The Origin of Stereoselectivity in the Hydrogenation of Oximes Catalyzed by Iridium Complexes: A DFT Mechanistic Study

Abstract

Herein the reaction mechanism and the origin of stereoselectivity of asymmetric hydrogenation of oximes to hydroxylamines catalyzed by the cyclometalated iridium (III) complexes with chiral substituted single cyclopentadienyl ligands (Ir catalysts A1 and B1) under acidic condition were unveiled using DFT calculations. The catalytic cycle for this reaction consists of the dihydrogen activation step and the hydride transfer step. The calculated results indicate that the hydride transfer step is the chirality-determining step and the involvement of methanesulfonate anion (MsO⁻) in this reaction is of importance in the asymmetric hydrogenation of oximes catalyzed by A1 and B1. The calculated energy barriers for the hydride transfer steps without an MsO⁻ anion are higher than those with an MsO⁻ anion. The differences in Gibbs free energies between TSA5−1fR/TSA5−1fS and TSB5−1fR/TSB5−1fS are 13.8/13.2 (ΔΔG^‡ = 0.6 kcal/mol) and 7.5/5.6 (ΔΔG^‡ = 1.9 kcal/mol) kcal/mol for the hydride transfer step of substrate protonated oximes with E configuration (E−2a−H⁺) with MsO⁻ anion to chiral hydroxylamines product R−3a/S−3a catalyzed by A1 and B1, respectively. According to the Curtin–Hammet principle, the major products are hydroxylamines S−3a for the reaction catalyzed by A1 and B1, which agrees well with the experimental results. This is due to the non-covalent interactions among the protonated substrate, MsO⁻ anion and catalytic species. The hydrogen bond could not only stabilize the catalytic species, but also change the preference of stereoselectivity of this reaction.

1. Introduction

The chiral hydroxylamines are one of the important organic intermediates in pharmaceutical and agricultural industries, which could be attributed to the reactive N–O bond in the structural motif. The molecules owning the N-alkoxy amine group are very common in a wide range of bioactive and pharmaceutical products [1]. Among the syntheses of chiral amines, the asymmetric hydrogenation of the C=N double bond catalyzed by transition-metal complexes is one of the most efficient methods to prepare enantiomers at an industrial scale [2]. However, the selective reduction of oximes to the corresponding chiral hydroxylamine derivatives remains a challenge because the undesired reductive cleavage of the labile N–O bond leads to undesired primary amines (see Scheme 1a) [3]. Chiral hydroxylamine could be obtained by partial oxidation of chiral amines, but this method could either involve multiple steps or be prone to overreaction [4,5,6,7]. In 2014, Oestreich et al. reported the reductive hydrogenation of oxime ether catalyzed by B(C₆F₅)₃ at room temperature [8]. Recently, Zhang et al. realized Ni-catalyzed asymmetric hydrogenation of oxides to chiral hydroxyl amines via weak attractive interactions between the catalyst and substrate (Scheme 1b) [9]. In 2020, Cramer et al. reported this reaction catalyzed by cyclometallated iridium (III) complexes with chiral substituted single cyclopentadienyl ligands and C,N-bidentate aryl imine ligand (see Scheme 1c) [10]. Despite the experimental reports above, the effective strategy of asymmetric hydrogenation of oxime to hydroxylamine is still relatively scarce. The reaction by Cramer et al. proceeded at room temperature with 98% ee and the turnover number (TON) reached 4000 before becoming inactive. This transformation was comprised of the quantitative protonation of the substrate and subsequent hydride transfer. The hydride could be transferred to the substrate either directly or with the help of the anion mediation by means of non-covalent contacts with the transition-metal catalytic species and the protonated substrates. Although it is obvious that the cyclometallated iridium complexes such as A1 and B1 owning chiral cyclopentadienyl ligand with binaphthyl as backbone and an achiral C,N-bidentate aryl imine ligand are of importance in the asymmetric hydrogenation of oximes, the reaction mechanism and the origin of stereoselectivity for this reaction are still unclear. Therefore, herein a DFT mechanistic study was performed in order to unveil the nature of this reaction catalyzed by catalysts A1 and B1.

As shown in Scheme 2, the catalytic cycle of this reaction consists of the dihydrogen activation (DA) step and the hydride transfer (HT) step. In the dihydrogen activation step, one molecular hydrogen interacts with 1 by replacing the weakly coordinated methanesulfonate anion (MsO⁻) ligand to afford intermediate 2, then the dihydrogen is heterolytically split with the assistance of the MsO⁻ anion via the transition state TS2−3 to form intermediate 3. Before the addition of oxime substates in the catalytic cycle, the original oxime substrate E−1a is protonated by MsOH to achieve protonated adduct E−2a−H⁺ without MsO⁻ anion or E−2a with MsO⁻ anion. In the subsequent hydride transfer step, the methanesulfonic acid (MsOH) is released to produce intermediate 4; the protonated substrates (E−2a−H⁺ or E−2a) could interact with 4 to achieve the chiral hydroxylamine product 3a and regenerate 1 completing the catalytic cycle. In this work, the reaction mechanism of the asymmetric hydrogenation of oximes to hydroxylamines was discussed, the origin of the stereoselectivity was unveiled, and the important role of the MsO⁻ anion in this reaction was described. This might shed lights on the development of transition-metal catalysts based on the mechanism of the asymmetric hydrogenation of oximes in the future.

2. Results and Discussion

The catalytic cycle of the asymmetric hydrogenation of oxime to hydroxyamine catalyzed by Ir catalysts (A1 and B1) includes the dihydrogen activation step and the hydride transfer step. In this work, the (S)-Ir catalysts bearing a chiral binaphthyl-derived cyclopentadienyl ligand was used as the catalytic species for this reaction, which was reported by experiments [10]. Figure 1 presents the Gibbs free energy profiles for the asymmetric hydrogenation of E-oxime (E−1a) catalyzed by A1; the energies of the stationary points are relative to corresponding starting point A1 (see that of hydrogenation of Z-oxime in Figure S1 of SM). E-oxime was used because it is more stable than Z-oxime. In the dihydrogen activation step, one molecule hydrogen coordinates with Ir center by replacing the weakly coordinated MsO⁻ anion to form intermediate A2, which is endergonic by 14.7 kcal/mol. Then the coordinated dihydrogen causes a heterolytic splitting with the assistance of the MsO⁻ anion. The free energy barrier from A1 to A3 is 14.5 kcal/mol via TSA2−3. A3 forms iridium hydride intermediate A4 with the release of one molecule of MsOH, which is exergonic by 4.9 kcal/mol. The following hydride transfer step could be the direct hydride transfer without the involvement of the MsO⁻ anion or the MsO⁻-assisted hydride transfer. In the direct hydride transfer step, the substrate tethering with one molecule MsOH, E−2a, could produce the protonated oxime E−2a−H⁺ with the removal of one MsO⁻ anion. Then E−2a−H⁺ approaches the Ir center of A4 along the si-face direction to form intermediate A5uR leading to the hydroxylamine product R−3a. If it approaches the Ir center of A4 along the re-face direction, the intermediate A5uS is generated leading to S−3a hydroxylamine. The “u” in the names of stationary points such as A5uS represents stationary points along an unfavorable pathway, the “f” is used for those along a favorable pathway, and “S” or “R” denotes the reaction pathways leading to S or R products, respectively. In the direct hydride transfer step from A5uR/A5uS to A6, the hydride is transferred from an Ir center to the carbon atom of pronated oxime moiety of A5uR/A5uS directly via TSA5−6uR/TSA5−6uS. The free energy barriers for this step are 5.2/8.1 kcal/mol from A5uR/A5uS to A6uR/A6uS and hydroxylamines, respectively.

Finally, A6uR/A6uS combine with the MsO⁻ anion to regenerate the active catalytic species A1 and the catalytic cycle is completed. It should be noted that the energetic span [11] of generating R−3a/S−3a products are 27.0/29.2 kcal/mol (A1→TSA5−6uR/TSA5−6uS), respectively. That means that R−3a is the dominant product instead of S−3a, and the energetic span is high, which is contrary with the conditions for the experiment to be carried out at room temperature.

Therefore, the direct hydride transfer mode cannot explain well the experimental observation. The non-covalent interactions have been proposed to be of importance in the asymmetric reactions, which might stabilize the transition states, accelerate chemical reaction and regulate the stereoselectivity [9,12,13]. Note that the Gibbs free energy of the oxime substrate tethering with one molecule MsOH (E−2a) is 17.3 kcal/mol lower than that of protonated oxime E−2a−H⁺, which could be more stable to form E−2a for E−2a−H⁺ combining with the conjugated base MsO⁻ anion. There are non-covalent interactions including hydrogen bonds formed by E−2a−H⁺, MsO⁻ and Ir catalytic species after E−2a interacts with A4. It is exergonic by 4.3 and 5.0 kcal/mol to generate A5fR and A5fS from the interaction of E−2a and A4. In the subsequent MsO⁻-assisted hydride transfer step from A5fR/A5fS to A1+R−3a/A1+S−3a, the hydride is transferred from the Ir center to the carbon atom of oxime moiety. The free energy barriers for this step are 10.9 and 11.0 kcal/mol for the formation of R−3a and S−3a via TSA5−1fR/TSA5−1fS, respectively. Based on the energetic span model and Curtin–Hammet principle [14], the energetic spans of the reaction pathways to achieve R-3a and S-3a products adopting the MsO⁻-assisted hydride transfer mode are 13.8 (A1→TSA5−1fR) and 13.2 (A1→TSA5−1fS) kcal/mol, respectively. It is obvious that the reaction pathways with the involvement of the MsO⁻ anion is much favorable than those without the MsO⁻ anion. The difference of the energetic spans for the reactions leading to R and S products are 0.6 kcal/mol (ΔΔG^‡ = 0.6 kcal/mol). The calculated results show that S−3a is the dominant product. The predicted enantioselectivity is 73:27 e.r., which agrees well with that obtained in the experiment (70:30 e.r.) [10]. This indicates that the non-covalent interactions including hydrogen bonds among E−2a−H⁺, MsO⁻ and Ir catalytic species are very important, as they could not only stabilize the catalyst but also turn over the stereoselectivity of the reaction.

Meanwhile, the asymmetric hydrogenation of oxime E−1a to hydroxylamine catalyzed by Ir catalyst B1 was also investigated using DFT method in this work. The free energy profiles for this reaction catalyzed by B1 are shown in Figure 2. Following the similar reaction pathways as discussed for catalyst A1 above, in the dihydrogen activation step from B1 to B4 molecular hydrogen replaces weakly coordinated the MsO⁻ anion of B1 to afford B2 at first, which is endergonic by 9.3 kcal/mol. Then B2 proceeds with a heterolytic dihydrogen splitting to achieve B3 via TSB2−3 with a free energy barrier of 3.0 kcal/mol. Subsequently, B3 releases MsOH to form iridium hydride intermediate B4. This step is exergonic by 4.6 kcal/mol. In the following hydride transfer step (the chirality-determining step), the hydride of metal center of B4 could be stereoselectively transferred to the protonated substrate with or without MsO⁻ anion assistance. Similarly, MsOH tethered substrate E−2a releases the MsO⁻ anion to afford protonated substrate E−2a−H⁺ before the interaction with the iridium hydride intermediate B4. In the situation without the participation of the MsO⁻ anion, E−2a−H⁺ interacts with B4 along its Si/Re face to form B5uR/B5uS intermediates, which could achieve the final chiral hydrogenated products. Instead, the intermediate B5fR/B5fS could be formed by the combination of B4 and E−2a with the participation of the MsO⁻ anion. Similarly, the intermediates B5fR/B5fS are more stable than B5uR/B5uS due to the formation of hydrogen bonds among E−2a−H⁺ and MsO⁻ anion in the former ones. The calculated free energy barriers of the MsO^—assisted hydride transfer step of this reaction to achieve R-3a and S-3a products are 6.6 (B5fR→TSB5f−1R) and 9.4 (B5fS→TSB5f−1S) kcal/mol, respectively. Compared to the MsO⁻-assisted hydride transfer, those adopting the direct hydride transfer mode without the involvement of the MsO⁻ anion are 18.7 (B1→TSB5−1uR) and 22.8 (B1→TSB5−1uS) kcal/mol, respectively. It is obvious that the MsO⁻-assisted hydride transfer with MsO⁻ anion is much more favorable than the direct hydride transfer without MsO⁻ anion. In addition, according to the Curtin–Hammet principle, the difference in relative free energies between TSB5−1fS and TSB5−1fR is 1.9 kcal/mol (ΔΔG^‡ = 1.9 kcal/mol). This implies that the hydroxylamine S−3a is the favorable product and the predicted enantioselectivity is 96:4 e.r., this also agrees well with the result reported by experiment (89:11 e.r.) [10].

In general, the calculated results above for the asymmetric hydrogenation of oximes to hydroxylamines catalyzed by iridium complexes A1 and B1 could predict that the final favorable product would be S−3a, which is consistent with experimental results. The stereoselectivity of B1 is better than that of A1. The hydride transfer step is the chirality-determining step and the MsO⁻ anion is proposed to be of importance in this asymmetric hydrogenation of oximes. The MsO⁻-assisted hydride transfer mode is superior to the direct hydride transfer mode.

In order to explore the origin of stereoselectivity of asymmetric hydrogenation of oximes to hydroxylamines catalyzed by chiral Ir catalysts (A1 and B1), the independent gradient model based on Hirshfeld partition of molecular density (IGMH) analysis were employed using Multiwfn software and were visualized using VMD software [15]. IGMH analysis could characterize and visualize the non-covalent interactions of key structures along reaction pathways for this reaction. IGMH analysis of key transition state structures TSB5−1fR/TSB5−1fS for the hydride transfer step in the asymmetric hydrogenation of oximes to hydroxylamines catalyzed by B1 are shown in Figure 3. It could be seen that there existed a strong hydrogen bond interaction between the MsO⁻ anion and the substrate in the anion-assisted hydride transfer step. This hydrogen bond interaction could stabilize the catalyst and make the MsO⁻ anion-assisted hydride transfer mode superior to the direct hydride transfer. In addition, we found that the distance between the hydrogen atom of benzene ring moiety of catalyst B1 and the hydrogen atom of the substrate is close (marked by a red cycle in Figure 3). The distance between the two hydrogen atoms in TSB5−1fR/TSB5−1fS is 1.904 Å/2.191 Å, respectively. IGMH analysis visualized this van der Waals repulsion, and the distance in TSB5−1fR is closer than that in TSB5−1fS, implying that the repulsion interaction in TSB5−1fR is greater than that in TSB5−1fS. This might be the origins of stereoselectivity of this reaction and the instability of TSB5−1fR compared to TSB5−1fS. However, this recognition effect is weakened in the catalyst A1 due to the smaller ligand and could not be visualized by IGMH analysis (see Figure S4 of SM). The distances between the two hydrogen atoms of TSA5−1fR and TSA5−1fS are 2.221 Å and 2.291 Å, respectively, which are longer than those of corresponding TSB5−1fR and TSB5−1fS.

3. Computational Methods

Being consistent with our previous research on the mechanisms of hydrogenation reaction catalyzed by transition-metal complexes [16,17,18,19,20,21,22,23,24], all the calculations were performed using the ωB97X-D/BSI method employing the Gaussian 09 program [25,26]. BSI indicates that the LANL2DZ basis set was used for Ir atom and 6-31G* basis set for all other non-metal atoms [27,28,29]. Additionally, single point energies were calculated with ORCA package at the ωB97M-V/def2-TZVP level using optimized geometries at the ωB97X-D/BSI level [30,31,32,33]. The solution model based on the density (SMD) solvation model using tert-amyl alcohol as the solvent (ε = 5.78) was employed in calculations [34]. The frequency analyses were performed to verify that all transition states have one and only one imaginary frequency. The intrinsic reaction coordinate (IRC) calculations were performed for key steps to confirm transition states connecting two desired minima [35]. The quasi-rigid-rotor harmonic oscillator was used to consider low-frequency contributions utilizing the Shermo code [36,37]. To correct the overestimations of entropy contributions due to using the ideal gas phase model in the Gaussian program, we applied a correction of (n-m)*1.9 for a process from m components to n components according to the reference [38]. The non-covalent interactions were observed by mapping the independent gradient model based on Hirshfeld partition (IGMH) surfaces using Multiwfn [39,40]. Unless otherwise stated, all energies of stationary points are the Gibbs free energies calculated at 298.15 K, 1.0 atm and compared to A1 or B1. Total energies, Cartesian coordinates of all optimized structures and the calculation formula for enantioselectivity ratios (e.r.) are given in the Supplementary Materials (SM).

4. Conclusions

In summary, the reaction mechanism and the origin of stereoselectivity of asymmetric hydrogenation of oximes to hydroxylamines catalyzed by the cyclometalated iridium (III) complexes (A1 and B1) with chiral substituted single cyclopentadienyl ligand were investigated using DFT method. The hydride transfer step is proposed to be the chirality-determining step. The reaction could proceed via the direct hydride transfer or the MsO⁻-assisted hydride transfer. The calculated results showed that the major product should be the hydroxylamines S−3a along the MsO⁻-assisted hydride transfer pathway, and that the involvement of the MsO⁻ anion be of importance in this asymmetric hydrogenation of oximes catalyzed by A1 and B1. This agrees well with experimental results and implies that the non-covalent interactions including hydrogen bonds formed by the protonated substrate, MsO⁻ anion and Ir catalytic species could not only stabilize the catalyst but also turn over the stereoselectivity of the reaction. This work might provide theoretical insights for the mechanism-based development of transition-metal catalysts for the asymmetric hydrogenation of oximes in the future.