Catalytic Hydrogenation of Sorbic Acid using Pyrazolyl Palladium(II) and Nickel(II) Complexes as Precatalysts

We have prepared several pyrazolyl palladium and nickel complexes ([(L1)PdCl2] (1), [(L2) PdCl2] (2), [(L3) PdCl2] (3), [(L1) NiBr2] (4), [(L2) NiBr2] (5) and [(L3) NiBr2] (6)) by reacting 3,5-dimethyl-1H-pyrazole (L1), 3,5-di-tert-butyl-1H-pyrazole (L2) and 5-ferrocenyl-1H-pyrazole(L3) with [PdCl2(NCMe)2] or [NiBr2(DME)] to afford mononuclear palladium and nickel complexes, respectively. These complexes were then investigated as pre-catalysts in the hydrogenation of 2,4-hexadienoic acid (sorbic acid). The active catalysts from these complexes demonstrate significant activities under mild experimental conditions. Additionally, the active catalysts show that the hydrogenation of sorbic acid proceeds in a sequential manner, where the less hindered C=C bond (4-hexenoic acid) is preferentially reduced over the more hindered C=C bond (2-hexenoic acid).


Introduction
Hydrogenation of α,β-unsaturated compounds has been widely employed in the vitamins, fragrances, pharmaceuticals, petrochemicals, agrochemicals, and cosmetics industries. 1 One of the extensively used transition metal catalysts in these hydrogenation reactions is chlorotris(triphenylphosphine)rhodium(I), [RhCl(PPh 3 ) 3 ]. 2,3 The catalyst, RhCl(PPh 3 ) 3 , catalyzes the chemo-specific hydrogenation of C=C bonds in the presence of other easily reduced groups, like nitro (NO 2 ) or carbonyl (CHO), as well as terminal alkenes even when the substrate has internal alkenes. 2,3 Other transition metals complexes have been extensively studied as heterogeneous and homogeneous catalysts in the catalytic hydrogenation of olefins and α,β-unsaturated compounds. Among these metal complexes are ruthenium, 4,5 rhodium, 6 iridium, 7 and platinum. 8 However, nickel 9 and palladium 10 complexes have in recent times gained considerable attention as efficient catalysts in hydrogenation reactions. Shevlin et al. described the first homogeneous nickel-catalyzed asymmetric hydrogenation of α,β-unsaturated esters, using molecular hydrogen, that gave high yield and high enantioselective products. 11 Apart from the good reactivity and selectivity of the nickel catalysts, ligand manipulation makes them attractive for homogeneous catalysts. Nickel catalysts are cheap and cost-effective.
On the other hand, palladium catalysts are more expensive but exhibit superior catalytic and selectivity properties in the hydrogenation of unsaturated compounds. 12 For example, P ∧ C ∧ P palladium pincer complexes are highly active catalysts for the chemo-selective transfer hydrogenation of α,β-unsaturated ketones. 13 These highly reactive and selective palladium pincer complexes afforded saturated ketones from α-enones. 13 Similarly, Bacci et al. reported hydrazinic-phosphine(P ∧ N)palla-dium(II) complexes as efficient catalysts for C=C bonds hydrogenation under mild experimental condition. 14 However, because phosphines are sensitive to air and moisture, 15 palladium complexes with nitrogen-donor ligands are emerging as an alternative to phosphorus-donor palladium complexes as hydrogenation catalysts. For example, {bis(arylimino)acenaphthene}-palladium(0) complexes are known to be efficient and highly chemo-selective in the hydrogenation of C=C bonds of α,β-unsaturated aldehydes. 16 But despite numerous nitrogen-donor nickel(II) and palladium(II) applications in catalysis, little work has been reported on their catalytic properties in the hydrogenation of α,β-unsaturated compounds.
In this study, we report on pyrazolyl nickel(II) and palladium(II) complexes as catalysts for the hydrogenation of 2,4-hexanoic acid (sorbic acid), which is an α,β-unsaturated acid. This study forms part of a bigger project on partial hydrogenation of biofuels from triglycerides.
All hydrogenation reactions were carried out in PPV-CTR01-CE (Eyela, Japan) high-pressure autoclave reactor with a stirring pact, heating and cooling systems. 20(b) The course of hydrogenation reactions involving palladium catalysts were followed by 1 H NMR spectroscopy, using dioxane as an internal standard, which was used to determine percentage conversions. The conversions were determined using the diagnostic peaks, following the integrations of the products from the hydrogenation reaction compared to the integration of dioxane.

Molecular Structure Determination
A mixture of CH 2 Cl 2 (0.5 mL) and n-hexane (0.1 mL) for 2 or CHCl 3 (0.5 mL) and n-hexane (0.1 mL) for 3 was used to obtain single crystals that were subsequently used for X-ray diffraction data collection for molecular structure determination.
Crystal data were collected using Bruker APEX-II CCD diffractometer with Mo Kα (λ = 0.71073 Å). The diffractometer to crystal distance was 4.00 cm, and all crystal data collected at 100 K. Data reduction measurement was performed using SAINT+, 21 and the intensity correction for absorption using SADABS. 21 Refinement of structures, with least square minimization, was performed using the SHELXT 22 and SHELXL 23 software packages. All non-hydrogen atoms were refined with anisotropic displacement coefficients and placed in geometrically idealized positions, and constrained to ride on their parent atoms with relative isotropic coefficients. 22,23

Hydrogenation with Molecular Hydrogen
Hydrogenation reactions using molecular hydrogen were studied in reactors with stainless steel vessels coupled with magnetic stirrers. In a typical experiment, the contents of the vessel consist of sorbic acid (0.5 mmol), catalyst (2.5 µmol, 0.5 mol%), hydrogen gas (5 bar) and methanol (5 mL). The solution mixture was purged twice with nitrogen gas, followed by the introduction of hydrogen gas (5 bar) and constant stirring of the mixture at 40°C for 2 h. At the end of the reaction period, the reaction vessel was cooled, and the excess pressure generated was vented off slowly. The resulting hydrogenation products were withdrawn and filtered with MS nylon syringe filter (0.22 µm, 13 mm). Their % conversions were then determined by 1 H NMR spectroscopy, using dioxane as an internal standard.

Hydrogenation with Formic Acid
In a typical experiment, sorbic acid (0.5 mmol), catalyst (2.5 µmol, 0.5 mol%) formic acid (20 mmol), KOH (4 mmol) and methanol (5 mL) were introduced into the reactor vessel. The solution mixture was purged twice with nitrogen gas, followed by stirring at 90°C for 12 h. At the end of the reaction period, the reaction vessel was cooled, and the excess pressure generated was vented off slowly. The hydrogenation products were drawn out of the reactor vessel, filtered using MS ® nylon syringe filter (0.22 µm, 13 mm) and their % conversions determined by 1 H NMR spectroscopy, using dioxane as an internal standard.

Synthesis of Palladium and Nickel Complexes.
The pre-catalysts 1-6 were synthesized with compounds L1-L3 and the corresponding metal precursors, as shown in Scheme 1. Characterization of the palladium complexes was achieved using a combination of 1 H NMR, IR, mass spectrometry and elemental analysis. The structures of the two new complexes, 2 and 3, were confirmed by single-crystal X-ray crystallography. Characterization of the nickel complexes, on the other hand, was carried out using mainly IR spectroscopy and elemental analysis. The structure of one of them, 4, was confirmed by single-crystal X-ray crystallography showing a structure result similar to that reported earlier. 20 Characteristic chemical shifts of the pyrazolyl nitrogen protons (N-H) confirmed the successful ligand complexations. For instance, there was a downfield shift in the position of N-H proton from 11.20 ppm to 11.82 ppm in 1; from 10.19 ppm to 11.65 in 2; and from 10.62 ppm to 11.53 in 3. All other spectroscopic data were as expected and similar to those previously reported for 1 24 , 2 24 , and 3 25 .

Molecular Structures of 2 and 3
Slow evaporation of solutions of 2 in CHCl 3 and 3 in CH 2 Cl 2 at room temperature produced orange crystals good enough for single-crystal X-ray crystallographic analysis. The crystals and structure refinement information are shown in Table 1, and the molecular structures in Figs. 1 and 2. Complexes 2 and 3 crystallized in C2/c and P-1 space groups, respectively. Figures 1 and 2 show that the pyrazole nitrogen atom is coordinated to the palladium in square planar geometries. However, the t Bu groups in 2 are disordered. This disorder was handled during the refinement process by rotating the H atoms around the t Bu axis to minimize the restraint caused by the tert-butyl groups.

Transfer Hydrogenation of Sorbic Acid
The pyrazolyl palladium and nickel complexes (1-6) were evaluated as catalysts for the transfer hydrogenation of sorbic acid, with formic acid as the source of hydrogen. Many phosphino nickel(II) and palladium(II) complexes are well-known catalysts for the hydrogenation of olefinic double bonds, 26-29 some of them featuring P ∧ N donor ligands. 29,30 These hydrogenation reactions were first carried out in the presence of a base, such as KOH, which facilitates the deprotonation of the formic acid to formate ion. The formate ion then coordinates to the metal centre leading to decomposition of the formic acid with the aid of the intermediate, [M]-OOCH as an active catalyst, to produce H 2 and CO 2 . 29 This accelerates H 2 heterolysis and causes the catalytic process. 20(b), 30 A typical hydrogenation reaction was performed with sorbic acid and a complex present in a 200:1 mole ratio at 90°C for 12 h (Fig. SI-3). It is worth noting that the hydrogenation process did not take place when the pre-catalyst was not added. Catalytic activities (in terms of conversions of sorbic acid) are very good with the palladium complexes and in the order 2 > 3 >1; but poor for the nickel complexes, except for 4, which gave 62 % conversion (Table 2), having a turnover number (TON) of 124 and turnover frequency (TOF) of 10. However, the selectivity towards the distribution of the products is quite significant for all the complexes. Considering the effect of substituents (methyl, tertiary-butyl, and ferrocenyl) on complexes 1-3, the order of catalytic activity is 2 > 3 > 1 in terms of conversion of sorbic acids. Complex 2 gave 100 % conversion of sorbic acid in 12 h compared to 67 % and 74 % for complexes 1 and 3, respectively. A similar trend is observed in their TON and TOF values. However, there is not much difference in their selectivities towards 2-hexenoic and 4-hexenoic acids ( Table 2, Fig. SI-4). The excellent activity of complex 2 might be due to the solubility of this compound provided by the tertiary-butyl substituent on the pyrazolyl ligand.
The pyrazolyl nickel and palladium catalyzed hydrogenation of sorbic acid, in this study, proceeds via a two-step (sequential) reaction. The first step involves the formation of the intermediates, 4-hexenoic and 2-hexenoic acids, and further hydrogenation of the hexenoic acid isomers in the second step produces hexanoic acid (Scheme 2). 31 In an attempt to find out if the C=C bonds of the intermediates 4-hexenoic and 2-hexenoic acids would be fully saturated, the hydrogenation reaction was run for 24 h using complex 2. Interestingly, the product distribution after 24 h was only slightly different from when the reaction was run for 12 h, namely 82 % hexanoic acid, 13 % 2-hexenoic acid and 5 % 4-hexenoic acids for 24 h vs 79 % hexanoic acid, 14 % 2-hexenoic acid and 7 % 4-hexenoic acids for 12 h (Table 2, entries 2 and 3).

3.3.2.. Hydrogenation of Sorbic Acid with Molecular Hydrogen
We also carried out the hydrogenation of sorbic acid with molecular hydrogen and complexes 1-6. In a typical experiment, 5 mmol of sorbic acid and 2.5 µmol (0.05 mol%) of the precatalysts were added to a reactor and ran for periods ranging from 0.5 h to 2 h, at 5 bar and 40°C. The complex to sorbic acid ratio was 1:200, and the reaction was with palladium catalysts monitored by 1 H NMR spectroscopy. All six complexes produced active catalysts, leading to the same products observed with   formic acid (Scheme 2); the product distribution from the molecular hydrogenation reactions depicted in Table 3 is not so different from that of the formic acid reactions shown in Table 2 and Fig. 3. As expected, the palladium complexes (1-3) are the most active, having the highest TOF of 400, all with complete conversion of sorbic acid within 0.5 h. Only two of the nickel complexes (4 and 5) had a complete conversion, and only after 2 h (Table 3, entries  7 and 11). Reactions, where the catalyst was a palladium complex, could be followed by 1 H NMR spectroscopy (Fig. 4). Furthermore, all the nickel and palladium catalysts did not completely hydrogenate the sorbic acid to hexanoic acid within 0.5 h, and selectivities for 2-hexenoic and 4-hexenoic acids were as high as 37 % and 30 %, respectively (Table 3). These results indicate the conditions under which any of these intermediate products can optimally be obtained if partial hydrogenation compounds are the targeted products. This product distribution is depicted in Table 3 and in Fig. 4, which shows the 1 H NMR spectral time study with complex 2.
At 0.5 h, 20 % conversion of sorbic acid was observed with complex 4 (Fig. 4). This conversion resulted in 37 % and 9 % selectivity towards 2-hexenoic acid and 4-hexenoic acid, respectively (Table 3, entry 4). After 1.5 h, the conversion was greatly increased to 99 % with 21 % selectivity towards 2-hexenoic acid, and 2 % of 4-hexenoic was detected (Table 3, entry 6). Further increment in the reaction time (after 2 h) only formed hexanoic acid with 100 % conversion of sorbic acid (Table 3, entry 7). Similar observations were also seen using complexes 5 and 6 with 22 % and 18 % conversions, respectively, at 0.5 h. 32 Results for the catalytic tests are summarized in Table 3   show that the less hindered C=C bond in 4-hexenoic acid is preferentially reduced over the more hindered one in 2-hexenoic acid. A detailed time-dependence study employing 1 H NMR spectroscopic technique was carried out on a sample of complex 2 with 100 % conversion of sorbic acid (Fig. 4). The hydrogenation reaction's progress, with complex 2, produced a betterresolved spectrum in the chemical shifts compared to its nickel analogue (complex 5), which produced a broad spectrum due to its paramagnetic property. The hydrogenation reaction proceeded with the formation of the intermediates (2-hexenoic and 4-hexenoic acid) (Fig. SI-6) and hexanoic acid after 0.5 h. It is noted that by this time (0.5 h), the sorbic acid had been completely converted in all cases, and no traces of the substrate detected. Further increment in the time consumed one of the intermediates (4-hexenoic acid), followed by 2-hexenoic acid until all the C=C bonds were fully saturated after 2 h with 100 % selectivity towards hexanoic acid (Fig. 5).

Conclusions
Our study has shown how the nature of metal centres in pre-catalysts can affect hydrogenation reactions. This influence is demonstrated by the higher efficiency of (pyrazolyl)palladium(II) complexes as compared to their corresponding nickel(II) counterparts. The activities of the catalysts with the same ligand are: 1 > 4, 2 > 5 and 3 > 6. Furthermore, all the complexes investigated as sorbic acid hydrogenation catalysts, 1-6, gave appreciable conversions for catalytic hydrogenation of sorbic acid (i.e. greater than 52 %) compared with what was reported in the literature using ruthenium 32,33 and rhodium. 33 The product distributions for the hydrogenation of sorbic acid using pre-catalysts (1-4), with ones with significant conversions, is shown in Fig. 3. in solving the crystal structures, data collections and X-ray analysis.

Supplementary Material
Supplementary information is provided in the online supplement, and additional data can be obtained free of charge from The Cambridge Crystallographic Data Centre at the addresses below. CCDC numbers 1878431 and 1878429 contain the supplementary crystallographic data for complexes 2 and 3, respectively.