One-pot synthesis of ethylene glycol by oxidative hydration of ethylene with hydrogen peroxide over titanosilicate catalysts

Journal of Catalysis 358 (2018) 89–99

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

One-pot synthesis of ethylene glycol by oxidative hydration of ethylene with hydrogen peroxide over titanosilicate catalysts Xinqing Lu a, Hao Xu a,⇑, Jiaying Yan a, Wen-Juan Zhou a,b, Armin Liebens b, Peng Wu a,⇑ a b

Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200062, China Eco-Efficient Products and Processes Laboratory (E2P2L), UMI 3464CNRS—Solvay, 3066 Jin Du Road, Xin Zhuang Ind. Zone, Shanghai 201108, China

a r t i c l e

i n f o

Article history: Received 27 October 2017 Revised 2 December 2017 Accepted 3 December 2017

Keywords: Titanosilicate Ti-MWW Oxidative hydration Ethylene Ethylene glycol

a b s t r a c t The oxidative hydration of ethylene with aqueous hydrogen peroxide was investigated over various titanosilicate catalysts for the purpose of one-pot synthesis of ethylene glycol (EG). The effect of titanosilicate topology (Ti-MWW, TS-1, Ti-MCM-68 and Ti-MOR), solvent, Ti content, catalyst amount, H2O/H2O2 ratio, reaction temperature and time on the EG production have been studied in detail. The Ti-MWW/ H2O2/H2O catalytic system showed the highest EG yield together with high H2O2 conversion and utilization efficiency for the oxidative hydration of ethylene. The mechanism for the titanosilicate-catalyzed hydration of ethylene oxide (EO) has also been considered, which then shed light on the active sites for the second step in the oxidative hydration of ethylene. The catalyst deactivation could be ascribed to that the target product of EG and other heavy by-products with high boiling points were deposited inside the channels. The used Ti-MWW can be reusable when subjected to the regeneration by hightemperature calcination. Amine-assisted structural rearrangement of Ti-MWW not only enhanced the catalytic activity but also improved its stability in the oxidative hydration of ethylene. Ó 2017 Elsevier Inc. All rights reserved.

1. Introduction Ethylene glycol (EG) as the simplest diol is widely used as engine coolants, antifreezes as well as raw material for the manufacture of polyesters, cosmetics, and other down-stream products [1]. The global demand for EG had reached 25 million tons in 2014 and continues to grow at an annual increasing rate of around 5% [2]. The non-catalytic hydration of ethylene oxide (EO) to EG at elevated temperatures (423–493 K) still accounts for the major market share, since it was developed in 1937 by Union Carbide Corporation (UCC) [3,4]. However, large excess of water is required in most cases to inhibit the generation of the by-products of diethylene glycol (DEG) and tri-ethylene glycol (TEG) that are produced from self-condensation of EG, even though various homogeneous or heterogeneous catalysts have been developed over the past few decades for the selective hydration of EO, such as quaternary phosphonium halides [5], macrocyclic chelating compounds [6], polymeric organosilane ammonium salts [7], cation- and anion-exchange resins [8–11], supported metal oxides [12–15], and zeolites [12]. Recently, Yang et al. made a breakthrough in developing an efficient immobilized homogeneous catalyst with ⇑ Corresponding authors. E-mail addresses: [email protected] (H. Xu), [email protected] (P. Wu). https://doi.org/10.1016/j.jcat.2017.12.002 0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

high activity and EG selectivity for the hydration of EO at almost a stoichiometric H2O/EO molar ratio (as low as 2) and mild reaction temperature (313 K), wherein CoIII (salen) was encapsulated in FDU-12 mesosilica-based nanoreactors [13]. Additionally, Li et al. also designed a nanoreactor with active Sn sites confined in Chabazite cages using SSZ-13 zeolite as the EO hydration catalyst [14], which exhibited similar catalytic performance with FDU-12[CoIII(salen)] catalyst. Tandem catalysis, that enables multistep reactions to take place in one-pot, holds great potential for increasing the efficiency of chemical synthesis [15]. The direct synthesis of EG from ethylene and H2O2 through oxidative hydration, which formally involves a combination of ethylene epoxidation and subsequent EO hydration, could be a more economically viable process because it operates with the low-cost and easily available feedstock of alkenes, and requires no separation of epoxide after the first step. Titanosilicate containing tetrahedral Ti species in the framework is well known as efficient heterogeneous catalysts in the epoxidation of various alkenes [16–18]. It is a general phenomenon that the alkene epoxidation is accompanied with the ring-opening of epoxide in the presence of a protic solvent (methanol or water, etc.), due to the acidity related to titanosilicates and/or from catalytic systems [16,17,19–21]. Although some reports have made the attempts at direct synthesis of EG from ethylene and H2O2 over titanosilicates like TS-1 and Al-TS-1 [21], no substantial progress

90

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

has been made due to the difficulty of meeting the demand of high EG selectivity and H2O2 utilization simultaneously. Hence, it is highly desirable to develop a more suitable titanosilicate for clean production of EG through oxidative hydration of ethylene. Recently, we reported that MWW-type titanosilicate (Ti-MWW) was a more effective catalyst for the selective oxidation of ethylene to EO with H2O2 in comparison to other titanosilicates (TS-1, TiMOR and Ti-Beta) [16]. Moreover, Ti-MWW showed superior reactivity and selectivity in various oxidation reactions [18], including alkene epoxidation, ketones or aldehydes ammoximation, amines and sulfides oxidation. With the special lamellar structure, TiMWW reactivity could be improved by structural modifications including swelling [22], delamination [22,23], pillaring and silylation [24–26]. Additionally, it has been demonstrated that the reactivity of Ti-MWW in the alkenes epoxidation can be further enhanced by the fluorination treatment [17,27,28], or structural rearrangement [29,30]. Ti-MWW is thus considered to be a suitable candidate for the oxidative hydration of ethylene to ethylene glycol. In this present study, we applied Ti-MWW to the oxidative hydration of ethylene with the aim of producing EG efficiently and selectively. The effect of solvent on the oxidative hydration of ethylene has been investigated in detail by comparing TiMWW with conventional TS-1. The catalytic performance of other titanosilicates was also evaluated and compared with Ti-MWW. The effect of the reaction conditions on the oxidative hydration of ethylene over Ti-MWW were investigated in detail. In addition, the Ti-MWW catalyst was modified by structural rearrangement with the purpose to enhance the catalytic performance in oxidative hydration of ethylene.

2. Experimental 2.1. Catalyst preparation According to the procedures reported previously [31], Ti-containing MWW precursors with different Si/Ti ratios were hydrothermally synthesized, wherein boric acid and piperidine (PI) were used as crystallization-supporting agent and structuredirecting agent (SDA), respectively. The synthetic gels with molar compositions of 1.0 SiO2: 0.01–0.04 TiO2: 1.4 PI: 0.67 B2O3: 19 H2O were crystallized at 443 K for 7 days in a Teflon-lined stainless-steel autoclave under tumbling conditions. After crystallization, the obtained precursor was refluxed in 2 M HNO3 solution with the purpose of removing extra-framework Ti species as well as part of framework boron. The acid-treated products were filtered, washed with deionized water and dried at 353 K overnight. Subsequently, the sample was calcined in static air at 823 K for 6 h with a ramp rate of 1 K min
1, denoted as [Ti, B]-MWW. B-MWW was prepared in the same way as Ti-MWW lamellar precursors without the introduction of tetrabutyl orthotitanate (TBOT). In order to remove the framework B3+ ions, both [Ti, B]-MWW and B-MWW zeolites were then subjected to hydrothermal treatment in H2O at 373 K for 8 h and subsequent calcination in static air at 823 K for 6 h with a ramp rate of 1 K min
1. The obtained products almost free of B were denoted as Ti-MWW and deB-MWW, respectively. For control experiments, MWW-type all silica ITQ-1 (Si/Al = 1) [32], H-MCM-22 aluminosilicate (Si/Al = 18) [33], and other titanosilicates of Ti-MCM-68 [34], TS-1 [35], and Ti-MOR [36], were also prepared according to the established literature. In order to adjust the solid acidity of Ti-MWW, sodium ionexchange was performed in 1 M aqueous NaNO3 or NaOAc solution with a solid-to-liquid weight ratio of 1:100 at room temperature for 24 h. Then, the obtained sample was filtered, washed with deionized water until the Na+ amount in the filtrate is lower than

10 ppm, and dried at 353 K overnight to get the corresponding sample of NaNO3-Ti-MWW or NaOAc-Ti-MWW. The structural rearrangement treatment for Ti-MWW was performed with the composition of SiO2: 0.4 PI: 10 H2O at 443 K for 1 day under rotation condition (10 rpm) [29,30]. The treated product was washed with deionized water, dried at 393 K overnight, and calcined in static air at 823 K for 6 h with a ramp rate of 1 K min
1, denoted as ReTi-MWW. The adsorption of EG was carried out by vigorous stirring the mixture of 0.5 g of Ti-MWW or H-MCM-22 and 50 g of 1 M EG aqueous solution at 303 K for 24 h. The EG absorbed samples were separated from the above-mentioned mixture and dried at 353 K for 12 h, denoted as EG/Ti-MWW and EG/H-MCM-22, respectively. 2.2. Characterization methods The crystalline structures of various titanosilicates were confirmed by X-ray diffraction (XRD) on a Rigaku Ultima IV diffractometer using Ni-filtered Cu Ka radiation. Scanning electron microscopy (SEM) images were collected on a Hitachi S-4800 microscope to show the crystal morphologies. The amounts of Na, Si, Ti, Al and B were determined by inductively coupled plasma (ICP) on a Thermo IRIS Intrepid II XSP after the samples were dissolved in aqueous HF solution. The coordination states of Ti species were examined by UV–visible spectroscopy on a Shimadzu 2700PC spectrophotometer using BaSO4 as a reference. Infrared spectra were collected by a Nicolet Nexus 670 FT-IR spectrometer at a resolution of 2 cm
1. The spectra were obtained using self-supported wafers with a diameter of 2 cm (10 mg cm
2 thickness), except that the spectra in the region of framework vibration (500–1300 cm
1) were recorded using a KBr pellet technique (3 wt% diluted in KBr). In order to avoid the influence of absorbed water, the self-supported wafer or the KBr diluted wafer placed in a quartz IR cell sealed with CaF2 or KBr windows, respectively, was evacuated at 723 K for 2 h before measurement. For pyridine spectra measurement, the pretreated self-supported wafer was exposed to a pyridine vapor at 298 K for 20 min, and then the physisorbed and chemisorbed pyridine was then removed by evacuation at different temperature (323–523 K) for 0.5 h. In addition, the in-situ IR analysis of EG/Ti-MWW or EG/H-MCM-22 was carried out by evacuation at different temperature (373–473 K) for 5 min. Thermogravimetric (TG) analysis was carried out in a Netzsch Sta 4049 F3 apparatus in air with a heating rate of 10 K min
1 in the temperature range of 300–1073 K. 2.3. Catalytic reactions The oxidative hydration of ethylene to EG with hydrogen peroxide was carried out under vigorous stirring in an autoclave reactor equipped with a telfon-inner. In a typical run, 0.1 g titanosilicate, 10 g H2O, and 10 mmol H2O2 (30 wt%) were added into the reactor, and then ethylene was charged into the autoclave to replace the air inside three times, reaching a constant reaction pressure at of 2.5 MPa at 313 K. For the hydration of EO, it was also proceeded in the above-mentioned autoclave reactor. A mixture of catalyst, EO, H2O and H2O2 (if added) was vigorously stirred at 313 K under N2 pressure of 1.5 MPa. After specified reaction time, the reactor was cooled down with ice water to stop the reaction and depressurized slowly before opening. The amount of unconverted H2O2 was determined by standard titration method with 0.05 M Ce(SO4)2 solution. The products were separated from the reaction mixture by centrifugation and determined by a gas chromatograph (Shimadzu 2014) equipped with an RtxÒ-Wax capillary column (30 m  0.25 mm  0.25 lm) and FID detector using isopropanol as internal standard. The products formed were identified by a GC-MS (Agilent 6890 series GC system,

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

5937 network mass selective detector). The results of the oxidative hydration of ethylene were calculated as follows.

H2 O2 conversion ¼

moles of H2 O2 consumed  100% moles of H2 O2 in the feed

Product selectivity ¼

EG yield ¼

moles of a defined product  100% moles of products produced

moles of EG produced  100% moles of H2 O2 in the feed

H2 O2 efficiency ¼

moles of products produced  100% moles of H2 O2 consumed

3. Results and discussion 3.1. Characterization of various titanosilicate catalysts The XRD patterns confirmed that all titanosilicates hydrothermally synthesized or postsynthesized possessed the expected structures, and possessed a high crystallinity (Fig. 1 and Fig. S1). The SEM images further affirmed the four titanosilicates with different topologies contained no impurities or amorphous phase (Fig. 2). Ti-MCM-68, TS-1 and Ti-MOR crystals were aggregates of nanocrystals while [Ti, B]-MWW crystals showed a typical platelet-shaped morphology with the thickness of 50 nm. As shown in Fig. S2, all the titanosilicates contained essentially isolated Ti4+ species in the framework evidenced by the band at 220 nm and 960 cm
1 in UV–vis and IR spectra, respectively [37,38]. The band at 330 nm was hardly observed in UV–vis spectra, indicating anatase TiO2 was absent. In order to rule out possible negative effect of B species on catalytic performance of tetrahedral Ti ions, a further deboronation was carried out by hydrothermal treatment, and then the Si/B ratio increased greatly from 73 for [Ti, B]-MWW to 419 for Ti-MWW, while the Si/Ti ratio remained unchanged (Table 1). Furthermore, no significant difference was found between [Ti, B]-MWW and Ti-MWW either in XRD patterns (Fig. 1) or in UV–vis and IR spectra (Fig. S2). Additionally, the Ti-MWW samples with various Si/Ti ratios contained only isolated Ti species in the framework as evidenced by the predominant band at 220 nm in the UV–visible spectra, and the band intensity increased with increasing Ti content (Fig. 3A). The acidic properties of these catalysts were determined by the FT-IR spectra of adsorbed pyridine. As shown in Fig. S3A, pyridine

Intensity (a.u.)

e

c b a 10

15

20

25

30

exhibited the vibrations of stretching modes originating from different adsorption sites on ITQ-1; they are hydrogen-bonded pyridine (hb-Py) at 1598 cm
1 (mode 8a) and 1446 cm
1 (mode 19b) and physically adsorbed pyridine (ph-Py) at 1581 cm
1 (mode 8b) and 1437 cm
1 (mode 19b) [39]. The bands arising from phPy nearly diminished after desorption at 373 K. The bands arising from hb-Py related to the silanol groups (internal silanols of Si– OH/Si–OH, isolated external silanols and hydrogen-bonded silanol nests) (Scheme 1), were more resistant against evacuation, but decreased in intensity with rising desorption temperature, and disappeared totally at 473 K. In the case of B-MWW (Fig. S3B), new bands emerged at 1626, 1545 and 1490 cm
1. The bands at 1626 and 1545 cm
1 are assigned to pyridine adsorbed on Lewis acid sites and Brönsted acid sites, respectively, while the peak at 1490 cm
1 appears commonly for pyridine adsorbed on both Brönsted acid sites and Lewis acid sites [40]. These bands decreased in intensity with rising desorption temperature but they were still observable even after desorption at 473 K, and then disappeared totally at 523 K. It is deduced that B-MWW possessed stronger acidity than ITQ-1. As shown in Fig. S3C, deB-MWW (Si/ B = 420) showed similar Py-adsorption IR spectra to ITQ-1. The absence of the bands related to the Brönsted acid sites and Lewis acid sites in deB-MWW confirmed that the B species were completely removed from the framework after the acid-treatment of B-MWW. In the spectra of Ti-MWW (Fig. S3D), it showed no band corresponding to the Brönsted acid sites, but two new bands belonging to pyridine adsorbed on Lewis acid sites at 1606 and 1490 cm
1 in addition to similar bands shown by ITQ-1 and deBMWW. These new bands were more evacuation resistant, and remained even after desorption at 523 K, indicating that stronger acid sites existed in Ti-MWW in comparison to B-MWW. Meanwhile, it should be noted that the 1598 and 1446 cm
1 bands assigned to hydrogen-bonded pyridine were still visible in the Py-adsorption IR spectrum of Ti-MWW after desorption at 523 K. Thus, the Lewis acid sites and the Si-OH/Ti-OH groups, generated by the introduction of Ti species into the matrix of silicate, possess stronger acidity in comparison to other silanol groups (Si-OH/SiOH, isolated external silanols and hydrogen-bonded silanol nests) (Scheme 1). Pyridine adsorption was further carried out on the Ti-MWW catalysts with various Ti contents (Fig. 3B). After Pydesorption at 473 K for 0.5 h, the bands at 1606, 1490 and 1446 cm
1, except for Ti-free deB-MWW material, were observed for all Ti-MWW catalysts and the band intensity increased with increasing Ti content, indicating that these acid sites are closely related to the framework Ti species. In contrast to MWW-type titanosilicate, borosilicate and silicate, H-MCM-22 aluminosilicate showed pyridine adsorption bands characteristic of Brönsted and Lewis acidity which remained a very high intensity even after desorption at 523 K (Fig. S3E). Based on the above results, the acidity strength and amounts of various MWW-type zeolites could be determined as H-MCM-22  Ti-MWW > B-MWW > deB-MWW  ITQ-1. 3.2. Oxidative hydration of ethylene

d

5

91

35

2 Theta (deg.) Fig. 1. XRD patterns of [Ti, B]-MWW (a), Ti-MWW (b), Ti-MCM-68 (c), TS-1 (d) and Ti-MOR (e).

3.2.1. A comparison among various titanosilicates Significant solvent effects are generally involved because the character of solvents may have a great influence on the intrinsic activity of Ti species as well as the product distribution. H2O always exists in the Ti-zeolite/H2O2 catalytic system, either from aqueous solution of H2O2 oxidant or from its decomposition. As shown in Table 2, for [Ti, B]-MWW and TS-1, the most favorable solvent was water, where the highest EG selectivity and EG yield were achieved in comparison to other solvents, indicating that a certain amount of water was necessary to proceed the oxidative hydration of ethylene with H2O2. The direct synthesis of EG from

92

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

a

b

1 μm

1 μm

c

d

1 μm

1 μm

Fig. 2. SEM images of [Ti, B]-MWW (a), Ti-MCM-68 (b), TS-1 (c) and Ti-MOR (d).

Table 1 A comparison of the oxidative hydration of ethylene over various titanosilicates.a. Entry

1 2 3 4 5 a b c d

Catalyst

Structure

[Ti, B]-MWW Ti-MWWd Ti-MCM-68 TS-1 Ti-MOR

MWW MWW MSE MFI MOR

Si/Tib

Si/Bb

49 50 42 50 51

Si/Alb

1 1 102 1 110

73 419 1 1 1

Products distribution (%)c

H2O2 (%)

EG yield (%)

conv.

eff.

EO

EG

Others

82.7 87.9 13.3 57.1 1.9

66.3 68.8 37.6 44.9 36.1

56.4 60.0 88.0 27.0 100.0

40.1 36.7 12.0 69.2 0.0

3.5 3.3 0.0 3.8 0.0

22.0 22.2 0.6 17.7 0.0

Reaction conditions: cat., 0.1 g; ethylene, 2.5 MPa; H2O2, 10 mmol; H2O, 10 mL; temp., 313 K; time, 2 h. Molar ratio determined by ICP analysis. EO, ethylene oxide; EG, ethylene glycol. Prepared by hydrothermal deboronation: water treatment at 373 K for 8 h, and then calcination in air at 823 K for 6 h.

e

1446

A 1490

Absorbance (a.u.)

c Absorbance (a.u.)

B

1606

d

b

e d c b

a

200

a 300

400

Wavelength (nm)

500

1600

1550

1500

1450

Wavenumber (cm-1)

Fig. 3. UV–visible spectra (A) and pyridine-adsorption IR spectra after evacuation at 473 K for 0.5 h (B) of various Ti-MWW catalysts with a Si/Ti ratio of 1 (a), 160 (b), 84 (c), 63 (d) and 50 (e).

93

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

Scheme 1. Possible acid sites in titanosilicate/H2O2 systems.

Table 2 A comparison of oxidative hydration of ethylene over [Ti, B]-MWW and TS-1 in different solvents.a Entry

a b

Catalyst

Solvent

H2O2 (%)

Products distribution

b

(%)

EG yield (%)

conv.

eff.

EO

EG

EGMME

Others

1 2 3 4 5

[Ti, B]-MWW

H2O Acetone MeOH t-BuOH MeCN

82.7 91.1 3.8 66.0 43.6

66.3 77.8 87.9 53.4 80.5

56.4 80.9 61.4 91.2 98.2

40.1 12.1 0.0 6.8 1.7

– – 38.6 – –

3.5 7.1 – 2.0 0.1

22.0 8.5 – 2.4 0.7

6 7 8 9 10

TS-1

H2O Acetone MeOH t-BuOH MeCN

57.1 73.8 90.4 5.2 6.4

44.9 69.9 86.0 70.4 75.6

27.0 55.7 17.3 70.6 99.6

69.2 13.8 5.2 0.0 0.4

– – 77.5 – –

3.8 30.5 – 29.4 –

17.7 7.1 4.1 – –

Reaction conditions: cat., 0.1 g; ethylene, 2.5 MPa; H2O2, 10 mmol; solvent, 10 mL; temp., 313 K; time, 2 h. EO, ethylene oxide; EG, ethylene glycol; EGMME, ethylene glycol mono-methyl ether; others, di-ethylene glycol (DEG), tri-ethylene glycol (TEG).

ethylene and H2O2 through oxidative hydration, which formally involves a combination of ethylene epoxidation and subsequent EO hydration. The high reactivity in water is possibly due to the readily activate the H2O2 to form Ti-OOH groups, even though the solubility of ethylene in water is the lowest in comparison to that in other organic solvents due to the hydrophobicity of ethylene. On the other hand, the EG yield and the EG selectivity increased with increasing the H2O/EO ratio (Fig. S4), indicating that a certain amount of water was necessary to proceed the hydration of EO with H2O. Hence, water would be the best solvent to get high EG yield in the oxidative hydration of ethylene. Ethylene glycol mono-methyl ether (EGMME) was by-produced in the reaction system with protic solvent of MeOH due to the solvolysis of EO. The EGMME selectivity reached 38.6% and 77.5% for [Ti, B]MWW and TS-1, respectively (Table 2, Entries 3 and 8). Additionally, in the investigated solvents except MeOH, [Ti, B]-MWW

showed a higher reactivity for the oxidative hydration of ethylene than TS-1. Considering the facts that H2O was definitely necessary for the hydration of EO and showed better performance in particular EG selectivity than other solvents, it was used as the solvent in the following experiments. Various titanosilicates with different structure topologies and compositions were evaluated in the oxidative hydration of ethylene with H2O2 (Table 1). Ti-MOR and Ti-MCM-68 showed a much lower activity than titanosilicates with MWW or MFI structure topology. [Ti, B]-MWW and Ti-MWW, with the same structure topology but distinct Si/B ratio compositions, showed the same EG yield, which is higher than TS-1 catalyst. The coexisting framework boron species would increase the electronegativity of zeolite framework, that may lead to a negative effect on the catalytic oxidative performance of Ti species [16,41]. Thus, Ti-MWW almost free of boron species showed a higher H2O2 conversion (87.9%) and

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

utilization efficiency (68.8%) than that of [Ti, B]-MWW (82.7% and 66.3% for H2O2 conversion and utilization efficiency, respectively). Hence, B-free Ti-MWW/H2O2/H2O was considered as the most suitable reaction system with the highest EG yield together with high H2O2 conversion and utilization efficiency for the oxidative hydration of ethylene. 3.2.2. Hydration of ethylene oxide There is no doubt that the first step of ethylene oxidative hydration, that is epoxidation, is catalyzed by the Ti-OOH groups. The real active site for the hydration of EO to EG is still controversial, because the alkenes epoxidation is often accompanied with the hydration or the solvolysis of epoxides [16,17,19–21]. To further explore the reaction mechanism of the oxidative hydration of ethylene especially the second step, the EO hydration was carried out over MWW-type zeolites containing different heteroatoms (Fig. 4). The EG yield decreased in the order of Ti-MWW > BMWW > deB-MWW  ITQ-1, in accordance with the strength of acidity mentioned above in Section 3.1. A very low EG yield below 5%, comparable to that of blank trial, was observed for ITQ-1 and deB-MWW silicates, indicating that the acidity of the Si-OH/SiOH, isolated external silanols and hydrogen-bonded silanol nests in the all-silica zeolites are too weak to promote the hydration of EO. In contrast, B-MWW exhibited a considerable EG yield of 60.1%. Consistent with higher EG selectivity achieved by [Ti, B]MWW (Table 1), it showed a higher ability for converting EO to EG in comparison to Ti-MWW, which could be ascribed to the boron species existing in the framework. Then, one may speculate that the zeolites with strong Brönsted acid sites would be useful for EO hydration. However, it is interesting that H-MCM-22 showed a lower EG yield in comparison to B-MWW and Ti-MWW (Fig. 4), although H-MCM-22 reveals the strongest acidity among the three catalysts. In order to clarify this phenomenon, we have compared the time course of EO hydration between Ti-MWW and H-MCM22 (Fig. 5). At the beginning of this reaction, H-MCM-22 showed a higher EG yield than Ti-MWW, possibly due to its stronger acidity. The EG yield over Ti-MWW gradually increased from 15.9% to 91.1%, when the reaction time was prolonged from 1 h to 10 h. In contrast, the EG yield over H-MCM-22 remained unchanged, indicating the extremely fast deactivation of H-MCM-22 after 1 h. In order to illustrate the significantly different resistance to the deac-

EG yield

100

EG Sel.

Percentage (%)

80 60 40 20 0

nk

Bla

-1 W W W W -22 ITQ -MW -MW ]-MW i-MW CM B M B T B H de [Ti,

Fig. 4. A comparison of hydration of ethylene oxide (EO) over different catalysts. Reaction condition: cat. (if added), 0.1 g; EO, 10 mmol; H2O, 0.15 mol; N2, 1.5 MPa; temp., 313 K; time, 8 h. [Ti, B]-MWW (Si/Ti = 49, Si/B = 73) and Ti-MWW (Si/Ti = 50, Si/B = 419).

100

80

EG yield (%)

94

30

60 20

a

10

40

0 0.0 0.5 1.0 1.5 2.0

20

b 0

0

2

4

6

8

10

Reaction time (h) Fig. 5. Dependence of EG yield with time on stream of Ti-MWW (Si/Ti = 50, Si/B = 419) (a) and H-MCM-22 (b). Reaction conditions: cat., 0.1 g; EO, 10 mmol; H2O, 0.15 mol; N2, 1.5 MPa; temp., 313 K. Inset shows the initial reaction stage.

tivation in the hydration of EO between Ti-MWW and H-MCM-22, the stability of EG adsorbed in Ti-MWW and H-MCM-22 were compared by TG analysis and in-situ IR. As shown in Fig. S5, EG/TiMWW and EG/H-MCM-22 both showed three stages of weight loss in the temperature range 298–1000 K. In the case of EG/Ti-MWW, the first stage (298–400 K) and the second stage (400–500 K) correspond to the removal of the physically adsorbed water and the absorbed EG, respectively, while the third stage (500–1000 K) can be ascribed to the condensation of two adjacent Si-OH groups. With respect to EG/H-MCM-22, it showed a wider temperature range (400–600 K) at the second stage, indicating that the removal of adsorbed EG from H-MCM-22 are more difficult in comparison to Ti-MWW. As shown in Fig. S6A, two bands at 2885 and 2947 cm
1 were observed in the spectrum of EG/Ti-MWW, which are assigned to symmetrical and asymmetrical stretching CH2 groups, respectively [42]. These bands should be closely related to EG and decreased in intensity with rising desorption temperature. In the case of EG/H-MCM-22, these bands shift to lower frequencies of 2860 and 2924 cm
1 which are more resistant against evacuation, indicative of a stronger interaction between EG and H-MCM-22 (Fig. S6B). Hence, it was reasonable that H-MCM-22 suffered a faster deactivation than Ti-MWW, because the product of EG or poly glycol molecules in products couldn’t desorb easily from H-MCM22 and then possibly caused a carbon deposition on the Ti active sites. The kinetics was also studied for the EO hydration over TiMWW catalysts with various Ti contents. As depicted in Fig. 6A, the EG concentration (CEG) shows a linear dependence on the reaction time within 1 h. The initial reaction rates r0 thus determined by the CEG versus reaction time shows a linear dependence on the Ti content (Fig. 6B). This strongly implies that there is a close relationship between the catalytic activity in the EO hydration and the Ti active species, which are probably the Si–OH/Ti–OH groups and the Lewis acid sites both related to the framework Ti species. To further confirm the contribution of the Si–OH/Ti–OH groups and the Lewis acid sites to the hydration of EO on Ti-MWW, we firstly removed the hydroxyl-related active sites of Si–OH/Ti–OH groups via Na+ ion-exchange. As shown in Table 3, the ionexchange with basic solution of NaOAc resulted in a much higher Na/Ti ratio in comparison to NaNO3, indicating that the exchange of sodium ions for the protons in Ti-MWW occurred more easily at high pH, meanwhile the Si/Ti and Si/B ratios remained

95

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

10

A

c

300 200

-1

-1

d

400

8 6

r0 (mmol L min )

500

CEG (mmol L )

B

e

-1

600

b

4 2

100

a 0

0

0

10

20

30

40

50

60

0

5

Time (min)

10

15

20

Ti/(Ti+Si) molar ratio *10

25 -3

Fig. 6. Kinetic of EO hydration (A) and dependence of initial reaction rate (B) over Ti-MWW with a Si/Ti ratio of 1 (a), 160 (b), 84 (c), 63 (d) and 50 (e). Reaction conditions: cat., 0.1 g; EO, 10 mmol; H2O, 0.15 mol; N2, 1.5 MPa; temp., 313 K.

Table 3 The results of EO hydration over Ti-MWW after ion exchange with Na+.

c d

Catalyst

Si/Tia

Si/Ba

Na/Tia

EG yield

1 2 3 4

Blank Ti-MWWc NaNO3-Ti-MWWd NaOAc-Ti-MWWd

– 63 64 63

– 417 412 412

– – 0.08 0.48

4.1 64.2 35.4 15.6

b

(%)

Determined by ICP analysis. Hydration of EO. Reaction conditions: cat., 0.1 g; EO, 10 mmol; H2O, 0.15 mol; N2, 1.5 MPa; temp., 313 K; time, 8 h. Ti-MWW: Si/Ti = 63 and Si/B = 417 The ion exchange was carried out in 1 M NaNO3 or NaOAc aqueous solution with a solid to liquid weight ratio of 1:100 at room temperature.

hydroxyl-related active sites caused an apparent decrease in catalytic activity. With the most complete ion-exchange with basic NaOAc solution, the EG yield decreased to 15.6%, much lower than the parent Ti-MWW zeolite (64.2%), which is ascribed to a removal of a large amount of the acid sites related to the Si-OH/Ti-OH groups. Nevertheless, NaOAc-Ti-MWW was still active in the hydration of EO, indicating that the Lewis acidity also contributed to the hydration of EO. Hence, both the Si–OH/Ti–OH groups

3745

Absorbance (a.u.)

3720

3520

c b

0.4

a 4000

3750

3500

3250

3000

60

b

Wavenumber (cm-1)

unchanged. Fig. 7 showed the IR spectra in the region of hydroxyl stretching vibration for Ti-MWW before and after Na+ ionexchange. The bands at 3745, 3720 and 3520 cm
1 are attributed to isolated external silanols, internal silanols and hydrogenbonded silanol nests, respectively [28]. Compared with the parent Ti-MWW, the bands at 3720 and 3520 cm
1 decreased in intensity apparently after ion-exchange with NaNO3, whereas the 3745 cm
1 band remained almost unchanged. On the other hand, the ion-exchange with NaOAc led to almost vanishment of the two bands attributed to internal silanols and hydrogen-bonded silanol nests, indicating the completely removal of the acidity contributed by the hydroxyl groups. As shown in Table 3, the removal of

0.3

EG yield (%)

Fig. 7. IR spectra in the hydroxyl stretching region of Ti-MWW (Si/Ti = 63, Si/B = 417) (a), as a after NaNO3 treatment (b), or as NaOAc treatment (c). The spectra were recorded after in situ evacuation at 723 K for 1.5 h.

45

a 0.2

30

c 15

H2O2 conv. (mmol)

a b

Entry.

0.1 0

0

5

10

15

20

25

-3

Ti/(Ti+Si) molar ratio*10

Fig. 8. Change of the EG yield with the Ti content in Ti-MWW without (a) or with (b) H2O2 at various Ti content for the hydration of EO. H2O2 conversion in Ti-MWW/ H2O2 system at various Ti content for the hydration of EO (c). Reaction conditions: cat., 0.1 g; EO, 10 mmol; H2O, 0.15 mol; H2O2 (if added), 10 mmol; N2, 1.5 MPa; temp., 313 K; time, 4 h. The symbol 0 at the Ti/(Ti + Si) molar ratio represent de-BMWW.

96

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

(48.6%) and Lewis acid sites (15.6%) contributed to the hydration of EO, and the former one was supposed to be the main active sites. In addition, the hydration of EO was carried out over Ti-MWW with various Ti contents (Fig. 8). For the system without H2O2, Si– OH/Ti–OH and Lewis acidity of framework Ti species were responsible for the catalytic hydrolysis of EO. Once H2O2 was added into this reaction system, the Si–OH/Ti–OH groups should be changed to Ti-OOH groups. As shown in Fig. 8a, the EG yield remained unchanged over de-B-MWW without Ti species when H2O2 was added into the reaction system, indicating that the H2O2 are inactive for the EO hydration. In contrast, the EG yield over Ti-MWW increased when H2O2 was added into the reaction system (Fig. 8b). Hence, the increment of EG yield after the addition of H2O2 is ascribed to Ti-OOH groups. Moreover, the replace of SiOH/Ti-OH active species by Ti-OOH in the H2O2-containing system favored the hydration of EO, indicating that the generated Ti-OOH showed stronger acidity than Si–OH/Ti–OH. It can be deduced that the acidity form Ti-OOH and Lewis acid sites of framework Ti species were responsible for catalyzing the second step of oxidative hydration of ethylene in the Ti-MWW/H2O2/H2O reaction system.

3.2.3. Effect of reaction parameters on the oxidative hydration of ethylene over Ti-MWW As shown in Scheme 2, the oxidative hydration of ethylene over Ti-MWW in H2O solvent is a tandem reaction with ethylene firstly epoxidized to EO over Ti-OOH groups followed by the hydration of EO to EG on Ti related acidic sites. By-products of DEG and TEG due to self-condensation of EG can also be formed in this reaction system. To further optimize the oxidative hydration of ethylene over Ti-MWW, the reaction parameters were studied in detail. Fig. 9 showed that the oxidative hydration of ethylene depended greatly on the Ti contents of Ti-MWW catalysts. With the amount of Ti active sites increased, the H2O2 conversion increased and the H2O2 utilization efficiency was maintained at a steady level of 70%. A higher EG selectivity could be obtained over Ti-MWW with higher Ti content, while the EO selectivity changed in a tendency opposite to EG selectivity. The by-products (DEG and TEG) selectivity could be controlled to be below 5%. Fig. 10a and e shows the time course for the oxidative hydration carried out over Ti-MWW in water. The EG selectivity increased rapidly at the beginning of the reaction, and then slowed down as H2O2 was consumed gradually. The EG selectivity finally reached >92% at 3 h. On the other hand, the H2O2 utilization efficiency almost maintained unchanged during the whole oxidative hydration process. A significant temperature effect on the performance of the oxidative hydration of ethylene was observed when the reaction was carried out in the temperature range of 303–343 K (Fig. 10b and f). With the reaction temperature increasing from 303 to 343 K, the H2O2 conversion increased. The EO selectivity

decreased while the EG selectivity increased with increasing temperature, indicating that higher temperature favored the hydration of EO in comparison to the epoxidation of ethylene. However, the H2O2 utilization efficiency decreased with the increasing reaction temperature, probably due to the nonproductive decomposition of H2O2 at higher temperature. Fig. 10c and g shows the effect of catalyst amount on the performance of the oxidative hydration of ethylene in Ti-MWW/H2O2/H2O system. The H2O2 conversion increased rapidly with increasing the catalyst amount, and H2O2 were almost completely consumed when the catalyst amount was up to 0.15 g, while the H2O2 utilization efficiency was maintained at about 70%. The EO selectivity decreased from 82.2% to 7.6% with increasing the catalyst amount from 0.05 to 0.3 g while the EG selectivity increased as a result of successive consumption of EO. As shown in Fig. 10d and h, the H2O/H2O2 ratios play a significant role on the performance of the oxidative hydration of ethylene. The H2O2 conversion increased rapidly with increasing the H2O/H2O2 ratio when the H2O/H2O2 ratio was below 27, and then slowed as a low H2O2 concentration would decrease the efficiency of ethylene epoxidation. On the other hand, the H2O/H2O2 ratio as high as 27 is preferred for the purpose of achieving a relatively high H2O2 utilization efficiency. Moreover, increasing the H2O/H2O2 ratio not only resulted in an increase in EG selectivity but also inhibited the further condensation of EG to DEG and TEG. Additionally, the effect of ethylene pressure on the EG yield has also been investigated. As shown in Fig. S7, the EG yield increased with increasing the ethylene pressure, possibly due to the higher ethylene solubility in water under the higher ethylene pressure.

3.2.4. Stability and reusability of Ti-MWW The stability and reusability of heterogeneous catalysts was essential for actual application in industrial processes. We thus checked the catalytic cycles of Ti-MWW in the oxidative hydration of ethylene (Fig. 11). The used Ti-MWW was regenerated by washing with acetone or by further calcination at 823 K for 6 h, and then it was subjected to repeated oxidative hydration of ethylene at a constant ratio of catalyst-oxidant-solvent. The EG yield decreased by half at seventh reuse, when the used Ti-MWW catalyst was only washed with acetone and then dried, with the EG selectivity >35% during the whole reaction-regeneration cycles. Nevertheless, the EG yield was restored (Fig. 11, the eighth and ninth reuses), when the reused catalyst was regenerated by washing with acetone and further calcination at 823 K for 6 h. As shown in Fig. S8, no significant difference was found in the XRD patterns, the UV–vis spectra and the IR spectra between the fresh Ti-MWW and the regenerated one. This suggests that the deactivation of acetone-washed catalyst was mainly due to the deposition inside the channels with the heavy organic compounds with high boiling points, like EG molecules and DEG and TEG by-products.

Scheme 2. Reaction pathways in the oxidative hydration of ethylene. The ‘‘Ti species” in the EO hydration step indicates both the Ti-OOH groups and Lewis acid sites of the framework Ti species.

97

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

100

100

B

80

EO, EG & others sel. (%)

H2O2 conv. & eff. (%)

A H2O2 eff.

60 40

H2O2 conv.

20

EO

80 60 40

EG

20

Others 0

0

5

10

15

20

0

25

5

-3

10

15

20

25 -3

Ti/(Ti+Si) molar ratio *10

Ti/(Ti+Si) molar ratio *10

Fig. 9. Dependence of H2O2 conversion and utilization efficiency (A) and products selectivity (B) on the Ti content of Ti-MWW. Reaction conditions: cat., 0.1 g; ethylene, 2.5 MPa; H2O, 10 mL; H2O2, 10 mmol; temp., 313 K; time, 2 h.

H2O2 Conv.

H2O2 Conv. & Eff. (%)

100

80

60

40

b

a 1

2

Reaction time (h)

3 300

320

340

EO

e

d

c 0.1

0.2

0.3

EG

40

60

Others

g

f

20

Catalyst amount (g) H2O/H2O2( mol/mol)

Reaction temp. (K)

100

EO, EG & others Sel. (%)

H2O2 Eff.

h

80 60 40 20 0

1

2

Reaction time (h)

3 300

320

340

Reaction temp. (K)

0.1

0.2

0.3

20

40

60

Catalyst amount (g) H2O/H2O2 (mol/mol)

Fig. 10. Dependence of H2O2 conversion, utilization efficiency and products selectivity on reaction time (a, e), reaction temperature (b, f), the amount of Ti-MWW (c, g) and H2O/H2O2 molar ratio (d, h). Reaction conditions: Ti-MWW (Si/Ti = 50, Si/B = 419), 0.1 g; temp., 333 K (a, e), 313 K (c, d, g, h); others, see Fig. 9.

3.3. Catalyst modification by structural rearrangement It has been demonstrated that the hydration of EO can be catalyzed by the Si-OH/Ti-OH groups and the Lewis acid sites of the framework Ti species (Section 3.2.2). Recently, the hexacoordinated Ti active species of Ti(OSi)2(OH)2(H2O)2 with more Ti-OH could be produced by piperidine-assisted structural rearrangement [30]. The structurally rearranged Ti-MWW showed

higher catalytic activity in the alkene epoxidation and the ketone ammoximation than the parent Ti-MWW catalyst [29,30]. Thus, the structurally rearranged Ti-MWW was also expected to show higher catalytic activity in the oxidative hydration of ethylene. The structural rearrangement of Ti-MWW, including hydrothermal treatment in an aqueous solution containing piperidine (PI) at 443 K and further calcination at 823 K, triggered a structural change from 3D MWW to 2D lamellar precursor and then back to 3D

98

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

80 Ti-MWW

Re-Ti-MWW

Washed with acetone

Further calcined

EG yield (%)

60

40

20

0

0

1

2

3

4

5

6

7

8

9

Number of reuse Fig. 11. Changes of EG yield with the reaction-regeneration cycles on Ti-MWW (Si/ Ti = 50, Si/B = 419) and Re-Ti-MWW (Si/Ti = 50, Si/B = 417). Reaction conditions for the first run: cat., 0.15 g; ethylene, 2.5 MPa; H2O, 30 mL; H2O2, 30 mmol; temp., 333 K; time, 2 h. All the catalytic runs proceed at a constant ratio of catalystoxidant-solvent.

MWW, resulting in the Re-Ti-MWW catalyst. The treatment did not affect the amount of Ti species (Table S1), but made a new broad band emerge around 260 nm (Fig. S9), which is ascribed to hexa-coordinated Ti active species as Ti(OSi)2(OH)2(H2O)2 [30].

The FT-IR spectra of adsorbed pyridine were conducted to discern the effect of structure rearrangement on the acidic properties of the Ti-MWW catalysts (Fig. S10). The bands at 1606, 1598, 1490 and 1446 cm
1 were significantly increased by structural rearrangement in comparison to the parent Ti-MWW catalysts, indicating that the hexahedral Ti active species (Ti(OSi)2(OH)2(H2O)2) possessed higher acid content than the tetrahedral Ti species. With higher acid content, Re-Ti-MWW showed a higher EG yield than TiMWW in the hydration of EO (Table S1). The positive effect of structural rearrangement over Ti-MWW catalysts was also found in the oxidative hydration of ethylene, as the H2O2 conversion, the EG selectivity and the EG yield over Re-Ti-MWW were increased in comparison to Ti-MWW (Fig. 12). In addition, the EG yield over Re-Ti-MWW decreased by 20% while it decreased by 52% over Ti-MWW after 7 runs compared to their fresh samples, when the used catalysts were only washed with acetone and then dried (Fig. 11, the first to seventh reuses), which indicated the more resistance to deactivation for Re-Ti-MWW in the oxidation hydration of ethylene. Further calcination at 823 K for 6 h can also restore the activity (Fig. 11, the eighth and ninth reuses).

4. Conclusions Ti-MWW possesses a superior catalytic performance to other titanosilicates in the oxidative hydration of ethylene to EG with H2O as the solvent. In the Ti-MWW/H2O2/H2O reaction system, Ti-OOH groups are responsible for the oxidation of ethylene to

100

100

B 80

80

EO selectivity (%)

H 2 O 2 conversion (%)

A

60

60

40

20 40 0.4

0.8

1.2

1.6

0

2.0

0.4

1.2

1.6

2.0

70

100

C

D

60

80

50

EG yield (%)

EG selectivity (%)

0.8

Reaction time (h)

Reaction time (h)

60

40

40 30 20

20

0

10 0.4

0.8

1.2

1.6

Reaction time (h)

2.0

0

0.4

0.8

1.2

1.6

2.0

Reaction time (h)

Fig. 12. Comparison of H2O2 conversion (A), EO selectivity (B), EG selectivity (C) and EG yield (D) over Ti-MWW (Si/Ti = 50, Si/B = 419) (blue) and Re-Ti-MWW (Si/Ti = 50, Si/B = 417) (red) with time on stream. Reaction conditions: cat., 0.05 g; ethylene, 2.5 MPa; H2O, 10 mL; H2O2, 10 mmol; temp., 333 K.

X. Lu et al. / Journal of Catalysis 358 (2018) 89–99

EO while the Ti-OOH groups and Lewis acid sites of the framework Ti species prove to catalyze the hydration of EO to EG. The formation of heavy products such as EG, DEG and TEG may cause a reversible deactivation to the catalyst, while the deactivated catalyst is regenerated readily after removing the organic species by calcination. The structural rearrangement proves to be an effective way to increase the catalytic activity and the resistance of Ti-MWW to the deactivation in the oxidative hydration of ethylene. Acknowledgments The authors gratefully acknowledge financial support from the NSFC of China (21533002, 21373089, 21403069), and China Ministry of Science and Technology under contract of 2016YFA0202804. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.jcat.2017.12.002. References [1] S. Rebsdat, D. Mayer, Ethylene Glycol. In Ullmann’s Encyclopedia of Industrial Chemistry, seventh ed., Wiley-VCH, Weinheim, 2000. [2] J.F. Pang, M.Y. Zheng, R.Y. Sun, A.Q. Wang, X.D. Wang, T. Zhang, Green Chem. 18 (2016) 342. [3] T. Maihom, S. Namuangruk, T. Nanok, J. Limtrakul, J. Phys. Chem. C 112 (2008) 12914. [4] H.R. Yue, Y.J. Zhao, X.B. Ma, J.L. Gong, Chem. Soc. Rev. 41 (2012) 4218. [5] K. Kawabe, U.S. Patent 6080897, 2000. [6] E.M.G.A. van Kruchten, W.O. Patent 9923053, 1999. [7] E.M.G.A. van Kruchten, U.S. Patent 5874653, 1999. [8] M.R. Altiokka, S. Akyalcin, Ind. Eng. Chem. Res. 48 (2009) 10840. [9] V.F. Shvets, R.A. Kozloskiy, I.A. Kozloskiy, M.G. Makarov, J.P. Suchkov, A.V. Koustov, Chem. Eng. J. 107 (2005) 199. [10] G.R. Strickler, V.G. Landon, G.J. Lee, U.S. Patent 6 211 419, 2001. [11] G.R. Strickler, G.J. Lee, W.J. Rievert, D.J. LaPrairie, E.E. Timm, U.S. Patent 6479715, 2002. [12] Y.C. Li, B. Yue, S.R. Yan, W.M. Yang, Z.K. Xie, Q.L. Chen, H.Y. He, Catal. Lett. 95 (2004) 163.

99

[13] B. Li, S.Y. Bai, X.F. Wang, M.M. Zhong, Q.H. Yang, C. Li, Angew. Chem. Int. Ed. 51 (2012) 11517. [14] W.L. Dai, C.M. Wang, B. Tang, G.J. Wu, N.J. Guan, Z.K. Xie, M. Hunger, L.D. Li, ACS Catal. 6 (2016) 2955. [15] G.H. Posner, Chem. Rev. 86 (1986) 831. [16] X.Q. Lu, W.J. Zhou, H.H. Wu, A. Liebens, P. Wu, App. Catal. A: Gen. 515 (2016) 51. [17] X.Q. Lu, H.H. Wu, J.G. Jiang, M.Y. He, P. Wu, J. Catal. 342 (2016) 173. [18] P. Wu, H. Xu, L. Xu, Y.M. Liu, M.Y. He, in: S.K. Sharma, MWW-Type Titanosilicate, Springer-Verlag Berlin Heidelberg, Berlin, 2013, p. 65. [19] E.A. Eilertsen, F. Giordanino, C. Lamberti, S. Bordiga, A. Damin, F. Bonino, U. Olsbye, K.P. Lillerud, Chem. Commun. 47 (2011) 11867. [20] E.A. Eilertsen, S. Bordiga, C. Lamberti, A. Damin, F. Bonino, B. Arstad, S. Svelle, U. Olsbye, K.P. Lillerud, ChemCatChem 3 (2011) 1869. [21] H. Li, Q. Lei, X.M. Zhang, J.S. Suo, Catal. Commun. 10 (2009) 1936. [22] P. Wu, D. Nuntasri, J.F. Ruan, Y.M. Liu, M.Y. He, W.B. Fan, O. Terasaki, T. Tatsumi, J. Phys. Chem. B 108 (2004) 19126. [23] L.L. Wang, Y. Wang, Y.M. Liu, L. Chen, S.F. Chen, G.H. Gao, M.Y. He, P. Wu, Micropor. Mesopor. Mater. 113 (2008) 435. [24] S.Y. Kim, H.J. Ban, W.S. Ahn, Catal. Lett. 113 (2007) 160. [25] P. Wu, J.F. Ruan, L.L. Wang, L.L. Wu, Y. Wang, Y.M. Liu, W.B. Fan, M.Y. He, O. Terasaki, T. Tatsumi, J. Am. Chem. Soc. 130 (2008) 8178. [26] L.L. Wang, Y. Wang, Y.M. Liu, H.H. Wu, X.H. Li, M.Y. He, P. Wu, J. Mater. Chem. 19 (2009) 8594. [27] X.Q. Fang, Q. Wang, A.M. Zheng, Y.M. Liu, Y.N. Wang, X.J. Deng, H.H. Wu, F. Deng, M.Y. He, P. Wu, Catal, Sci. Technol. 2 (2012) 2433. [28] X.Q. Fang, Q. Wang, A.M. Zheng, Y.M. Yue, L.F. Lin, H.H. Wu, F. Deng, M.Y. He, P. Wu, Phys. Chem. Chem. Phys. 15 (2013) 4930. [29] L.L. Wang, Y.M. Liu, W. Xie, H.H. Wu, X.H. Li, M.Y. He, P. Wu, J. Phys. Chem. C 112 (2008) 6132. [30] L. Xu, D.D. Huang, C.G. Li, X.Y. Ji, S.Q. Jin, Z.C. Feng, F. Xia, X.H. Li, F.T. Fan, C. Li, P. Wu, Chem. Commun. 51 (2015) 9010. [31] P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, Chem. Lett. 29 (2000) 774. [32] M.A. Camblor, A. Corma, M.J. Díaz-Cabañas, J. Phys. Chem. B 102 (1998) 44. [33] M.K. Rubin, P. Chu, U.S. Patent 4954325, 1990. [34] Y. Kubota, Y. Koyama, T. Yamada, S. Inagaki, T. Tatsumi, Chem. Commun. (2008) 6224. [35] M. Taramasso, G. Perego, B. Notari, U.S. Patent 4410501, 1983. [36] P. Wu, T. Komatsu, T. Yashima, J. Catal. 168 (1997) 400. [37] F. Geobaldo, S. Bordiga, A. Zecchina, E. Giamello, G. Leofanti, G. Petrini, Catal. Lett. 16 (1992) 109. [38] P. Wu, T. Komatsu, T. Yashima, J. Phys. Chem. 100 (1996) 10316. [39] R. Buzzoni, S. Bordiga, G. Ricchiardi, C. Lamberti, A. Zecchina, Langmuir 12 (1996) 930. [40] D.B. Ravindra, Y.T. Nie, S. Jaenicke, G.K. Chuah, Catal. Today 96 (2004) 147. [41] P. Wu, T. Tatsumi, Chem. Commun. (2002) 1026. [42] G. Phaomei, E.S. Ningthoujam, W.R. Singh, N.S. Singh, M.N. Luwang, R. Tewari, R.K. Vatsa, Opt. Mater. 32 (2010) 616.