Rapid synthesis of Sn-Beta for the isomerization of cellulosic sugars

Abstract

Sn-Beta, a tin-containing molecular sieve with the zeolite beta topology, was rapidly synthesized in two days through a modified seeding method, and its high catalytic activity for the isomerization of cellulosic sugars was demonstrated.

---

Because of their unique catalytic activity and excellent hydrothermal stability, zeolites, aluminosilicate molecular sieves, have been extensively used in petrochemical processing and the production of high-value chemicals and biofuels from naturally abundant biomass.^1,2 Instead of being used as solid Brønsted acid catalysts, molecular sieves containing tetrahedrally coordinated Ti and Sn have been explored as solid Lewis acid catalysts for redox reactions.³ Ti-containing, high-silica molecular sieves with the zeolite beta topology (Ti-Beta) and MFI topology (TS-1) have been employed for various selective oxidation reactions, such as olefin epoxidation, selective oxidation of alcohols, hydroxylation of phenol and ammoximation of cyclohexanone.⁴ Sn-Beta, a tin-containing molecular sieve with the zeolite beta topology, has been used in the Meerwein–Ponndorf–Verley (MPV) reduction of aldehydes and ketones, the Meerwein–Ponndorf–Verley–Oppenauer (MPVO) oxidation of alcohols, and the Baeyer–Villiger oxidation reaction.⁵ Recently, due to its particular Lewis acidic properties, Sn-Beta has been shown to catalyze the isomerization reactions of triose sugars (dihydroxyacetone and glyceraldehyde), pentose sugars (xylose and xylulose) and hexose sugars (glucose and fructose) with activities that are comparable to biological processes.⁶ In particular, it has been revealed that Sn-Beta is a water tolerant Lewis acid catalyst, and can catalyze the isomerization reactions in aqueous phase at low pH, which is most likely due to its hydrophobic nature derived from the high-silica microporous structure.^6c,7 Because of its unique properties, Sn-Beta has also been used for the one-pot synthesis of 5-(hydroxymethyl)-furfural (HMF), an important precursor for the production of renewable polymers and biofuels, from glucose by combining with a homogeneous acid catalyst (HCl) in a biphasic system.

Although Sn-Beta has shown promising catalytic properties, its industrial applications and related researches in academia have been hindered by the difficulties in synthesizing this material, particularly the use of hydrofluoric acid and the long crystallization time. In general, active Sn-Beta is synthesized using the fluoride anion as a mineralizing agent under near-neutral conditions, with a crystallization time of around 40 days as reported by Moliner et al.^6c The long crystallization time could be due to the relatively low supersaturation degree, and the limited nucleation caused by the fluoride anion and the neutral pH used in the synthesis. To reduce the crystallization time, a seeded growth method was applied to the synthesis of Sn-Beta. However, it still requires from 22 days to 30 days to accomplish the synthesis.^5a

In this study, we report that the morphology and dispersion status of zeolite seeds in the synthesis gel can substantially affect the crystal growth kinetics of Sn-Beta. By uniformly distributing crystalline zeolite beta seeds in the synthesis gel, high-quality Sn-Beta can be synthesized in only 2 days with a nearly complete conversion (>90%) of the provided silica source. The Sn-Beta catalyst synthesized by this approach is highly active for the isomerization of triose (C3), pentose (C5) and hexose (C6) sugars.

In contrast to the previous seeded growth method, in this study 200 nm crystalline zeolite beta nanocrystals were used as seeds and added to the synthesis mixture as a suspension. Crystalline zeolite beta nanocrystals (Si : Al = 23) were prepared according to the previous literature.⁸ In order to avoid the irreversible aggregation caused by calcination and drying, dealumination of the zeolite seeds was carried out by directly treating the stable seed solution with a concentrated nitric acid solution. The dealuminated zeolite beta seeds were collected by centrifugation and thoroughly washed with deionized water until the pH of the supernatant was close to neutral. The final concentration of the obtained seed solution was adjusted to 0.145 g mL⁻¹ by dispersing the seeds in deionized water. During the whole process, no drying or calcination was performed on the sample, which enabled us to prepare a stable suspension with well-dispersed dealuminated zeolite seeds. The crystallinity of the seeds showed no sign of a significant change after the dealumination process, as illustrated by the XRD patterns (Fig. S1, ESI†). After the dealumination, no detectable Al was found in the seeds by elemental analysis. Details of the seed synthesis and the dealumination process can be found in the supplementary information.†

For the synthesis of Sn-Beta, a clear synthesis solution was made by adding tetraethylorthosilicate (TEOS) to tetraethylammonium hydroxide solution (TEAOH). Tin(IV) chloride was first dissolved in deionized water before being added to the prepared clear solution. The resulting solution was stirred in a hood until the ethanol generated from the hydrolysis of TEOS was completely evaporated. Next, HF was added with stirring, and the solution turned into a dry gel at this stage. Finally, the suspension containing dealuminated zeolite seeds (4.1 wt% seeds with respect to the silica content in the dry gel) were added directly into the dry gel and homogenized. The composition of the final gel was SiO₂ : 0.54 TEAOH : 0.54 HF : 0.008 SnO₂ : 7.5 H₂O. The hydrothermal synthesis was carried out in a Teflon-lined stainless steel autoclave at 140 °C with a rotation of 2 rpm. Synthesis details can be found in the supplementary information.†

Fig. 1 shows the characterization of the zeolite beta seeds and the Sn-Beta synthesized in this work. The SEM image of the zeolite beta indicates that the size of the spherical seeds is around 200 nm (Fig. 1a). The XRD pattern (Fig. 1b) indicates that highly crystalline Sn-Beta can be achieved after 52 h by this approach. The SEM images (Fig. 1c and Fig. S2, ESI†) reveal that the discrete Sn-Beta crystals are highly intergrown. The size of the primary crystal was around 1 μm, and the size of the secondary particles ranged from several to tens of μm. Nitrogen sorption measurement of the calcined crystals further confirmed the high crystallinity of the material. The adsorption–desorption isotherms are typical of microporous materials (type I). The noticeable hysteresis loop located at ∼0.5 < P/P₀ < 0.8 indicates that mesopores exist in the structure, which could be a result of the highly intergrown structure as shown in the SEM image (Fig. 1c). The calculated micropore volume and BET surface area (0.19 cm³ g⁻¹ and 488 m² g⁻¹, respectively, Table S1, ESI†) are similar to the values reported in the literature.^6d,9 It is also noteworthy that the yields of the Sn-Beta (calculated from the calcined samples) were consistently above 90% with respect to the amount of SiO₂ + SnO₂ in the synthesis gel when the synthesis time was longer than 52 h. The final Si : Sn ratio of the Sn-Beta made by this method after 52 h was 126 (Table S1, ESI†). In comparison, the samples collected after 12 h and 24 h were a mixture of the crystalline phase and unreacted amorphous phase as indicated by the XRD patterns and SEM images (Fig. S3, ESI†). The UV-vis spectrum of the Sn-Beta showed absorbance from 200 nm to 250 nm, indicating the presence of Sn in the sample, although the coordination state of Sn is not conclusive from the measurement (Fig. S4, ESI†). In addition, IR spectra of adsorbed pyridine on the Sn-Beta in the range of pyridine ring-stretching modes were measured to demonstrate the Lewis acidity of the Sn-Beta catalyst. The bands at 1610, 1490, and 1450 cm⁻¹, attributed to different vibrational modes of the pyridine molecule interacting with Sn species within the molecular sieve, were observed from the Sn-Beta sample, revealing the presence of Lewis acidity in the sample (Fig. S5, ESI†). These bands remained even after desorption at 723 K, indicating the stability of the Lewis acid sites at relatively high temperatures. Due to the high Si : Sn ratio in the Sn-Beta sample (Si : Sn = 126, 1.6 wt%) and the complex structure of the zeolite beta topology, the ¹¹⁹Sn MAS NMR measurement of the sample did not show a measurable signal rising above the noise, limiting the identification of the coordination environments of Sn in the framework. Further studies on the coordination environments of Sn using ¹¹⁹Sn-enriched reactants are under investigation.

Fig. 1 Characterization of zeolite beta seeds and synthesized Sn-Beta. (a) SEM image of zeolite beta seeds; (b, c, d) XRD pattern, SEM image and N₂ adsorption–desorption isotherms of the synthesized Sn-Beta (52 h).

The catalytic performance of the synthesized Sn-Beta was first tested on the isomerization of glucose in aqueous phase at 90 °C. The main products of the isomerization were fructose and mannose. The conversion of glucose after 2 h was 54%, and the yields of fructose and mannose were 36% and 9%, respectively (Fig. 2a). Similar results were reported by Moliner et al.^6c and Lew et al.^6d In their studies, the Sn-Beta catalysts were made by the conventional method, with a crystallization time of 22 days to 40 days, depending on if zeolite seeds were used. To evaluate the reproducibility of our synthesis method, Sn-Beta samples made from 6 different batches were tested on the isomerization of glucose. The standard deviation of their activities, an indicator of reproducibility, is less than 4% (Fig. S6, ESI†).

Fig. 2 Yields of major products as a function of reaction time for the reactions of cellulosic sugars catalyzed by the Sn-Beta. (a) Isomerization of glucose in aqueous phase. Reaction conditions: initial glucose concentration of 10 wt%, glucose to tin molar ratio of 50 : 1, 100 mg Sn-Beta, 90 °C; (b) reaction of dihydroxyacetone (DHA) in methanol. Reaction conditions: 1.25 mmol DHA, 4 g methanol, 80 mg Sn-Beta, 70 °C; (c) isomerization of xylose in aqueous phase. Reaction conditions: initial xylose concentration of 10 wt%, xylose to tin molar ratio of 70 : 1, 78 mg Sn-Beta, 100 °C.

It is also known that the Sn-Beta catalyst is highly active for the isomerization–esterification of triose sugar to methyl lactate. The Sn-Beta made in this study was used to catalyze the reaction of dihydroxyacetone (DHA) in methanol at 70 °C (Fig. 2b). As expected, DHA was selectively converted to methyl lactate with a small amount of glyceraldehyde (GLA), an isomer of DHA, in the first hour. After 7 h, DHA was fully converted to methyl lactate, resembling the results reported by Taarning et al.¹⁰

The catalytic performance of the synthesized Sn-Beta was further tested for the isomerization of xylose, a pentose sugar, in water at 100 °C (Fig. 2c). The products of the reaction were xylulose, lyxose, and byproducts from degradation reactions and/or polymerization reactions. The isomerization of xylose to xylulose catalyzed by Sn-Beta is analogous to the isomerization of glucose to fructose as proposed by Moliner et al.^6c and Choudhary et al.^6b Since the HPLC column used in the analysis cannot separate xylulose and lyxose, the total yield of the two isomers was plotted against the reaction time in Fig. 2c. The isomerization reaction reached equilibrium at 0.5 h, with a maximum xylulose + lyxose yield of 35% at a xylose conversion of 61%. The xylulose + lyxose yield decreased with time as a result of side reactions consuming xylose, xylulose and lyxose. This result is also comparable to the observation reported by Choudhary et al.^6b and Lew et al.^6d

In summary, we have shown that the crystallization time of Sn-Beta can be significantly reduced to two days through a modified seeding method. The main differences between the current approach and previous ones are: in this study (1) highly crystalline zeolite crystals (200 nm) were used as the seeds, and (2) to avoid the aggregation of the seeds, a stable suspension containing the crystalline zeolite seeds was prepared, and directly added into the synthesis mixture without calcination and drying. It is believed that these two efforts enable the crystalline seeds to be uniformly distributed in the synthesis mixture. The crystallization time for Sn-Beta can, thus, be effectively shortened. Although hydrofluoric acid was still used in the synthesis, limiting its industrial application, the significantly reduced crystallization time presents an important advance for the widespread laboratory use of Sn-Beta as an active Lewis catalyst. The structural properties of the Sn-Beta made by different approaches, including the coordination environments of Sn, hydrophobicity, structural defects and their relation to the catalytic activity, are under investigation. The resulting Sn-Beta can catalyze the isomerization reactions of triose, pentose and hexose sugars with activities that are comparable to the Sn-Beta catalysts synthesized by other methods. It is reasonable to expect that this synthesis method will be applicable to other zeolites, like siliceous zeolite beta, Ti-Beta, Zr-Beta, etc. To demonstrate the versatility of this synthetic approach, siliceous zeolite beta was synthesized by this method. It was found that the crystallization time of siliceous zeolite beta could be reduced from 300 h to 50 h (Fig. S7, ESI†).

Acknowledgements

This work is financially supported as a part of the Catalysis Center for Energy Innovation, an Energy Frontier Research Center funded by the U. S. Department of Energy, Office of Science, Office of Basic Energy Science, under Award Number DE-SC0001004. We thank Prof. Gonghu Li at the University of New Hampshire for the UV-vis measurement.

References

1. (a) A. Corma, Chem. Rev., 1997, 97, 2373; (b) M. E. Davis, Nature, 2002, 417, 813.
2. (a) T. R. Carlson, T. P. Vispute and G. W. Huber, ChemSusChem, 2008, 1, 397; (b) N. M. Xavier, S. D. Lucas and A. P. Rauter, J. Mol. Catal. A: Chem., 2009, 305, 84; (c) C. L. Williams, C.-C. Chang, P. Do, N. Nikbin, S. Caratzoulas, D. G. Vlachos, R. F. Lobo, W. Fan and P. J. Dauenhauer, ACS Catal., 2012, 2, 935.
3. (a) A. Corma and H. García, Chem. Rev., 2002, 102, 3837; (b) I. W. C. E. Arends, R. A. Sheldon, M. Wallau and U. Schuchardt, Angew. Chem. Int. Ed., 1997, 36, 1144.
4. (a) M. G. Clerici, G. Bellussi and U. Romano, J. Catal., 1991, 129, 159; (b) M. G. Clerici and P. Ingallina, J. Catal., 1993, 140, 71; (c) F. Maspero and U. Romano, J. Catal., 1994, 146, 476; (d) A. Thangaraj, R. Kumar and P. Ratnasamy, J. Catal., 1991, 131, 294; (e) A. Thangaraj, S. Sivasanker and P. Ratnasamy, J. Catal., 1991, 131, 394.
5. (a) A. Corma, L. T. Nemeth, M. Renz and S. Valencia, Nature, 2001, 412, 423; (b) M. Renz, T. Blasco, A. Corma, V. Fornes, R. Jensen and L. Nemeth, Chem.–Eur. J., 2002, 8, 4708; (c) A. Corma, M. E. Domine and S. Valencia, J. Catal., 2003, 215, 294.
6. (a) M. S. Holm, S. Saravanamurugan and E. Taarning, Science, 2010, 328, 602; (b) V. Choudhary, A. B. Pinar, S. I. Sandler, D. G. Vlachos and R. F. Lobo, ACS Catal., 2011, 1, 1724; (c) M. Moliner, Y. Román-Leshkov and M. E. Davis, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 6164; (d) C. M. Lew, N. Rajabbeigi and M. Tsapatsis, Microporous Mesoporous Mater., 2012, 153, 55.
7. E. Nikolla, Y. Romaán-Leshkov, M. Moliner and M. E. Davis, ACS Catal., 2011, 1, 408.
8. (a) S. Mintova, V. Valtchev, T. Onfroy, C. Marichal, H. Knozinger and T. Bein, Microporous Mesoporous Mater., 2006, 90, 237; (b) H. Y. Chen, J. Wydra, X. Y. Zhang, P. S. Lee, Z. P. Wang, W. Fan and M. Tsapatsis, J. Am. Chem. Soc., 2011, 133, 12390.
9. A. Corma, M. E. Domine, L. Nemeth and S. Valencia, J. Am. Chem. Soc., 2002, 124, 3194.
10. E. Taarning, S. Saravanamurugan, M. S. Holm, J. M. Xiong, R. M. West and C. H. Christensen, ChemSusChem, 2009, 2, 625.

---

Footnotes

† Electronic Supplementary Information (ESI) available. See DOI: 10.1039/c2ra21381h

‡ These authors contributed equally to this work.

---