Ionic Liquids Based on Oxidoperoxido-Molybdenum(VI) Complexes with a Chelating Picolinate Ligand for Catalytic Epoxidation

: Ionic oxidoperoxido-molybdenum(VI) complexes of the type [Cat][MoO(O 2 ) 2 (pic)], with pic = N,O-chelated picolinate ligand and Cat = monocation, were prepared in high yields (82–95%) from the precursor complex [H 3 O][MoO(O 2 ) 2 (pic)] via [H] + cation exchange for 1-ethyl-3-methylimidazolium [EMIM] + ,1-butyl-3-methylimidazolium[BMIM] + ,1-octyl-3-methylimidazolium[OMIM] + , N -cetylpyridinium [C 16 Py] + ,and N -methyl- N , N , N -trioctylammonium[Aliquat] + . Thestructureandpurityoftheioniccompounds were assessed by 1 H and 13 C NMR, FT-IR, and elemental analysis (C, H, N), and the electrochemical properties were studied by differential pulse voltammetry (DPV) and cyclic voltammetry (CV). The [Cat][MoO(O 2 ) 2 (pic)] compounds showed promising catalytic epoxidation activity based on the model reaction of cis -cyclooctene with tert -butyl hydroperoxide as oxidant. The type of cation influenced the physical state of the compound and the catalytic performance.

In this work, ionic complexes of the type [Cat][MoO(O 2 ) 2 (L)] were synthesized, characterized, and tested as epoxidation catalysts with TBHP, without the need of additional co-catalysts or co-additives.

Materials and Methods
Commercially available reagents were used for the syntheses of the complexes and catalytic tests. The anion exchange resin Amberlite IRA-400-OH (hydroxide ion-exchange capacity = 1.4 meq. mL −1 ) was purchased from Supelco. [OH] solution was slowly added to a 0.1 M aqueous solution (10 mL) of 1 (1 mmol; 1.4 eq.). The reaction mixture was stirred at room temperature for 1 h, and subsequently extracted several times using dichloromethane. The combined extracts were subjected to solvent evaporation followed by vacuum-drying at room temperature, giving the pure product as a yellow powder (0.28 g, 96%). m.p. 82-86 • C. 1  [Cl] (0.25 g, 1.43 mmol) was dissolved in distilled water (20 mL) and passed through an anion exchange column loaded with Amberlite IRA-400(OH) (5.2 mL; 5 eq., flux rate 8 BV h −1 ). The resultant [BMIM][OH] solution was slowly added to a 0.1 M aqueous solution (20 mL) of 1 (2 mmol; 1.4 eq.). The reaction mixture was stirred at room temperature for 1 h and then extracted several times using dichloromethane. The combined extracts were subjected to solvent evaporation followed by vacuum-drying at room temperature, giving the pure product as a yellow viscous liquid (0.51 g, 82%). m.p. 44-52 • C. 1  solution (10 mL) of 1 (1 mmol; 1.4 eq.). The reaction mixture was stirred at room temperature for 1 h and then extracted several times using dichloromethane. The combined extracts were subjected to solvent evaporation followed by vacuum-drying at room temperature, giving the pure product as an amorphous solid (0.32 g, 91%). m.p. 106-108 • C. 1  [C 16 Py][MoO(O 2 ) 2 (pic)] (5). [C 16 Py][Cl] (0.240 g, 0.71 mmol) was dissolved in distilled water (10 mL) and passed through an anion exchange column loaded with Amberlite IRA-400(OH) (2.5 mL; 5 eq., flux rate 8 BV h −1 ). The resultant N-cetylpyridinium hydroxide solution was slowly added to a 0.1 M aqueous solution (10 mL) of 1 (1 mmol; 1.4 eq.). The reaction mixture was stirred at room temperature for 1 h. The product was filtered using a Hirsh funnel, washed with distilled water, and vacuum-dried at room temperature to give the pure product as a pale-yellow powder (0.39 g, 91%). m.p. 89-92 • C. 1   ). The reaction mixture was stirred at room temperature for 1 h, and then subjected to solvent evaporation. The concentrated solution was diluted with distilled water and subsequently extracted several times using chloroform. The combined extracts were subjected to solvent evaporation followed by vacuum-drying at room temperature, giving the pure product as a yellow coloured, room temperature ionic liquid (RTIL) (0.45 g, 95%). 1  Catalytic epoxidation tests: The catalytic tests were carried out at 70 • C under air in a closed borosilicate batch reactor with 5 mL capacity, equipped with a V-shaped magnetic stirring bar and a valve for sampling. Typically, the reactor was loaded with 1 mol.% molybdenum complex relative to the substrate: the substrate (1.8 mmol), the oxidant (2.74 mmol) and a solvent (1 mL). The solvents used were α,α,α-trifluorotoluene (TFT), acetonitrile and 1,2-dichloroethane (DCE). The loaded reactor was immersed in a thermostatically controlled oil bath heated at 70 • C, with a stirring rate of 1000 rpm; this was taken as the initial instant of the epoxidation reaction.
Prior to sampling and analysis of freshly prepared samples, the reactor was quenched in cold water, and the reaction mixture was subjected to centrifugation (3500 rpm). The catalytic reactions were monitored using an Agilent 7820A GC equipped with a capillary column (HP-5 30 m × 0.320 mm × 0.25 mm), a flame ionization detector, and undecane was used as internal standard. Product identification was carried out using a GC-MS (Trace GC 2000 series (Thermo Quest CE instruments)-DSQ II mass detector (Thermo Scientific)) equipped with an Agilent J&W DB1 capillary column (DB-1 MS, 30 m × 0.25 mm × 0.25 mm), with He as carrier gas. TOF were based on catalytic results at 6 h reaction. Conversion was calculated using the formula 100 × [((initial molar concentration of Cy)-(molar concentration of Cy at time t))/(initial molar concentration of Cy)] and cyclooctene oxide (CyO) selectivity was calculated using the formula 100 × [(molar concentration of CyO at time t)/((initial molar concentration of Cy)-(molar concentration of Cy at time t))]. DSQ II mass detector (Thermo Scientific)) equipped with an Agilent J&W DB1 capillary column (DB-1 MS, 30 m × 0.25 mm × 0.25 mm), with He as carrier gas. TOF were based on catalytic results at 6 h reaction. Conversion was calculated using the formula 100 × [((initial molar concentration of Cy)-(molar concentration of Cy at time t))/(initial molar concentration of Cy)] and cyclooctene oxide (CyO) selectivity was calculated using the formula 100 × [(molar concentration of CyO at time t)/((initial molar concentration of Cy)-(molar concentration of Cy at time t))].  The electrochemical properties of the ionic complexes were studied by differential pulse voltammetry (DPV) and cyclic voltammetry (CV) (Figure 1) using as reference SCE (standard calomel  [27,62]. None of the compounds exhibited the characteristic band of the carbonyl group attributed to the free picolinic acid (2-pyridinecarboxylic acid) at 1722 cm −1 [64]. All spectral data (IR, 1 H, and 13 C NMR), as well as elemental analysis (C, H, N), suggested pure compounds containing an anionic metal complex.

Characterization of the Ionic Complexes
The electrochemical properties of the ionic complexes were studied by differential pulse voltammetry (DPV) and cyclic voltammetry (CV) (Figure 1) using as reference SCE (standard calomel electrode in 3 M KCl). In general, the voltammograms in DPV showed one predominate oxidation peak at approximately 1.45 V, while the corresponding wave in CV is irreversible.   (Table 1). Overall, the DPV and CV results suggested that the variation of the cation structure does not significantly affect the electrochemical profile of the anionic complex.

cis-Cyclooctene Catalytic Epoxidation
The  Table 2). The TBHP: Cy molar ratio was 1.6, and [Cat][MoO(O2)2(pic)] was added in an amount equivalent to 1 mol% of molybdenum relative to Cy. For all compounds already prepared, the turnover number (TON) was greater than unity (13−100 molCy molMo -1 ), and up to 100% Cy conversion was reached (Figure 2a). Without catalyst, conversion was negligible. Cyclooctene oxide (CyO) was the main product formed in up to 100% yield ( Figure  2b). These results demonstrate the ability of the catalysts [Cat][MoO(O2)2(pic)] to trigger the epoxidation process with TBHP and without requiring co-catalysts.
Mechanistic studies for molybdenum-catalysed epoxidation of olefins with alkyl hydroperoxide  (Table 1). Overall, the DPV and CV results suggested that the variation of the cation structure does not significantly affect the electrochemical profile of the anionic complex.

cis-Cyclooctene Catalytic Epoxidation
The  Table 2). The TBHP: Cy molar ratio was 1.6, and [Cat][MoO(O 2 ) 2 (pic)] was added in an amount equivalent to 1 mol% of molybdenum relative to Cy. For all compounds already prepared, the turnover number (TON) was greater than unity (13−100 mol Cy mol Mo olefin epoxidation using stoichiometric amounts of neutral complexes [MoO(O2)2(L)2] (L = pyrazole or 3,5-dimethylpyrazole), with or without a hydroperoxide oxidant, and suggested that the epoxidation reaction mechanism could proceed via Sharpless [65] or Thiel type mechanisms [2,[69][70][71]. In the present study, only 1 mol% complex was used, thus if the complex acted as oxidant (and considering one atom of active oxygen per peroxido ligand), it would lead to a maximum Cy conversion of 2%, which is negligible. The catalytic cycle of the system [Cat][MoO(O2)2(pic)]/TBHP likely involves a Lewis acid-base type mechanism with TBHP as the oxygen atom source. The influence of the type of N-containing cation, i.e., imidazolium, pyridinium or ammonium derivatives, on the performance of [Cat][MoO(O2)2(pic)] type catalysts was studied, using TFT as solvent at 70 °C. The turnover frequency (TOF, molCy molMo -1 h −1 ) increased in the order 2 (2.1) < 3 (2.9) ≈ 5 (2.9) < 4 (3.9) ≈ 6 (4.2) (Figure 2a). The parent compound 1 exhibited the highest activity, leading to 100% CyO yield at 3 h, compared to 42% CyO yield at 24 h with 1,2-cyclooctanediol (CyDOL) as by-product (13% yield) for the best-performing catalyst possessing a N-containing cation, namely 6 ( Figure 2b). The catalyst 6 was tested using different types of solvents (TFT, DCE, MeCN), which indicated that the most favourable one was TFT. Specifically, Cy conversion at 24 h increased in the order MeCN (27%) < DCE (39%) < TFT (55%). The poorer results for MeCN as solvent may be Mechanistic studies for molybdenum-catalysed epoxidation of olefins with alkyl hydroperoxide oxidants corroborate a Lewis acid-base reaction in which the molybdenum compound acts as a Lewis acid and the alkyl hydroperoxide (TBHP) as a Lewis base, leading to an active oxidising species that is involved in the oxygen atom transfer to the olefin, finally giving the epoxide product [65][66][67][68][69]. In the present study, the catalytic epoxidation system [Cat][MoO(O 2 ) 2 (pic)]/TBHP is an alternative strategy to that firstly reported by Herbert et al. [2] involving the use of [Bu 4 N][MoO(O 2 ) 2 (pic)] as stoichiometric oxidant (not as catalyst) and Co(acac) 2 as catalyst. Later, the same group reported olefin epoxidation using stoichiometric amounts of neutral complexes [MoO(O 2 ) 2 (L) 2 ] (L = pyrazole or 3,5-dimethylpyrazole), with or without a hydroperoxide oxidant, and suggested that the epoxidation reaction mechanism could proceed via Sharpless [65] or Thiel type mechanisms [2,[69][70][71]. In the present study, only 1 mol% complex was used, thus if the complex acted as oxidant (and considering one atom of active oxygen per peroxido ligand), it would lead to a maximum Cy conversion of 2%, which is negligible. The catalytic cycle of the system [Cat][MoO(O 2 ) 2 (pic)]/TBHP likely involves a Lewis acid-base type mechanism with TBHP as the oxygen atom source.
The influence of the type of N-containing cation, i.e., imidazolium, pyridinium or ammonium derivatives, on the performance of [Cat][MoO(O 2 ) 2 (pic)] type catalysts was studied, using TFT as solvent at 70 • C. The turnover frequency (TOF, mol Cy mol Mo -1 h −1 ) increased in the order 2 (2.1) < namely 6 ( Figure 2b). The catalyst 6 was tested using different types of solvents (TFT, DCE, MeCN), which indicated that the most favourable one was TFT. Specifically, Cy conversion at 24 h increased in the order MeCN (27%) < DCE (39%) < TFT (55%). The poorer results for MeCN as solvent may be partly due to its coordinating ability since, according to the above mechanistic considerations, coordinating solvents may compete with the oxidant molecules for coordination to the molybdenum centre.
To gain insights into the electronic features of the complexes, their electrochemical properties were studied (discussed above). No clear correlation could be established between catalytic activity and the electrochemical properties (Figure 3). The type of cation did not considerably affect the electrochemical properties of the anionic complex. Hence, other factors seem to influence the catalytic performance. Focusing on the set of compounds with N-containing cations, specifically those with imidazolium type cations, it seems that the activity increases with increasing carbon chain length of the alkyl substituent, being highest for [Cat] + = [OMIM] + . The compound with [Aliquat] + , which possesses three C8 carbon chains, led to the highest Cy conversion at 24 h (55%). Possibly, for this set of compounds, longer carbon chains favoured the dissolution of the catalyst in the reaction medium, enhancing the overall reaction rate. On the other hand, the transition state formed via coordination of TBHP to the metal centre (to give an active oxidizing species) may be influenced by the presence of the hydronium cation in the best-performing catalyst [H 3 O][MoO(O 2 ) 2 (pic)] (1). For example, the mechanism may involve partial dissociation of the pic ligand and the hydronium cation may influence the protonation of an oxido ligand by TBHP. However, validation of this hypothesis would require more detailed studies, e.g., computational chemistry. partly due to its coordinating ability since, according to the above mechanistic considerations, coordinating solvents may compete with the oxidant molecules for coordination to the molybdenum centre.
Overall, it is not trivial to make clear comparisons of the results between different studies due to the different reaction conditions used. Nevertheless, a rough comparison to literature data for    Catalyst (1) was further explored using H 2 O 2 as oxidant and acetonitrile as solvent (forming a single liquid phase), at 70 • C, which led to slower Cy reaction kinetics, without affecting epoxide selectivity (100% CyO selectivity): conversion at 1 h/3 h/6 h was 18%/30%/47%, compared to 82%/100% conversion at 1 h/3 h using TBHP as oxidant (the initial molar ratio Mo:Cy:oxidant was the same for the two oxidising systems). Possibly, the type of active species may be different for the two oxidants [68,70,71,77] and/or the presence of coordination solvents (water added together with H 2 O 2 and acetonitrile) may lead to competition with the reactants in the coordination to the molybdenum centre. A similar trend was reported previously for Mo-catalysed epoxidation systems [78].
The catalytic system of 1/TBHP was tested with different substrates (cyclododecene and trans-2-octene), at 70 • C ( Table 3). The corresponding epoxides were the main reaction products. The substrate reactivity decreased in the order Cy > trans-2-octene > cyclododecene. The higher reactivity of Cy compared to cyclododecene may be partly associated with steric effects (higher steric hindrance for the latter). Nevertheless, a comparative study for the cyclic versus linear C8 substrates suggests that other factors may also be important, such as electronic ones. A higher energy density of the C=C double bond may favour the attack of the nucleophilic olefin on an electrophilic oxidising species [79], to give the epoxide product. The higher reactivity of Cy than cyclododecene may be related with important steric effects in the reaction of the bulkier olefin, as observed in a previous work of some of us [80]. The epoxide selectivity was excellent (100%) in the case of Cy and very high for the remaining olefins (95-97%) Table 3.  16 Py] + and [Aliquat] + , were prepared in high yields. The CV and DPV analysis indicated that variation of the cation structure does not significantly affect the electrochemical properties of the complex anion. This is the only catalytic study which follows the studies by Garah et al. [31,32,34,60] (1) was also active for the epoxidation of (linear) trans-2-octene and (bulkier) cyclododecene. A comparative study for H 2 O 2 or TBHP as oxidant, indicated that the latter was more favorable, enhancing the reaction kinetics. This catalyst is one of the most active complexes consisting of neutral or ionic [MoO(O 2 ) 2 (L) n ] reported so far. The cation exchange for an N-containing cation, i.e., imidazolium, pyridinium, or ammonium derivatives, does not necessarily lead to an improvement in catalytic performance, which highlights the favourable simplicity of (1).

Conflicts of Interest:
The authors declare no conflict of interest.