Chiral amido-oxazoline functionalized MCM-41: A sustainable heterogeneous catalyst for enantioselective Kharasch–Sosnovsky and Henry reactions

Abstract

In this study, a series of chiral amido-oxazoline ligands was synthesized with a primary focus on immobilizing the most effective ligands on MCM-41 mesoporous material. Following several attempts, the para-nitro group of the chiral amido-oxazoline ligands was successfully reduced to amino group, enabling their immobilization on MCM-41. The resulting chiral heterogeneous amido-oxazoline ligands were characterized using various techniques, including FT-IR, XRD, TGA, SEM, TEM, EDX, and BET-BJH, confirming the successful immobilization of the amido-oxazoline ligands. A comparison of the efficiency of the homogeneous and heterogeneous amido-oxazoline-based ligands in the Kharasch-Sosnovsky and Henry reactions revealed better performance of the heterogeneous ligand. The immobilized amido-oxazoline-copper complexes exhibited remarkable catalytic activity, achieving excellent yields and enantioselectivities (up to 88 % ee) in the Kharasch-Sosnovsky reaction, and delivering excellent yields with moderate enantioselectivities in the Henry reaction. Notably, the Henry reaction proceeded with moderate diastereoselectivity, favoring the syn diastereomer, under solvent-free conditions, highlighting the sustainability of the process. The heterogeneous nature of the catalysts facilitated effortless recovery and efficient reusability.

Graphical abstract

1. Introduction

Oxazoline-based ligands have garnered considerable attention in asymmetric transformations due to their significant potential advantages, such as easy accessibility, modular nature, and suitability for a broad range of metal-catalyzed reactions. Typically, this class of ligands can be prepared from readily available chiral β-amino alcohols and carboxylic acids or nitriles, and can be modified by incorporating a diverse range of functional groups. Moreover, the close proximity of the chiral center to the donor nitrogen atom creates a well-organized chiral environment at the catalytic center, which greatly enhances asymmetric induction [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12]].

Introducing sustainable, environmentally friendly, and efficient chiral catalytic systems is a central theme in modern synthetic chemistry. While homogeneous oxazoline-based ligands often exhibit high reactivity and enantioselectivity, their practical application is mainly hindered by challenges related to separation, recovery, and reuse. The immobilization of oxazoline ligands onto suitable supports provides a reasonable approach to overcoming such limitations, aligning with the principles of sustainability and green chemistry. Mesoporous silica materials such as MCM-41 have emerged as promising supports due to their high surface area, tunable pore size, and ease of functionalization [[13], [14], [15], [16], [17], [18], [19], [20], [21]]. These outstanding features facilitate the immobilization of chiral catalysts, potentially enhancing their performance through confinement effects and site isolation, consequently improving the catalysts’ efficiency and significantly impacting enantioselectivity [[22], [23], [24], [25], [26], [27], [28], [29], [30], [31]].

The enantioselective Kharasch-Sosnovsky reaction, which involves the enantioselective oxidation of allylic C-H bonds in olefins primarily using a perester in the presence of a chiral copper complex, is a powerful and cost-effective approach in organic synthesis. This process is distinct from hydroxylation and epoxidation, as a second functional group is formed without altering the C=C bond. Additionally, the resulting chiral allylic ester can be further functionalized to produce a diverse range of products [[32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55]].

The asymmetric Henry reaction, also referred to as the nitroaldol reaction, is a highly atom-economical process that involves the enantioselective nucleophilic addition of the α-carbon of a nitroalkane to the carbonyl group of an aldehyde or ketone in the presence of a chiral environment, typically involving copper-ligand complexes, resulting in the formation of a C−C bond that consequently yields a chiral β-nitroalcohol [[56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83]]. The newly formed compounds can be readily transformed into other valuable synthetic intermediates by dehydration to produce nitroalkenes, oxidation of the hydroxy group to produce α-nitro ketones, or reduction of the nitro group to generate β-amino alcohols [[84], [85], [86], [87]]. It is noteworthy that β-amino alcohol derivatives have great potential for the synthesis of important pharmaceutical compounds, such as sphingosine [88], ephedrine [89], and chloramphenicol [89]. Additionally, they have been utilized in the production of β-blockers such as (S)-propranolol [90,91], (S)-pindolol [92], and (S)-metoprolol [93,94], as well as β-receptor agonists such as (−)-denopamine and (−)-arbutamine [95].

Building upon our recent success in utilizing amido-oxazoline ligands 1 for the enantioselective Kharasch-Sosnovsky reaction [96], we aimed to develop a sustainable and recyclable catalytic system by covalently immobilizing amido-oxazoline ligand derivatives on MCM-41. Herein, we report the synthesis and characterization of amido-oxazoline ligands immobilized on MCM-41, and their application in two important enantioselective transformations: the Kharasch-Sosnovsky reaction and the Henry reaction.

2. Result and discussion

2.1. Synthesis of chiral homogeneous amido-oxazoline based ligands

In our previous publication, we described the synthesis of amido-oxazoline ligands 1. This process begins with the preparation of amino-oxazolines 2 through the reaction of chiral (S)-amino alcohols 3 [97] and cyanogen bromide [[3], [4], [5], [6],96,[98], [99], [100], [101], [102], [103], [104], [105], [106]]. The amino-oxazolines 2 was then reacted with benzoyl chlorides 4 to yield the desired amido-oxazoline ligand derivatives 1. It was proven that the main structure is oxazolidine-imine form 1′ (Scheme 1) (Pages S2-S5 and Figs. S1-S10, Supplementary Material).

Scheme 1.

Synthesis of amido-oxazolines 1.

Despite this straightforward strategy, the heterogenization of this class of ligand on MCM-41 poses a challenge. Two practical approaches to covalently immobilize the ligands on MCM-41 mesoporous materials was considered. One strategy involves the introduction of a functional group, such as an amino group, at the para position of benzoic acid. The alternative entails the reduction of the nitro group in amido-oxazoline ligands 1 to form an amino group [[13], [14], [15]]. Therefore, to perform the first approach, para-amino benzoic acid 5 was chosen as a suitable starting material. Initially, the amino group of benzoic acid 5 was protected using di-tert-butyl dicarbonate (Boc₂O) [107,108]. In the ¹H NMR spectrum, the appearance of a singlet signal at δ = 1.57 ppm corresponds to the tert-butyl hydrogens. In the subsequent step, similar to the procedure mentioned earlier, the protected benzoic acid 6 was converted to its corresponding benzoyl chloride 7 using oxalyl chloride [96,108]. The resulting benzoyl chloride 7 was then promptly reacted with the amino-oxazolines 2 to produce the desired protected amido-oxazolines 8 (Scheme 2) (Pages S5 and S6, and Figs. S11-S19, Supplementary Material).

Scheme 2.

Preparation of the protected amido-oxazolines 8 from para-amino benzoic acid through benzoyl chloride.

Despite numerous attempts, this method did not yield satisfactory results. Therefore, an alternative strategy was employed, in which the protected carboxylic acid 6 was directly reacted with amino-oxazolines 2 using the Appel reagent (PPh₃/CCl₄) [52,109]. This procedure was successful in producing the desired products 8, achieving significantly higher yields than the previous attempt (Scheme 3). It is apparent that the in-situ activation of the carboxylic group led to a remarkable decrease in the formation of unwanted by-products. The presence of signals for aromatic and oxazoline rings at δ = 7–8 ppm and at δ = 4–5 ppm, respectively, in the ¹H NMR spectra confirmed the successful formation of the protected amido-oxazolines 8 (Pages S6 and S7, Supplementary Material).

Scheme 3.

Preparation of the protected amido-oxazolines 8 from para-amino benzoic acid using Appel reagent.

In the final step, amido-oxazolines 8 were treated with trifluoroacetic acid (TFA) to remove the Boc protecting group. However, despite several attempts, the deprotection process was unsuccessful, which is in contrast to our previous study where we effectively eliminated the Boc group [108] (Scheme 4).

Scheme 4.

Unsuccessful deprotection of Boc group from the protected amido-oxazolines 8.

As a result of these failures, the second option was explored, involving the reduction of the para-nitro group of the amido-oxazolines 1 into an amino group. Our previous experiments revealed that the use of hydrogen gas in the presence of a catalytic amount of Pd/C (10 %) is the most effective method for reducing nitro groups [109,110] (Scheme 5). This procedure can provide the corresponding aniline-oxazoline ligands 9 in 97 % and 93 % yield, respectively (Pages S7 and S8, and Figs. S20-S27, Supplementary Material). In the ¹H NMR spectra of compound 1a, the ortho and meta protons appeared at δ = 8.41 ppm and 8.26 ppm, respectively, while in 9a, these protons shift to δ = 6.65 ppm and 8.09 ppm, respectively. In fact, in the ¹H NMR spectra of compounds 1a and 1b, the ortho protons appeared at downfield values, while in compounds 9a and 9b, the meta protons are observed in downfield regions.

Scheme 5.

Direct conversion of the nitro group of amido-oxazolines 1 into amino group using H₂, Pd/C.

2.2. Heterogenization of the synthesized chiral amido-oxazoline ligands

Based on our previous study [96], which found that the amido-oxazoline ligands 1 containing phenyl and benzyl groups on the oxazolidine ring, demonstrated the highest efficiency, we opted to immobilize them on cylindrical MCM-41 mesoporous support. Initially, MCM-41 silica was synthesized hydrothermally, utilizing sodium silicate as a silica source and tetradecyl trimethyl ammonium bromide (TTAB) as a template agent to regulate pore size [18,108].

To immobilize the aniline-oxazoline ligands 9 on the MCM-41, it is essential to have a linker between the support and the ligand. Therefore, the mesoporous silica was reacted with 3-chloropropyltrimethoxysilane (CPTMS) as a linker to create Cl-MCM-41 material. In the subsequent step, the prepared aniline-oxazoline ligands 9a and 9b were treated with the Cl-MCM-41 in refluxing toluene (Page S8, Supplementary Material). During this process, the chloro group is replaced with the amino group, resulting in the formation of heterogeneous oxazoline based ligands 10 (OX-R-MCM-41) [98,108,111] (Scheme 6).

Scheme 6.

Immobilization of the aniline-oxazoline ligands 9 on mesoporous MCM-41.

2.3. Characterization of the chiral heterogeneous ligands 10

Various techniques were employed to study the structure of the prepared chiral heterogeneous ligands 10. These methods included Fourier transform infrared spectroscopy (FT-IR), powder X-ray diffraction (XRD), thermogravimetric analysis (TGA), differential thermal analysis (DTA), CHN elemental analysis, scanning electron microscopy (SEM), energy dispersive X-ray analysis (EDX), transmission electron microscopy (TEM), and BET/BJH nitrogen adsorption-desorption methods.

The results of the experimental analysis indicated that the aniline-oxazoline ligands 9 were successfully immobilized on the MCM-41. A comparison of the FT-IR spectra of MCM-41, Cl-MCM-41, and the chiral heterogeneous ligands 10 (OX-Ph-MCM-41 10a and OX-Bn-MCM-41 10b) showed notable differences in their respective spectra (Fig. 1).

Fig. 1.

The FT-IR spectra of MCM-41, Cl-MCM-41, and chiral heterogeneous ligands 10.

The FT-IR spectrum of MCM-41 exhibited the following spectral characteristics: A broad band observed at 3430 cm⁻¹ indicated the stretching vibration of O-H bonds in adsorbed water, accompanied by O-H bending vibration at 1634 cm⁻¹. The bands around 1221 and 1084 cm⁻¹ were identified as the asymmetric stretching of the Si-O-Si group, while the symmetric stretching modes of the Si-O-Si group were observed at 809 and 591 cm⁻¹. The presence of Si-OH groups was indicated by a band at 956 cm⁻¹, and an absorption band at 451 cm⁻¹ corresponded to the bending vibration of Si-O-Si groups.

In the FT-IR spectrum of Cl-MCM-41, a distinctive absorption band at 707 cm⁻¹ indicated the stretching of the C-Cl bond. Absorption bands at about 2953 and 2864 cm⁻¹ corresponded to the asymmetric and symmetric stretching vibrations of C-H bonds in the CH₂ groups of the propyl chain. Additionally, the intensity of the broad band at around 3400 cm⁻¹ is decreased, confirming the attachment of CPTMS on the mesoporous MCM-41.

The FT-IR spectra of the chiral heterogeneous ligands 10 exhibit characteristic vibrational bands indicative of the presence of functional groups. These bands are typically observed at around 1700 cm⁻¹ and correspond to the stretching vibrations of C=N and C=O groups present in the oxazoline ligands. Broad bands around 3400 cm⁻¹ correspond to the N-H groups of the oxazoline. These bands are a clear indication of the successful immobilization of the oxazoline ligands on the MCM-41. It is worth noting that distinct bands in the ranges of 1000–1250 cm⁻¹ and 450-950 cm⁻¹ were observed in all of the materials analyzed [108,111].

The powder XRD pattern of the synthesized mesoporous MCM-41 showed the presence of four distinct peaks at low angle of 2θ. The peak at (1 0 0) was particularly striking, exhibiting a significantly higher intensity in comparison to the other peaks such as (1 1 0), (2 0 0), and (2 1 0), which displayed relatively lower intensities [108,111] (Fig. 2).

Fig. 2.

The XRD patterns of MCM-41, Cl-MCM-41, and chiral heterogeneous ligands 10.

These characteristic peaks collectively confirm the creation of a well-structured hexagonal mesoporous substance with a p6mm space group [112]. The XRD patterns of the Cl-MCM-41 and the chiral heterogeneous ligands 10 displayed a general decrease in intensity of the peaks, potentially due to the contrast matching between the organic moieties that are integrated into the mesoporous channels of MCM-41 and the silica framework. However, the organic functionalization did not significantly alter the hexagonal structure of the channels. The lower intensity of OX-Ph-MCM-41 10a compared to OX-Bn-MCM-41 10b is likely attributed to a greater quantity of the aniline-oxazoline ligand 9a being immobilized on the MCM-41 than 9b.

The amounts of the immobilized the aniline-oxazoline ligands 9a and 9b on the MCM-41 were determined using thermogravimetric analysis (TGA) and CHN elemental analysis. The TGA diagrams showed that the chiral heterogeneous ligands 10a and 10b experienced significant weight loss between 150 and 600 °C, as illustrated in Fig. 3. Minor weight losses of around 1–2 wt% at temperatures below 150 °C were due to the elimination of physically adsorbed water. Additionally, the DTA analysis revealed a two-step weight loss at around 380 °C and 555 °C as a result of the oxazoline decomposition. The findings confirmed that the aniline-oxazoline ligands 9a and 9b are covalently bonded to the MCM-41 and demonstrated thermal stability up to 380 °C. Furthermore, the CHN elemental analysis revealed that the amounts of ligands 9a and 9b attached on the MCM-41 are nearly 0.5 mmol/g and 0.3 mmol/g, respectively.

Fig. 3.

The TGA of chiral heterogeneous ligands 10.

Fig. 4 displays scanning electron micrographs (SEM) of the MCM-41, the Cl-MCM-41 and the chiral heterogeneous ligands 10, aiming to examine potential morphological alterations. The analysis of the SEM images reveals that there are no noticeable differences in the size and morphology of the material before and after the immobilization of the aniline-oxazoline ligands 9.

Fig. 4.

SEM images of MCM-41, Cl-MCM-41, and chiral heterogeneous ligands 10.

Immobilization of the synthesized chiral ligands 9a and 9b on the MCM-41 was also confirmed using the EDX spectrum (Fig. 5). The presence of nitrogen and carbon elements serves as confirmation for the efficient immobilization of the ligands. It is worth noting, however, that the presence of chlorine suggests that there are still some unreacted linkers remaining.

Fig. 5.

The EDX spectra of chiral heterogeneous ligands 10.

The TEM micrograph of the chiral heterogeneous ligand 10a revealed the presence of a hexagonal structure, as shown in Fig. 6.

Fig. 6.

TEM image of chiral heterogeneous ligand 10a.

The specific surface areas (S_BET) of the chiral heterogeneous ligands 10a and 10b were determined using the BET method. The pore diameter at the peaks of the pore size distributions (D_p) was derived from the adsorption branches of the isotherms, employing Barrett-Joyner-Halenda (BJH) method; and the total pore volumes (V_p) were obtained based on the nitrogen uptake amount at a p/p₀ value of 0.99.

The summarized data can be found in Table 1. It was observed that all of the parameters decreased from the MCM-41 to the Cl-MCM-41 and to the chiral heterogeneous ligands 10, suggesting pore blocking caused by the linker and oxazoline ligands overloading. The chiral heterogeneous ligand 10a exhibited higher overloading compared to 10b, as previously confirmed through CHN and low-angle XRD analysis, which led to a decrease in specific surface area, lower total pore volumes, and narrower pore size distributions.

Table 1.

Textural properties of MCM-41, Cl-MCM-41, and chiral heterogeneous ligands 10.

Sample	SBET (m2/g)a	Vp (cm3/g)b	Dp (nm)c
MCM-41	1105	1.14	4.32
Cl-MCM-41	543	0.623	3.17
10a	229	0.188	2.40
10b	397	0.3903	2.93

^aS_BET is the surface area calculated from the BET equation.

^bV_p is the pore volume determined from the adsorption branch of the isotherm at p/p₀ = 0.99.

^cD_p represents the pore diameter calculated using the BJH method.

Fig. 7 presents the nitrogen adsorption-desorption isotherms of the chiral heterogeneous ligands 10a and 10b. The isotherms demonstrate a type IV behavior with H₄ hysteresis loops, indicating the presence of a narrow slit mesoporous structure, as classified by IUPAC [113]. The heterogeneous materials 10 display two distinct capillary condensation steps at low and high relative pressures. In the case of mesoporous structure 10a, capillary condensation occurs within the range of approximately 0.01 < p/p₀ < 0.30, and another condensation step is observed around 0.82 < p/p₀ < 0.99. This structure exhibits an H₄ hysteresis loop spanning from approximately p/p₀ = 0.25 to 0.95. Conversely, mesoporous structure 10b demonstrates capillary condensations at nearly 0.01 < p/p₀ < 0.25 and another at approximately 0.90 < p/p₀ < 0.99. Moreover, it displays an H₄ hysteresis loop within the range of approximately p/p₀ = 0.1 to 0.99. The pore size distribution curve, shown in Fig. 8, demonstrates a narrow distribution of pore sizes for mesoporous structure 10a.

Fig. 7.

The nitrogen adsorption–desorption isotherms of chiral heterogeneous ligands 10 at 77 K.

Fig. 8.

Pore size distribution curve of chiral heterogeneous ligand 10a.

2.4. Catalytic activity of the copper complexes of homogenous and heterogeneous amido-oxazoline ligands

2.4.1. Examination of the catalysts in the enantioselective Kharasch-Sosnovsky reaction

Based on our previous studies in the enantioselective Kharasch-Sosnovsky reaction, it has been firmly established that the inclusion of mesoporous additives has a profound influence on the overall outcomes, especially in terms of enantioselectivity [108,109,114]. Therefore, before studying the catalytic activity of the prepared heterogeneous catalysts, the homogeneous ones were investigated in the presence of mesoporous MCM-41 as an additive (Fig. 9).

Fig. 9.

Enantioselective allylic oxidation of cyclohexene using the chiral homogeneous oxazoline-copper(I) complexes.

To carry out the homogeneous reaction, cyclohexene was enantioselectively oxidized using t-butyl p-nitroperbenzoate 11a [96] in the presence of oxazoline ligands 1, 8, and 9, Cu(CH₃CN)₄PF₆, a highly effective copper source for this transformation [96,108,109,114] and a catalytic amount of phenyl hydrazine as a reducing agent to regenerate Cu(I) [96,108,114].

Based on Fig. 9, it can be inferred that the homogeneous catalysts with electron-withdrawing groups (EWG), such as NO₂, and Br on the phenyl ring, exhibited higher enantioselectivity (Fig. 1, Fig. 9). These results align with our previous research [96], which identified amido-oxazoline ligands containing a NO₂ group as the most effective. In contrast, the catalysts with NH₂, an electron-releasing group (ERG), displayed the lowest enantioselectivity and yield (Fig. 9, Ligands 9a and 9b), requiring three times the reaction time compared to the NO₂-containing catalysts (Fig. 1, Fig. 9). Catalysts with no substituent or an NHBoc group demonstrated efficiency lower than those with EWGs but higher than those with an NH₂ group (Fig. 1, Fig. 8, Fig. 9). It should be noted that the catalysts with a phenyl or benzyl group on the oxazoline ring outperformed those with i-propyl or i-butyl substitutions (Pages S8 and S9, Supplementary Material).

Moreover, the experimental evaluation of other cyclic and acyclic olefins under these conditions provided results similar to our prior findings [96,98]. Following the above experiments, the efficacy of the prepared heterogeneous catalysts was evaluated using various peresters (as depicted in Fig. 10) under similar conditions.

Fig. 10.

Comparison of the efficiency of the chiral heterogeneous catalysts using different peresters in the enantioselective allylic oxidation of cyclohexene.

It is noteworthy that while the homogeneous catalyst with the NH₂ group produced the poorest results (Fig. 9, Ligands 9a and 9b), the corresponding immobilized catalysts on the MCM-41 demonstrated outstanding enantioselectivity and yield. Remarkably, the heterogeneous catalysts 10a-Cu(I) and 10b-Cu(I) gave even more favorable results than the homogeneous ones (Fig. 9 vs. Fig. 10).

A comparison of the efficiency of the heterogeneous catalysts 10a-Cu(I) and 10b-Cu(I) showed that while the use of heterogeneous catalyst 10a-Cu(I), containing phenyl groups at the oxazoline rings, generally gave the best results, it typically required a longer reaction time. In general, under optimized conditions, the conversion of cyclohexene to enantioenriched (S)-allylic ester using t-butyl p-nitroperbenzoate 11a in the presence of heterogeneous catalyst 10a-Cu(I) led to remarkable enantioselectivity of 86 % and a high yield of 92 % within a 48-h timeframe.

Furthermore, in line with our prior researches [96,108,109,114] employing peresters containing EWGs (Fig. 10, Xʹ = NO₂, Cl) led to the formation of chiral allylic esters with higher enantioselectivity and yield in a shorter time compared to those with ERGs (Fig. 10, Xʹ = Me, OMe). In other words, there is a direct relationship between the strength of the EWGs on the perester and the enantioselectivity and yield of the reaction, meaning that an increase in the former leads to an improvement in the latter.

With optimized reaction conditions in hand, the generality of this protocol was investigated using other cyclic and acyclic olefins. The experimental data presented in Table 2 revealed that, under these conditions, both heterogeneous catalysts (10a-Cu(I) and 10b-Cu(I)) exhibited good efficiency for the studied substrates. Notably, allylic oxidation of 1,5-cyclooctadiene (1,5-COD) yielded better results than other olefins (entries 7 and 8). This reaction proceeded with the shortest reaction time, producing the (S)-enantiomer of the allylic ester product with the highest enantioselectivity and yield in comparison to other olefin substrates. This behavior can be linked to the catalyst complex cavity, which can effectively accommodate the twist-boat conformation of 1,5-COD [[115], [116], [117]].

Table 2.

Enantioselective copper catalyzed allylic oxidation of olefins using heterogeneous oxazoline based ligands.

Entry	Olefin	Heterogeneous Ligand	Allylic Ester	Time (h)	Yield (%)	ee (%)
1	Cyclopentene	10a	13	55	84	68
2	Cyclopentene	10b	13	52	77	71
3	Cycloheptene	10a	14	65	86	77
4	Cycloheptene	10b	14	60	90	68
5	Cyclooctene	10a	15	72	72	62
6	Cyclooctene	10b	15	72	65	65
7	1,5-COD	10a	16	34	95	88
8	1,5-COD	10b	16	34	95	83
9	1-Hexene	10a	17	132	42	28
10	1-Hexene	10b	17	108	35	33
11	1-Octene	10a	18	148	30	27
12	1-Octene	10b	18	148	33	22
13	1-Allyl benzene	10a	19	156	37	32
14	1-Allyl benzene	10b	19	156	29	26

Additionally, consistent with our expectations [12,96,98,118], the cyclic substrates produced higher enantioselectivity and yield in a shorter reaction time compared to the acyclic ones (Table 2, entries 1–8 vs. 9–14). The observed remarkable differences between cyclic and acyclic olefins can be attributed to the steric congestion encountered by the cyclic reactants at the key intermediate stage of the reaction. In contrast, acyclic olefins exhibit higher flexibility, allowing them to easily overcome the steric hindrance caused by the oxazoline ring by adopting a wide range of geometries [12].

Based on literature [[119], [120], [121], [122], [123], [124], [125], [126], [127]], the proposed mechanism of the reaction involves the complexation of the heterogeneous oxazoline based ligands 10 with Cu(I) to form a chiral catalyst. It is hypothesized that the chiral catalyst predominantly coordinates with cyclohexene, resulting in the formation of complex I (Scheme 7) [128].

Scheme 7.

Proposed mechanism of the enantioselective allylic oxidation of cyclohexene.

The addition of perester 11 initiates the reaction through a catalytic cycle that induces changes in the copper oxidation state. In the first step, cyclohexene undergoes substitution by perester 11. Subsequently, an oxidative-addition step occurs where the concerted cleavage of the oxygen-oxygen bond in perester 11 generates a crypto-tert-butoxyl radical and oxidizes Cu(I) to Cu(III). Following this, cyclohexene once again coordinates with the Cu(III). Then, in the rate-determining step, an intramolecular process takes place wherein the tert-butoxy group bonded to the copper complex selectively removes a prochiral allylic hydrogen from cyclohexene. Consequently, the elimination of tert-butyl alcohol leads to the formation of a key intermediate. A subsequent attack by carboxyl on the Si-face of cyclohexenyl triggers a pericyclic rearrangement involving the migration of the π-bond and stereospecific reductive-elimination step. Eventually, upon substitution with cyclohexene, Cu(I)-cyclohexene is regenerated, resulting in the release of the enantioenriched (S)-allylic ester 12.

2.4.2. Examination of the catalysts in the enantioselective Henry reaction

To further investigate the effectiveness of the prepared heterogeneous catalysts, they were also examined in the enantioselective Henry reaction (Table 3) [77,129,130]. Initially, the reaction between p-chloro benzaldehyde and nitromethane was conducted at room temperature in toluene, using Et₃N as a base, in the presence of a catalytic amount of the heterogeneous ligand containing phenyl group (10a) and Cu(CH₃CN)₄PF₆ (Table 3, entry 1). The reaction under these conditions resulted in the production of enantioenriched (S)-β-nitroalcohol 20a with low yield and enantioselectivity, and the reaction time was lengthy, taking 40 h to complete (Pages S9 and S10, and Figs. S28-S32, Supplementary Material).

Table 3.

Optimizing the enantioselective Henry reaction of nitromethane under various conditions.

Entry	X	Base (mmol)	Ligand	Cu salt	Solvent	T (°C)	Time (h)	Yield (%)	ee (%)
1	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	Toluene	r.t.	40	42	20
2	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	CH2Cl2	r.t.	33	53	13
3	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	THF	r.t.	22	58	22
4	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	Dioxane	r.t.	20	63	27
5	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	MeCN	r.t.	48	20	16
6	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	EtOH	r.t.	32	71	35
7	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	MeOH	r.t.	37	55	30
8	Cl	Et3N (1)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	15	70	40
9	Cl	TMG (1)a	10a (25)	Cu(CH3CN)4PF6	–	r.t.	20	68	24
10	Cl	DMA (1)b	10a (25)	Cu(CH3CN)4PF6	–	r.t.	14	23	32
11	Cl	Piperazine (1)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	36	32	13
12	Cl	NMPZ(1)c	10a (25)	Cu(CH3CN)4PF6	–	r.t.	30	25	20
13	Cl	Pyridine (1)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	48	43	32
14	Cl	Et3N (0.5)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	16	72	36
15	Cl	Et3N (0.25)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	12	80	33
16	Cl	Et3N (0.1)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	12	83	38
17	Cl	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	10	90	44
18	Cl	Et3N (0.025)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	18	65	40
19	Cl	Et3N (0.01)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	30	48	33
20	Cl	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	0	41	53	51
21	Cl	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	40	8	50	8
22	Cl	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	Reflux	8	10	0
23	Cl	Et3N (0.05)	10a (25)	Cu2O	–	r.t.	24	70	20
24	Cl	Et3N (0.05)	10a (25)	CuCl	–	r.t.	16	75	18
25	Cl	Et3N (0.05)	10a (25)	Cu(OAc)2	–	r.t.	12	86	32
26	Cl	Et3N (0.05)	10a (25)	Cu(NO3)2	–	r.t.	12	78	26
27	Cl	Et3N (0.05)	10a (25)	CuSO4	–	r.t.	24	82	12
28	Cl	Et3N (0.05)	10a (25)	CuCl2	–	r.t.	10	85	15
29	Cl	Et3N (0.05)	10a (25)	–	–	r.t.	14	70	6
30	Cl	Et3N (0.05)	10a (30)	Cu(CH3CN)4PF6	–	r.t.	13	79	48
31	Cl	Et3N (0.05)	10a (20)	Cu(CH3CN)4PF6	–	r.t.	10	85	33
32	Cl	Et3N (0.05)	10a (15)	Cu(CH3CN)4PF6	–	r.t.	20	72	15
33	Cl	Et3N (0.05)	10a (10)	Cu(CH3CN)4PF6	–	r.t.	24	66	5
34	H	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	20	60	21
35	OMe	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	27	57	23
36	NO2	Et3N (0.05)	10a (25)	Cu(CH3CN)4PF6	–	r.t.	5	95	50
37	NO2	Et3N (0.05)	10b (25)	Cu(CH3CN)4PF6	–	r.t.	5	95	43
38	NO2	Et3N (0.05)	1a (0.055)	Cu(CH3CN)4PF6	–	r.t.	2.5	85	35
39	NO2	Et3N (0.05)	1b (0.055)	Cu(CH3CN)4PF6	–	r.t.	2.5	90	30
40	NO2	Et3N (0.05)	1c (0.055)	Cu(CH3CN)4PF6	–	r.t.	4	85	25
41	NO2	Et3N (0.05)	1d (0.055)	Cu(CH3CN)4PF6	–	r.t.	4	80	28

^aTetramethylguanidine.

^bN,N-dimethylamine.

^cN-methyl piperazine.

To improve the yield and enantioselectivity of the product, various factors were investigated in this reaction, including the effects of solvent, base, temperature, Cu salt, and the heterogeneous ligand. By testing the reaction in polar and non-polar solvents, it was found that protic solvents, especially EtOH, enhanced both enantioselectivity and yield (Table 3, entries 1–7). However, the most favorable results were obtained under solvent-free conditions, where the reaction time decreased, and the enantioselectivity and yield of the chiral product were significantly improved (Table 3, entry 8).

When testing various bases, including tetramethylguanidine, N, N-dimethylaniline, N-methyl piperazine, piperazine, and pyridine, it was found that employing Et₃N as the base yielded the most favorable results [131] (Table 3, entries 9–13). Subsequently, different quantities of Et₃N were used in the reaction to ascertain the optimal amount (Table 3, entries 14–19). The experimental data revealed that decreasing the amount of Et₃N to 0.05 mmol had a significant positive impact on the reaction. Under these conditions, the reaction achieved a substantial increase in the yield, up to 90 %, along with an improvement in the enantioselectivity, which rose to 44 %. Additionally, the reaction time was notably shortened to just 10 h (Table 3, entry 17).

Further investigation revealed that the reaction temperature also has a significant impact on the outcomes of the reaction. While decreasing the temperature resulted in a slight improvement in enantioselectivity, it also led to reduced yield and a longer reaction time. Conversely, increasing the temperature of the reaction resulted in a shorter reaction time but, unfortunately, at the expense of a noticeable decrease in both yield and enantioselectivity (Table 3, entries 20–22).

Testing different copper salts showed that Cu(CH₃CN)₄PF₆ was the most effective (Table 3, entries 17 vs 23–28). Other Cu (I) and Cu (II) salts resulted in a slight decrease in yield, along with poor enantioselectivity and longer reaction times. It is worth noting that, in the absence of a Cu salt, the enantioselectivity was very low, while the yield and reaction time remained acceptable (Table 3, entry 29). It was found that increasing the ratio of the ligand to Cu(CH₃CN)₄PF₆ had a marginal positive effect on the enantioselectivity (Table 3, entry 30). However, this improvement came at the cost of a lower yield and a longer time. Experimental outcomes were inferior when the ratio of ligand to Cu salt was reduced (Table 3, entries 31–33).

Utilizing various aromatic aldehyde derivatives under optimal conditions indicated that the presence of an electron-withdrawing group on the aromatic aldehyde improved the results, whereas electron-donating groups greatly diminished the results (Table 3, entries 17, 36 vs. 34, 35). It should be noted that neither the utilization of the heterogeneous catalyst containing a benzyl group nor the homogeneous catalysts containing nitro groups led to any significant enhancement in the yield or enantioselectivity of the corresponding product (Table 3, entries 37–41).

Considering the aforementioned optimal results, the nitroethane substrate was also investigated in the asymmetric Henry reaction utilizing both chiral homogeneous and heterogeneous ligands, Cu(CH₃CN)₄PF₆ and Et₃N (Table 4). It is worth noting that this reaction yielded two diastereomeric mixtures, syn (1S, 2S) and anti (1S, 2R), with the resulting products obtained in high yield, moderate enantioselectivity, and diastereoselectivity (de = syn:anti), with the predominant syn diastereomer [6,30,31,[132], [133], [134], [135], [136], [137]].

Table 4.

Effect of reaction conditions on ee and de in the enantioselective Henry reaction of nitroethane.

Entry	X	Ligand	T (°C)	T (h)	Yield (%)	syn:anti (%)	ee (%)
1	Cl	1a	rt	24	85	60:40	27:22
2	Cl	1b	rt	20	77	63:37	17:25
3	Cl	1c	rt	30	75	43:57	11:23
4	Cl	1d	rt	30	88	50:50	19:14
5	Cl	10a	rt	12	92	70:30	30:25
6	Cl	10b	rt	10	83	67:33	42:33
7	Cl	10a	0	27	70	73:27	40:33
8	Cl	10a	−10	45	46	76:24	50:43
9	Cl	10a	40	8	95	50:50	13:8
10	Cl	-	rt	36	75	35:65	0
11	OMe	10a	rt	18	75	62:38	17:23
12	H	10a	rt	20	80	70:30	29:20
13	NO2	10a	rt	5	95	78:22	43:37

Similar to the nitromethane reaction, the heterogeneous catalysts showed better results than the homogeneous ones. Moreover, among the homogeneous catalysts, 1a exhibited the best efficiency. (Table 4, entries1-4). Regarding the heterogeneous catalysts, while the use of the heterogeneous catalyst containing a benzyl group resulted in the corresponding product in a slightly shorter reaction time and favorable enantioselectivity, the yield and diastereoselectivity were actually better when the heterogeneous catalyst containing a phenyl group was used (Table 4, entry 5 vs. 6).

Lowering the temperature could improve both de and ee; however, this improvement comes at the expense of a decrease in the yield and a longer reaction time (Table 4, entries 7 and 8). Conversely, increasing the temperature had an inverse effect (Table 4, entry 9).

Furthermore, similar to the above results, examination of different substituted aromatic aldehydes revealed that products with electron-withdrawing groups could be obtained with higher yields, ees and des (Table 4, entries 11–13).

The catalytic cycle of the enantioselective Henry reaction, as proposed in Scheme 8, begins with the deprotonation of the nitroalkane by Et₃N at the α-carbon position, resulting in the formation of a nitronate anion, which then coordinates to the oxazoline copper complex.

Scheme 8.

Plausible mechanism of the enantioselective Henry reaction.

Subsequently, the aldehyde coordinates to the copper atom, and in a rationalized transition state based on the obtained products (TS1), the nitronate attacks the aldehyde on the re face, predominantly leading to the formation of the chiral syn-(1S, 2S)-β-nitroalcohol isomer. The increased syn:anti ratio of the products can be attributed to the steric hindrance between the methyl group in the nitronate anion and the aromatic group of the aldehyde in the transition state (TS1 vs. TS2) [77,79,94,136,137].

3. Catalyst reusability

The reusability of chiral heterogeneous catalyst 10a in both asymmetric reactions was evaluated under optimal conditions. After completion, the catalyst was filtered, washed with CH₂Cl₂, and the resulting precipitate was dried at 50 °C. The recovered catalyst was then used for the next run of the same reaction under similar conditions. Repeating this process revealed that the activity and enantioselectivity of the heterogeneous catalyst remained relatively constant even after four consecutive cycles, demonstrating its potential for long-term application and reduced environmental impact.

Fig. 11 depicts the results for the enantioselective allylic oxidation of cyclohexene in the optimum conditions over five sequential cycles. Moreover, the XRD pattern (Fig. S33, Supplementary Material) of the catalyst after four cycles showed that the structure of the recycled catalyst did not change noticeably.

Fig. 11.

Reusability of catalyst 10a in the enantioselective allylic oxidation of cyclohexene.

4. Conclusions

In summary, we successfully prepared and characterized a class of recyclable heterogeneous oxazoline-based ligands by immobilizing synthesized chiral amido-oxazoline on MCM-41 mesoporous silica. The performance of these chiral heterogeneous ligands was evaluated in two enantioselective reactions: the allylic oxidation of olefins and the Henry reaction. In both cases, heterogeneous catalyst outperformed their homogeneous counterparts in terms of enantioselectivity. Significantly, the enantioselective Henry reaction proceeded efficiently under solvent-free conditions, aligning with green chemistry principles. Additionally, the use of nitroethane in the Henry reaction yielded predominantly the syn diastereomer. Future investigations will explore immobilization on alternative supports such as SBA-15 and MOFs, with the aim of optimizing reaction conditions and broadening the scope of these heterogeneous catalysts.

CRediT authorship contribution statement

Niloofar Tavakoli: Methodology, Investigation, Formal analysis. Hamid Arvinnezhad: Writing – original draft, Methodology, Investigation, Formal analysis. Shiva Majidian: Methodology, Investigation, Formal analysis. Mahsa Mahramasrar: Methodology, Investigation, Formal analysis. Khosrow Jadidi: Supervision, Methodology, Investigation, Formal analysis, Conceptualization. Saadi Samadi: Writing – original draft, Supervision, Methodology, Investigation, Formal analysis, Conceptualization.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used GPT3.5 in order to improve writing. After using this service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: