Wenjing
Xiao†
,
Zhenyu
Sun†
,
Sha
Chen
,
Hongye
Zhang
,
Yanfei
Zhao
,
Changliang
Huang
and
Zhimin
Liu
*
Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing, 100190, P. R. China. E-mail: liuzm@iccas.ac.cn
First published on 7th August 2012
A new ionic liquid (IL), 1-butyl-3-methylimidazolium cholate, was first synthesized through an ion exchange reaction of 1-butyl-3-methylimidazolium chloride with sodium cholate. Stable aqueous dispersions of graphene were achieved by exfoliating graphite in the presence of the IL under ultrasonication. Both transmission electron microscopy and Raman measurements showed that the IL-stabilized graphene (IL–G) sheets existed with only a few (<5) layers. Furthermore, the IL–G was used to immobilize noble metal nanoparticles (Pt, Pd, Ru, Rh, etc), and a series of graphene–metal (G–M) composites with metal size ≤2 nm and very narrow size distributions were obtained. The resulting G–M exhibited superior catalytic performance with respect to hydrogenation of arenes. In particular, the as-prepared G–Ru with Ru content of 5% was very active for the hydrogenation of benzene to hexane with a turnover frequency as high as 6000 h−1. The catalysts could be reused without detectable loss of activity, a result of their stable structure.
In recent years, ionic liquids (ILs), a kind of organic salt with melting points below 100 °C, have been paid much attention owing to their unique properties, such as high charge density, high polarity, high dielectric constant, excellent solvent power for organic and inorganic compounds and supramolecular network formation, very low vapor pressure, etc. More importantly, the properties of ILs can be designed through judicious combination of anions and cations, and thereby ILs with unique functions can be obtained. Imidazolium cation-based ILs have been widely investigated, showing potential applications in many areas. In a recent work, dispersion of graphene oxide in IL, 1-butyl-3-methylimidazolium hexafluorophosphate, was successfully achieved with the aid of a polymerized ionic liquid (PIL).25 The surfactant-like property of some functional ILs may make the IL suitable to be applied in the production of graphene via exfoliating graphite. However, exfoliation of pristine graphite into graphene in ILs has not been found in a literature survey. On the other hand, very few studies have been carried out concerning the combination of ILs with graphene in catalysis.
Graphene as a 2D carbon material is expected to be an ideal support for metal and/or metal oxide nanocatalysts because of its excellent physical and chemical properties, including high mechanical strength, high thermal conductivity, outstanding electron mobility, and chemical inertness. Indeed graphene-based catalysts have shown promising applications in catalysis. For example, graphene-supported Pd particles displayed extraordinary high activities for Suzuki–Miyaura coupling reactions.26 Pt particles deposited on graphene exhibited good catalytic activities for both methanol oxidation and hydrogen conversion reactions.17 Graphene-supported Ru and Rh nanocomposites showed high efficiency in hydrogenations of cyclohexene and benzene.27 Despite recent advances in this regard, synthesis of graphene supported catalysts remains an ongoing challenge in terms of control of particle size, size distribution, and their thermal and/or electrochemical stability under different conditions.
In this work, a new functional IL, 1-butyl-3-methylimidazolium cholate, was designed and used to exfoliate pristine graphite to make graphene. The resulting IL-stabilized graphene sheets were further applied in the immobilization of noble metal particles including Ru, Pt and Rh. The as-obtained graphene and graphene-supported metal composites were characterized by IR and Raman spectroscopy, transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and thermogravimetric analysis (TGA). Moreover, the catalytic performances of the metal/graphene nanocomposites for the arene hydrogenation were investigated as well.
The IL–G was used to prepare G–M nanocomposites. In a typical experiment to prepare the G–Ru composite, 5 mL of RuCl3 ethanol solution at a designated concentration was mixed with 20 mL ethanol suspension of IL–G (1 mg mL−1), followed by dropwise addition of 5 mL of an NaBH4 ethanol solution (its concentration being 5 times higher than that of the metal precursor) under tip ultrasonication (Vibra Cell CVX, 500 W, 20 kHz, 25% of amplitude). The mixture was then centrifuged, and the collected precipitate was washed repeatedly with absolute ethanol and distilled water, and subsequently vacuum-dried at 60 °C. Using similar procedures, other G–M (including Rh, Pt, etc) composites with metal loading at 5 wt% were prepared as well by varying metal precursors.
Scheme 1 The chemical structure of the as-synthesized IL. |
The efficiency of the as-synthesized IL for graphite exfoliation was investigated under sonicating irradiation. It was found that stable graphene dispersions were achieved after being subjected to sonication in the IL aqueous solution for 24 h. The dispersion remained highly stable for several months. Fig. 1 shows the Raman spectra of graphite and IL–G. Compared to the spectrum of the used graphite, the spectrum of IL–G exhibited characteristics of graphene sheets with three peaks at 1350 cm−1, 1580 cm−1, and 2700 cm−1, which are ascribed to the G, D, and 2D peaks of graphene, respectively. The Raman analysis for IL–D confirms the formation of graphene sheets. Moreover, the shape of the 2D peak (∼2700 cm−1) was consistent with that of graphene sheets with fewer than 5 layers;23 especially, the symmetric 2D band further confirms the formation of single-layer graphene. The formation of graphene was also supported by the observation that a shift (∼29 cm−1) towards lower wavenumbers occurred for the 2D band of IL–G accompanied by an enhancement of its intensity when compared to that of graphite.
Fig. 1 (a) Raman spectra of graphite and IL–G; (b) IR spectra of graphite, IL–G and IL. |
Fig. 1b shows the IR spectra of the used graphite, IL–G and the IL. In comparison with the spectra of the IL and the graphite, the spectrum of IL–G displayed the characteristic peaks assigned to the IL, suggesting the presence of IL in this sample. This also suggests that the IL played important roles in exfoliating graphite and stabilizing the resultant graphene sheets.
To determine the amount of IL in IL–G, TGA analysis was carried out for both IL–G and IL. As shown in Fig. 2, the IL started to decompose at 270 °C and completely decomposed at about 570 °C. In contrast, there are three weight loss peaks in the TGA curve of IL–G. The weight loss of about 3% below 250 °C was ascribed to the removal of adsorbed water. The sharp weight loss between 250 and 570 °C was assigned to the decomposition of the IL, from which the amount of IL in IL–G was estimated to be approximately 52%. This is indicative of strong interactions between the graphene and the IL, resulting in stabilization of exfoliated graphene against reaggregation. The further weight loss in the range of 570–850 °C was attributed to the oxidation of the graphene sheets.
Fig. 2 TGA curves of IL and IL–G. |
A typical TEM image of IL-stabilized graphene flakes is illustrated in Fig. 3a, which clearly showed a large number of thin graphene sheets. Monolayer and bilayer graphene were observed as well, as displayed in Fig. 3b, indicating the successful exfoliation of graphene in the IL. The mechanism for the formation of stable graphene dispersion in the IL aqueous solution under sonicating irradiation can be explained as follows. The mechanical energy due to intensive sonication is sufficient to overcome the van der Waals attraction between graphite layers, thus the graphite was exfoliated into single- and few-layered sheets under the sonication irradiation. The attraction between the graphene surface and the hydrophobic groups of the IL molecules led to the adsorption of a large amount of IL molecules onto the graphene sheets, preventing them from restacking. The stabilization of the graphene dispersion in the IL aqueous solution may result from the electrostatic repulsion between the hydrophilic head ions coating adjacent graphene. Depending on the exfoliating extent of graphite, graphene sheets with different layers can be obtained. It should be pointed out that the dried IL–G can be redispersed in aqueous solutions to form stable suspension without aggregation.
Fig. 3 TEM images of IL–G sheets (a) and 2-layer graphene (b). |
Fig. 4 TEM images, particle size distribution and EDS spectra of G–M: (a) and (b) G–Ru (5 wt.%, 1.1 nm), (c) and (d) G–Rh (5 wt.%, 1.6 nm), (e) and (f) G–Pt (5 wt.%, 1.7 nm). |
Fig. 5 illustrates the XRD patterns of IL–G and G–Ru in the wide range of 2θ = 10–80°. The diffraction peak at 26.7° in both cases was assigned to the (111) reflection of the graphitic structure from graphene. Note that no diffraction peaks attributable to metal nanoparticles was observed for G–Ru with a Ru loading of 5 wt.% in the XRD patterns. This may be an indication that the Ru particles were very small.
Fig. 5 XRD patterns of G–Ru and IL–G. |
XPS measurements were employed to analyze the surface composition and oxidation states of species in the G–M composites. Fig. 6a shows the wide survey XPS pattern of G–Ru, which indicates the coexistence of C, Ru, N and O in the catalyst, and no Cl was detectable, suggesting its high purity without the presence of the metal precursor. The element N was originated from the IL. The deconvoluted Ru 3d XPS profile shows a doublet with peak binding energies (BEs) at 281.5 (Ru 3d5/2) and 285.5 eV (Ru 3d3/2), which were ascribed to a Ru–O species. Two peaks appearing at BEs of 485.5eV (Ru 3p1/2) and 463.4eV (Ru 3p3/2) were also attributed to a Ru–O species. These results suggest that the Ru species were present in the form of an oxide state in the composite. As expected, Ru species should be present in the form of a Ru(0) state because they were obtained via the reduction of RuCl3 by NaBH4. The presence of Ru oxide in G–Ru may result from the subsequent oxidation of Ru(0) particles as they were exposed to air. Moreover, the absence of Ru(0) in the G–Ru composite indicates that the initially obtained Ru(0) particles were completely converted into oxide, probably due to their tiny size. In addition, the other peaks at BEs of 285.4, 286.5 and 288.6 eV were ascribed to C–C, C–O and O–CO, respectively, which were originated from the IL.
Fig. 6 XPS spectra of G–Ru: (a) survey spectrum, (b) Ru 3d and C 1s, (c) Ru 3p. |
G–Pt and G–Rh were also examined by XPS analysis, and the metal oxides were detectable in these G–M composites as well, suggesting the oxidation of metal particles after the composites were exposed to air.
Entry | Catalyst | Substrate | T/°C | PH2/MPa | t/min | Conv./% | TOFa |
---|---|---|---|---|---|---|---|
a Turnover frequency (TOF) of substrate was defined as moles of substrate converted per moles of Ru per hour. b The catalyst was reused for the sixth run for benzene hydrogenation. | |||||||
1 | G–Ru | Benzene | 110 | 8 | 40 | >99.9 | 6060 |
2 | G–Rub | Benzene | 110 | 8 | 40 | >99.9 | 6060 |
3 | G–Ru | Benzene | 110 | 4 | 65 | >99.9 | 3730 |
4 | G–Ru | Benzene | 110 | 2 | 60 | 10 | 420 |
5 | G–Ru | Benzene | 60 | 4 | 170 | 68.5 | 990 |
6 | G–Ru | Benzene | 25 | 4 | 300 | 32.9 | 270 |
7 | G–Rh | Toluene | 110 | 8 | 68 | >99.9 | 3630 |
8 | G–Pt | Nitrobenzene | 60 | 2 | 13 | >99.9 | 36000 |
9 | TMG-MMT/Ru28 | Benzene | 110 | 8 | 150 | >99.9 | 4000 |
The G–Ru catalyst was very active for benzene hydrogenation to hexane, and the turnover frequency (TOF) reached over 6000 h−1 at 110 °C and H2 pressure at 8 MPa, which was much higher than that of most of the reported Ru catalysts.28 We note that the catalyst exhibited high activity even at room temperature (entry 6), and its activity increased with temperature and H2 pressure. It is known that Ru(0) is the active component for catalyzing hydrogenations. As analyzed above, the Ru particles were present in the form of oxide in the as-prepared G–Ru catalyst. Therefore, it can be deduced that the Ru oxide with sizes less than 2.0 nm can be easily reduced to the Ru(0) state during the hydrogenation process, which served as the actual active species for catalyzing benzene hydrogenation.
Moreover, the G–Ru catalyst could be reused 6 times without detectable loss of activity, suggesting the high stability of the catalyst. To explore the reason for its high stability, the recovered catalyst after 6 times was examined by TEM. It was shown that the metal particles remained highly distributed on the graphene surfaces without aggregation, and the particle size was kept unchanged, compared to the fresh G–Ru catalyst. The final product after reaction was examined by ICP, and it was found that the metal species were undetectable, indicating that there was no leaching of the metal species during the reaction process. As discussed above, the metal particles were immobilized onto the graphene sheets via the IL. The high activity and good stability of the catalyst indicates that the moderate interaction existed between the particles and the IL, which prevented particles from leaching into the reaction mixture and/or aggregating into larger ones.
Similarly, the as-prepared G–Rh and G–Pt catalysts were applied to the hydrogenation of toluene and nitrobenzene, respectively, which showed remarkably high activities (entries 7 and 8) and good stabilities without activity loss even after 5 times recovered use. The high performance of the as-synthesized nanocatalysts may result from their unique structures, in which small metal particles with high monodispersibility were strongly immobilized on the graphene sheets with the assistance of the IL.
Footnote |
† W. J. Xiao and Z. Y. Sun equally contributed to this work. |
This journal is © The Royal Society of Chemistry 2012 |