Nickel-Catalyzed Transfer Hydrogenation of Benzonitriles with 2-Propanol and 1,4-Butanediol as the Hydrogen Source.

The homogeneous transfer hydrogenation of benzonitrile with 2-propanol or 1,4-butanediol produced N-benzylidene benzylamine (BBA, 85% yield) using 5 mol % [Ni(COD)2] as a catalytic precursor and a mixture of Cy2P(CH2)2PCy2 and Cy2P(CH2)2P(O)Cy2 as ancillary ligands, under mild reaction conditions (120 °C, 96 h, tetrahydrofuran). 1,4-Butanediol performed better than 2-propanol as a hydrogen donor and yielded γ-butyrolactone as the product of transfer dehydrogenation. Selectivity toward dibenzylamine (DBA, 62% yield) was achieved by varying the amount of 1,4-butanediol in the catalytic system. A reaction mechanism was proposed, involving a ligand-assisted O-H bond activation, end-on substrate coordination, and a key dihydrido-Ni(II) intermediate, leading to the in situ formation of primary imines and amines to ultimately yield the secondary imines observed.

Introduction

The catalytic pan> class="Chemical">hydrogenation of nitriles is a reaction with a high atom economy to synthesize amines and Schiff bases.[1] Because of the extended use of these derivatives as starting materials for pharmaceuticals, fungicides, agrochemicals, dyes, fragrances, and plastics,[1a,1b] the catalytic hydrogenation of nitriles is an active field and has been reviewed recently.[1c,1d] This reaction is challenging in terms of selectivity (Scheme ). The formation of primary imine intermediates (a) combined with primary amines (b) leads to the production of secondary imines (c), which undergo further hydrogenation to yield the corresponding secondary amines (d).

Scheme 1

Hydrogenation of Nitriles

Selectivity control has mainly focusepan> class="Chemical">d on primary amines, and in homogeneous catalysis, most of the reports used Ru complexes for this purpose.[1c,1d] Exploiting the use of earth-abundant metals,[2] Beller and Milstein pioneered the synthesis of primary amines with Fe,[3a−3c] Co,[3d] and Mn[3e] PNP and NNP pincers and Co[3f] phosphino complexes. Reports on the selective homogeneous partial hydrogenation of nitriles to yield secondary imines are scarce and comprise the use of Ru catalysts by Sabo-Etienne,[4a] Milstein,[4b] and Prechtl[4c] and W and Mo PNP pincers by Berke.[4d] Our group reported the hydrogenation of benzonitrile (BN) to selectively yield N-benzylidene benzylamine (BBA) using [(Pr2P(CH2)2PPr2)Ni(H)]2 as a catalytic precursor.[5]

Hydrogenation wipan> class="Chemical">th hydrogen gas generates no residues in stoichiometric amounts; however, safety and environmental concerns arise with it. Transfer hydrogenation (TH) allows using hydrogen-donating molecules that are easily available and safe to handle.[6] The catalytic TH of nitriles yielding primary amines was achieved using Pd/C,[7a] NiPd alloy nanoparticles,[7b] and CuNi nanoparticles[7c] with either HCOOH/NEt3 or ammonia borane adduct (AB). Co-Supported catalysts allowed for the synthesis of both primary amines and secondary imines with isopropyl alcohol.[7d] The homogeneous Ru-catalyzed TH with 2-butanol led to primary amines[8a] and that with 2-propanol produced N-isopropyl[8b] and N-isopropylidene-amines.[8c−8e] These N-substituted products were also obtained with the heterogeneous Raney Ni TH from 2-propanol.[8f] The use of soluble Co pincers with AB selectively yielded primary, secondary, and tertiary amines.[8g]

1,4-Butanediol is a pan> class="Chemical">hydrogen source whose production from biomass derivatives has been reported recently.[9] Its dehydrogenation yields γ-butyrolactone (GBL) along with two H2 units (Scheme ).[10] Lactones are an important functional group as they occur in natural products and polymer synthesis.[11] Thus, it has been reported that their synthesis by catalytic (transfer) dehydrogenation of diols is also an environmentally friendly protocol.[12]

Scheme 2

Dehydrogenation of 1,4-Butanediol

Regarding pan> class="Chemical">the TH from 1,4-butanediol to BN, there is only one example in the literature, in which a dihydrido-Ru-NHC complex (5 mol %) catalyzed nitrile and diol conversion to benzylamine and GBL, respectively, to undergo further condensation, followed by dehydrogenation, to yield N-phenyl succinimide.[13] In recent years, our group has been interested in the catalytic TH of the unsaturated compounds using low-valent Ni complexes and both amines and alcohols as hydrogen donors.[14] To the best of our knowledge, there are no reports on the use of low-valent Ni species in the homogeneous TH of nitriles with the use of 2-propanol or 1,4-butanediol. Herein, we report a novel methodology to yield BBA by the Ni-catalyzed TH of BN with both 2-propanol and 1,4-butanediol as the hydrogen source.

Results and Discussion

The pan> class="Chemical">TH of a benchmark BN was first assessed with an excess of 2-propanol as the hydrogen source in the presence of 5 mol % [Ni(COD)2] (COD = 1,5-cyclooctadiene) and a mixture of dcype/dcypeO/dcypeO2 (dcype = 1,2-bis(dicyclohexylphosphino)ethane; dcypeO = dcype monoxide; and dcypeO2 = dcype dioxide) in a [Ni(COD)2]/dcype/dcypeO/dcypeO2 3:3:2:1 ratio as a catalyst precursor (1) (Scheme ). Under these optimized conditions, we observed 88% conversion of BN with 97% selectivity toward the formation of secondary imine BBA (85% yield). Notably, in this catalytic system, the formation of both N-alkenyl and N-alkylated byproducts was prevented almost completely. Such species are formed by the condensation of primary amines with acetone, which is produced by the transfer dehydrogenation of 2-propanol. Additionally, the secondary amine formation [i.e., dibenzylamine (DBA)] was also negligible.

Scheme 3

Ni-Catalyzed TH of BN with Excess 2-Propanol

Encouragepan> class="Chemical">d by this result, we switched the hydrogen source to less explored 1,4-butanediol, reacting an excess of it with BN (Scheme ). In this case, the quantitative conversion of nitrile was observed after 94 h, with a selectivity of 62% toward the formation of DBA. This reaction was monitored by analyzing the conversion and selectivity using gas chromatography–mass spectrometry (GC–MS) at regular time intervals (Figure ). During the first 23 h, BBA was formed as the only product in 63% yield. In the next 71 h, the secondary imine was gradually hydrogenated to yield DBA, reaching the highest conversion of BN and a maximum yield of 62% for the secondary amine, limited by the appearance of other byproducts.[15] In view of this, we lowered the amount of 1,4-butanediol used for the TH. Compared with the use of excess 1,4-butanediol (Table , entry 1), this change enhanced the selectivity, up to 97%, selective to the production of BBA (85% yield) after 96 h (Table , entry 2).

Scheme 4

Ni-Catalyzed TH of BN with Excess 1,4-Butanediol[15]

Figure 1

GC–MS monitoring of the TH of BN with excess 1,4-butanediol.

Table 1

Ni-Catalyzed TH of BN with 1,4-Butanediol

entry	equiv of 1,4-butanediol	t/h	% conversiona	% BBA
1	11	23	63	63
2	1	96	88	85
3	0.75b	96	54	54
4c	1	96	85	85
5d	1	96	85	85

Determined using GC–MS. In all entries, GBL was observed.

Considering a 1,4-butanediol to benzonitrile molar ratio of 0.75:1 = 3:4.

Hg0 drop test.

1,4-Butanediol-d2 was used.

GC–MS monitoring of the pan> class="Chemical">TH of BN with excess 1,4-butanediol.

Determinepan> class="Chemical">d using GC–MS. In all entries, GBL was observed.

Consipan> class="Chemical">dering a 1,4-butanediol to benzonitrile molar ratio of 0.75:1 = 3:4.

Hg0 n class="Chemical">drop test.

n class="Chemical">1,4-Butanediol-d2 was usepan> class="Chemical">d.

Comparison between pan> class="Chemical">2-propanol and 1,4-butanediol shows that 1,4-butanediol is a more efficient hydrogen donor because the same BN conversion and selectivity toward BBA were achieved using a BN to 1,4-butanediol 1:1 molar ratio (Table , entry 2), whereas for 2-propanol, a 1:20 ratio, relative to BN, was used (Scheme ). At this point, we reasoned that the in situ formed benzylamine (vide supra) might also participate in the TH process. Without adding 2-propanol or 1,4-butanediol, external benzylamine was reacted with BN in a 1:1 molar ratio (see Scheme ). After 96 h, BBA was found to be the major component of the reaction mixture. BN and benzylamine were also quantified, thus showing that BBA might indeed be formed both by the condensation of in situ produced primary imine with benzylamine and by the oxidative self-coupling of benzylamine. We then used a stoichiometric 1,4-butanediol to BN ratio, but it turned out to be detrimental for the conversion of BN (Table , entry 3), thus showing that a slight excess of 1,4-butanediol is required to have an optimal performance and also that the in situ formed benzylamine is not an efficient hydrogen donor in the presence of 1,4-butanediol. Thus, the main role of benzylamine is to act as a nucleophile carrying out the condensation reaction needed to produce the observed BBA (vide infra).

Scheme 5

Ni-Catalyzed TH of BN with Benzylamine

To provide insight into pan> class="Chemical">the nature of the hydrogen transfer from 1,4-butanediol, we performed a 1H NMR monitoring of the Ni-catalyzed reaction between BN and 1,4-butanediol in a 1:1 molar ratio (Figure ). The integration of signature signals for BBA (δ 8.41)[16a] and GBL (δ 4.22)[16b] accounted for an approximate 1:3 proton ratio through all over the reaction time. Given the coefficients of the balanced redox reaction shown in Figure , these integration values are consistent with a simultaneous production of both BBA and GBL, thus discarding a dehydrogenation/hydrogenation pathway and demonstrating that an actual TH takes place. The final product of the hydrogen source is GBL, which is produced after formally losing two H2 units from 1,4-butanediol by transfer dehydrogenation. We observed 74% conversion of BN with 100% selectivity toward BBA using GC–MS analysis of the final reaction crude. This was in good agreement with the previously observed 85% chromatographic yield for the model reaction.[17]

Figure 2

1H NMR (300 MHz, THF-d8) monitoring of the Ni-catalyzed TH of BN with 1,4-butanediol.

Control tests were assayepan> class="Chemical">d excluding [Ni(COD)2], dcype, dcypeO, or dcypeO2 out of the reaction mixture, observing a significant loss in the catalytic activity when no [Ni(COD)2], dcype, or dcypeO was added (Table S1). Importantly, no inhibition of the catalytic activity occurred upon the addition of a mercury drop to the model reaction, consistent with a homogeneous behavior of the system (Table , entry 4). Additionally, with the use of deuterated 1,4-butanediol (1,4-butanediol-d2) bearing terminal O–D moieties, the yield of BBA remained unchanged (Table , entry 5), but a mixture of isotopologues was detected using an EI–MS analysis (i.e., BBA + BBA-d1, Table S2). Given these facts, no kinetic isotope effect is due to the use of O–D fragments; therefore, in the mechanistic proposal displayed in Scheme , O–H/D bond activations (a,g) are not rate-determining nor the insertions into Ni–H/D bonds (e,k) or reductive eliminations from a Ni–H/D-type intermediate (f,m), as they all involve such an isotope. Thus, in the current system, β-hydride eliminations (b,h) should be rate-determining. On the other hand, the unselective incorporation of deuterium into the carbonated skeleton of BBA accounts for a dihydrido-Ni(II) intermediate, which, according to Scheme , is formed by the oxidative addition (a,g) of the O–H/D bonds to Ni(0) followed by β-hydride elimination (b,h) from the corresponding alkoxo-ligand.

Scheme 6

Mechanistic Proposal for the TH of BN with 1,4-Butanediol

With regarpan> class="Chemical">d to the influence of the substrate electronic properties, we carried out the reaction of p-methoxy BN and p-trifluoromethyl BN with 1,4-butanediol (Scheme ). We observed a preference for the nitrile bearing an electron-releasing substituent, for which the reaction exhibited 80% yield at a lower time (72 h) than that required for BN to produce 85% BBA (96 h). On the contrary, this TH became more sluggish for the electron-withdrawing-substituted nitriles as only 30% yield of secondary imine was reached after 168 h.

Scheme 7

Influence of Electronic Properties on the TH of BN with 1,4-Butanediol

These experimental observations fit well wipan> class="Chemical">th the proposal given in Scheme , invoking a dihydrido-Ni(II) intermediate to which nitrile end-on coordinates (d). The more coordinating the nitrile, the more probable the insertion reaction into a Ni-hydride-type bond (e). The same stands for the in situ formed primary imine to yield the corresponding primary amine (j,k). The in situ generation of both primary imine and amine allows for the formation of secondary imines by condensation with the corresponding extrusion of ammonia (n). The simultaneous production of GBL (vide supra) is consistent with two dehydrogenation steps: one yielding an aldol (b) that undergoes 5-exo-trig cyclization to the corresponding lactol (c) and the other in which hemiacetal oxidizes to the lactone (g,h).

With regarpan> class="Chemical">d to the role of dcype and dcypeO, because these two species are involved in the TH (Table S1), we reasoned that dcype acts as an ancillary ligand because its strong σ-donor character favors key activation of the hydrogen source by oxidative addition and stabilizes the Ni(II) intermediates invoked in the catalytic cycle (Scheme ). With regard to the role of dcypeO in the TH, given the high hydrogen-bond-accepting character of the P=O moiety, better than that of ethers or tertiary amines,[18] such species might promote O–H bond cleavage. We thus found it reasonable to use only dcypeO2 instead of dcypeO; nevertheless, a significant decrement in the catalytic activity was observed, producing 5% BBA (Table S1, entry 4). In view of this, we propose that dcypeO also acts as a monodentate ligand, bound to Ni(0) through its trivalent phosphorus atom. All in all, dcypeO behaves as a labile ligand, allowing for the coordination of nitrile to the Ni(II) centers and promoting 1,4-butanediol and hemiacetal activation onto the Ni(0) species (Scheme ).

Scheme 8

Proposal for the Activation of the Hydrogen Source

Conclusions

We successfully performed pan> class="Chemical">the selective synthesis of BBA by the catalytic homogeneous TH of BN with 2-propanol or 1,4-butanediol under mild reaction conditions (120 °C, 96 h) using 5 mol % [Ni(COD)2] as a catalytic precursor along with a mixture of dcype and dcypeO ligands. A significant excess of 2-propanol could be readily replaced by 1,4-butanediol in a slight excess to achieve the very same performance for both catalytic systems (88% conversion of BN and 97% selectivity for BBA). GBL was shown to be the final product of 1,4-butanediol, proving that this dehydrogenates irreversibly, which accounts for its better performance. The added amount of 1,4-butanediol allowed us to control selectivity either toward BBA (85% yield), whether used in a BN to 1,4-butanediol 1:1 molar ratio, or DBA (62% yield) in a 1:11 ratio. Control tests, assays with p-substituted BNs, isotopic labeling, and the use of external benzylamine provided mechanistic insights, which stand for dcypeO-assisted O–H bond activation and a key dihydrido-Ni(II) intermediate, leading to the in situ formation of primary amines and imines to ultimately yield secondary imines. Current studies are underway in our group to develop novel Ni-based catalytic settings for related selective nitrile TH reactions.

Experimental Section

General Considerations

Unless otherwise notepan> class="Chemical">d, all manipulations were performed using standard Schlenk techniques in an inert-gas/vacuum double manifold or under an argon atmosphere (Praxair 99.998) in a MBRAUN UNIlab Pro SP glovebox (<1 ppm H2O and O2). All liquid reagents were purchased as reagent grade and degassed before use. 1,4-Butanediol was stored in the glovebox over 3 Å molecular sieves. Anhydrous BN, p-trifluoromethyl BN, p-methoxy BN, [Ni(COD)2], and 1,2-bis(dicyclohexylphosphino)ethane were purchased from Aldrich and stored in a glovebox for their use. Regular THF and isopropyl alcohol (J. T. Baker, reagent grade) were dried, distilled from sodium/benzophenone and magnesium/iodine, respectively, and stored in the glovebox before use. Deuterated solvents were purchased from Cambridge Isotope Laboratories and stored over 3 Å molecular sieves for 24 h before use. NMR spectra were recorded at room temperature on a 300 MHz Varian Unity spectrometer.

1H chemical shifts (δ/ppm) are reportepan> class="Chemical">d relative to the residual proton resonance of the solvent THF-d8 3.58 (s), 1.73 (s); CDCl3 7.26 (s). 31P{1H} NMR chemical shifts (δ/ppm) are reported relative to external 85% H3PO4. Coupling constants (J values) are given in hertz. The following abbreviations are used for the NMR data: s = singlet; d = doublet; t = triplet, m = multiplet, and br = broad. GC–MS determinations were performed using an Agilent Technologies G3171A equipped with the following column: 5% phenylmethylsilicone, 30 m × 0.25 mm × 0.25 μm.

Preparation of the Mixture dcype/dcypeO/dcypeO2

In the inert-gas/vacuum double manifold, 5 μL of H2O2 (30% v/v) was added dropwise under vigorous stirring to a THF solution of 1,2-bis(dicyclohexylphosphino)ethane (100 mg, 0.24 mmol). The solvent was evaporated under reduced pressure immediately after the addition, and the remaining white solid was dried under vacuum for 4 h. 31P{1H} NMR (121 MHz, CDCl3, δ/ppm) dcypeO2: 50.2 (s); dcypeO 49.2 (d, 3JPP = 38 Hz), 3.5 (d, 3JPP = 38 Hz); dcype 2.5 (s). Upon NMR analysis, a mixture of dcype/dcypeO/dcypeO2 in the molar ratio of 3:2:1 was obtained.

Use of 1 as a Catalytic Precursor in the TH of BN with Excess 2-Propanol

In the glovebox, a mixture of pan> class="Chemical">dcype/dcypeO/dcypeO2 in the molar ratio of 3:2:1 (6 mg, 7 μmol dcype, 4 μmol dcypeO, and 2 μmol dcypeO2) was dissolved in THF and added dropwise to [Ni(COD)2] (2 mg, 7 μmol) under stirring. To this mixture were added THF solutions of BN (14.8 mg, 0.14 mmol) and 2-propanol (172.4 mg, 2.9 mmol). In a THF total volume of 1 mL, the reaction mixture was transferred to a Schlenk tube and heated at 120 °C in an oil bath under continuous stirring. After 96 h, the crude reaction mixture was exposed to air, and 1 μL of aliquot was analyzed using GC–MS.

Use of 1 as a Catalytic Precursor in the TH of BN with 1,4-Butanediol

In the glovebox, a mixture of pan> class="Chemical">dcype/dcypeO/dcypeO2 in the molar ratio of 3:2:1 (6 mg) was dissolved in THF and added dropwise to [Ni(COD)2] (2 mg) under stirring. To this mixture were added THF solutions of BN (14.8 mg, 0.14 mmol) and 1,4-butanediol (either 12.9 mg, 0.14 mmol; 142.3 mg, 1.6 mmol; or 9.7 mg, 0.11 mmol). In a THF total volume of 1 mL, the reaction mixture was transferred to a Schlenk tube and heated at 120 °C in an oil bath under continuous stirring. After 96 h, the crude reaction mixture was exposed to air, and 1 μL of aliquot was analyzed using GC–MS.

Homogeneity Test with 1 as a Catalytic Precursor in the TH of BN with 1,4-Butanediol

In the glovebox, a mixture of pan> class="Chemical">dcype/dcypeO/dcypeO2 (6 mg) was dissolved in THF and added dropwise to [Ni(COD)2] (2 mg) under stirring. To this mixture were added THF solutions of BN (14.8 mg, 0.14 mmol) and 1,4-butanediol (12.9 mg, 0.14 mmol). In a THF total volume of 1 mL, the reaction mixture was transferred to a Schlenk tube, and a drop of distilled mercury was added. After heating for 96 h at 120 °C in an oil bath under continuous stirring, the crude reaction mixture was exposed to air and filtered to remove Hg, and 1 μL of aliquot was analyzed using GC–MS.

1H NMR Monitoring of the TH of BN with 1,4-Butanediol

In the glovebox, a mixture of pan> class="Chemical">dcype/dcypeO/dcypeO2 (6 mg) was dissolved in THF-d8 and added dropwise to [Ni(COD)2] (2 mg) under stirring. To this mixture were added THF-d8 solutions of BN (14.8 mg, 0.14 mmol) and 1,4-butanediol (12.9 mg, 0.14 mmol). In a THF-d8 total volume of 1 mL, the reaction mixture was transferred to a Wilmad tube equipped with a J. Young valve. A 1H NMR spectrum of this just prepared mixture was acquired at room temperature (t = 0 h trace in Figure ). Then, the tube was heated at 120 °C in an oil bath during successive time intervals of 24 h (t = 24 and 48 h traces) and 48 h (t = 96 h trace). After each interval, the corresponding 1H NMR spectrum was acquired at room temperature. 1H NMR (300 MHz, THF-d8, δ/ppm) benzonitrile: 7.48–7.72 (m); 1,4-butanediol: 3.46–3.58 (m), 1.49–1.59 (m); BBA: 8.41 (t), 7.71–7.81 (m), 7.30–7.41 (m), 4.77 (br); GBL: 4.21 (t), 2.20–2.37 (m). In addition, after completing 96 h of heating, the crude reaction mixture was exposed to air, and 1 μL of aliquot was analyzed using GC–MS.

TH of p-Methoxy Benzonitrile (MeOPhCN) and p-Trifluoromethyl Benzonitrile (CF3PhCN) Using Precatalyst 1 and 1,4-Butanediol

In the glovebox, a mixture of pan> class="Chemical">dcype/dcypeO/dcypeO2 in the molar ratio of 3:2:1 (6 mg) was dissolved in THF and added dropwise to [Ni(COD)2] (2 mg) under stirring. To this mixture were added THF solutions of either MeOPhCN (18.6 mg, 0.14 mmol) or CF3PhCN (23.9 mg, 0.14 mmol) and 1,4-butanediol (12.9 mg, 0.14 mmol). In a THF total volume of 1 mL, each reaction mixture was transferred to a Schlenk tube and heated at 120 °C in an oil bath under continuous stirring. After 72 h for MeOPhCN and 168 h for CF3PhCN, each crude reaction mixture was exposed to air, and 1 μL of aliquot was analyzed using GC–MS.

Preparation of 1,4-Butanediol-d2

Under anpan> unpan> class="Chemical">controlled atmosphere, 1,4-butanediol (200 mg, 2.2 mmol) was dissolved in 1 mL of deuterium oxide (D2O ≥99%, Sigma-Aldrich). 1H NMR (300 MHz, D2O, δ/ppm) 1,4-butanediol-d2: 3.47 (bs, 2H), 1.46 (bs, 2H). An additional signal was observed in δ 4.79, corresponding to H2O (bs, 1H)[19] formed by isotope exchange. The solvent was evaporated, and 1,4-butanediol-d2 was dried under vacuum for 4 h, after which it was stored in the glovebox. A new NMR spectrum was acquired. 1H NMR (300 MHz, CDCl3, δ/ppm) 1,4-butanediol-d2: 5.09 (bs, 1H), 3.56 (bs, 20H), 1.58 (bs, 20H). From these data, 95% deuterium incorporation was estimated.

Isotopic Labeling with 1,4-Butanediol-d2 Using 1 as a Precatalyst in the TH of BN

In the glovebox, a mixture of pan> class="Chemical">dcype/dcypeO/dcypeO2 in the molar ratio of 3:2:1 (6 mg) was dissolved in THF and added dropwise to [Ni(COD)2] (2 mg) under stirring. To this mixture were added THF solutions of BN (14.8 mg, 0.14 mmol) and 1,4-butanediol-d2 (12.9 mg, 0.14 mmol). In a THF total volume of 1 mL, the reaction mixture was transferred to a Schlenk tube and heated at 120 °C in an oil bath under continuous stirring. After 96 h, the crude reaction mixture was exposed to air, and 1 μL of aliquot was analyzed using GC–MS. Deuterium incorporation was confirmed by comparing the normalized abundance of the [M+]-isotopic peaks of the BBA obtained in this experiment with the average normalized abundance of the [M+]-isotopic peaks of five different MS of nondeuterated BBA obtained from five different experiments.[20] On the basis of this analysis, a mixture of isotopologues, namely, BBA + BBA-d1 is proposed.