Structural reconstruction: a milestone in the hydrothermal synthesis of highly active Sn-Beta zeolites

Zhiguo Zhu , Hao Xu *, Jingang Jiang , Haihong Wu and Peng Wu *
Shanghai Key Laboratory of Green Chemistry and Chemical Process, School of Chemistry and Molecular Engineering, East China Normal University, North Zhongshan Rd. 3663, Shanghai 200062, China. E-mail: pwu@chem.ecnu.edu.cn; hxu@chem.ecnu.edu.cn; Fax: +86-21-62232292

Received 11th September 2017 , Accepted 31st October 2017

First published on 31st October 2017


A novel structural reconstruction strategy is proposed to prepare an active Sn-Beta catalyst with high Sn contents and a hydrophobic nature. Compared with post-synthesized Sn-Beta and state-of-the-art classic fluoride-mediated Sn-Beta-F, this Sn-Beta zeolite exhibits unparalleled active site-based turnover frequency for desirable products and in particular catalyst weight-based space-time-yields in various redox reactions of ketones.


Sn-Beta zeolite, containing isolated Sn ions isomorphously substituted in the *BEA framework and possessing an open pore system consisting of 3-dimensional 12-membered ring channels,1 is regarded as one of the state-of-the-art solid Lewis acid catalysts among various homogeneous and heterogeneous catalysts.2 So far, a number of extraordinary catalytic results have been obtained with the Sn-Beta catalyst, which demonstrated incomparable performance in many transformation reactions including the Meerwein–Ponndorf–Verley (MPV) reduction, Baeyer–Villiger (B–V) oxidation, Diels–Alder reaction, and especially the conversion of biomass and renewables.3–5

Nonetheless, long synthetic periods, upper-limited isolated Sn active sites, and relatively bulky crystal sizes are often encountered in the conventional hydrothermal synthesis of Sn-Beta in fluoride media.3 It is generally accepted that the excellent catalytic performance of Sn-Beta is closely related to the high content of isolated Sn (i.e. high active sites) actually incorporated into the framework, the absence of undesired SnO2-like particles as well as the small crystal sizes beneficial to reduction of intracrystal diffusion limitation.6,7 Therefore, many alternative methods such as dry gel conversion,8 interzeolite transformation,9 using Sn–Si composite oxide as a precursor,10 and post-synthesis6,7,11,12 have been developed to shorten the synthesis time and/or to reduce the crystal sizes as well as to increase the framework Sn contents. Although some points are enhanced to some extent, they still limit the large-scale industrially-relevant applicability of this promising zeolite material. For example, the synthesis process of interzeolite transformation is tedious, which requires a special silica source of ITQ-1 silicate that is synthesized using expensive N,N,N-trimethyl-1-adamantammonium hydroxide and hexamethyleneimine as organic structure directing agents (OSDAs).9 Sn-Beta zeolites prepared via the solid-state ion-change post-synthesis method exhibit poor hydrophobicity which is significantly unfriendly to catalytic reactions especially in aqueous media.7 Consequently, there is still an urgent need and a significant challenge to develop a convenient and facile strategy that is capable of directly synthesizing hydrophobic nanosized Sn-Beta zeolites with high Sn contents and within a short crystallization period as well.

In this communication, we develop a straightforward structural reconstruction route for the direct hydrothermal synthesis of hydrophobic Sn-Beta, which involves three steps as illustrated in Scheme 1. The highly dealuminated Beta zeolite (Beta-DA) was dissolved in basic media, and then the degraded Beta debris and Sn source self-assembled into Sn-Beta with the assistance of tetraethylammonium ions (TEA+) and fluoride. This innovative method not only significantly decreases the synthesis time (1 h) by two orders of magnitude in comparison to the conventional hydrothermal synthesis approach (more than 10 days), but also affords hydrophobic nanosized Sn-Beta zeolites with unpredictably high framework Sn contents (Si/Sn ratio low as 33). Additionally, the crystallization mechanism, physicochemical properties, and corresponding catalytic activity of the prepared Sn-Beta-Re zeolite are studied systematically via various techniques.


image file: c7cc06778j-s1.tif
Scheme 1 Schematic representation of the proposed crystallization mechanism for fluoride-assisted synthesis of hydrophobic nanosized Sn-Beta with unpredictably high Sn contents.

Firstly, the HNO3 treatment was performed over commercial nanosized Beta aluminosilicate zeolite (60–80 nm and Si/Al = 11) to obtain a highly siliceous Beta (Beta-DA) in which nearly all the framework Al ions were extracted (Si/Al > 1900) based on ICP analysis. Besides, the crystalline structure of Beta zeolite remained completely intact (Fig. 1Aa). Beta-DA exhibited an intensive Raman vibration band centered at 965 cm−1 (Fig. 1Ba), attributed to silanol groups,13 and a resonance band around −101.8 ppm in the 29Si MAS NMR spectrum (Fig. S1a, ESI), associated with the (SiO)3Si(OH) groups (Q3),6 confirming the appearance of vacancies after removal of the Al ions in the zeolites. Subsequently, a synthetic gel was prepared by mixing together Beta-DA zeolite, TEAOH, SnCl4·5H2O and water. After hydrothermal treatment at 413 K for 45 min, the crystalline Beta phase was dissolved completely. Both XRD and 29Si MAS NMR investigations indicated that due to this treatment the corresponding signals, closely related to the *BEA-type zeolite,9 nearly disappeared in comparison to the parent Beta-DA sample (Fig. 1 and Fig. S1b, ESI). Additionally, the SEM images showed that the Beta-DA nanocrystals were degraded and transformed into a platelet-like amorphous phase at the very beginning (Fig. S2b, ESI). A conclusion could be made from these results that the parent Beta-DA silicate was dissolved in TEAOH-containing alkaline medium, forming silica fragments, which could serve as a nutrient for facilitating the Sn-Beta-Re zeolite growth in the following crystallization process.


image file: c7cc06778j-f1.tif
Fig. 1 (A) XRD patterns and (B) UV Raman spectra of the parent highly dealuminated sample of Beta-DA (Si/Al > 1900) (a) and the corresponding solid Sn-Beta-Re products synthesized for 0 min (b), 20 min (c), 40 min (d), 60 min (e), and 1 d (f) under the crystallization conditions of SiO2/SnO2 = 30; TEAOH/SiO2 = 0.5; NH4F/SiO2 = 0.5; H2O/SiO2 = 7.5; temp., 413 K. The calcined samples were used for UV Raman characterization.

Then, various characterization techniques were employed to investigate the structure growth process for Sn-Beta-Re zeolite at a representative Si/Sn ratio, corresponding to a high Sn content never achieved by the existing direct hydrothermal synthesis method.9Fig. 1 shows the XRD patterns and UV Raman spectra of the Sn-Beta-Re samples obtained at different stages. With the crystallization proceeding, the reflection intensities around 2θ = 7.9 and 22.6°, characteristic of the *BEA structure,9 were gradually enhanced and the crystallinity was levelled off stage after heating for 1 h (Fig. S3, ESI). Besides, the characteristic diffraction in the high angle region was slightly shifted from 22.9° for Beta-DA to a lower angle of 22.6° for Sn-Beta-Re probably due to lattice expansion, which was indicative of the incorporation of Sn ions into the framework.14,15 Additionally, the UV Raman spectra of the calcined products showed the bands attributed to primary structural units of Beta zeolite, namely, five-membered rings (321 and 348 cm−1), six-membered rings (405 cm−1), and four-membered rings (469 cm−1).9,16 These bands became more and more sharp and evident. Comparably, the corresponding band intensities appeared to be constant after a synthesis time of 1 h and the Raman bands at about 631 and 773 cm−1 were not observed in Sn-Beta-Re-30, implying the absence of the SnO2-like species. The results of the crystallization time-dependent 29Si MAS NMR spectra (Fig. S1, ESI), SEM and TEM images (Fig. S2 and S4, ESI) were in good agreement with the structural evolution achieved by XRD and UV Raman measurements.

Sn-Beta-Re zeolites with various Sn contents were also prepared via this structural reconstruction method. As shown in Fig. S5 (ESI), well-crystalline Sn-Beta-Re zeolites with molar Si/Sn ratios of 30–200 were obtained without any impurity phase. Besides, it was found that Sn-Beta-Re zeolites could not be crystallized or fully crystallized when the Sn contents in the synthetic gels were too high (Si/Sn ≤ 20), probably due to the retarding effects of large Sn ions on the crystallization.17 Whereas the minimum Si/Sn ratio of Sn-Beta-F hydrothermally synthesized by the conventional method could not reach below 100. The Raman spectrum of the calcined gel, that is the Sn-Beta-Re-30P sample, obviously exhibited the primary structural units of Beta zeolite (Fig. S6a, ESI), including five-membered rings (321 and 348 cm−1), six-membered rings (405 cm−1), and four-membered rings (469 cm−1),9 which intensively facilitated nucleation and growth for the Sn-Beta-Re zeolite. However, the primary structural units of Beta zeolite were not observed in the Raman spectrum for the gel of Sn-Beta-F-30P prepared from TEOS, a Sn source and TEA+ in fluoride media (Fig. S6b, ESI).

The Sn contents actually incorporated into the Sn-Beta-Re products were almost equal to those in the gels (Table S1, ESI), indicating a high utilization efficiency of the Sn source. The micropore volume of the Sn-Beta-Re zeolites was comparable to that of conventionally synthesized Sn-Beta-F-150, while the Sn-Beta-Re zeolites possessed higher BET surface areas (∼650 m2 g−1) and external surface areas (∼210 m2 g−1) than Sn-Beta-F-150 (SBET = 552 m2 g−1, Sext = 118 m2 g−1), which is considered to be due to the unprecedented smaller crystal size obtained by structural reconstruction (30–50 nm vs. > 1 μm), as evidenced in Fig. S2e and S7a (ESI). More importantly, the time for a full crystallization of Sn-Beta-Re was as short as 1 h, two orders of magnitude lower than that of Sn-Beta-F-150 in conventional hydrothermal synthesis (14 d).

A combination of characterization techniques were employed to determine the hydrophilicity/hydrophobicity and the Sn coordination state, which are closely associated with their catalytic performances.9,18 FT-IR spectra in the hydroxyl stretching region (Fig. 2A) indicated that post-synthesized Sn-Beta-GPS-36 and Sn-Beta-SSIE-30 samples possessed comparable silanol groups, including internal silanol groups (3400–3600 cm−1) and terminal silanol groups (ca. 3745 cm−1).7 These contents were far higher than those for Sn-Beta-F-150 and Sn-Beta-Re-30. Additionally, the amount of internal silanol groups in the framework for Sn-Beta-F-150 was higher than that for Sn-Beta-Re-30, while the amount of terminal silanol groups was less for Sn-Beta-F-150 (Fig. 2A). Therefore, the comparison of hydrophilicity/hydrophobicity between Sn-Beta-Re-30 and Sn-Beta-F-150 could not be made. Then it could be speculated from 29Si MAS NMR spectra that Sn-Beta-Re-30 possessed a slightly higher hydrophobicity than Sn-Beta-F-150, far outperforming Sn-Beta-GPS-36 and Sn-Beta-SSIE-30, which resulted from the (SiO)3SiOH (Q3) contents (Sn-Beta-Re-30, 4.4%; Sn-Beta-F-150, 4.8%; Sn-Beta-GPS, 9.5%; Sn-Beta-SSIE-30, 10.3%) (Fig. 2B). Similar conclusions could also be obtained by thermogravimetric analysis (Fig. S8, ESI).


image file: c7cc06778j-f2.tif
Fig. 2 (A) FT-IR spectra in the hydroxyl stretching vibration region and (B) 29Si MAS NMR spectra for Sn-Beta-Re-30 (a), Sn-Beta-F-150 (b), Sn-Beta-GPS-36 (c), and Sn-Beta-SSIE-30 (d) samples. The dotted line indicates the recorded signals.

The dehydrated Sn-Beta-Re zeolites showed two Sn 3d XPS peaks at 487.7 and 496.2 eV (Fig. S9A, ESI) as well as a characteristic UV-vis band around 208 nm which was sharp even at an extremely high Sn loading (Si/Sn = 33 or 6.06 wt% Sn) (Fig. S9B, ESI). These signals or adsorptions are all assigned to the isolated four-fold coordination Sn species in the *BEA framework.19,20119Sn MAS NMR spectroscopy verified that the Sn species existed as isolated ones in the framework (Fig. S10, ESI). No SnO2-like species were observed in the XRD patterns, UV Raman, XPS, UV-vis spectra, and SEM-EDX (Fig. 1, Fig. S9 and S11, ESI). These characteristic results indicated that the Sn-Beta-Re zeolites synthesized using a structural reconstruction protocol, with tetrahedral Sn4+ within the framework even at Sn contents five-times higher than the traditional material, exhibited a slightly higher hydrophobicity in comparison to Sn-Beta-F-150 prepared by the conventional fluoride method, which would tremendously favor the improvement of catalytic performances of solid Lewis acids.

Isolated tetrahedral coordinated Sn4+ in the Sn-Beta framework endowed it with Lewis acidity which played a significant role in various Lewis acid-catalyzed reactions.21 The Lewis acidity of Sn-Beta-Re was detected using an FT-IR technique. Several spectroscopic bands, closely related to Lewis acidity,9,21 were clearly observed in the pyridine (1451, 1490, and 1611 cm−1) and CD3CN (2308 and 2316 cm−1) adsorption–desorption IR spectra of the Sn-Beta-Re-30 sample (Fig. S12 and S13, ESI).

The catalytic activities of the prepared zeolite catalysts were evaluated to illustrate their unparalleled advantages, including nanosized crystals, high hydrophobicity, extremely high framework Sn contents, in the Baeyer–Villiger (B–V) oxidation of 2-adamantanone with aqueous hydrogen peroxide. As shown in Table S2 (ESI), the ketone conversion (26.9%), turnover frequency per Sn site (TOF, 195 h−1), and space-time-yield per catalyst weight (STY, 3.6 h−1) obviously increased for Sn-Beta-Re-150 in comparison to the blank test, siliceous Beta-Re prepared by structural reconstruction without Sn ions in the synthetic gels, and Sn-impregnated SnO2/Beta-Re samples, illustrating that the isolated tetrahedral Sn4+ in the zeolite framework was the active site for the production of lactone. At comparable Sn contents, the catalytic activity of Sn-Beta-Re-150 (Conv. = 26.9%, TOF = 195 h−1, STY = 3.6 h−1) was found to be superior to that of Sn-Beta-F-150 (Conv. = 16.6%, TOF = 123 h−1, STY = 2.2 h−1) which was hydrothermally synthesized by the conventional fluoride method, mostly because the former has the advantages of intracrystal diffusion and mass transfer derived from its nanocrystals (Fig. S2 and S7, ESI) and high hydrophobicity (Fig. 3). On the other hand, for the control experiment, Beta-GPS-36, and Sn-Beta-SSIE-30 were post-synthesized with solid–gas isomorphous substitution with SnCl4 vapor, and solid-state ion-exchange with SnCl4·5H2O, respectively. Although a comparable crystal size was observed for Sn-Beta-Re-30, Sn-Beta-GPS-36, and Sn-Beta-SSIE-30 (Fig. S2 and S7, ESI), Sn-Beta-Re-30 with excellent hydrophobicity still exhibited higher performances than the two post-synthesized Sn-Beta zeolites at similar Sn contents (Table S1, No. 8 and 9, ESI). With the increase in the Sn loading, the initial reaction rate for lactone production increased linearly (Fig. S14, ESI), further evidencing that the isolated framework Sn species are responsible for the catalytic performances. The initial reaction rate for Sn-Beta-Re was beyond that for Sn-Beta-F, Sn-Beta-GPS, and Sn-Beta-SSIE. Simultaneously, the catalytic activities in terms of conversion and space-time-yield were improved incredibly, in the case of the Sn-Beta-Re-30 catalyst, far outperforming the benchmark Sn-Beta-F-150 material.


image file: c7cc06778j-f3.tif
Fig. 3 (A) Dependence of the relative conversion of 2-adamantanone on the amount of extra added water and (B) reusability of Sn-containing Beta over Sn-Beta-Re-30 (a), Sn-Beta-F-150 (b), Sn-Beta-GPS-36 (c), and Sn-Beta-SSIE-30 (d) in the Baeyer–Villiger oxidation of 2-adamantanone. Relative conversion indicates the percentage of conversion without adding water into the reaction system. Reaction conditions: (A) cat, 50 mg; 2-adamantanone, 2 mmol; TBHP (5.5 M in decane), 4 mmol; chlorobenzene, 10 mL; temp., 363 K; time, 8 h. (B) See Table S2 (ESI). The used catalysts were regenerated by washing with chlorobenzene after each run, but by further calcination (823 K, 6 h) after the fourth run.

To further demonstrate the preponderance of the high hydrophobicity of the Sn-Beta-Re catalyst, the dependence of the relative conversion of 2-adamantanone on the amount of extra added water in the B–V oxidation using TBHP (in decane) as the oxidant was studied. It was generally accepted that the higher hydrophobicity of the catalyst, the less the impact of water in the reaction system on the catalytic performances under suitable conditions.22 As can be seen from Fig. 3A, with the increase in the amount of extra added water, the relative conversion of 2-adamantanone gradually decreased. Note that the slowest reduction in respect of relative conversion was observed for Sn-Beta-Re in comparison to other Sn-Beta catalysts, resulting from the high hydrophobicity of Sn-Beta-Re.

The reusability of these Sn-Beta zeolites in the B–V oxidation of 2-adamantanone with aqueous H2O2 was also tested (Fig. 3B). The conversion of Sn-containing Beta catalysts decreased gradually when the used catalyst was washed only with chlorobenzene after each run. After the fourth run, the used catalyst was regenerated by calcination in air, the activity for Sn-Beta-Re and Sn-Beta-F could be nearly restored to the initial level, whereas the post-synthesized Sn-Beta-GPS-36 and Sn-Beta-SSIE-30 catalysts could not be regenerated totally. The structures of all Sn-Beta zeolites were well retained after the reaction (Fig. S15, ESI) and the Sn contents were nearly the same as the initial values according to the ICP data. Surprisingly, Sn-Beta-Re-30 and Sn-Beta-F-150 showed nearly unaltered Sn coordination states (Fig. S16A and B, ESI). Conversely, a clear increase in extra-framework Sn species was observed for post-synthesized Sn-Beta-GPS-36 and Sn-Beta-SSIE-30 (Fig. S16C and D, ESI). At this point, it was inferred that the coke deposition mainly contributed to the deactivation of Sn-Beta-Re-30 and Sn-Beta-F-150, whereas that of the post-synthesized Sn-Beta was ascribed to pore blocking by organics and Sn active site restructuring.

The generalized applicability of Sn-Beta-Re in the isomerization–esterification reaction of dihydroxyacetone (DHA) in ethanol and the Meerwein–Ponndorf–Verley (MPV) reaction in isopropanol with cyclohexanone further confirmed that Sn-Beta-Re zeolites hydrothermally synthesized via structural reconstruction are productive and reliable Lewis catalysts (Fig. S17 and S18, ESI).

In conclusion, an innovative and convenient direct hydrothermal synthesis strategy for heteroatom-containing zeolites, in this case Sn-Beta, via structural reconstruction is developed. The Sn-Beta-Re zeolite with incredibly high framework Sn contents (Si/Sn = 33 or 6.06 wt%), nanosized crystals (30–50 nm), and outstanding hydrophobicity was crystallized completely within 1 h at 413 K, the time of which was shortened by two orders of magnitude in comparison to conventional methods. Additionally, the main raw materials required for this synthesis, including an organic structure directing agent (tetraethylammonium hydroxide) and a silica source (siliceous Beta from Beta aluminosilicate), are simple and commercially available. Furthermore, compared with post-synthesized Sn-Beta and state-of-the-art Sn-Beta-F materials, Sn-Beta-Re zeolites exhibited unparalleled catalytic performances in terms of turnover frequency and space-time-yield. This hydrothermal synthesis route has potential for highly productive zeolite synthesis on an industrial scale.

The authors gratefully acknowledge the financial support from the NSFC of China (21533002, 21373089, 21603075), China Ministry of Science and Technology (2016YFA0202804).

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. J. M. Newsam, M. M. J. Treacy, W. T. Koetsier and C. B. De Gruyter, Proc. R. Soc. Lond. A, 1988, 420, 375 CrossRef CAS .
  2. M. S. Holm, S. Saravanamurugan and E. Taarning, Science, 2010, 328, 602 CrossRef CAS PubMed .
  3. A. Corma, L. T. Nemeth, M. Renz and S. Valencia, Nature, 2001, 412, 423 CrossRef CAS PubMed .
  4. A. Corma, M. E. Domine, L. Nemeth and S. Valencia, J. Am. Chem. Soc., 2002, 124, 3194 CrossRef CAS PubMed .
  5. M. Moliner, Y. Román-Leshkov and M. E. Davis, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 6164 CrossRef CAS PubMed .
  6. P. Li, G. Liu, H. Wu, Y. Liu, J. Jiang and P. Wu, J. Phys. Chem. C, 2011, 115, 3663 CAS .
  7. C. Hammond, S. Conrad and I. Hermans, Angew. Chem., Int. Ed., 2012, 51, 11736 CrossRef CAS PubMed .
  8. Z. Kang, X. Zhang, H. Liu, J. Qiu and K. L. Yeung, Chem. Eng. J., 2013, 218, 425 CrossRef CAS .
  9. Z. Zhu, H. Xu, J. Jiang, Y. Guan and P. Wu, J. Catal., 2017, 352, 1 CrossRef CAS .
  10. T. Iida, A. Takagaki, S. Kohara, T. Okubo and T. Wakihara, ChemNanoMat, 2015, 1, 155 CrossRef CAS .
  11. W. N. P. van der Graaff, G. Li, B. Mezari, E. A. Pidko and E. J. M. Hensen, ChemCatChem, 2015, 7, 1152 CrossRef CAS .
  12. A. Al-Navili, K. Yakabi and C. Hammond, J. Mater. Chem. A, 2016, 4, 1373 Search PubMed .
  13. X. Yang, L. Zhou, C. Chen, X. Li and J. Xu, Mater. Lett., 2009, 63, 1754 CrossRef CAS .
  14. B. Tang, W. Dai, G. Wu, N. Guan, L. Li and M. Hunger, ACS Catal., 2014, 4, 2801 CrossRef CAS .
  15. J. E. Schmidt, D. Fu, M. W. Deem and B. M. Weckhuysen, Angew. Chem., Int. Ed., 2016, 55, 16044 CrossRef CAS PubMed .
  16. T. Ikuno, W. Chaikittisilp, Z. Liu, T. Iida, Y. Yanaba, T. Yoshikawa, S. Kohara, T. Wakihara and T. Okubo, J. Am. Chem. Soc., 2015, 137, 14533 CrossRef CAS PubMed .
  17. S. Tolborg, A. Katerinopoulou, D. D. Falcone, I. Sádaba, C. M. Osmundsen, R. J. Davis, E. Taarning, P. Fristrup and M. S. Holm, J. Mater. Chem. A, 2014, 2, 20252 CAS .
  18. J. D. Lewis, S. Van de Vyver and Y. Román-Leshkov, Angew. Chem., 2015, 127, 9973 CrossRef .
  19. M. P. Pachamuthu, K. Shanthi, R. Luque, A. Ramanathan and Y. Román-Leshkov, Green Chem., 2012, 15, 2158 RSC .
  20. I. Hermans, S. Conrad, P. Wolf, P. Mueller and H. Orsted, ChemCatChem, 2017, 9, 175 CrossRef .
  21. S. R. Bare, S. D. Kelly, W. Sinkler, J. J. Low, F. S. Modica, S. Valencia, A. Corma and L. T. Nemeth, J. Am. Chem. Soc., 2005, 127, 12924 CrossRef CAS PubMed .
  22. G. M. Lari, P. Y. Dapsens, D. Scholz, S. Mitchell, C. Mondelli and J. Pérez-Ramírez, Green Chem., 2016, 18, 1249 RSC .

Footnote

Electronic supplementary information (ESI) available: Details of experimental procedures and material characterization. See DOI: 10.1039/c7cc06778j

This journal is © The Royal Society of Chemistry 2017