Journal of Energy Chemistry 22(2013)769–777

Catalytic performance of hierarchical H-ZSM-5/MCM-41 for methanol dehydration to dimethyl ether Yu Sang,

Hongxiao Liu, Shichao He, Hansheng Li∗ , Qingze Jiao, Qin Wu,

Kening Sun

School of Chemical Engineering and the Environment, Beijing Institute of Technology, Beijing 100081, China [ Manuscript received January 3, 2013; revised February 22, 2013 ]

Abstract Micro-mesoporous composite molecular sieves H-ZSM-5/MCM-41 were prepared by the hydrothermal technique with alkali-treated H-ZSM-5 zeolite as the source and characterized by scanning electron microscopy, transmission electron microscopy, energy dispersive spectroscopy, X-ray diffraction, N2 adsorption-desorption measurement and NH3 temperature-programmed desorption. The catalytic performances for the methanol dehydration to dimethyl ether over H-ZSM-5/MCM-41 were evaluated. Among these catalysts, H-ZSM-5/MCM-41 prepared with NaOH dosage (nNa /nSi ) varying from 0.4 to 0.47 presented excellent catalytic activity with more than 80% methanol conversion and 100% dimethyl ether selectivity in a wide temperature range of 170−300 ◦ C, and H-ZSM-5/MCM-41 prepared with nNa /nSi = 0.47 showed constant methanol conversion of about 88.7%, 100% dimethyl ether selectivity and excellent lifetime at 220 ◦ C. The excellent catalytic performances were due to the highly active and uniform acidic sites and the hierarchical porosity in the micro-mesoporous composite molecular sieves. The catalytic mechanism of H-ZSM-5/MCM-41 for the methanol dehydration to dimethyl ether process was also discussed. Key words hierarchical porosity; H-ZSM-5; composite molecular sieve; methanol dehydration; dimethyl ether

1. Introduction Recently, environmentally benign and economical alternative fuel has received global attention because of the limited oil reserves in the world and stringent environmental regulations. Dimethyl ether (DME) can be used as the most promising candidate for diesel engines due to its excellent behavior in compression ignition combustion [1]. Accordingly, the synthesis of DME has drawn wide attention. At present, there are two main methods to produce DME: two-step DME synthesis with methanol as interim material (MTD) [2] and direct DME synthesis from synthesis gas (STD) over a hybrid catalyst comprising a methanol-synthesis catalyst and a solid acid in a single reactor [3,4]. Both of these two processes involve methanol dehydration, which plays a key role in catalyst longevity, DME productivity and production costs. Besides, methanol dehydration prefers low temperature reaction environment, as it is an exothermic reaction. [5]. Thus, in view of reaction thermodynamics and economy of DME production, low-temperature, high activity and high stability are essential for developing a solid acidic catalyst for the MTD process or for the process which fits well with the methanol-synthesis catalyst such as Cu/ZnO/Al2 O3 [6] in the STD process. ∗

Up to now, many methanoldehydration catalysts have been examined, for instance, γ-alumina [7−10], aluminasilica mixtures [11], aluminium phosphate [12], molecular sieves [13], etc. It is generally accepted that Br¨onsted acid or Lewis acid sites are the active sites for methanol dehydration to DME [14]. The stronger the acidity of the active sites, the higher the catalytic activity for methanol dehydration to DME. However, many secondary reactions, especially coking reaction, usually take place on the sites with strong acidity during the catalytic reaction of methanol to DME process, due to the non-uniform distribution of acidic strength on the surface of solid acids, and result in the decrease of DME selectivity and catalyst deactivation [15,16]. Molecular sieves have been widely used in heterogeneous catalysis and H-ZSM-5 molecular sieves have been reported by many research groups to be excellent dehydration catalysts with low-temperature activity superior to γ-Al2 O3 . Many research results [17−19] show that H-ZSM-5 with the SiO2 /Al2 O3 ratio of 15−25 present a good catalytic activity and stability for the MTD process. Besides, the catalytic activity decreases while DME selectivity shows an upward trend as the Si/Al ratio increases. However, H-ZSM-5 has a certain anti-coking capability due to the shape effect, secondary

Corresponding author. Tel/Fax: +86-10-68918979; E-mail: [email protected]

Copyright©2013, Dalian Institute of Chemical Physics, Chinese Academy of Sciences. All rights reserved.

770

Yu Sang et al./ Journal of Energy Chemistry Vol. 22 No. 5 2013

products of hydrocarbons and coke are also generated when H-ZSM-5 is used as the catalyst, lowering the DME selectivity to be less than 100%, due to the existing strong acid sites on the surface. Selective poisoning H-ZSM-5 with Na [20] or modifying it with Ti [21] or P [22] can effectively improve the selectivity to DME. Na-modified H-ZSM-5 catalysts (0−80 mol%) prepared by the impregnation method have higher activities than γ-Al2 O3 , and still present more than 80% of methanol conversion, 100% DME selectivity and a good water resistance at 230−340 ◦ C after 65 h, due to the elimination of acid sites by partly replacing with Na and avoiding the production of coke and/or hydrocarbons [20]. Timodified H-ZSM-5 catalyst (Ti-ZSM-5, Si/Al = 50−200 and Si/Ti = 70, prepared by the sol-gel process associated with microwave radiation) shows a moderate activity and selectively produces DME because of the insertion of Ti4+ and Al3+ into the framework of Ti-ZSM-5 [21]. In P-modified H-ZSM-5 catalyst (P/ZSM-5), the interaction of P with the framework of H-ZSM-5 results in decreasing the acid strength and generating new acid sites, and P/ZSM-5 shows lower acidity, higher hydrothermal stability and improved DME selectivity. The optimal activity was obtained at the P/Al ratio of 1.05 in respect to H-ZSM-5 [21]. However, the modified H-ZSM-5 are still microporous zeolites like H-ZSM-5, and their small microporous channels in which the size of the reactants and the micropore diameter are comparable and products like DME does not diffuse quickly enough [23]. This causes modified H-ZSM-5 to lose catalytic activity and selectivity quickly because many byproducts and carbon deposits are produced during the catalytic process. For a given zeolitic material, the basic strategy to improve diffusion is to shorten the length of the micropore channels or to widen the pore diameter [24,25]. Zheng et al. [26] combine β-zeolite and mordenite zeolite to form composite molecular sieves (BMZ) with microporous and mesoporous structures, along with the controllable Lewis/Br¨onsted acid by two-stage hydrothermal crystallization to improve the deficiency of microporous molecular sieve. H-BMZ shows an excellent performance that: 90% methanol conversion and 100% DME selectivity are obtained during 197−275 ◦ C, and more than 80% methanol conversion and 100% DME selectivity at 275 ◦ C are achieved after 72 h. The authors attributed the high activity of H-BMZ to the high amount of total acid and the secondary mesoporous structure, and the high DME selectivity was ascribed to high Lewis/Br¨onsted acid ratio and microcrystals in H-BMZ. Tang et al. [27] synthesize micro-mesoporous composite molecular sieves H-ZSM-5/MCM-41 by self-assembly in nanoscale which show high methanol conversion, 100% selectivity and a long lifetime in a wide temperature range. In this work, a series of micro-mesoporous composite molecular sieves H-ZSM-5/MCM-41 were prepared by the hydrothermal technique with alkli-treated H-ZSM-5 zeolite as the source. The influences of the dosage of NaOH solution on the structure and surface acidity of H-ZSM-5/MCM-41, and the catalytic performance for the methanol dehydration to DME were discussed. Besides, a catalytic mechanism for methanol dehydration to DME over H-ZSM-5/MCM-41was put forward based on the obtained results.

2. Experimental 2.1. Catalyst preparation The micro-mesoporous H-ZSM-5/MCM-41 composite molecular sieves were prepared by the hydrothermal technique using alkli-treated H-ZSM-5 zeolite as the source [28,29]. Samples of 2.0 g H-ZSM-5 zeolite with SiO2 /Al2 O3 of 38 (The Catalyst Plant of Nankai University) were alkalitreated with 1.5 mol/L NaOH solution (Sinopharm Chemical Reagent Co., Ltd.) with the molar ratio nNa /nSi of 0.4, 0.47, 0.6, 0.8 and 1.0 at 40 ◦ C for 60 min. A zeolite solution consisting of aluminosilicate fragments was formed. 25 mL 10 wt% cetrimonium bromide (CTAB, Sinopharm Chemical Reagent Co., Ltd.) solution was added into the above solution and stirred for 60 min. Then, the resulting solution was placed in an autoclave with trifluoroethanol lining and crystallized at 110 ◦ C for 24 h. Further crystallization under 110 ◦ C for 24 h was carried out after cooling the reactor and adjusting the pH value of the solution to 8.5. As crystallization was completed, Na-ZSM-5/MCM-41 composite molecular sieves were obtained after the solid product was filtered, washed, dried, and calcined in air at 550 ◦ C for 6 h. Finally, the Na-ZSM-5/MCM-41 was treated with 1.0 mol/L NH4 Cl solution, and then filtered, washed, dried and calcined in air at 550 ◦ C for 2 h to obtain the composite molecular sieves H-ZSM-5/MCM-41. 2.2. Catalyst characterization Scanning electron microscopy (SEM) was performed with an FEI Quanta FEG 250 scanning electron microscope operated at 20 KV to examine the surface topography of the samples. Energy dispersive spectroscopy (EDS) equipped on this instrument was used to obtain the SiO2 /Al2 O3 ratio of H-ZSM-5 and H-ZSM-5/MCM-41 composite zeolites. Transmission electron microscopy (TEM) was operated at 120 kV on a JEM-2010 transmission electron microscope. X-ray diffraction (XRD) analysis was performed on an X’Pert Pro MPD powder X-ray diffractometer system (40 kV, 40 mA) using a Cu Kα radiation source and a nickel filter in the 2θ range of 0.5o –6o and 5o –80o. Infrared spectroscopic (IR) analysis was carried out on a PekinElmer spectrometer with a resolution of 4 cm−1 and scanning range from 4000 cm−1 to 450 cm−1 . Before test, the samples were mixed with KBr and then pressed into self supporting wafers. N2 adsorption-desorption measurement was performed on a Quantachrome autosorb iQ instrument at 77 K. The samples were degassed in vacuum for 3 h at 573 K. The total surface area was determined by the BET method, based on p/p0 data in the range of 0.05–0.3. The micropore volume was obtained from the t-plot method. The non local density functionol theory (NLDFT) model applied to the adsorption branch of the isotherm was used to obtain the pore size distribution of composite molecular sieves and the Saito-Foley

771

Journal of Energy Chemistry Vol. 22 No. 5 2013

(SF) model was used to obtain the pore size distribution of H-ZSM-5. NH3 temperature-programmed desorption (NH3 -TPD) was carried out using a Quantachrome ChemBET 3000 with a thermal conductivity detector (TCD). 80 mg sample was placed in a quartz tubular reactor and pretreated at 600 ◦ C with a N2 flow of 30 mL/min for 1 h and then cooled to 100 ◦ C. Ammonia diluted with Ar (5% v/v NH3 ) was then introduced at a flow rate of 30 mL/min for 1 h at 100 ◦ C and then a He stream was fed in until a constant signal of TCD was obtained. NH3 -TPD was carried out with the reactor temperature at a ramp rate of 10 ◦ C/min from 100 ◦ C to 700 ◦ C.

the components in the effluent with a sampling frequency of 0.05 min−1 . The atomic balances were satisfied with a deviation of less than 5%. The methanol conversion (Xmethanol ) and DME selectivity (SDME ) was defined as follows:

2.3. Catalytic performance evaluation

3. Results and discussion

The performance of catalysts for methanol dehydration was evaluated on a micro-reactor system. 0.5 g catalyst (about 1.5 cm height) was placed in the middle of a stainless steel tubular reactor with quartz sand and glass beads packed at the two ends. Thermocouple was placed in the middle of catalyst. The catalyst was pre-treated at 400 ◦ C in a N2 flow of 30 mL/min for 4 h. As the reactor temperature was cooled to 170 ◦ C, the methanol feed was input by a micro-liquid pump (LabAlliance Series II, America) at a flow rate of 0.1 mL/min, vaporized and then reacted on H-ZSM-5/MCM-41 or H-ZSM5 catalysts for 6 h at each set reaction temperature during the range from 170 ◦ C to 300 ◦ C at 0.1 MPa. The lifetime of HZSM-5/MCM-41 was investigated at 220 ◦ C and the reaction ran 500 h. An online Techcomp GC 7890T gas chromatograph equipped with a TCD detector and a Porapak T column (60−80 mesh, φ3×5000 mm) was connected to analyze

3.1. Structure of ZSM-5/MCM-41

Xm = (Fm,in − Fm,out) ÷ Fm,in

(1)

SDME = 2FD,out ÷ (Fm,in − Fm,out)

(2)

where, Fm,in , Fm,out and FD,out are the molar flow of methanol at inlet, outlet and the molar flow of DME at outlet, respectively.

The surface topography, chemical component, crystal phase and pore size distribution of the composite molecular sieves H-ZSM-5/MCM-41 prepared with H-ZSM-5 were characterized by SEM, TEM, EDS, XRD, IR analysis, and N2 adsorption-desorption measurement, in comparison with the corresponding H-ZSM-5. H-ZSM-5 has typical micropores with the characteristics of being circular Z type with a crossover structure, but has no ordered mesopores. Figure 1 and Figure 2 show the SEM and TEM images of H-ZSM-5, respectively, and their corresponding H-ZSM-5/MCM-41 composite molecular sieves prepared with nNa /nSi of 0.4, 0.47, 0.6, 0.8 and 1.0. Figure 2 also showed the TEM images of alkli-treated H-ZSM-5 prepared under the same conditions as the first crystallization

Figure 1. SEM images of (a) H-ZSM-5 and H-ZSM-5/MCM-41 prepared by alkli-treatment with nNa /nSi of (b) 0.4; (c) 0.47; (d) 0.6; (e) 0.8; and (f) 1.0

772

Yu Sang et al./ Journal of Energy Chemistry Vol. 22 No. 5 2013

Figure 2. TEM images of (a) H-ZSM-5, (b) alkli-treated H-ZSM-5; H-ZSM-5/MCM-41 prepared by alkli-treatment with nNa /nSi of (c) 0.4, (d) 0.47, (e) 0.6, (f) 0.8, and (g) 1.0; (h) MCM-41

process without CTAB for preparing H-ZSM-5/MCM-41 and MCM-41. H-ZSM-5 molecular sieves are easily aggregated to form cubic particles, as shown in Figure 1(a). As H-ZSM-5 was treated by NaOH, H-ZSM-5 cubic particles were disintegrated gradually with the increase of the dosage of NaOH solution, and when nNa /nSi was equal to 1.0, most of the cubic particles were disintegrated and assembled into new particles as shown in Figure 1(b) to Figure 1(f). The difference in the zeolite samples became more obvious from Figure 2(c) to Figure 2(g) and the characteristics of MCM-41 became more evident, reflecting the effect of the alkali-treatment. The pores arrayed hexagonally along the direction of the pore while one-dimensional lines were found in a regular arrangement in the direction perpendicular to the pores in H-ZSM-5/MCM41, which is characteristic of the pores of MCM-41 [30]. Furthermore, the TEM image of the alkali-treated H-ZSM-5 displayed nanosized H-ZSM-5 nanoparticles. The H-ZSM5/MCM-41 composite molecular sieves and MCM-41 had a similar morphology as shown in TEM images. These proved that the alkali-treatment of H-ZSM-5 with the tested dosages of NaOH solution can surely bring a change in the zeolites that MCM-41 with a mesoporous structure is introduced around H-ZSM-5 crystal particles. Figure 3 shows the XRD patterns of H-ZSM-5, MCM-41 and H-ZSM-5/MCM-41 samples that were prepared by alkalitreatment with different dosages of NaOH solution. From the high-angle XRD results as shown in Figure 3, the two diffraction peaks between 7o and 10o and three diffraction peaks between 22.5o and 25o were found in H-ZSM-5 and H-ZSM-5/MCM-41 samples that were prepared by alkalitreatment with NaOH dosage increasing from 0.4 to 0.6, which were the characteristic peaks of H-ZSM-5 molecular sieve, corresponding to the (101), (020), (501), (151) and (303) crystal faces [31], respectively. As NaOH dosage was

increased from 0.4 to 0.6, the diffraction peaks of the (101), (020), (501), (151) and (303) crystal faces remained but became weaker. When NaOH dosage was 0.8, the diffraction peaks of the (151) and (303) crystal faces disappeared. None of the five diffraction peaks were seen as NaOH dosage was 1.0, indicating that the skeleton structure of ZSM-5 was destroyed completely by NaOH. From the low-angle XRD results as shown in Figure 3, when NaOH dosage increased from 0.4 to 1.0, the characteristic peak of the (100) crystal face assigned to the hexagonal mesopore structure of MCM-41 appeared [32]. In comparison with MCM-41, the diffraction peak of the (100) crystal face of H-ZSM-5/MCM-41 samples shifted to the high angle and the diffraction peaks of the (110) and (200) crystal faces disappeared. It is due to that the alkali-treatment exerted certain influence on the assembling micelles in the hydrothermal process and the following formation of mesopores. The results indicate that part of HZSM-5 were disintegrated into Si-Al nanoclusters and formed the hexagonal mesopore structure in the presence of CTAB templates [33]. It indicated that NaOH dosage had an important effect on the skeleton structure of H-ZSM-5 and the proper NaOH dosage in alkali-treatment was favored to the formation of composite molecular sieves. Figure 4 shows the IR spectra of H-ZSM-5, MCM-41 and H-ZSM-5/MCM-41 samples that were prepared by alkalitreatment with different dosages of NaOH solution. The absorption band appared at 1230 cm−1 as shown in Figure 4(a)– Figure 4(f), which was assigned to the T–O–T (T is Al or Si) asymmetric stretching mode. As shown in Figure 4(a), the absorption band at 550 cm−1 appeared in ZSM-5, which was the characteristic peak of the five-membered ring of HZSM-5 skeleton structure [34]. Comparing with H-ZSM5, as NaOH dosage increased from 0.4 to 0.8, the absorption band at 550 cm−1 as shown in Figure 4(a)–4(e) became weaker. It indicated that these samples contained the

Journal of Energy Chemistry Vol. 22 No. 5 2013

primary or secondary structural units of H-ZSM-5, but the skeleton structure of H-ZSM-5 changed gradually with the increase of NaOH dosage. When NaOH dosage was 1.0, the absorption band at 550 cm−1 in Figure 4(f) disappeared,

773

which implied that the skeleton structure of H-ZSM-5 was destroyed completely. These were consistent with the XRD results.

Figure 3. Wide- (a) and small-angle (b) XRD patterns of (1) H-ZSM-5 and H-ZSM-5/MCM-41 prepared by alkli-treatment with nNa /nSi of (2) 0.4, (3) 0.47, (4) 0.6, (5) 0.8, and (6) 1.0; and MCM-41 (7)

Figure 4. IR spectra of (1) H-ZSM-5 and H-ZSM-5/MCM-41 prepared by alkli-treatment with nNa /nSi of (2) 0.4, (3) 0.47, (4) 0.6, (5) 0.8, and (6) 1.0; and MCM-41 (7)

The N2 adsorption-desorption isotherms of H-ZSM-5 and H-ZSM-5/MCM-41 and their pore size distributions obtained based on the adsorption branch of isotherm were displayed in Figure 5. The N2 adsorption-desorption isotherm of HZSM-5 as shown in Figure 5 was type I, which was typical of microporous zeolites. The adsorbed amount of N2 kept at the level of 100 cm3 /g, which was the value of the filling volume of micropores. The N2 adsorption-desorption isotherm of H-ZSM-5/MCM-41 was type IV, which was typical of mesoporous zeolites, and it indicated the existence of meso-

pores. In the low pressure range (p/p0