Catalytic Performance of Fe ( II )-Scorpionate Complexes towards Cyclohexane Oxidation in Organic , Ionic Liquid and / or Supercritical CO 2 Media : A Comparative Study

The catalytic activity of the iron(II) C-scorpionate complexes [FeCl2{HC(pz)3}] 1 (pz = pyrazol-1-yl) and [FeCl2{HOCH2C(pz)3}] 2, and of their precursor FeCl2·2H2O 3, towards cyclohexane oxidation with tert-butyl hydroperoxide was evaluated and compared in different media: acetonitrile, ionic liquids (1-butyl-3-methylimidazolium hexafluorophosphate, [bmim][PF6], and 1-butyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate, [bmim][FAP]), supercritical carbon dioxide (scCO2), and scCO2/[bmim][X] (X = PF6 or FAP) mixtures. The use of such alternative solvents led to efficient and selective protocols for the oxidation of cyclohexane. Moreover, tuning the alcohol/ketone selectivity was possible by choosing the suitable solvent.


Introduction
The efficient application of a catalyst, instead of a stoichiometric conversion, as well as the use of supercritical fluids (e.g., carbon dioxide, scCO 2 ) or ionic liquids (IL) as safer solvents for chemical reactions are very important principles of green chemistry [1].The combination of both is, therefore, a promising way to design more sustainable chemical conversions.
Currently, supercritical fluids (SCFs) have not only proven their environmental benefits, but also have offered substantial advantages in several features of catalytic processes [2].Oxidation reactions have been investigated in supercritical carbon dioxide, mostly those of ethanol [3] and toluene [4,5].Selective aerobic oxidation of cyclohexane in scCO 2 has also been attempted [6][7][8], relying on the oxidative stability of carbon dioxide and the miscibility of dioxygen above the critical point to drive free radical chemistry in scCO 2 .Temperature and pressure have been shown to affect the reaction rate and conversion of cyclohexane [6], providing tunability of reactions in SCFs.Despite these encouraging results, homogeneous catalytic oxidations remain relatively unexplored in scCO 2 [9][10][11].
Within our interest on the oxidation of alkanes catalyzed by C-scorpionate complexes [22][23][24], the main objective of the present study was to evaluate the suitability of the above alternative greener media as solvents for such reactions.The low environmental impact and the chemical inertness of scCO2 and ILs under most catalytic process conditions are expected to promote safer reactions in these media.Moreover, IL-scCO2 systems address separation problems, facilitating the recycling and re-use of catalysts.
Within our interest on the oxidation of alkanes catalyzed by C-scorpionate complexes [22][23][24], the main objective of the present study was to evaluate the suitability of the above alternative greener media as solvents for such reactions.The low environmental impact and the chemical inertness of scCO2 and ILs under most catalytic process conditions are expected to promote safer reactions in these media.Moreover, IL-scCO2 systems address separation problems, facilitating the recycling and re-use of catalysts.
Cyclohexanol and cyclohexanone were the only products detected by gas chromatography-mass spectrometry (GC-MS) analysis, revealing the high selectivity of the oxidation systems under the chosen conditions, which was confirmed by the molar balance between the converted substrate and those products.
Complex 1 led to the highest yield and TON values in all tested media (Tables 1 and 3).The yield values achieved in acetonitrile were comparable to those previously obtained at room temperature [28] without any added promotor.Higher yields were reported in very different reaction conditions, such as when 1 was supported on carbon nanotubes or zeolites (total yield up to 38%) [32,33] or using microwave (MW) assistance (total yield of 33%) [12,34], always in the presence of pyrazinecarboxylic acid (Hpca) [32][33][34][35].However, for the present study, to compare the catalysts' performances in different media, promotor-free conditions were preferred.
The FeCl2•2H2O 3 salt exhibited significant lower catalytic activity (Table S1, Supplementary Materials) than complexes 1 and 2 in all media, suggesting the favorable involvement of the scorpionate ligands in the metal-assisted steps [23] of this catalytic oxidation reaction.Moreover, the known radical trap 2,2,6,6-tetramethylpiperidyl-1-oxyl radical (TEMPO) was used in the different media and in all cases a marked decrease (between 80% and 93%) in the products yield was observed, suggesting the involvement of a radical mechanism as proposed in Scheme 2 [36].The involvement of iron (III) species was also suggested by the Fe 2p XPS spectrum of 1 after the fourth recycling run Fe III + H2O2 → HOO • + H + + Fe II (2) CyOOH + Fe III → CyOO • + H + + Fe II (7) Scheme 2. Proposed mechanism for the oxidation of cyclohexane catalyzed by 1-3.The catalytic performances of the C-scorpionate complexes [FeCl 2 {RC(pz) 3 }] (R = H, 1 or CH 2 OH, 2) were compared in terms of products yield, turnover number (TON) or frequency (TOF, h −1 ), and selectivity as a function of reaction time and solvent media (Tables 1-3).
Cyclohexanol and cyclohexanone were the only products detected by gas chromatography-mass spectrometry (GC-MS) analysis, revealing the high selectivity of the oxidation systems under the chosen conditions, which was confirmed by the molar balance between the converted substrate and those products.
Complex 1 led to the highest yield and TON values in all tested media (Tables 1 and 3).The yield values achieved in acetonitrile were comparable to those previously obtained at room temperature [28] without any added promotor.Higher yields were reported in very different reaction conditions, such as when 1 was supported on carbon nanotubes or zeolites (total yield up to 38%) [32,33] or using microwave (MW) assistance (total yield of 33%) [12,34], always in the presence of pyrazinecarboxylic acid (Hpca) [32][33][34][35].However, for the present study, to compare the catalysts' performances in different media, promotor-free conditions were preferred.
The FeCl 2 •2H 2 O 3 salt exhibited significant lower catalytic activity (Table S1, Supplementary Materials) than complexes 1 and 2 in all media, suggesting the favorable involvement of the scorpionate ligands in the metal-assisted steps [23] of this catalytic oxidation reaction.Moreover, the known radical trap 2,2,6,6-tetramethylpiperidyl-1-oxyl radical (TEMPO) was used in the different media and in all cases a marked decrease (between 80% and 93%) in the products yield was observed, suggesting the involvement of a radical mechanism as proposed in Scheme 2 [36].The involvement of iron (III) species was also suggested by the Fe 2p XPS spectrum of 1 after the fourth recycling run (see below) in [bmim][PF 6 ] IL.
The cyclohexanol and cyclohexanone yields obtained in this work (under promotor-free conditions) were considerably higher than those provided by the current industrial process that uses a homogeneous cobalt species as a catalyst, dioxygen as an oxidant, needs a considerably high temperature (150 • C), and is only able to generate 5-12% yields to assure a selectivity of ca.80-85% [26].[38] and, as expected, mass transfer is higher in the less viscous IL, promoting a higher yield of cyclohexane oxidation products.
The following order is observed for the activity of the above C-scorpionate Fe(II) catalysts under promoter-free conditions in the solvents used: ) 2:1 molar mixtures were also tested as a reaction medium for 1 and 2 (Table 3), and were found to improve the solubility of the scorpionate complexes.A significant increase in the products yield was found in the scCO 2 -[bmim][FAP] mixture, for both catalysts, relative to the values obtained in [bmim][FAP] or scCO 2 (compare e.g., entry 3 of Table 3 with entries 9 and 12 of Table 1).The dissolution of CO 2 in the IL reduces its viscosity and enhances the mass transfer of the IL, promoting higher yields.Moreover, in scCO 2 -IL systems, oxygenates were separated without contamination of catalyst or solvents, making them good candidates for a continuous flow strategy able to further improve the efficiency and industrial potential of these systems.
Cyclohexanone was the main product of the cyclohexane oxidation in all tested media.Regarding pure solvents, the highest selectivity towards cyclohexanone was achieved in [bmim][FAP].The possibility of tuning the selectivity towards alcohol or ketone formation by choosing the appropriate IL was previously found [12] for 1 in [bmim][N(CN) 2 ] and [bmim][BF 4 ] and was now extended to the media of the present study.
In scCO 2 -IL, the selectivity towards cyclohexanone increased significantly (up to 96% in scCO 2 -[bmim][PF 6 ] for 2, Table 3, entry 2) when compared with that obtained in pure solvents (maximum of 77% in [bmim][FAP], Table 2, entry 9).Such a feature can be of applied relevance.In fact, the current production process of polyamide 6 involves the conversion of cyclohexanol and cyclohexanone mixture obtained from cyclohexane oxidation to pure cyclohexanone (at high pressure, 250 • C, in the presence of copper and chromium oxides), which is then used to form caprolactam, the monomer of polyamide 6 [26,27].
The influence of scCO 2 density on the selectivity for oxygenated products from the oxidation of cyclohexane has been reported previously [37].A higher density leads to a higher selectivity, since the formed oxygenates are believed to be removed more easily from the catalyst surface without being totally oxidized due to the higher solvent power of the denser CO 2 in the supercritical phase [39].
The attempts to recycle catalysts 1 or 2 in acetonitrile or scCO 2 were not successful.However, recycling of these catalysts was performed up to five cycles in each pure ionic liquid medium or scCO 2 -IL mixed solvents (Figure 3).
The attempts to recycle catalysts 1 or 2 in acetonitrile or scCO2 were not successful.However, recycling of these catalysts was performed up to five cycles in each pure ionic liquid medium or scCO2-IL mixed solvents (Figure 3).While the catalytic activity of compound 2 is well preserved in all (pure or mixture) tested solvents at least for the first three consecutive cycles (compare Figure 3b-d), catalyst 1 is better preserved in pure ILs than in the scCO2-IL mixed solvents (compare Figure 3a-c).The decrease in the catalyst activity after the third cycle is probably associated to its decomposition or deactivation.A deeper understanding of how the ILs and SCFs properties can affect the performance of the catalytic systems, in particular their recyclability, will be addressed in future studies.

Materials and Methods
All the reagents and solvents were purchased from Sigma-Aldrich, Munich, Germany, and used as received.The scorpionate complexes 1 [28] and 2 [29] were synthesized and characterized according to methods described in the literature.Liquid CO2, purity of 99.995%, in a cylinder with a dip tube was supplied by Air Liquid and used without further purification.[bmim][PF6] was synthesized according to a published procedure [40], dried and degassed under vacuum at 70 °C for 48 h prior to use.[bmim][FAP] was kindly given by Merck Chemicals and no further treatment was performed.While the catalytic activity of compound 2 is well preserved in all (pure or mixture) tested solvents at least for the first three consecutive cycles (compare Figure 3b-d), catalyst 1 is better preserved in pure ILs than in the scCO 2 -IL mixed solvents (compare Figure 3a-c).The decrease in the catalyst activity after the third cycle is probably associated to its decomposition or deactivation.A deeper understanding of how the ILs and SCFs properties can affect the performance of the catalytic systems, in particular their recyclability, will be addressed in future studies.

Materials and Methods
All the reagents and solvents were purchased from Sigma-Aldrich, Munich, Germany, and used as received.The scorpionate complexes 1 [28] and 2 [29] were synthesized and characterized according to methods described in the literature.Liquid CO 2 , purity of 99.995%, in a cylinder with a dip tube was supplied by Air Liquid and used without further purification.[bmim][PF 6 ] was synthesized according to a published procedure [40], dried and degassed under vacuum at 70 • C for 48 h prior to use.[bmim][FAP] was kindly given by Merck Chemicals and no further treatment was performed.

Procedure for the Catalytic Oxidations Using a Molecular or an Ionic Liquid Solvent
The cyclohexane oxidation using [bmim][X] (X = PF 6 or FAP) or CH 3 CN as a solvent (up to 5 mL total volume) was carried out in air, in a round bottom flask, with vigorous stirring.The catalyst was added to the solvent as a solid or in the form of a stock solution in the reaction medium.Cyclohexane (5 mmol) was then introduced, and the reaction started when TBHP (10 mmol) was added in one portion.The reaction was stopped after the desired time and 5 mL of diethyl ether were added for the extraction of the organic products.
The reaction mixtures were analyzed by gas chromatography (GC) upon the addition of an excess of triphenylphosphine to reduce cyclohexyl hydroperoxide to cyclohexanol, following a method developed by Shul'pin [41].Nitromethane (0.05 mL) was used as an internal standard.Chromatographic measurements were performed in a Fisons Instruments GC 8000 series gas chromatograph with a BP20/SGE (30 m × 0.22 mm × 0.25 mm) capillary column (flame ionization detector (FID) detector) and Jasco-Borwin v.1.50software (Jasco, Toquio, Japan) using helium as a carrier gas.All products formed were identified and their retention times confirmed with those of commercially available samples.A Perkin-Elmer Clarus 600 gas chromatograph, equipped with two capillary columns (SGE BPX5; 30 m × 0.32 mm × 25 mm), one with an electron impact-mass spectrometry EI-MS detector and the other one with an FID detector, was also used for analyzing the reaction mixtures.Helium was used as the carrier gas.Blank experiments in CH 3 CN and ILs were performed and it was confirmed that no cyclohexane oxidation products (or only traces, below 1%) were obtained in the absence of the iron catalyst.

Procedure for the Catalytic Oxidations in scCO 2 or in scCO 2 -IL
The oxidation reaction was carried out in batch mode using a supercritical fluid apparatus as previously described [42], placed inside a fume hood for safety reasons.
Oxidation reactions were carried out in batch mode using a stainless steel reactor (16 mL, 316 stainless steel hollow cylinder, custom-fabricated with a mechanic stirrer).The reactor was charged with the catalyst, cyclohexane, and the oxidant.The system was then brought up to the desired CO 2 pressure and temperature.The supercritical temperature and pressure were measured with a platinum resistance thermometer and a digital pressure transducer, respectively.The reaction mixtures were stirred for 1-10 h at 70 • C and at CO 2 pressure up to 90 bar.Products were collected by trapping the reaction in cold ethyl ether and carefully vented to the atmospheric pressure through a cold trap cooled with liquid nitrogen.Products were analyzed by GC upon the addition of an excess of triphenylphosphine (see above) [41].Blank experiments in scCO 2 were performed and it was confirmed that no cyclohexane oxidation products (or only traces, below 1%) were obtained in the absence of the iron catalyst.

Catalyst Recycling
Catalyst recyclability was investigated in all media, but only the ILs and scCO 2 -IL media led to successful results.Each new cycle was initiated after the preceding one upon the addition of new typical (see above) portions of all other reagents.After the completion of each run, the ionic liquid with the catalyst was recovered by drying in vacuum overnight at 80 • C and the products were analyzed as previously described.

Table 3 .
Selected data a for the cyclohexane oxidation with TBHP in scCO 2 -IL medium.