3.1 Synthesis and Characterization
Figure 1 is the XRD patterns of the as-synthesized samples yielded from the different gel precursors. It can be seen that all of the as-synthesized ZSM-5 samples display diffraction peaks at 7.88°, 8.76°, 23.0°, 23.84°, and 24.3°, which can be ascribed to the characteristic diffraction peaks of MFI topology (JCPDS: 44 − 00033). And no impure phase can be detected, indicating that the as-prepared samples are pure ZSM-5 zeolite [2, 10]. Figure 1 also shows that an increased Si/Al ratio of precursors results in an obviously enhanced diffraction peaks strength (along with the shrinked half-peak breadth) at 7.88°, 8.76° for the samples ZSM-5-20, ZSM-5-40, and ZSM-5-80. In addition, a precursor containing ethyl alcohol yields the sample ZSM-5-40-5.32 with increased diffraction peaks strength at 7.88°, 8.76°, indicating that the ethyl alcohol in the gel precursor contributes to the crystallization of the ZSM-5 crystals in the sample. It was supposed that ethanol and sodium ions form tetrahedron and played a template role, which accelerates the nucleation and crystallization [37, 38]. Moreover, as compared with the samples ZSM-5-40 and ZSM-5-40-5.32, ZSM-5-40-5.32-2.1 has an obviously increased half-peak breadth, suggesting that the added TPOAC in the precursor does not affect the nucleation of ZSM-5 zeolite, while may bring an effect on the growth of the crystals in the as-synthesized sample ZSM-5-40-5.32-2.1. [39]
As shown in Fig. S1, the Si/Al ratio of the gel precursors has a prominent effect on the as-synthesized zeolite samples. It can be seen that all of the as-synthesized ZSM-5-x yielded from the synthesis gels with a series of Si/Al ratios (10–80) display the typical MFI topological characteristic diffraction peaks at 7.88°, 8.76°, 23.0°, 23.84°, and 24.3°. Fig. S1 also shows that when the Si/Al ratio of the gel precursors is lower than 18, impure peaks at about 9.06°, 12.68°, 21.88°, 28.38°, and 33.54° appear; Combining with the results as shown in Fig. 1, it can be inferred that the Si/Al ratio of the gel precursors higher than 20 is a key condition for yielding the pure ZSM-5 zeolite with a typical MFI topology in the present work.
As shown in Fig. 2 and Fig. S2, the crystal morphologies of the as-synthesized zeolite samples are determined by the Si/Al ratios of the gel precursors. It is seen that all of the crystals in the as-synthesized samples ZSM-5-x (x = 30–50) are chain-like polycrystalline ZSM-5 zeolite aggregates consisting of about several to more than ten primary crystals, which connect to each other and then forms chain-like polycrystalline aggregates (Fig. 2A-D; Fig. S2). The crystals in the sample ZSM-5-20 consists of three structural levels: the primary crystals with a size of about less than 100 nm, and hundreds of the primary nanocrystals together into a polycrystalline aggregate with a size of about 600 nm, which further connect each other and then forms the third level, namely, the chain-like polycrystalline ZSM-5 zeolite. It can be seen from Fig. 2 and Fig. S2 that when the Si/Al ratio of the precursor is lower than 30, the chain-like crystals is shorter, consisting of about 3–4 primary ZSM-5 zeolite crystals (Fig. S2); Increased Si/Al ratio of the precursor offers the as-synthesized sample with longer chain-like crystals. For example, the chain-like polycrystalline aggregates in ZSM-5-40 and ZSM-5-50 are composed of more than a dozen of primary ZSM-5 zeolite crystals. (Fig. 2D and Fig. S2). It can be seen that the primary ZSM-5 crystals in sample ZSM-5-40-5.32 is about 250 nm, obviously larger than the 120 nm of the primary crystals in the chain-like ZSM-5-40 zeolite sample (Fig. 2D and 2F). That can be ascribed to the template effect of the ethanol [37, 38], which promotes the faster growth of the primary crystals in the as-synthesized, and is in good agreement with the results as shown in Fig. 1. However, the crystals in the as-synthesized sample ZSM-5-80 are the typical monodispersed coffin-like ZSM-5 zeolite crystals with a size of about 3 µm×0.8 µm×0.4 µm (Fig. 2I, 2G). The secondary template TPOAC signally changes the morphology of the primary crystals in the chain-like aggregates while does not change the chain length. As shown in Fig. 2G and 2H, due o the added of TPOAC in the precursor, the chain-like polycrystals in sample ZSM-5-40-5.32-2.1 become rather coarse, and the chain-like morphology endows the crystals more like “caterpillars” (Fig. 2G). As exhibited in Fig. 3, “caterpillars” crystals in sample ZSM-5-40-5.32-2.1 are rather loosened and many interspaces, resulted from the pore-forming role of the TPOAC [39], can be seen indistinctly. While the crystals in the ZSM-5-40 rare rather smooth and dense.
Figure 4 is the N2 adsorption-desorption isotherms and the corresponding DFT (Density Functional Theory) pore size distribution curves of the as-synthesized samples. It can be seen from Fig. 4A that the N2 sorption isotherm curves exhibit a steep increase at the relative pressure p/p0 < 0.02 and a hysteresis loop at a relative pressure p/p0 of 0.8–1.0, indicating the co-existence of intrinsic micropores and meso- or/and macropores in the samples. The hysteresis loop at a relative pressure p/p0 of 0.8–1.0 can be attributed to the capillary condensation [10–11, 30, 39] in open mesopores or macropores obtained by filling the primary MFI zeolite interparticles spaces. In addition, a hysteresis loop at a relative pressure p/p0 of 0.2–0.3, which is present in the nitrogen isotherms of sample ZSM-5-80, does not suggest a real porosity in this relative pressure range. It has been reported that this hysteresis loop can be ascribed to a fluid-to-crystal-like phase transition of nitrogen molecules in the micropores [40–42] and the appearance of hysteresis ring is related to the Si/Al ratio of the as-synthesized ZSM-5 zeolite [41, 42]. This phase transition phenomenon is limited to MFI-type zeolites, due to the specific framework properties. From another point of view, the presence of this hysteresis loop indicates that hierarchical zeolite is well crystallized [41]. Moreover, a hysteresis loop at p/p0 = 0.45–0.8 is present in the nitrogen isotherm of sample ZSM-5-40-5.32-2.1, which can be ascribed to the capillary condensation [39] mesopores left by the secondary template TPOAC after removing by calcination.
The DFT pore structure distribution calculated from the adsorption branch of the isotherms are shown in Fig. 4B. The pore distribution centers at about 3.7 nm can be the so-called pseudo pore caused by fluid-to-crystal-like phase transition of nitrogen molecules in the micropores [40–42]. A pore distribution range from 4 nm to 14 nm is obviously observed in the sample ZSM-5-40-5.32-2.1, suggesting that the sample is a hierarchically porous ZSM-5 zeolite.
Table 1
Physical structural properties of samples
Sample | SBET (m2/g) | Smic (m2/g) | Sext (m2/g) | VTotal (cm3/g) | Vmic (cm3/g) | Vmeso (cm3/g) |
ZSM-5-20 | 393 | 300 | 93 | 0.20 | 0.12 | 0.08 |
ZSM-5-40 | 453 | 358 | 95 | 0.26 | 0.14 | 0.12 |
ZSM-5-40-5.32 | 419 | 343 | 76 | 0.21 | 0.14 | 0.07 |
ZSM-5-40-5.32-2.1 | 451 | 289 | 162 | 0.39 | 0.11 | 0.28 |
ZSM-5-80 | 391 | 340 | 51 | 0.17 | 0.13 | 0.04 |
As shown in Table 1, the samples ZSM-5-40, ZSM-5-40-5.32, and the ZSM-5-80 have lager microporous areas, suggesting the high crystalline structure, which agrees the observed results as shown in Fig. 1 very well. The sample ZSM-5-20 has a small microporous area can be ascribed to the small primary nano-sized ZSM-5 crystals as shown in Fig. 2B. Obviously, smaller granule of the zeolite crystals means more defects and lower order in the crystal framework, which will lead to a weakened microporous property [6, 12]. Moreover, it can be observed from Table 1 that as compared to ZSM-5-40 and ZSM-5-40-5.32, the sample ZSM-5-40-5.32-0.21 has a weakened microporous area of 289 m2/g, while a larger external surface area of 162 m2/g and a mesoporous volume of 0.28 cm3/g. That indicates that slightly sacrificial microporous property because of the decreased three-dimensional order [35–36] of the zeolite framework due to the introduction of the secondary template TPOAC can effectively produce an interpenetrated hierarchical pore system in ZSM-5 zeolite.
It is generally recognized that lower crystallization temperature is conductive to the formation of small crystal particles, the effect of crystallization temperature on the as-synthesized sample was done. As shown in Fig. S3 and Fig. S4, a lower hydrothermal temperature than 140 ℃ for treating the precursor is not appropriate to prepare chain-like hierarchical ZSM-5 zeolite samples, and results in a sample with a lower crystallinity. The result about the effect of different crystallization time on the as-synthesized sample is displayed in Fig. S5. It can be seen that all of the characteristic diffraction peaks attributing to the MFI topological structure appears in the XRD pattern of the sample yielded from the precursor after hydrothermally treated for 6 h, and with the prolonged crystallization time, the strength of the characteristic diffraction peaks at 7.88°, 8.76°, 23.0°, 23.84°, and 24.3°consistently increases, suggesting an enhanced crystallinity. Fig. S6 shows that a shorter time offers the sample, for example the one yielded from the precursor after treated for 6 h, with abundant amorphous-form phase. Fig. S6 also exhibits that the increased treatment time, the length of the chain-like polycrystalline aggregates is obviously elongated, proposing that the gradual growth process (in-situ self-assembly) of the so-called “caterpillar”. As shown in Fig. 5, the HRTEM images of the as-synthesized ZSM-5-40 suggests that the growth of the “caterpillar” is able to realize by orderly self-assembly of ZSM-5 primary nano-sized crystals along the (010) lattice plane. And this leads to the growth of “caterpillar” along the b-orientation of MFI crystals.[43–44] Hu et al [44]reported a chainlike hierarchical ZSM-5 was synthesized by using sucrose as a template, and the chain length and mesoporous structure of ZSM-5 were controlled by the amount of sucrose in the precursor. They therefore suggested that the presence of sucrose in the precursor of the zeolite was critical for the formation of the chainlike and mesoporous structure because polar C–OH groups in the sucrose interacted with polar species. The sucrose template interacted with the aluminosilicate species through hydrogen bonding, and such an interaction played a significant role in directing the formation of a chainlike and hierarchical structure. Wang et al [43] considered that the formation of disk crystals with proper dimension and flat surface having abundant hydroxyl groups was crucial to the growth of chainlike ZSM-5 crystals, and the condensation of Si − OH groups on the (010) facet was energetically more favorable than that on other facets.
It can be inferred from Fig. S7-S8 that aging temperature of the precursors also works well in the formation of the chain-like hierarchical ZSM-5 zeolite. The higher aging temperature results in the shorter length of the chain. When the aging temperature is 80 ℃ provides the as-synthesized sample with the mono-dispersed ZSM-5 zeolite crystals with a size of about 30–100 nm. As shown in Fig. S9-S10, a aging time for 3 h is enough to produce a chain-like hierarchical ZSM-5 zeolite.
Figure 6 is NH3-TPD curves of the H-formed ZSM-5 zeolite samples, all of the H-formed catalysts display two desorption peaks, one centers at about 218–245℃, the other centers around 429–456℃, attributing to the desorbed NH3 from the weak acid sites and the strong acid sites of the catalysts [11–12, 45], respectively. It can be seen that for the sample ZSM-5-20, ZSM-5-40, and ZSM-5-80, both of the desorption peaks from the weak acid sites and the strong acid sites shift toward high temperature with the deceased Si/Al ratio. It can be observed that the desorption peaks from the weak acid sites shift from the 223 ℃ of ZSM-5-80 to the 235 ℃ of ZSM-5-40, and to the 245 ℃ of ZSM-5-20, and the strong acid desorption peaks shift from the 436 ℃ of ZSM-5-80, to the 446 ℃ of ZSM-5-40, and to the 456 ℃ of ZSM-5-20, respectively. It can be seen from Table 2 that the acid density whether the weak acid sites or the strong acid sites consistently decrease from the sample ZSM-5-20, to ZSM-5-40, and to ZSM-5-80 because of the increased Si/Al ratio.
Table 2
Acid strength distribution of the H-ZSM-5 zeolite catalysts
Sample | Si/Al ratio① | Acid sites (µmol/g) |
Weak | Strong | Total |
ZSM-5-20 | 20.9 | 250 | 264 | 514 |
ZSM-5-40 | 29.7 | 144 | 182 | 326 |
ZSM-5-40-5.32 | 32.6 | 120 | 138 | 258 |
ZSM-5-40-5.32-2.1 | 58.9 | 59 | 87 | 146 |
ZSM-5-80 | 34.2 | 86 | 128 | 214 |
①the Si/Al ratio was determined by XRD patterns
It can be inferred from Fig. 6 that for the samples ZSM-5-40, ZSM-5-40-5.32, and ZSM-5-40-5.32-2.1, both of the ethanol and TPOAC in the gel precursors affect the aluminium species entering the frameworks of ZSM-5 [45–46]. As shown in Table 2, the Si/Al ratio of ZSM-5-40, ZSM-5-40-5.32, and ZSM-5-40-5.32-2.1 samples determined by XRD patterns of Fig. 1 are 29.7, 32.6, and 58.9, respectively. It can be speculated that for the three catalysts, the acid density follows the order as: ZSM-5-40 > ZSM-5-40-5.32 > > ZSM-5-40-5.32-2.1. As shown in Table 2, the calculated acid density of ZSM-5-40 is 326 µmol/g, larger than the 258 µmol/g of ZSM-5-40-5.32, and much larger than the 146 µmol/g of catalyst ZSM-5-40-5.32-2.1. As exhibited in the “Synthesis” section, the three samples are obtained by hydrothermal treatment of the gel precursors with the same Si/Al ratio. Such a huge difference in the Si/Al ratios along with the acid density of the three samples strongly suggests that the added ethanol especially the added TPOAC affect the interaction between silicon and aluminium species, and TPOAC may contribute to more silicon species entering the zeolite framework [45]. The negative impact of TPOAC on incorporating Al species into the zeolite framework [45–46] results in the reduced acid sites and the weaker acidity of hierarchical ZSM-5-40-5.32-2.1.
3.2 Catalytic test
ZSM-5 is usually used as an active additive in the FCC catalyst for elevating the selectivity of olefin, while, the inherent microporous channel system in ZSM-5 zeolite often weakens its performance involving the catalytic cracking of heavy reactant. [34] The aforementioned result as shown in Table 1 exhibits the as-synthesized hierarchical ZSM-5 with the increased external surfaces. It is obvious that the increased external surfaces of is beneficial to promoting the accessibility of active sites to reactants.[5, 7, 12] For the hierarchical zeolite catalyst, the higher external surface conduces to the pre-cracking of the heavy reactants, while the acid centers confined in the micropores is conducive to a deeply cracking of the pre-cracking intermediate products into the destination products.[32, 34] Catalytic cracking of TIPB is chosen as a probe reaction so as to explore the effect of the increased external surfaces on the catalytic performances of the hierarchical ZSM-5 zeolite catalyst.
As shown in Fig. 7, due to the different structural and textural properties, the catalysts display an obvious difference in the catalytic activity. Catalyst ZSM-5-80, which has the biggest zeolite crystals as shown in Fig. 3, displays the lowest catalytic activity, may be ascribed to its low external surfaces along with the lower acid density because of the higher Si/Al. For the hierarchical zeolite catalysts, it can be speculated that ZSM-5-40 should be more active than the corresponding ZSM-5-40-5.32-2.1 catalyst because of higher density of acid sites in the former catalyst as shown in Fig. 6. While, catalyst ZSM-5-40-5.32-2.1 unexpectedly has a higher catalytic activity in TIPB cracking than the corresponding ZSM-5-40 catalyst. That can be ascribed to its higher external surfaces for the former catalyst. A higher external surface means more pore windows can be accessed by the heavy reactant molecules. [10] For the catalytic cracking of TIPB, the kinetic diameter of reactant is about 0.94 nm, which is obviously much larger than the inherent micropore channels (about 0.58 nm) of ZSM-5 zeolites. As a consequence, the bulky guest molecules, for example TIPB can approach to the microporous windows but can not enter the microporous channels of the ZSM-5 catalyst [47]. And the most of the acid sites confined in the micropores are inaccessible for the heavy reactant molecules, only the acid sites on the external surfaces closed to the microporous windows are in accessibility and can be utilized by the reactant molecules. In other words, the cracking reactions of TIPB can only take place on the external surface of the catalysts. This means that not all acid centers can be effectively utilized, and only the acid sites located the external surfaces are effective and accessible to the bulky reactants. [12, 34]
Furthermore, the product selectivity provides powerful information to explain their catalytic performances. It is widely accepted that DIPB, isopropylbenzene (IPB), benzene and propylene are the major products during TIPB cracking [34]. The result shows that the products mainly contain benzene, DIPB, propylene during the catalytic cracking of TIPB, and the IPB in the products is negligible. It can be inferred from Fig. 7B, 7D, the DIPB in the cracking products over the catalyst ZSM-5-40-5.32-2.1 is much lower than those over the ZSM-5-80 and the corresponding ZSM-5-40, while the selectivity of benzene over the hierarchical ZSM-5-40-5.32-2.1 is the highest among the three catalysts. That means that TIPB over the catalyst ZSM-5-40-5.32-2.1 suffers from a deeply cracking, also suggests that when used as a cracking catalyst, the hierarchical ZSM-5-40-5.32-2.1 can offer the bulky reactants a multi-stage cracking process [34] because not only TIPB but the pre-cracking products, for example, the DIPB and IPB can also be further cracked by the acid in the intercrystalline mesopores as detected by SEM, TEM images and the N2 adsorption-desorption isotherms.
In addition, as can be seen from Fig. 7C, for all of the catalysts, propylene selectivity over the ZSM-5-40-5.32-2.1 is about the same as that over the corresponding ZSM-5-40. It can be inferred from Fig. 7B that the selectivity of benzene over the hierarchical ZSM-5-40-5.32-2.1 is much higher than that over ZSM-5-40. During the catalytic cracking of TIPB, higher benzene selectivity should mean a higher theoretical selectivity of propylene. The depressed propylene selectivity in the cracking products may be caused by the more serious hydrogen transfer reaction [48–49]. Because the reaction involved in hydrogen transfer is a bimolecular reaction process, which need much more spaces than the monomolecular reaction. The additional mesopores in the hierarchical ZSM-5-40-5.32-2.1 offer more capacious spaces contributing to the hydrogen transfer of propylene, which reduces the selectivity of propylene in the final products.