Preparation, Characterization, and Catalytic Activity of Tin (Antimony) Substituted and Lacunar Dawson Phosphotungstomolybdates for The Adipic Acid Synthesis

Tin (antimony) substituted and lacunar Dawson phosphotungstomolybdates (1-K10P2W12Mo5O61, 1K8P2W12Mo5SnO61 and -Cs4SnP2W12Mo6O62, and -Cs3SbP2W12Mo6O62) were synthesized and characterized by Fourier Transform Infra Red (FTIR), nuclear magnetic resonance (31P NMR), Visible Ultra Violet (UV-Vis) spectroscopy, and X-ray diffraction (XRD). Their catalytic properties were examined in the oxidation reaction of cyclohexanone at 90 °C and that of cyclohexene at 70 °C to adipic acid (AA), in presence of hydrogen peroxide and in free solvent. The effects of catalyst/substrate molar ratios, hydrogene peroxide flow rate, heteropolysalt composition, and cyclohexanol addition on AA yields were studied. The Cs4SnP2W12Mo6O62 (the most efficient) led to 61 % of AA yield from the cyclohexanone oxidation using a catalyst/substrate molar ratio of 13.3×10-4, H2O2 flow rate of 0.5 mL/h, and a reaction time of 20 h. Copyright © 2019 BCREC Group. All rights reserved

and antimony into this type of material. Dawson-type POMs examined in the AA synthesis have formula of -K6P2W18O62, -K6P2Mo6W12O62, -H6P2W12Mo6O62, 1-K10P2W12Mo5O61, 1-K8P2W12Mo5SnO61, -Cs4SnP2W12Mo6O62, and -Cs3SbP2W12Mo6O62. The AA synthesis was carried out from liquidphase oxidation of cyclohexanone and/or cyclohexanol and cyclohexene in the presence of hydrogen peroxide (30 %) without solvent, acidic additives, and phase transfer agents. The method is based on that proposed by Nomiya et al. [28]. POMs were characterized by FT-IR, 31 P NMR, UV-Vis spectroscopy, and XRD.
The mixed cesium salts, Cs4SnP2W12Mo6O62 and Cs3SbP2W12Mo6O62, were prepared by precipitation from -H6P2W12Mo6, SnCl2 or SbCl3, and CsCl according with the stoichiometric ratios (Equations (1) and (2)). Cs4SnP2W12Mo6 is dark blue and the Cs3SbP2W12Mo6 is green.

Characterization
Infrared spectra were recorded on the 4000-400 cm −1 range on Bruker IFS 66 FT-IR spectrometer using samples prepared as KBr disks. 31 P MAS NMR spectra were measured at room temperature on Bruker Avance 400 spectrometer. The 85 % H3PO4 was used as an external reference. UV-Vis Diffuse Reflectance spectra were recorded in the 800-200 nm region on Specord 210 Plus Analytic Jena spectrometer equipped with a polytetrafluoroethylene (PTFE) integration sphere. PTFE was used as a reference. X-ray Diffraction analysis was obtained on a BRUKER D8 ADVANCE X-ray diffractometer using a Cu-K (k = 1.54178 Å) radiation, in the range of 2θ = 5-60° at a rate of 0.02 °.s -1 .

Catalytic Methods
The synthesis method is based on that described in the literature [27]. The liquid-phase oxidation was carried out at 90 °C in the case of cyclohexanone (-one) and at 70 °C in the case of cyclohexene (-ene), using a 100 mL roundbottomed flask equipped with a magnetic stirring bar and a reflux condenser. The whole is stirred at 1000 rpm for 20 h reaction time. The reaction mixture is constituted by a calculated amount of POM catalyst and substrate. Hydrogen peroxide (30 %) is added drop wise whenever the POM is reduced. The state of the latter is visualized by the presence of a blue color corresponding to Mo(V) atoms. After adding of H2O2, the catalyst shows a color change from blue to yellow, color characteristic of Mo(VI). It should be pointed out that only two Mo(VI) per Keggin anion can undergo a reduction at a time and the resultant homogeneous mixture was cooled at 0 °C overnight. The AA, one of oxidation products, was isolated as white crystals and identified by FT-IR and 1 H-NMR spectroscopy and melting point (~151 °C). The AA yield is given by the following relationship: AA yield (%) = AA recovered mass × 100 / theoretical AA mass. Figure 2 shows FT-IR spectra of the Dawson potassium salts (-P2W18, -P2W12Mo6, 1-P2W12Mo5 and 1-P2W12Mo5Sn), heteropolyacid -H6P2W12Mo6 and cesium salts (Cs4SnP2W12Mo6 and Cs3SbP2W12Mo6). Dawson anion characteristic vibration bands were observed in the low wave number region (500-100  cm -1 ). The metal-oxygen vibration bands corresponding to inter group M-Ob-M and the intragroup M-Oc-M appear at 898-907 cm -1 and 725-749 cm -1 , respectively. The elongation of M=Od band appear at 930-951 cm -1 range. In the Dawson anion, the phosphorus-oxygen vibration band is very sensitive to its environment. An intense vibration band appears around 1073-1083 cm -1 assigned to the junction phosphorus-bimetallic group M2O10 (noted (P-Oa)b) and another around 1011-1017 cm -1 of low intensity attributed to the junction phosphorus-trimetallic group M3O13 (noted (P-Oa)t). FT-IR band observed around 510-520 cm -1 is attributed to (P-O) vibration. These results are in agreement with those of the literature [32,33,35]. In addition, in the cases of 1-P2W12Mo5 and 1-P2W12Mo5Sn, vibration bands were observed at 1117 and 1140 cm -1 , respectively, resulting of the local symmetry decrease attributed to the vacancy presence and to the Sn atom introduction in this vacancy situated in bimetallic group (number 4 as seen in Figure 1).  Figure 3. Only single resonance peak recorded at -9.8, -13.3, -10.2 and -11.3 ppm was observed for -H6P2W12Mo6, -P2W18O62, -P2W12Mo6, and Cs4SnP2W12Mo6, respectively, showing that the two half-anions of the Dawson structure are identical. Therefore, these results evidenced the purity of prepared Dawson POMs. The presence of the vacancy in 1-P2W12Mo5 and the Sn atom insertion in the case of 1-P2W12Mo5Sn, were highlighted by the two signals observed at -7.6 and -10.06 ppm as well as -8.9 and -11.1 ppm, respectively, results according with those observed in FT-IR spectroscopy. For Cs3SbP2W12Mo6, no signal observed, lied probably to the antimony element paramagnetic.

Catalysts Characterization
The UV-Vis spectra of -P2W18 (a), and Cs3SbP2W12Mo6 (f) ( Figure 4) show a large band in 200-500 nm wavelengths domain assigned to oxygen-metal charge transfer (LMCT), corresponding to the oxidation state VI of metal [8,[36][37][38][39][40][41]. In the case of substituted POMs, the introduction of tin and antimony led to the appearance to another LMCT band above 700 nm that can attributed to molybdenum atoms in an oxidation state V. This observation was already reported in the case of substituted Keggin-type POMs [37][38][39][40][41][42]. The intensity of this band increases following the sequence Cs4SnP2W12Mo6 > 1-P2W12Mo5Sn > Cs3SbP2W12Mo6. It was also reported that the Mo(VI) reduced amount increases with the band intensity [37,38]. These observations suggest a partial reduction of POMs confirmed by the observed blue color during their preparation. This suggests that an electron exchange takes place between Sn(II) or Sb(III) and Mo(VI) (Equations (3) and (4)). Figure 5 shows the X-ray patterns of prepared salts. The XR pattern of K6P2W18 is characteristic of a triclinic system, with the following parameters: a = 12.8600 Å, b = 14.8300 Å, c = 22.3400 Å,  = 94.400°,  = 116.870° and  = 115.600° and spacial group P -1 (2), according to the literature data [43][44][45]. The -K6P2W12Mo6O62 (Figure 5b) presents also Xray pattern characteristic of a triclinic system. X-ray diffraction patterns of 1-P2W12Mo5Sn   (Figure 5f) salts that present totally different X-ray patterns which would suggest that these two salts would also crystallize in different unidentified systems in the frame of this work. These results evidenced the influence of the composition of the POM on its crystalline structure.

Catalytic Tests
The catalytic performance of -H6P2W12Mo6, -P2W12Mo6, 1-P2W12Mo5, 1-P2W12Mo5Sn, -Cs4SnP2W12Mo6 and -Cs3SbP2W12Mo6 Dawson type salts were examined in adipic acid synthesis from oxidation of both cyclohexanone at 90 °C and cyclohexene at 70 °C, in the presence of hydrogen peroxide (30 %) and free solvent. It was showed that the substrate oxidation to AA did not take place in absence of catalyst and when the reaction was carried out in one pot (substrate + catalyst + hydrogen peroxide) [12,[20][21][22].
Catalytic tests were repeated three times in the case of cyclohexanone oxidation and twice in the case of the oxidation of alcohol/ketone and that of alkene to verify the reproducibility of the results. As shown by the results in the Table 1, the AA yields vary very little from one test to another, demonstrating the reproducibility of the catalytic test.

Cyclohexanone oxidation
In order to optimize the reaction parameters favoring to AA formation, the effect of catalyst/substrate molar ratio was examined with -P2W12Mo6 catalyst. The cyclohexanone oxidation was carried out at 90 °C, with a fixed flow rate of H2O2 of 1 mL every 150 min, a reaction time of 20 h and a magnetic stirring of 1000 rpm. Table 2 shows an increase of AA yield from 12 to 39 % with the increase of catalyst/cyclohexanone molar ratio (noted ncatalyst/n-one) from 4.0×10 -4 to 13.3×10 -4 . Up to this latter value, AA yield decreases to 32%. For all following catalytic experiments, the catalyst/-one molar ratio will fixed at 13.3×10 -4 , corresponding to 90 mg of catalyst and 15 mmol of substrate.
The effect of H2O2 flow rate on the adipic acid formation from cyclohexanone was exam-  The oxidation reaction was conducted at 90 °C with a catalyst / cyclohexanone molar ratio of 13.3×10 -4 , a reaction time of 20 h and a magnetic stirring of 1000 rpm. Hydrogen peroxide (30%) was added during the reaction via a continual automatic mode using a device with a syringe pump that controls the H2O2 flow rate (0.2-1.0 mL/h) and a manual mode that consists to add H2O2 by fraction of 0.5 or 1 mL/h. For both modes, the injected total H2O2 volume is of 6 mL. Figure 6 shows the obtained results from automatic mode. Two distinct domains were observed, with a maximum of AA yield (52 %), obtained with a flow rate of 0.5 mL/h. From 0.2 to 0.5 mL/h, the AA yield increases from 16 to 52 % and from 0.5 to 1.0 mL/h, it decreases from 52 to 23 %. These results emphasize the hydrogen peroxide flow rate importance on the AA production. So, a low flow rate (<0.5 mL/h), slows down the oxidation reaction of substrate and up to this value (>0.5 mL/h), probably favoured the others oxidation products.
The results of the Table 3 confirms that a H2O2 flow rate of 0.5 mL/h leads to best results regardless the used adding mode. So, when H2O2 flow rate increases from 0.5 to 1 mL/h, AA yield decreases from 61 to 42 % for manually adding and from 52 to 23 % for automatically adding. Moreover, regardless H2O2 flow rate, the manual mode leads to the highest yields. For the following catalytic tests, a volume of 6 mL of H2O2 with a flow rate of 0.5 mL/h and manual addition mode will used.
The time oxidation cyclohexanone effect on AA formation was examined on -P2W12Mo6 and Cs4SnP2W12Mo6. The oxidation reaction was conducted at 90 °C with a catalyst / cyclohexanone molar ratio of 13.3×10 -4 , a magnetic stirring of 1000 rpm and 6 mL of H2O2 (30 %) added with 0.5 mL/h flow rate using manual mode. Table 4 shows that after 10 h of reaction, the AA yields obtained with both POMs are similar (28-29 %), inferior to those obtained after 20 h of reaction (39-61 %) suggesting an active site increase with the reaction time, favoring thus, the AA formation. Whereas, the AA yield increase is more important in the case of Cs4SnP2W12Mo6 (from 28 to 61 %) compared to that observed with -P2W12Mo6 (from 29 to 39 %). These results evidenced the tin action efficiency with the reaction time, on the catalytic performances, lied probably to the presence of redox couples Mo(VI)/Mo(V) and Sn(IV)/Sn(II). This would promote the oxidation either of the substrate or intermediate products to AA. Table 5 shows the AA yield, obtained from cyclohexanone oxidation, as function of the Dawson-type POM composition. The catalytic performances were investigated under the optimized conditions. The proton total substitution of -H6P2Mo6W12 heteropolyacid by the different elements (Cs, K, Sb, Sn) favour the adipic     30-46 %). In this series, Cs4SnP2W12Mo6 exhibits the best catalytic performance with an AA yield of 61 %, evidencing the importance of the role played by the tin as counter-ion, results in agreement with those obtained with the Keggin-type POMs [22]. The obtained AA yields (Table 5) from a quimolar mixture, cyclohexanol (-ol) and cyclohexanone (-one) are sensitive to POM composition. In the presence of -P2W12Mo6 and 1-P2W12Mo5, the alcohol addition to ketone does not seem to influence the formation of AA, thus similar AA yields were obtained, respectively, with 46 and 30% from (-one) oxidation and 47 and 34% from (-ol)/(-one) mixture oxidation. While in the case of 1-P2W12Mo5Sn, Cs4SnP2W12Mo6 and Cs3SbP2W12Mo6, a strong decrease of AA yield from 33 to 17, 61 to 27, and 52 to 37 % was observed, respectively after (-ol) addition to (-one) suggesting that the alcohol inhibits AA formation. This is can be attributed to the hydrogen bonds formation between the C=O group of -one and the hydrogen of C-OH group of -ol, that makes difficult the oxidation of both substrates. Results are in agreement with those observed in literature [12,[18][19][20][21][22].
The AA yield increases with oxidative power increase of POM, following sequence: Cs4 SnP 2 W 12Mo6 > C s 3 Sb P 2 W 12Mo6 > -P2W12Mo6 > 1-P2W12Mo5Sn ~ 1-P2W12Mo5 > -H6P2Mo6W12. Therefore, these results show that the AA formation requires oxidative sites coming from the presence several redox couples as Mo(VI) / Mo(V), Sn(IV) / Sn(II), and Sb(V) / Sb(III) and different peroxo-POM species, resulting of hydrogen peroxide action on reduced POM.

Cyclohexene oxidation
The catalytic performances of POMs were examined in the cyclohexene oxidation to adipic acid in the same operation conditions than those used in the cyclohexanone oxidation. The reaction temperature was fixed at 70 °C, temperature inferior to that of boiling point (83 °C). The results of Table 6 show that the POMs are less active in cyclohexene oxidation to AA compared to those obtained from cyclohexanone oxidation with yields of 11-32 % against 30-61%. 1-P2W12Mo5, and Cs3SbP2W12Mo6 have a similar behaviour with 30 and 32 % of AA yield, superior to those obtained with others catalysts (11-23 %).
Compared to the AA formation from cyclohexanone oxidation that requires of strong oxidative sites, in the case of cyclohexene oxidation, its formation seems to be independent oxidative character of POM. Therefore, a strong oxidative power not favored the cyclohexene oxidation to adipic acid.

Reusability of the catalyst
The catalytic performance of used POM catalyst was also evaluated in order to test its activity as well as its stability. The results are represented in Table 7 Table 7. AA yield (%) obtained from cyclohexanone oxidation with fresh and used POM Reaction parameters: reaction temperature: 90°C, stirring: 1000 rpm; ncatayst/n-one: 13.3×10 -4 and 6 mL of H2O2 (30%) added with 0.5 mL/h flow rate using manual mode lasts 20 h. After recovery of the adipic acid, 15 mmol of cyclohexanone were added to reaction mixture and the oxidation reaction was carried out with the soluble used catalyst, under the optimized conditions. Obtained AA yields after the first run are 46, 61, and 52 % in the presence of -P2W12Mo6, Cs4SnP2W12Mo6, and Cs3SbP2W12Mo6, respectively. When the test was repeated a second one with the same used catalyst, AA was not observed. These results seem to suggest a total deactivation of the catalyst.

Conclusion
In this work, the Dawson structure and purity of salt were confirmed by FT-IR and 31 P NMR spectroscopies, respectively, for potassium salts, -K6P2W18O62, -K6P2W12Mo6O62, -1-K10P2W12Mo5O61, 1-K8P2W12Mo5SnO61 and cesium mixed salts, Cs4SnP2W12Mo6O62, and Cs3SbP2W12Mo6O62. The UV-Vis spectroscopy showed that tin and antimony based heteropolysalts were partially reduced. The XRD results evidenced the effect of Dawson polyoxometalate composition on its crystalline structure.
The operation conditions of the cyclohexanone oxidation in the presence of 30 % hydrogen peroxide toward adipic acid were optimiz ed.
A mo ng , the te sted P OMs, -Cs4SnP2W12Mo6O62 and -Cs3SbP2W12Mo6O62 exhibit the best catalytic performances with 61 and 52 % of adipic acid yield, respectively. In the case of the cyclohexene oxidation, 1-P2W12Mo5 and Cs3SbP2W12Mo6 were found to be the most active with 30-32 % of AA yield. The absence of additives as phase transfer agent, organic solvents and mineral acids, in this process makes the synthesis of adipic acid more environments respectful.