1 M PhCH₃.

We evaluated the performance of both 1b and 1c in the hydroalkoxylation reactions of diphenylacetylene (2a) with various secondary benzylic alcohols 3a–l (Scheme 2). Substituents at the ortho, meta, and para positions of the aryl group of 1-phenylethanol derivatives were tolerated, and the corresponding vinyl ethers 4a–e were obtained in good yields. Interestingly, the electronic properties of the aryl group of the alcohol did not affect the stereoselectivity of the reaction, and the (Z)-isomer was obtained predominantly in all cases. Importantly, the hydroalkoxylation reaction of (S)-1-phenylethanol was found to produce one enantiomer of the vinyl ether with 1b or 1c as catalysts.

Scheme 2. Hydroalkoxylation of Alkynes Using Various Secondary Benzylic Alcohols.

Reaction conditions: 2a (0.50 mmol), 3a–o (0.55 mmol, 1.1 equiv), 1b or 1c (0.3 mol %), neat, 80 °C, in air. The catalyst giving the best result, reaction time, yield of isolated product and Z/E ratio of product are given in parentheses. The results for the other catalyst are given in the Supporting Information.

Changing the methyl moiety at the α′-position of the alcohol (3f–i) required longer reaction times for the transformation to reach completion. Interestingly, substitution of this methyl group with an isopropyl (3f) or trifluoromethyl (3g) group led to a reversal in selectivity and gave instead (E)-vinyl ethers 4f and 4g as main products. We hypothesized that this change in selectivity was linked to the electronic nature of the substituent in the α′-position. The addition of methyl mandelate 3h to 2a, however, resulted again in the predominant formation of (Z)-vinyl ether 4h. Mandelonitrile 3i did not undergo the reaction, most likely because of competitive coordination of the catalyst to the triple bond of the nitrile moiety. We also tested benzhydrol (3j) and derivatives 3k,l under these reaction conditions. Longer reaction times were required to reach completion in these instances, as these possess increased steric bulk. Nonetheless, the corresponding (Z)-vinyl ethers 4j–l were obtained in modest yields with excellent stereoselectivities.

The reactivity of various symmetrical and unsymmetrical internal alkynes 2b–j was evaluated next (Scheme 3). Despite the need for longer reaction times,²² the hydroalkoxylation reactions of unsymmetrical diaryl-substituted alkynes 2b–f proceeded well. The corresponding vinyl ethers 4m–p were isolated in good yields as mixtures of regioisomers with high stereoselectivity favoring the (Z)-isomer. A preferential addition to the less electron-rich center was observed when NO₂ or MeO substituents were present, (alkynes 2b,c, vinyl ethers 4m,n), whereas a 1:1 mixture of regioisomers was obtained with 1-chloro-4-(phenylethynyl)benzene (alkyne 2d, vinyl ether 4o) that lacks such a substituent. With both MeO and Cl substituents at the para positions of the phenyl rings (alkyne 2e, vinyl ether 4p), the preference of addition to the less electron-rich center was restored.

Scheme 3. Substrate Scope for the Hydroalkoxylation of Alkynes Using Various Symmetrical/Unsymmetrical Alkynes.

Reaction conditions: 2b–2k (0.50 mmol), 3a (0.55 mmol, 1.1 equiv), 1b (0.3 mol %), neat, 80 °C, in air. Reaction time and yield are given. Z/E ratios are given in parentheses.

Alcohol condensation side-product (oxybis(ethane-1,1-diyl))dibenzene was observed as the major product.

5 was observed as the major product.

A complex mixture formed.

Next, hydroalkoxylation with symmetrical alkynes was evaluated. The reaction of strongly activated dimethylacetylene dicarboxylate (DMAD, 2f) with 3a afforded a 1:1 mixture of the desired vinyl ether 4q, with complete stereoselectivity toward the (E)-vinyl ether, along with alcohol condensation side-product (oxybis(ethane-1,1-diyl))dibenzene.²³ In agreement with the report of Teles and co-workers, the hydroalkoxylation reactions of 1,4-bis(2-thiophene)butyne 2g, 4-octyne 2h and 1,4-dichlorobutyne 2i to form vinyl ethers 4r–t were unsuccessful under these reaction conditions.^18a Replacing one phenyl group of diphenylacetylene 2a with a methyl (2j) hampered the hydroalkoxylation reaction and led to the formation of a complex mixture of products instead of the vinyl ether 4u. The hydroalkoxylation reaction of phenylacetylene 2k led to the formation of a complex mixture of products. Digold hydroxide catalysts [{Au(NHC)}₂(μ–OH)][BF₄] are known to be able to dissociate into a Lewis acidic [Au(NHC)][BF₄] and a Brønsted basic [Au(NHC)(OH)] component.²⁴ Competitive deprotonation of the acetylenic proton of phenylacetylene by the latter might explain the incompatibility with this substrate.²⁵ Attempts to form vinyl ether 4v by using non-Brønsted basic catalysts 1d or 1e, however, were unsuccessful and ketone 5 and (oxybis(ethane-1,1-diyl))dibenzene formed as the sole products.

To assess the efficiency of catalyst 1b, once the reaction between alkyne 2a and alcohol 3a was complete, iterative additions of both substrates (0.5 and 0.55 mmol, respectively) were performed. As a result, 2.5 mmol of 2a was converted to 4a over 12 h, affording a high turnover number (TON) of 840 and a modest turnover frequency (TOF) of 70 h^–1. In addition, the performance of the new digold hydroxide catalyst (1c) was evaluated by conducting the hydroalkoxylation reaction between 2a (5.0 mmol) and 3a (5.5 mmol) on a gram scale using 0.3 mol % 1c. This reaction afforded vinyl ether 4a in excellent yield (1.4 g, 94%) without loss of stereoselectivity (Z/E = 98/2).

Previous studies examining gold-catalyzed hydroalkoxylation reactions have used aliphatic alcohols such as MeOH, i-PrOH, n-BuOH, and BnOH as model substrates.^18a,18c Their addition to diphenylacetylene (2a) proceeded smoothly at room temperature using 1c as catalyst and the corresponding vinyl ethers were obtained in excellent yields and selectivities (Table 2). As reported previously, the reactivity increases from i-PrOH to MeOH by 1 order of magnitude (Table 2, entries 1 and 4).^18a These results constitute a significant improvement compared to previous catalyst systems with regards to reaction conditions, catalyst loading, and chemo- and stereoselectivity.^18c

Table 2. Addition of Aliphatic Alcohols to Diphenylacetylene 2a^a.

entry	R–OH	catalyst (mol %)	t (h)	product	yield (%)b (Z/E)c
1	i-PrOH	1c (0.3)	12	6a	>99 (100/0)
2	n-BuOH	1c (0.1)	4	6b	96 (100/0)
3	BnOH	1c (0.3)	3	6c	98 (100/0)
4	MeOH	1c (0.3)	2	6d	98 (100/0)
5	MeOH	1e (0.6)	1	6d	96 (85/15)
6	MeOH	1c (0.025)	28	6d	50 (100/0)
7	MeOH	1e (0.020)	16	6d	>99 (100/0)

^aReaction conditions: 2a (0.5 mmol), 3 (0.5 mmol, 1 equiv), neat, in air.

^bYield of isolated product.

^cDetermined by ¹H NMR spectroscopy.

We compared the performance of monogold catalyst 1e to digold hydroxide catalyst 1c for the addition of MeOH to diphenylacetylene (2a). We found that monogold catalyst 1e was more active in the addition of MeOH than digold hydroxide catalyst 1c, but the selectivity toward the (Z)-vinyl ether decreased to 85% (Table 2, entry 5). We found that monogold catalyst 1e was much more active than digold hydroxide catalyst 1c when the catalyst loading was drastically decreased (Table 2, entries 6, 7). Indeed, while the formation of vinyl ether 6d stopped after 50% conversion using 250 ppm of digold 1c, this product could be isolated in quantitative yield and complete stereoselectivity after 16 h using only 200 ppm of monogold catalyst 1e. This catalyst loading enables a very high TON of 35 000 and a TOF of 2188 h^-1.

We continued to examine the difference between catalysts 1c and 1e at different loadings in the hydroalkoxylation reaction of MeOH and DMAD (2f) (Table 3). This transformation proceeded rapidly at room temperature and reached completion after 3 h using either catalyst (Table 3, entries 1, 2). Interestingly, the corresponding (E)-vinyl ether 6e was obtained selectively. At reduced catalyst loadings, however, we observed the formation of (Z)-6e after 2 h (Z/E = 15/85 at 55% conversion for digold catalyst 1c and Z/E = 33/67 at 24% conversion for monogold catalyst 1e), which was then predominantly converted to (E)-6e after prolonged reaction time (Table 3, entries 3, 4). These results suggest that the hydroalkoxylation reaction and the subsequent isomerization are competitive processes.

Table 3. Addition of MeOH to DMAD 2f^a.

entry	catalyst (mol %)	t (h)	conversionb (%) (Z/E)c
1	1c (0.3)	3	>99 (4/96)
2	1e (0.6)	3	>99 (4/96)
3	1c (500 ppm)	6	87 (7/93)
4	1e (0.1)	6	34 (26/74)

^aReaction conditions: 2f (0.5 mmol), MeOH (0.5 mmol, 1 equiv), neat, in air.

^bConversion with respect to 2f.

^cDetermined by ¹H NMR spectroscopy.

Corma and co-workers have proposed a mechanism to account for the conversion of (Z)-vinyl ethers into (E)-vinyl ethers.^18c This mechanism involves a trans-addition of a second molecule of alcohol to the (Z)-vinyl ether and subsequent cis-elimination to form the (E)-vinyl ether (Scheme 4). Alternatively, a thermal process or rotation around the C–C bond of the vinylgold intermediate could be envisioned.²⁶ This latter route would be particularly fast for vinyl ethers from DMAD because of its ability to be involved in keto–enol tautomerization.

Scheme 4. Proposed Isomerization of (Z)-Vinyl Ethers to (E)-Vinyl Ethers^18c.

To shed light on the isomerization process, the direct isomerization reactions of pure vinyl ethers (Z)-4a and (Z)-6d catalyzed by digold catalyst 1c and monogold catalyst 1e were surveyed (Table 4). As expected from the high stereoselectivity obtained in reactions (Schemes 2 and 3) involving 1-phenylethanol (3a), isomerization of vinyl ether (Z)-4a was found to be slow (Table 4, entries 1, 2). Appreciable isomerization was only observed in the presence of 1-phenylethanol (3a) and monogold catalyst 1e (Table 4, entry 3). We found that isomerization of (Z)-6d occurred spontaneously at 80 °C and was accelerated when catalytic amounts of either the mono- or digold hydroxide catalyst was added (Table 4, entries 4–6). No appreciable isomerization was observed at lower temperatures, and the proposed acetal intermediates were never observed. These results suggest that the isomerization process from (Z)-vinyl ethers to (E)-vinyl ethers occurs spontaneously at elevated temperature and is accelerated by cationic gold species, but the process does not involve or require the addition of a second molecule of alcohol.

Table 4. Isomerization Reactions of Pure (Z)-4a and (Z)-6d^a.

entry	ether	3 (equiv)	catalyst (mol %)	t (h)	Z/Eb
1	(Z)-4a	3a (1.5)	none	24	100/0
2	(Z)-4a	3a (1.5)	1c (0.3)	24	96/4
3	(Z)-4a	3a (1.5)	1e (0.6)	24	89/11
4	(Z)-6d	none	none	1	93/7
5	(Z)-6d	none	1c (0.3)	1	80/20
6	(Z)-6d	none	1e (0.6)	1	80/20

^aReaction conditions: neat, in air.

^bDetermined by ¹H NMR spectroscopy.

We further probed whether the vinyl ether products could be transformed into other vinyl ethers. To this end, we subjected vinyl-ether 6d and CD₃OD to catalytic conditions (Scheme 5). Apart from the previously observed Z/E isomerization, no formation of acetal or incorporation of CD₃O was observed. The reverse experiment, the reaction of d₄-6d with MeOH, gave an analogous result. The incorporation of small amounts of deuterium in the vinylic position suggests the formation of a alkylgold species (as in Scheme 4) that is subsequently deuterodeaurated under these conditions.

Scheme 5. Reaction of (Z)-6d with CD₃OD.

Z/E ratios are determined by ¹H NMR analysis. Deuterium content was determined by ¹H NMR analysis and confirmed by ²D NMR analysis.

In conclusion, we have demonstrated that both [Au(NHC)(CH₃CN)][BF₄] and [{Au(NHC)}₂(μ–OH)][BF₄] complexes are highly effective catalysts for the stereoselective intermolecular hydroalkoxylation of alkynes. Their use under solvent-free conditions constitutes a practical, operationally simple, and scalable strategy for the assembly of a range of new vinyl ethers in high yields. In particular, [Au(IPr^Cl)(CH₃CN)][BF₄] (1e) has been shown to be highly active in the addition of aliphatic alcohols to internal alkynes. Experiments have also revealed that monogold and digold hydroxide catalysts display different behavior in the isomerization of the two stereoisomers of the vinyl ethers at different catalyst loadings. Further synthetic and mechanistic studies focusing on the catalytic uses of these complexes are ongoing in our laboratories.

Supporting Information Available

The following file is available free of charge on the ACS Publications website at DOI: 10.1021/cs501976s.

• Experimental procedures, optimization studies, mechanistic studies, and characterization data (PDF)

Author Contributions

^† R.M.P.V. and S.D. contributed equally to this work.

The authors declare no competing financial interest.