Catalytic Dehydration of Methanol to Dimethyl Ether Catalyzed by Aluminum Phosphate Catalysts

Page 1

CZ

Qa

b

c

a

ARRAA

KDMCMD

1

timrccmsm

2

3

catrc

v

b

0d

Applied Catalysis A: General 413– 414 (2012) 36– 42

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General

jo u r n al hom epage: www.elsev ier .com/ locate /apcata

atalytic dehydration of methanol to dimethyl ether over micro–mesoporousSM-5/MCM-41 composite molecular sieves

iang Tanga, Hang Xua,b,∗, Yanyan Zhenga, Jinfu Wanga,∗∗, Hansheng Li c, Jun Zhangb

Department of Chemical Engineering, Tsinghua University, Beijing 100084, PR ChinaChemical Engineering and Pharmaceutics School, Henan University of Science and Technology, Luoyang 471003, PR ChinaSchool of Chemical Engineering and the Environment, Beijing Institute of Technology, Beijing 100081, PR China

r t i c l e i n f o

rticle history:eceived 13 September 2011eceived in revised form 23 October 2011ccepted 26 October 2011vailable online 3 November 2011

a b s t r a c t

A series of micro–mesoporous ZSM-5/MCM-41 composite molecular sieves were prepared by combininga microporous zeolite silica source with nano self-assembly methods, and characterized by SEM, TEM,XRD, N2 adsorption and desorption, TPD of ammonia, and their catalytic performance for the dehydrationof methanol to dimethyl ether (DME) in a fixed bed microreactor at atmospheric pressure. Among thesecatalysts, the ZSM-5/MCM-41 sample alkali-treated with 1.5 mol/L NaOH solution, in which the relative

eywords:ehydrationicro–mesoporous

omposite molecular sieves

crystallinity of ZSM-5 and MCM-41 are 54.5% and 30.5%, respectively, gave the best activity (XMeOH > 80%)with 100% selectivity and a long lifetime in a wide range of temperature from 190 ◦C to 300 ◦C. Fromthe characterization and activity data, the formation mechanism of the micro–mesoporous compositemolecular sieves was proposed to be a liquid-crystal templating mechanism.

ethanolimethyl ether

. Introduction

In recent years, dimethyl ether (DME) has received global atten-ion as a green alternative fuel for diesel engines because of thencreasingly inadequate petroleum supply and stringent environ-

ental regulations [1,2]. Additional uses for DME include theeplacement of chlorofluorocarbons as the aerosol propellant in theosmetic industry [3], as a family cooking gas, and as a hydrogenarrier for fuel cell [4,5]. There are two main ways to produce DME:ethanol dehydration over a solid acid catalyst (Eq. (1)) and direct

ynthesis from synthesis gas over a hybrid catalyst comprising aetal oxide and a solid acid (Eq. (2)) [6–8].

CH3OH � CH3OCH3 + H2O (1)

CO + 3H2 � CH3OCH3 + CO2 (2)

The direct synthesis of DME usually uses a bifunctional catalystontaining a metal oxide to catalyze synthesis gas to methanol and

solid acid for methanol dehydration in a single reactor. However,

he optimum catalytic reaction temperatures of the two componenteactions are different [9], and it is difficult to operate the pro-ess in a way which ensures that the two reactions both have high

∗ Corresponding author at: Department of Chemical Engineering, Tsinghua Uni-ersity, Beijing 100084, PR China. Tel.: +86 10 62796109; fax: +86 10 62772051.∗∗ Corresponding author. Tel.: +86 10 62796109; fax: +86 10 62772051.

E-mail addresses: [email protected] (H. Xu), [email protected],[email protected], [email protected] (J. Wang).

926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.apcata.2011.10.039

© 2011 Elsevier B.V. All rights reserved.

activity and selectivity. Kim et al. [3] have tried a mixed catalystof Cu/ZnO/Al2O3 and ZSM-5, but because Cu/ZnO/Al2O3 has highactivity and selectivity at 270–290 ◦C while ZSM-5 showed poorselectivity when the reaction temperature is higher than 260 ◦C,there was a limit on the performance of the integrated reactionsand process. Thus, it is important to develop a catalyst for convert-ing methanol to DME that also has good activity and selectivity ina temperature range that fits well with that of Cu/ZnO/Al2O3 forconverting synthesis gas to methanol.

Methanol dehydration to DME over a solid acid as catalyst in afixed bed reactor was first reported by Mobil in 1965 [10]. Sincethen, many methanol dehydration catalysts have been examined,for instance, �-alumina [11], alumina–silica mixtures [12], crys-talline aluminosilicates [13], crystalline zeolites [1], clay [2], andphosphates [14] such as aluminium phosphate. The catalytic mech-anisms on solid acids are determined by their properties. It isgenerally accepted that BrØnsted acid and Lewis acid sites arethe active sites for the dehydration of methanol to DME [15]. TheBrØnsted acid site is a strong acid with high catalytic activity. Asa result, on a catalyst that contains many BrØnsted acid sites, thecatalytic reaction of methanol to DME process has many secondaryreactions that reduce the catalyst selectivity.

Molecular sieves are widely used in heterogeneous catalysis, buttheir small microporous channels in which the size of the react-

ing molecules and the micropore diameter are similar can slowdown their reactions [16]. For example, although ZSM-5 is a micro-porous zeolite that has high activity, it also some disadvantageslike low selectivity and short catalytic lifetime [8]. The reason for

Page 2

is A: G

tmtTbd

stw1ton4mcwsss

nacdMMbm

alstgsnmXtmos

2

2

sio13t1Ztauacmt

Q. Tang et al. / Applied Catalys

his is that most of the dehydration catalytic reaction occurs in theicroporous pores. A narrow and slender microporous pore has

he disadvantage that DME does not diffuse quickly enough in it.his causes ZSM-5 to lose catalytic activity and selectivity quicklyecause many by-products and carbon deposits will then be pro-uced in the catalytic process [17].

For a given zeolitic material, the basic strategy to improve diffu-ion is to shorten the length of the micropore channels [18–21] oro widen the pore diameter. Mesoporous molecular sieves, amonghich MCM-41 is representative, have pore diameters from 1.5 to

0 nm. As compared with ZSM-5, MCM-41 has a much decreasedransport resistance that would increase the effective utilizationf acidic sites by bulky molecules, but unfortunately MCM-41 hasearly no acidity and poor hydrothermal stability. A ZSM-5/MCM-1 micro/mesoporous composite molecular sieve could have aicropore and mesopore dual pore size distribution, which would

ombine the channel advantage of a mesoporous molecular sieveith that of having the acidity of ZSM-5. A composite molecular

ieve would allow both the actions of the two types of molecularieves, and a composite molecular sieve can improve hydrothermaltability of MCM-41.

Vishwanathan and Jun [22] tried to modify H-ZSM-5 by impreg-ating with NaNO3 to reduce the catalytic activity of the BrØnstedcid sites to improve the catalyst selectivity. However, this did nothange the pore structure of the zeolite, and their work did notiscuss the mechanism of Na-modification. Naik et al. [23] usedCM-41 as a methanol dehydration catalyst, but pure siliceousCM-41 does not have appreciable acidity. Although acidity can

e introduced through the incorporation of metal ions such as alu-inium, the conversion of methanol could not exceed 80%.In this work, an alkaline hydrothermal condition was used that

llowed micro–mesoporous ZSM-5/MCM-41 composite molecu-ar sieves to be synthesized by combining a microporous zeolitesilica source with nano self-assembly methods. The effect ofhe alkali-treatment condition on the structure was investi-ated. The formation of the ZSM-5/MCM-41 composite molecularieve was suggested to be by a liquid-crystal templating mecha-ism. Methanol dehydration over the ZSM-5/MCM-41 compositeolecular sieve showed high methanol conversion (optimum

MeOH = 86.6%), and 100% selectivity and a long lifetime in a wideemperature range that matched well with that of a Cu/ZnO/Al2O3

ethanol synthesis catalyst. This catalyst can be used as a partf an integrated catalyst system for producing DME directly fromynthesis gas [10].

. Experimental

.1. Catalyst preparation

The micro–mesoporous ZSM-5/MCM-41 composite molecularieves were prepared by combining a microporous zeolite sil-ca source with nano self-assembly methods as follows. Samplesf 2.0 g ZSM-5 zeolite (SiO2/Al2O3 = 20) were alkali-treated with0 ml NaOH solution of various concentrations from 0.5 mol/L to.5 mol/L, respectively, at 40 ◦C for 60 min. A ZSM-5 zeolite solu-ion consisting of silicon aluminium fragments was formed. 25 ml0 wt% cetrimonium bromide (CTAB) solution was added into theSM-5 solution and stirred for 60 min. Then, the resulting solu-ion was placed in an autoclave with trifluoroethanol (TFE) liningnd crystallized under 110 ◦C for 24 h. A further crystallizationnder 110 ◦C for 24 h was carried out after cooling the reactor

nd adjusting the pH of the crystallization liquid to 8.5. When therystallization was complete, Na type ZSM-5/MCM-41 compositeolecular sieves were obtained after the solid product was fil-

ered, washed, dried, and heated in air at 550 ◦C for 6 h. Finally,

eneral 413– 414 (2012) 36– 42 37

the Na-ZSM-5/MCM-41 was treated with 1.0 mol/L NH4Cl solutionfor exchanging ions, then filtered, washed, dried and heated in airat 550 ◦C for 2 h to get the H-ZSM-5/MCM-41 composite molecularsieves.

2.2. Catalyst characterization

The physicochemical properties of the ZSM-5/MCM-41 com-posite molecular sieves were characterized by scanning electronmicroscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), N2 adsorption and desorption (BET and DFTisotherms) and NH3 temperature-programmed desorption (NH3-TPD). SEM was performed with a JSM-7401F scanning electronmicroscope operated at 3 kV to examine the surface topographyof the samples. TEM, used to examine crystallization and disper-sion characteristics of the nano-materials, was operated at 120 kVon a JEM-2010 transmission electron microscope.

XRD analysis was performed on an automated powder X-raydiffractometer system (40 kV, 40 mA) using a Cu Ka radiation sourceand a nickel filter in the 2� range of 1–90◦.

NH3-TPD was carried out using an automated adsorption systemwith an online thermal conductivity detector (TCD). A 100 mg sam-ple was placed in a quartz tubular reactor and pretreated at 600 ◦Cwith a N2 flow of 30 ml/min for 1 h and then cooled to 50 ◦C. Ammo-nia (20% NH3–80% N2) was introduced at a flow rate of 30 ml/minfor 0.5 h at 50 ◦C and then a N2 stream was fed in until a constantTCD level was obtained. The reactor temperature was programmedat a ramp rate of 30 ◦C/min and the effluent was dried by powderedKOH to remove moisture. Ammonia concentration was measuredusing the TCD and recorded as function of temperature.

2.3. Catalytic performance evaluation

Methanol dehydration was performed in a microreactor sys-tem. 1.0 g catalyst was placed in the middle of a long, steel, fixedbed reactor, and packed between quartz sand and glass beads. Thiswas pretreated at 400 ◦C in N2 for 4 h and then cooled to 170 ◦C.The feed was then changed to analytically pure CH3OH, with theflow rate controlled by a micro-liquid pump and it was gasified bya vaporizer. The reaction temperature was from 170 ◦C to 310 ◦Cand the gas hourly space velocity (GHSV) was varied from 4 to60 ml/gcat h−1. The lifetimes of the ZSM-5/MCM-41 and pure ZSM-5 catalysts were investigated at 210 ◦C at a GHSV of 12 ml/gcat h−1.An online Tianmei 7900 gas chromatograph equipped with a TCDdetector was connected to the effluent. A Porapak T column wasused in a series arrangement for the complete separation of CH3OH,H2O and DME. The atomic balances were satisfied with a deviationof less than 5%. CH3OH conversion (Xmethanol) and DME selectivity(SDME) was defined as follow:

Xmethanol = Fmethanol,in − Fmethanol,out

Fmethanol,in

SDME = 2FDME,out

Fmethanol,in − Fmethanol,out

where Fmethanol,in, Fmethanol,out and FDME,out are the molar flow ofmethanol at inlet, outlet and the molar flow of DME at outlet.

3. Results and discussion

3.1. Physical properties of the catalysts

3.1.1. SEM and TEM analysisFig. 1 shows the SEM images of ZSM-5 and the series of ZSM-

5/MCM-41 composite molecular sieves that were alkali-treated

Page 3

38 Q. Tang et al. / Applied Catalysis A: General 413– 414 (2012) 36– 42

H solu

wtFhbgtb

4scF

F

Fig. 1. SEM images of (a) ZSM-5 and the ZSM-5/MCM-41 alkali-treated with NaO

ith NaOH. Pure ZSM-5 molecular sieve has a uniform pore struc-ure and easily aggregated to form cubic particles, as shown inig. 1(a), which was possibly due to the high surface energy andigh capillary force of the Si–Al groups. After an alkali-treatmenty 0.5 mol/L NaOH solution, only a few cubic particles were disinte-rated. When the concentration of the NaOH solution was increasedo 2.5 mol/L, most of the cubic particles were disintegrated andecame irregular, as shown in the figures from Fig. 1(b)–(f).

Fig. 2 shows the TEM images of ZSM-5 and the ZSM-5/MCM-1 composite molecular sieves that were alkali-treated with NaOH

olution. The ZSM-5 pores had the characteristics of being cir-ular Z type with a crossover structure, as shown in Fig. 2(a).ig. 2(b) is similar to Fig. 2(a). The difference in the zeolite

ig. 2. TEM images of (a) ZSM-5 and the ZSM-5/MCM-41 alkali-treated with NaOH soluti

tions of (b) 0.5 mol/L; (c) 1.5 mol/L; (d) 2.5 mol/L; (e) 3.0 mol/L; and (f) 3.5 mol/L.

samples became more evident from Fig. 2(c)–(f) and the charac-teristics of MCM-41 became more evident, reflecting the effect ofthe alkali-treatment. As shown in Fig. 2(c)–(f), the pores becamecylindrical or hexagonal along the direction of the pore while in thedirection perpendicular to the pore, one-dimensional lines can beseen in a regular arrangement, which is characteristic of the poresof MCM-41. In addition, the pore diameter increased as the alkali-treatment became stronger. The pore diameter increased from 5 nmto 12 nm, as seen in Fig. 2(c)–(f).

According to the SEM and TEM analysis, the alkali-treatment

of ZSM-5 could surely bring a change in the zeolites from thatZSM-5 a microporous structure to MCM-41 with a macroporousstructure.

ons of (b) 0.5 mol/L; (c) 1.5 mol/L; (d) 2.5 mol/L; (e) 3.0 mol/L; and (f) 3.5 mol/L.

Page 4

Q. Tang et al. / Applied Catalysis A: G

76543210

AIn

tens

ity (a

.u.)

fe

d

c

b

2 theta/degrees

a

100

110 200

403530252015105

Inte

nsity

(a.u

.)

f

e

d

c

b

2 theta/degrees

a

101 020 501 151303

B

Fs

3

stprtto(

t(of(wtlop3d

tbg

ig. 3. XRD patterns of (a) ZSM-5 and the ZSM-5/MCM-41 alkali-treated with NaOHolutions of (b) 0.5 mol/L; (c) 1.5 mol/L; (d) 2.5 mol/L; (e) 3.0 mol/L; and (f) 3.5 mol/L.

.1.2. XRD analysisFig. 3 shows the XRD patterns of ZSM-5 and the ZSM-5/MCM-41

amples that were alkali-treated by NaOH solution. Fig. 3(A) showshe low angle diffraction range where there were three diffractioneaks due to the (1 0 0), (1 1 0) and (2 0 0) crystal faces of MCM-41,espectively. Fig. 3(B) shows the high angle diffraction range. Thewo diffraction peaks between 2� = 7◦ and 10◦ and three diffrac-ion peaks between 2� = 22.5◦ and 25◦ were the characteristic peaksf the ZSM-5 molecular sieve, corresponding to the (1 0 1), (0 2 0),5 0 1), (1 5 1) and (3 0 3) crystal faces, respectively.

As shown in Fig. 3(B), when the concentration of the NaOH solu-ion used was 0.5 mol/L, the diffraction peaks of the (1 0 1), (0 2 0),5 0 1), (1 5 1) and (3 0 3) crystal faces were nearly the same as thosef pure ZSM-5. As the strength of the NaOH solution was increasedrom 1.5 mol/L to 2.5 mol/L, the diffraction peaks of the (1 0 1),0 2 0), (5 0 1), (1 5 1) and (3 0 3) crystal faces remained but becameeaker. When the strength of the NaOH solution was 3.0 mol/L,

he diffraction peaks of the (1 5 1) and (3 0 3) crystal faces were noonger seen, which indicated the main effect of the NaOH treatmentn the skeleton structure of ZSM-5. None of the five diffractioneaks were seen when the strength of the NaOH solution was.5 mol/L, indicating that the skeleton structure of ZSM-5 had beenestroyed completely by the NaOH.

As shown in Fig. 3(A), when the strength of the NaOH solu-ion was 0.5 mol/L, no characteristic peak of MCM-41 appearedecause there were not enough disintegration of the ZSM-5 toive enough Si–Al nanoclusters to form the MCM-41 hexagonal

eneral 413– 414 (2012) 36– 42 39

mesopore structure. When the strength of the NaOH solution was1.5 mol/L, the characteristic peaks of the (1 0 0) and (1 1 0) crystalfaces appeared, showing the existence of the hexagonal mesoporestructure. Another characteristic peak of the (2 0 0) crystal faceappeared and its intensity increased as the concentration of theNaOH was increased from 2.5 mol/L to 3.5 mol/L. The probable pro-cess was that the ZSM-5 had been disintegrated into a matrix ofSi–Al nanoclusters, which can form the hexagonal mesopore struc-ture in the presence of CTAB templates [24].

Table 1 shows relative crystallinity of ZSM-5 and MCM-41 in theseries of composite molecular sieves in which there was a changein the zeolites from that with a microporous structure to that witha macroporous structure. The sample alkali-treated with 0.5 mol/LNaOH did not change structure, showing 100% relative crystallinityof ZSM-5. With the use of alkali-treatment solutions from 1.5 to3.0 mol/L, ZSM-5 and MCM-41 co-existed together, the relativecrystallinity of ZSM-5 decreased from 54.5% to 15.0% while that ofMCM-41 increased from 30.5% to 73.5%. When the concentrationof NaOH solution was increased to 3.5 mol/L, all ZSM-5 turned toMCM-41. The increasing alkali-treatment strength resulted in the2� angles of the XRD peaks becoming smaller and the peaks gotsharper. Taking into consideration both the acidity and the porestructure, it can be expected that the catalyst from the 1.5 mol/Ltreatment can provide enough acidity and a macroporous structureat the same time, which would give it excellent catalytic activityand selectivity.

3.1.3. N2 adsorption and desorptionAs shown in Fig. 4(a), the N2 adsorption and desorption isotherm

of ZSM-5 is Type I, which is typical of microporous zeolites. Inthe whole pressure range (0 < P/Po < 0.95), the adsorbed amountof N2 kept at the initial level of 100 cm3/g, because N2 was onlyadsorbed on the surface of ZSM-5 and could not enter the microp-ores in ZSM-5. While the N2 adsorption and desorption isotherm ofZSM-5/MCM-41 is Type IV, which is typical of mesoporous zeo-lites, and that indicated the existence of mesopores. In the lowpressure range (P/Po < 0.44), the adsorbed amount of N2 increasedlinearly with pressure. This was due to monolayer adsorption of N2on the walls of the pores. In range of 0.44 < P/Po < 0.95, a jump inthe adsorbed amount appeared because N2 began filling the meso-pores. Multilayer adsorption of N2 in the mesopores occurred whenP/Po became larger.

The pore size distribution in ZSM-5 in Fig. 4(b) showed that mostof the pores had a diameter of 0.74 nm and some had diameters of2.2–2.7 nm The BET surface area of the ZSM-5 sample was 407 m2/gand the pore volume was 0.19 cm3/g. In the ZSM-5/MCM-41 samplein Fig. 4(b) mesopores with pore diameters of 3.5–5.2 nm appeared,and the BET surface area was 435 m2/g and the pore volume was0.40 cm3/g. This is further evidence that the alkali-treatment by1.5 mol/L NaOH solution resulted in some microporous structureconverting to a mesoporous structure, and the pore volume andsurface area were increased.

3.1.4. NH3-TPD analysisNH3-TPD characterization of the catalyst was used to measure

the distribution of surface acidity and strength of acid sites. Thetotal amount of acid sites in ZSM-5 is related to the Si/Al ratio ofthe skeleton structure. The higher the Si/Al ratio is, the strongerthe acidity is but the lesser the total amount of acid sites is [25,26].Fig. 5 shows the NH3-TPD spectra of ZSM-5 and the ZSM-5/MCM-41 combined molecular sieves alkali-treated with 1.5 mol/L NaOHsolution.

In the spectra in Fig. 5, the lower temperature peaks were dueto weak acid sites and the high temperature peaks were due strongacid sites. Regions I (30–230 ◦C), II (230–410 ◦C) and III (410–570 ◦C)reflect weak, medium and strong acid sites, respectively. The area

Page 5

40 Q. Tang et al. / Applied Catalysis A: General 413– 414 (2012) 36– 42

Table 1Relative crystallinity of ZSM-5 and MCM-41 in the series of composite molecular sieves.

NaOH solution (mol/L) Relative crystallinity (%) MCM-41 Product components

ZSM-5 MCM-41 2� d1 0 0 (nm) ao (nm)

0.5 100 – – – – ZSM-51.5 54.5 30.5 2.450 3.54 4.16 ZSM-5/MCM-412.5 35.3 52.6 2.412 3.68 4.33 ZSM-5/MCM-413.0 15.0 73.5 2.387 3.75 4.41 ZSM-5/MCM-413.5 – 100 2.294 3.86 4.54 MCM-41

Table 2Distribution of acidity on ZSM-5 and the ZSM-5/MCM-41 alkali-treated with 1.5 mol/L NaOH solution.

Sample Distribution of acidity (mmol/g) NH3 Total

Weak (I = 30–230 ◦C) Medium (II = 230–410 ◦C) Strong (III = 410–570 ◦C)

umtTpsott

Ffs

ZSM-5 0.143 0.634

ZSM-5/MCM-41 0.287 0.459

nder the peaks gave the amounts of acid sites. Table 2 shows theeasured amounts of acid sites and their strengths on ZSM-5 and

he ZSM-5/MCM-41 alkali-treated with NaOH solution of 1.5 mol/L.hree factors affect the distribution of acid sites: presence of amor-hous materials, structure and Lewis/Bronsted acidity. It can be

een that after the alkali-treatment, the amounts and the strengthsf medium and strong acid sites decreased, while the strength ofhe weak acid sites increased. This was a change in the distribu-ion of acidity to become one that is good for methanol dehydration

1.00.80.60.40.20.00

50

100

150

200

250

300

ZSM-5/MCM-41

ZSM-5

Vol

ume

adso

rbed

(cm

3 /g)

Relative Pressure (p/p0)

1086420

Inte

nsity

ZSM-5

ZSM-5/MCM-41

Pore diameter (nm)

A

B

ig. 4. N2 adsorption and desorption isotherms (a) and pore diameter distributionrom the BJH model (b) of ZSM-5 and the ZSM-5/MCM-41 alkali-treated with NaOHolutions of 1.5 mol/L.

0.138 0.9150.039 0.785

because its total amount of acid sites was enough for effective catal-ysis while the decrease in strong acid sites will stop by-productformation and improve selectivity.

3.2. Formation mechanism of ZSM-5/MCM-41

Concerning the formation mechanism of the micro–mesoporouscomposite molecular sieves, Beck et al. [27] proposed theliquid-crystal templating (LCT) mechanism for the formation ofmesoporous molecular sieves. In the LCT mechanism, lyotropic liq-uid crystals from the surfactant work as the template agent forthe MCM-41 structure. This process is shown in Fig. 6. First, thealkali-treatment dissolved the Si–Al nano clusters of ZSM-5 intoaluminosilicates. These aluminosilicates aggregated around theCTAB template and formed the walls of mesopores. Then, the alu-minosilicates with the mesoporous structure were deposited ontozeolite nanocrystallites and finally formed the micro–mesoporouscombined molecular sieves.

3.3. Catalytic performance

High activity, high selectivity and long life time are the threeessential requirements for an excellent catalyst. The catalytic per-

formances of the series of micro–mesoporous ZSM-5/MCM-41composite molecular sieves were tested for the methanol dehy-dration to DME process.

500400300200100

Inte

nsity

Temperature(ºC)

a

b

Fig. 5. NH3-TPD spectra of (a) ZSM-5 and (b) ZSM-5/MCM-41 alkali-treated by1.5 mol/L NaOH solution.

Page 6

Q. Tang et al. / Applied Catalysis A: General 413– 414 (2012) 36– 42 41

-5/M

3

ZhmsammtNosaoonTbcot

F52

Fig. 6. Formation mechanism of ZSM

.3.1. Catalytic activityThe activities of ZSM-5 and the series of micro–mesoporous

SM-5/MCM-41 zeolites are shown in Fig. 7. Pure ZSM-5 hadigh activity, when the temperature was increased to 190 ◦C, theethanol conversion exceeded 80%, and got the highest conver-

ion as 86.7% at 210 ◦C; then a slight decrease of conversion to 78.4%ppeared along with the increase of temperature to 310 ◦C. Becauseethanol dehydration is exothermic reaction, 210 ◦C is the opti-um reaction temperature, too high temperature is unfavourable

o reaction and catalysts. Activity of the sample alkali-treated byaOH solution of 0.5 mol/L was lower than ZSM-5 due to decreasef some BrØnsted acid. When alkali-treated with 1.5 mol/L NaOHolution, the sample showed outstanding activity nearly the sames pure ZSM-5 due to acid catalysis effect and hierarchical effectf pore structure. From the XRD and TPD characterization previ-usly discussed, ZSM-5/MCM-41 (1.5 mol/L) has a proper acidityeeded for effective catalysis and a proper macroporous structure.he decrease of acidity could decrease BrØnsted acid in some extent

ut increase Lewis acid, meanwhile the mesoporous structureould improve reactants diffusing into pores and products diffusingut pores resulting in high activity. The further stronger alkali-reatments caused evident decrease of activity because too much

3203002802602402202001801600.0

0.2

0.4

0.6

0.8

1.0

Con

verti

on

Tempreture/ºC

f

ed

a

b

c

ig. 7. Conversion of methanol in the dehydration to DME: (a) ZSM-5 and the ZSM-/MCM-41 alkali-treated with NaOH solutions of (b) 0.5 mol/L; (c) 1.5 mol/L; (d).5 mol/L; (e) 3.0 mol/L; and (f) 3.5 mol/L.

CM-41 composite molecular sieves.

acidity was neutralized and too many mesopores were formed.When the concentration of NaOH was 3.5 mol/L, the conversion ofmethanol was nearly nil because all ZSM-5 turned to MCM-41.

In order to further compare ZSM-5/MCM-41 with pure ZSM-5,their selectivity and stability were studied.

3.3.2. Catalytic selectivityAs shown in Fig. 8, in the wide range of temperature of

170–290 ◦C, the ZSM-5/MCM-41 alkali-treated with NaOH solutionof 1.5 mol/L kept 100% selectivity, but pure ZSM-5 showed a rapiddecrease of selectivity from 98% to 86% as the reaction temperaturewas increased from 170 ◦C to 310 ◦C. The bad selectivity of ZSM-5was proposed due to two reasons. First, the surface acidity of ZSM-5 was too strong; then, microporous structure was unfavourableto diffusion of products, which gave more by-products. So afterchanging the acid strength, acid distribution and pore structure,the ZSM-5/MCM-41 alkali-treated by 1.5 mol/L NaOH solution gavehigher selectivity.

3.3.3. Catalytic stabilityUnlike FCC catalyst, methanol dehydration catalyst cannot be

regenerated online. Thus, a long life time is of special significance

3025201510500.4

0.5

0.6

0.7

0.8

0.9

30028026024022020018080

85

90

95

100

ZSM-5/MCM-41(1.5mol/L)

Sele

ctiv

ity (%

)

Temperature ( ºC)

ZSM-5

ZSM-5

ZSM-5/MCM-41(1.5mol/L)

Con

verti

on(m

ol/%

)

Time/day

Fig. 8. Stability and selectivity in methanol dehydration over ZSM-5 and ZSM-5/MCM-41 composite molecular sieves alkali-treated by NaOH solution 1.5 mol/L.

Page 7

4 is A: G

fbarId4bcwap

4

sdpwdBhmsdawiaaa

A

dP

[

[[[

[[

[[[

[

[

[

[

[

[

2 Q. Tang et al. / Applied Catalys

or fixed bed DME synthesis. ZSM-5/MCM-41 (1.5 mol/L) had theetter stability compared to pure ZSM-5, as seen in Fig. 8. Over

period of 30 days, the activity of ZSM-5/MCM-41 (1.5 mol/L)emained above 85% methanol conversion without any decrease.n contrast, the activity of pure ZSM-5 could be maintained for 15ays at about 85% methanol conversion, but it decreased to nearly5% after 30 days. The deactivation of ZSM-5 may be caused by car-on deposition because microporous pores are easily blocked byarbon deposits. Mesoporous pores effectively improve diffusion,hich would also decrease carbon deposits. These data showed the

dvantages of the ZSM-5/MCM-41 composite molecular sieves overure ZSM-5, especially in selectivity and stability.

. Conclusion

Micro–mesoporous ZSM-5/MCM-41 composite molecularieves, which have a micropore and mesopore dual pore sizeistribution, that combine the channel advantage of a meso-orous molecular sieve with the acidity of a microporous zeoliteere successfully synthesized. The LCT mechanism was used toescribe this synthesis process. Compared with pure ZSM-5, theET surface area and pore volume of ZSM-5/MCM-41 were bothigher and there was vastly improved diffusion in the pores. Theicro–mesoporous ZSM-5/MCM-41 composite molecular sieves

howed high activity, selectivity and stability for methanol dehy-ration to DME in the wide range of temperature of 190–300 ◦C,nd especially at 210 ◦C it gave an optimum activity (XMeOH = 86.6%)ith 100% selectivity and long lifetime. This gives this catalyst that

s a good match with a Cu/ZnO/Al2O3 catalyst that has high activitynd selectivity at 270–290 ◦C. The combination of Cu/ZnO/Al2O3nd ZSM-5/MCM-41 should give satisfying catalytic performancet 270–290 ◦C for producing DME directly from synthesis gas.

cknowledgments

This work is supported by the National Nature Science Foun-ation of China (nos: 21006057 and 21076063) and Chinaostdoctoral Science Foundation (no: 20100470351).

[[

[

eneral 413– 414 (2012) 36– 42

References

[1] S. Hassanpour, F. Yaripou, M. Taghizadeh, Fuel Processing Technology 91 (2010)1212–1221.

[2] K.S. Ha, Y.J. Lee, J.W. Bae, Y.W. Kim, Applied Catalysis A: General 395 (2011)95–106.

[3] J.H. Kim, M.J. Park, S.J. Kim, Applied Catalysis A: General 264 (2004)37–41.

[4] J.H. Fei, Z.Y. Hou, B. Zhu, H. Lou, X.M. Zheng, Applied Catalysis A: General 304(2006) 49–52.

[5] Q. Xu, P. Lan, K. Huang, Y. Yan, Petroleum Science and Technology 29 (2011)1080–1092.

[6] J.H. Fei, Z.Y. Hou, B. Zhu, H. Lou, X.M. Zheng, Applied Catalysis A: General 304(2006) 47–49.

[7] F. Yaripour, F. Baghaei, I. Schmidt, J. Perregaard, Catalysis Communications 6(2005) 147.

[8] Z. Hosseini, M. Taghizadeh, F. Yaripour, Journal of Natural Gas Chemistry 20(2011) 128–134.

[9] M. Stiefel, R. Ahmad, U. Arnold, M. Doring, Fuel Processing Technology 92 (2011)1466–1474.

10] F. Raoof, M. Taghizadeh, A. Eliassi, F. Yaripour, Fuel Processing Technology 87(2008) 2967–2971.

11] S.D. Kim, S.C. Baek, Applied Catalysis A: General 309 (2006) 139–143.12] L.W. Zhang, J.H. Wang, P. Wu, Chinese Journal of Catalysis 31 (2010) 987–992.13] D.M. Sung, Y.H. Kim, E.D. Park, Research on Chemical Intermediates 36 (2010)

653–660.14] W.L. Dai, W.B. Kong, L.D. Li., Catalysis Communications 12 (2011) 535–538.15] N. Khandan, M. Kazemeini, M. Aghaziarati, Applied Catalysis A: General 349

(2008) 6–12.16] J.J. Zheng, Q.H. Zeng, R.F. Li., Catalysis Today 168 (2011) 124–132.17] D.F. Jin, B. Zhu, Z.Y. Hou, et al., Fuel 86 (2007) 2707–2713.18] J.C. Groen, T. Bach, U. Ziese, A.M.P. Donk, K.P. de Jong, J.A. Moulijn, et al., Journal

of the American Chemical Society 127 (2005) 10792.19] J.C. Groen, W. Zhu, S. Brouwer, S.J. Huynink, F. Kapteijn, J.A. Moulijn, et al.,

Journal of the American Chemical Society 129 (2007) 355.20] J.C. Groen, L.A.A. Peffer, J.A. Moulijn, et al., Colloids and Surfaces A 241 (2004)

53.21] J.C. Groen, J.A. Moulijn, J. Pérez-Ramírez, Microporous and Mesoporous Mate-

rials 87 (2005) 153.22] V. Vishwanathan, K.W. Jun, Applied Catalysis A: General 276 (2004)

251–255.23] S.P. Naik, V. Bui, T. Ryu, Applied Catalysis A: General 381 (2010)

183–190.24] J.B. Peng, S.J. Xie., Chinese Journal of Catalysis 23 (2002) 63–366.

25] Z.Y. Hou, T. Okuhara, Applied Catalysis A: General 216 (2001) 147–155.26] R.G. Luz, F. Hermes, M. Bertmer, et al., Applied Catalysis A: General 328 (2007)

174–182.27] J.S. Beck, J.C. Vartuli, W.J. Roth, et al., Journal of the American Chemical Society

114 (1992) 10834–10843.