Ju Huang,
Yaqi Fan,
Guanqun Zhang* and
Yanhang Ma*
School of Physical Science and Technology, ShanghaiTech University, Shanghai 201210, China. E-mail: zhanggq1@shanghaitech.edu.cn; mayh2@shanghaitech.edu.cn
First published on 3rd April 2020
Introduction of meso-/macropores into the intrinsic microporous framework of zeolites has raised substantial interest in catalytic reactions with bulky reactants. Herein, we report the formation of secondary meso-/macropores in Silicalite-1 zeolite by a solvent-free mechanochemical grinding process. The strategy allows the preservation of high crystallinity and microporosity of the pristine zeolite, and the generation of mesopores at room temperature and marcopores at higher temperatures. The roles of the tetrapropylammonium bromide (TPABr) and ammonium fluoride (NH4F) have been proposed and demonstrated. A protective layer is formed by TPA+ ions bonded with the surficial defects to shield the outer surface from the direct attack by F−. Instead, F− diffuses into the micropore system in a local aqueous environment within zeolite formed by the mechanochemical reaction. As a result, freely diffused F− selectively dissolves zones with structural defects to form secondary pores inside the zeolite. Moreover, this strategy proves highly effective in encapsulation of nanoparticles (Pt, Co) in the meso-/macropores of Silicalite-1 zeolite, forming a yolk–shell composite catalyst for potential applications.
Up to now, various types of top-down methods have been developed, including hydrothermal steaming,10 inorganic base treatment,11 organic base treatment (dissolution–recrystallization)12,13 and fluoride-etching strategies.14,15 These methods can be used alone or combined depending on the specific requirements for meso-/macropores,16 and it has been successful in the fabrication of many types of hierarchical zeolites with superior diffusion property, high stability and adjustable acidity.5,17–19 Despite the great synthetic success, research on the formation mechanism of the meso-/macropores in zeolites remains preliminary. Previous reports reveal that the generation behavior of secondary porosity is highly influenced by the post-treatment environment (etching chemical, solvent) and the nature (size, Si/Al ratio) of parent zeolites.10,20
The etching or dissolution process of a crystal is usually viewed as an inverse process of crystallization, and the generation of regular-shaped meso-/macropores is often considered as the result of a “dissolution–recrystallization” process.15,21 However, a recent study on the rectangular voids formed inside ZSM-5 zeolite using NH4F solution in mild condition suggests that these pores are resulted from the preferential dissolution of sub-structure domains lined along the defect zones of the zeolite.22 Although NH4F is not considered as a template for synthesizing zeolitic materials, recent studies demonstrate its cooperation with tetrapropylammonium bromide (TPABr) can lead to the formation of many types of zeolites in solvent-free condition.23–25
Because of the similarity between dissolution and crystallization of zeolites, a dissolution process is reasonably expected by using a combination of organic ammonium salt and NH4F in the solvent-free condition. In this work, introduction of secondary porosity into Silicalite-1 zeolite is succeeded in a solvent-free approach by grinding calcined zeolite with TPABr and NH4F. The strategy preserves the high crystallinity and microporosity of the pristine zeolites, and forms mesopores at room temperature and macropores at higher temperatures. To the best of our knowledge, it is the first report on the introduction of secondary pores under the solvent-free condition.
The solvent-free mechanochemical treatment was carried out as follows: 0.3 g of the calcined Silicalite-1, 0.1 g of ammonium fluoride (NH4), and 0.1 g of tetrapropylammonium bromide (TPABr) were added into the mortar, with a molar composition of 1SiO2:0.54NH4:0.075TPABr, and grinded. After grinding for 5 min, the powder was transferred to a Teflon-lined steal autoclave, sealed and heated at 453 K. After treatment, the powder was washed thoroughly by centrifugation and dried overnight at 363 K. Finally, organic species was removed by calcination in static air at 823 K for 6 h.
Co@Silicalite-1 samples were prepared by firstly impregnation of aqueous solution of Co(NO3)2·6H2O (5 wt% of the Silicalite-1) in Silicalite-1. After being dried at 373 K overnight, the samples were mixed with NH4F and TPABr, ground at room temperature for 5 min, and then transferred to autoclaves for heating at 180 °C for 15 h. The products were recovered by washing, drying, calcination at 550 °C, and then reduced by H2 at 750 °C for 3 h. Pt@Silicalite-1 sample was prepared in a similar way, expect that the impregnation of H2PtCl6 solution (5 wt% of the Silicalite-1) was carried out after the mixing-grinding process.
The effects of experimental conditions, such as the mixture composition and the heating temperature, were also studied (Fig. S3 and S5†). Of note, a too high-temperature heating would cause a surface fusion of the zeolites, suggesting an etching on both the inner and outer parts of Silicalite-1 zeolite (Fig. S6†). Furthermore, this method was also applied to commercial Silicalite-1 and ZSM-5 with different Si/Al ratios, in which meso/macropores were also obtained (Fig. S7†).
Fig. 1e shows the N2 adsorption/desorption isotherms of the parent and post-treated Silicalite-1 samples. All isotherms exhibit two hysteresis loops at low pressure zone (P/P0 ≈ 0.2) and high-pressure zone (P/P0 = 0.6–0.9), respectively. The first hysteresis loop is attributed to the presence of internal structural defects29 and the second to the presence of inter-crystal mesopores.30 The hysteresis loops at low pressure zone slightly increase after post-treatment, indicating the increase of internal structural defects. On the other hand, the hysteresis loops at high pressure zone first increase after mechanical grinding and heating for 1 h, but then decrease with further heating, which may be caused by the merging of mesopores into macropore. In addition, the micropore properties including BET surface area and micropore volume of the post-treated samples show slight decrease, while the external surface area and mesopore volume increase at first but decrease after prolonged heating for 15 h (Table 1).
Sample | SBET (cm2 g−1) | Sext (cm2 g−1) | Vmic (cm3 g−1) | Vmes (cm3 g−1) |
---|---|---|---|---|
a SBET (specific surface area) was obtained by BET method. Sext (external surface area), Vmic (micropore volume) and Vmes (mesopore volume) was obtained by t-plot method of the desorption branch. | ||||
Silicalite-1 | 434 | 28.0 | 0.18 | 0.04 |
G5 min | 412 | 31.0 | 0.19 | 0.03 |
G5 min + H1 h | 403 | 32.5 | 0.21 | 0.01 |
G5 min + H15 h | 399 | 30.0 | 0.18 | 0.03 |
Of note, if there is no heating treatment, more mesopores would be formed at room temperature with the time, but would not merge into large macropores (Fig. 2a). On the other hand, upon heating the merging process takes places very rapidly (Fig. 2). This is a clear indication that the merging process is driven by the high temperature. Post-treated Silicalite-1 samples were analyzed by Powder XRD before washing (Fig. 3). After simple grinding, the diffraction peaks of TPABr are present while NH4F is absent in powder XRD profile, indicating that NH4F is completely consumed. This is also supported by the release of strong ammonium scent during the grinding process. Upon heating, the diffraction peaks of TPABr gradually disappear while the peaks of NH4Br start to emerge and intensify. On such basis, we propose the following chemical reactions:
6NH4F + SiO2 → (NH4)2SiF6 + 4NH3 + 2H2O | (1) |
(NH4)2SiF6 + 2TPABr → (TPA)2SiF6 + 2NH4Br | (2) |
Fig. 2 TEM images of Silicalite-1 after grounded with TPABr–NH4F at (a) room temperature for 15 h, (b) 453 K for 1 h, (c) 453 K for 15 h, (d) 453 K for 48 h. |
Fig. 3 Powder XRD patterns of parent and unwashed post-treated Silicalite-1 zeolites under different conditions. G and H in (d–f) stand for grinding and heating treatment, respectively. |
It is noteworthy that NH4F is in excess amount to TPABr, so the majority of SiF62− species is still present as (NH4)2SiF6, which agrees well with the powder XRD results. Considering the SEM and XRD results mentioned above, it is concluded that the reaction (2) is the cause of the merging of multiple mesopores to macropore. On the other hand, powder XRD profiles of washed samples show only diffraction peaks of MFI-type zeolite, and these post-treated samples maintain the high crystallinity as the parent Silicalite-1 (Fig. 1d). The reaction (1) was also mentioned in the solvent-free synthesis of zeolites as the starting reactions,31 confirming the similarity of the fundamental mechanism between crystallization and dissolution of zeolites. However, transformation from TPABr to NH4Br was not detected in the solvent-free crystallization.31,32 This is probably because in crystallization process the TPA+ is pre-surrounded by silica precursor; while in dissolution process TPA+ is kept outside the framework due to diffusion limit rather than being surrounded by zeolitic framework as a template.
Qin has demonstrated the strong etching effect of NH4F solution to form cleavages and hollow mosaic structures inside MFI zeolites,22 but when NH4F was used alone in solvent-free condition, dissolution only happens on the surface even at elevated temperatures, with the resultant zeolite becoming roughly rounded or even “melted” to merge with each other (Fig. 4). On the other hand, the sole use of TPABr seems to have no effect on the zeolite, since no meso-/macropores are formed and no obvious morphological change can be observed in the Silicalite-1 (Fig. S8†). However, the rapid generation of mesoporosity by the combined use of TPABr–NH4F in solvent-free mechanochemical approach suggests that F− was effectively delivered into the zeolite framework. It should be noted although liquid water was not intentionally added, the presence of trace amount of water is inevitable in the system due to the hydroscopic nature of NH4F and TPABr.24,33 Moreover, the reaction between NH4F and SiO2 also releases water in the system as shown in eqn (1). The stoichiometric amount of water creates a local aqueous environment within the Silicalite-1. Under such environment NH4F readily diffuses into some structural defect zones such as intergrowth interfaces and/or grain boundaries which are more vulnerable to chemical attack by NH4F,22 and preferentially etches them to form multiple mesopores. At elevated temperatures, F− becomes more reactive, and therefore dissolution tends to inter-connect the multiple mesopores after the complete dissolution of defective zones. The further dissolution will undergo the inverse fashion of the layer-by-layer SiO2 accumulation during crystallization,26 that is, a layer-by-layer etching on certain lattice planes, to form the single macropore with well-defined edges and facets.
The state of fluoride species was monitored by 19F MAS NMR spectra of the post-treated Silicalite-1 samples (Fig. 1f). A remarkable band at −128 ppm was observed corresponding to the significant presence of (NH4)2SiF6.34 Upon heating, three new bands gradually emerge at −65, −69 and −80 ppm, respectively. The first signal is attributed to F− trapped in the [415262] cage of Silicalite-1, the second signal probably to ionically bonded TPA+–F− pair34 and the third signal to the presence of F− bonded to the structural defect.35 19F MAS NMR investigation on the solvent-free synthesis of zeolite discovers the simultaneous decrease of SiF62− and increase of F−, indicating a mass transformation from SiF62− to F− via the templating process of TPA+31−32. However, no such transformation were detected in the present study, which further confirms the non-templating effect of TPABr in generating the secondary pores.
Although not acting as a template, TPA+ plays an essential role in forming secondary porosity in Silicalite-1. To gain deeper understanding of the effect of TPA+ during solvent-free post-treatment, tetramethylammonium bromide (TMABr), tetraethylammonium bromide (TEABr) and cetyltrimethylammonium bromide (CTABr) with increasing ionic radius were also attempted (Fig. 5). It should be mentioned that only TMA+ is small enough (∼0.44 nm) to diffuse into the micropores of Silicalite-1 (∼0.57 nm), while TEA+, TPA+ and CTA+ are too bulky to enter the micropores.36,37 The results show that the secondary porosities of similar shape and size are also observed by using TEA+ and CTA+ but not by TMA+ (which shows only a slightly melted morphology). The inability of TMA+ to generate secondary pores is probably due to its easy diffusion into the channels, which renders surface vulnerable to F− attack while the pore is blocked by TMA+ and less accessible to F−. It is also noteworthy that TEA+ and CTA+ has no structure-directing effect towards MFI zeolites, so the formation of secondary mesopores is not a recrystallization process but a preferred dissolution. Compared with TPABr-treated Silicalite-1, the TEABr-treated Silicalite-1 shows smaller, multiple macropores after the same heating treatment, while CTABr-treated Silicalite-1 shows remarkably lager macropores. Such features can be reasonably attributed to, compared with TPABr, the weaker ability of TEABr and stronger ability of CTABr to shield the surface and to keep the micropore open for F− diffusion. On such basis, the unique dissolution process in the solvent-free environment was proposed to proceed in following steps: (i) the initial reaction between NH4F and silica framework forms a surficial layer of silanol anion Si–O−, which then adsorb bulky organic cations like TPA+ and prevent further etching on the surface. (ii) The presence of stoichiometric amount of water forms a local aqueous environment, delivering F− into the zeolite and selectively etch the defective zones.
Fig. 6 shows the elemental distribution map of the unwashed post-treated samples. The map of carbon shows a very thin enriched layer in the peripheral part of the Silicalite-1, confirming that TPA+ are preferentially adsorbed on the outer surface of zeolite. The fluoride distribution, however, is quite similar with silicon distribution, confirming the free diffusion of F− into the silica framework of zeolite. After thorough washing, a large amount of TPA+ and F− that are physically adsorbed on the outer surface are removed, as evidenced by the less intense carbon and fluoride signals on the outer part of zeolite. Nevertheless, the enriched carbon distribution in the peripheral part is still present for post-treated Silicalite-1, attributed to TPA+ bonded to surface Si–O− of zeolites; while a relatively uniform fluoride distribution is present across the whole framework domain. The above observation supports our proposal that TPA+ are kept on the surface while F− freely diffuses into the zeolite framework.
Fig. 6 STEM images and EDS maps of (a–d) unwashed and (e–h) washed samples by post-treatment. Silicon (b and f), carbon (c and g) and fluoride (d and h) distribution map were separately displayed. |
Hierarchical zeolites can serve as the support to disperse38 or entrap39 active metal nanoparticles for specific catalytic applications. Especially, the encapsulation of noble and transition metal nanoparticles into hollow zeolites as core–shell or yolk–shell composites have been extensively studied for their great potential in catalytic applications.26,40 The metal nanoparticles encapsulated in such composite materials are expected to have high mono-dispersion of active sites that are strongly resistant against sintering at catalytic conditions.
Herein this solvent-free post-treatment was attempted to encapsulate platinum and cobalt nanoparticles, respectively.41–44 The loading of metal species was achieved by an impregnation method with the corresponding aqueous solution of precursor. TEM images were taken from the post-treated samples before calcination-reduction (Fig. S9†) and after calcination-reduction (Fig. 7). The results show that after grinding-heating treatment, the metal species were already transformed to nanoparticles (<10 nm) uniformly dispersed inside and outside Silicalite-1. After subsequent calcination-reduction process, larger nanoparticles emerge inside the meso-/macropores (Pt@Silicalite-1) or along the edges of meso-/macropores (Co@Silicalite-1), as well as on the outer surface (readily been washed out later). We propose the following hypothesis: firstly the metal salts are dissolved in the solution; during the impregnation, they adhere to the outer surface and enter the channels of the zeolites; as the F−-motivated dissolution proceeds, the metal species dissolved in trace amount of water diffuse with F− into the framework of Silicalite-1 and preferentially reside around the etched zones. During the heating treatment at 180 °C, these metal species convert to nanoparticles that are either sparsely dispersed around the meso-/macropores (Pt@Silicalite-1) or located on the edges of the meso-macropores (Co@Silicalite-1). The different morphologies of Pt and Co nanoparticles seem to be caused by the nature of Pt and Co elements, as similar results have been reported for Pt@Silicalite-1 and Co@Silicalite-1 zeolites prepared with the conventional hydrothermal post-treatment.13,45
Fig. 7 STEM images and EDS maps of (a and b) Pt@Silicalite-1 and (c and d) Co-Silicalite-1 samples. Note that in EDS map green colour stands for Si while red for Pt or Co. |
Upon high-temperature reduction, the nanoparticles are prone to assemble and grow larger. In the case of Pt–Silicalite-1, Pt nanoparticles diffuse into meso-/macropores to agglomerate into larger particles; while in the case of Co@Silicalite-1, the Co species agglomerate along the edges of secondary pores. The encapsulation of Pt and Co nanoparticles is further supported by the STEM-EDS results and TEM results of tilted samples (Fig. 7b, d and S10†). It is noteworthy that the absence of the water in the solvent-free route can effectively improve the utilization of metal precursors.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ra00670j |
This journal is © The Royal Society of Chemistry 2020 |