Ligand and Substrate Effects on Regio- and Stereoselective Ru(II)-Catalyzed Hydroacyloxylations to Vinylic Esters

Abstract

Two Ru(II)-catalytic systems were developed for anti-Markovnikov regioselective hydroacyloxylations of terminal alkynes to vinylic esters. [Ru(NCCH₃)₆][(BF₄)₂] favors (E)-vinylic ester products with arylacetylenes and select carboxylic acids, whereas a Ru scorpionate complex with two electron-withdrawing ligands favors (Z)-vinylic ester isomers.

Introduction

A direct and atom-economic route for synthesizing vinylic ethers and esters is the reaction of alkynes with alcohols (hydroalkoxylation) or carboxylic acids (hydroacyloxylation), respectively. For these reactions with terminal alkynes, nucleophilic addition may occur in a Markovnikov or anti-Markovnikov fashion, and for anti-Markovnikov addition, providing (E) and/or (Z)-stereoisomers. Methods for Markovnikov functionalizations of alkynes are generally associated with Lewis acid catalysis.^[1] Anti-Markovnikov hydroacyloxylations are catalyzed by various ruthenium and rhodium catalysts, with vinylidene carbenes proposed as reactive intermediates.^[2] (E)-Vinylic esters are favored with Ru₃(CO)₁₂ and Ru(CO)₃-diene catalysts,^[3] whereas phosphine ligands with Ru or Rh catalysts generally produce the (Z)-vinylic ester isomers (Scheme 1a).^[4]

Scheme 1.

Anti-Markovnikov syntheses of vinylic esters, and a vinylic ether from allyl alcohol.

However, the corresponding anti-Markovnikov hydroalkoxylation reactions are not well developed. For example, several Rh(I)-catalyzed methods for anti-Markovnikov alkyne hydroalkoxylation are known, but all require the alcohol component in excess, generally as solvent.^[5] In the 1990s, the Kirchner laboratory reported ruthenium tris(pyrazolyl)borate-catalyzed hydroacyloxylation, and also described the corresponding hydroalkoxylation of phenylacetylene (1) with allyl alcohol (2, Scheme 1b). In related work, they isolated Ru vinylidenes including 6, and (allyloxy)carbene 7 with the alkene pi-bonded with Ru (Scheme 1c).^[6]

Because of the variations possible for modifying the parent tris(pyrazolyl)borate (HB(pz)₃) “scorpionate” ligand,^[7] we initially elected to build on the Kirchner results, seeking an intermolecular hydroalkoxylation process in which terminal alkynes reacted with aliphatic alcohols. Although aliphatic alcohols proved unreactive with these catalysts, we observed interesting trends with the corresponding hydroacyloxylation process with carboxylic acids. Herein we describe new applications of Ru(II) complexes with and without scorpionate ligands, evaluating their reactivities for regio- and stereoselective additions of oxygen nucleophiles with terminal alkynes.

Results and Discussion

Catalyst Synthesis

We evaluated several published procedures for preparing Ru(HB(pz)₃(COD)Cl (14), finding that refluxing a suspension of [RuCl₂(COD)(CH₃CN)₂] (8) with K[HB(pz)₃] (10) was the most reliable preparation for compound 14 (Table 1).^[8] To increase electrophilic reactivity of a possible vinylidene intermediate with nucleophiles, exchanging the COD ligand of 14 with 5,5’-bis(trifluoromethyl)-2,2’-bipyridine (11) produced the novel complex 15. For a Ru scorpionate with more labile ligands, the known tris(acetonitrile) complex 16 was prepared from [Ru(NCCH₃)₆][(BF₄)₂] (9) and ligand 10.^[9] The corresponding methods were applied with metal precursors 8 or 9 and scorpionate ligands 12^[10] or 13,^[11] derived respectively from benzotriazole and 3-nitropyrazole. The non-coordinated nitrogen atoms of the benzotriazolyl ligands in scorpionates 17–18 were designed as prospective hydrogen bond acceptors with an incoming oxygen nucleophile. The electron-withdrawing nitro groups of 19–20 potentially offered another avenue to increase electrophilicity of the ruthenium catalyst and catalytic intermediates.

Table 1.

Ru scorpionate catalysts in this study.

metal source	ligand	solvent	temperature, time	compound, yield
8	10	THF	65 °C, 6 h	14, 70%
14	11	DMF	153 °C, 1 h	15, 88%
9	10	CH3CN/CH3OH	20 °C, 2 h	16, 50%
8	12	Toluene	110 °C, 4 h	17, 82%
9	12	CH3CN/CH3OH	20 °C, 2 h	18, 100%
8	13	THF	65 °C, 24 h	19, 95%
9	13	ch3cn/ch3oh	20 °C, 2 h	20, 42%

Ru scorpionate complex 15 was characterized by ¹H NMR spectroscopy in d₆-DMSO, confirming the predicted symmetry of this complex (see Supporting Information). However, the other Ru scorpionate complexes 17–20 were insoluble at room temperature, preventing effective characterization. Crystals of novel complexes 15 and 17–20 were not sufficient quality for structure determination by X-ray diffraction. Therefore, the structures proposed for 17–20 are based on analogy to established preparations of 14 and 16.^[8,9] The various solids exhibited variable reactivity, consistent with in situ preparation.

Hydroalkoxylation Experiments

Ru scorpionates 17–20 were tested for hydroalkoxylation activity on phenylacetylene (1), using equimolar amounts of methanol and benzyl alcohol as test substrates. Under a variety of conditions, from solvent-free combinations as well as solutions in toluene and dimethylacetamide, no new products were observed, even upon heating to 110 °C (see Supporting Information).

To ensure that the Ru scorpionates were reactive species with phenylacetylene (1), we screened each complex 14–20 in hydroacyloxylation with benzoic acid (22), which produced vinylic esters. The reactivity validation experiments revealed some unexpected differences in (E)/(Z)-selectivity, which we further explored.

Hydroacyloxylation (E)/(Z)-Selectivity

As a control experiment with hydroacyloxylation catalyst 14, the reaction of phenylacetylene (1) with an equimolar amount of benzoic acid (22) proceeded in excellent yield, giving a mixture of (E)-23 and (Z)-23 (Table 2, entry 1). The (E)/(Z) ratio was similar to the literature report (62 : 38),^[6a] and we also detected a trace amount of the Markovnikov vinylic ester isomer 24. 5,5’-Bis(trifluoromethyl)-2,2’-bipyridine (11) was the optimal co-ligand for CpRu-catalyzed anti-Markovnikov hydration of terminal alkynes,^[12] however Ru complex 15 gave very low conversion, although we noted that the stereoselectivity switched to favor (Z)-23 (entry 2). The hydroacyloxylation reactivity of Ru scorpionate 16 was not previously reported, but we observed that 16 was catalytically active, producing vinylic esters in modest yield, favoring (E)-23 (entry 3). The combination of 14 with co-ligand 11 increased the proportion of product (Z)-23 (entry 4). Tris(benzotriazolyl)borate catalysts 17 and 18 gave modest yields and poor (E)/(Z)-selectivity (entries 5–7), and were not further explored. In contrast, tris(nitropyrazolyl)borate catalyst 19 showed good reactivity for hydroacyloxylation, although with diminished anti-Markovnikov regioselectivity and without (E)/(Z)-selectivity (entry 8). The addition of co-ligand 11 now gave high (Z)-selectivity without substantially lowering yield (entry 9). The trisacetonitrile scorpionate catalyst 20 also gave good yield and favored (E)-23 (entry 10), with stereoselectivity switching to (Z)-23 when combined with co-ligand 11 (entry 11). As the hydroacyloxylation reactivity of the hydration catalyst [CpRu(NCMe)₃][PF₆] (21)^[12] was not reported, we observed moderate reactivity (entry 12), with a strong preference for (Z)-23 in the presence of co-ligand 11 (entry 13). Unexpectedly, the parent [Ru(NCMe)₆][(BF₄)₂] complex 9 also catalyzed alkyne hydroacylation, with significant regioselectivity for anti-Markovnikov product, and preference for the (E)-23 vinylic ester stereoisomer (entry 14).

Table 2.

Hydroacyloxylation activity of Ru catalysts.

entry	catalyst	co-ligand	yield[a]	isomer ratio[b] (E)-23 : (Z)-23 : 24
1	14	–	96%	61 : 36 : 3
2	15	–	1%	24 : 74 : 2
3	16	–	41%	68 : 25 : 7
4	16	11	29%	24 : 72 : 4
5	17	–	39%	35 : 59 : 6
6	18	–	25%	43 : 50 : 7
7	18	11	35%	31 : 66 : 3
8	19	–	79%	46 : 42 : 12
9	19	11	71%	12 : 88 : 0
10	20	–	64%	68 : 24 : 8
11	20	11	85%	22 : 70 : 8
12	21 [c]	–	56%	60 : 40 : 0
13	21	11	35%	17 : 83 : 0
14	9	–	100%	73 : 20 : 7

^[a]Combined yield of isomers (E)-23, (Z)-23, and 24, after chromatographic separation from other materials.

^[b]Vinylic ester isomers (E)-23, (Z)-23, and 24 were chromatographically inseparable. Isomer ratios were determined by ¹H NMR integration of diagnostic alkene resonances in the product mixture, prior to chromatographic purification.

^[c]21=[CpRu(NCMe)₃][PF₆].

The absence of hydroalkoxylation reactivity with these catalysts was disappointing. However, the hydroacyloxylation activity of [Ru(NCCH₃)₆][(BF₄)₂] (9) was serendipitous. Previous applications of [Ru(NCMe)₆]²⁺ complexes have been limited to ligand substitution reactions.^[9][13] The preference for the (E)-vinylic ester stereoisomer with this simple catalyst 9 aligns with literature reports for terminal alkyne hydroacyloxylations catalyzed by Ru₃(CO)₁₂ and Ru(CO)₃-diene complexes.^[3] With the ruthenium scorpionate complexes 14–20, we observed that (E)-selectivity increased when COD and Cl ligands were replaced by (CH₃CN)₃ (Table 2, entries 1 vs. 3, and entries 8 vs. 10). (E)-selectivity decreased with other scorpionate ligands relative to the parent complex 14 (compare entry 1 vs. entries 5 and 8). A complementary trend for (Z)-selectivity with Ru scorpionate 19 enhanced by the bipyridine co-ligand 11 was consistent with other reports in which sterically hindered ligands on ruthenium were associated with higher selectivity for the anti-Markovnikov (Z)-vinylic ester.^[4,14]

Hydroacyloxylation Substrate Scope

We explored the scope of hydroacyloxylations catalyzed by the (E)-selective [Ru(NCMe)₆][(BF₄)₂] catalyst 9, compared with our most (Z)-selective catalyst from Ru scorpionate 19+co-ligand 11 (Table 3). Overall vinylic ester yields were generally higher with catalyst 9, although the combination of 19+co-ligand 11 uniformly exhibited higher regioselectivity for anti-Markovnikov products. With phenylacetylene and cyclohexanecarboxylic acid (25), catalyst 9 vs. 19+11 exhibited complementary (E) vs (Z)-stereoselectivities (Table 3, entries 1, 2). Reactions with electron-rich p-methoxybenzoic acid (26) gave higher (E)-selectivity with 9 (entry 3), and slightly lower (Z)-selectivity with 19+11 (entry 4). In contrast, electron-deficient p-trifluoromethylbenzoic acid (27) and sterically hindered 2,4,6-trimethylbenzoic acid (28) gave lower (E)-selectivity with 9 (entries 5, 7), but high (Z)-selectivity with 19+11 (entries 6, 8).

Table 3.

Substrate scope, stereo- and regioselectivities of Ru-catalyzed hydroacyloxylations.

entry	alkyne	carboxylic acid	catalyst	combined yield[a]	isomer ratio[b] (E) : (Z) : Markovnikov	vinylic ester products
1			9	79%	68 : 27 : 5	(E)-33 : (Z)-33 : 34
2			19+11	65%	19:81 :0
3			9	63%	81 : 13 : 6	(E)-35 : (Z)-35 : 36
4			19+11	41%	22:77 :1
5			9	53%	57 : 23 : 20	(E)-37 : (Z)-37 : 38
6			19+11	41%	15 : 83 : 2
7			9	83%	55 : 39 : 6	(E)-39 : (Z)-39 : 40
8			19+11	73%	15 : 85 : 0
9			9	73%	71 : 25 : 4	(E)-41 : (Z)-41 : 42
10			19+11	70%	20:79 :1
11			9	81%	33 : 13 : 54	(E)-43 : (Z)-43 : 44
12			19+11	82%	17:68 :15
13			9	69%	50 : 35 : 15	(E)-45 : (Z)-45 : 46
14			19+11	35%	37:50 :13
15			9	41%	33 : 29 : 38	(E)-47 : (Z)-47 : 48
16			19+11	73%	31:48 :21

^[a]Combined yield of (E), (Z), and Markovnikov isomers, after chromatographic separation from other materials.

^[b]Vinylic ester isomers were chromatographically inseparable. Isomer ratios were determined by ¹H NMR integration of diagnostic alkene resonances in the product mixture, prior to chromatographic purification.

The effective scope of terminal alkynes was more limited: electron-deficient p-trifluoromethylphenylacetylene (29) maintained the trend of (E)- vs. (Z)-selectivity with the two complementary catalysts (entries 9, 10), however, electron-rich p-methoxyphenylacetylene (30) gave substantial amounts of Markovnikov addition isomer 44 with catalyst 9 (entry 11), with only modest selectivity for (Z)-43 with catalyst 19+11 (entry 12). Anti-Markovnikov regioselectivity returned with the branched cyclohexylacetylene (31), but with poor (E)/(Z)-selectivity (entries 13, 14). Catalytic reactions of 1-decyne (32) gave non-selective formation of vinylic ester products 47 and 48 (entries 15, 16).

Conclusion

Although the ruthenium catalysts described in this study did not promote our original goal of alkyne hydroalkoxylation, they catalyze anti-Markovnikov alkyne hydroacyloxylation favoring (E)- or (Z)-vinylic esters. The ligand effects on (E)/(Z) stereoselectivity in the hydroacyloxylation reaction are noteworthy. The simple [Ru(NCMe)₆][(BF₄)₂] complex 9 generally gives (E)-selectivity, whereas Ru(II)-complexes of electron-deficient scorpionate ligands in combination with an electron-deficient bipyridine favor (Z)-selectivity. The best substrate combinations involve electron-deficient aromatic terminal alkynes or electron-rich carboxylic acids.

Experimental Section

General Procedure for Alkyne Hydroacyloxyation with [Ru(NCMe)₆][(BF₄)₂] (9)

Under argon atmosphere, ruthenium catalyst 9 (2 mol %), terminal alkyne (1 mmol) and carboxylic acid (1 mmol) were added to anhydrous oxygen-free toluene (1.5 mL). The mixture was then heated to reflux for 24 h. After cooling, the solvent was removed by rotary evaporation. Isomer ratios were determined by ¹H NMR integration of diagnostic alkene resonances in the crude reaction mixture. Isolated yields were determined after silica gel column chromatography with 8 : 2 pentane : ethyl acetate eluent, to provide the vinylic ester products as chromatographically inseparable mixtures of (E)-, (Z)-, and Markovnikov isomers.

General Procedure for Alkyne Hydroacyloxyation with Ru[HB(3-NO₂pz)₃](COD)Cl (19) and Co-Ligand 5,5’-bis(trifluoromethyl)-2,2’-bipyridine (11)

Under argon atmosphere, ruthenium catalyst 19 (2 mol %), co-ligand 11 (2 mol %), terminal alkyne (1 mmol) and carboxylic acid (1 mmol) were added to anhydrous oxygen-free toluene (1.5 mL). The mixture was then heated to reflux for 24 h. After cooling, the solvent was removed by rotary evaporation. Isomer ratios were measured by ¹H NMR integration of the reaction mixture, and isolated yields were obtained as inseparable mixtures of (Z)-, (E)-, and Markovnikov isomers.

Supplementary Material

Supporting information for this article is available on the WWW under https://doi.org/10.1002/adsc.202301222