H3PMo12O40 Immobilized on Amine Functionalized SBA-15 as a Catalyst for Aldose Epimerization

In this work various amount of phosphomolybdic acid (PMo) were immobilized on amine functionalized SBA-15 and used as heterogeneous catalysts in the epimerization of glucose in aqueous solution. 13.3PMo/NH2-SBA-15 exhibited the best catalytic performance with a glucose conversion of 34.8% and mannose selectivity of 85.6% within two hours at 120 °C. The activation energy of 80.1 ± 0.1 kJ·mol−1 was lower than that of 96 kJ·mol−1 over the homogeneous H3PMo12O40 catalyst. The catalytic activities of 13.3PMo/NH2-SBA-15 for the transformation of some other aldoses including mannose, arabinose and xylose were also investigated.


Introduction
As a renewable carbon resource, biomass is considered to be an ideal substitute for traditional fossil resources [1,2]. Carbohydrates are easily available in the utilization of plant biomass and glucose obviously prevails as the main building block [3,4]. As the most abundant C6 monosaccharide in nature, glucose can be easily extracted in large volumes from cellulose and hemicellulose, so many efforts have been put into valorizing glucose [5]. Among the various chemical transformations of glucose, epimerization is a carbon-efficient pathway for the production of chiral counterparts (Scheme 1) [6][7][8]. Epimerization is widely used to produce some rare sugars with valuable properties since these sugars could not be largely obtained from biopolymers. For example, sugars like D-lyxose and L-ribose, which are produced based on their abundant C-2 epimers, can be employed to synthesize antitumor agents and hepatitis B virus resistance drugs, respectively [9,10]. Currently, the large-scale production of these rare monosaccharides mainly relies on enzyme catalytic process, which suffers from the strict operating conditions such as specific temperature and pH value [11]. In addition, epimerases are only active on the monosaccharides pretreated by phosphate or nucleotide and difficult to be separated

Preparation of Catalysts
Preparation of mesoporous SBA-15 was carried out according to the literature method [32]. Typically, P123 (4.0 g), deionized water (30 g) and 2.9 mol·L −1 HCl (120 g) were added into a round bottom flask, and the mixture was kept stirring for 4 h at 35 • C. TEOS (8.5 g) was added into the system, then the mixture was kept stirring for 20 h. The resulting mixture was transferred into a Teflon-lined autoclave and was then aged at 100 • C for 24 h. After that, the white product was filtered, washed three times with deionized water and dried at 60 • C overnight. Mesoporous SBA-15 was obtained after the dried sample was calcined at 550 • C for 5 h in a muffle furnace with 1 • C·min −1 of the heating rate.
The amine functionalized SBA-15 was prepared via a post-grafting method following the reported procedure [33]. In a typical process, SBA-15 (3.0 g) was dried overnight at 110 • C under vacuum and quickly dispersed in toluene (75 mL). The mixture was refluxed at 120 • C for 4 h to remove residual water molecules in the system. APTES (1.5 g) was dissolved in toluene (15 mL) and the solution was added dropwise to the suspension with stirring for another 4 h at 120 • C. After cooling to room temperature, the product was filtered, washed with warm ethanol three times and then dried at 60 • C. The resulting material was denoted as NH 2 -SBA-15.
The immobilization of PMo on the surface of NH 2 -SBA-15 was performed as follows. Different amount of PMo (10,20,30,40, and 50 mg) were dissolved in deionized water (30 mL) to obtain a clear solution and then NH 2 -SBA-15 (0.3 g) was dispersed in the solution. The mixture was stirred at room temperature for 24 h. After centrifugation, washing three times with water and drying at 60 • C overnight, xPMo/NH 2 -SBA-15 catalyst was obtained, where x represents the weight percentage of PMo (3.3, 6.7, 10, 13.3, and 16.7 wt%) anchored on NH 2 -SBA-15 in the experimental preparation process.

Catalyst Characterization
Fourier transform infrared (FT-IR) spectra were measured on a ThermoFisher Nicolet iS10 infrared instrument (Waltham, MA, USA) using KBr discs. X-ray powder diffraction (XRD) patterns were recorded on a Bruker D8 Advance X-ray diffractometer (Karlsruhe, Germany) using Cu-Kα radiation with a voltage of 40 kV and a current of 40 mA. The nitrogen content of catalysts was determined with a Vario EL elemental analyzer (Analysemsysteme GmbH, Langenselbold, Germany) and the molybdenum content was measured on a Perkin Elmer 8000 inductively coupled plasma atomic emission spectrometer (ICP-AES, Waltham, MA, USA). N 2 adsorption-desorption measurements were performed at 77 K on a Micromeritic Tristar II 3020 apparatus (Norcross, GA, USA). SBA-15 was outgassed at 200 • C for 3 h and the other samples were outgassed at 100 • C for 3 h before the measurements. The specific surface area was calculated with the BET equation. The pore size and pore volume were calculated from the desorption branch of the isotherm based on the BJH model. 31 P magic-angle spinning nuclear magnetic resonance (MAS NMR) studies were carried out on a Bruker 400 WB AVANCE III solid-state nuclear magnetic resonance instrument (Karlsruhe, Germany), and the chemical shift was determined using 85 wt% H 3 PO 4 as a standard.

Catalytic Tests
The catalytic reaction of glucose epimerization was conducted as follows: catalyst (40 mg) and 5 mL of 5 wt% aqueous solution of aldose (glucose, mannose, arabinose or xylose) were added into a 50 mL autoclave and the reactor was placed in a 120 • C oil bath with magnetic stirring. After reaction with a desired time, the reactor was cooled to room temperature with flowing cold water and the solid catalyst was separated by a 0.22 µm filter. Then the reaction solution was diluted 10 times with deionized water and analyzed on an Agilent 1100 high-performance liquid chromatography (HPLC, Santa Clara, CA, USA) with a refractive index detector. A Biorad©Aminex HPX-87 sugar column was employed at 35 • C and 5 mmol·L −1 H 2 SO 4 solution was used as the mobile phase at a flow rate of 0.6 mL·min −1 . The conversion of aldose, the yield and selectivity of the corresponding epimer were calculated as follows: Aldose conversion = moles of converted aldose moles of initial aldose Epimer yield = moles of formed epimer moles of initial aldose Epimer selectivity = moles of formed epimer moles of converted aldose Figure 1 shows the FT-IR spectra of SBA-15, NH 2 -SBA-15 and xPMo/NH 2 -SBA-15 catalysts. A broad band ranging from 3000 to 3600 cm −1 and a band around 1620 cm −1 observed for all samples is attributed to the vibration of O-H of physisorbed water [34,35]. Compared with SBA-15, NH 2 -SBA-15 had two new peaks at 2950 and 1520 cm −1 , which were assigned to the stretching mode of C-H and bending mode of N-H, respectively, indicating that the aminopropyl groups were successfully grafted on the surface of SBA-15 [34,36]. The decreased intensity of Si-OH band around 965 cm −1 also implies the successful amine functionalization [37]. As for SBA-15 or NH 2 -SBA-15, the bands at 1080 and 798 cm −1 were attributed to the symmetric and asymmetric stretching vibration of Si-O-Si, respectively [38]. The typical Keggin structure of PMo was identified by four characteristic bands in the range of 1300-600 cm −1 . The band at 1064 cm −1 was attributed to the asymmetric stretching vibration of P-O t , and the bands at 965, 870, and 784 cm −1 were attributed to the stretching modes of terminal Mo=O t , edge sharing of Mo-O b -Mo, and corner sharing of Mo-O c -Mo, respectively [39].

XRD
The small-angle XRD patterns of SBA-15, NH 2 -SBA-15 and xPMo/NH 2 -SBA-15 catalysts are depicted in Figure 2A. Three diffraction peaks, one strong peak around 1 • attributed to (100) plane and two weak peaks around 1.7 • and 1.9 • attributed to (110) and (200) planes, respectively, are shown for all the samples [32]. It suggests that the primary two-dimensional hexagonal structure of SBA-15 was maintained after the modification with APTES and the further immobilization of PMo. Compared with SBA-15 and NH 2 -SBA-15, no diffraction peak of PMo was detected in the wide-angle XRD patterns of xPMo/NH 2 -SBA-15, indicating the high dispersion of PMo species on the support [36].  Figure S1 and the textural properties are listed in Table 1. All the samples had typical type IV isotherms and H1 type hysteresis loops, indicating the uniform mesoporous structures [40]. After the modification with APTES, the BET surface area decreased from 599 to 343 m 2 ·g −1 , and the pore volume declined from 0.80 to 0.54 cm 3 ·g −1 . The pore size calculated from the isotherm also reduced from 5.5 to 5.1 nm. When PMo was immobilized on NH 2 -SBA-15, the surface area showed a gradually decreased trend with increased amount of PMo. A similar phenomenon was observed for pore volume. However, the introduction of PMo had little effect on the pore size due to low loading of phosphomolybdic acid.

31 P MAS NMR Spectroscopy
As displayed in Figure 3, 31 P MAS NMR spectrum of HPMo showed a strong resonance at −5.7 ppm and a weak resonance at −5.3 ppm. In the case of 13.3PMo/NH 2 -SBA-15 catalyst, there was a single resonance at 0.1 ppm, further confirming the successful immobilization of PMo species. The broadening of NMR peak and shifting to down-field of 0.1 ppm were related to the decrease of physisorbed and structural water during immobilization [41]. Also, the strong interaction between PMo and NH 2 -SBA-15 may also have been responsible for the shift of NMR peak [42].

Catalytic Activity of xPMo/NH 2 -SBA-15 for Glucose Epimerization
The catalytic activity of various catalysts was investigated in the glucose epimerization. As shown in Table 2, no glucose conversion could be found in the blank experiment without catalyst or only with the bare SBA-15. When NH 2 -SBA-15 was used as a catalyst, the conversion of glucose was 8.9%, and the selectivity of fructose and mannose were 38.2% and 5.6%, respectively. According to the research of Carraher et al. [43], OHgenerated from the hydrolysis of -NH 2 group can simultaneously catalyze the isomerization and epimerization of glucose, with the isomeric fructose being the main product [43]. As a result, a little fructose was observed when NH 2 -SBA-15 was used as the catalyst. In addition, a large number of side reactions existed in the base catalyst system, and these products could not be precisely analyzed by HPLC. The color of the solid catalysts changed from white to dark brown after the reaction, indicating that Maillard reaction occurred in the system [44]. When the PMo polyanions were introduced to NH 2 -SBA-15 via the neutralization of -NH 2 groups with H + of HPMo during the preparation process [45], no fructose was observed for the xPMo/NH 2 -SBA-15 catalysts. The glucose conversion increased from 6.8% over 3.3PMo/NH 2 -SBA-15 to 34.8% over 13.3PMo/NH 2 -SBA-15 and the selectivity of mannose increased dramatically from 29.4% to 85.6%. The mannose yield of 29.8% is close to the theoretical equilibrium yield reported in the literature [28,46]. The glucose conversion continuously increased with the decrease of mannose selectivity over 16.7PMo/NH 2 -SBA-15, attributing to the enhanced side reactions. These results reveal that the PMo is of great significance for the epimerization of glucose. The catalytic performance of 13.3PMo/NH 2 -SBA-15 was superior to that of Sn-β zeolite or porous tin-organic frameworks and comparable to that of layered niobium molybdates in the aqueous reaction system [16][17][18]28,47].

Calculation of the Activation Energy
Experiments with controlled reaction temperature were conducted to determine the activation energy (Table S1 in the Supplementary Information). Based on the assumption that the epimerization of glucose is a first order reaction, the apparent activation energy (E a ) of the reaction over 13.3PMo/NH 2 -SBA-15 was calculated to be 80.1 ± 0.1 kJ·mol −1 (Figure 4), which is lower than the 96 kJ·mol −1 over homogeneous PMo catalyst [29]. Rellan-Pineiro et al. [48] reported that the reducibility of Mo atoms on the surface Mo-based catalysts plays an important role in the process of 1,2 carbon shift. The strong interaction between PMo species and the support in 13.3PMo/NH 2 -SBA-15 may influence the reducibility of the Mo atom, so that the glucose molecule is more easily be transformed into a transition state. 13.3PMo/NH 2 -SBA-15 catalyst may not only catalyze the epimerization reaction of glucose, but also catalyze its reverse reaction. Figure 5 shows the time courses of mannose concentration in epimerization reaction of glucose and its reverse reaction of mannose epimerization. It is clearly shown that mannose concentration in the glucose epimerization reached theoretical equilibrium at a shorter time than that in the reverse reaction. This phenomenon is consistent with the research of Ju et al. [29] over HPMo catalyst. Hayes et al. [49] reported that some unreactive complexes could be formed in the system of mannose, which did not epimerize to glucose.

Reusability of the Catalyst for Glucose Epimerization
In order to investigate the reusability of the catalyst, 13.3PMo/NH 2 -SBA-15 catalyst with the best catalytic activity was chosen for recovery experiments. After each reaction, the catalyst was separated by centrifugation, washed three times with water and dried at 60 • C, and then used again for the glucose epimerization. As shown in Figure 6, the 13.3PMo/NH 2 -SBA-15 exhibited no significant decrease of the catalytic activity after three cycles in glucose epimerization. As anticipated, no apparent leaching of PMo species (<0.1%) occurred during the reaction as confirmed by ICP-AES tests, which is much less than that in the reported system catalyzed by Ag 3 PMo 12 O 40 . These results may be related to the strong interaction between PMo and NH 2 -SBA-15. Compared with insoluble salts of PMo, immobilization of the PMo on amine functionalized support may have been more beneficial for heterogenization of HPMo in the epimerization of glucose.

Catalytic Activity of 13.3PMo/NH 2 -SBA-15 for Other Aldoses Epimerization
We also investigated the catalytic activity of the 13.3PMo/NH 2 -SBA-15 catalyst for the epimerization of other aldoses at the C-2 position (Scheme 2). As shown in Table 3, the arabinose conversion and ribose yield were 27.7% and 15.9%, respectively. The xylose conversion and the lyxose yield were 44.6% and 33.3%, respectively. Similar to glucose epimerization, these two reactions were also limited by thermodynamics, with arabinose/ribose and xylose/lyxose theoretical equilibrium ratios being 69:31, 67:33, respectively, according to the Gibbs free energy calculations [28]. The yield of lyxose was close to the equilibrium yield, but the selectivities of the epimeric aldoses in the both reaction systems were lower than mannose due to the presence of more side reactions. As a result, besides C6 aldoses, the C5 aldoses, like arabinose and xylose could also be effectively transformed into their C-2 epimers over the 13.3PMo/NH 2 -SBA-15 catalyst.

Conclusions
A series of xPMo/NH 2 -SBA-15 heterogeneous catalysts were prepared and tested in the glucose epimerization. The reaction over 13.3PMo/NH 2 -SBA-15 reached almost equilibrium within 2 h with a glucose conversion of 34.8% and mannose yield of 29.8%. The activation energy of the reaction over 13.3PMo/NH 2 -SBA-15 was lower than that over the homogeneous HPMo catalyst. The leaching of PMo was negligible in the reaction system. Mannose, arabinose and xylose could also be transformed to their corresponding C-2 epimeric aldoses successfully over the catalyst. Our work provides a promising heterogeneous catalyst for the epimerization of aldoses in aqueous solution.

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