Ruthenium-Catalyzed E-Selective Alkyne Semihydrogenation with Alcohols as Hydrogen Donors

Selective direct ruthenium-catalyzed semihydrogenation of diaryl alkynes to the corresponding E-alkenes has been achieved using alcohols as the hydrogen source. The method employs a simple ruthenium catalyst, does not require external ligands, and affords the desired products in > 99% NMR yield in most cases (up to 93% isolated yield). Best results were obtained using benzyl alcohol as the hydrogen donor, although biorenewable alcohols such as furfuryl alcohol could also be applied. In addition, tandem semihydrogenation–alkylation reactions were demonstrated, with potential applications in the synthesis of resveratrol derivatives.


■ INTRODUCTION
The alkene motif is present in a variety of important molecules, including natural products, pharmaceuticals, and fragrances ( Figure 1). 1 Stereoselective installation of this functionality therefore remains central to organic synthesis. Semihydrogenation of alkynes is a natural synthetic transformation to obtain alkenes. However, E-selective alkyne semihydrogenations have historically been more difficult to achieve than Z-selective. The former transformation has typically been limited to alkynes bearing alcohols, amines, or ketones in the propargylic position, generally requiring stoichiometric reagents 2 or proceeding via two-step methods such as trans-hydrosilylation followed by protodesilylation. 3 Lately, hydrogenation based on homogeneous transitionmetal catalysis has begun to offer a remedy to these limitations. 2 For example, iron, 4 cobalt, 5 nickel, 6 palladium, 7 manganese, 8 and iridium 9 have been used to obtain alkenes from alkynes with E-selectivity. In particular, an iridiumcatalyzed method for alkyne semihydrogenation recently reported by Yang et al. deserves highlighting 9a as it allows the selective formation of either the E-or the Z-alkene isomer simply by adding a bulkier ligand (COD) to the reaction system in the latter case. In addition, an inexpensive and sustainable alcohol (ethanol) is used as the hydrogen source. A few accounts of direct ruthenium-based E-selective semihydrogenations of alkynes have also been published in the past decade (Scheme 1). 10 Several of the reported ruthenium systems require elevated temperatures (145−180°C) 10c,h or stoichiometric or excess amounts of organic acids (1−50 equiv). 10c,k Despite displaying good substrate scope in the presence of other reductive-sensitive functional groups, these harsh reaction conditions could limit the utility of these methods. Milder methods for semihydrogenations based on ruthenium have recently been described. 10d,j Furstner and coworkers used high pressures of dihydrogen (10 bar) with silver triflate as an additive at ambient temperatures 10d or propargylic alcohols as substrates at lower pressures (1 bar), 10l while Lindhardt recently published a method proceeding at 45°C with dihydrogen generated in situ in a closed two-chamber system. 10j Djukic et al. have shown that μ-chlorido and μhydroxobridged ruthenacycles can affect the hydrogenation of triple bonds using isopropanol as the hydrogen donor at 90°C , 10b while Gelman has reported an elegant semihydrogenation of alkynes involving ligand−metal cooperation as the mode of action, using a ruthenium catalyst and a mixture of formic acid and sodium formate as the hydrogen source. 10i By adding D 2 O, this procedure could also be applied toward deuterium labeling.
While conducting a ruthenium-catalyzed "borrowing hydrogen" reaction involving alcohols and amines in the presence of an alkyne functionality, 11 we noticed that small amounts of the corresponding alkene were formed. We envisioned that a transfer hydrogenation between the alcohol and the alkyne competed with the borrowing hydrogen reaction to a minor extent. Indeed, in 1981, Shvo and co-workers presented a ruthenium-catalyzed oxidative ester formation from alcohols using diphenylacetylene as a hydrogen acceptor. 10a Despite recent reports on alkyne semihydrogenations, the scope of ruthenium-catalyzed transfer hydrogenation between alcohols and alkynes has, to the best of our knowledge, not been investigated in detail. 10b We herein present a relatively mild semihydrogenation of alkynes, which can be performed without the necessity of external ligands or stoichiometric amounts of organic/inorganic acids or bases. The procedure performs well with diaryl acetylenes and is experimentally facile, using only commercially available reagents and without the need for any special equipment.

■ RESULTS AND DISCUSSION
For the initial investigation of the transfer hydrogenation between alcohols and alkynes, diphenylacetylene (1a) was used as a model substrate (Scheme 2). A selection of different alcohols were screened for efficiency, E/Z selectivity, and their ability to avoid over-reduction. The reaction was performed in the presence of a simple ruthenium catalyst, Ru 3 (CO) 12 , and initially with stoichiometric amounts of tBuOK as a base, using toluene as the solvent and heating the reactions in a heating block. Experiments were analyzed by 1 H NMR using 2,5dimethylfuran as an internal standard. 12 Of the screened hydrogen donors, benzylic alcohols (benzyl and furfuryl alcohol, 1-phenylethanol) stood out both in terms of selectivity and efficiency. In particular, benzyl alcohol produced Estilbene ((E)-2a) with 100% selectivity over Z-stilbene ((Z)-2a) while only generating ∼2% bibenzyl (3) via overreduction. A number of other alcohols also displayed good compatibility with the reaction. Cyclopentanol generated the semihydrogenation product in good yields, with only minor over-reduction, while longer noncyclic secondary aliphatic alcohols (2-butanol, 3-pentanol) reacted sluggishly. Isopropyl alcohol and ethanol both showed a good conversion to alkene, while the more hindered neopentyl alcohol and glycerol reacted slowly. Interestingly, the reactivity of isopropyl alcohol could be greatly enhanced by introducing a methoxy group in the 1-position, generating a substantial amount of bibenzyl. Control experiments were also performed. Excluding the base from the reaction significantly lowered the efficiency, affording 13% of (Z)-2a and no other products. The alcohol and catalyst, as expected, proved to be essential to the reaction, with no products formed in their absence.
While benzyl alcohol outperformed the other hydrogen donors, the generation of reactive benzaldehyde in situ could under some circumstances be problematic due to its potential reactivity with nucleophiles. Isopropyl alcohol, on the other hand, forms acetone, which is less prone to adduct formation with nucleophiles. Additionally, compared to benzyl alcohol and benzaldehyde, both isopropyl alcohol and acetone can be easily removed through evaporation, thus expediting the purification of the product. Further optimization was thus performed using isopropyl alcohol as the hydrogen donor, aiming to improving the E/Z selectivity and yield.
In addition to Ru 3 (CO) 12 , nine other commercially available ruthenium catalysts were screened using isopropyl alcohol as the hydrogen donor (Table 1). The reactivity of the catalysts varied from very low when using RuCl 3 (entry 2), Cp*RuCl-(COD) (entry 5), or the Shvo catalyst (entry 9 and Figure 2) to being higher but unselective for the semihydrogenation product when RuCl 2 (PPh 3 ) 3 was employed (entry 7). The Grubbs first-generation catalyst ( Figure 2) gave the fully reduced bibenzyl product 3 with nearly complete selectivity in a good yield (entry 4). However, our interest lays in the selective semihydrogenation to form the (E)-2a. In this context, catalyst RuCl 2 (DMSO) 4 displayed good properties, with a combined yield of 91% and 6:1 in terms of the E/Z selectivity (entry 3). RuCl(CO)H(PPh 3 ) 3 also performed well, affording only (E)-2a in a good yield, albeit with some concomitant over-reduction to 3 (entry 6). Viable catalysts for the E-selective semihydrogenation of diphenylacetylene, using isopropyl alcohol as the hydrogen donor, were thus found to be Ru 3 (CO) 12 , RuCl 2 (DMSO) 4 , and RuCl(CO)H(PPh 3 ) 3 . RuCl 2 (DMSO) 4 was selected for further studies when using isopropyl alcohol as the hydrogen donor.
With two catalyst systems in hand, that is, Ru 3 (CO) 12 / benzyl alcohol and RuCl 2 (DMSO) 4 /iPrOH, further studies concerning the loading of catalyst, base, and hydrogen donor were performed (Table 2). For Ru 3 (CO) 12 /benzyl alcohol, using a catalyst amount corresponding to 2 mol % Ru and reducing the amount of base to 0.2 equiv did not affect the yield (entries 1 and 4), while lowering the amount of alcohol (entry 5) or temperature (entry 6) had a negative effect on the yield and E/Z selectivity. Interestingly, reducing the amount of catalyst while maintaining the base at 1 equiv decreased the yield of the alkene (entry 2). Hence, the activity of the catalyst is related to the relative amount of base. The same behavior was observed when using a 5 mol % catalyst (entry 3). RuCl 2 (DMSO) 4 /iPrOH was also evaluated but displayed a much slower reaction rate. Reducing the amount of catalyst, base, and alcohol dramatically reduced the yield within the investigated time frame of 24 h (entries 7−10).
The optimized conditions for Ru 3 (CO) 12 /benzyl alcohol (Method A) were then applied to a series of alkynes (1a−n, Scheme 3) to investigate the scope. Diaryl acetylenes with varying electronic properties were well tolerated and formed their corresponding hydrogenated E-isomers selectively, with close to quantitative conversion (as determined by 1 H NMR) and high isolated yields (compounds 2a−f). Electron-rich compounds such as 1c reacted slightly slower, and longer reaction times were needed to achieve full conversion. Primary amines and pyridines (1g−i) proved to be more challenging substrates. The hydrogenation of the p-amino derivative proceeded sluggishly under the standard conditions. Increasing the catalytic loading 4-fold gave a satisfactory hydrogenation yield, accompanied, however, by the formation of substantial amounts of another compound (Scheme 4). Interestingly, further analysis showed that this compound resulted from a hydrogen borrowing process 13 between benzaldehyde, formed in situ from the benzyl alcohol hydrogen donor and the primary amine, to form an intermediate imine that could be reduced to the corresponding amine 4 (Scheme 4) using a second equivalent of hydrogen. The fact that a concomitant semihydrogenation−amine alkylation process is feasible is not surprising as Ru 3 (CO) 12 has been employed for the direct amination of alcohols via hydrogen borrowing under similar conditions. 14 This tandem process could potentially be applied toward the synthesis of resveratrol derivatives such as 5, reported as a promising lead compound for the treatment of Alzheimer's disease. 15 Switching to different reaction conditions, utilizing iPrOH as the hydrogen donor with RuCl 2 (DMSO) 4 as the catalyst (Method B), suppressed the competing hydrogen borrowing reaction, allowing isolation of alkene 2g in a moderate yield. The more sterically challenging ortho-amine could be reduced using Method A but required a higher catalytic loading to proceed (compound 2h). In this case, the hydrogen borrowing product was not observed, most likely owing to the more hindered position of the amino group in the substrate. Similar to the other nitrogen-containing compounds, 3-(phenylethynyl)pyridine also required a higher catalytic loading and also a longer reaction time but afforded 2i in a high NMR yield. The lack of reactivity is most likely due to deactivation of the catalyst through coordination by nitrogen. This could also explain the lack of reports on the rutheniumcatalyzed semihydrogenation of aniline-containing compounds. The protons ortho to the nitrogen displayed broad signals in 1 H NMR after completion of the reaction, indicating coordination. The stability of this interaction was further validated as it was maintained even after column chromatography on silica. The ruthenium could be removed by chromatography on amine-functionalized silica, supplying the pure semihydrogenation product with some loss in yield due to the more elaborate purification required. Other heterocyclic alkyne substrates were more successful, with indole-and thiophene-derivatives 2j and 2k formed in 56 and 76% yields, respectively. A ferrocenyl-substituted E-alkene (2l) could be obtained in a moderate yield, while appending an ester substituent to diphenyl acetylene was unproblematic (2m), Reaction and conditions as in Scheme 2 but with different catalysts. Isopropyl alcohol was used as the hydrogen donor. Catalyst amount corresponds to 10 mol % Ru. b NMR yield (2,5-dimethylfuran as internal standard). c See Figure 2.   (Figure 3), instead of the expected semihydrogenation product. This product is most likely also the result of a hydrogen borrowing-type mechanism (as for 4) but in this case involving carbon−carbon bond formation instead of amine alkylation. Method B instead affected   The Journal of Organic Chemistry pubs.acs.org/joc Article concomitant alkyne semihydrogenation and transfer hydrogenation of the ketone, producing alcohol 2n in a moderate yield. In terms of limitations of the reaction, alkyl/aryl substitution of alkynes and dialkylacetylenes was unsuccessful, showing both low reactivity and formation of byproducts. Analysis of the crude products by 1 H NMR showed that while some alkene was formed in the reaction, double bond isomerization had also occurred, resulting in a mixture of products. In addition, while a p-CF 3 substituent on diphenylacetylene was well tolerated (2d), the corresponding p-NO 2 compound afforded a complex mixture, where some concomitant reduction of the nitro group had taken place.
Terminal alkynes such as 1-ethynyl-4-methoxybenzene afforded a complex mixture, with only trace amounts of products.
The reaction of diphenylacetylene with benzyl alcohol, using Ru 3 (CO) 12 as the catalyst, could be monitored over time using 1 H NMR, which revealed an initial hydrogenation to form the Z-isomer that underwent an isomerization process to the Eisomer ( Figure 4). This observation is in line with previous reports. 10c,k,16 The isomerization was further investigated by subjecting cisstilbene to the standard reaction conditions in the presence of deuterated benzyl alcohol (Bn-OD). Z-Stilbene ((Z)-2a) was isomerized into E-stilbene ((E)-2a) under these conditions but without incorporation of deuterium (Scheme 5). This observation differs from the recent study by Lindhardt and co-workers 10j in which they found that isomerization of (Z)-2a in the presence of a ruthenium catalyst and D 2 results in incorporation of deuterium at the alkenylic positions. We further found that the isomerization to (E)-2a occurred in the presence of the catalyst alone. These results indicate that the isomerization process does not proceed via a hydrogenation/ rotation/β-hydride elimination route. No isomerization was observed when omitting the catalyst while including the other reactants. Both benzaldehyde and benzyl benzoate were observed as side products after the transfer−hydrogenation reaction. Benzyl benzoate is likely formed via a second reaction between benzaldehyde and benzyl alcohol with subsequent oxidation, as previously reported by Shvo. 10a

■ CONCLUSIONS
In conclusion, a methodology for the selective semihydrogenation of diaryl alkynes to E-alkenes was developed, involving the use of a simple Ru catalyst, a low catalyst loading, ligandfree conditions, and alcohols as the source of hydrogen. While benzyl alcohol gave the most favorable E-selectivity and conversion, renewable alcohols such as furfuryl alcohol could also be applied as hydrogen donors with good results. A tandem semihydrogenation−amine alkylation reaction, the latter via hydrogen borrowing, was also demonstrated, using 4-(phenylethynyl)aniline (1g) as the substrate. Reaction monitoring indicates that the high E-selectivity in the semihydrogenation is due to isomerization of initially formed Z-alkene by the catalyst, rather than a result of the semihydrogenation process itself.

■ EXPERIMENTAL SECTION
General Remarks. All reactions were carried out under an argon atmosphere with dry solvents in oven-dried glassware, unless otherwise noted. Toluene, triethylamine (Et 3 N), ethanol (EtOH), ethyl acetate (EtOAc), and petroleum ether were bought from commercial vendors. Toluene was purchased in anhydrous form and used without further purification. Et 3 N was dried over molecular sieves (3 Å). EtOH, EtOAc, and petroleum ether were used as received. Reagents as well as alkynes 1a and 1c were purchased from commercial vendors and used as received, unless otherwise stated. For the Sonogashira reaction, oxygen-free Et 3 N was obtained by bubbling argon through the solvent for 15 min. Reactions were monitored by thin-layer chromatography carried out on 0.25 mm E. Merck silica gel plates (60F-254) using UV light as the visualizing agent. Flash chromatography was performed on a Biotage Isolera One using Biotage KP-Sil columns (packed with 50 μm irregular silica) using 254 nm and 280 nm UV light for monitoring. NMR spectra were recorded on samples in deuterated chloroform (CDCl 3 ) or DMSO (DMSO-d 6 ) on an Agilent 400 MHz (101 MHz for 13 C) instrument. Residual undeuterated chloroform ( 1 H: δ = 7.26 ppm, 13 C: δ = 77.2 ppm) or DMSO ( 1 H: δ = 2.50 ppm, 13 C: δ = 39.5 ppm) were used as the internal reference. The following abbreviations, or a combination thereof, were used to characterize the multiplicities: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, and br = broad. Melting points (mp) were recorded on a Mettler FP 90/82 melting point apparatus and were uncorrected. IR spectra were recorded with a PerkinElmer Spectrum ONE FT-IR spectrometer using KBr pellet sample preparation. High-resolution mass determinations were obtained with an Agilent QTOF 6520 with Infinity UHPLC and electrospray ionization.
General Procedure for the Preparation of Internal Alkynes 1b and 1d−n via Sonogashira Reaction. Arylhalide, bis-(triphenylphosphine)palladium(II) dichloride (Pd(PPh 3 ) 2 Cl 2 ), and copper(I) iodide (CuI) (see each compound for amounts) were transferred to a dry 20 mL Biotage microwave reaction vial equipped with a cross-shaped magnetic stirring bar. The vial was sealed using a cap with septum, evacuated of air, and refilled with argon (three cycles). The alkyne and dry deoxygenated Et 3 N were thereafter

Scheme 5. Investigation of the Isomerization Process
The Journal of Organic Chemistry pubs.acs.org/joc Article transferred to the vial. The obtained mixture was further deoxygenated by bubbling argon through for 5 min while stirring. The argon inlet was removed and the reaction was heated in a Radleys Heat-On block to 80°C for an indicated amount of time. The reaction was cooled to room temperature and concentrated under reduced pressure. The crude product was taken up in approximately 5 mL CH 2 Cl 2 and the slurry was transferred to a 3 g Biotage KP-Sil samplet. After allowing the samplet to dry, it was transferred to a 25 g column and purified by flash chromatography. 1-Methoxy-4-(phenylethynyl)benzene (1b). 17 The reaction was performed according to the general procedure using 4-iodoanisole (1. 2-(Phenylethynyl)aniline (1h). 22 The reaction was performed according to the general procedure using 2-iodoaniline (1.1 g, 5 mmol), Pd(PPh 3 ) 2 Cl 2 (70 mg, 0.1 mmol), CuI (19 mg, 0.1 mmol), phenylacetylene (0.66 mL, 6 mmol), and Et 3 N (15 mL). Flash chromatography gradient: petroleum ether/EtOAc, 1:0 to 1:0 (5 column volumes) to 85:15 ( (4-Chlorophenylethynyl)ferrocene (1l). 26 The reaction was performed according to the general procedure using 4-bromochlorobenzene (0.618 mL, 3.2 mmol), Pd(PPh 3 ) 2 Cl 2 (105 mg, 0.15 mmol), CuI (28 mg, 0.15 mmol), ethynylferrocene (1.0 g, 4.8 mmol), and Et 3 N (16 mL). Flash chromatography gradient: petroleum ether/ EtOAc, 1:0 to 98:2 (10 column volumes) to 98:2 (10 column volumes). Product 1l was obtained as a red crystalline solid (634 mg, Semihydrogenation of Internal Alkynes, Method A. To an oven-dried 5 mL Biotage microwave reaction vial equipped with a magnetic stirring bar was transferred alkyne, Ru 3 (CO) 12 , and tBuOK (see each compound for amounts). The vial was sealed with a Biotage cap and connected to a Schlenk line. The atmosphere was evacuated and the vial was refilled with argon (three cycles). Dry toluene and benzyl alcohol were subsequently transferred (no special precautions were taken to exclude air from these components). The Schlenk connection was removed and the sealed system was heated in a Radleys Heat-On block to 100°C. After being stirred at that temperature for an indicated period of time, the reaction was allowed to cool to room temperature and toluene was removed under a stream of N 2 . The reported NMR yields were obtained using 2,5dimethylfuran as the internal standard. 29 Everything was taken up in ∼2 mL of CDCl 3 and a 1 H-NMR was recorded (no. of transients: 2, relaxation delay: 60 s). The spectrum was phase-corrected and baseline-corrected before being integrated. The amount of product was calculated as previously described using the 2,5-dimethylfuran H Ar peak at δ 5.87 ppm and appropriate product peaks. As an example, the hydrogenation of phenylacetylene with cyclopentanol as the hydrogen donor can be found in the Supporting Information ( Figure S1). The CDCl 3 was after analysis evaporated and the crude mixture was taken up in ∼1−3 mL of DCM and transferred to either a 1 g or 3 g Biotage KP-Sil samplet. After drying the samplet, it was transferred to a 10 g or a 25 g column and purified through flash chromatography.
Method B. Method B was the same as Method A but used RuCl 2 (DMSO) 4 (10 mol %) as the catalyst, iPrOH (10 equiv) as the hydrogen donor, and 50 mol % tBuOK as the base instead.