HZSM-5 zeolite capsule catalyst

Journal of

Materials Chemistry A View Article Online

Published on 12 April 2017. Downloaded by Dalian Institute of Chemical Physics, CAS on 11/05/2017 07:31:59.

PAPER

View Journal | View Issue

Cite this: J. Mater. Chem. A, 2017, 5, 8599

A hollow Mo/HZSM-5 zeolite capsule catalyst: preparation and enhanced catalytic properties in methane dehydroaromatization† Pengfei Zhu,ab Guohui Yang,b Jian Sun, *c Ronggang Fan,d Peipei Zhang,b Yoshiharu Yoneyamab and Noritatsu Tsubaki *b A hollow Silicalite-1-HZSM-5 zeolite capsule structure (H-S-Z) embedded with Mo nanoparticles was designed and prepared. The H-S-Z was obtained by a dual-layer hydrothermal synthesis method using activated carbon (AC) as a hard template via self-assembly combined hydrothermal crystallization. One layer of Silicalite-1 zeolite shell was first synthesized on the AC core under neutral-like synthesis conditions. Then, a HZSM-5 zeolite shell was synthesized on this Silicalite-1 zeolite shell enveloping the AC core, followed by calcination in air to remove the core AC. After being loaded with Mo, this hollow zeolite capsule catalyst was tested in a methane dehydroaromatization (MDA) reaction. Compared with

Received 16th March 2017 Accepted 12th April 2017

the conventional solid catalyst, the Mo/H-S-Z catalyst significantly improved the methane conversion, formation rate of the benzene product, and catalytic stability, and inhibited the carbon deposition, due to

DOI: 10.1039/c7ta02345f

the accelerated mass-transfer rate in the hollow structure. Furthermore, the catalyst deactivation derived

rsc.li/materials-a

from the side reactions such as poly-aromatics or coke formation at the inner zeolite could be eliminated.

1. Introduction Direct catalytic conversion of methane, which mainly comes from shell gas or natural gas resources, to desired chemical products or liquid fuel is a great challenge and a long-term project in catalysis science.1,2 Wang et al. rst reported that methane could be directly converted to benzene in the absence of oxygen, named methane dehydroaromatization (MDA), using HZSM-5 zeolite supported molybdenum catalysts.3 Among a series of former studies,4–7 it was reported that Mo exhibited best performance for MDA and the HZSM-5 structure was the most-selective zeolite for benzene formation. However, according to its thermodynamic constraints, this MDA reaction can only proceed at temperatures higher than 873 K,5–8 which leads to the facile formation of heavy carbon deposits in the pores of the as-used catalysts. The catalyst activity and stability signicantly decrease because of the coverage of active sites and the inuence of mass transfer by heavy carbon deposits. Until now, numerous attempts have been focused on the improvement of a

School of Environmental Science and Engineering, Shaanxi University of Science and Technology, Xian 710021, PR China

b

Department of Applied Chemistry, School of Engineering, University of Toyama, Gofuku 3190, Toyama 930-8555, Japan. E-mail: [email protected]

c Dalian National Laboratory for Clean Energy (DNL), Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, PR China. E-mail: sunj@dicp. ac.cn d

Shinka Co. Ltd., Gofuku 3383-4, Toyama 930-8555, Japan

† Electronic supplementary 10.1039/c7ta02345f

information

(ESI)

This journal is © The Royal Society of Chemistry 2017

available.

See

DOI:

the activities, selectivities and stabilities of MDA catalysts.9,10 For the industrial application of the MDA reaction, the poor catalyst stability has been usually considered one of the major obstacles, caused by coking in the reaction process. Therefore, it is very necessary to re-design the catalyst structure, to accelerate the mass transfer rate of the formed benzene, and to stop the side reactions of benzene which lead to the formation of coke or poly-aromatics, the reason for catalyst deactivation. Recently, varied encapsulated catalysts with various and controllable diameters from 0.2 to 1.7 mm have been developed by our group. HZSM-5 or H-beta zeolite shell enwrapped on catalyst pellets acted as the solid acidic membrane catalyst providing acidic sites for reactions.11 In our previous reports, the zeolite capsule catalysts have exhibited excellent properties for some reactions.12–14 In general, these encapsulated catalysts were more stable and easily operated because of their much larger sizes than that of microencapsulated ones. It is considered that these may offer us an opportunity to apply these capsule catalysts (core–shell structure) in methane dehydroaromatization (MDA). In this work, we have developed a new reconstruction strategy to synthesize a hollow Silicalite-1-HZSM-5 zeolite capsule structure (H-S-Z) with lots of irregular holes and pores, which could facilitate the diffusion and migration of impregnated Mo species into the inner surface of the hollow structure. Generally, MDA reactions occur at the HZSM-5 channels near the outer surface or near the channel mouth, especially at the high-speed mode of the methane feed. The zeolite at the center of the catalyst pellet could not be utilized and unexpectedly, it

J. Mater. Chem. A, 2017, 5, 8599–8607 | 8599

View Article Online


Journal of Materials Chemistry A

could promote the secondary reactions of the benzene, deactivating the catalyst. So, it is expected that the hollow structure of our new Mo/H-S-Z catalyst can enhance the activity because of the enriched Mo in the inner surface and lower the deactivation rate by suppressing the side reactions, eliminating the possibility of the formation of coke or poly-aromatics at the inner zeolite due to the existence of irregular holes and pores. A dual-layer hydrothermal synthesis method was used to prepare the hollow Silicalite-1-HZSM-5 zeolite capsule structure by using activated-carbon (AC, a size from 0.42 to 0.84 mm) as a hard template. That is, one layer of Silicalite-1 zeolite shell was rst synthesized on the hard core AC under close to neutral synthesis conditions. Here, the formed Silicalite-1 layer, as an intermediate zeolite layer, could induce the growth of the following HZSM-5 zeolite shell on its surface effectively. Then, the HZSM-5 zeolite shell was synthesized on this Silicalite-1 zeolite layer covering the AC core support, followed by calcination in air to remove the AC core. The preparation scheme of the dual-layer structure is shown in Scheme 1. This dual-layer method can make the growth of the following HZSM-5 zeolite shell easier on the Silicalite-1 zeolite layer acting as a seeding layer. Furthermore, the catalytic performance of the HZSM-5 zeolite capsule structure was demonstrated by the MDA reaction.

2. 2.1

Experimental Catalyst preparation

The activated-carbon (AC, Kanto Chemical Co.) with a size from 0.42 to 0.84 mm was vacuumed at 120 C for 4 h, which was used as a hard template for the capsule structure. A Silicalite-1 zeolite layer was rst synthesized on the astreated naked AC pellets, acting as an intermediate layer for HZSM-5 zeolite shell growth on its surface. 0.4 g AC and the Silicalite-1 zeolite synthesis solution with a molar ratio of 2TEOS : 0.50TPAOH : 120H2O : 8EtOH : 0.25HNO3 were sealed together into a Teon lined autoclave with a rotation rate of 2 rpm for hydrothermal synthesis. The temperature and time of crystallization were 180 C and 48 h, respectively. Aer that, the obtained ones were washed with distilled water and dried in air at 120 C for 12 h, and named AC-S, which would be employed for the next step. The fresh AC-S sample aer washing and drying but without calcination in air was utilized here as a new core for the

Paper

synthesis of the second layer of HZSM-5 zeolite shell. The composition of the HZSM-5 zeolite shell direct synthesis solution was at a molar ratio of 2TEOS : 0.5TPAOH : 120H2O : 8EtOH : 0.025Al2O3, in which the aluminum resource was Al(NO3)3$9H2O.13,14 For the process of hydrothermal synthesis, 0.4 g of fresh AC-S and the above HZSM-5 precursor solution were simultaneously added into in a Teon lined autoclave, operated at 180 C with a rotation speed of 2 rpm for 48 h. Then, the obtained solid was ltered from the solution and washed with distilled water until its pH value was less than 8. Next, the above solid was dried in air at 120 C for 12 h and named AC-SZ. Finally, the obtained AC-S-Z was directly calcined in air from room temperature to 500 C with a ramp rate of 1 C min1 and kept at 500 C for 12 h, forming a new hollow capsule Silicalite1-HZSM-5 shell structure, denoted as H-S-Z. For comparison, the conventional HZSM-5 zeolite or Silicalite1 zeolite was synthesized with the synthesis parameters similar to those of the zeolite capsule structure except for the presence of an AC hard template. Briey, the Silicalite-1 zeolite synthesis solution with a molar ratio of 2TEOS : 0.50TPAOH : 120H2O : 8EtOH : 0.25HNO3 was sealed into a Teon lined autoclave for hydrothermal synthesis, and the composition of the HZSM-5 zeolite direct synthesis solution was at a molar ratio of 2TEOS : 0.5TPAOH : 120H2O : 8EtOH : 0.025Al2O3, in which the aluminum resource was Al(NO3)3$9H2O. The temperature and time of crystallization were 180 C and 48 h, respectively. Next, the above solid was dried in air at 120 C for 12 h, and directly calcined in air at 500 C for 12 h. The obtained HZSM-5 zeolite (Z) and Silicalite-1 (S) were physically mixed well with the weight ratio of 1.67 : 1, the same as that of the HZSM-5 zeolite and the Silicalite-1 content of hollow capsule structure H-S-Z, and then pressed and sorted into sizes of 0.42–0.84 mm. The crystal sizes of the HZSM-5 and Silicalite-1 used for the preparation of the Mo/ S-Z-M catalyst are 1 mm and 3–5 mm, respectively, as displayed in Fig. S1 (ESI†). The physical mixture sample is named S-Z-M, where the “M” stands for the physical mixing of Silicalite-1 and HZSM-5 zeolites. Then, Mo-containing catalysts (Mo wt% ¼ 6%) were prepared by incipient wetness impregnation of the obtained HS-Z and S-Z-M with an aqueous solution of ammonium molybdate. Aer impregnation, the catalysts were dried at 120 C for 4 h, and then both Mo/H-S-Z and Mo/S-Z-M catalysts were calcined in air at 500 C for 4 h. Besides, all the conditions for the two catalysts were kept the same for comparison.

2.2

Scheme 1 Catalyst preparation scheme of the dual-layer hydrothermal method.

8600 | J. Mater. Chem. A, 2017, 5, 8599–8607

Catalyst characterization

The surface morphology and elemental composition of the prepared samples (AC-S, AC-S-Z and H-S-Z) were investigated using a JEOL JSM-6360LV scanning electron microscope (SEM) coupled with an energy-diffusive X-ray spectroscope (EDS). The X-ray diffraction (XRD) patterns were obtained on a Rigaku RINT 2400 equipment with Cu-Ka radiation (l ¼ 0.154 nm). The X-ray tube was operated at 40 kV and 40 mA. The acidic properties of the samples were measured by temperature-programmed desorption of ammonia (NH3-TPD) using a BELCAT-B-TT (BEL, Japan) instrument.


View Article Online

Paper



The carbon deposition behavior of the spent catalysts aer MDA was investigated by thermogravimetric and differential thermal analysis (DTA/TGA 60, Shimadzu) at a heating rate of 10 C min1 from room temperature to 800 C in air ow. 2.3

Catalyst evaluation

The catalytic tests were carried out in a quartz tubular xed-bed reactor (i.d. 9 mm, length 300 mm) at 700 C under atmosphere pressure, with a space velocity of 1500 mL gcat1 h1. Firstly, the catalyst mixed with quartz sand was loaded at the center of the reactor. Secondly, the catalyst zone was pre-treated with the feed of He at 30 mL min1; meanwhile the temperature was raised from ambient to 700 C. Then, the gaseous feed comprising 90.3 vol% CH4 and 9.7 vol% Ar was introduced into the catalyst bed, while Ar in the feed was used as an internal standard for analyzing products. The gas effluent was analyzed by using an online gas chromatograph (GC) equipped with a thermal conductivity detector (TCD), where a Porapak Q column was utilized for the separation of Ar, CH4, CO, C2H4, and C2H6. Aromatic hydrocarbons were collected by using an ice-water trap and then analyzed by using an offline gas chromatograph equipped with a hydrogen ame ionization detector (FID). The conversion and aromatics formation rates were calculated on the carbon number basis. Aromatic hydrocarbons were collected by using an ice-water trap (isooctane as the solvent); next, a certain volume of the sample (V) was taken out and the internal standard (n-decane) was added, and then were analyzed using an offline gas chromatograph equipped with a hydrogen ame ionization detector (FID); the rates of formation of benzene (RB, mmol-C g1 min1) were estimated according to eqn (1). RB ¼ f

V AB mI CI 1 1 AI MI CB mcat T V 00

(1)

where f (mmol per GC peak area) represents a calibration factor for the aromatic product (benzene) and was determined using the external calibration technique; V (mL) refers to the certain volume of the sample, and V00 (mL) stands for the injected volume of the sample for the GC analysis; AB refers to the GC peak area measured for the product (benzene); AI refers to the GC peak area measured for the internal standard (n-decane); mI and MI stand for the mass and mass fraction of the internal standard used, respectively; C is the carbon number in one molecule (6 for the benzene (CB) and 10 for the internal standard (CI)); mcat the weight of the catalyst sample and T the reaction time.

3. 3.1

Results and discussion X-ray diffraction (XRD) analysis

X-ray diffraction (XRD) was very benecial to determine whether or not the zeolite crystal was formed on the hard core AC. The XRD patterns of the naked AC, calcined capsule structure H-S-Z, hybrid sample S-Z-M (mixture of Silicalite-1 and HZSM-5 zeolites) and pure HZSM-5 zeolite are presented in Fig. 1. For the formed hollow capsule structure H-S-Z, by comparing XRD


Fig. 1 XRD patterns of the naked AC, calcined capsule structure H-SZ, hybrid sample S-Z-M (mixture of Silicalite-1 and HZSM-5 zeolites) and pure HZSM-5 zeolite.

patterns with those of the pure HZSM-5 zeolite, the typical peaks belonging to the HZSM-5 zeolite appear in the range of 2q from 5 to 10, as well as from 21 to 25, which conrms that the HZSM-5 zeolite layer was successfully formed using activated carbon as the solid core template by the dual-layer synthesis method. The formed Silicalite-1 zeolite, as a structural layer, was prepared under close-to-neutral synthesis conditions, which avoided the structural destruction of the solid core and favored the in situ growth of the HZSM-5 zeolite. 3.2

SEM and EDS comparison

Fig. 2 provides the overall SEM images of the prepared capsule structures AC-S and AC-S-Z, and the hollow capsule structure HS-Z. Aer the hydrothermal treatment for 48 h, the Silicalite-1 zeolite layer was successfully coated on the core AC pellets, as shown in Fig. 2a, and no obvious defect was observed. Furthermore, its overall surface (Fig. 2a) presents a relatively rough morphology, being contributed by the formation of the Silicalite-1 zeolite layer. The SEM image of the external surface of the AC-S is shown in Fig. 3a, which further indicated that the crystallites of the Silicalite-1 zeolite were formed on the core AC. Here, the formed Silicalite-1 layer, as an intermediate zeolite layer, could induce the growth and fabrication of the following HZSM-5 zeolite shell on its surface effectively.13,17 By the second hydrothermal synthesis, the capsule structure AC-S-Z was prepared, constructing a HZSM-5 zeolite shell on the surface of AC-S. Moreover, when the HZSM-5 zeolite shell was synthesized on the Silicalite-1 zeolite layer covering the AC core, as in Fig. 2b (AC-S-Z, without calcination in air), the overall external surface of the AC-S-Z sample became smoother than that of AC-S (Fig. 2a), indicating that the HZSM-5 zeolite shell was homogeneously coated on AC-S. The surface SEM image of the AC-S-Z pellet is shown in Fig. 3b, and it is easy to observe some differences between this HZSM-5 shell and the intermediate layer of Silicalite-1 shell (Fig. 3a) with respect to their zeolite crystallite shapes. The reason may be the different pH values of the synthesis solution for the Silicalite-1 or HZSM-5

J. Mater. Chem. A, 2017, 5, 8599–8607 | 8601

View Article Online



Fig. 2

Overall surface SEM images of (a) AC-S, (b) AC-S-Z, and (c) H-S-Z.

zeolites, being derived from the existence or not of the aluminum resource in their synthesis recipes.12,15,16 According to the synthesis process, the new hollow capsule H-S-Z pellet was obtained by controlled calcination of the AC-S-Z pellet in air, which made the hard template core to be removed. The overall and surface SEM images of the H-S-Z pellet are presented in Fig. 2c and 3c, respectively. Comparing Fig. 2b

Fig. 3

Paper

(before calcination) and Fig. 2c (aer calcination), the overall external surface of the HZSM-5 layer was smoother before calcination (Fig. 2b) than that aer calcination (Fig. 2c). The phenomenon can be explained as follows. For the H-S-Z pellet aer calcination, the external surface of the sample exhibited a lot of irregular holes and pores because of CO2 escape during the combustion of the hard AC core at high temperature, as

Surface SEM images of (a) AC-S, (b) AC-S-Z, and (c) H-S-Z, and EDS surface analysis of the H-S-Z sample (d).

8602 | J. Mater. Chem. A, 2017, 5, 8599–8607


View Article Online


Paper

clearly shown in Fig. 3b and c, which provided further evidence on this new hollow capsule structure, demonstrating that the capsule structure changed from the hard core (AC)-shell (Silicalite-1 and HZSM-5) to a new hollow hierarchical shell (Silicalite-1 and HZSM-5, H-S-Z) along with some irregular holes and pores during the process of removing the hard AC core in air at high temperature. The surface EDS analysis of H-S-Z determined its surface elemental composition, as shown in Fig. 3d, which suggested that the HZSM-5 zeolite capsule structure coated on the hard template core could be prepared successfully by this dual-layer hydrothermal method. Moreover, according to this EDS surface analysis result, the surface Si/Al ratio (Si/Al ¼ 64) of H-S-Z structure was also obtained. The cross-sectional SEM image and EDS line analysis of the AC-S-Z capsule structure are exhibited in Fig. 4. From this SEM image, a compact and defect-free zeolite shell enwrapping the AC core structure can be observed clearly, which also proves that the zeolite shell was successfully synthesized on the surface of AC and covered the core completely without obvious damage to the core template, indicating the success of this dual-layer method. The EDS line analysis was performed along the line in the SEM image, as shown in Fig. 4. The changes of Al Ka, Si Ka and C Ka signals in intensity from the core to the zeolite shell were exhibited by the analysis results. From this EDS line analysis, it is clearly observed that the radial distribution of Si Ka increased sharply while that of C Ka decreased to zero in the interface region between the core and zeolite shell, which


demonstrated the solid phase being changed from the core to the zeolite shell. At the same time, it was also noticed that Al signal could be detected in the zeolite shell. Moreover, during the second hydrothermal synthesis, some Al source might permeate into the Silicalite-1 zeolite, while parts of Silicalite-1 zeolite might be dissolved under these strong alkaline conditions.17 Therefore, the dissolved silica and introduced Al ions formed the HZSM-5 zeolite during the second hydrothermal synthesis. The HZSM-5 and Silicalite-1 zeolite shell thickness was calculated to be 13.3 mm (distance: 34–47.3 mm) and 11 mm (distance: 23–34 mm) for AC-S-Z, according to the Si and Al EDS line analysis, as displayed in Fig. 4. A transition layer is shown from the distance 23 to 34 mm with increasing Si content and decreasing C content, which is attributed to the gradual coverage of the Silicalite-1 zeolite shell on the carbon core. The thickness ratio of HZSM-5 and Silicalite-1 zeolites is very close to the theoretical weight ratio of 1.67 : 1, demonstrating the signicant growth of the dual-layer zeolite structure.

3.3

Distribution of Mo active sites

As is well known, the MDA reaction usually occurs on bifunctional catalysts, including the active Mo species and Brønsted acid sites inside the zeolite pores, respectively.18,19 It is very important to dene the active Mo species distribution, which signicantly affects the catalytic performance for the MDA reaction. In this case, a cracked sample of the Mo/H-S-Z hollow

Fig. 4 Cross-sectional SEM image and the EDS line analysis of zeolite capsule structure AC-S-Z.


J. Mater. Chem. A, 2017, 5, 8599–8607 | 8603

View Article Online



capsule catalyst was chosen for performing further EDS elemental analysis. As marked in Fig. 5a, the points 1, 2–3 and 4–6 stand for the outer surface, zeolite shell and inner surface of the hollow capsule structure, respectively. Through the EDS statistics (Fig. 5b), it is clearly observed that the Mo content initially gradually increased from the outer to the inner surface and subsequently remained at a stable value on the inner surface, indicating that the inner surface had a slightly higher Mo loading amount. The above nding suggested that the Mo particles were slightly enriched on the inner surface of the zeolite shell. The reason for this should be the fact that the external surface of the sample exhibited some irregular pores and channels because of CO2 escape during the combustion of the hard AC core at high temperature, which facilitated the diffusion and migration of impregnated Mo species into the inner surface. This enriched Mo-containing structure at the inter surface could promote CH4 conversion, because a part of primary product (CHx; C2Hy) obtained inside could undergo further reaction quickly. MoCx inside the zeolite pores is more catalytically active to convert methane to aromatics than that outside.18 The Mo species enriched on the inner surface provide more opportunities for methane and its intermediates to pass through the 10-membered-ring micropores of the MFI zeolite.

Paper

Furthermore, we determined metallic active sites (i.e. Mo dispersion) for the Mo/H-S-Z and the Mo/S-Z-M catalyst with a chemisorption method.20 The percentage of Mo active sites for the hollow Mo/H-S-Z catalyst reaches 5.8%, higher than 3.9% of the mixed Mo/S-Z-M, further demonstrating that the inner surface enriched Mo species are benecial to exposing more active Mo species. 3.4

Acidic property analysis of the hollow capsule H-S-Z

Besides MoCx sites, the acidic sites of the zeolite are the active sites for the MDA reaction. It is important to measure the acidic properties of the hollow capsule structure H-S-Z. The NH3-TPD proles of the hollow capsule H-S-Z and conventional HZSM-5 zeolite are shown in Fig. 6. It can be found in Fig. 6 that the proles of both HZSM-5 and H-S-Z exhibit typical double-peak characteristics of zeolites with the MFI-structure.21,22 The peak at about 207 C was attributed to the desorption of NH3 from weak acidic sites, having low activity for the MDA reaction. The peak at about 405 C could be assigned to the strong Brønsted acidic sites, the active sites for the reaction. The strong acidic sites of H-S-Z played an important role in the C6H6 formation. 3.5

Catalyst performance evaluation

To compare the catalytic performances of the two catalysts, Mo/ H-S-Z and Mo/S-Z-M catalysts with the same 6 wt% Mo loading were prepared. Fig. 7 compares methane conversions and formation rates of benzene for both catalysts. As illustrated in Fig. 7a, methane conversions over the two catalysts displayed a typical trend of continuous decrease with the reaction time, due to the catalyst deactivation caused by accumulation of carbon deposition.23 However, the hollow capsule Mo/H-S-Z catalyst had a lower deactivation rate than Mo/S-Z-M during the reaction. The formation rates of benzene on the two catalysts are shown in Fig. 7b. The hollow capsule Mo/H-S-Z catalyst always exhibited a remarkably higher formation rate of benzene than Mo/S-Z-M during the reaction. Moreover, compared with the solid capsule catalyst Mo/H-S-Z(Ar) calcined with Ar and common catalysts (Table 1), the hollow capsule Mo/H-S-Z

Fig. 5 Mo loading amount by the selected EDS points over the zeolite capsule structure catalyst.

8604 | J. Mater. Chem. A, 2017, 5, 8599–8607

NH3-TPD profiles of the hollow capsule H-S-Z and conventional HZSM-5.

Fig. 6


View Article Online


Paper


Fig. 7 (a) Catalytic performances of Mo/S-Z-M and Mo/H-S-Z for the MDA reaction; (b) formation rates of benzene at 700 C on the two

catalysts Mo/S-Z-M and Mo/H-S-Z with a space velocity of 1500 mL g1 h1.

Catalytic performance of various catalysts being subjected to the MDA reaction for 180 min

Table 1

Catalyst

T ( C)

CH4 conv. (%)

Mo/HZSM-5a Mo/HZSM-5b Mo/H-S-Z(Ar)c Mo/H-S-Z

700 700 700 700

5.72 6.00 5.86 6.29

Formation rate of benzene (mmol gcat1 min1)

Ref.

2.1 2.4 1.3 3.5

This work 24 This work This work

a

The common Mo/HZSM-5. b The parent zeolite catalyst used in ref. 24. The encapsulated Mo/AC-S-Z calcined in an argon atmosphere; the space velocity for all reactions is 1500 mL g1 h1, and the Mo loading amount of these catalysts is the same as 6 wt%. c

catalyst exhibits remarkably higher CH4 conversion and a higher formation rate of benzene. In fact, when CH4 contacts the catalyst surface, CH4 is rst dehydrogenated on the active Mo carbide (MoCx) to form the surface species as CHx (0 < x < 3). Then, the active CHx and a coupled C2 species (C2Hy), as the primary intermediate products, are oligomerized and dehydro-cyclized to form C6H6 on the acidic sites inside channels of HZSM-5,25 as presented in Scheme 2. In this work, for the Mo/S-Z-M catalyst, the solid structure hindered the mass transfer of feed gas from the surface to inner MoCx active sites due to diffusion resistance inside zeolite channels. CH4 molecules might only contact the outside surface randomly and escape. Consequently, a large number of active sites are not utilized efficiently compared with the hollow capsule catalyst Mo/H-S-Z. However, in the case of the hollow

Scheme 2 (a and b) On the Mo site of carbide; (c) on the acidic sites of HZSM-5; (d) on the inner sites of HZSM-5.


capsule Mo/H-S-Z catalyst described in Scheme 3, with the driving force of feed gas ow, the primary intermediate products (CHx and C2Hy) and unreacted CH4 molecules were introduced and accumulated into the hollow cavity inside, where MoCx was enriched and these reactants were further mixed and reacted to form C6H6. Compared with the conventional solid catalyst, the hollow capsule catalyst could suppress the secondary reactions of benzene.26 Benzene was formed at HZSM-5 channels near the outer surface or near the channel mouth, especially at the high-speed mode of methane feed. The zeolite at the center of the catalyst pellet could not be utilized and unexpectedly, it could promote the secondary reactions of the benzene, deactivating the catalyst, because the formed benzene readily reacted to become poly-aromatics and coke if diffused into the inner zeolite. For our hollow case, the inner surface enriched Mo sites are responsible for methane dehydrogenation to CHx in the rst step of the MDA process. And B-acid sites are mainly concentrated on the outer surface due to coverage of the HZSM-5 layer, responsible for the oligomerization and dehydro-cyclization of intermediates (CHx or C2Hy). Assisted by the novel design of inner surface enriched Mo sites and outer surface enriched Bacid sites, the two-step MDA reaction can be efficiently proceeded, achieving a high production rate of benzene. It is considered that the hollow structure of our new Mo/H-SZ catalyst can enhance the activity and lower the deactivation

Schematic representation of the MDA reaction on the hollow capsule catalyst.

Scheme 3

J. Mater. Chem. A, 2017, 5, 8599–8607 | 8605

View Article Online



rate by suppressing the side reactions. Firstly, as the catalyst weight for each experiment was the same, the hollow structure meant that more HZSM-5 thin layers containing MoCx could be provided for methane conversion, than that in the solid catalyst, and these layers were easy for methane to diffuse inside, guaranteeing higher conversion. Secondly, the existence of the inner part of the zeolite pellet could not be fully utilized for the MDA reaction due to the slow methane diffusion rate, and reversely it deactivated the catalyst due to the side reactions of benzene occurring in this core area and nally lowered the benzene selectivity. Resultantly, a high CH4 conversion and high formation rate of benzene were obtained for the hollow capsule Mo/H-S-Z catalyst. Furthermore, regarding the catalytic stability of MDA catalysts, the hollow capsule catalyst Mo/H-S-Z exhibited a lower deactivation rate and better stability than Mo/S-Z-M (Fig. 7a). During the preparation process of the hollow capsule structure, the hard core AC was oxidized and decomposed to steam and CO2 blast by calcination in air at high temperature, forming irregular holes and pores of varied sizes at the zeolite shell. As a result, the hollow interior structure as well as the irregular holes and pores on the zeolite shell accelerated the diffusion rate of products, improving the catalyst activity, enhancing the benzene selectivity and lowering the catalyst deactivation rate.27–29 Thus, the enhanced benzene formation rate and extended catalyst lifetime were determined by the hollow capsule and hierarchical structure, which functionally facilitated the formation of more active sites being quickly accessible by methane, accompanied by the fast diffusion of reaction products off the reaction sites.30 On the other hand, carbon deposition is considered one of the major reasons for the catalyst deactivation during the MDA reaction. Traditionally, coke accumulation on the external surface of the zeolite crystallites or at the pore mouths of the zeolite channels enables the Mo/HZSM-5 catalyst to rapidly deactivate since it narrows and eventually blocks the zeolite pore opening.31 To overcome this, different types of reactors such as a uidized bed reactor and a riser reactor (which is a modication of the uidized bed reactor) have to be applied in this reaction. Zhang et al. have successfully demonstrated that the cyclic reaction/H2-regeneration operation of Mo/HZSM-5 enables the catalyst to keep the high initial activity for long time frames under severe conditions.9,10 To investigate the carbon deposition behavior of the used catalysts aer MDA, TG-DTA was used for testing. Fig. 8 compares TG-DTA curves in air for the conventional solid Mo/SZ-M and the hollow capsule Mo/H-S-Z catalysts aer MDA. When the two spent catalysts contacted air at high temperature (from 100 to 800 C), the oxygen reacted with both the carbon in molybdenum carbide (MoCx) and the coke formed during the MDA. If all of the molybdenum existed in the form of molybdenum carbide aer the reaction, the weight should increase in air at high temperature due to the formation of molybdenum oxide (MoO3) by oxidation. However, according to the TG and DTA data, an increase for the two spent catalysts in weight was not observed while only an obvious weight loss and one corresponding exothermic peak were observed. The reason is

8606 | J. Mater. Chem. A, 2017, 5, 8599–8607

Paper

Fig. 8 TG-DTA curves of the spent Mo/S-Z-M catalyst and the spent capsule Mo/H-S-Z catalyst after the MDA reaction.

considered to be the fact that the MoCx oxidation occurred simultaneously with the combustion of the coke and nally the weight increase of the MoCx oxidation was masked.32 Therefore, the weight loss of the spent catalysts appearing in the TG curve mainly represented the carbon deposition of the catalysts. Clearly, the TG proles illustrated that there was a weight loss ratio of 6.6% on the spent conventional catalyst Mo/S-Z-M, higher than that of 4.8% over the spent hollow capsule catalyst Mo/H-S-Z. Thus, the hollow capsule structure loaded Mo catalyst (Mo/H-S-Z) effectively inhibited the carbon deposition from the MDA reaction.

4. Conclusions A HZSM-5 zeolite shell coated activated-carbon (AC) core with a size from 0.42 to 0.84 mm was successfully fabricated using a simple dual-layer hydrothermal synthesis method under rotating crystallization conditions. Then, a new hollow core Silicalite-1-HZSM-5 zeolite shell capsule structure (H-S-Z) was obtained by removing the core template AC via controlled calcination in air. Aer loading Mo, methane dehydroaromatization (MDA), as an application reaction, was selected to test the Mo/H-S-Z catalyst. Compared with the conventional solid catalyst, the reaction results showed that the hollow capsule Mo/HS-Z catalyst exhibited signicantly enhanced catalyst activity and relatively decreased the carbon deposition in the MDA reaction. It is considered that the promoted catalytic performance was due to the hollow hierarchical structure. The hollow structure provided more HZSM-5 thin layers containing MoCx for methane conversion, than that in the solid catalyst, and these layers were easy for methane to diffuse inside, guaranteeing higher conversion. On the other hand, some irregular holes and pores which were formed via the removal of the AC template led to a higher concentration of Mo sites on the internal surface, which suggested that MoCx particles being enriched on the inner surface of the zeolite shell supplied an extra opportunity for the unreacted feed gas to contact active sites when they escaped away from the cavity.


View Article Online

Paper

Acknowledgements


This work was nancially supported by the grant from NSFC of China (No. 21528302), as well as JST and NEDO of Japan.

References 1 J. H. Lunsford, Angew. Chem., 1995, 34, 970–980. 2 J. H. Lunsford, Stud. Surf. Sci. Catal., 1993, 103, 75. 3 L. Wang, L. Tao, M. Xie, G. Xu, J. Huang and Y. Xu, Catal. Lett., 1993, 21, 35–41. 4 S. Wong, Y. Xu, W. Liu, L. Wang and X. Guo, Appl. Catal., A, 1996, 136, 7–17. 5 K. Wong, J. Thybaut, E. Tangstad, M. St¨ ocker and G. Marin, Microporous Mesoporous Mater., 2012, 164, 302–312. 6 P. Wu, Q. Kan, D. Wang, H. Xing, M. Jia and T. Wu, Catal. Commun., 2005, 6, 449–454. 7 D. Wang, Q. Kan, N. Xu, P. Wu and T. Wu, Catal. Today, 2004, 93, 75–80. 8 C. Zhang, S. Li, Y. Yuan, W. Zhang, T. Wu and L. Lin, Catal. Lett., 1998, 56, 207–213. 9 Y. Xu, J. Lu, J. Wang, Y. Suzuki and Z. Zhang, Chem. Eng. J., 2011, 168, 390–402. 10 Y. Cui, Y. Xu, J. Lu, Y. Suzuki and Z. Zhang, Appl. Catal., A, 2011, 393, 348–358. 11 J. Bao, J. He, Y. Zhang, Y. Yoneyama and N. Tsubaki, Angew. Chem., 2008, 47, 353. 12 G. Yang, N. Tsubaki, J. Shamoto, Y. Yoneyama and Y. Zhang, J. Am. Chem. Soc., 2010, 132, 8129–8136. 13 X. Li, J. He, M. Meng, Y. Yoneyama and N. Tsubaki, J. Catal., 2009, 265, 26–34. 14 J. He, Z. Liu, Y. Yoneyama, N. Nishiyama and N. Tsubaki, Chem.–Eur. J., 2006, 12, 8296–8304. 15 A. Persson, B. Schoeman, J. Sterte and J. Otterstedt, Zeolites, 1994, 14, 557–567.



16 A. Persson, B. Schoeman, J. Sterte and J. Otterstedt, Zeolites, 1995, 15, 611–619. 17 Y. Jin, G. Yang, Q. Chen, W. Niu, P. Lu, Y. Yoneyama and N. Tsubaki, J. Membr. Sci., 2015, 475, 22–29. 18 N. Kosinov, F. J. A. G. Coumans, E. A. Uslamin, A. S. G. Wijpkema, B. Mezari and E. J. M. Hensen, ACS Catal., 2017, 7, 520–529. 19 N. Kosinov, F. J. A. G. Coumans, E. Uslamin, B. Mezari, A. S. G. Wijpkema, E. A. Pidko and E. J. M. Hensen, J. Catal., 2017, 346, 125–133. 20 P. Da Costa, C. Potvin, J.-M. Manoli, M. Breysse and G. Dj´ egaMariadassou, Catal. Lett., 2001, 72, 91–97. 21 H. Ichibashi and M. Kitamura, Catal. Today, 2002, 73, 23–28. 22 R. Long and R. Yang, J. Catal., 2001, 198, 20–28. 23 W. Liu, Y. Xu, S. Wong, L. Wang, J. Qiu and N. Yang, J. Mol. Catal. A: Chem., 1997, 120, 257–265. 24 K. Honda, X. Chen and Z. Zhang, Appl. Catal., A, 2008, 351, 122–130. 25 H. Ma, R. Kojima, S. Kikuchi and M. Ichikawa, Catal. Lett., 2005, 104, 63–66. 26 J. Sun, C. Xing, H. Xu, F. Meng, Y. Yoneyama and N. Tsubaki, J. Mater. Chem. A, 2013, 1, 5670–5678. 27 C. Christensen, K. Johannsen, I. Schmidt and C. Christensen, J. Am. Chem. Soc., 2003, 125, 13370–13371. 28 F. Xiao, L. Wang, C. Yin, K. Lin, Y. Di, J. Li, R. Xu, D. Su, R. Schl¨ ogl, T. Yokoi and T. Tstsumi, Angew. Chem., 2006, 118, 3162–3165. 29 K. Egeblad, C. Christensen, M. Kustova and C. Christensen, Chem. Mater., 2008, 20, 946–960. 30 N. Chu, J. Wang, Y. Zhang, J. Yang, J. Lu and D. Yin, Chem. Mater., 2010, 22, 2757–2763. 31 Y. Xu, J. Wang, Y. Suzuki and Z. Zhang, Appl. Catal., A, 2011, 409, 181–193. 32 E. Matus, I. Ismagilov, O. Sukhova, V. Zaikovskii, L. Tsikoza and Z. Ismagilov, Ind. Eng. Chem. Res., 2007, 46, 4063–4074.

J. Mater. Chem. A, 2017, 5, 8599–8607 | 8607