Synthesis of boron doped polymeric carbon nitride solids and their use as metal-free catalysts for aliphatic C–H bond oxidation

Abstract

Oxidation of hydrocarbons with clean oxidants is a crucial process for the development of products ranging from commodity chemicals to speciality pharmaceuticals. While selective oxidation is commonplace in nature and mainly occurs through a broad variety of cytochrome P450 enzymes, the synthetic catalysts that have been identified so far are relatively inactive with aliphatic C–H bonds. We present here a promising strategy for modifying polymeric graphitic carbon nitride by replacing the carbon atoms in the “melon” network using boron. ¹¹B solid-state MAS NMR and XPS spectral studies confirmed the incorporation of the boron atoms in the polymeric g-C₃N₄ as well as the degree of doping, while X-ray diffraction and IR indicated that the characteristic structural properties of polymeric g-C₃N₄, such as the original graphitic structure, and the C–N heterocycle units were essentially retained. The boron-doped polymeric carbon nitride shows promising catalytic activity and high selectivity (≥87%) in the oxidation of benzylic aromatics. The use of this metal-free, organic, semiconducting polymer as an oxidation catalyst and H₂O₂ or O₂ as oxidants makes the reactions interesting from both an economic and environmental point of view. It is also hoped that our findings will stimulate further investigation and will open new possibilities to form hybrid carbon nitride materials with new and probably exciting applications.

---

Introduction

Oxidation, especially oxidation of aliphatic (sp³) C–H bonds is a core technology for converting petrochemical based materials to commodity chemicals in the chemical industry.^1–5 However, selective catalytic oxidation of sp³ hybridized carbons, within hydrocarbons, to alcohols, aldehydes, or ketones remains a challenging topic in contemporary chemical research.^6–9 The increased reactivity of the products relative to the starting material makes selective oxidation very difficult. The catalysts available today for heterogeneous oxidation are generally far from ideal. Industrially performed catalytic oxidation reactions often suffer from drawbacks such as poor selectivity due to over-oxidation, lack of catalyst recycling, and negative environmental impact due to the use of toxic metal catalysts, etc.^10–13 For example, in industry oxidation of toluene is achieved by using heavy metal based catalysts at high temperature (>200 °C). Generally, the conversion has to be kept at less than 4% to attain 70% selectivity of benzaldehyde and to avoid the formation of the carboxylic acid.^14,15 The oxidation catalysts of the future have to be highly selective, environmentally benign and highly stable. Investigations into exquisitely selective biological catalysts, such as the oxidase and oxygenase enzymes, and some biomimetic oxidative catalysts are guiding us towards the ideal oxidation catalyst.^16–21 Due to the efficiency of enzymes, the development of molecular compounds that can mimic the reactivity of enzymes have been extensively studied in recent years.^22–29 However, most of the biomimetic catalysts that have been identified so far are relatively inactive toward aliphatic C–H bonds, and catalyst preparation often remains critical. Moreover, reducing agents are usually needed to compensate the excessive oxidation.^30,31 Therefore, alternative metal-free heterogeneous catalysts for the selective oxidation of aliphatic C–H bond are highly desirable.

In our search for biomimetic systems, we focus on heterogeneous organocatalysis, essentially adjusting the HOMO and LUMO positions and the coupled redox properties along with the natural models. Recently, a nitrogen-rich graphitic carbon nitride (g-C₃N₄) material has attracted much attention.^32–37 This material has a heterocyclic macrocycle structure with a N–C–N– bonding pattern resembling a porphyrin (Scheme S1 ESI†). It was found that this polymeric material exhibits high chemical and thermal stability and an appealing electronic structure, being a wide-band gap, indirect semiconductor.³⁸ This allows its direct use in heterogeneous catalysis.^8,39–41 In addition, the catalytic, electronic and optical properties of the structure are by principle adjustable, using the rich chemistry of carbon and nitrogen.^42–45 Indeed, similar “doping” protocols have been widely used to control the performance of more traditional materials. For example, various lightweight materials with diverse nanomorphologies that contain heteroatoms such as nitrogen, boron, or fluorine, have been actively pursued because of their unusual properties, say as semiconductors or in catalytic applications.^46,47 Among the different non-metal dopants, boron seems to be especially promising in the materials field.^48,49 A promising strategy for modifying g-C₃N₄ exists by replacing the C atoms in the network by boron to form a π-bonded, planar layered configuration like melon (Scheme 1). The active site, the heterocyclic macrocycle structure, is still there and the boron functional groups on the surface might act as strong Lewis acid sites, thus complementing the basic nitrogen sites in a bifunctional catalysis.

Scheme 1 Proposed copolymerization processes of dicyandiamide with BH₃NH₃ at high temperature.

We herein demonstrated the synthesis of boron-doped polymeric carbon nitride (CNB) solids using BH₃NH₃ as a molecular doping source which avoids the use of oxygen-containing precursors or the generation of HF, as with BF₄⁻.^8,43 We also demonstrate the catalytic activity of these CNB materials in the heterogeneous, metal-free oxidation of aliphatic C–H bonds. The CNB materials show superior performance (as compared to the previous homogeneous biomimetic oxidation catalysts) in the oxidation of toluene and other substituted aromatics with good conversion and an excellent aldehyde or ketone selectivity.

Experimental

Materials and methods

Unless otherwise stated, all solvents and chemicals used were of commercially available analytical grade. WAXS spectra were recorded on a Bruker D8 Advance diffractometer. UV-Vis spectra were recorded on a Varian Cary 4E UV-Vis system equipped with a Labsphere diffuse reflectance accessory. FTIR spectra were recorded on a BioRad FTS 6000 spectrometer, equipped with an attenuated total reflection (ATR) setup. ¹¹B solid-state MAS NMR was recorded on a Bruker Avance III 400M NMR. X-ray photoelectron spectroscopy (XPS) data were obtained on Thermo ESCALAB250 instrument with a monochromatized Al-Kα line source (200 W). All binding energies were referenced to the C^1s peak at 288.2 eV³⁹ corresponding to a C–N–C coordination in CNB.

Synthetic procedures for CNB_x materials

The synthesis of CNB samples has been carried out employing a directly incorporating of different amount of BH₃NH₃ (x = 0.01, 0.025, 0.05, 0.1, 0.15, 0.20, 0.25 g, where x is the exact weight of BH₃NH₃ used per 3 g dicyandiamide) into the classical carbon nitride condensation (the resulting samples are denoted CNB_x). In a typical synthetic procedure, BH₃NH₃ (0.15 g) was dissolved in 10 ml water and stirred for 5 min. Then 3 g dicyandiamide was added into the aqueous solution and the mixture was heated in an oil bath at 80 °C until the water was removed and a white solid formed. The white solid was then transferred into a crucible and heated for 4 h to reach a temperature of 600 °C and tempered at this temperature for 4 h under high purity nitrogen with a flow rate of 500 ml min⁻¹. The sample was then allowed to cool in the oven to room temperature.

Typical procedure for the catalytic oxidation of aromatics

The oxidation of toluene or ethylbenzene over a CNB_x sample using H₂O₂ as the oxidant was carried out in a teflon-lined stainless steel batch reactor (40 ml total volume), using acetonitrile as the reaction solvent. Toluene or ethylbenzene (0.8 ml), catalyst (50 mg), 30% hydrogen peroxide in water solution and acetonitrile (8 ml) were introduced into the reactor. The reactor heated to 150 °C within 30 min with stirring, and the reactor was held at this temperature for 2 h. Oxidations of substituted aromatics over CNB_0.15 using O₂ as oxidant were also carried out in a Teflon-lined stainless steel batch reactor. 1 mmol Substrate, 4 ml acetonitrile, 50 mg CNB_0.15 were loaded into the reactor. The reactor was sealed, purged with O₂ to remove the air for 3 times, increased the O₂ pressure to 1 MPa, and then the reactor was heated to the desired temperature. After reaction the reactor was placed in ice water to quench the reaction and the products were analyzed with a GC and GC-MS (Agilent Technologies, GC6890N, MS5970) with benzyl chloride as the internal standard.

Reusing of the catalysts

For the recycling experiments, the reaction mixture was centrifuged for 10 min after the first reaction and the liquid layer was siphoned out. The residual solid was washed with anhydrous ethanol and centrifuged twice and finally dried in a vacuum oven at 80 °C for 6 h. Aromatics, H₂O₂ and acetonitrile were then charged for subsequent round of reaction under otherwise identical conditions.

GC and GC-MS analysis

All GC experiments were carried out and recorded using a HP 6890 Series. The GC conditions for the analysis of the aromatic oxidation products were as follows. Injection port temperature 280 °C, column temperature: initial temperature: 50 °C (1 min); Gradient rate: 50 °C (4 min); Final temperature 250 °C (5 min); Flow rate: 80 ml min⁻¹. All GC-MS experiments were carried out and recorded using an Agilent GC6890N, MS5970 series. The GC-MS conditions for the products analysis was kept the same as that used on GC. Selectivities and conversions were calculated from the equations: Selectivity = [aldehyde or ketone]/[Consumed aromatics] ×100; Conversion = [Consumed aromatics]/[Initial aromatics] × 100, respectively.

Results and discussion

The structures of these hybrid materials were first investigated by powder X-ray diffraction (XRD). The XRD pattern (Fig. 1a) supports the preservation of the graphite-like packing of practically all products, showing the typical (002) interlayer-stacking peak around 27.4°, as is well known for the melon network.^39,50 The FTIR spectra (Fig. S1 ESI†) of the CNB samples showed typical C–N heterocycle stretches in the region of ca. 1100–1600 cm⁻¹ and the breathing mode of the tri-s-triazine units at 800 cm⁻¹ evidencing that the original graphitic C–N network remained. The typical vibration of B–N at 1370 cm⁻¹ is presumably overlapped by that of the C–N stretch.⁵¹

Fig. 1 (a) XRD patterns of the B doped and pristine g-C₃N₄; (b) ¹¹B solid-state MAS NMR spectra of CNB_0.15, inset: the building unit of C₃N₄ with B doping at the bay-carbon site (right) and at the corner-carbon site (left), the dotted lines imply that the tricyclic system is a repeat unit in a regular lattice framework; (c) XPS spectra of CNB_0.15.

Structural details on the incorporation of boron into the C/N matrix were obtained with ¹¹B solid-state MAS NMR experiment. Fig. 1b suggested two peaks, one strong peak at 4.99 ppm and one small peak at 10.13 ppm, which corresponded to two different positions in the framework structure, say as the bay-carbon site and the corner carbon site (Fig. 1b, inset). Further evidence of the boron doping came from the X-ray photoelectron spectroscopy (XPS) measurements. The XPS result gives direct evidence of a boron heteroatom in the doped g-C₃N₄ as well as the degree of doping (Fig. 1c).The B_1s binding energy (BE) peaks of CNB_0.15 is centered at ca. 191.6 eV, which is at a higher binding energy than that for h-BN (190.0 eV) and at a lower binding energy than that for Kawaguchi's BCN(H) (192.1 eV, boron surrounded by three nitrogen atoms), which indicates that the boron atoms in our CNB network are more electropositive than that in h-BN, but are not as electropositive as that in Kawaguchi's BCN(H).⁵² The boron content (CNB_0.15) as determined by XPS is near 9 mol%, i.e. a tentative sum formula of the structure is C₃N_4.7B_0.8H_1.9.

UV-Vis spectra (Fig. S2 ESI†) indicate that the optical band gap and thereby the semiconductor properties of the CNB have indeed changed, with an extension of the spectrum towards absorption in the visible range. This can in principle enhance the ability to harvest visible light, beneficial for heterogeneous photocatalysis (Table S1 ESI†), which is however not the subject of the present work. It is also to be said that the incorporation of boron is expected to lower the HOMO position of the resulting material, thus increasing the potential oxidation strength as compared to bare carbon nitrides.

Catalytic testing was first performed for the oxidation of toluene using CNB with varying boron content as metal-free catalysts. The corresponding conversions are summarized in Table 1. In all the cases, 2 to 6.3% conversions were found, while the blank run without catalyst gave almost no product; that is, CNB (including the non doped g-C₃N₄) indeed acts as an oxidation catalyst. The doping of boron atoms in the “melon” texture undoubtedly increased the activity. The g-C₃N₄ itself gave minor conversion of 2%, while all the boron doped materials appeared to be more active as illustrated in Table 1. Among all the catalysts tested, CNB_0.15 gave the best conversion of up to 6.3%. The catalysts are highly stable and can be reused for several cycles without loss of activity (Table 1, entry 5, 6, 7), which is a prerequisite for practical applications. Also this fact however proves that the amount of reaction is presumably limited by the availability of an oxidation agent, only. For practical applications, this would be ensured by a constant influx of H₂O₂, which however cannot be performed in our autoclave setup. Increasing the ratio of n_toluene/n_H2O2 to 1 : 5, yielded a 12.4% conversion of toluene (Table 1, entry 9). GC-MS analyses of the post-reaction mixture showed that the CNB catalysts (beside CNB_0.25) afford remarkably high selectivities towards the formation of benzaldehyde, i.e. almost all the oxidation reaction gave a selectivity above 99%. It should be noted, that the activity of CNB did not simply increase with the amount of boron doped, as exceeding 9 mol% appears to be detrimental (Table S2 ESI†).

Table 1 Conversion and selectivity of toluene oxidation over CNB_x catalysts^a

Entry	Catalyst	Conv. [%]	Sel. [%] BA^b
a Reaction conditions: toluene 0.8 ml, H2O2 (30% in water solution) 0.77 ml, catalyst 50 mg, acetonitrile 8 ml, reaction temperature 150 °C, reaction time 2 h. b BA = benzaldehyde. c Second run to test the reusability of catalyst. d Third run to test the reusability of the catalyst. e Forth run to test the reusability of the catalyst. f Mole ratio nH2O2/ntoluene = 3. g Mole ratio nH2O2/ntoluene = 5. h Mole ratio nH2O2/ntoluene = 7.
1	g-C3N4	2.0	>99.0
2	CNB0.025	4.2	>99.0
3	CNB0.1	5.9	>99.0
4	CNB0.15	6.3	>99.0
5^c	CNB0.15	6.2	>99.0
6^d	CNB0.15	6.1	>99.0
7^e	CNB0.15	6.3	>99.0
8^f	CNB0.15	9.8	>99.0
9^g	CNB0.15	12.4	>99.0
10^h	CNB0.15	11.5	>99.0

---

The C–H bond energy of the aliphatic toluene bond is 90 kcal mol⁻¹, while the corresponding C–H bond of ethylbenzene has 87 kcal mol⁻¹.⁵³ This should enable similar organocatalytic oxidation chemistry of ethylbenzene towards acetophenone. In fact, using H₂O₂ as the oxidant and keeping T = 150 °C for 2 h, GC-MS analysis indeed disclosed the formation of acetophenone in ∼4.6% and 10.6% yield, respectively, based on the CNB_x species (see Table 2), with the same remarkable selectivities exceeding 99%.

Table 2 Conversion and selectivity of ethylbenzene oxidation over g-C₃N₄ and CNB_x catalysts^a

Entry	Catalyst	Conv. [%]	Sel. [%] AP^b
a Reaction conditions: ethylbenzene 0.8 ml, H2O2 (30% in water solution) 0.66 ml, catalyst 50 mg, acetonitrile 8 ml, reaction temperature 150 °C, reaction time 2 h. b AP = acetophenone.
1	g-C3N4	4.6	>99.0
2	CNB0.025	5.9	>99.0
3	CNB0.1	7.4	>99.0
4	CNB0.15	10.6	>99.0

---

With the preliminary success in the oxidation of toluene and ethylbenzene using H₂O₂ as oxidant, we also addressed the oxidation of substituted benzylic aromatics using molecular oxygen (O₂) an oxidant which is much harder to activate and tame. Since CNB_0.15 was found to be the best catalyst for the oxidation of toluene and ethylbenzene, it was used in all further experiments, the results of which are summarized in Table 3.

Table 3 Molecular oxygen oxidation of substituted aromatics to ketone using CNB_0.15 as metal free catalyst^a

Entry	Substrate	Main product	BDE ^b	T (°C)	Conv. (%)	Sel. (%)
a Reaction conditions: substrates 1 mmol, O2 pressure 1 MPa, catalyst 50 mg, acetonitrile 4 ml, reaction time 24 h. b C–H bond dissociation energy from the literature.^54 c Using g-C3N4 as catalyst.
1			81	160	20.3	95.1
2			80	130	45.7	>99.0
3^c			80	130	28.8	>99.0
4			83	130	72.2	87.5
5			82	115	36.7	>99.0

---

Although no products had been detected in the case of toluene and ethylbenzene using O₂ as the oxidant, oxidation of other substrates with an activated benzylic position went smoothly. Using oxidation of fluorene (C–H binding energy:⁵⁴ 80 kcal mol⁻¹) to fluorenone as an example, 45.7% conversion and 99% selectivity to fluorenone were achieved in 24 h at 130 °C using 1 MPa O₂ pressure. As a control experiment, by simply heating the same reaction mixture without the adding of catalyst, almost no fluorenone could be detected by GC-MS.

Conclusions

To conclude, we have presented a new synthesis of boron doped carbon nitrides, using BH₃NH₃ as the boron source. The catalytic properties of these materials were tested in the oxidation of activated aliphatic C–H bonds. These materials exhibit good catalytic performance for the selective oxidation of toluene and substituted benzylic aromatics. Compared to traditional metal catalysts, the metal-free, boron doped, organic semiconducting polymers afford a remarkably high selectivity towards the formation of aldehydes or ketones. In addition they are highly stable and could be recycled several times without losing their reactivity. The range of substrates depends on the doping of the catalyst as well as on the oxidation agent applied. Using H₂O₂, the best catalyst of this series is already able to oxidize methylaromatic systems, while with molecular oxygen, the reaction is restricted to more activated methylene groups. The catalysts are therefore rather selective, with respect to both the substrate and the product they create. This gives hope that further modifications of similar catalytically active polymers and solid state materials will allow access to other more demanding substrates such as alkanes or even methane (C–H bond energy:⁵⁴ 105 kcal mol⁻¹) while keeping the selectivity. It is hoped that our findings will stimulate such further investigations, both on the materials as well as on the applications.

Acknowledgements

This work was supported by ENERCHEM project of MPI and the National Science Foundation of China (No. 20806065 and No. 20990221). We thank Dr Yan Di for testing the hydrogen evolution reactions.

Notes and references

1. M. Christmann, Angew. Chem., Int. Ed., 2008, 47, 2740–2742.
2. G. Dyker, Angew. Chem., Int. Ed., 1999, 38, 1698–1712.
3. C. G. Jia, T. Kitamura and Y. Fujiwara, Acc. Chem. Res., 2001, 34, 633–639.
4. A. E. Shilov and G. B. Shul'pin, Chem. Rev., 1997, 97, 2879–2932.
5. J. Zhang, X. Liu, R. Blume, A. H. Zhang, R. Schlogl and D. S. Su, Science, 2008, 322, 73–77.
6. K. Kamata, K. Yonehara, Y. Nakagawa, K. Uehara and N. Mizuno, Nat. Chem., 2010, 2, 478–483.
7. M. S. Chen and M. C. White, Science, 2010, 327, 566–571.
8. Y. Wang, J. S. Zhang, X. C. Wang, M. Antonietti and H. R. Li, Angew. Chem., Int. Ed., 2010, 49, 3356–3359.
9. J. A. Labinger and J. E. Bercaw, Nature, 2002, 417, 507–514.
10. T. Punniyamurthy, S. Velusamy and J. Iqbal, Chem. Rev., 2005, 105, 2329–2363.
11. F. Rajabi, J. H. Clark, B. Karimi and D. J. Macquarrie, Org. Biomol. Chem., 2005, 3, 725–726.
12. S. M. Yiu, Z. B. Wu, C. K. Mak and T. C. Lau, J. Am. Chem. Soc., 2004, 126, 14921–14929.
13. D. R. Dreyer, H.-P. Jia and C. W. Bielawski, Angew. Chem., Int. Ed., 2010, 49, 6686–6816.
14. F. Wang, J. Xu, X. Q. Li, J. Gao, L. P. Zhou and R. Ohnishi, Adv. Synth. Catal., 2005, 347, 1987–1992.
15. F. Konietzni, U. Kolb, U. Dingerdissen and W. F. Maier, J. Catal., 1998, 176, 527–535.
16. S. Das, C. D. Incarvito, R. H. Crabtree and G. W. Brudvig, Science, 2006, 312, 1941–1943.
17. L. Que and W. B. Tolman, Nature, 2008, 455, 333–340.
18. M. Yamashita, H. Furutachi, T. Tosha, S. Fujinami, W. Saito, Y. Maeda, K. Takahashi, K. Tanaka, T. Kitagawa and M. Suzuki, J. Am. Chem. Soc., 2007, 129, 2–3.
19. B. J. Wallar and J. D. Lipscomb, Chem. Rev., 1996, 96, 2625–2657.
20. J. Valton, M. Fontecave, T. Douki, S. G. Kendrew and V. Niviere, J. Biol. Chem., 2005, 281, 27–35.
21. I. Siewert, C. Limberg, S. Demeshko and E. Hoppe, Chem.–Eur. J., 2008, 14, 9377–9388.
22. C. Limberg, Angew. Chem., Int. Ed., 2003, 42, 5932–5954.
23. M. Fontecave, S. Menage and C. Duboc-Toia, Coord. Chem. Rev., 1998, 178–180, 1555–1572.
24. M. W. Grinstaff, M. G. Hill, J. A. Labinger and H. B. Gray, Science, 1994, 264, 1311–1313.
25. F. Bedioui, Coord. Chem. Rev., 1995, 144, 39–68.
26. M. H. So, Y. G. Liu, C. M. Ho and C. M. Che, Chem.–Asian J., 2009, 4, 1551–1561.
27. D. Mansuy, Coord. Chem. Rev., 1993, 125, 129–141.
28. M. Costas, M. P. Mehn, M. P. Jensen and L. Que, Chem. Rev., 2004, 104, 939–986.
29. R. F. Parton, I. F. J. Vankelecom, M. J. A. Casselman, C. P. Bezoukhanova, J. B. Uytterhoeven and P. A. Jacobs, Nature, 1994, 370, 541–544.
30. I. Siewert and C. Limberg, Chem.–Eur. J., 2009, 15, 10316–10328.
31. N. Kitajima, M. Ito, H. Fukui and Y. Morooka, J. Chem. Soc., Chem. Commun., 1991, 102–104.
32. Z. H. Zhang, K. Leinenweber, M. Bauer, L. A. J. Garvie, P. F. McMillan and G. H. Wolf, J. Am. Chem. Soc., 2001, 123, 7788–7796.
33. Y. Wang, X. C. Wang, M. Antonietti and Y. J. Zhang, ChemSusChem, 2010, 3, 435–439.
34. T. Komatsu, J. Mater. Chem., 2001, 11, 802–805.
35. J. Gracia and P. Kroll, J. Mater. Chem., 2009, 19, 3013–3019.
36. J. R. Holst and E. G. Gillan, J. Am. Chem. Soc., 2008, 130, 7373–7379.
37. J. L. Zimmerman, R. Williams, V. N. Khabashesku and J. L. Margrave, Nano Lett., 2001, 1, 731–734.
38. M. Antonietti and P. Fratzl, Macromol. Chem. Phys., 2010, 211, 166–170.
39. A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J. O. Muller, R. Schlogl and J. M. Carlsson, J. Mater. Chem., 2008, 18, 4893–4908.
40. X. Jin, V. V. Balasubramanian, S. T. Selvan, D. P. Sawant, M. A. Chari, G. Q. Lu and A. Vinu, Angew. Chem., Int. Ed., 2009, 48, 7884–7887.
41. J. Zhu, Y. Wei, W. Chen, Z. Zhao and A. Thomas, Chem. Commun., 2010, 46, 6965–6967.
42. Y. Wang, Y. Di, M. Antonietti, H. Li, X. Chen and X. Wang, Chem. Mater., 2010, 22, 5119–5121.
43. S. C. Yan, Z. S. Li and Z. G. Zou, Langmuir, 2010, 26, 3894–3901.
44. G. Liu, P. Niu, C. Sun, S. C. Smith, Z. Chen, G. Q. Lu and H.-M. Cheng, J. Am. Chem. Soc., 2010, 132, 11642–11648.
45. Y. J. Zhang, T. Mori, J. H. Ye and M. Antonietti, J. Am. Chem. Soc., 2010, 132, 6294–6295.
46. Z. J. Li, G. D. Del Cul, W. F. Yan, C. D. Liang and S. Dai, J. Am. Chem. Soc., 2004, 126, 12782–12783.
47. S. In, A. Orlov, R. Berg, F. Garcia, S. Pedrosa-Jimenez, M. S. Tikhov, D. S. Wright and R. M. Lambert, J. Am. Chem. Soc., 2007, 129, 13790–13791.
48. O. Stephan, P. M. Ajayan, C. Colliex, P. Redlich, J. M. Lambert, P. Bernier and P. Lefin, Science, 1994, 266, 1683–1685.
49. W. L. Wang, X. D. Bai, K. H. Liu, Z. Xu, D. Golberg, Y. Bando and E. G. Wang, J. Am. Chem. Soc., 2006, 128, 6530–6531.
50. T. Komatsu and T. Nakamura, J. Mater. Chem., 2001, 11, 474–478.
51. A. Y. Liu, R. M. Wentzcovitch and M. L. Cohen, Phys. Rev. B: Condens. Matter, 1989, 39, 1760–1765.
52. M. Kawaguchi, T. Kawashima and T. Nakajima, Chem. Mater., 1996, 8, 1197–1201.
53. H. R. Lucas, L. Li, A. A. N. Sarjeant, M. A. Vance, E. I. Solomon and K. D. Karlin, J. Am. Chem. Soc., 2009, 131, 3230–3245.
54. L. J. J. Laarhoven and P. Mulder, J. Phys. Chem. B, 1997, 101, 73–77.

---

Footnote

† Electronic supplementary information (ESI) available: Additional analytical and experimental data. See DOI: 10.1039/c0sc00475h

---