ε-Caprolactam (CL) is an important precursor of nylon-6 and several other plastics, and the global market for this material reached 5 million metric tons per annum in 2015 ^[1]. CL is currently manufactured from cyclohexanone, which is converted to cyclohexanone oxime (CHO) before undergoing a rearrangement process to give CL via a three step process. The yield for the Beckmann rearrangement of CHO to CL is high when fuming sulfuric acid is used as catalyst and solvent. However, there are several disadvantages associated with this process, including (1) the formation of a large amount of ammonium sulfate, which is generated as a by-product during the neutralization of the hazardous sulfuric acid waste with ammonia, (2) the corrosion of the reactors because of the strongly acidic conditions, and (3) the environmental pollution caused by the use of fuming sulfuric acid. Although cyclohexanone can be obtained by the hydrogenation of phenol ^[2], the preferred route for the industrial manufacture of this material involves the direct oxidation of cyclohexane. However, the aerobic oxidation of cyclohexane to cyclohexanone has limited the overall efficiency of this three-step process because of the low yields achieved with each pass, which are usually in the range of 3%-8% with a selectivity of around 80% for cyclohexanone over cyclohexanol ^[3]. Based on the increasing demand for environmentally friendly processes, the development of an ecofriendly and economically viable procedure for the one-pot synthesis of CL is highly desirable. Several groups have investigated the direct synthesis of CL from cyclohexane using various strategies, including conducting the reaction in the presence of nitrosyl sulfuric acid ^[4-9] or using nitrocyclohexane as a substrate ^[10]. One group even investigated the possibility of using cyclohexanone in combination with ammoximation and Beckmann rearrangement reactions over Nb-MCM-41 catalysts ^[11], which resulted in a CL yield of about 10%. Ishii’s group ^[12] reported the synthesis of CL from cyclohexane and t -butyl nitrite, using N -hydroxyphthalimide (NHPI) as a catalyst, which gave a CL yield of 43%. However, the conditions used in this reaction resulted in the decomposition of NHPI to give phthalic acid and several other related impurities, thereby limiting the utility of this approach. Raja et al. ^[13] developed a one-step process for the conversion of cyclohexanone to CL using bimetallic Mn(III)Mg(II)AlPO-5 as a bifunctional catalyst. The maximum conversion of cyclohexanone achieved under these conditions was 68.3%, with selectivity of 5.7% and 77.9% for CHO and CL, respectively. More recently, Shin et al. ^[14] reported a multifunctional Pd/Sc(OTf)₃/ionic liquid catalyst for the tandem one-pot conversion of phenol to CL in 67% yield using a large excess (300 mol%) of 1-butyl-methylimidazolium hexafluorophosphate ([bmim]PF₆) as an additive. Although this reaction gave a high phenol conversion and good selectivity for CL, it was difficult to separate the IL from the reaction mixture, thereby limiting the effectiveness of this approach.

We envisaged that it would be possible to overcome the issues described above by directly synthesizing CL from cyclohexanol, which is an unavoidable by-product of the oxidation of cyclohexane. Thus, the oxidation of cyclohexanol in the liquid phase using H₂O₂, followed by sequential oximation and Beckmann rearrangement reaction in a single pot would represent an attractive alternative for the synthesis of CL (Scheme 1). This route would be particularly interesting from an industrial perspective, because all three steps, including (1) the oxidation of cyclohexanol to cyclohexanone, (2) the oximation of cyclohexanone to CHO, and (3) the rearrangement of CHO to CL could be performed as a one-pot procedure in the presence of a multifunctional catalysts.

The “green” oxidation of alcohols using tungsten peroxide as an oxidant, which may be prepared by the reaction of H₂O₂ with a catalytic amount of sodium tungstate (Na₂WO₄), has attracted considerable attention ^[15]. However, it is only possible to obtain high yields of the alcohol products in this way under acidic conditions with a phase transfer catalyst ^{[15, 16]}. We recently investigated the direct synthesis of ε-caprolactam from cyclohexanol using [ n -C₁₆H₃₃N(CH₃)₃]H₂PW₁₂O₄₀ as a catalyst ^[17].

In this study, we synthesized nine Brnsted acidic ILs (Fig. 1) and investigated their use as catalysts for the one-pot synthesis of CL from cyclohexanol using H₂O₂ and hydroxylamine. In addition to CL, we also obtained several other valuable products, including cyclohexanone and CHO. The results ultimately revealed that Na₂WO₄-sulfonic acid-functionalized ionic liquids could be used as catalyst for the synthesis of CL. Furthermore, the final products were readily separated from the catalyst system by a simple extraction.

All of the chemicals used in the current study were purchased as the AR grade from The Tianjin Fu Chen Chemical Reagent Factory (Tianjin, China).

[Hnmp]HSO₄, [Hmin]HSO₄, [Tma]HSO₄, [Tea]HSO₄, [BSMin]HSO₄, [BSTea]HSO₄, [BSTma]HSO₄, [BSTma]H₂PO₄ and [BSTma] p -TA were prepared using previously reported procedures ^[18-24]. The resulting ILs were identified by ¹H NMR and IR analyses.

In a typical procedure, Na₂WO₄ (2.75 mmol) and an ionic liquid were mixed with cyclohexanol (50 mmol) in a 100-mL three-necked flask equipped with a reflux condenser, magnetic stirrer and a thermometer. The resulting mixture was then heated to the required temperature, and treated with 30% H₂O₂ (75 mmol), which was added in a dropwise manner to the reaction mixture with vigorous stirring over 5 h. Hydroxylamine sulfate (25 mmol) was then added to the flask with vigorous stirring over a certain period of time. Upon completion of the reaction, the mixture was allowed to settle to give two phases. The aqueous phase was collected and extracted with dichloromethane, and the extracted products were identified by gas chromatography-mass spectrometry using a Thermo Trace DSQ gas chromatograph-mass spectrometer (Thermo Electron, Massachusetts, America). The products were analyzed on an SP-3420A gas chromatograph equipped with a KB-Wax column (30 m × 0.32 mm × 0.25 μm).

The cyclohexanol conversion (X _ANOL), product selectivity (S_i ) and carbon balance were calculated as follows:

X _ANOL = (Moles of cyclohexanol added - moles of unconverted cyclohexanol)/moles of cyclohexanol added × 100%.

S_i = Moles of product i /(moles of cyclohexanol added - moles of unconverted cyclohexanol) × 100%.

Carbon balance = (Moles of unconverted cyclohexanol + moles of carbon in cyclohexanone + moles of carbon in CHO + moles of carbon in CL)/moles of carbon in cyclohexanol × 100%.

The direct synthesis of CL from cyclohexanol would require three steps. The first step would involve the selective oxidation of cyclohexanol to cyclohexanone. The product cyclohexanone would then react with hydroxylamine to form CHO, which would undergo a rearrangement reaction to give CL. To identify a suitable catalytic system for the oxidation of cyclohexanol to cyclohexanone using Na₂WO₄, we screened a variety of different ionic liquids (Table 1). In the absence of the catalyst, the yield of cyclohexanone was only 1.08% (Table 1, entry 1). Among the four non-sulfonic acid-functionalized ILs tested in the current study (Table 1, entries 2-5), [Tma]HSO₄ showed the highest catalytic activity. The catalytic activities of ILs featuring alkanesulfonic acid groups (Table 1, entries 6-10) were much higher than those of the ILs without an alkanesulfonic acid group (Table 1, entries 2-4). In particular, [BSTma]HSO₄-Na₂WO₄ gave an excellent catalytic performance with a cyclohexanol conversion of 97.7% and a selectivity for cyclohexanone of 96.2%.

The oxidation reaction proceeded smoothly in an organic-aqueous biphasic system. In the aqueous phase, the catalyst precursor Na₂WO₄ was rapidly oxidized by H₂O₂, as follows: Na₂WO₄ + 2H₂O₂ = Na₂[WO(O₂)₂(OH)₂] + 2H₂O ^[25]. According to Noyori’s report ^[25], the resulting bisperoxotungstate compound A would exist in equilibrium with compounds B and C during the oxidation process (Fig. 2). Noyori et al. ^[25] also indicated that the dianion in A exhibited very little activity toward alcohols, whereas the mono- and di-protonated forms in B and C were sufficiently reactive. The bisperoxotungstate compound was readily transferred to the organic phase by a Na⁺-Q⁺ ion exchange mechanism. The catalytic activity observed during the oxidation process would therefore be highly dependent on the acidity of the reaction medium, whereas the efficiency of the aqueous-organic phase transfer would be dependent on the active species aided by Q⁺ (Q refers to the cation of the ILs) ^[25]. The high catalytic performance of sulfonic acid-functionalized ionic liquids can therefore be attributed to the improved hydrophobicity of their cation, whilst their improved acidity can be attributed to the introduction of alkanesulfonic acid groups. A similar effect was also observed for changes in the alkyl chain length during a study of the selective oxidation of cyclohexanol to cyclohexanone ^{[26, 27]}.

As shown in Table 1, the catalytic activities of [BSTma]H₂PO₄ and [BSTma] p -TA were poor. The acidity of an IL is dependent on the nature of its cations and anions. The acidities of ILs bearing the same cations are dependent on the difference in their anions. ILs featuring [H₂PO₄]^- as a conjugate base are much less acidic than those featuring [HSO₄]^- as a conjugate base, according to H ₀ ^[27]. The catalytic activity of [BSTma] p -TA was poor, most likely because its p -CH₃C₆H₄SO₃^- anion cannot provide protons, resulting in the low acidity of the IL. Based on these results, we concluded that the hydrophobic cation, hydrophilic anion and acidity characteristics of the ILs were responsible for the oxidation of cyclohexanol ^[28].

It seems almost contradictory to suggest that a strong acid and dehydrating media are required to ensure the protonation of the oxime intermediate during the rearrangement step, when water from the H₂O₂ could lead to the hydrolysis of the oxime or the formation of undesirable by-products under the strongly acidic conditions ^{[29, 30]}. With this in mind, we investigated the effect of the molar ratio of cyclohexanone to [BSTma]HSO₄ on the outcome of the one-pot conversion of cyclohexanone to CL (Fig. 3). In the absence of [BSTma]HSO₄, the yields of CL and CHO were 62.7% and 26.8%, respectively. The yield of CL increased as the molar ratio of cyclohexanone to [BSTma]HSO₄ increased from 1:0.04 to 1:0.08. However, further increasing the molar ratio to 1:0.16 did not lead to any further changes in the yield of CL and the carbon balance gradually decreased, which was attributed to the formation of a high boiling condensation by-product. The one-pot conversion of cyclohexanone to CL most likely proceeds via the non-catalytic oximation of cyclohexanone to CHO, followed by the Beckmann rearrangement of CHO to CL. The mechanism for the rearrangement of CHO involves the Brnsted acid-mediated protonation of the nitrogen atom of the oxime, followed by a proton transfer from the nitrogen atom to the oxygen atom of the oxime ^{[31, 32]}. The formation of an O-protonated oxime in this way is a 1, 2-H-shift reaction, and reactions of this type generally require a high activation energy. For this reason, this step is considered to be the rate-determining step in the Beckmann rearrangement.

According to previous reports from the literature, solvents with high dielectric constants such as benzonitrile, DMSO and MeCN can play an important role in lowering the activation energy of a Beckmann reaction by interacting directly with the N -protonated oxime to form a cyclic transition state ^[33-35]. In the one-pot solvent free conversion of cyclohexanone to CL, the sulfonic acid groups associated with the cations on the sulfonic acid-functionalized ionic liquid could facilitate the formation and stabilization of the positively charged intermediate based on their strong acidity and high polarity ^[36]. A cyclohexanone to [BSTma]HSO₄ molar ratio of 1:0.08 was therefore selected as the optimum molar ratio for the one-pot conversion of cyclohexanone to CL.

Finally, we investigated the one-pot conversion of cyclohexanol to CL using Na₂WO₄-[BSTma]HSO₄ as a catalyst (Table 2). In the absence of a catalyst, this reaction did not afford any CL (Table 2, entry 1). When the same reaction was conducted in the presence of Na₂WO₄-[BSTma]HSO₄, the yield of CL reached 49.1% (Table 2, entry 3). Na₂WO₄-[BSTma]HSO₄ was therefore used as a catalyst in the one-pot synthesis of CL. During the oxidation process, the catalyst acted as a phase-transfer catalyst based on its ion exchange behavior, and the acidity of the catalyst system improved the coordination of the substrate to bisperoxotungstate. Our newly developed catalyst also lowered the activation energy for the Beckmann rearrangement by stabilizing the charged intermediate formed during the rearrangement, to ultimately afford CL ^[14].

We subsequently proceeded to investigate the effects of several key parameters, including the reaction temperature, reaction time and the molar ratio of cyclohexanone to [BSTma]HSO₄ on the outcome of the Na₂WO₄-[BSTma]HSO₄-catalyzed one-pot conversion of cyclohexanol to CL. The tandem reaction was divided into three steps. The first and the third steps were determined to be exothermic, whereas the second one was slightly exothermic. The cyclohexanol conversion and yield of CL increased with increasing temperature from 70 to 80 ℃, where the yield of CL reached its maximum value of 49.1% (Table 2, entries 2 and 3). Although further increases in the temperature resulted in a decrease in the yield of CL (Table 2, entries 4 and 5), the selectivity for cyclohexanone continued to increase. The high selectivity for cyclohexanone, which was formed by the hydrolysis of the CHO, resulted in a decrease in the selectivity for CL for reaction temperatures above 80 ℃.

The cyclohexanol conversion remained unchanged for oximation and rearrangement reaction times in the range of 60 to 180 min. The CL yield initially increased and reached its maximum value of 49.1% (Table 2, entries 3, 6, and 7), before decreasing to 41.4% after extending the reaction time by 30 min (Table 2, entry 8). Furthermore, the selectivity for cyclohexanone increased from 10.3% to 19.1% during this period. This result indicated that the CHO formed during the reaction was being hydrolyzed in the presence of [BSTma]HSO₄ and water during the prolonged reaction times ^[37]. When the molar ratio of cyclohexanone to IL7 was 1:0.08, Na₂WO₄-[BSTma]HSO₄ showed high catalytic activity, with the conversion of cyclohexanol and yield of CL reaching 97.3% and 76%, respectively.

With the optimized reaction conditions in hand, we proceed to investigate the scope of this reaction by synthesizing a series of N -substituted amides (Table 3). Pleasingly, a wide range of alcohol substrates, including various aromatic and cyclic alcohols, reacted smoothly under the optimized conditions to give the desired products in excellent yields. Furthermore, these results compared well with those previously reported for the reactions of the corresponding ketoximes using [ n -C₁₆H₃₃N(CH₃)₃]H₂PW₁₂O₄₀ ^[17], Cs _x H_{3- x} PW₁₂O₄₀ ^[33], propane sulfonic acid-functionalized imidazolium salt of phosphotungstate ^[34], CL-based Brnsted acidic ionic liquid ^[38], heteropolyanion-based ionic liquid ^[39] and H₃PW₁₂O₄₀ ^[40] as catalysts (shown in Table 3).

Notably, our newly developed catalyst gave an excellent performance for the conversation of isopropanol and sec-butyl alcohol. This result was particularly impressive because previous reports had suggested that it was not possible to convert these substrates to amides using novel Brnsted acidic ILs bearing double -SO₃H cations ^[41]. The CL yield achieved in this work was lower than that reported by Zhang et al. ^[29] (99%), where cyclohexanone was directly converted to CL using trifluoroacetic acid as a catalyst with acetonitrile as an additive.

Table 4 summarizes the catalytic performance of various catalysts previously reported for the synthesis of CL. The yield of CL over Na₂WO₄-[BSTma]HSO₄ was much higher than that of Nb-MCM-41 using cyclohexanone as a reactant (Table 4, entry 1). The yield of CL over Na₂WO₄-[BSTma]HSO₄ was higher than that of Pd-Sc(OTf)₃/[bmim]PF₆ (Table 4, entry 2), where 300 mol% 1-butyl-methylimidazolium hexafluorophosphate ([bmim]PF₆) was used as an additive. The yield of CL over Na₂WO₄-[BSTma]HSO₄ was much lower than that of cyclohexanone when cyclohexanone was used as a reactant (Table 4, entry 4). However, the application of this process is limited by the difficulties associated with the recovery of the homogeneous CF₃COOH catalyst from the reaction mixture. Na₂WO₄-[BSTma]HSO₄ exhibited comparable catalytic activity to all of the other catalysts using CHO as a reactant (Table 4, entries 5-10). These results therefore demonstrate that Na₂WO₄-[BSTma]HSO₄ is an efficient catalyst for the one-pot conversion of cyclohexanol to CL.

Upon completion of the reaction, the mixture was allowed to settle to form two distinct phases. The aqueous phase was collected and extracted with dichloromethane, with the ILs and Na₂WO₄ completely dissolving in the aqueous phase. To investigate the reusability of the catalyst, the aqueous phase from the first run was concentrated to dryness using a rotary evaporator, and the resulting residue was stored at 60 ℃ under vacuum to obtain the recycled catalyst. This process was repeated for four cycles. Notably, the yield of CL obtained using the recycled catalyst decreased only slightly (3.5%) after four runs. This loss in activity can most likely be attributed to the physical losses incurred during the recycling process.

A new process has been developed for the one-pot synthesis of CL from cyclohexanol using H₂O₂ and hydroxylamine in the presence of a Na₂WO₄-acidic ionic liquid catalyst. Sulfonic acid-functionalized ionic liquids bearing HSO₄^- anions showed excellent catalytic activities. Under the optimized conditions, we achieved a cyclohexanol conversion of 97.3% with a CL yield of 76.0%. This new reaction system has several advantages such as ease of operation, mild reaction conditions and environmental friendliness. Furthermore, the catalyst can be readily recycled without suffering any major losses in its catalytic activity. Notably, the results of our one-pot procedure are comparable with the yield achieved for the three-step synthesis of CL, highlighting its potential application. We have therefore shown that the sequential oxidation of cyclohexanol to cyclohexanone, oximation of cyclohexanone to CHO and rearrangement of CHO to CL can be conducted as a one-pot reaction.