Synthesis of phosphotungstic acid-supported versatile metal–organic framework PTA@MIL-101(Fe)–NH₂–Cl

Abstract

The catalyst PTA@MIL-101(Fe)–NH₂–Cl was synthesized in various solvents to investigate the effect of solvents on the catalyst crystal structure. PTA@MIL-101(Fe)–NH₂–Cl can only be obtained in DMF because of its excellent ability to deprotonate the organic ligands and promotion to yield MIL-101 crystals by forming a Cl–H–DMF complex. The water-containing solvents were unfavorable to yield MIL-101 crystals since MIL-101 crystals reorganized and other crystals were formed in this kind of solvents. The metal cation concentration kinetically controlled the crystal structure. MIL-101 crystals were fabricated only at low metal cation concentration. In the DMF synthesis system, the supported phosphotungstic acid (PTA) retained its Keggin structure. The acidity dramatically decreased compared to PTA caused by the double electrostatic interaction of amino groups and positively charged exterior surface of carrier MIL-101(Fe)–NH₂–Cl. The roles of both amino groups as PTA immobilization sites and chlorine groups as substrate absorption sites were confirmed in starch hydrolysis catalyzed by PTA@MIL-101(Fe)–NH₂–Cl.

---

1. Introduction

Phosphotungstic acid (PTA, H₃PW₁₂O₄₀) is one kind of Keggin-type heteropolyacid anion of which consists of Keggin units in which a central P atom in tetrahedral coordination (PO₄) is surrounded by 12 metal–oxygen octahedra (WO₆). The anion structure of PTA leads to complete dissociation of the counter cations H⁺ in normal solvents. Therefore, PTA possesses strong acidity and shows great performance in acid catalytic reactions.¹ However, as for the majority of homogeneous catalysts, PTA is difficult to isolate from the products which is unavoidable to greatly reduce its acidity in the recycling progress. Furthermore, it is non-oriented to substrates which is unfavorable for selective conversion. Supporting PTA to an appropriate carriers is the main road to overcome these drawbacks and thus investigation of PTA supported catalyst is promising in the acid-catalytic research field. Hitherto, neutral or basic carriers such as zeolites, activated carbon, SiO₂ and Al₂O₃ have been attempted to support PTA through pore-size limitation or acid–base interaction.² Y-type zeolites whose size of pore channel is smaller than PTA (1.4 nm) can encapsule PTA into their pores to avoid loss of PTA. These catalysts are only applicable to catalytic reactions of small-molecule reactants. The loading of PTA is limited by pore volume. Other zeolites with pore channel size larger than PTA are inevitable to lose PTA. Activated carbon, SiO₂ and Al₂O₃ support PTA through acid–base interaction between hydroxyl groups and PTA which can obviously decrease the PTA leaching. Meanwhile, the acid–base interaction greatly reduces acidity of PTA causing its poor catalytic performance.

Metal–Organic Frameworks (MOFs) are coordination polymers which are self-assembled by transition metal cations and polydentate organic ligands. Compared to the carriers mentioned above, MOFs take advantages of tunability of pore size and convenience of functioning.³ Among numbers of MOFs, MIL-101 (MIL, Material Institute Lavoisier) has unique pore structure – mesoporous cages (2.9 nm and 3.4 nm) which are accessible through microporous windows (1.2 nm and 1.6 nm)⁴ and compatible to PTA (1.4 nm diameters). Furthermore, MIL-101 can be functioned by selecting organic ligands with functional groups which not only can strengthen immobilization of PTA through interaction between the functional groups and PTA, but also can equip MIL-101 with selective adsorption ability to improve the oriented conversion of reactants in PTA catalysis. Therefore, MIL-101 has been considered as an appropriate alternative to support PTA.

Proceeding the immobilization of PTA and synthesis of MIL-101 simultaneously is an important method for supporting PTA to MIL-101. Directly added to synthesis system, PTA can be encapsuled into pore structure of MIL-101 in coordination reaction between metal cations and organic ligands (denoted as PTA@MIL-101).^5–8 The obtained catalyst has an excellent activity in desulfurization,^7,8 alcoholysis,⁸ denitrogenation,⁷ et al. As the carrier of PTA@MIL-101, MIL-101 supported PTA through window-size limitation. There was still partial PTA leaching out because the pore channel size of MIL-101 was 1.6 nm which was larger than PTA. Wang supported PTA to amino groups-functionalized MIL-101–NH₂ through electrostatic interaction between H⁺ of PTA and amino groups (denoted as PTA@MIL-101–NH₂) recently. The catalytic activity of PTA@MIL-101–NH₂ did not lose after six consecutive cycles in the oxidative desulfurization of dibenzothiophene by H₂O₂ and the leaching rate of PTA was only 4%.² It demonstrated that amino groups contributed to strengthen PTA immobilization and reduced its leaching. Although PTA is widely used as an acidic catalyst in many fields, but there is no work in the publishing literature concentrating on the acid catalytic activity of PTA@MIL-101–NH₂. Thus, it is necessary to study the effect of PTA encapsulation on its acidity. In theory, the electrostatic interaction between basic functional groups of MIL-101–NH₂ and H⁺ of PTA can make H⁺ difficult to dissociate and reduce its acidity. However, the effect of electrostatic interaction between PTA and basic functional groups on acidity of PTA has hardly been discussed until now. Besides, the opening of anion of PTA ([PW₁₂O₄₀]³⁻) may occur in synthesis process followed by formation of lacunary Keggin anion ([PW₁₁O₃₉]⁷⁻) which can affect dissociation of H⁺ and consequently change acidity of PTA. The acidity decrease of the supported PTA was reported in Haber's study of acidic performance of PTA supported SiO₂ applied to dehydration of ethanol.⁹ It was inferred that in water as solvent synthesis system, anion of PTA hydrolyzed into lacunary Keggin anion and loosed its tie to H⁺. The amount of H⁺ decreased and PTA acidity reduced obviously. Nevertheless, PTA exhibits stability and it has not been proved the opening of PTA Keggin anion during PTA@MIL-101–NH₂ synthesis process. Thus, the effect of structure change of PTA on its acidity needs to be further discussed.

The crystal structure change of MIL-101 during synthesis process probably affects the size of pore channel as well as integrity of pore structure which may change the limitation of pore structure to PTA and the stability of PTA immobilization. Furthermore, the crystal structure change of MIL-101 may affect the thermal stability. The structure of MIL-101 may collapse on the heating condition and lose the ability to support PTA as a carrier. Hence the crystal structure change of MIL-101 in synthesis process needs to be paid attention to. Proceeding PTA loading and MIL-101 synthesis simultaneously ensures uniform distribution of PTA in pore structures of MIL-101-type catalysts. The supported PTA does not affect the crystal structure of MIL-101.¹⁰ Therefore, the crystal structure change of MIL-101 caused by PTA encapsulation can be neglected.¹¹ As a transient crystal, MIL-101 crystal is strict with synthesis conditions. The solvents are considered to have significant effect on crystal synthesis. Bauer¹² and Lin¹³ attempted to synthesize MIL-101–NH₂ in various solvents including acetonitrile, methanol, N,N-dimethylformamide (DMF), water and NaOH aqueous solution. It was found that MIL-101 crystals were only fabricated in DMF and NaOH aqueous solution. MIL-53 crystal or MIL-88 crystal were fabricated in other solvents. It stated that solvents controlled crystal type in synthesis process. According to theoretical analysis, solvents can affect crystal synthesis in two aspects. Firstly, solvents dissolving capacity for organic ligands. The dissolving rate of organic ligands in solvent affects the coordination between organic ligands and metal cation and subsequently affects crystal structure. Secondly, the reaction between solvents and the intermediate products during synthesis process. This reaction promotes the intermediate products to decompose and transform into specific crystals. Until now, researchers have focused on selecting solvents to obtain MIL-101 crystals with specific functional groups and have paid less attention to discuss the effect of solvents in the crystal synthesis process. In addition, the changing rule of crystal structure affected by precursor concentrations as an important factor has been rarely shed light on. Precursor concentrations can kinetically control the rate of crystal growth. Increasing precursor concentrations accelerates crystal growth of MIL-101 and increases the crystallinity of MIL-101 which probably stimulates MIL-101 crystals to transform to thermodynamically stable products.¹² It is necessary to discuss the dissolving capacity of solvent for organic ligands, the reaction between solvents and the intermediate products occurring during synthesis process and the changing rule between precursor concentrations and crystal structure, so that the optimized solvents, precursor concentration as well as the mechanisms of these two factors in MIL-101 crystal synthesis process can be obtained.

In order to increase the conversion efficiency of reactants, synthesizing MIL-101 crystals with organic ligands grafted functional groups can equip MIL-101 with the ability to selectively adsorb the reactants which can help PTA to get more opportunity to catalyze the reactants. The negative atoms, such as N, O and Cl were always used to adsorb the reactants with –OH by forming hydrogen bond.^14,15 This work aims to fabricate versatile catalyst PTA@MIL-101–NH₂–Cl whose amino groups strengthen PTA immobilization through electrostatic interaction and chlorine groups enhance the catalyst with selective adsorption ability by forming hydrogen bond with reactants (as seen in Fig. 1). The effect of solvents dissolving capacity for organic ligands and reaction between solvents and intermediate products on crystal structure has been investigated. Fe³⁺ concentration was set to investigate the effect of precursor concentration on crystal structure. The acidity change of PTA@MIL-101(Fe)–NH₂–Cl was evaluated by acid density and acid strength. The electrostatic interaction between amino groups and PTA as well as formation of lacunary Keggin anion were discussed to understand the PTA structure change and acidity change during encapsulation. Starch was chosen to represent substrates with hydroxyl groups. The hydrolysis of starch catalyzed by PTA@MIL-101(Fe)–NH₂–Cl was carried out to detect the functions of grafted amino groups and chlorine groups in oriented conversion.

Fig. 1 Schematic view of PTA@MIL-101(Fe)–NH₂–Cl, electrostatic interactions between amino groups and PTA and the hydrogen bond between chlorine groups and substrates with hydroxyl groups.

2. Material and methods

2.1 Material

2-Nitroterephthalic acid (TA-NH₂, 98%), 2,5-dichloroterephthalic acid (TA-Cl, 98%) were purchased from Micxy Chemical Reagent Co., Ltd. (Chengdu, China). 1,4-Dicarboxybenzene (TA-H, AR), phosphotungstic acid (PTA, AR), FeCl₃·6H₂O (AR), N,N-dimethylformamide (DMF, AR) and other common chemicals and solvents used were all purchased from commercial sources and were analytical reagents. All the chemicals in this study were used without further purification.

2.2 PTA@MIL-101(Fe)–NH₂–Cl synthesis

The mixed-linker MOFs material PTA@MIL-101(Fe)–NH₂–Cl was synthesized by autoclaving FeCl₃·6H₂O, TA-NH₂, TA-Cl and PTA mixture in DMF through one-pot synthesis according to the reported method.¹² Typically, 0.75 g (4.14 mmol) of TA-NH₂ and 0.97 g (4.14 mmol) of TA-Cl were dissolved in 25 mL of DMF and a solution of 2.25 g (8.32 mmol) of FeCl₃·6H₂O, 1.00 g (0.3 mmol) PTA in 25 mL of DMF was added. The mixture solution was heated to 110 °C for 16 h without stirring in a Teflon-lined steel autoclave bomb with a total volume of 100 mL. The product was separated by centrifugation, washed with DMF at ambient temperature and finally dried at 80 °C for not less than 10 h under vacuum. To investigate the solvent effect on the crystal structures of PTA@MIL-101(Fe)–NH₂–Cl, DMF was replaced by water or 0.4 M NaOH aqueous solution and the mixture of DMF and water with the volume ratio of 0.95 : 0.05 and 0.90 : 0.10. To understand the effect of precursors concentration on the crystal structures of PTA@MIL-101(Fe)–NH₂–Cl, the molar concentration of FeCl₃·6H₂O in DMF was varied (0.16 M, 0.74 M, 1.47 M).

Samples with various molar ratio of amino groups to PTA (TA-NH₂/PTA) were synthesized as described below. In order to obtain samples with TA-NH₂/PTA = 0, 1, 2, 3, 10, 13, 0 mmol, 0.3 mmol, 0.6 mmol, 0.9 mmol, 3 mmol, 4 mmol of TA-NH₂, 0.97 g (4.14 mmol) of TA-Cl, 2.25 g (8.32 mmol) of FeCl₃·6H₂O and 1.00 g (0.3 mmol) PTA were added in 50 mL of DMF. 4 mmol, 3.7 mmol, 3.4 mmol, 3.1 mmol, 1 mmol, 0 mmol TA-H were also added to the above solution to ensure the same molar mass of organic ligands in the synthesis of every samples. The mixture solution was heated to 110 °C for 16 h without stirring in a Teflon-lined steel autoclave bomb with a total volume of 100 mL. The product was separated by centrifugation, washed with DMF at ambient temperature and finally dried at 80 °C for not less than 10 h under vacuum.

2.3 Characterization

The crystal structures of the sample were determined by an X-ray diffractometer (XRD) D8 Advance X-ray Diffraction system and Bruker AXS with a copper target tube radiation (Cu Kα1) producing X-rays at a wavelength of 0.15418 nm. Materials were placed on a quartz plate and were scanned from 5° to 60° (2θ) at scan speed of 1.2° min⁻¹, a scan step of 0.02°, 1.0° DS-SS slits, 8.0 mm RS slit for monochromator. Fourier transform infrared spectra (FTIR) was recorded with KBr disk containing the powder sample with the FT-IR spectrometer (Nicolet Magna 550) in the range of 4000–500 cm⁻¹. The morphology and element analysis of the sample were examined by a field emission scanning electron microscope (Leo 1530 VP) together with an EDS (Energy-dispersive X-ray spectroscopy, Inca 300). C, N, H element analysis were examined by Vario EI III.

The acid density of the samples was determined by the acid–base titration method as follow, NaCl aqueous solution (2.0 mM, 20 mL) was added to a sample (0.1 g). The mixture was stirred for 24 h at room temperature. After centrifugal separation, the supernatant solution was titrated by NaOH aqueous solution (0.001 M) using phenolphthalein as the indicator. The acid strength of the sample was evaluated by the initial potential detected by a pH meter of the solution of the sample (0.1 g) after stirred with acetonitrile (30 mL) for 12 h. It was reported that the initial potential of super strong acid was beyond 100 mV, strong acid was between 0 and 100 mV, weak acid was between −100 mV and 0 mV and super weak acid was below −100 mV.

3. Results and discussion

3.1 Solvent effect on crystal structure of PTA@MIL-101(Fe)–NH₂–Cl

3.1.1 Crystal structure of samples synthesized in various solvent. PTA@MIL-101(Fe)–NH₂–Cl was attempted to fabricate in DMF, water, mixture of DMF and water and NaOH aqueous solution to investigate the effect of the solvents on the crystal structure formation. Fig. 2 demonstrated XRD patterns of the samples synthesized in various solvents. Considering that both MIL-101 and MIL-53 may be synthesized in the chosen solvents. XRD patterns of the obtained samples were analyzed referring to the simulated XRD patterns of these two crystals. The simulated XRD patterns of MIL-101 (Fig. 2(b)) exhibited five peaks at 2θ = 5.1°, 5.5°, 5.9°, 8.4°, 9.1° respectively corresponding to the 511, 440, 351, 822, 911 reflection. The simulated XRD patterns of MIL-53 (ref. 16) (Fig. 2(g)) exhibited a peak at 2θ = 8.9° corresponding to the 110 reflection and a peak at 2θ = 18.2° corresponding to both the 211 and 220 reflection and the peak at 2θ = 8.9° was the highest.

Fig. 2 XRD patterns of samples synthesized in various solvents (a) DMF, (b) simulated XRD patterns of MIL-101, (c) water, (d) DMF/water (v/v) = 0.95 : 0.05, (e) 0.90 : 0.10, (f) NaOH aqueous solution, (g) simulated XRD patterns of MIL-53.

As can be seen in Fig. 2(a), the patterns of the sample synthesized in DMF corresponded to the one of MIL-101 which meant that the sample was MIL-101 crystal. The patterns exhibited low peaks around 2θ = 7.6° which the simulated patterns of MIL-101 without PTA (Fig. 2(b)) did not exhibit. None of these peaks overlapped with those of PTA which exhibited no peaks below 2θ = 10°. It was suggested that MIL-101 without PTA had a null intensity around 2θ = 7.6° and the supported PTA created observed intensity around 2θ = 7.6° by adding electron density to the crystal structure of MIL-101.¹⁷ The highest peak of simulated patterns of MIL-101 was at 2θ = 5.1° while intensity of three peaks peak of sample at 2θ = 5°–6° declined greatly and its highest peak was at 2θ = 9.1°. The intensity decline of peaks at low 2θ may demonstrated either the occupation of pore channels by amino groups or the reduction of crystal size of samples synthesized in DMF. The simulated patterns of MIL-101 crystal were assumed that it was obtained in water as solvent system. DMF had the stronger organic ligands dissolving ability than water which accelerated the deprotonation of the organic ligands to offer more deprotonated organic ligands and increased the rate of crystal growth. The crystal size of the MIL-101 synthesized in DMF became smaller than the one synthesized in water. It elucidated the wide and weak peaks of the MIL-101 crystal synthesized in DMF at low 2θ. Compared with the simulated MIL-101, the highest peak of samples in DMF was located at different 2θ which demonstrated that DMF promoted the growth of 911 crystal face. Therefore, PTA@MIL-101(Fe)–NH₂–Cl was obtained in DMF which spurred to synthesize small-size MIL-101 crystals and grow the 911 crystal reflection.

The XRD patterns of the samples synthesized in water (Fig. 2(c)) exhibiting the peaks at 2θ = 9.0° and 18.2° were analogy to patterns of MIL-53. The retained water in the pores of the samples caused the peak at 2θ = 19.5°. As the organic ligands, TA-NH₂ was difficult to dissolve in water. The slow rate of TA-NH₂ deprotonation and crystal growth were beneficial for nuclei to arrange to MIL-53 crystal. Hence water dissolving capacity for organic ligands was bound to the fabrication of MIL-53 crystal instead of MIL-101 crystal in water.

Fig. 2(d) and (e) showed the XRD patterns of the samples synthesized in the DMF/water solution at the ratio (v/v) of 0.95 : 0.05 and 0.90 : 0.10 respectively. Except the small peaks around 2θ = 7°, the patterns completely matched to the simulated patterns of MIL-53 which demonstrated that these two samples were MIL-53 crystals. The peak around 2θ = 9° corresponding to 110 reflection became sharp and intensive. It was likely that water in the DMF/water solvent reduced the association effect of the organic ligands by interspersing among them and resulted in faster deprotonation and crystal growth in specific direction.¹⁸ Thus, MIL-53 crystals were obtained and the growth of 110 crystal reflection was accelerated in the DMF/water solvent. The reason why the DMF/water solvent was beneficial to synthesize MIL-53 crystal instead of MIL-101 crystal still remained to be further discussed.

The XRD patterns of samples synthesized in NaOH aqueous solution revealed two adjacent and separate peaks at 2θ = 8° and 9° while the simulated pattern of MIL-101 revealed two continuous peaks with shoulders at 2θ = 8° and 9°. The peak at 2θ = 9.1° shifted to higher 2θ compared with the simulated MIL-53. It was inferred that the samples synthesized in NaOH aqueous solution were the mixture of MIL-53 and MIL-101 crystals. Both MIL-53 and MIL-101 crystals were probably fabricated in 0.4 M NaOH aqueous solution. Since MIL-101 and MIL-53 crystals were kinetically and thermodynamically favored products separately,¹² there was a transient stage in which MIL-101 transformed to MIL-53 and MIL-101 as well as MIL-53 crystals were obtained. In the XRD patterns of the sample, the peak at 2θ = 8.5° belonging to MIL-101 became insensitive and wide while the peak at 2θ = 9.1° belonging to MIL-53 became a single peak without a shoulder. It indicated that the sample stayed in this transient stage. In consequence, shortening synthesis time may prevent formation of MIL-53 and obtain MIL-101 which probably led to the low crystallinity and imperfect crystals.

3.1.2 Morphology of PTA@MIL-101(Fe)–NH₂–Cl synthesized in DMF. PTA@MIL-101(Fe)–NH₂–Cl can be obtained only in DMF from the results mentioned above. The morphology was examined by SEM as shown in Fig. 3. Fig. 3(a) and (c) were the images of MIL-101 synthesized according to Ferey's report which were magnified 10 thousand times and 30 thousand times.⁴ The MIL-101 crystals were uniform octahedrons with smooth surface and the size ranged from 825 nm to 1050 nm. Fig. 3(b) and (d) were SEM images of PTA@MIL-101(Fe)–NH₂–Cl synthesized in DMF magnified 10 thousand times and 30 thousand times. The samples were uniform octahedrons and the surface were ambiguous and rough compared with MIL-101. It was because the grafted chlorine groups easily adsorbed water to the surfaces of crystals through hydrogen bond. Furthermore, the size of PTA@MIL-101(Fe)–NH₂–Cl was between 300 nm and 525 nm which was smaller than the MIL-101. It indicated that the fast nuclei and crystal growth rate in DMF with strong dissolving ability of organic ligands led to small-size crystals.¹³ This observation evidenced the inference that reduction of crystal size caused intensity decline of peaks at low 2θ.

Fig. 3 SEM image of typical MIL-101(Cr) (a and c), sample synthesized in DMF (b and d).

3.1.3 Mechanism of DMF for PTA@MIL-101(Fe)–NH₂–Cl synthesis. FT-IR spectrum of precursors solution was referred to analyze the reaction of DMF occurring in synthesis process and to discuss the mechanism of DMF for PTA@MIL-101(Fe)–NH₂–Cl synthesis. Fig. 4 was the FT-IR spectrum of precursors solution including DMF, FeCl₃·6H₂O, PTA and organic ligands. Compared with the spectrum of DMF, the stretching vibration band of C=O which was weakened by the hydrogen bond Cl–H–O shifted from 1669 cm⁻¹ to 1642 cm⁻¹ and became wider. The C–H bond shifted from 1390 cm⁻¹ to 1385 cm⁻¹ and exhibited a new band at 1366 cm⁻¹ which resulted from the electron density reduction of the oxygen atom caused by hydrogen bond.¹⁹ Bending vibration band of O=C–N at 664 cm⁻¹ represented that a part of DMF did not interact with Cl⁻ in the solution. The emergence of a new band at 698 cm⁻¹ represented that the other part of DMF interacted with Cl⁻ and formed Cl–H–O. Compared with the spectrum of DMF, the vibration bands of C=O and C–H shifted to smaller wavenumber and became less intensive in the spectrum of precursors solution. It suggested that DMF can connect with Cl⁻ forming hydrogen bond in precursor solution which promoted to yield MIL-101 during synthesis process. As shown in Fig. 5, in synthesis process, DMF formed hydrogen bond with Cl⁻ through interaction of Cl⁻ + H₂O + DMF → OH⁻ + Cl–H–DMF. Since there was only a small amount of water in DMF as solvent synthesis system, the interaction of DMF can impel terminal aqua ligands of intermediate product MOF-235 to dissociate. Formed OH⁻ substituted unsaturated sites of intermediate products which promoted transformation of MOF-235 to MIL-101. This deduction was in line with the Gosten's result that obtained in situ ¹H and ²⁷Al NMR analysis of crystal growth of MIL-101–NH₂.²⁰ In the DMF/water solvent, water was available to interact with DMF. DMF cannot play its promotional role because terminal aqua ligands of MOF-235 did not dissociate. It exposited that the phenomenon observed in Section 3.1.1 that water was beneficial to form MIL-53 instead of MIL-101. Accordingly, DMF promoted to fabricate MIL-101 by forming Cl–H–DMF in DMF.

Fig. 4 FT-IR spectrum of (a) DMF and (b) precursors solution.

Fig. 5 Promotion of DMF in MIL-101(Fe)–NH₂–Cl fabrication (L denoted organic ligands and S denoted solvents).²⁰

3.2 Precursors concentration effect on crystal structure of PTA@MIL-101(Fe)–NH₂–Cl

The molar ratio of metal cation to organic ligands was set over 1 to ensure the polydentate coordination. In order to study the effect of metal cation concentrations on the crystal structure under this condition, PTA@MIL-101(Fe)–NH₂–Cl was attempted to synthesize at 0.16 M, 0.74 M and 1.47 M Fe³⁺ concentration in DMF. Since either MIL-101 crystal or MIL-53 crystal can be obtained at various Fe³⁺ concentration, the crystal structures were investigated based on simulated XRD patterns of MIL-101 and MIL-53. Fig. 6(a)–(c) were the XRD patterns of the samples synthesized at various Fe³⁺ concentrations. As shown in Fig. 6(a), the sample obtained at 0.16 M exhibited the characteristic peaks around 2θ = 8° and 9° belonging to MIL-101 crystals. The patterns of samples obtained at 0.74 M and 1.47 M were similar to the patterns of MIL-53. The sample obtained at 0.74 M shown that there were only two low peaks at 2θ = 7.7° and 9.1°. The patterns of sample obtained at 1.47 M perfectly matched to the one of MIL-53.

Fig. 6 XRD patterns of samples synthesized with varied Fe³⁺ concentrations (a) 0.16 M, (b) 0.74 M, (c) 1.47 M, simulated XRD patterns of (d) MIL-101 and (e) MIL-53.

It was speculated that MIL-101 and MIL-53 were likely to be fabricated in DMF and low Fe³⁺ concentration (0.16 M) favored MIL-101 as a kinetic product. With the increase of Fe³⁺ concentration, the formed MIL-101 gradually converted to MIL-53 as a thermodynamic product. MIL-53 was obtained at 0.74 M Fe³⁺ concentration with the low peak intensity because of immature crystal growth. The sample obtained at the 1.47 M Fe³⁺ concentrations showed all characteristic peaks of MIL-53 due to the higher Fe³⁺ concentration which stimulated the crystal growth. In DMF, Fe³⁺ concentration kinetically controlled the crystal structure. High Fe³⁺ concentration was helpful for MIL-53 synthesis. MIL-101 can be obtained only at the low Fe³⁺ concentration.

3.3 Effect of PTA encapsulation to MIL-101(Fe)–NH₂–Cl on its acidity

3.3.1 Acidity change of PTA after encapsulation. In order to investigate the effect of electrostatic interaction on dissociation degree (acid strength) and available amount of H⁺ (acid density) from PTA, the molar ratio of amino groups to PTA (TA-NH₂/PTA) of the carrier MIL-101(Fe)–NH₂–Cl was adjusted to tune the intensity of their electrostatic interaction. In Table 1, the mass ratio of C to N (C/N) based on element analysis of the samples was close to the one based on theory. It indicated that the samples with designed TA-NH₂/PTA were synthesized. PTA content decreased from 12.8 wt% to 10.2 wt% with the increase of TA-NH₂/PTA from 0 to 10. As TA-NH₂/PTA = 13, PTA content increased to14.6 wt%. In theory, PTA content was affected by both pore volume and the electrostatic interaction between PTA and amino groups. When TA-NH₂/PTA was from 0 to 10, the electrostatic interaction was not strong enough to attract more PTA. Moreover, the pore volume decreased with the increase of amino groups. Thereby, PTA content decreased. When TA-NH₂/PTA = 13, since the electrostatic interaction was strong enough to show affinity to PTA, more PTA was encapsuled and thus PTA content increased. The acid strength and acid density of the samples were measured by the initial potential and acid–base titration. The results were shown in Table 1. Compared with the PTA of which the initial potential was 511 mV, the initial potential of the samples greatly decreased. The initial potential of the samples declined dramatically with the increase of TA-NH₂/PTA. When TA-NH₂/PTA was 13, the initial potential dropped to 125 mV. It stated that the electrostatic interaction between amino groups and PTA was strengthened with the increase of TA-NH₂/PTA. Thus H⁺ was tightened and more difficult to dissociate which reduced the acidity strength of the samples. It was noteworthy that the initial potential of the sample without amino groups also dramatically declined which indicated that the supported PTA was affected by other part of MIL-101(Fe)–NH₂–Cl except for amino groups. The acid density of the samples decreased from 0.62 mmol g⁻¹ to 0.77 mmol g⁻¹ with the increase of TA-NH₂/PTA. The dissociation difficulty of H⁺ improved with the increase of TA-NH₂/PTA and the available amount of H⁺ reduced accordingly leading to the drop of acid density.

Table 1 Acid density and acid strength of PTA@MIL-101(Fe)–NH₂–Cl at various NH₂/PTA

TA-NH2/PTA	Formula^12 of MIL-101(Fe)–NH2–Cl	C/N based on theory	C/N based on element analysis	PTA content (wt%)	Initial potential (mV)	Acid density (mmol g^−1)
0	FeO3(DMF)3Cl(TA-Cl)1.5(TA-H)1.5	9.42	9.87	12.8	195	0.77
1	FeO3(DMF)3Cl(TA-Cl)1.5(TA-NH2)0.1125(TA-H)1.3875	9.08	9.09	12.8	180	0.70
2	FeO3(DMF)3Cl(TA-Cl)1.5(TA-NH2)0.225(TA-H)1.275	8.77	8.14	12.2	180	0.70
3	FeO3(DMF)3Cl(TA-Cl)1.5(TA-NH2)0.3375(TA-H)1.1625	8.47	8.01	11.5	170	0.66
10	FeO3(DMF)3Cl(TA-Cl)1.5(TA-NH2)1.125(TA-H)0.375	6.85	6.43	10.2	160	0.63
13	FeO3(DMF)3Cl(TA-Cl)1.5(TA-NH2)1.5	6.28	5.69	14.6	125	0.62

---

3.3.2 Structure change of supported PTA and mechanism for its acidity change. The interaction between PTA and MIL-101(Fe)–NH₂–Cl was investigated by FT-IR spectrum. The FT-IR spectrum of PTA and MIL-101(Fe)–NH₂–Cl with different ratio of TA-NH₂/PTA were shown in Fig. 7. As seen in Fig. 7(g), PTA was distinguished by four characteristic absorption bands appearing at 1078 cm⁻¹, 983 cm⁻¹, 889 cm⁻¹ and 803 cm⁻¹ respectively corresponding to stretching vibrations of P–O, W–Od (Od as terminal oxygen atom), W–Ob–W (Ob as oxygen atom bridging between corner sharing octahedras) and W–Oc–W (Oc as oxygen atom bridging between edge sharing octahedras). Fig. 7(a)–(f) were the FT-IR spectrums of six samples with TA-NH₂/PTA ranging from 0 to 13. The bands at 3468 cm⁻¹ and 3358 cm⁻¹ ascribed to –NH₃⁺ stretching vibration. The bands at 1257 cm⁻¹and 1050 cm⁻¹ belonged to C–N and C–Cl stretching vibration. The bands attributing to P–O and W–Ob–W at 1078 cm⁻¹ and 889 cm⁻¹ were identified for PTA. It can be seen from Fig. 7(a)–(f) that the bands representing W–Od at 983 cm⁻¹ shifted to lower wavenumber and became insensitive. The bands ascribed to W–Oc–W split into a maximum at 826 cm⁻¹ with a shoulder at 792 cm⁻¹. Formation of Keggin lacunary anion was significantly evidenced by W–Ob–W band disappearance. In Fig. 7(a)–(f), the W–Ob–W bands of the samples were obviously observed. It demonstrated that Keggin lacunary anion did not form and the PTA retained complete Keggin anion after encapsulation. Hence DMF can stabilize Keggin anion of PTA to prevent acidity decrease caused by lacunary anion in addition to promote fabrication of MIL-101. As shown in Fig. 7(a)–(f), with the increase of TA-NH₂/PTA, the band at 1257 cm⁻¹ assigned to C–N vibration became intensive which indicated the increase of amino groups. The bands at 3468 cm⁻¹ and 3358 cm⁻¹ assigned to –NH₃⁺ vibrations²¹ became strong which demonstrated the enhancement of electrostatic interaction between amino groups and PTA. It can be concluded that raising the amount of amino groups can enhance electrostatic interaction between amino groups and PTA. Furthermore, the location and intensity of bands at 803 cm⁻¹ and 983 cm⁻¹ assigned to W–Oc–W and W–Od changed obviously with the increase of electrostatic interaction since H⁺ connected with anion of Oc and Od in PTA. Combined acidity evaluation with FT-IR spectrums of the samples in Section 3.3.1, a conclusion can be drawn as follows. The occurred electrostatic interaction between amino groups and H⁺ of PTA did not break anion structure of PTA but made H⁺ difficult to dissociate which led to the acidity reduction of PTA.

Fig. 7 FT-IR spectra of samples at TA-NH₂/PTA = (a) 0, (b) 1, (c) 2, (d) 3, (e) 10, (f) 13 and (g) PTA.

As shown in Fig. 7(a), although without amino groups, the vibrations representing W–Od and W–Oc–W of the samples exhibited differences from PTA. It suggested that not only electrostatic interaction of amino group but also other interaction between the supported PTA and carrier affected the structure of the supported PTA which was compatible to inference based on acidity change in Section 3.3.1. In synthesis process, the exterior of MIL-101(Fe)–NH₂–Cl was positively charged after dissociation of Cl⁻ or OH⁻ and moreover, PTA was negatively charged after dissociation of H⁺. The positively-charged MIL-101(Fe)–NH₂–Cl and negatively-charged PTA connected with each other through Oc and Od as the following chemical formula,

n[Fe₃O(organic ligands)₃(solvents)₂(OH, Cl)] + H₃PW₁₂O₄₀ → [Fe₃O(organic ligands)₃(solvents)₂]H_3−nPW₁₂O₄₀ + nH₂O or HCl (n ≤ 3).

This speculation agreed with the one that Keggin anion supported to MIL-101 through interaction with its surface in Maksimchuk's²² and Ferey's⁴ reports.

The acidity of supported PTA was affected by the carrier MIL-101(Fe)–NH₂–Cl in two ways, even though the supported PTA retained Keggin structure in synthesis process. Firstly, grafted amino groups and H⁺ of the supported PTA formed –NH₃⁺H₂PW₁₂O₄₀ through electrostatic interaction which made H⁺ difficult to dissociate and then reduced available amount of H⁺. Secondly, anion of the supported PTA connected with the positively-charged exterior of MIL-101(Fe)–NH₂–Cl and the available amount of H⁺ declined by forming hydrides.

3.4 Function evaluation of amino groups and chlorine groups of PTA@MIL-101(Fe)–NH₂–Cl

3.4.1 Function of amino groups of PTA@MIL-101(Fe)–NH₂–Cl. The catalytic stability of PTA@MIL-101(Fe)–NH₂–Cl was tested by hydrolyzing the starch which represents the substrates with hydroxyl groups. Fig. 8 illustrated the reducing sugar yield in three consecutive hydrolysis of starch catalyzed by PTA@MIL-101(Fe)–NH₂–Cl at various TA-NH₂/PTA. It can be seen from Fig. 8 that the supported PTA presented the best catalytic capability in starch hydrolyzing in the first time using. The yield of reducing sugar decreased with the increase of the TA-NH₂/PTA. The highest yield of reducing sugar 25.81% was obtained as the TA-NH₂/PTA was 0. With the increase of PTA@MIL-101(Fe)–NH₂–Cl using times, the yield of reducing sugar decreased which indicated the decrease of the supported PTA catalytic capability. According to the increase of the reducing sugar yield as well as the increase of the TA-NH₂/PTA in the PTA@MIL-101(Fe)–NH₂–Cl at the second and third using times, it can be concluded that the amino groups in the PTA@MIL-101(Fe)–NH₂–Cl was beneficial to the catalytic stability of the supported PTA. The more TA-NH₂/PTA, the stronger the stability of the catalysts with the increase of the using time. The dramatic decline of the reducing sugar yield was observed at low TA-NH₂/PTA with the increase of the using times. As the TA-NH₂/PTA was 13, the yields of the reducing sugar were similar at each using time ranging from 19.48 to 19.98. It was proved that the grafted amino groups played its role as PTA immobilization in PTA@MIL-101(Fe)–NH₂–Cl.

Fig. 8 Yield of reducing sugar in starch and avicel hydrolysis catalysed by samples with varied molar ratio of TA-NH₂ and PTA (condition: 10 mL water, 0.1 g catalyst, 0.02 g starch at 120 °C for 5 h).

FT-IR was used to analyze the chemical structure of the catalysts after starch hydrolysis to further discuss the function of amino groups. The FT-IR spectrum of the catalysts at TA-NH₂/PTA = 13 and TA-NH₂/PTA = 0 after reused 3 times in starch hydrolysis was shown in Fig. 9. Both Fig. 9(a) and (b) exhibited PTA characteristic peaks at 1078 cm⁻¹, 967 cm⁻¹, 889 cm⁻¹, 803 cm⁻¹ which demonstrated PTA remained in the catalysts after reused 3 times. The intensity of the PTA characteristic peaks at TA-NH₂/PTA = 13 was more intensive than TA-NH₂/PTA = 0. Based on element analysis, PTA percentage reduced from 14.6% to 10.7% as TA-NH₂/PTA = 13 and reduced from 12.8% to 5.7% as TA-NH₂/PTA = 0. The leaching rate of PTA was smaller in the catalyst at TA-NH₂/PTA = 13 than that at TA-NH₂/PTA = 0. It indicated that the amino groups helped to decrease the PTA leaching in hydrolysis. Therefore, amino groups were confirmed to immobilize PTA in PTA@MIL-101(Fe)–NH₂–Cl.

Fig. 9 FT-IR spectra of samples at TA-NH₂/PTA = (a) 13, (b) 0 after starch hydrolysis.

3.4.2 Function of chlorine groups of PTA@MIL-101(Fe)–NH₂–Cl. To clarify the role of chlorine groups in PTA@MIL-101(Fe)–NH₂–Cl, the relationship between chlorine groups percentage and reducing sugar yield in starch hydrolysis was investigated (Fig. 10). Under the condition that dosage of starch was set throughout, the reducing sugar yield had a liner relation with chlorine groups percentage (R² = 0.9652) which increased from 20.62% to 36.45% accordingly. It was indicated that the more starch was connected to PTA@MIL-101(Fe)–NH₂–Cl with chlorine groups as absorption sites through hydrogen bond, the more reducing sugar can be obtained by hydrolyzing starch with PTA as catalytic sites. In avicel hydrolysis, 13.06% reducing sugar yield was obtained with 6.27 Cl wt% catalyst which was higher than 5.3% reducing sugar yield with MIL-101–SO₃H without chlorine groups.²³ With chlorine groups, PTA@MIL-101(Fe)–NH₂–Cl can adsorb avicel which improved the catalyst to contact more avicel and was beneficial to increase the catalytic ability. Therefore, PTA@MIL-101(Fe)–NH₂–Cl with chlorine groups showed better catalytic performance in avicel hydrolysis than MIL-101–SO₃H without chlorine groups. In conclusion, it was the chlorine groups that enhanced the substrates absorption ability and improved the catalytic performance of PTA@MIL-101(Fe)–NH₂–Cl.

Fig. 10 Effect of chlorine percentage on reducing sugar yield in starch hydrolysis (condition: 10 mL water, 0.06 g catalyst, 0.04 g starch at 120 °C for 5 h).

4. Conclusion

In this work, PTA@MIL-101(Fe)–NH₂–Cl was designed and synthesized according to the demand of acid-catalytic conversion of reactants with hydroxyl groups. The effect of the solvents and the metal cation concentration on catalysts crystal structure in synthesis process was focused on. DMF played a promotional role to synthesize MIL-101 crystals because of its strong ability to dissolve organic ligands and the formation of Cl–H–DMF. Water had the poor ability to dissolve organic ligands and impelled intermediate products to transform to MIL-53 crystals. Therefore, water-containing solvents were not in favor of MIL-101 crystals. Metal cation concentration kinetically controlled crystal growth of PTA@MIL-101(Fe)–NH₂–Cl which can be obtained at low metal cation concentration.

Anion of the supported PTA retained its Keggin structure. However, H⁺ was tightened by electrostatic interaction with amino groups and was difficult to dissociate. In addition, the available amount of H⁺ reduction resulted from the formation of hydrides through electrostatic interaction between PTA and positively-charged exterior of carrier.

In starch hydrolysis catalyzed by PTA@MIL-101(Fe)–NH₂–Cl, the grafted amino groups and chlorine groups functioned PTA@MIL-101(Fe)–NH₂–Cl with PTA immobilization ability and substrates absorption ability through electrostatic interaction between amino groups and PTA as well as hydrogen bond between chlorine groups and hydroxyl groups of starch.

Acknowledgements

We acknowledge the financial support from the National Science Foundation of China (Grant No. 31570568), State Key Laboratory of Pulp and Paper Engineering Foundation (201535) and Guangdong High Level Talent Project (201339).

References

1. J. Yang, M. J. Janik, D. Ma, A. Zheng, M. Zhang, M. Neurock, R. J. Davis, C. Ye and F. Deng, J. Am. Chem. Soc., 2005, 127, 18274.
2. X. Wang, Y. Huang, Z. Lin and R. Cao, Dalton Trans., 2014, 43, 11950.
3. A. Schneemann, V. Bon, I. Schwedler, I. Senkovska, S. Kaskel and R. A. Fischer, Chem. Soc. Rev., 2014, 43, 6062.
4. G. Ferey, Science, 2005, 309, 2040.
5. S. Jia, Y. Zhang, Y. Liu, F. Qin, H. Ren and S. Wu, J. Hazard. Mater., 2013, 262, 589.
6. X. Hu, Y. Lu, F. Dai, C. Liu and Y. Liu, Microporous Mesoporous Mater., 2013, 170, 36.
7. S. Ribeiro, A. D. S. Barbosa, A. C. Gomes, M. Pillinger, I. S. Gonçalves, L. Cunha-Silva and S. S. Balula, Fuel Process. Technol., 2013, 116, 350.
8. F. B. C. L. Lik, H. Wee and J. A. Martensa, Green Chem., 2014, 1351.
9. J. Haber, K. Pamin, L. Matachowski and D. Mucha, Appl. Catal., A, 2003, 256, 141.
10. J. Juan-Alcañiz, E. V. Ramos-Fernandez, U. Lafont, J. Gascon and F. Kapteijn, J. Catal., 2010, 269, 229.
11. J. Juan-Alcañiz, M. G. Goesten, E. V. Ramos-Fernandez, J. Gascon and F. Kapteijn, New J. Chem., 2012, 36, 977.
12. S. Bauer, C. Serre, T. Devic, P. Horcajada, J. Marrot, G. Férey and N. Stock, Inorg. Chem., 2008, 47, 7568.
13. Y. Lin, C. Kong and L. Chen, RSC Adv., 2012, 2, 6417.
14. J. Chen, W. Zhang, L. Chen, L. Ma, H. Gao and T. Wang, ChemPlusChem, 2013, 78, 142.
15. L. S. A. X. Pan, Energy Environ. Sci., 2012, 5, 6889.
16. J. M. Chin, E. Y. Chen, A. G. Menon, H. Y. Tan, A. T. S. Hor, M. K. Schreyer and J. Xu, CrystEngComm, 2013, 15, 654.
17. L. Bromberg, Y. Diao, H. Wu, S. A. Speakman and T. A. Hatton, Chem. Mater., 2012, 24, 1664.
18. X. Cheng, A. Zhang, K. Hou, M. Liu, Y. Wang, C. Song, G. Zhang and X. Guo, Dalton Trans., 2013, 42, 13698.
19. S. G. S. O. V. D. Maiorov, Russ. Chem. Bull.+, 1993, 42, 1511.
20. M. G. Goesten, P. C. M. M. Magusin, E. A. Pidko, B. Mezari, E. J. M. Hensen, F. Kapteijn and J. Gascon, Inorg. Chem., 2014, 53, 882.
21. G. Luo, L. Kang, M. Zhu and B. Dai, Fuel Process. Technol., 2014, 118, 20.
22. N. Makismchuk, M. Timofeeva, M. Melgunov, A. Shmakov, Y. Chesalov, D. Dybtsev, V. Fedin and O. Kholdeeva, J. Catal., 2008, 257, 315.
23. G. Akiyama, R. Matsuda, H. Sato, M. Takata and S. Kitagawa, Adv. Mater., 2011, 23, 3294–3297.

---