Zhenhao Weia,
Kake Zhua,
Lanyu Xinga,
Fan Yanga,
Yunsheng Lia,
Yarong Xub and
Xuedong Zhu
*a
aUNILAB, State Key Lab of Chemical Engineering, School of Chemical Engineering, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China. E-mail: xdzhu@ecust.edu.cn; Fax: +86-21-64252386; Tel: +86-21-64252386
bResearch Institute of Urumchi Petrochemical Company, PetroChina Company Limited, Urumchi 830019, China
First published on 3rd May 2017
Transforming natural kaolin into pure-phase hierarchical aggregates of nano ZSM-11 has been achieved by using a novel tetrabutylphosphonium hydroxide as a structure-directing agent via steam-assisted crystallization. The hierarchical ZSM-11 possesses a small particle size (about 240 nm), large surface area (428 m2 g−1), and abundant mesopores (0.30 cm3 g−1). Moreover, the hierarchical ZSM-11 exhibits prolonged catalytic lifetime and promoted aromatic selectivity in the methanol-to-aromatics reaction. The superior catalyst stability is attributed to the better capacity of accommodating the coke and a lowered coking rate.
Kaolin, an inexpensive natural mineral, has been increasingly used to replace aluminum- and silicon-containing chemicals as the precursor to synthesize various zeolites, such as A,8 Y,9 NaX,10 SAPO-34,11 mordenite12 and ZSM-5.13 To date, there are no reports on the synthesis of ZSM-11 from kaolin, to the best of our knowledge. Generally, kaolin-derived zeolites, prepared by conventional hydrothermal methods, exhibit micrometer-sized particles. Zhu et al. reported a facile method to prepare mesoporous ZSM-5 zeolites via a steam-assisted crystallization (SAC) method.14 When the method is extrapolated to synthesize ZSM-11, ZSM-5/11 intergrowth was achieved from natural kaolin using the conventional tetrabutylammonium hydroxide (TBAOH) as the structure-directing agent (SDA) via SAC (cf. Experimental and results in ESI†).
Recently, Tsapatsis et al. proposed the one-step synthesis of self-pillared pentasil (SPP) hierarchical zeolite using tetrabutylphosphonium hydroxide (TBPOH) as a single template.15 In this work, we report the generation of hierarchical ZSM-11 made up of nano-sized aggregates from kaolin by using this novel TBPOH as SDA. The product is phase-pure and is prepared via a porogen-free SAC process. The obtained samples were analyzed by X-ray diffraction (XRD), FT-IR spectroscopy, FT-IR spectra of pyridine, scanning electron microscopy, nitrogen physisorption, ammonia temperature-programmed desorption (NH3-TPD) and thermogravimetric (TG) analysis. Furthermore, their catalytic performance in the MTA reaction was evaluated and compared with a standard ZSM-11 with similar composition.
For comparison, conventional ZSM-11 zeolites were prepared with the composition 1SiO2:0.019Al2O3:0.2TBAOH:8.65H2O via SAC, according to the procedure reported by Song et al.16
The as-synthesized Na-ZSM-11 zeolites were converted to the NH4+ salt form by ion-exchange treatment with a 1 M NH4NO3 solution at 80 °C for 24 h. The NH4+-exchanged zeolites were then dried and calcined at 550 °C for 6 h to obtain the H+ form. The conventional ZSM-11 and kaolin-derived ZSM-11 zeolites (H+ form) were named ZSM-11-C and ZSM-11-K, respectively.
FT-IR spectra were recorded with a Nicolet 6700 spectrometer in the 400–4000 cm−1 range by the KBr method.
The FT-IR spectra of pyridine adsorption (Py-IR) of all samples were recorded by a Bruker Tensor 27 spectrometer. Firstly, about 50 mg of the sample was pressed into a self-supporting wafer (Φ 20) and activated at 500 °C for 2 h under vacuum. Then, the wafer was cooled to 100 °C and saturated in pyridine vapor for 0.5 h. Finally, the wafer was evacuated at 350 °C for 1 h to obtain IR spectra.
Scanning electron microscopy (SEM, Nova NanoSEM 450) and transmission electron microscopy (TEM, JEM-2100) were applied to determine the morphology and crystal size.
N2 adsorption–desorption isotherms were obtained using a Micromeritics ASAP 2020 V3 instrument. Prior to measurement, all samples were degassed at 280 °C for 12 h. The adsorption–desorption isotherms were collected at −196 °C in liquid N2, the total surface areas (SBET) were obtained from the Brunauer–Emmett–Teller (BET) method in the adapted pressure range (p/p0 = 0.01–0.15), which was used for comparative purposes.17 The mesopore size distribution was inferred from the adsorption branch of the isotherms, using a standard Barrett–Joyner–Halenda (BJH) model.18 The total pore volume data was deduced from the amount adsorbed at a relative pressure of 0.99. The t-plot method was employed to estimate micropore volume by extrapolating the interception to zero.19
NH3-TPD analysis was performed using a Micromeritics ChemiSorb 2720 with a thermal conductivity detector. Adsorption of NH3 was conducted at 25 °C under a 5% NH3–He atmosphere, and desorption of NH3 was recorded from 150 °C to 650 °C at a heating rate of 10 °C min−1 in a He stream.
The chemical composition of the samples was determined by X-ray fluorescence (XRF) with a Shimadzu Model XRF-1800 instrument.
TG analysis was conducted on a Cahn TherMax 700 instrument. About 10 mg of deactivated catalysts was preheated at 100 °C for 1 h under flowing air (100 mL min−1). Then the sample was heated to 550 °C and held for 2 h.
The XRD patterns in Fig. 2a show that LMK was X-ray amorphous. Moreover, a spot of crystalline quartz was observed corresponding to its characteristic peaks at 2θ of 26.6° and 20.8°. By contrast, ZSM-11-C and ZSM-11-K displayed the characteristic peaks of ZSM-11 zeolites (JCPDS card no. 38-246, 55-349). The enlarged XRD patterns at 2θ of 20–27° and 43–47° are illustrated in Fig. 2b and c. Compared to standard ZSM-5 (JCPDS card no. 00-037-0361), the (421), (133) and (0,10,0) peaks belonging to ZSM-5 were not observed for ZSM-11-C and ZSM-11-K. This suggested that as-synthesized ZSM-11 zeolites were pure-phase and were free from ZSM-5 intergrowths.21,22 Furthermore, ZSM-11-K showed slightly lower crystallinity and broadening diffraction peaks at 2θ = 23–25° as compared to ZSM-11-C, which could be attributed to the smaller relative particle size of ZSM-11-K.
![]() | ||
Fig. 2 (a) XRD pattern, (b) enlarged 2θ region from 20° to 27°, (c) enlarged 2θ region from 43° to 47°, and (d) FTIR spectra of LMK, ZSM-11-C and ZSM-11-K. |
Fig. 2d displays the FT-IR spectra of samples LMK, ZSM-11-K and ZSM-11-C. In the FT-IR spectrum of LMK, the band at 954 cm−1 was ascribed to the Si–O stretching vibration in surface silanol groups.23 However, in the FT-IR spectrum of ZSM-11-K, this band vanished and new bands at 547 and 1225 cm−1 appeared, suggesting the formation of double five-membered rings characteristic of the pentasil family of zeolites.24 Moreover, both ZSM-11-K and ZSM-11-C exhibited characteristic bands at 449, 547, 800, 1099, and 1225 cm−1, indicating that ZSM-11 zeolites with high crystallinity were indeed obtained.25
As shown in Fig. 3, ZSM-11-C exhibited a type I isotherm, corresponding to the typical microporous structure. The N2 adsorption isotherms of ZSM-11-K (Fig. 3) represent a hybrid type I and IV with H3 hysteresis loops, indicating the presence of slit-shaped mesopores,26,27 which could be ascribed to the aggregation of nanocrystals. Furthermore, the mesopore-size-distributions calculated from the adsorption branch of N2 isotherms with the Barrett–Joyner–Halenda (BJH) model18 are also shown in Fig. 3 (insets). The mesopore-size-distribution for ZSM-11-K was inferred from the adsorption branch of the isotherm, to avoid the tensile-strength-effect.28 From the mesopore-size-distribution data, it was clear that the majority of mesopores were larger than 3 nm, albeit with a rather broad distribution due to the packing of primary crystallites that formed the agglomerated morphology (Fig. 1b2 and S2†). These results indicated that ZSM-11-K possibly had a hierarchical porosity: the slit-shaped intercrystalline mesopores that were produced from the aggregation of brick-type nanocrystals and the micropores intrinsically possessed by these brick-type nanocrystals. Moreover, ZSM-11-K exhibited larger BET surface areas (SBET) than did ZSM-11-C, which matched well with the relatively small particle size of ZSM-11-K (Table 1).
Sample | Si/Ala | SBETb (m2 g−1) | Smesoc (m2 g−1) | Vtotald (cm3 g−1) | Vmicroc (cm3 g−1) | Vmesoe (cm3 g−1) |
---|---|---|---|---|---|---|
a Determined by XRF with a Shimadzu Model XRF-1800 instrument.b BET method.c t-Plot method.d Volume adsorbed at p/p0 = 0.99.e Vmeso = Vtotal − Vmicro. | ||||||
ZSM-11-C | 29 | 320 | 128 | 0.19 | 0.12 | 0.07 |
ZSM-11-K | 27 | 428 | 228 | 0.41 | 0.11 | 0.30 |
The strength and concentration of the acidity of ZSM-11-C and ZSM-11-K were determined by NH3-TPD, as shown in Fig. 4a and Table 2. Two desorption peaks were evident at about 200 °C and 400 °C, which were associated with ammonia physically adsorbed onto weak acid sites (WAS) and chemisorbed onto strong acid sites (SAS), respectively.29 From Fig. 4a and Table 2, it was confirmed that ZSM-11-K exhibited fewer SAS and more WAS than did ZSM-11-C. Notably, ZSM-11-K possessed a lower concentration of total acid sites than did ZSM-11-C, even though they shared close Si/Al ratios (Table 1). This should be attributed to the inability of some Al species in LMK to be incorporated into the ZSM-11 framework.30
Sample | Acidity by strengtha (mmol g−1) | Acidity by typeb (mmol g−1) | ||||
---|---|---|---|---|---|---|
SAS | WAS | Total | LAS | BAS | Total | |
a The concentration of strong acid sites (CSAS) and weak acid sites (CWAS) determined by NH3-TPD.b The concentration of Lewis acid sites (CLAS) and Brønsted acid sites (CBAS) calculated by Py-IR after evacuation at 350 °C. | ||||||
ZSM-11-C | 0.686 | 0.518 | 1.204 | 0.256 | 0.220 | 0.476 |
ZSM-11-K | 0.569 | 0.590 | 1.159 | 0.215 | 0.224 | 0.439 |
Py-IR was employed to analyze the Lewis acid sites (LAS) and Brønsted acid sites (BAS) of the catalysts. As demonstrated in Fig. 4b, the IR bands detected at 1454 and 1542 cm−1 were attributed to coordinate interaction of bound pyridine with LAS and to pyridinium ions generated on BAS, respectively.31 Moreover, the concentrations of LAS (CLAS) and BAS (CBAS) were calculated according to the integrated area of the IR bands at about 1454 and 1542 cm−1, as shown in Table 2. It was found that the CBAS was similar for ZSM-11-K and ZSM-11-C, whereas the CLAS of ZSM-11-C was obviously higher than that of ZSM-11-K. Additionally, ZSM-11-K had a lower concentration of total acid sites than did ZSM-11-C, which was consistent with the results of NH3-TPD.
To obtain a comprehensive idea of the intrinsic acidic properties of ZSM-11-K and ZSM-11-C, the acidity obtained by NH3-TPD and Py-IR was also normalized by surface areas (Table S3†). It has been proposed that the BAS is mainly located in micropores.32,33 Hence the BAS was normalized by micropore surface areas. Notably, it was hard to probe the exact location of LAS, WAS and SAS. Thus, the LAS, WAS and SAS were roughly normalized by SBET. ZSM-11-K showed obviously lower concentrations of LAS, WAS and SAS than did ZSM-11-C, due to its large SBET. Furthermore, it was found that the two samples exhibited similar CBAS, which was consistent with the results obtained from the acidity normalized by mass.
To explore the potential of the as-synthesized zeolites as catalysts, we compared the catalytic performance of ZSM-11-K and ZSM-11-C in the MTA reaction. As shown in Fig. 5a, ZSM-11-K and ZSM-11-C displayed high methanol conversion (above 98%) within about 180 h. After that, ZSM-11-C deactivated rapidly, while ZSM-11-K maintained methanol conversion above 98% until 236 h. This indicated that ZSM-11-K exhibited better catalyst stability than did ZSM-11-C. Previously, it has been reported that the hierarchical nanocrystalline ZSM-5 zeolites showed high catalyst stability, due to the improved diffusion.2,34 With time-on-stream (TOS) increasing, the aromatic selectivity of the two catalysts decreased monotonously. However, ZSM-11-K displayed higher aromatic selectivity than ZSM-11-C (Fig. 5a and b). Notably, it was well accepted that aromatization activity could be enhanced by increasing the CBAS and improving diffusion properties.2,35,36 The results showed that the two catalysts exhibited similar CBAS (Table 2 and S3†), while ZSM-11-K had more mesopores and smaller crystallite size than ZSM-11-C. As small crystallite size (Thiele theorem) and the presence of auxiliary porosity enhanced diffusion properties, that was also a determinant of the catalyst lifetime in the MTA reaction.1,2 Therefore, the high aromatic selectivity for ZSM-11-K should be attributed to the high mesopore volume (Table 1) and small crystallite size (Fig. 1b2 and S2†). In summary, these reactivity results suggest that ZSM-11-K is more promising for catalytic applications than ZSM-11-C.
![]() | ||
Fig. 5 (a) Methanol conversion and the selectivity for aromatics as a function of time-on-stream (TOS) and (b) initial product distribution obtained at 2 h of TOS (the reaction became stable). |
Technically, industrial catalysts are rarely permitted to run until complete loss of activity; thus the total conversion capacities were supplemented to qualitatively evaluate the catalytic performance of ZSM-11-K and ZSM-11-C. As explained by Bjørgen et al.,37 this was done by plotting the methanol conversion versus the grams of methanol converted per gram of catalyst, and extrapolating the curves towards zero conversion, thereby obtaining the value for the total capacity of the catalyst for the methanol to hydrocarbons reaction (MTH) until complete deactivation; this is illustrated in Fig. 6. For ZSM-11-C, a total conversion capacity of 428 g (methanol) (g (catalyst))−1 was observed, and for ZSM-11-K it was 539 g g−1. These calculated values suggested that the catalyst utilization of ZSM-11-K was obviously higher than that of ZSM-11-C.
The unsaturated species, including alkenes, aromatics and cyclic alkenes, were transformed into coke through dehydrogenation and condensation. Coke was trapped inside the cavity and deposited on the external surface of the catalysts, which resulted in catalyst deactivation.38 The TG was used to estimate the content and rate of coke deposited on the catalysts. Fig. 7 shows that ZSM-11-K had a higher coke content than ZSM-11-C, which can be attributed to the abundance of mesopores in ZSM-11-K. This was also reported for other coked hierarchical catalysts.39–41 The mesopores could enhance the capacity of accommodating the coke of the catalysts and thus result in high catalyst utilization.39 Furthermore, the coking rate, defined as the coke production per hour, was calculated. It was about 2.37 mg g−1 h−1 for ZSM-11-C and 2.20 mg g−1 h−1 for ZSM-11-K. The lower coking rate of ZSM-11-K was ascribed to the lower CLAS. Guisnet et al. proposed the role of acid sites in the formation of coke: (i) the higher concentration of acid sites favors the condensation reactions, resulting in a faster coking rate; (ii) the stronger acid sites could speed up these reactions and therefore result in a faster coking rate.38 These results suggest that it was reasonable for ZSM-11-K to exhibit outstanding catalyst stability, owing to its better capacity of accommodating the coke and the lower coking rate.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra03141f |
This journal is © The Royal Society of Chemistry 2017 |