Fluorine-planted titanosilicate with enhanced catalytic activity in alkene epoxidation with hydrogen peroxide

Abstract

Fluorine species were successfully implanted into a MWW-type titanosilicate via post-treatment, which generates the SiO_3/2F units in the zeolite framework. The incorporated fluorine species significantly improved the catalytic performance in the epoxidation of alkenes as a result of enhancing the Lewis acid strength and the hydrophobicity of the zeolite.

---

Titanosilicate zeolites exhibit remarkable activity in a variety of oxidation reactions which use H₂O₂ as an oxidizing agent and have already found real commercial applications in several environmentally benign selective oxidation processes. The framework tetrahedral Ti⁴⁺ ions, which act as Lewis acid sites, may interact with H₂O₂ to form the Ti–OOH species, which are considered to be the active intermediates.¹ Many publications support that the catalytic activity of titanosilicates is attributed to the Lewis acidic sites related to the framework Ti species.^2–4 Nevertheless, it remains a big challenge to improve the oxidation catalytic activity of titanosilicates through modifying their Lewis acidity.

When fluorine, the most electronegative element, is implanted into a zeolite skeleton, the electropositivity of the Ti species is increased and consequently their Lewis acid strength would be enhanced. The fluoride ions incorporated into the zeolites as-synthesized in a fluoride media are usually connected to the frameworks mainly as penta-coordinated SiO_4/2F⁻ units although they sometimes exist in the form of tetrahedrally coordinated SiO_3/2F units.^5–7 The presence of SiO_4/2F⁻ units with a high electronegativity modifies the electron density around the neighbouring Si, which then decreases the catalytic properties of the zeolite.^6,7 In this communication, choosing a typical titanosilicate with the MWW topology (Ti-MWW) as a representative, we have intentionally implanted the fluorine species into the Ti-MWW framework with different coordination sites. The SiO_3/2F units thus introduced were shown to increase the Lewis acid strength of Ti-MWW. After selectively eliminating the negative effect of the SiO_4/2F⁻ unit, the catalytic activity of Ti-MWW was greatly enhanced in the epoxidation of alkenes with H₂O₂.

Conventional Ti-MWW with a 3-dimensional MWW structure was prepared by treating hydrothermally synthesized lamellar precursors (Si/Ti molar ratios of 100, 70, 50, 25 and 20) with a 2 M HNO₃ solution following previous procedures.⁸ The F-implanted materials, F-Ti-MWW, were prepared by adding NH₄F (Si/F = 26) into the above mentioned acid treatment system for the precursors. The acid-treated Ti-MWW and F-Ti-MWW samples were calcined in air at 823 K for 6 h to remove any residual organic species.

The XRD measurements indicate that both Ti-MWW and F-Ti-MWW are highly crystalline materials with a pure MWW phase (see Fig. S1A, ESI†). The specific surface area of the samples was in the range of 520–610 m² g⁻¹. The UV-visible diffuse reflection spectra verified that they were dominated by the tetrahedrally coordinated Ti ions in the framework, showing a principal adsorption at 210 nm (see Fig. S1B, ESI†). Meanwhile, the adsorption at 330 nm, attributed to the anatase phase, was extremely weak. These physicochemical properties and the Ti active site characterization indicated that F-Ti-MWW qualified as a catalyst for liquid-phase oxidation reactions like Ti-MWW. The XPS spectrum of calcined F-Ti-MWW detected a peak at 688 eV, assigned to F 1s (see Fig. S2, ESI†), clearly indicating the F species remained even after calcinations at 823 K. The surface Si/F molar ratio given by the XPS data was 33, which is close to the amount of F added in the acid treatment.

A direct examination of the fluorine environments in F-doped zeolites can be realized using ¹⁹F NMR spectroscopy. A typical ¹⁹F MAS NMR spectrum of F-Ti-MWW is given in Fig. 1a. It shows that two kinds of F species exist in F-Ti-MWW as Si–F and B–F interaction forms. The formation of the B–F species (−160.8 ppm) is due to the fact that the Ti-MWW precursors were synthesised using boric acid as a crystallization-supporting agent.⁸ The resonances at −129.1, −140.6 and −152.8 are attributable to the SiF₆²⁻, SiO_2/4F⁻ and SiO_2/3F species, respectively.^5,9,10 Consistent with the XPS measurement, the ¹⁹F MAS NMR spectrum identified that the F species were incorporated into Ti-MWW by the NH₄F modification and that the F⁻ ion which was favorably introduced consisted of the SiO_2/4F⁻ and SiO_2/3F units located in the framework.

Fig. 1 ¹⁹F MAS NMR spectra of F-Ti-MWW (a) and F-Ti-MWW-K (b). The asterisks indicate the spin side bands.

The IR spectrum of Ti-MWW showed absorptions at 3745 cm⁻¹, 3720 cm⁻¹ and 3510 cm⁻¹ in the hydroxyl stretching region (see Fig. S3, ESI†), which are attributed to unperturbed external silanols, internal silanols and hydrogen-bonded silanol nests, respectively.¹¹ After the F modification, the OH stretching vibrations, particularly the 3720 cm⁻¹ and 3510 cm⁻¹ bands, clearly decreased in intensity in comparison to the unmodified Ti-MWW sample. This implies that the F incorporation was probably realized through the interaction with Si–OH groups. This then improved the surface hydrophobicity of the zeolite, which is in agreement with the ²⁹Si NMR investigation. F-Ti-MWW showed very similar resonances due to the Q⁴ sites but a less intense resonance of the Q³ site than Ti-MWW (see Fig. S4, ESI†).

Fig. 2 shows the FTIR spectra in the pyridine vibration region after pyridine adsorption and desorption at different temperatures. The bands at 1580 and 1446 cm⁻¹ are attributed to the pyridine species interacting with the Lewis acid sites, while the 1540 cm⁻¹ band is assigned to the pyridium ions.¹² The pyridine adsorption confirmed that Ti-MWW and F-Ti-MWW possessed both Brønsted acid sites and Lewis acid sites but the Brønsted acidity is negligible as the 1540 cm⁻¹ band was extremely weak in comparison to the others. The bands related to Lewis acidity decreased sharply upon evacuation at elevated temperatures. In the case of Ti-MWW, only a weak band at 1446 cm⁻¹ was observed after desorption at 523 K, whereas it was more distinct for F-Ti-MWW (Fig. 2e). This implies that F-Ti-MWW possesses stronger Lewis acid sites than Ti-MWW. The stronger Lewis acidity in F-Ti-MWW is probably due to the substitution of framework Si–OH by Si–F during the NH₄F modification process. The fluorine species that had a high electronegativity and existed as SiO_3/2F units would withdraw the electrons of neighbouring Ti ions. As shown in Scheme 1, the Ti ions should become more positively charged, serving as stronger Lewis acid sites in F-Ti-MWW. In actual catalytic oxidation reactions, the electrophilic interaction of these Ti sites with H₂O₂ molecules are presumed to be intensified, giving rise to more active Ti-peroxo species.

Fig. 2 FTIR spectra in the pyridine region of Ti-MWW (A) and F-Ti-MWW (B) before (a) and after pyridine adsorption and desorption at 323 K (b), 373 K (c), 423 K (d) and 523 K (e), respectively.

Scheme 1 Electropositive function of F in F-Ti-MWW.

After choosing 1-hexene epoxidation of as a probe reaction, we investigated the catalytic performance of Ti-MWW and F-Ti-MWW with various Ti contents. From the dependence of the initial epoxidation rate on the Ti content and reaction time courses (Fig. 3 and Fig. S5, ESI†), it is obvious that F-Ti-MWW shows an outstanding catalytic activity in comparison to conventional Ti-MWW. Table 1 shows a comparison between these two types of catalysts that results in the epoxidation of alkenes which differ in molecular structure and functional group. The catalysts were prepared from the same lamellar precursor synthesized at Si/Ti = 25. Again, the F modification enhanced the oxidation activity and peroxide selectivity significantly (Table 1, Nos. 1–4).

Fig. 3 Dependence of the initial epoxidation rate on the Ti content in Ti-MWW (a), F-Ti-MWW (b) and F-Ti-MWW-K (c). Reaction conditions: cat., 0.05 g; CH₃CN, 10 mL; 1-hexene, 10 mmol, H₂O₂, 10 mmol; temp., 333 K.

Table 1 Catalytic properties of Ti-MWW and F-Ti-MWW in the epoxidation of alkenes with H₂O₂

No.	Substrate	Ti-MWW (%)			F-Ti-MWW (%)
		Xalkene	XH2O2	Oxide sel.	Xalkene	XH2O2	Oxide sel.
Reaction conditions: time, 2 h; others, see Fig. 3.a Catalysed by Re-MWW prepared via piperidine treatment.b Catalysed by F-Ti-MWW-K prepared via KCl treatment.
1	1-Hexene	41.6	49.4	98.6	58.5	67.4	98.9
2	Allyl chloride	21.5	24.5	99.6	28.7	31.2	99.8
3	1-Heptene	26.0	46.7	98.1	34.0	46.6	97.2
4	Cyclopentene	15.9	17.1	67.1	17.4	19.2	81.7
5	1-Hexene^a	47.2	53.2	98.7	—	—	—
6	1-Hexene^b	42.2	55.1	98.6	86.7	94.7	98.8

---

As shown above, when fluorine was implanted into the Ti-MWW framework as SiO_3/2F units, the Lewis acid strength was enhanced and simultaneously the surface hydrophilicity was decreased. In order to clarify which aspect contributes the most to the improved catalytic activity of F-Ti-MWW, we first investigated the hydrophobic/hydrophilic issue. A structural rearrangement with piperidine treatment has been proven to remove the hydrophilic silanols on defect sites, possibly by 35%, and then to effectively enhance the oxidation activity of Ti-MWW.¹³ Here, Ti-MWW was similarly treated with piperidine at 443 K, leading to a defectless sample with a higher hydrophobicity, denoted as Re-Ti-MWW. The conversion of 1-hexene then increased from 41.6% to 47.2% (Table 1, Nos. 1 and 5). However, it was far below the activity level of F-Ti-MWW (58.5%). Presumably, the increased Lewis acid strength is a dominant factor which govens the outstanding activity of F-Ti-MWW.

With respect to the epoxidation of different types of alkenes, both Ti-MWW and F-Ti-MWW showed lower conversions for allyl chloride and cyclopentene in comparison to other alkenes. Allyl chloride contains a chlorine atom, a strong electronegative element, which lowers the electron density in the C=C double bond. Thus, it is intrinsically less reactive. As a cycloalkene with a relatively large molecular size, cyclopentene may suffer from steric limitation imposed by the distorted 10-membered ring opening of the MWW zeolites. It is reasonable therefore, that these two alkenes showed much lower conversions. In the case of 1-hexene and 1-heptene, they are both of a linear shape and have almost the same molecular cross section but an increase in carbon number would make a great difference in the case of catalysis within microporous materials. The longer the alkyl chain, the lower the diffusion rate inside the zeolite channels and therefore the lower the conversion.

Ishihara et al. once reported that F⁻ is removable via anion-exchange with aqueous sodium hydroxide.¹⁴ The K ions may create an equilibrium with the F ions on solid materials. To eliminate the negative effect of the SiO_4/2F⁻ units while keeping the SiO_2/3F units intact, we have attempted the selective removal of the F species by employing a more moderate KCl treatment instead of KOH because a severe alkali treatment unquestionably poisons the active sites of titanosilicates.^15,16 The F-Ti-MWW-K catalyst obtained via KCl treatment showed a further improved epoxidation activity. The conversion of 1-hexene increased from 58.5% to 86.7% with a well preserved selectivity for the epoxide product (Table 1, No. 6). The initial epoxidation rate of F-Ti-MWW became much higher, especially in high Ti content regions (Fig. 3c). The K treatment also increased the conversion of F-Ti-MWW in the epoxidation of other alkenes (see Table S1, ESI†). In a control experiment, the same treatment made almost no difference to F-free Ti-MWW (Table 1, No. 6). Based on these facts, it is deduced that there must be an interaction between F-Ti-MWW and KCl, which led to enhanced active sites in F-Ti-MWW-K. In fact, the ¹⁹F NMR investigation clearly evidenced that the SiO_4/2F⁻ units (−140.6 ppm) were remarkably reduced by the KCl treatment but the SiO_3/2F units (−152.8 ppm) were almost completely retained (Fig. 1b). Once the negative effect concerning the SiO_4/2F⁻ units was eliminated, the benefits of the remaining SiO_3/2F units became outstanding in the oxidation reactions. This result further confirmed the aforementioned speculation from another perspective.

In summary, we have successfully substituted the Si–OH groups in Ti-MWW for Si–F through a postsynthesis modification with NH₄F. The incorporation of F species significantly improved its catalytic activity in the epoxidation of various alkenes. This study may provide a versatile way of functionalizing other titanosilicates.

The authors gratefully acknowledge the financial support from the National Science Foundation of China (20973064, 20925310, U1162102), Ministry of Science and Technology (2012BAE05B02), Shanghai Municipal Education Commission (11CXY20) and Shanghai Leading Academic Discipline Project (B409).

Notes and references

1. S. Bordiga, F. Bonino, A. Damin and C. Lamberti, Phys. Chem. Chem. Phys., 2007, 9, 4854.
2. A. Corma, U. Diaz, M. E. Domine and V. Fornś, J. Am. Chem. Soc., 2000, 122, 2804.
3. A. Corma and H. Garcia, Chem. Rev., 2003, 103, 4307.
4. G. Yang, J. Q. Zhuang, D. Ma, X. J. Lan, L. J. Zhou, X. C. Liu, X. H. Han and X. H. Bao, J. Mol. Struct., 2008, 882, 24.
5. R. E. Youngman and S. Sen, J. Non-Cryst. Solids, 2004, 349, 10.
6. Y. Goa, P. Wu and T. Tatsumi, J. Phys. Chem. B, 2004, 108, 4242.
7. L. Benoît and K.-M. Lioubov, Microporous Mesoporous Mater., 2004, 74, 171.
8. P. Wu, T. Tatsumi, T. Komatsu and T. Yashima, J. Phys. Chem. B, 2001, 105, 2897.
9. I. Ogino, M. M. Nigra, S.-J. Hwang, J.-M. Ha, T. Rea, S. I. Zones and A. Katz, J. Am. Chem. Soc., 2011, 133, 3288.
10. M. Melaimi and F. P. Gabbaï, J. Am. Chem. Soc., 2005, 127, 9680.
11. G. P. Heitmann, G. Dahlhoff and W. F. Hölderich, J. Catal., 1999, 186, 12.
12. G. A. H. Mekhemer, A. K. H. Nohman, N. E. Fouad and H. A. Khalaf, Colloids Surf., A, 2000, 161, 439.
13. L. L. Wang, Y. M. Liu, W. Xie, H. H. Wu, X. H. Li, M. Y. He and P. Wu, J. Phys. Chem. C, 2008, 112, 6132.
14. T. Ishihara, Y. Shuto, S. Ueshima, H. L. Ngee and H. Nishiguchi, J. Ceram. Soc. Jpn., 2002, 110, 801.
15. C. B. Khouw and M. E. Davis, J. Catal., 1995, 151, 77.
16. T. Tatsumi, K. A. Koyano and Y. Shimizu, Appl. Catal., A, 2000, 200, 125.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis, characterizations and catalysis. DOI: 10.1039/c2cy20446k

---