Nano-sized mesoporous phosphated tin oxide as an efficient solid acid catalyst

Herein, we prepared a mesoporous tin oxide catalyst (mSnO2) activated with phosphate species by the adsorption of phosphate ions from a phosphoric acid solution onto tin oxyhydroxide (Sn(OH)4) surface. The phosphate content ranged from 3 to 45 wt%. The nonaqueous titration of n-butylamine in acetonitrile was used to determine the total surface acidity level. FTIR of chemically adsorbed pyridine was used to differentiate between the Lewis and Brönsted acid sites. Thermal and X-ray diffraction analysis indicated that the addition of phosphate groups stabilized the mesostructure of mSnO2 and enabled it to keep its crystalline size at the nanoscale. FTIR analysis indicated the polymerization of the HPO42− groups into P2O74−, which in turn reacts with SnO2 to form a SnP2O7 layer, which stabilizes the mesoporous structure of SnO2. The acidity measurements showed that the phosphate species are distributed homogeneously over the mSnO2 surface until surface saturation coverage at 25 wt% PO43−, at which point the acid strength and surface acidity level are maximized. The catalytic activity was tested for the synthesis of hydroquinone diacetate, where it was found that the % yield of hydroquinone diacetate compound increased gradually with the increase in PO43− loading on mSnO2 until it reached a maximum value of 93.2% for the 25% PO43−/mSnO2 catalyst with 100% selectivity and excellent reusability for three consecutive runs with no loss in activity.


Introduction
The replacement of liquid acid or acid halide catalysts, such as H 2 SO 4 , H 3 PO 4 , HF and AlCl 3 , is a critical factor for the design of cleaner processes for green catalytic action. 1 The problems of homogeneous liquid acid catalysis include toxicity, corrosion, catalyst waste and difficulty in separation and recovery. 2 The economics of the separation cost majorly determine the viability of such a process. 3 Efforts have been made to design environmentally safe processes and technologies to achieve these goals. The best catalytic technology for solving the problems of the liquid acid catalyst is to replace it by a heterogeneous solid acid catalyst. 4 The reuse and recycling of the catalyst in heterogeneous processes are the main advantages of the increased catalyst efficiency over homogeneous ones. 5 The presence of an interconnected network of a large number of pores, acid sites of different strength and the hydrophobic properties of solid catalyst surfaces support inhibiting the processes of corrosion and separation and for limiting pollution problems. 2 Oil rening and pharmaceutical industries utilize solid catalysts with acidic properties. 6 , Cl À and F À supported on metal oxides have been used to increase their thermal stability, mesoporosity, and more critically, the acidity level of such catalysts. [9][10][11][12] Most of the published works to date have been focused on SO 4 2À -and WO 4 2À -supported metal oxide systems, [13][14][15][16][17][18] which have been found to be suitable for specic acid-catalyzed reactions. The acidity of such systems strongly depend on the preparation method. 11,14,16 A template pathway for preparing mesoporous materials is a novel process to produce materials with high activity for the synthesis of industrially relevant chemicals. [19][20][21][22] Mesoporous materials are characterized by a high surface area, ordered pore channels, and large pore volume value. These advantages can facilitate active components achieving a highly dispersed state on the support, which raises their activity for reacting molecules. 23,24 For better performance, an efficient heterogeneous catalyst needs a higher acid strength level and density as well as thermal stability to minimize catalyst poisoning and leaching. Therefore, mesoporous materials, primarily metal oxides, have attracted growing interest as a support in the eld of catalysis. 25 Tin oxide is one of the most attractive functional materials because of its potential applications, mainly as a catalyst and as a carrier in supported catalysts. 26 Besides its intrinsic nonstoichiometry and crystal defects, tin oxide nanoparticles can be used more efficiently as a catalyst due to the faster migration of the oxide ions within the SnO 2 nanoparticles to increase the surface to volume ratio. Many authors have used high valence cations of metal oxide to promote the thermal stability and catalytic property of pure tin oxide. 27,28 Therefore, it is of interest to study the novelties of phosphate-supported SnO 2 as a solid acid catalyst. Various chemical routes have been used to synthesize nanoparticles of mesoporous metal oxides, e.g., precipitation, sol-gel, hydrothermal, microwave-assisted syntheses and ultrasonic spray pyrolysis. 29,30 The addition of anion groups to a mesostructure metal oxide during the preparation methods prevent the sintering process and thus lead to improvements in their thermal stability and textural properties. 31,32 Further improvement is necessary, however, for the preparation of thermally stable mesoporous tin oxide. This process is still a challenge to researchers. 9,33 Unlike sulfated metal oxides, phosphated metal oxides have received inadequate attention. Albeit, phosphate-based solid acids have proven quite efficacious in several industrially important acid-demanding reactions.
The present study aimed to adopt a phosphate-modied mesoporous tin oxide (mSnO 2 ) system as a catalyst. The modi-ed catalytic systems (H 3 PO 4 /m-SnO 2 ) were then applied to the preparation of hydroquinone diacetate, which is widely used as an ingredient for skin bleaching cream. DTA was used to systematically elucidate the physico-chemical properties of the H 3 PO 4 /mSnO 2 catalysts, while TGA, XRD, TEM and FTIR techniques were applied for characterization, and the texture properties were determined from N 2 adsorption at À196 C. The surface acidity was also examined.

Results and discussion
The thermograms of DTA and the TGA of dried PO 4 3À /Sn(OH) 4 samples are shown in Fig. 1. The sample showed endothermic peaks at low temperature, which are assigned to the evolution of physically adsorbed water. The sample also showed two exothermic peaks. The rst exothermic effect with a weight loss of 8.9 wt% at around 312 C was attributed to the decomposition of the surfactant template and partial removal of the OH groups (dehydroxylation), 34,35 while the second exothermic peak at 423 C corresponded to a weight loss of 3.1 wt%, which may be due to the crystallization of tin oxyhydroxide into tetragonal cassiterite SnO 2 . 36 In the case of PO 4 3À /Sn(OH) 4 dried samples, the DTA in Fig. 2 shows exothermic peaks at 321 C, 334 C and 342 C as the phosphate content increases for the 8%, 25% and 45% PO 4 3À /Sn(OH) 4 samples, respectively. Also, the exothermic peaks of the crystallization of tin oxyhydroxide into mSnO 2 were also shied to higher temperatures as the phosphate content increased, reaching 427 C, 484 C and 504 C for the 8%, 25% and 45% PO 4 3À /mSnO 2 samples, respectively, indicating that the addition of phosphate to mSnO 2 may enhance the surfactant-tin oxyhydroxide surface interaction, 37 thus increasing the thermal stability of the tin oxyhydroxide phase and shiing its crystallization to higher temperature. These results show no effect that could be assigned to the decomposition of PO 4 3À species, which was attributed to the thermal stability when treated with phosphoric acid. 38 Fig. 2S The degree of crystallization of cassiterite mSnO 2 gradually decreases with the increase in the phosphate content, while the width of the reections is considerably broadened, indicating a small crystalline domain size ( Table 1). The addition of phosphate anions to the template synthesis reaction medium of mSn(HO) 4 , substantially slows down the crystallization process via retarding the crystal size growth by occupying the defects on the surface of mSn(HO) 4 , 43 thus increasing its thermal stability and thus protecting the mesoporous structure from collapsing. 41 The crystallite size data were determined from Xray diffraction and are listed in Table 1, showing that the crystallite size of PO 4 3À /mSnO 2 is smaller in comparison to the pure mSnO 2 nanocrystalline (13.70-11.30 nm). Furthermore, the crystallite size of 3% PO 4 3À /mSnO 2 increases from 12.17 to 16.66 nm with the increase in calcination temperature (400-650 C), which indicates that pure mSnO 2 and PO 4 3À /mSnO 2 are present in the nanoscale. In the phosphate sample annealed between 400 C and 650 C, the crystallite size slightly changes from 13.70 to 16.66 nm, suggesting that the PO 4 species remain bonded to the mSnO 2 surface, stabilizing it against sintering. It was reported that the order of the mesostructured materials strongly depends on the texture properties of the pores. 44 An analogous phenomenon was observed with TiO 2 , ZrO 2 and Fe 2 O 3 . 44,45 TEM images of mSnO 2 show an ordered structure with a uniform mesoporous arrangement into hexagons (Fig. 4a). It is apparent that mSnO 2 has a signicant void fraction, which may be due to the mesoporous structure of SnO 2 , and concomitantly a rather low density. Fig. 4c Fig. 5A and B. The band at 1625 cm À1 appearing in the   spectra of the samples even aer calcination at different temperatures corresponds to the OH bending vibrations. The FTIR spectra of phosphate mSnO 2 generally show absorption bands at 1035 and 1150, 1360, 1410 and 1520 cm À1 . The absorption bands at 1360, 1410 and 1520 cm À1 could be assigned to P-O stretching frequency bonds in which their order is close to two P]O, phosphoryl groups. 48 By increasing the temperature to 650 C, the broadband at 1550-1300 cm À1 is reduced to one sharp band and one shoulder, due to the formation of bidentate-bound phosphate ions (C 2 V point group). The band at 1035 and the shoulder at 1150 cm À1 are frequencies for PO 4 3À and the position of such a fundamental band was observed in phosphated zirconia. 49 The fading of the shoulder and the presence of a broad and sharp band located at 1055 cm À1 when the PO 4 3À content was increased to 45 wt% may be due to the dropping of the symmetry in the free PO 4 3À (Td point group) to either C 3 V or C 2 V. 48 This may be due to the PO 4 3À groups being bound to the tin oxide surface and being distributed homogeneously. 50,51 Here, the 1055 cm À1 band is divided into two bands at 1035 and 1150 cm À1 at 650 C, which could be assigned to the bidentate bound phosphate ions (C 2 V point group), which motivate the polymerization of the HPO 4 2À groups into P 2 O 7 4À , which is in a good agreement with the appearance of a new peak due to SnP 2 O 7 in the X-ray pattern (Fig. 3A).
These results indicate that phosphate groups are chemically adsorbed on the surface of mSnO 2 and still exist on the surface even aer calcination of the samples at high temperatures. The mechanism, as explained in Scheme 1, follows the dehydroxylation reaction of P/mSnO 2 samples, with amorphous Sn(HPO 4 ) 2 formed on the surface, which is further heated to form tin pyrophosphate, SnP 2 O 7 , which is in a good agreement with the XRD pattern at 650 C. Fig. 6 shows the nitrogen adsorption-desorption isotherms of some representative samples of PO 4 3À /mSnO 2 , in which the isotherms exhibit a type IV feature, which is a characteristic feature of mesoporous materials. 52 Barrett-Joyner-Halenda (BJH) analysis demonstrated that the average pore diameter (d pore,BJH ) of mSnO 2 calcined at 550 C was 13.21 nm ( , which stabilize the defects and inhibit the collapsing of the mSnO 2 surface, 53 which also explains the role of PO 4 3À in enhancing the porous structure of PO 4 3À /mSnO 2 . The increase in PO 4 3À content beyond 25 wt% led to the production of more phosphate layers and/or the phosphate groups attacking the pore mouths blocking them, and consequently the surface area values decreased. 54,55 The above results suggest that the interaction between PO 4 3À and mSnO 2 crystallites not only stabilizes the crystallite structure in lower dimensions, but also it keeps phosphate groups at the SnO 2 surfaces, inhibiting the crystallite aggregation and acting as a structure progeny director mediating nanoparticle growth and assembly and stabilizing the mesostructure walls and   increasing the thermal stability. 56 Again, the increase in surface area and pore volume and the decrease in the average pore diameter of the 3% to 25% PO 4 3À /mSnO 2 samples over pure mSnO 2 with increasing the phosphate content indicate a relatively high thermal stability of the prepared samples. The decrease in the pore volume and d pore,BJH upon increasing the phosphate content over 25 wt% may be due to the formation of multi-phosphate layers, as proposed in Scheme 1, whereby the surface hydroxyl groups of Sn(OH) 4 are replaced by phosphate groups and therefore there is an increased energy barrier and less crystal growth sites, consequently inhibiting the crystal growth.
The samples showed a bimodal pore size distribution consisting of smaller and larger interparticle pores, as illustrated in Fig. 6B. The addition of PO 4 3À to mSnO 2 was associated with an increase in the number of mesopores and micropores, as evident from the increase in the height of the distribution maxima and slight shi to lower rH values until a maximum at 25 wt% phosphate was reached, where it begins to decrease. Upon increasing the calcination temperature, (Fig. 3S †), both the smaller and larger interparticle pores decreased, which is in line with the decrease in the surface ( Table 2). The width of the mesopore type increased with phosphate content, which conrmed the role of PO 4 3À in increasing the porosity of mSnO 2 . 56 The nonaqueous potentiometric titration curves of n-butylamine of Fig. 4S † were used to calculate the total number of acid sites and the acid strength of the PO 4 3À /mSnO 2 catalysts. 57,58 As evident from Fig. 4S † and the data in Table 3, the acidity and the acid strength values of PO 4 3À /mSnO 2 changed with the phosphate loading and reached their maximum value for 25% PO 4 3À /mSnO 2 (Ei ¼ +480.6 mV) calcined at 400 C, which may be due to the completion of a bidentate phosphate monolayer on the tin oxide surface, as evident from the FTIR data. The decrease in the total acidity and the acid strength upon increasing the PO 4 3À loading may be due to the formation of more phosphate layers on the mSnO 2 surface. The calcination temperature 400 C was the most suitable calcination temperature for the generation of strong acidic sites. The loss of surface acidity at higher calcination temperatures up to 650 C may be due to the loss of H + by dehydration with the surface OH or OH of the next phosphate groups, forming a bridging phosphate group on the surface, as shown in Scheme 1, which causes a reduction in acidic hydrogen and a decrease in the total acidity.
Pyridine adsorption on the PO 4 3À /mSnO 2 samples was used to specify the Brönsted (B) and Lewis (L) acid sites on the sample surfaces. The FTIR spectra of pyridine adsorbed over the Brönsted (B) and Lewis (L) acid sites of the samples are shown in Fig. 7A and B. Bands at around 1444 and 1452 cm À1 were characteristic of pyridine adsorbed on Lewis acid sites. 59 The bands at 1536 and 1553 cm À1 were due to the pyridinium ion adsorbed on Brönsted acid sites. 60 Moreover, other FTIR bands appear at 1484 and 1503 cm À1 , attributed to the overlap of Lewis and Brönsted acid sites. 61 The integrated areas of the bands at 1536 and 1452 cm À1 were used to calculate the number of Brönsted and Lewis acid sites. It was evident from Fig. 7C  . Moreover, the Brönsted and Lewis acid sites change with the same trend as the formation of hydroquinone diacetate compound, which may indicate that Lewis and/or Brönsted acid sites can catalyse the formation of hydroquinone diacetate compound. No activity was observed for pure mSnO 2 , which may be due to its low acidity and site strength (+74 mV). As evident from Table 3 and Scheme 2, increasing the calcination temperature from 400 C to 650 C lead to a decrease in the catalytic activity, which is in line with the decrease in surface area and the decrease of the number of Brönsted and Lewis acid sites. The calcination temperature is an essential aspect in the synthesis of phosphated metal oxide 3.82 0.14 catalyst due to its role in the interaction between the metal oxide surface and phosphate groups for the generation of active acid sites. The optimum operating conditions for an anionmetal oxide catalyst was found to depend on the type of metal oxides, supported anions and the calcination temperature. 52 It was observed that with increasing the molar ratio from 1 : 1 to 1 : 2 and 1 : 3, the % yield of hydroquinone diacetate increased gently from 60.2%, to 77.5% to 93.2%, respectively, with 100% selectivity. However, a further increase in the acetic anhydride molar content was accompanied by a decrease in the percentage yield of hydroquinone diacetate to 83.1% (molar ratio 1 : 4). As acetic anhydride acts as both a solvent and reactant, an excess of acetic anhydride is used to facilitate the reaction until the molar ratio of hydroquinone to acetic anhydride is 1 : 3. Many factors may affect the catalyst activity; the dilution effect of acetic anhydride (solvent effect) and/or the catalyst active sites may be blocked due to a higher acetic anhydride concentration, thereby decreasing the yield of hydroquinone diacetate. Because the activation of acetic anhydride is the key step in the formation of hydroquinone diacetate, the decrease in % yield of hydroquinone diacetate may be due to the presence of low (molar ratio 1 : 1 and 1 : 2) or excess (molar ratio 1 : 4) activated acetic anhydride -Brönsted acid sites and/or activated acetic anhydride -Lewis acid sites   intermediates. Therefore, 1 : 3 molar ratio of reactants was considered optimum for the further studies.
The reusability of the 25% PO 4 3À /mSnO 2 catalyst was checked with three consecutive experiments by using the recovered catalyst. The catalyst was separated by ltration, washed with acetone several times and dried overnight at 100 C. The reaction was reperformed on this reactivated catalyst. It was observed that there was no signicant loss in activity, as shown in Fig. 9, and the catalytic activity of the catalyst was 93.2%, 92.4% and 91.1%, respectively. Moreover Fig. 7SA and B † show the TEM images of the 25% PO 4 3À /mSnO 2 aer the third run, which showed no change in the morphology of the reused catalyst; also the FTIR of the pyridine adsorbed on the reused sample showed no signicant loss in the acidity. The obtained results indicated that the catalyst had excellent reusability over three consecutive runs.

Reaction mechanism
The rst step in the formation of hydroquinone diacetate is chemisorption of the carbonyl group of acetic anhydride on the Brönsted acid sites or Lewis acid sites of the catalyst, then on the adsorbed acetic anhydride on the Brönsted acid sites, Scheme 3A shows a nucleophilic attack by OH of hydroquinone to form an intermediate that breaks down to lose acetic acid in the next step. Finally, the proton of the Brönsted acid site is eliminated to give hydroquinone acetate. The hydroquinone acetate reacts with another acetic anhydride -Brönsted acid site to give the nal product. The mechanism of the formation of hydroquinone diacetate on Lewis acid sites starts by the chemisorption of the carbonyl group on the acetic anhydride, which breaks down to form acylium ion. Then the OH of hydroquinone attacks acylium ion to form an intermediate, which lose a proton to give hydroquinone acetate, which undergoes the same steps to give the product hydroquinone diacetate, Scheme 3B.

Experimental
TEM images of the calcined PO 4 3À /mSnO 2 samples were obtained using a Joel JEM-1230 operated at 120 KV. FTIR spectra of the calcined samples were recorded by using MATTSON 5000 FTIR spectrophotometer (4 cm À1 resolution and 16 scans) in dried KBr (Merck). Physically adsorbed nitrogen at À196 C was used for elucidation of the texture properties of the calcined P/ mSnO 2 samples. The total acidity of the solid samples was measured by suspending 0.05 g of solid catalyst in 10 ml of acetonitrile for 3 h and titrating with 0.1 N n-butylamine base in acetonitrile. 35,57 An Orion 420 digital A pH-meter was used for measuring the potential variation. Lewis and Brönsted acid levels were determined from the exposure of the degassed solid samples at 200 C to dried pyridine. 68 Aer the removal of excess pyridine, a MATTSON 5000 FTIR spectrophotometer was used to record the spectra. For the NH 3 -TPD experiment, 200 mg of the samples were kept under ammonia gas for 4 h at 100 C. Then the samples were ushed with helium, and then the temperature was raised from 50 C to 700 C at a ramping rate of 5 C min À1 in the presence of helium, and the amount of NH 3 desorbed was recorded using a TCD detector.

Catalytic activity
The synthesis of hydroquinone diacetate (1,4-diacetoxybenzene) was used to test the catalytic activity of the samples. 0.1 g of the activated catalyst at 120 C was added to 3 ml (0.032 mol) of acetic anhydride and 1.1 g (0.01 mol) of hydroquinone in a 50 ml ask, and the mixture then stirred for 15 min. Aer 15 min, the clear solution was ltered to separate the catalyst, and the ltrate was poured into 500 ml of crushed ice. The crystalline solids were dried to constant weight. The product was identied by FTIR, NMR spectroscopy and melting point determination. The % yield of hydroquinone diacetate was then calculated.

Conclusions
Mesoporous SnO 2 was successfully prepared through a simple method and loaded with different amounts of phosphate species. The XRD and TEM images showed that the samples were mesoporous. Phosphate species were homogeneously adsorbed on the surface of mesoporous SnO 2 up to surface saturation coverage at 25 wt% PO 4 3À at which point the surface, area surface acidity level and acid strength were at a maximum. The acidity of the samples was found to decrease when the phosphate loading exceeded 25 wt%, at which point polyphosphate multilayers were formed. Both Lewis and/or Brönsted acid sites were present on the catalyst surface. The synthesis of hydroquinone diacetate was used as a model reaction for testing the activity of the solid sample for the formation of bulky molecules. The maximum formation of hydroquinone diacetate and the surface acidities were found at 25 wt% phosphate loading and 400 C calcination temperature, respectively. The catalyst could be successfully used many times without signicant loss of its catalytic activity.

Conflicts of interest
There are no conicts to declare.