Selective Hydration of Nitriles to Corresponding Amides in Air with Rh(I)-N-Heterocyclic Complex Catalysts

: A new synthetic method for obtaining [RhCl(cod)(NHC)] complexes ( 1 – 4 ) (cod = η 4 -1,5-cyclooctadiene, NHC = N -heterocyclic carbene: IMes, SIMes, IPr, and SIPr, respectively) is reported together with the catalytic properties of 1 – 4 in nitrile hydration. In addition to the characterization of 1 – 4 in solution by 13 C NMR spectroscopy, the structures of complexes 3 , and 4 have been established also in the solid state with single-crystal X-ray di ﬀ raction analysis. The Rh(I)-NHC complexes displayed excellent catalytic activity in hydration of aromatic nitriles (up to TOF = 276 h − 1 ) in water / 2-propanol (1 / 1 v / v ) mixtures in air.

Transition metal-free processes have been described, too, such as the CsOH/DMSO superbase system [30], NaOH as catalyst [31], or tBuOK under anhydrous conditions [32]. Nitrile hydratases catalyze the hydration of nitriles to the corresponding amides under softer conditions and have been successfully used, for example, for production of levetiracetam (Keppra ® ) for the treatment of epilepsy [3]. However, the application of most NHases is limited because of their substrate Wilkinson's catalyst with acetaldoxime as the water source; various functional groups were compatible with the reaction conditions [51].
To the best of our knowledge, there are no [RhCl(cod)(NHC)] type Rh(I)-catalysts reported until now for the selective hydration reaction of nitriles with water in aqueous or partly aqueous solvents. Therefore, we initiated a study of catalytic nitrile hydration with the use of the known [RhCl(cod)(NHC)] (1)(2)(3)(4) complexes with the NHC ligands IMes, SIMes, IPr, and SIPr, respectively ( Figure 1) [44,46,48,49,[52][53][54][55][56]. In this article, we report on a simple, one-step synthetic procedure for obtaining these complexes using [RhX(cod)] 2 (X = Cl -, OH -) as a metal precursor, the respective imidazolium/imidazolinium chlorides, and K 2 CO 3 as the deprotonating agent, in toluene at 70 • C. Successful application of complexes 1-4 for the selective hydration of several aromatic and heteroaromatic nitriles to the corresponding amides is also described in detail below.
To the best of our knowledge, there are no [RhCl(cod)(NHC)] type Rh(I)-catalysts reported until now for the selective hydration reaction of nitriles with water in aqueous or partly aqueous solvents. Therefore, we initiated a study of catalytic nitrile hydration with the use of the known [RhCl(cod)(NHC)] (1)(2)(3)(4) complexes with the NHC ligands IMes, SIMes, IPr, and SIPr, respectively ( Figure 1) [44,46,48,49,[52][53][54][55][56]. In this article, we report on a simple, one-step synthetic procedure for obtaining these complexes using [RhX(cod)]2 (X = Cl -, OH -) as a metal precursor, the respective imidazolium/imidazolinium chlorides, and K2CO3 as the deprotonating agent, in toluene at 70 °C. Successful application of complexes 1-4 for the selective hydration of several aromatic and heteroaromatic nitriles to the corresponding amides is also described in detail below.

Synthesis and Characterization of the [RhCl(cod)(NHC)] Complexes 1-4
In this work, we explored the applicability of Rh(I)-N-heterocyclic complexes 1-4 ( Figure 1) for catalysis of hydration of aromatic nitriles. We developed a synthetic method for obtaining these known compounds [44,46,48,49,[52][53][54][55][56], which does not require the use of the isolated free carbenes or the use of the corresponding Ag(I)-NHC transmetallating agents. In general, the synthesis of 1-4 (Scheme 1) involved stirring of the respective 1,3diarylimidazolium or 1,3-diarylimidazolinium salt in toluene at 70 °C together with [RhCl(cod)]2 and K2CO3 as an efficient and mild base (A) [47] or with [Rh(OH)(cod)]2 (no base added; B). After removal of the toluene solvent the products were dissolved in CH2Cl2-ethyl acetate and purified by passing through a short silica column; complexes 1-4 were isolated in 58-88% yield. The purity of the complexes was checked by 1 H and 13 C{ 1 H} NMR spectroscopy. The 13 C{ 1 H} NMR spectra of all complexes displayed the diagnostic Rh(I)-C(carbene) doublet resonances at 183.2 and 185.5 ppm (1 and 3), and 212.4 and 214.9 ppm (2 and 4), respectively (further spectral details in the Materials and Methods Section).
Single-crystals of 4 could be obtained by crystallization from chloroform at room temperature. In addition, both 3 and 4 yielded single-crystals from benzene, however, these crystals contained solvating benzene molecules, too. (Further experimental details of the X-ray structure analysis can be found in Supplementary Materials). The crystals were subjected to X-ray diffraction measurements.
In general, the synthesis of 1-4 (Scheme 1) involved stirring of the respective 1,3-diarylimidazolium or 1,3-diarylimidazolinium salt in toluene at 70 • C together with [RhCl(cod)] 2 and K 2 CO 3 as an efficient and mild base (A) [47] or with [Rh(OH)(cod)] 2 (no base added; B). After removal of the toluene solvent the products were dissolved in CH 2 Cl 2 -ethyl acetate and purified by passing through a short silica column; complexes 1-4 were isolated in 58-88% yield.
Catalysts 2020, 10, x FOR PEER REVIEW 3 of 16 To the best of our knowledge, there are no [RhCl(cod)(NHC)] type Rh(I)-catalysts reported until now for the selective hydration reaction of nitriles with water in aqueous or partly aqueous solvents. Therefore, we initiated a study of catalytic nitrile hydration with the use of the known [RhCl(cod)(NHC)] (1-4) complexes with the NHC ligands IMes, SIMes, IPr, and SIPr, respectively ( Figure 1) [44,46,48,49,[52][53][54][55][56]. In this article, we report on a simple, one-step synthetic procedure for obtaining these complexes using [RhX(cod)]2 (X = Cl -, OH -) as a metal precursor, the respective imidazolium/imidazolinium chlorides, and K2CO3 as the deprotonating agent, in toluene at 70 °C. Successful application of complexes 1-4 for the selective hydration of several aromatic and heteroaromatic nitriles to the corresponding amides is also described in detail below.

Synthesis and Characterization of the [RhCl(cod)(NHC)] Complexes 1-4
In this work, we explored the applicability of Rh(I)-N-heterocyclic complexes 1-4 ( Figure 1) for catalysis of hydration of aromatic nitriles. We developed a synthetic method for obtaining these known compounds [44,46,48,49,[52][53][54][55][56], which does not require the use of the isolated free carbenes or the use of the corresponding Ag(I)-NHC transmetallating agents. In general, the synthesis of 1-4 (Scheme 1) involved stirring of the respective 1,3diarylimidazolium or 1,3-diarylimidazolinium salt in toluene at 70 °C together with [RhCl(cod)]2 and K2CO3 as an efficient and mild base (A) [47] or with [Rh(OH)(cod)]2 (no base added; B). After removal of the toluene solvent the products were dissolved in CH2Cl2-ethyl acetate and purified by passing through a short silica column; complexes 1-4 were isolated in 58-88% yield. Single-crystals of 4 could be obtained by crystallization from chloroform at room temperature. In addition, both 3 and 4 yielded single-crystals from benzene, however, these crystals contained solvating benzene molecules, too. (Further experimental details of the X-ray structure analysis can be found in Supplementary Materials). The crystals were subjected to X-ray diffraction measurements. The purity of the complexes was checked by 1 H and 13 C{ 1 H} NMR spectroscopy. The 13 C{ 1 H} NMR spectra of all complexes displayed the diagnostic Rh(I)-C(carbene) doublet resonances at 183.2 and 185.5 ppm (1 and 3), and 212.4 and 214.9 ppm (2 and 4), respectively (further spectral details in the Materials and Methods Section).
Single-crystals of 4 could be obtained by crystallization from chloroform at room temperature. In addition, both 3 and 4 yielded single-crystals from benzene, however, these crystals contained solvating benzene molecules, too. (Further experimental details of the X-ray structure analysis can be found in Supplementary Materials). The crystals were subjected to X-ray diffraction measurements.

[RhCl(cod)(SIPr)]_benzene_4
Rh-C carbene 2.028 (7) 2.034 (7) 2.046 (7) 2.044 (7)  The solid-state crystal structure of 3 has already been determined by single crystal X-ray diffraction and was resolved without solvent [54]. This gives a possibility to compare the structures of 3 and 4 ( Table 1). The unit cell of [RhCl(cod)(SIPr)] (4), obtained from chloroform, does not contain solvent molecules, and, in contrast to [RhCl(cod)(IPr)] (3) [54] (P2 1 / c ), it crystallizes in the monoclinic P2 1 / n space group. There are two neutral Rh-complex molecules in the unit cells of both compounds. The lengths of the unit cell edges, and the unit cell angles show only slight differences. This is not surprising, since the sp 2 or sp 3 C-atoms in the IPr, and SIPr ligands, respectively, do not influence significantly the measures of the unit cell (the same is true for the two extra hydrogen atoms in SIPr). There are no significant differences in the Rh-C carbene and in the Rh-Cl bond lengths, either, however, the C2-C3 bond lengths in the [RhCl(cod)(SIPr)] (4) molecules are 1.501(7) Å, and 1.487(7) Å, respectively, which unambiguously refers to sp 3 carbon atoms. Saturation of the heterocyclic ring does not alter significantly the C carbene -Rh-Cl angles, either. Interestingly, the data of the unit cells of 3 and 4 are almost identical to those of [IrCl(cod)(IPr)] [57]; the data are compared in Tables S2 and S3

Hydration of Aromatic Nitriles Catalyzed by the [RhCl(cod)(NHC)] Complexes 1-4
Due to the importance of amides in the synthesis of important pharmaceuticals, there is a strong incentive to develop new transition metal catalysts which are able to facilitate the selective hydration reaction of aliphatic, as well as aromatic nitriles to corresponding amides (Scheme 2) with high activity under mild conditions (i.e., at temperatures below 100 • C, and preferably close to room temperature).

Hydration of Aromatic Nitriles Catalyzed by the [RhCl(cod)(NHC)] Complexes 1-4
Due to the importance of amides in the synthesis of important pharmaceuticals, there is a strong incentive to develop new transition metal catalysts which are able to facilitate the selective hydration reaction of aliphatic, as well as aromatic nitriles to corresponding amides (Scheme 2) with high activity under mild conditions (i.e., at temperatures below 100 °C, and preferably close to room temperature). Scheme 2. General scheme of the selective hydration of benzonitriles to benzamides.
It was found that complexes 1-4 efficiently catalyzed the hydration of benzonitrile to benzamide in a water/2-propanol = 1/1 mixture in air and under mild conditions (≤80 °C). The choice of 2propanol as the organic component of the solvent was based on its unique favourable effects on certain reactions, e.g., hydrogenation and transfer hydrogenation of ketones [58]. The reactions did not display an induction period ( Figure S1) and they proved completely selective; no products other than benzamide were detected by GC-MS or 1 H NMR spectroscopy. With these catalysts, fast hydration of benzonitrile was observed only in the presence of bases. The data in Table 4 show that in the lack of a base no reaction of benzonitrile was observed in 1.5 h, and even after 2 h the conversion reached only 3%. Conversely, with bases such as tBuOK, KOH, Na2CO3, and NaOH at a [base]/[Rh] = 1/1 ratio, the conversions in 1.5 h were in the 50-60% range and were not strongly dependent on the choice of the particular base. The use of NaOH resulted in the highest conversion, and therefore it was chosen for further studies. The possible catalytic effect of the bases in Table 4 were also checked in the hydration of benzonitrile in the absence of catalysts 1-4. Under the conditions used, conversion of benzonitrile to benzamide was < 1% with all four bases (only a trace of product could be detected by gas chromatography). These results show that the contribution of base-catalyzed hydration is negligible compared to the metal-complex catalyzed transformation. The effects of various reaction parameters for the hydration of benzonitrile were studied in detail using complex 1 as the catalyst. The progress of the reactions could be conveniently monitored by gas chromatography. Representative results are summarized in Table 5. It was found that complexes 1-4 efficiently catalyzed the hydration of benzonitrile to benzamide in a water/2-propanol = 1/1 mixture in air and under mild conditions (≤80 • C). The choice of 2-propanol as the organic component of the solvent was based on its unique favourable effects on certain reactions, e.g., hydrogenation and transfer hydrogenation of ketones [58]. The reactions did not display an induction period ( Figure S1) and they proved completely selective; no products other than benzamide were detected by GC-MS or 1 H NMR spectroscopy. With these catalysts, fast hydration of benzonitrile was observed only in the presence of bases. The data in Table 4 show that in the lack of a base no reaction of benzonitrile was observed in 1.5 h, and even after 2 h the conversion reached only 3%. Conversely, with bases such as tBuOK, KOH, Na 2 CO 3 , and NaOH at a [base]/[Rh] = 1/1 ratio, the conversions in 1.5 h were in the 50-60% range and were not strongly dependent on the choice of the particular base. The use of NaOH resulted in the highest conversion, and therefore it was chosen for further studies. The possible catalytic effect of the bases in Table 4 were also checked in the hydration of benzonitrile in the absence of catalysts 1-4. Under the conditions used, conversion of benzonitrile to benzamide was < 1% with all four bases (only a trace of product could be detected by gas chromatography). These results show that the contribution of base-catalyzed hydration is negligible compared to the metal-complex catalyzed transformation. The effects of various reaction parameters for the hydration of benzonitrile were studied in detail using complex 1 as the catalyst. The progress of the reactions could be conveniently monitored by gas chromatography. Representative results are summarized in Table 5. The data in Table 5 show that [RhCl(cod)(IMes)] (1) is an active catalyst for benzonitrile hydration. The TOF values (up to 276 h −1 ) compare well with those of most transition metal catalysts although fall behind the highest activities [33]. With increasing temperatures, the yield of benzamide increased and reached a maximum (98%) at 80 • C. It is also evident from Table 5, that under the applied reaction conditions, 2 h is the optimum reaction time for the catalytic hydration of benzonitrile to benzamide. For the entries 1-10 of Table 5, the effect of the base (NaOH) alone (i.e., in the absence of the Rh(I)-complex catalyst) has been checked and the results are shown in parentheses in the Conversion (%) column of the Table, next to the values obtained with catalyst 1 + NaOH. Here, again, it can be concluded, that the base-catalyzed hydration increases the total benzamide yield only to a minor extent even at higher reaction temperatures and longer reaction times (entries 5 and 10). In order to determine the efficiency of the catalyst in the absence of NaOH, we had to increase the catalyst concentration to 5 mol% (entries [11][12][13]. Even then, no reaction was observed at reflux conditions (approximately 81 • C, see Experimental) in 60 min, and only 26% conversion of benzonitrile was obtained after 180 min reaction time. In contrast, the reaction with 1 + NaOH led to 86% conversion already after 10 min (entry 14). Table 5 also shows the effect of the water-soluble tertiary phosphines PTA (1,3,5-triaza-7-phosphaadamantane) and mtppms-Na (sodium diphenylphosphinobenzene-3-sulfonate or monosulfonated triphenylphosphine Na-salt). Compared to catalyst 1 (entry 11, 0% conversion in 60 min), both PTA and mtppms increased the reaction rate and at a [phosphine]/[Rh] ratio their effect is about the same (entries 18 and 21). In general, however, mtppms proved to be more effective. Nevertheless, with regard to the rate increase, both phosphines were much inferior to NaOH (entry 14). The precedents in the literature show that with bmim (1-butyl-3-methyl-imidazole-2-ylidene) as the NHC ligand, PTA and mtppms form [Rh(cod)(bmim)(PTA)]Cl, and [Rh(cod)(bmim)(mtppms)] (a neutral zwitterionic complex), respectively [59]. In accordance with these earlier results, we expect that tertiary phosphines coordinate to the central Rh(I) ion in [RhCl(cod)(NHC)] complexes. However, the resulting complex species are coordinatively saturated and coordination of the nitrile substrate and/or H 2 O or OHto the metal ion in a Rh(I)-complex seems unlikely. In the case of [RuCl 2 (PTA) 4 ]-catalyzed nitrile hydration, Frost suggested that the increased catalytic activity in the presence of a large excess of PTA was due to the pH shift into the alkaline region in concentrated PTA solutions caused by the protonation of PTA [34]. This may happen in our reactions with added PTA, too, however, it is certainly not the case with mtppms which is protonated only in concentrated aqueous acid solutions. Nevertheless, since the roles of PTA and mtppms were not clarified in detail, our observations on the effect of PTA and mtppms on the Rh(I)-complex catalyzed hydration of benzonitrile can be regarded only as an information of practical importance. Details of these phosphine effects were not scrutinized. Table 6 presents the results of benzonitrile hydration with [RhCl(cod)(NHC)] complexes 1-4. It can be seen that in the presence of NaOH, high conversions (93 ->99%) could be obtained in reasonable reaction times (1-3 h) with all four catalysts (entries 2,5,8,11). Conversely, in the absence of NaOH, each catalyst showed only low activity, and the highest conversion under such conditions was only 26% in 3 h (entry 1). It seems from the conversion data for the first hour of the reactions, that the evolution of the real catalytic species in the water/2-propanol mixed solvent from the precursor complexes 1-4 and NaOH needs noticeable time. It is fast with 1 and 2 (entries 2, 5), somewhat slower with 4 (entry 11) and significantly slower in the case of 3 (entry 8). Note, that even with catalyst 3, the conversion of benzonitrile to benzamide reached 93% in 3 h. Compared to NaOH, lower rates were achieved with PTA in the case of all four catalysts, similar to the observations discussed above in conjunction with Table 5. [RhCl(cod)(IMes)] (1) proved suitable for hydration of benzonitriles with both electron donating and electron withdrawing substituents (Table 7). High conversions were achieved with as low as 1 mol% of catalyst. Para-chlorobenzonitrile showed more efficient conversion to p-chlorobenzamide than p-methylbenzonitrile which has an electron donating group in 4-position. Electron-withdrawing groups make the nitrile carbon more susceptible to nucleophilic attack by the activated water molecule or OH − . These findings are in agreement with the previously reported observations [6,10]. The conversions of various pyridine-carbonitriles to the corresponding amides (picolinamide, nicotinamide, isonicotinamide) were explored with 5 mol% catalyst 1 and the results are summarized in Table 8. Remarkably, the reactions of 3-and 4-pyridinecarbonitrile proceeded efficiently even in the absence of NaOH; apparently the pyridine moiety provided the sufficient basicity. The coordinating ability of the pyridyl functionality of 2-pyridinecarbonitrile reduced the activity as a catalyst of the complex and the reaction resulted only in 9% picolinamide. However, heteroaromatic nitriles with the N heteroatom adjacent to the β or γ position of the CN group (3-pyridinecarbonitrile and 4-pyridinecarbonitrile) showed high reactivity. Addition of three equivalents of PTA increased the catalytic activity in all cases, and 3-pyridinecarbonitrile, too, was hydrated with > 99% conversion in only 1 h. Finally, we studied the hydration of benzonitrile at 25 • C. It was found, that the use of 1 mol% catalyst 1 was sufficient to give a reasonable yield in 40 h (Table 9, entry 3). However, with a higher catalyst loading (2.5 mol%) 99% conversion was reached in 24 h (Table 9, entry 8). Lowering the concentration of 2-PrOH in the aqueous solvent mixture from 50% to 20% v/v, lead to a decrease in the conversion (entries 9, 10 vs. 1-3). The origin of this latter effect is presently unclear, since even at the lower 2-propanol concentration the reaction mixtures were homogeneous, and-formally-2-propanol is not involved in the hydration of benzonitrile. The above results did not allow the suggestion of a detailed reaction mechanism. Nevertheless, the findings are in accord with the nucleophilic attack of a Rh(I)-coordinated hydroxide onto the nitrile carbon atom (Scheme 3), similar to the mechanism suggested in [17]. It is an important observation, that the hydration reactions proceed with high rate already with 1 equivalent of base per Rh(I). Since there is hardly any conversion of benzonitrile in the absence of a base, this points to an intermediate formation of a Rh(I)-OH hydroxo-complex. On the other hand, the complete selectivity of the reaction to benzamide shows that most probably the nitrile also coordinates to the Rh-based catalyst, thereby activating the nitrile carbon against a nucleophilic attack. It should also be mentioned, that at the moment the role of the cod ligand is unclear. It may stay coordinated to the rhodium throughout the catalytic cycle, but in the reductive milieu of basic 2-propanol it may also be hydrogenated and replaced by other ligands present in the solution. The above results did not allow the suggestion of a detailed reaction mechanism. Nevertheless, the findings are in accord with the nucleophilic attack of a Rh(I)-coordinated hydroxide onto the nitrile carbon atom (Scheme 3), similar to the mechanism suggested in [17]. It is an important observation, that the hydration reactions proceed with high rate already with 1 equivalent of base per Rh(I). Since there is hardly any conversion of benzonitrile in the absence of a base, this points to an intermediate formation of a Rh(I)-OH hydroxo-complex. On the other hand, the complete selectivity of the reaction to benzamide shows that most probably the nitrile also coordinates to the Rh-based catalyst, thereby activating the nitrile carbon against a nucleophilic attack. It should also be mentioned, that at the moment the role of the cod ligand is unclear. It may stay coordinated to the rhodium throughout the catalytic cycle, but in the reductive milieu of basic 2-propanol it may also be hydrogenated and replaced by other ligands present in the solution.

Materials
All chemicals and reagents used in this work were purchased from Sigma-Aldrich, St. Louis, Missouri, USA; Molar Chemicals Kft., Halásztelek, Hungary and VWR International, West Chester,
Method A: To a solution of 0.88 mmol of the imidazolium/imidazolinium salt in 30 mL of toluene were added 0.44 mmol of [RhCl(cod)] 2 and 8.88 mmol of K 2 CO 3 in one portion. The mixture was stirred at 70 • C for 3-24 h (followed by TLC). After removal of the solvent the product was purified by passing through a short silica column in dichloromethane:ethyl acetate = 1:1 as solvent. The coloured fraction of the complex was collected, evaporated to dryness, and the yellow solid was vacuum-dried, characterized by 1 H, 13

General Methods
1 H and 13 C{ 1 H} NMR spectra were recorded on a Bruker Avance 360 MHz spectrometer (Bruker, Billerica, MA, USA) and were referenced to residual solvent peaks. Single crystal X-ray diffraction (SCXRD) measurements were performed using a Bruker D8 Venture diffractometer and the methods and software were described in [64][65][66][67][68][69][70]. Gas chromatographic measurements were done with the use of an Agilent Technologies 7890 A instrument (HP-5, 0.25 µm × 30 m × 0.32 mm, FID 300 • C (Agilent Technologies, Santa Clara, CA, USA); carrier gas: Nitrogen 1.9 mL/min). ESI-TOF-MS measurements were carried out on a Bruker maXis II MicroTOF-Q type Qq-TOF-MS instrument (Bruker Daltonik, Bremen, Germany) in positive ion mode. The mass spectra were calibrated internally using the exact masses of sodium formate clusters. The spectra were evaluated using the Compass Data Analysis 4.4 software from Bruker.
All catalytic reactions were carried out under air. The reaction temperatures were kept constant either by using a thermostated circulator (set e.g., to 80.0 ± 0.1 • C), or by running the reactions under reflux (lit. b.p. of 50% aqueous 2-propanol: 81.1 • C [71]). The products were identified by comparison of their retention time with known standard compounds. a) Hydration of Benzonitrile without Product Isolation 100 µL (1.0 mmol) benzonitrile, 5.5 mg (0.01 mmol) [RhCl(cod)(IMes)] (1), 0.4 mg (0.01 mmol) NaOH, and 12.8 mg (0.1 mmol, 10 mol% of the substrate) naphthalene (internal standard) were dissolved in a mixture of 1.5 mL 2-propanol and 1.5 mL deionized water. This reaction mixture was placed into a temperature-controlled bath and stirred at 80 • C for 2 h. A 0.10 mL part of the resulting hot solution was extracted with 2 mL CH 2 Cl 2 , passed through a short MgSO 4 column and subjected to gas chromatography. Conversion of benzonitrile: 98%.

Conclusions
We have realized one-step syntheses of the [RhCl(cod)(NHC)] complexes 1-4 without generation of the free carbene ligands or the silver-NHC complexes. The metal precursor [RhCl(cod)] 2 and the respective imidazolium/imidazolinium salt was stirred overnight in toluene at 70 • C, and the desired complexes were produced in good to excellent yields. An efficient catalytic system for the selective hydration of nitriles to the corresponding amides in a water/2-propanol solvent, with tolerance of air and several functional groups, is also described. The suggested reaction mechanism considers the nucleophilic attack of a Rh(I)-coordinated OHonto the nitrile carbon atom activated by the N-coordination of nitrile group to the metal ion.