- Email: [email protected]
Synthesis, characterization and growth mechanism of mesoporous hollow carbon nanospheres by catalytic carbonization of polystyrene
Microporous and Mesoporous Materials 176 (2013) 31–40
Contents lists available at SciVerse ScienceDirect
Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso
Synthesis, characterization and growth mechanism of mesoporous hollow carbon nanospheres by catalytic carbonization of polystyrene Jiang Gong a,b, Jie Liu a, Xuecheng Chen a, Xin Wen a, Zhiwei Jiang a, Ewa Mijowska c, Yanhui Wang a, Tao Tang a,⇑ a b c
State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, China University of Chinese Academy of Sciences, Beijing 100049, China Institute of Chemical and Environment Engineering, West Pomeranian University of Technology, Szczecin ul. Pulaskiego 10, 70-322 Szczecin, Poland
a r t i c l e
i n f o
Article history: Received 15 January 2013 Received in revised form 22 March 2013 Accepted 27 March 2013 Available online 6 April 2013 Keywords: Mesoporous hollow carbon nanospheres Polystyrene Organically-modified montmorillonite Cobalt catalyst Growth mechanism
a b s t r a c t Mesoporous hollow carbon nanospheres (HCNs) were synthesized through the carbonization of polystyrene (PS) under the combined catalysis of organically-modified montmorillonite (OMMT)/cobalt catalyst at 700 °C. The morphology, microstructure, phase structure, textural property and surface composition of the obtained mesoporous HCNs were investigated by field-emission scanning electron microscope, transmission electron microscope (TEM), high-resolution TEM, X-ray diffraction, Raman spectroscopy, N2 sorption and X-ray photoelectron spectroscopy. It was found that OMMT not only promoted the dispersion of cobalt catalyst in the PS matrix but also affected the degradation of PS into light hydrocarbons and aromatics. The lattice oxygen of the cobalt catalyst facilitated the decomposition of light hydrocarbons and aromatics into atomic carbon during the formation of the mesoporous HCNs. A possible mechanism was proposed to explain the growth of mesoporous HCNs through the carbonization of PS under the combined catalysis of OMMT/cobalt catalyst. More importantly, this approach also offers a new potential way to transform waste polymer materials into mesoporous HCNs, which may be used as catalyst supports, adsorbents, storage media and templates for the synthesis of other useful hollow materials. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Hollow carbon spheres (HCSs) with a structure of hollow core and carbon shell have attracted great attention, due to their unique physicochemical properties such as low density, large inner space and specific surface area, and wide applications in catalyst supports, adsorbents, storage media and templates for the synthesis of other useful hollow materials [1–9]. The common methods for the preparation of HCSs are chemical vapor deposition [10,11], hydrothermal treatment [7] and template-assisted method [4,7,12–16] using ethylene [10], acetonitrile [11], sucrose [11], dopamine [14], acetone [17] or furfuryl alcohol [18] etc. as carbon source. From a sustainable point of view, reutilization of waste polymer materials to synthesize carbon spheres not only shows advantages with the nontoxic, cheap and abundant sources, but also provides a potential way to recycle waste polymer materials. Recently, some studies have demonstrated that virgin or waste polymer materials including polyethylene (PE), polypropylene (PP), polystyrene (PS) ⇑ Corresponding author. Tel.: +86 (0) 431 85262004; fax: +86 (0) 431 85262827. E-mail address: [email protected] (T. Tang). 1387-1811/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.micromeso.2013.03.039
and poly(ethylene terephthalate) (PET) can be converted into high value-added solid carbon microspheres (SCMs) [19–23]. For example, Pol et al. converted waste PE, PS and PET into SCMs using autoclave as reactor at 700 °C [19,20]. Qian et al. converted PP into SCMs using autoclave as reactor at 700 °C [21]. Chen et al. prepared SCMs by pyrolyzing waste PET in supercritical carbon dioxide at 500–650 °C [22]. However, to the best of our knowledge, so far, there has been only one report about synthesizing hollow carbon nanospheres (HCNs) using polymer material as carbon source. Chen et al. prepared HCNs using PP as carbon source and the combined organically-modified montmorillonite (OMMT)/Co(Ac)2 as catalyst at 900 °C [23]. But the HCNs showed extreme heterogeneity in the size distribution, and they did not provide information about the textural property of HCNs, nor further study the growth mechanism of HCNs. Hence, it is of great desire to synthesize HCNs with a narrow diameter distribution and further investigate the growth mechanism of HCNs using polymer as carbon source. Herein, based on our ‘‘combined catalysis’’ strategy to prepare carbon nanotubes with micro/mesoporous structure [24–29], mesoporous HCNs were synthesized by catalytic carbonization of PS under the combined catalysis of OMMT and cobalt catalyst.
32
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 1. Photographs of PS pellets (a, middle) and the HCN-1 (a, right) synthesized by pyrolyzing PS/10OMMT-10Co2O3 at 700 °C and typical SEM images of the HCN-1 (b–d) at various magnifications.
Fig. 2. Typical TEM (a and b) and HRTEM (c and d) images of the HCN-1.
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Firstly, the morphology, microstructure, phase structure, textural property and surface composition of the mesoporous HCNs were investigated. Subsequently, the effects of OMMT on the dispersion of cobalt catalyst in the PS matrix and the degradation products of PS were studied. Finally, a possible mechanism was put forward to explain the growth of mesoporous HCNs using PS as carbon source. This approach offers a new potential way to transform waste polymer materials into valuable mesoporous HCNs, which may be used as catalyst supports, adsorbents, storage media and templates for the synthesis of other useful hollow materials.
33
2.2. Preparation of samples PS pellets (40.00 g) were mixed with designed amounts of OMMT and cobalt catalyst in a Brabender mixer at 100 rpm and 180 °C for 10 min. The resultant sample was designed as PS/ xOMMT-yCat, where x and y represented the contents of OMMT and cobalt catalyst, respectively, and Cat represented the cobalt catalyst. For comparison, PS/10OMMT and PS/10Cat mixtures were also prepared.
2.3. Preparation of mesoporous HCNs 2. Experimental part 2.1. Materials PS pellets were supplied by Zhenjiang Qimei Chemical Co., Ltd., China. Four common cobalt catalysts: Co2O3, Co3O4, Co(Ac)24H2O and metallic cobalt (Co) powder, purchased from commercial vendors (Sinopharm Chemical Reagents Co., Ltd., China and Alfa Aesar), were of analytical-grade quality and used without further purification. OMMT (Closite 15A, organic modifier: dimethyl-dihydrogenated tallow quarternary ammonium; modifier concentration: 125 mequiv per 100 g clay) was purchased from Southern Clay. All other chemicals were of analytical-grade quality.
Mesoporous HCNs were prepared according to our previous method ‘‘combustion experiment’’ [27]. Briefly, a piece of sample (about 5.0 g) was placed into a crucible, which was heated at 700 °C until the flame from the crucible upper brim went out (after about 5 min). Subsequently, the charred residue in the crucible was purified with hydrofluoric acid and nitric acid to eliminate montmorillonite (MMT), amorphous carbon and most of cobalt catalysts. The yield of mesoporous HCNs was calculated by dividing the amount of purified carbon products by that of carbon element in the PS from the sample. All of the samples were weighed by analytical balance. The mesoporous HCNs from PS/ 10OMMT-10Co2O3, PS/10OMMT-10Co3O4 and PS/10OMMT10Co(Ac)2 were denoted as HCN-1, HCN-2 and HCN-3, respectively.
Fig. 3. Typical TEM images of the HCN-2 from PS/10OMMT-10Co3O4 after purification (a and b) and the HCN-3 from PS/10OMMT-10Co(Ac)2 (c and d) after purification at 700 °C.
34
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 4. XRD patterns (a), Raman spectra (b), nitrogen adsorption–desorption isotherms (c) and pore size distributions (d) of the resultant HCNs.
Table 1 Textural parameters of the obtained HCNs.
a b c d e
Property
HCN-1
HCN-2
HCN-3
SBET (m2/g) a Smeso (m2/g) b Vtotal (cm3/g) c Vmeso (cm3/g) d DAV (nm) e
111.6 88.7 0.490 0.479 3.77
149.8 149.8 0.509 0.509 3.80
227.3 227.3 1.020 1.020 3.77
The The The The The
total specific surface area. specific surface area of mesopores. total volume. volume of mesopores. average diameter of pores.
The degradation products of PS are the carbon feedstock for the formation of the mesoporous HCNs. To study the effect of OMMT on the degradation products of PS, pyrolysis experiments [27] for PS and PS/10OMMT mixture were conducted at 700 °C. The liquid pyrolyzed products were collected using a cold trap and the gas pyrolyzed products were collected using a sample bag. 2.4. Characterization The morphology of the mesoporous HCNs was observed by means of field-emission scanning electron microscope (SEM, XL30ESEM-FEG). The microstructure of the mesoporous HCNs was investigated using transmission electron microscope (TEM, JEM-1011) at an accelerating voltage of 100 kV and high-resolution
TEM (HRTEM) on a FEI Tecnai G2 S-Twin transmission electron microscope operating at 200 kV. The phase structure of the mesoporous HCNs was analyzed by X-ray diffraction (XRD) using a D8 advance X-ray diffractometer with Cu Ka radiation operating at 40 kV and 200 mA. Raman spectroscopy (T6400, excitationbeam wavelength: 514.5 nm) was used to characterize the vibrational properties of the mesoporous HCNs. The textural properties of the mesoporous HCNs were measured by N2 sorption at 77 K using a Quantachrome Autosorb-1C-MS analyzer. The specific surface area was calculated by BET method, and the contribution of micropores to both volume and surface area was evaluated by means of the t-plot method. The surface element composition of the mesoporous HCNs was characterized by means of X-ray photoelectron spectroscopy (XPS) carried out on a VG ESCALAB MK II spectrometer using an Al Ka exciting radiation from an X-ray source operated at 10.0 kV and 10 mA. Thermal gravimetric analysis (TGA) was performed using TA Instruments SDT Q600 at a heating rate of 10 °C/min. The rheological properties of PS and its mixture were conducted on a controlled strain rate rheometer (ARES rheometer) under nitrogen atmosphere. Round samples 25 mm (diameter) 1 mm (thickness) were run at 180 °C. Frequency sweep was performed from 0.01 to 100 rad/s, with a strain of 1% in order to make the materials be in linear viscoelastic response. Ultrathin sections of PS and its mixture were cryogenically cut using a Leica Ultracut and a glass knife at room temperature. The samples were collected on carbon-coated copper TEM grids. To study the effect of OMMT on the degradation products of PS, the collected liquid products were weighed and analyzed by gas chromatography–mass spectrometry (GC–MS, AGILENT
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
35
Fig. 5. XPS patterns of the mesoporous HCNs (a) and C1s high-resolution XPS spectra of HCN-1 (b), HCN-2 (c) and HCN-3 (d).
5975MSD). The volume of collected gas products was determined by the displacement of water. The hydrocarbon gas products were analyzed by a GC (Kechuang, GC 9800) equipped with a FID, using a KB-Al2O3/Na2SO4 column (50 m 0.53 mm ID). H2, CO and CH4 were analyzed by a GC (Kechuang, GC 9800) equipped with a TCD, using a packed TDX-01 (1 m) and molecular sieve 5A column (1.5 m). 3. Results and discussion 3.1. Morphology and microstructure of the resultant HCNs It is found from Fig. 1a that the HCN-1 from the carbonization of PS/10OMMT-10Co2O3 mixture at 700 °C can be produced on a gram scale by using this approach. The morphology observation of the HCN-1 by SEM at various magnifications showed that it possessed spherical-shape structure (Fig. 1b–d). In order to gain more detailed information about the internal microstructure of the HCN-1, TEM and HRTEM observations were conducted. Fig. 2a and b clearly showed that the HCN-1 was comprised of carbon nanospheres with central hollow inside. The main diameter of the HCN-1 was in the range of 60–90 nm with an average diameter of 73.5 nm. HRTEM images of the HCN-1 are shown in Fig. 2c and d, further revealing the microstructure of hollow carbon nanospheres. The carbon shell had an ordered and curved graphitic structure. The discontinuous graphitic layers and the numerous defects in the shells of the HCN-1 are not only propitious to the
Li-ion insertion and extraction but also to the diffusion of Li-ions, so it is expected that the obtained HCN-1 may be used in the lithium electrochemical cells. Furthermore, the interlayer spacing between graphitic layers was 0.33–0.35 nm, consistent with the ideal graphitic interlayer spacing. The thicknesses of the graphitic shells fell in the range of 6–12 nm. The yield of the HCN-1 was 11.1 wt.%. Other two common cobalt catalysts including Co3O4 and Co(Ac)2 were also used as catalysts for synthesizing HCNs. Fig. 3 presents the typical TEM images of the HCN-2 from PS/10OMMT10Co3O4 and the HCN-3 from PS/10OMMT-10Co(Ac)2 at 700 °C. Similar to the HCN-1, both HCN-2 and HCN-3 were hollow carbon nanospheres. In the case of the HCN-2 (Fig. 3a and b), the main nanosphere diameters and the graphitic shell thicknesses were in the range of 60–85 nm and 5–12 nm, respectively, which were close to those of the HCN-1. The HCN-3 showed smaller diameters (20–40 nm) and thinner graphitic shells (3–8 nm) (Fig. 3c and d). The yield of the HCN-2 and HCN-3 was 9.8 and 4.8 wt.%, respectively. The reason for the low yield of HCNs was analyzed in Section 3.3.3 and then a simple way was also proposed to increase the yield of HCNs. 3.2. Phase structure, textural property and surface composition of the resultant HCNs The graphitization degree of the resultant HCNs were characterized by XRD and Raman spectroscopy. Fig. 4a displays the XRD patterns of the resultant HCNs. The two diffraction peaks at about
36
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 6. Typical TEM images of PS/10Co2O3 mixture (a) and PS/10OMMT-10Co2O3 mixture (b). Rheological curves of PS and its mixture: complex viscosity (c) and storage modulus (d).
Table 2 Yields of the different fractions through the pyrolysis of PS and PS/10OMMT at 700 °C. Product
PS
PS/10OMMT
Carbon (g/100 g PS) Liquid (g/100 g PS) Gas (g/100 g PS) a Gas (g/100 g PS) b Gas (mL/100 g PS) c
0 99.873 0.127 0.084 1.848
1.015 94.595 4.390 3.887 59.145
a
Calculated by the material balance. Calculated by the volume of the gas (Supplementary Table S1) first divided by 22.4 L/mol, then multiply the molar number of the gas by its molar mass, finally obtained the yield of the total gases by adding the mass of each gas together. c Calculated by the volume of the gas divided by the mass of PS, finally obtained the yield of the total gases by adding the volume of each gas together. b
26.0° and 43.2° were assigned to the typical graphitic (0 0 2) and (1 0 1) planes, respectively. The broadening of the two peaks suggested the low graphitization degree. Compared to the HCN-2 and HCN-3, the relative sharp and strong graphite diffraction peaks of the HCN-1 indicated the relative higher degree of graphitization among these HCNs [11]. The weak diffraction peaks of metallic cobalt at 44.4°, 51.6° and 75.9° were also observed, implying that a
trace amount of metallic cobalt catalysts remained in the resultant HCNs. Fig. 4b shows the Raman spectra of the resultant HCNs. The peak at about 1580 cm 1 (G band) corresponds to an E2g mode of hexagonal graphite and is related to the vibration of sp2-bonded carbon atoms in a graphite layer, and the D band at about 1350 cm 1 is associated with the vibration of carbon atoms with dangling bonds in the plane terminations of disordered graphite [30]. The intensity of the G and D bands provides information about the crystallinity of the HCNs. Larger IG/ID ratio indicates a higher degree of structural ordering for the HCNs. Compared to the HCN-2 and HCN-3, the relative larger IG/ID ratio for the HCN-1 reflected the relative high degree of graphitization and the presence of a relative low amount of disordered carbon in the HCN-1, consistent with the XRD results (Fig. 4a). The nitrogen adsorption–desorption isotherms of the obtained HCNs (Fig. 4c) showed the type-IV curve and exhibited a hysteresis loop associated to capillary condensation in the range of P/P0 being from 0.5 to 1.0. This means that the porosity of the obtained HCNs is essentially made up of mesopores. The BET surface area (SBET), mesopore surface area (Smeso), total pore volume (Vtotal), mesopore volume (Smeso) and average pore diameter (DAV) of the HCNs are summarized in Table 1. Interestingly, the values of SBET, Smeso, Vtotal
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
37
and Vmeso increased in the following sequence: HCN-1 < HCN2 < HCN-3, possibly due to the decrease of the diameter of HCNs. The pore-size distributions of the HCNs were calculated using the Barrett–Joyner–Halenda (BJH) model from the desorption branches of the isotherms (Fig. 4d). This clearly showed that the size of the mesopores in the HCNs was in the narrow range of 2–5 nm (centered on 3.8 nm). The mesopores could be attributed to the cavities in the HCNs. The formation of the mesoporous HCNs and their mesopores was discussed in Section 3.3.3. XPS measurements were used to characterize the surface element composition of the mesoporous HCNs. It revealed that the surface of mesoporous HCNs mainly consisted of C (about 82 at.%) and O (about 18 at.%), with no evidence of any other elements (Fig. 5a). To determine the chemical components and the oxidation states of C element, high-resolution XPS spectra of C1s were curve-fitted into three individual peaks: graphitic carbon (284.6 eV), C-OH (285.6–285.8 eV) and COOH (288.9 eV) (Fig. 5b– d) [29]. These results showed that the carbon on the surface of the mesoporous HCNs existed in the presence of graphitic carbon with relatively small amounts of C–OH and COOH. These surface functional groups could contribute to the removal of heavy metallic ions [31] or organic dyes [32] when using the mesoporous HCNs as adsorbents in wastewater treatment. 3.3. Discussion about the growth mechanism of the mesoporous HCNs 3.3.1. OMMT promoting the dispersion of cobalt catalyst in the PS matrix To elucidate the roles of OMMT and cobalt catalyst in the formation of the mesoporous HCNs, firstly, PS/10Co2O3 and PS/ 10OMMT mixtures were pyrolyzed under the same condition. The yield of carbon product from PS/10Co2O3 and PS/10OMMT was 1.2 and 1.1 wt.%, respectively, much lower than that from PS/10OMMT-10Co2O3 (11.1 wt.%), suggesting that the combined OMMT/cobalt catalyst promoted the carbonization of PS. Furthermore, the carbon product from PS/10OMMT was amorphous carbon (not shown), demonstrating that OMMT was not the real active site for the formation of mesoporous HCNs. The pyrolysis of PS/10Co2O3 also led to the formation of hollow carbon nanospheres (Fig. S1). This indicated that the cobalt catalysts were the real active sites for the growth of mesoporous HCNs. However, the diameter of hollow carbon nanospheres from PS/10Co2O3 was in the range of 100–250 nm, obviously larger than that from PS/ 10OMMT-10Co2O3 (HCN-1, 60–90 nm). This was possibly because that OMMT promoted the dispersion of Co2O3 catalysts into smaller size in the PS matrix. To confirm the above speculation, the dispersion states of Co2O3 catalysts in the PS matrix with or without the addition of OMMT are shown in Fig. 6. It could be observed that without the addition of OMMT, a large amount of Co2O3 catalysts were agglomerated and the mean size was 379 nm (Fig. 6a). Interestingly, after the addition of OMMT, the dispersion degree of Co2O3 catalysts was significantly higher than that without OMMT, especially, the size of Co2O3 catalysts was evidently smaller (180 nm) than that without OMMT (Fig. 6b). There are two possible reasons for the above phenomenon. One of them results from the viscosity increase of the PS matrix due to the presence of dispersed OMMT (Fig. 6c and d), which increases shear force during melt mixing and improves the dispersion degree of the cobalt catalysts. The other is possible interaction between the OMMT and cobalt catalysts, which can prevent the re-aggregation of the cobalt catalysts. 3.3.2. OMMT affecting the degradation products of PS The degradation products of PS are carbon sources for the formation of mesoporous HCNs. The addition of OMMT may influence the composition of the degradation products from PS. In order
Fig. 7. GC–MS profiles of liquid degradation products from PS and PS/10OMMT at 700 °C.
Fig. 8. XRD patterns of PS/10Co2O3 mixture (a) and its carbonized product at 700 °C before purification (b), and PS/10OMMT-10Co2O3 mixture (c) and its carbonized product at 700 °C before purification (d).
to study the influence of OMMT on the degradation products of PS, GC and GC–MS measurements were conducted to analyze the composition of the degradation products from PS and PS/10OMMT mixture at 700 °C. Table 2 presents the material balance of the degradation products from PS and PS/10OMMT mixture. It was found that after adding OMMT, the quantity of gas degradation products significantly increased from 0.127 to 4.390 (g/100 g PS), indicating that OMMT promoted the degradation of PS into light hydrocarbons. Table S1 shows the composition of gas degradation products from these mixtures in detail. The gas degradation products mainly consisted of hydrogen, methane, ethane, ethylene, propane, propylene and i-butene. Compared to pure PS, the yields of methane,
38
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 9. Possible mechanism about the formation of the mesoporous HCNs through the carbonization of PS under the combined catalysis of OMMT/cobalt catalyst at 700 °C.
ethylene and propylene increased significantly. For example, the yield of methane increased from 0.644 to 24.970 (mL/100 g PS). Fig. 7 displays the GC–MS profiles of liquid degradation products from PS and PS/10OMMT. The components of the liquid degradation products were quantitatively analyzed and the results were listed in Table S2. The main liquid degradation products from pure PS were toluene (3.118% area), styrene (65.381% area) and 2phenyl-1H-indene (14.893% area). Interestingly, after adding 10 wt.% OMMT, the content of styrene significantly decreased to 20.366 (%area), meanwhile other aromatic compounds (such as toluene, xylene and naphthalene) increased obviously. The above results demonstrated that OMMT promoted the degradation of
PS into light hydrocarbons and aromatics, which are widely used as carbon sources for the synthesis of carbon nanospheres [10,33,34]. 3.3.3. Real process for the formation of the mesoporous HCNs It is necessary to clear whether metallic cobalt catalysts are real active sites for understanding the growth mechanism of the mesoporous HCNs. To confirm the role of metallic cobalt catalysts during the growth of the mesoporous HCNs, XRD was used to monitor states of cobalt catalysts before and after the growth of mesoporous HCNs (Fig. 8). Compared to the XRD patterns of PS/ 10Co2O3 and PS/10OMMT-10Co2O3 mixtures, the diffraction peaks
Fig. 10. TEM (a and b) and HRTEM (c and d) images of Co@C-1 core/shell composite from PS/10OMMT-10Co2O3 at 700 °C before purification.
39
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
of PS (2h = 20.4°) and Co2O3 (2h = 19.5°, 31.7°, 37.3°, 39.0°, 42.8°, 45.2°, 56.1°, 59.8°, 62.0° and 65.6°) disappeared completely after the carbonization of PS, meanwhile the diffraction peaks of metallic cobalt (2h = 44.3°, 51.5° and 76.0°) appeared obviously. This indicated that Co2O3 catalysts were reduced into metallic cobalt catalysts, which were the real active sites for the formation of the mesoporous HCNs. Similar results were found using Co3O4 or Co(Ac)2 as catalyst with or without OMMT (not shown). To get more information about the formation process of the mesoporous HCNs, metallic cobalt powder (Co) in place of Co2O3 was added into PS/10OMMT mixture and the resultant PS/ 10OMMT-10Co mixture was also pyrolyzed under the same condition. The yield of carbon product from PS/10OMMT-10Co was 12.8 wt.%, similar to that from PS/10OMMT-10Co2O3, suggesting that the combined OMMT/metallic cobalt powder also promoted the carbonization of PS. However, the carbon product from PS/ 10OMMT-10Co was actually irregular carbon rather than mesoporous HCNs (Fig. S2). This result suggested that the reduction of cobalt oxide catalysts to metallic cobalt catalysts was an essential process for the formation of the mesoporous HCNs. Qian et al. reported that the unreduced Ni/Cu/Al2O3 catalyst showed much higher efficiency on the decomposition of methane than reduced Ni/Cu/Al2O3 catalyst, and they believed that the lattice oxygen on the unreduced catalyst accelerated the decomposition of methane into carbon [35]. Similarly in this work, the lattice oxygen on the cobalt oxide may promote the decomposition of light hydrocarbons and aromatics into atomic carbon. Hence, we speculated that
Fig. 11. Nitrogen adsorption–desorption isotherms (a) and pore size distributions (b) of the resultant Co@C core/shell composites. Co@C-2 and Co@C-3 were from PS/ 10OMMT-10Co3O4 and PS/10OMMT-10Co(Ac)2, respectively.
Table 3 Textural parameters of the resultant Co@C core/shell composites. Property 2
a
SBET (m /g) Smeso (m2/g) b Vtotal (cm3/g) c Vmeso (cm3/g) d DAV (nm)e a b c d e
The The The The The
Co@C-1
Co@C-2
Co@C-3
68.5 59.0 0.227 0.222 3.76
123.5 106.7 0.336 0.330 3.77
137.1 119.5 0.368 0.359 3.75
total specific surface area. specific surface area of mesopores. total volume. volume of mesopores. average diameter of pores.
the formation process of the mesoporous HCNs through the carbonization of PS included the reduction process of cobalt oxide catalysts into metallic cobalt catalysts and the simultaneous carbonization process of the degradation products from PS (such as methane and benzene) into carbon shells on the surface of in situ formed metallic cobalt catalysts. Based on the above results, a possible mechanism about the formation of the mesoporous HCNs through the catalytic carbonization of PS was put forward (Fig. 9), similar to the generally accepted dissociation–diffusion–precipitation mechanism for the growth of carbon nanotubes [36]. Firstly, OMMT promoted the dispersion of cobalt oxide catalysts in the PS matrix (step 1); subsequently, OMMT also promoted the degradation of PS into light hydrocarbons and aromatics (step 2). Secondly, the resultant light hydrocarbons and aromatics were dehydrogenated and aromatizated on the surface of cobalt oxide catalysts, meanwhile the cobalt oxide catalysts (such as Co2O3) were reduced into metallic cobalt catalysts (step 3), which have some (about 1 at.%) carbon solubility in the solid solution [37]. After further carbonization of carbon sources, once supersaturated, carbon precipitated from the surface of metallic cobalt nanoparticles (step 4) to form Co@C core/shell composites (step 5, Fig. 10 and Fig. S3), which were essentially made up of mesopores according to the nitrogen adsorption–desorption isotherms and the pore-size distributions (Fig. 11 and Table 3). Obviously, the formation of Co@C core/shell composites played an important role in the synthesis of mesoporous HCNs and the metallic cobalt nanoparticles from the reduction of cobalt oxide catalysts actually acted as templates for the growth of HCNs. In the case of cobalt catalyst such as Co(Ac)2, cobalt catalyst was firstly in situ decomposed into smaller cobalt oxide nanoparticles such as Co2O3 or Co3O4 (Fig. S4). This is the possible reason why the average diameter of HCN-3 from PS/10OMMT10Co(Ac)2 is much smaller than that of HCN-1 from PS/10OMMT10Co2O3 or HCN-2 from PS/10OMMT-10Co3O4. In addition, during the formation of Co@C core/shell composites, the lattice oxygen on the cobalt oxide catalysts promoted the decomposition of light hydrocarbons and aromatics into atomic carbon. Subsequently, the yielded carbon dissolved into the metallic cobalt nanoparticles. This is the possible reason why the carbon product using metallic cobalt powder directly as catalyst was irregular carbon rather than mesoporous HCNs (Fig. S2). Finally, after the removals of metallic cobalt catalysts and MMT, the mesoporous HCNs were obtained (step 6). Owing to the short synthesis time (about 5 min, see Section 2.3) and the low content of cobalt oxide (10 wt.%), the carbon shell of Co@C core/shell composites could not grow very thick (Fig. 10 and Fig. S3). Hence, the yield of mesoporous HCNs was limited. But the yield can be increased by increasing the content of cobalt oxide. For example, when the content of Co2O3 increased from 10 to 30 wt.% into PS/10OMMT, the yield of mesoporous HCNs dramatically increased from 11.1 to 35.0 wt.%. Besides, further studies are needed to evaluate the ef-
40
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
fect of additives in the waste PS on the formation of mesoporous HCNs and more investigation will be conducted in the further work. According to the above mechanism, the formation of mesopores was promoted by water etching and HNO3 oxidation. Water was in situ yielded as byproduct during cobalt oxide catalyzing carbonization of PS degradation products. At high temperature (such as 700 °C), water could efficiently etch the carbon shell of Co@C core/shell composites and favor the formation of mesopores. Similar results were reported by Pol et al. [34] and they believed that water could help the formation of pores in carbon microspheres. In addition, during the purification of Co@C core/shell composites, HNO3 permeated into internal structure of HCNs after the removal of cobalt core and further etched the inner surface carbon to form more mesopores. Han et al. [3] also found that mesopores were generated during the removal of the residual Zn and ZnO dispersed in the HCNs or intercalated into the graphitic shells by acid purification. 4. Conclusion We reported a simple approach to synthesize mesoporous HCNs through the carbonization of PS by the combined catalysis of OMMT/cobalt catalyst at 700 °C. The diameter of the mesoporous HCNs was in the range of 60–90 nm using OMMT/Co2O3, 60–85 nm using OMMT/Co3O4, or 20–40 nm using OMMT/Co(Ac)2. The total surface area and pore volume of the mesoporous HCNs were 111.6–227.3 m2/g and 0.49–1.02 cm3/g, respectively. The size of mesopores in the HCNs was in the narrow range of 2–5 nm. OMMT not only promoted the dispersion of cobalt catalyst in the PS matrix but also affected the degradation of PS into light hydrocarbons and aromatics. The lattice oxygen on the cobalt oxide catalysts promoted the decomposition of light hydrocarbons and aromatics into atomic carbon during the formation of the mesoporous HCNs. At last, a possible growth mechanism of the mesoporous HCNs using PS as carbon source and combined OMMT/cobalt catalyst as catalyst was proposed. This mechanism will be favorable to understand the growth of the mesoporous HCNs using polymer as carbon source. More importantly, this approach offers a new potential way to transform waste polymers into mesoporous HCNs, which may be used as catalyst supports, adsorbents, storage media and templates for synthesis of other useful hollow materials. Also, further studies are needed to evaluate the effect of additives in the waste polymers on the formation of mesoporous HCNs. Acknowledgments This work was supported by the National Natural Science Foundation of China (2124079, 50873099 and 20804045) and Polish Foundation (No. 2011/03/D/ST5/06119). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.micromeso.2013. 03.039.
References [1] A. Nieto-Márquez, R. Romero, A. Romero, J. Valverde, J. Mater. Chem. 21 (2011) 1664–1672. [2] A.A. Deshmukh, S.D. Mhlanga, N.J. Coville, Mater. Sci. Eng. R 70 (2010) 1–28. [3] F.D. Han, Y.J. Bai, R. Liu, B. Yao, Y.X. Qi, N. Lun, J.X. Zhang, Adv. Energy Mater. 1 (2011) 798–801. [4] S.B. Yang, X.L. Feng, L.J. Zhi, Q. Cao, J. Maier, K. Müllen, Adv. Mater. 22 (2010) 838–842. [5] A.H. Lu, G.P. Hao, Q. Sun, X.Q. Zhang, W.C. Li, Macromol. Chem. Phys. 213 (2012) 1107–1131. [6] G.H. Sun, K.X. Li, L.J. Xie, J.L. Wang, Y.Q. Li, Micropor. Mesopor. Mat. 151 (2012) 282–286. [7] K. Tang, L.J. Fu, R.J. White, L.H. Yu, M.-M. Titirici, M. Antonietti, J. Maier, Adv. Energy Mater. 2 (2012) 873–877. [8] J.H. Kim, B.Z. Fang, S.B. Yoon, J.-S. Yu, Appl. Catal. B: Environ. 88 (2009) 368– 375. [9] B.Z. Fang, M. Kim, J.H. Kim, J.-S. Yu, Langmuir 24 (2008) 12068–12072. [10] X.C. Chen, K. Kierzek, Z.W. Jiang, H.M. Chen, T. Tang, M. Wojtoniszak, R.J. Kalenczuk, P.K. Chu, E. Borowiak-Palen, J. Phys. Chem. C 115 (2011) 17717– 17724. [11] Z.X. Yang, R. Mokaya, Micropor. Mesopor. Mat. 113 (2008) 378–384. [12] A.H. Lu, T. Sun, W.C. Li, Q. Sun, F. Han, D.H. Liu, Y. Guo, Angew. Chem. Int. Ed. 50 (2011) 11765–11768. [13] K.-M. Choi, K. Kuroda, Chem. Commun. 47 (2011) 10933–10935. [14] R. Liu, S.M. Mahurin, C. Li, R.R. Unocic, J.C. Idrobo, H.J. Gao, S.J. Pennycook, S. Dai, Angew. Chem. Int. Ed. 50 (2011) 6799–6802. [15] G. Sun, K.X. Li, J. Wang, H.W. He, J.L. Wang, Y.Q. Li, Y. Liu, J.Y. Gu, Micropor. Mesopor. Mat. 139 (2011) 207–210. [16] M. Kim, S.B. Yoon, K. Sohn, J.Y. Kim, C.-H. Shin, T. Hyeon, J.-S. Yu, Micropor. Mesopor. Mat. 63 (2003) 1–9. [17] G.D. Li, C.L. Guo, C.H. Sun, Z.C. Ju, L.S. Yang, L.Q. Xu, Y.T. Qian, J. Phys. Chem. C 112 (2008) 1896–1900. [18] Y.S. Li, Y.Q. Yang, J.L. Shi, M.L. Ruan, Micropor. Mesopor. Mat. 112 (2008) 597– 602. [19] V.G. Pol, M.M. Thackeray, Energy Environ. Sci. 4 (2011) 1904–1912. [20] V.G. Pol, Environ. Sci. Technol. 44 (2010) 4753–4759. [21] J.H. Zhang, J. Li, J. Cao, Y.T. Qian, Mater. Lett. 62 (2008) 1839–1842. [22] L.Z. Wei, N. Yan, Q.W. Chen, Environ. Sci. Technol. 45 (2011) 534–539. [23] X.C. Chen, H. Wang, J.H. He, Nanotechnology 19 (2008) 325607. [24] T. Tang, X.C. Chen, X.Y. Meng, H. Chen, Y.P. Ding, Angew. Chem. Int. Ed. 44 (2005) 1517–1520. [25] Z.W. Jiang, R.J. Song, W.G. Bi, J. Lu, T. Tang, Carbon 45 (2007) 449–458. [26] R.J. Song, Z.W. Jiang, W.G. Bi, W.X. Cheng, J. Lu, B.T. Huang, T. Tang, Chem. Eur. J. 13 (2007) 3234–3240. [27] J. Gong, J. Liu, L. Ma, X. Wen, X.C. Chen, D. Wan, H.O. Yu, Z.W. Jiang, E. Borowiak-Palen, T. Tang, Appl. Catal. B: Environ. 117–118 (2012) 185–193. [28] J. Gong, K. Yao, J. Liu, X. Wen, X.C. Chen, Z.W. Jiang, E. Mijowska, T. Tang, Chem. Eng. J. 215–216 (2013) 339–347. [29] J. Gong, J. Liu, D. Wan, X.C. Chen, X. Wen, E. Mijowska, Z.W. Jiang, Y.H. Wang, T. Tang, Appl. Catal. A: Gen. 449 (2012) 112–120. [30] M. Ignat, C.J. Van Oers, J. Vernimmen, M. Mertens, S. Potgieter-Vermaak, V. Meynen, E. Popovici, P. Cool, Carbon 48 (2010) 1609–1618. [31] R. Demir-Cakan, N. Baccile, M. Antonietti, M.-M. Titirici, Chem. Mater. 21 (2009) 484–490. [32] X.H. Song, Y.B. Wang, K.A. Wang, R. Xu, Ind. Eng. Chem. Res. 51 (2012) 13438– 13444. [33] Y.Z. Jin, C. Gao, W.K. Hsu, Y.Q. Zhu, A. Huczko, M. Bystrzejewski, M. Roe, C.Y. Lee, S. Acquah, H. Kroto, D.R.M. Walton, Carbon 43 (2005) 1944–1953. [34] V.G. Pol, M. Motiei, A. Gedanken, J. Calderon-Moreno, M. Yoshimura, Carbon 42 (2004) 111–116. [35] W.Z. Qian, T. Liu, F. Wei, Z.W. Wang, Y.D. Li, Appl. Catal. A: Gen. 258 (2004) 121–124. [36] S. Amelinckx, X.B. Zhang, D. Bernaerts, X.F. Zhang, V. Ivanov, J.B. Nagy, Science 265 (1994) 635–639. [37] C.P. Deck, K. Vecchio, Carbon 44 (2006) 267–275.
Contents lists available at SciVerse ScienceDirect
Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso
Synthesis, characterization and growth mechanism of mesoporous hollow carbon nanospheres by catalytic carbonization of polystyrene Jiang Gong a,b, Jie Liu a, Xuecheng Chen a, Xin Wen a, Zhiwei Jiang a, Ewa Mijowska c, Yanhui Wang a, Tao Tang a,⇑ a b c
State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, China University of Chinese Academy of Sciences, Beijing 100049, China Institute of Chemical and Environment Engineering, West Pomeranian University of Technology, Szczecin ul. Pulaskiego 10, 70-322 Szczecin, Poland
a r t i c l e
i n f o
Article history: Received 15 January 2013 Received in revised form 22 March 2013 Accepted 27 March 2013 Available online 6 April 2013 Keywords: Mesoporous hollow carbon nanospheres Polystyrene Organically-modified montmorillonite Cobalt catalyst Growth mechanism
a b s t r a c t Mesoporous hollow carbon nanospheres (HCNs) were synthesized through the carbonization of polystyrene (PS) under the combined catalysis of organically-modified montmorillonite (OMMT)/cobalt catalyst at 700 °C. The morphology, microstructure, phase structure, textural property and surface composition of the obtained mesoporous HCNs were investigated by field-emission scanning electron microscope, transmission electron microscope (TEM), high-resolution TEM, X-ray diffraction, Raman spectroscopy, N2 sorption and X-ray photoelectron spectroscopy. It was found that OMMT not only promoted the dispersion of cobalt catalyst in the PS matrix but also affected the degradation of PS into light hydrocarbons and aromatics. The lattice oxygen of the cobalt catalyst facilitated the decomposition of light hydrocarbons and aromatics into atomic carbon during the formation of the mesoporous HCNs. A possible mechanism was proposed to explain the growth of mesoporous HCNs through the carbonization of PS under the combined catalysis of OMMT/cobalt catalyst. More importantly, this approach also offers a new potential way to transform waste polymer materials into mesoporous HCNs, which may be used as catalyst supports, adsorbents, storage media and templates for the synthesis of other useful hollow materials. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Hollow carbon spheres (HCSs) with a structure of hollow core and carbon shell have attracted great attention, due to their unique physicochemical properties such as low density, large inner space and specific surface area, and wide applications in catalyst supports, adsorbents, storage media and templates for the synthesis of other useful hollow materials [1–9]. The common methods for the preparation of HCSs are chemical vapor deposition [10,11], hydrothermal treatment [7] and template-assisted method [4,7,12–16] using ethylene [10], acetonitrile [11], sucrose [11], dopamine [14], acetone [17] or furfuryl alcohol [18] etc. as carbon source. From a sustainable point of view, reutilization of waste polymer materials to synthesize carbon spheres not only shows advantages with the nontoxic, cheap and abundant sources, but also provides a potential way to recycle waste polymer materials. Recently, some studies have demonstrated that virgin or waste polymer materials including polyethylene (PE), polypropylene (PP), polystyrene (PS) ⇑ Corresponding author. Tel.: +86 (0) 431 85262004; fax: +86 (0) 431 85262827. E-mail address: [email protected] (T. Tang). 1387-1811/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.micromeso.2013.03.039
and poly(ethylene terephthalate) (PET) can be converted into high value-added solid carbon microspheres (SCMs) [19–23]. For example, Pol et al. converted waste PE, PS and PET into SCMs using autoclave as reactor at 700 °C [19,20]. Qian et al. converted PP into SCMs using autoclave as reactor at 700 °C [21]. Chen et al. prepared SCMs by pyrolyzing waste PET in supercritical carbon dioxide at 500–650 °C [22]. However, to the best of our knowledge, so far, there has been only one report about synthesizing hollow carbon nanospheres (HCNs) using polymer material as carbon source. Chen et al. prepared HCNs using PP as carbon source and the combined organically-modified montmorillonite (OMMT)/Co(Ac)2 as catalyst at 900 °C [23]. But the HCNs showed extreme heterogeneity in the size distribution, and they did not provide information about the textural property of HCNs, nor further study the growth mechanism of HCNs. Hence, it is of great desire to synthesize HCNs with a narrow diameter distribution and further investigate the growth mechanism of HCNs using polymer as carbon source. Herein, based on our ‘‘combined catalysis’’ strategy to prepare carbon nanotubes with micro/mesoporous structure [24–29], mesoporous HCNs were synthesized by catalytic carbonization of PS under the combined catalysis of OMMT and cobalt catalyst.
32
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 1. Photographs of PS pellets (a, middle) and the HCN-1 (a, right) synthesized by pyrolyzing PS/10OMMT-10Co2O3 at 700 °C and typical SEM images of the HCN-1 (b–d) at various magnifications.
Fig. 2. Typical TEM (a and b) and HRTEM (c and d) images of the HCN-1.
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Firstly, the morphology, microstructure, phase structure, textural property and surface composition of the mesoporous HCNs were investigated. Subsequently, the effects of OMMT on the dispersion of cobalt catalyst in the PS matrix and the degradation products of PS were studied. Finally, a possible mechanism was put forward to explain the growth of mesoporous HCNs using PS as carbon source. This approach offers a new potential way to transform waste polymer materials into valuable mesoporous HCNs, which may be used as catalyst supports, adsorbents, storage media and templates for the synthesis of other useful hollow materials.
33
2.2. Preparation of samples PS pellets (40.00 g) were mixed with designed amounts of OMMT and cobalt catalyst in a Brabender mixer at 100 rpm and 180 °C for 10 min. The resultant sample was designed as PS/ xOMMT-yCat, where x and y represented the contents of OMMT and cobalt catalyst, respectively, and Cat represented the cobalt catalyst. For comparison, PS/10OMMT and PS/10Cat mixtures were also prepared.
2.3. Preparation of mesoporous HCNs 2. Experimental part 2.1. Materials PS pellets were supplied by Zhenjiang Qimei Chemical Co., Ltd., China. Four common cobalt catalysts: Co2O3, Co3O4, Co(Ac)24H2O and metallic cobalt (Co) powder, purchased from commercial vendors (Sinopharm Chemical Reagents Co., Ltd., China and Alfa Aesar), were of analytical-grade quality and used without further purification. OMMT (Closite 15A, organic modifier: dimethyl-dihydrogenated tallow quarternary ammonium; modifier concentration: 125 mequiv per 100 g clay) was purchased from Southern Clay. All other chemicals were of analytical-grade quality.
Mesoporous HCNs were prepared according to our previous method ‘‘combustion experiment’’ [27]. Briefly, a piece of sample (about 5.0 g) was placed into a crucible, which was heated at 700 °C until the flame from the crucible upper brim went out (after about 5 min). Subsequently, the charred residue in the crucible was purified with hydrofluoric acid and nitric acid to eliminate montmorillonite (MMT), amorphous carbon and most of cobalt catalysts. The yield of mesoporous HCNs was calculated by dividing the amount of purified carbon products by that of carbon element in the PS from the sample. All of the samples were weighed by analytical balance. The mesoporous HCNs from PS/ 10OMMT-10Co2O3, PS/10OMMT-10Co3O4 and PS/10OMMT10Co(Ac)2 were denoted as HCN-1, HCN-2 and HCN-3, respectively.
Fig. 3. Typical TEM images of the HCN-2 from PS/10OMMT-10Co3O4 after purification (a and b) and the HCN-3 from PS/10OMMT-10Co(Ac)2 (c and d) after purification at 700 °C.
34
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 4. XRD patterns (a), Raman spectra (b), nitrogen adsorption–desorption isotherms (c) and pore size distributions (d) of the resultant HCNs.
Table 1 Textural parameters of the obtained HCNs.
a b c d e
Property
HCN-1
HCN-2
HCN-3
SBET (m2/g) a Smeso (m2/g) b Vtotal (cm3/g) c Vmeso (cm3/g) d DAV (nm) e
111.6 88.7 0.490 0.479 3.77
149.8 149.8 0.509 0.509 3.80
227.3 227.3 1.020 1.020 3.77
The The The The The
total specific surface area. specific surface area of mesopores. total volume. volume of mesopores. average diameter of pores.
The degradation products of PS are the carbon feedstock for the formation of the mesoporous HCNs. To study the effect of OMMT on the degradation products of PS, pyrolysis experiments [27] for PS and PS/10OMMT mixture were conducted at 700 °C. The liquid pyrolyzed products were collected using a cold trap and the gas pyrolyzed products were collected using a sample bag. 2.4. Characterization The morphology of the mesoporous HCNs was observed by means of field-emission scanning electron microscope (SEM, XL30ESEM-FEG). The microstructure of the mesoporous HCNs was investigated using transmission electron microscope (TEM, JEM-1011) at an accelerating voltage of 100 kV and high-resolution
TEM (HRTEM) on a FEI Tecnai G2 S-Twin transmission electron microscope operating at 200 kV. The phase structure of the mesoporous HCNs was analyzed by X-ray diffraction (XRD) using a D8 advance X-ray diffractometer with Cu Ka radiation operating at 40 kV and 200 mA. Raman spectroscopy (T6400, excitationbeam wavelength: 514.5 nm) was used to characterize the vibrational properties of the mesoporous HCNs. The textural properties of the mesoporous HCNs were measured by N2 sorption at 77 K using a Quantachrome Autosorb-1C-MS analyzer. The specific surface area was calculated by BET method, and the contribution of micropores to both volume and surface area was evaluated by means of the t-plot method. The surface element composition of the mesoporous HCNs was characterized by means of X-ray photoelectron spectroscopy (XPS) carried out on a VG ESCALAB MK II spectrometer using an Al Ka exciting radiation from an X-ray source operated at 10.0 kV and 10 mA. Thermal gravimetric analysis (TGA) was performed using TA Instruments SDT Q600 at a heating rate of 10 °C/min. The rheological properties of PS and its mixture were conducted on a controlled strain rate rheometer (ARES rheometer) under nitrogen atmosphere. Round samples 25 mm (diameter) 1 mm (thickness) were run at 180 °C. Frequency sweep was performed from 0.01 to 100 rad/s, with a strain of 1% in order to make the materials be in linear viscoelastic response. Ultrathin sections of PS and its mixture were cryogenically cut using a Leica Ultracut and a glass knife at room temperature. The samples were collected on carbon-coated copper TEM grids. To study the effect of OMMT on the degradation products of PS, the collected liquid products were weighed and analyzed by gas chromatography–mass spectrometry (GC–MS, AGILENT
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
35
Fig. 5. XPS patterns of the mesoporous HCNs (a) and C1s high-resolution XPS spectra of HCN-1 (b), HCN-2 (c) and HCN-3 (d).
5975MSD). The volume of collected gas products was determined by the displacement of water. The hydrocarbon gas products were analyzed by a GC (Kechuang, GC 9800) equipped with a FID, using a KB-Al2O3/Na2SO4 column (50 m 0.53 mm ID). H2, CO and CH4 were analyzed by a GC (Kechuang, GC 9800) equipped with a TCD, using a packed TDX-01 (1 m) and molecular sieve 5A column (1.5 m). 3. Results and discussion 3.1. Morphology and microstructure of the resultant HCNs It is found from Fig. 1a that the HCN-1 from the carbonization of PS/10OMMT-10Co2O3 mixture at 700 °C can be produced on a gram scale by using this approach. The morphology observation of the HCN-1 by SEM at various magnifications showed that it possessed spherical-shape structure (Fig. 1b–d). In order to gain more detailed information about the internal microstructure of the HCN-1, TEM and HRTEM observations were conducted. Fig. 2a and b clearly showed that the HCN-1 was comprised of carbon nanospheres with central hollow inside. The main diameter of the HCN-1 was in the range of 60–90 nm with an average diameter of 73.5 nm. HRTEM images of the HCN-1 are shown in Fig. 2c and d, further revealing the microstructure of hollow carbon nanospheres. The carbon shell had an ordered and curved graphitic structure. The discontinuous graphitic layers and the numerous defects in the shells of the HCN-1 are not only propitious to the
Li-ion insertion and extraction but also to the diffusion of Li-ions, so it is expected that the obtained HCN-1 may be used in the lithium electrochemical cells. Furthermore, the interlayer spacing between graphitic layers was 0.33–0.35 nm, consistent with the ideal graphitic interlayer spacing. The thicknesses of the graphitic shells fell in the range of 6–12 nm. The yield of the HCN-1 was 11.1 wt.%. Other two common cobalt catalysts including Co3O4 and Co(Ac)2 were also used as catalysts for synthesizing HCNs. Fig. 3 presents the typical TEM images of the HCN-2 from PS/10OMMT10Co3O4 and the HCN-3 from PS/10OMMT-10Co(Ac)2 at 700 °C. Similar to the HCN-1, both HCN-2 and HCN-3 were hollow carbon nanospheres. In the case of the HCN-2 (Fig. 3a and b), the main nanosphere diameters and the graphitic shell thicknesses were in the range of 60–85 nm and 5–12 nm, respectively, which were close to those of the HCN-1. The HCN-3 showed smaller diameters (20–40 nm) and thinner graphitic shells (3–8 nm) (Fig. 3c and d). The yield of the HCN-2 and HCN-3 was 9.8 and 4.8 wt.%, respectively. The reason for the low yield of HCNs was analyzed in Section 3.3.3 and then a simple way was also proposed to increase the yield of HCNs. 3.2. Phase structure, textural property and surface composition of the resultant HCNs The graphitization degree of the resultant HCNs were characterized by XRD and Raman spectroscopy. Fig. 4a displays the XRD patterns of the resultant HCNs. The two diffraction peaks at about
36
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 6. Typical TEM images of PS/10Co2O3 mixture (a) and PS/10OMMT-10Co2O3 mixture (b). Rheological curves of PS and its mixture: complex viscosity (c) and storage modulus (d).
Table 2 Yields of the different fractions through the pyrolysis of PS and PS/10OMMT at 700 °C. Product
PS
PS/10OMMT
Carbon (g/100 g PS) Liquid (g/100 g PS) Gas (g/100 g PS) a Gas (g/100 g PS) b Gas (mL/100 g PS) c
0 99.873 0.127 0.084 1.848
1.015 94.595 4.390 3.887 59.145
a
Calculated by the material balance. Calculated by the volume of the gas (Supplementary Table S1) first divided by 22.4 L/mol, then multiply the molar number of the gas by its molar mass, finally obtained the yield of the total gases by adding the mass of each gas together. c Calculated by the volume of the gas divided by the mass of PS, finally obtained the yield of the total gases by adding the volume of each gas together. b
26.0° and 43.2° were assigned to the typical graphitic (0 0 2) and (1 0 1) planes, respectively. The broadening of the two peaks suggested the low graphitization degree. Compared to the HCN-2 and HCN-3, the relative sharp and strong graphite diffraction peaks of the HCN-1 indicated the relative higher degree of graphitization among these HCNs [11]. The weak diffraction peaks of metallic cobalt at 44.4°, 51.6° and 75.9° were also observed, implying that a
trace amount of metallic cobalt catalysts remained in the resultant HCNs. Fig. 4b shows the Raman spectra of the resultant HCNs. The peak at about 1580 cm 1 (G band) corresponds to an E2g mode of hexagonal graphite and is related to the vibration of sp2-bonded carbon atoms in a graphite layer, and the D band at about 1350 cm 1 is associated with the vibration of carbon atoms with dangling bonds in the plane terminations of disordered graphite [30]. The intensity of the G and D bands provides information about the crystallinity of the HCNs. Larger IG/ID ratio indicates a higher degree of structural ordering for the HCNs. Compared to the HCN-2 and HCN-3, the relative larger IG/ID ratio for the HCN-1 reflected the relative high degree of graphitization and the presence of a relative low amount of disordered carbon in the HCN-1, consistent with the XRD results (Fig. 4a). The nitrogen adsorption–desorption isotherms of the obtained HCNs (Fig. 4c) showed the type-IV curve and exhibited a hysteresis loop associated to capillary condensation in the range of P/P0 being from 0.5 to 1.0. This means that the porosity of the obtained HCNs is essentially made up of mesopores. The BET surface area (SBET), mesopore surface area (Smeso), total pore volume (Vtotal), mesopore volume (Smeso) and average pore diameter (DAV) of the HCNs are summarized in Table 1. Interestingly, the values of SBET, Smeso, Vtotal
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
37
and Vmeso increased in the following sequence: HCN-1 < HCN2 < HCN-3, possibly due to the decrease of the diameter of HCNs. The pore-size distributions of the HCNs were calculated using the Barrett–Joyner–Halenda (BJH) model from the desorption branches of the isotherms (Fig. 4d). This clearly showed that the size of the mesopores in the HCNs was in the narrow range of 2–5 nm (centered on 3.8 nm). The mesopores could be attributed to the cavities in the HCNs. The formation of the mesoporous HCNs and their mesopores was discussed in Section 3.3.3. XPS measurements were used to characterize the surface element composition of the mesoporous HCNs. It revealed that the surface of mesoporous HCNs mainly consisted of C (about 82 at.%) and O (about 18 at.%), with no evidence of any other elements (Fig. 5a). To determine the chemical components and the oxidation states of C element, high-resolution XPS spectra of C1s were curve-fitted into three individual peaks: graphitic carbon (284.6 eV), C-OH (285.6–285.8 eV) and COOH (288.9 eV) (Fig. 5b– d) [29]. These results showed that the carbon on the surface of the mesoporous HCNs existed in the presence of graphitic carbon with relatively small amounts of C–OH and COOH. These surface functional groups could contribute to the removal of heavy metallic ions [31] or organic dyes [32] when using the mesoporous HCNs as adsorbents in wastewater treatment. 3.3. Discussion about the growth mechanism of the mesoporous HCNs 3.3.1. OMMT promoting the dispersion of cobalt catalyst in the PS matrix To elucidate the roles of OMMT and cobalt catalyst in the formation of the mesoporous HCNs, firstly, PS/10Co2O3 and PS/ 10OMMT mixtures were pyrolyzed under the same condition. The yield of carbon product from PS/10Co2O3 and PS/10OMMT was 1.2 and 1.1 wt.%, respectively, much lower than that from PS/10OMMT-10Co2O3 (11.1 wt.%), suggesting that the combined OMMT/cobalt catalyst promoted the carbonization of PS. Furthermore, the carbon product from PS/10OMMT was amorphous carbon (not shown), demonstrating that OMMT was not the real active site for the formation of mesoporous HCNs. The pyrolysis of PS/10Co2O3 also led to the formation of hollow carbon nanospheres (Fig. S1). This indicated that the cobalt catalysts were the real active sites for the growth of mesoporous HCNs. However, the diameter of hollow carbon nanospheres from PS/10Co2O3 was in the range of 100–250 nm, obviously larger than that from PS/ 10OMMT-10Co2O3 (HCN-1, 60–90 nm). This was possibly because that OMMT promoted the dispersion of Co2O3 catalysts into smaller size in the PS matrix. To confirm the above speculation, the dispersion states of Co2O3 catalysts in the PS matrix with or without the addition of OMMT are shown in Fig. 6. It could be observed that without the addition of OMMT, a large amount of Co2O3 catalysts were agglomerated and the mean size was 379 nm (Fig. 6a). Interestingly, after the addition of OMMT, the dispersion degree of Co2O3 catalysts was significantly higher than that without OMMT, especially, the size of Co2O3 catalysts was evidently smaller (180 nm) than that without OMMT (Fig. 6b). There are two possible reasons for the above phenomenon. One of them results from the viscosity increase of the PS matrix due to the presence of dispersed OMMT (Fig. 6c and d), which increases shear force during melt mixing and improves the dispersion degree of the cobalt catalysts. The other is possible interaction between the OMMT and cobalt catalysts, which can prevent the re-aggregation of the cobalt catalysts. 3.3.2. OMMT affecting the degradation products of PS The degradation products of PS are carbon sources for the formation of mesoporous HCNs. The addition of OMMT may influence the composition of the degradation products from PS. In order
Fig. 7. GC–MS profiles of liquid degradation products from PS and PS/10OMMT at 700 °C.
Fig. 8. XRD patterns of PS/10Co2O3 mixture (a) and its carbonized product at 700 °C before purification (b), and PS/10OMMT-10Co2O3 mixture (c) and its carbonized product at 700 °C before purification (d).
to study the influence of OMMT on the degradation products of PS, GC and GC–MS measurements were conducted to analyze the composition of the degradation products from PS and PS/10OMMT mixture at 700 °C. Table 2 presents the material balance of the degradation products from PS and PS/10OMMT mixture. It was found that after adding OMMT, the quantity of gas degradation products significantly increased from 0.127 to 4.390 (g/100 g PS), indicating that OMMT promoted the degradation of PS into light hydrocarbons. Table S1 shows the composition of gas degradation products from these mixtures in detail. The gas degradation products mainly consisted of hydrogen, methane, ethane, ethylene, propane, propylene and i-butene. Compared to pure PS, the yields of methane,
38
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
Fig. 9. Possible mechanism about the formation of the mesoporous HCNs through the carbonization of PS under the combined catalysis of OMMT/cobalt catalyst at 700 °C.
ethylene and propylene increased significantly. For example, the yield of methane increased from 0.644 to 24.970 (mL/100 g PS). Fig. 7 displays the GC–MS profiles of liquid degradation products from PS and PS/10OMMT. The components of the liquid degradation products were quantitatively analyzed and the results were listed in Table S2. The main liquid degradation products from pure PS were toluene (3.118% area), styrene (65.381% area) and 2phenyl-1H-indene (14.893% area). Interestingly, after adding 10 wt.% OMMT, the content of styrene significantly decreased to 20.366 (%area), meanwhile other aromatic compounds (such as toluene, xylene and naphthalene) increased obviously. The above results demonstrated that OMMT promoted the degradation of
PS into light hydrocarbons and aromatics, which are widely used as carbon sources for the synthesis of carbon nanospheres [10,33,34]. 3.3.3. Real process for the formation of the mesoporous HCNs It is necessary to clear whether metallic cobalt catalysts are real active sites for understanding the growth mechanism of the mesoporous HCNs. To confirm the role of metallic cobalt catalysts during the growth of the mesoporous HCNs, XRD was used to monitor states of cobalt catalysts before and after the growth of mesoporous HCNs (Fig. 8). Compared to the XRD patterns of PS/ 10Co2O3 and PS/10OMMT-10Co2O3 mixtures, the diffraction peaks
Fig. 10. TEM (a and b) and HRTEM (c and d) images of Co@C-1 core/shell composite from PS/10OMMT-10Co2O3 at 700 °C before purification.
39
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
of PS (2h = 20.4°) and Co2O3 (2h = 19.5°, 31.7°, 37.3°, 39.0°, 42.8°, 45.2°, 56.1°, 59.8°, 62.0° and 65.6°) disappeared completely after the carbonization of PS, meanwhile the diffraction peaks of metallic cobalt (2h = 44.3°, 51.5° and 76.0°) appeared obviously. This indicated that Co2O3 catalysts were reduced into metallic cobalt catalysts, which were the real active sites for the formation of the mesoporous HCNs. Similar results were found using Co3O4 or Co(Ac)2 as catalyst with or without OMMT (not shown). To get more information about the formation process of the mesoporous HCNs, metallic cobalt powder (Co) in place of Co2O3 was added into PS/10OMMT mixture and the resultant PS/ 10OMMT-10Co mixture was also pyrolyzed under the same condition. The yield of carbon product from PS/10OMMT-10Co was 12.8 wt.%, similar to that from PS/10OMMT-10Co2O3, suggesting that the combined OMMT/metallic cobalt powder also promoted the carbonization of PS. However, the carbon product from PS/ 10OMMT-10Co was actually irregular carbon rather than mesoporous HCNs (Fig. S2). This result suggested that the reduction of cobalt oxide catalysts to metallic cobalt catalysts was an essential process for the formation of the mesoporous HCNs. Qian et al. reported that the unreduced Ni/Cu/Al2O3 catalyst showed much higher efficiency on the decomposition of methane than reduced Ni/Cu/Al2O3 catalyst, and they believed that the lattice oxygen on the unreduced catalyst accelerated the decomposition of methane into carbon [35]. Similarly in this work, the lattice oxygen on the cobalt oxide may promote the decomposition of light hydrocarbons and aromatics into atomic carbon. Hence, we speculated that
Fig. 11. Nitrogen adsorption–desorption isotherms (a) and pore size distributions (b) of the resultant Co@C core/shell composites. Co@C-2 and Co@C-3 were from PS/ 10OMMT-10Co3O4 and PS/10OMMT-10Co(Ac)2, respectively.
Table 3 Textural parameters of the resultant Co@C core/shell composites. Property 2
a
SBET (m /g) Smeso (m2/g) b Vtotal (cm3/g) c Vmeso (cm3/g) d DAV (nm)e a b c d e
The The The The The
Co@C-1
Co@C-2
Co@C-3
68.5 59.0 0.227 0.222 3.76
123.5 106.7 0.336 0.330 3.77
137.1 119.5 0.368 0.359 3.75
total specific surface area. specific surface area of mesopores. total volume. volume of mesopores. average diameter of pores.
the formation process of the mesoporous HCNs through the carbonization of PS included the reduction process of cobalt oxide catalysts into metallic cobalt catalysts and the simultaneous carbonization process of the degradation products from PS (such as methane and benzene) into carbon shells on the surface of in situ formed metallic cobalt catalysts. Based on the above results, a possible mechanism about the formation of the mesoporous HCNs through the catalytic carbonization of PS was put forward (Fig. 9), similar to the generally accepted dissociation–diffusion–precipitation mechanism for the growth of carbon nanotubes [36]. Firstly, OMMT promoted the dispersion of cobalt oxide catalysts in the PS matrix (step 1); subsequently, OMMT also promoted the degradation of PS into light hydrocarbons and aromatics (step 2). Secondly, the resultant light hydrocarbons and aromatics were dehydrogenated and aromatizated on the surface of cobalt oxide catalysts, meanwhile the cobalt oxide catalysts (such as Co2O3) were reduced into metallic cobalt catalysts (step 3), which have some (about 1 at.%) carbon solubility in the solid solution [37]. After further carbonization of carbon sources, once supersaturated, carbon precipitated from the surface of metallic cobalt nanoparticles (step 4) to form Co@C core/shell composites (step 5, Fig. 10 and Fig. S3), which were essentially made up of mesopores according to the nitrogen adsorption–desorption isotherms and the pore-size distributions (Fig. 11 and Table 3). Obviously, the formation of Co@C core/shell composites played an important role in the synthesis of mesoporous HCNs and the metallic cobalt nanoparticles from the reduction of cobalt oxide catalysts actually acted as templates for the growth of HCNs. In the case of cobalt catalyst such as Co(Ac)2, cobalt catalyst was firstly in situ decomposed into smaller cobalt oxide nanoparticles such as Co2O3 or Co3O4 (Fig. S4). This is the possible reason why the average diameter of HCN-3 from PS/10OMMT10Co(Ac)2 is much smaller than that of HCN-1 from PS/10OMMT10Co2O3 or HCN-2 from PS/10OMMT-10Co3O4. In addition, during the formation of Co@C core/shell composites, the lattice oxygen on the cobalt oxide catalysts promoted the decomposition of light hydrocarbons and aromatics into atomic carbon. Subsequently, the yielded carbon dissolved into the metallic cobalt nanoparticles. This is the possible reason why the carbon product using metallic cobalt powder directly as catalyst was irregular carbon rather than mesoporous HCNs (Fig. S2). Finally, after the removals of metallic cobalt catalysts and MMT, the mesoporous HCNs were obtained (step 6). Owing to the short synthesis time (about 5 min, see Section 2.3) and the low content of cobalt oxide (10 wt.%), the carbon shell of Co@C core/shell composites could not grow very thick (Fig. 10 and Fig. S3). Hence, the yield of mesoporous HCNs was limited. But the yield can be increased by increasing the content of cobalt oxide. For example, when the content of Co2O3 increased from 10 to 30 wt.% into PS/10OMMT, the yield of mesoporous HCNs dramatically increased from 11.1 to 35.0 wt.%. Besides, further studies are needed to evaluate the ef-
40
J. Gong et al. / Microporous and Mesoporous Materials 176 (2013) 31–40
fect of additives in the waste PS on the formation of mesoporous HCNs and more investigation will be conducted in the further work. According to the above mechanism, the formation of mesopores was promoted by water etching and HNO3 oxidation. Water was in situ yielded as byproduct during cobalt oxide catalyzing carbonization of PS degradation products. At high temperature (such as 700 °C), water could efficiently etch the carbon shell of Co@C core/shell composites and favor the formation of mesopores. Similar results were reported by Pol et al. [34] and they believed that water could help the formation of pores in carbon microspheres. In addition, during the purification of Co@C core/shell composites, HNO3 permeated into internal structure of HCNs after the removal of cobalt core and further etched the inner surface carbon to form more mesopores. Han et al. [3] also found that mesopores were generated during the removal of the residual Zn and ZnO dispersed in the HCNs or intercalated into the graphitic shells by acid purification. 4. Conclusion We reported a simple approach to synthesize mesoporous HCNs through the carbonization of PS by the combined catalysis of OMMT/cobalt catalyst at 700 °C. The diameter of the mesoporous HCNs was in the range of 60–90 nm using OMMT/Co2O3, 60–85 nm using OMMT/Co3O4, or 20–40 nm using OMMT/Co(Ac)2. The total surface area and pore volume of the mesoporous HCNs were 111.6–227.3 m2/g and 0.49–1.02 cm3/g, respectively. The size of mesopores in the HCNs was in the narrow range of 2–5 nm. OMMT not only promoted the dispersion of cobalt catalyst in the PS matrix but also affected the degradation of PS into light hydrocarbons and aromatics. The lattice oxygen on the cobalt oxide catalysts promoted the decomposition of light hydrocarbons and aromatics into atomic carbon during the formation of the mesoporous HCNs. At last, a possible growth mechanism of the mesoporous HCNs using PS as carbon source and combined OMMT/cobalt catalyst as catalyst was proposed. This mechanism will be favorable to understand the growth of the mesoporous HCNs using polymer as carbon source. More importantly, this approach offers a new potential way to transform waste polymers into mesoporous HCNs, which may be used as catalyst supports, adsorbents, storage media and templates for synthesis of other useful hollow materials. Also, further studies are needed to evaluate the effect of additives in the waste polymers on the formation of mesoporous HCNs. Acknowledgments This work was supported by the National Natural Science Foundation of China (2124079, 50873099 and 20804045) and Polish Foundation (No. 2011/03/D/ST5/06119). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.micromeso.2013. 03.039.
References [1] A. Nieto-Márquez, R. Romero, A. Romero, J. Valverde, J. Mater. Chem. 21 (2011) 1664–1672. [2] A.A. Deshmukh, S.D. Mhlanga, N.J. Coville, Mater. Sci. Eng. R 70 (2010) 1–28. [3] F.D. Han, Y.J. Bai, R. Liu, B. Yao, Y.X. Qi, N. Lun, J.X. Zhang, Adv. Energy Mater. 1 (2011) 798–801. [4] S.B. Yang, X.L. Feng, L.J. Zhi, Q. Cao, J. Maier, K. Müllen, Adv. Mater. 22 (2010) 838–842. [5] A.H. Lu, G.P. Hao, Q. Sun, X.Q. Zhang, W.C. Li, Macromol. Chem. Phys. 213 (2012) 1107–1131. [6] G.H. Sun, K.X. Li, L.J. Xie, J.L. Wang, Y.Q. Li, Micropor. Mesopor. Mat. 151 (2012) 282–286. [7] K. Tang, L.J. Fu, R.J. White, L.H. Yu, M.-M. Titirici, M. Antonietti, J. Maier, Adv. Energy Mater. 2 (2012) 873–877. [8] J.H. Kim, B.Z. Fang, S.B. Yoon, J.-S. Yu, Appl. Catal. B: Environ. 88 (2009) 368– 375. [9] B.Z. Fang, M. Kim, J.H. Kim, J.-S. Yu, Langmuir 24 (2008) 12068–12072. [10] X.C. Chen, K. Kierzek, Z.W. Jiang, H.M. Chen, T. Tang, M. Wojtoniszak, R.J. Kalenczuk, P.K. Chu, E. Borowiak-Palen, J. Phys. Chem. C 115 (2011) 17717– 17724. [11] Z.X. Yang, R. Mokaya, Micropor. Mesopor. Mat. 113 (2008) 378–384. [12] A.H. Lu, T. Sun, W.C. Li, Q. Sun, F. Han, D.H. Liu, Y. Guo, Angew. Chem. Int. Ed. 50 (2011) 11765–11768. [13] K.-M. Choi, K. Kuroda, Chem. Commun. 47 (2011) 10933–10935. [14] R. Liu, S.M. Mahurin, C. Li, R.R. Unocic, J.C. Idrobo, H.J. Gao, S.J. Pennycook, S. Dai, Angew. Chem. Int. Ed. 50 (2011) 6799–6802. [15] G. Sun, K.X. Li, J. Wang, H.W. He, J.L. Wang, Y.Q. Li, Y. Liu, J.Y. Gu, Micropor. Mesopor. Mat. 139 (2011) 207–210. [16] M. Kim, S.B. Yoon, K. Sohn, J.Y. Kim, C.-H. Shin, T. Hyeon, J.-S. Yu, Micropor. Mesopor. Mat. 63 (2003) 1–9. [17] G.D. Li, C.L. Guo, C.H. Sun, Z.C. Ju, L.S. Yang, L.Q. Xu, Y.T. Qian, J. Phys. Chem. C 112 (2008) 1896–1900. [18] Y.S. Li, Y.Q. Yang, J.L. Shi, M.L. Ruan, Micropor. Mesopor. Mat. 112 (2008) 597– 602. [19] V.G. Pol, M.M. Thackeray, Energy Environ. Sci. 4 (2011) 1904–1912. [20] V.G. Pol, Environ. Sci. Technol. 44 (2010) 4753–4759. [21] J.H. Zhang, J. Li, J. Cao, Y.T. Qian, Mater. Lett. 62 (2008) 1839–1842. [22] L.Z. Wei, N. Yan, Q.W. Chen, Environ. Sci. Technol. 45 (2011) 534–539. [23] X.C. Chen, H. Wang, J.H. He, Nanotechnology 19 (2008) 325607. [24] T. Tang, X.C. Chen, X.Y. Meng, H. Chen, Y.P. Ding, Angew. Chem. Int. Ed. 44 (2005) 1517–1520. [25] Z.W. Jiang, R.J. Song, W.G. Bi, J. Lu, T. Tang, Carbon 45 (2007) 449–458. [26] R.J. Song, Z.W. Jiang, W.G. Bi, W.X. Cheng, J. Lu, B.T. Huang, T. Tang, Chem. Eur. J. 13 (2007) 3234–3240. [27] J. Gong, J. Liu, L. Ma, X. Wen, X.C. Chen, D. Wan, H.O. Yu, Z.W. Jiang, E. Borowiak-Palen, T. Tang, Appl. Catal. B: Environ. 117–118 (2012) 185–193. [28] J. Gong, K. Yao, J. Liu, X. Wen, X.C. Chen, Z.W. Jiang, E. Mijowska, T. Tang, Chem. Eng. J. 215–216 (2013) 339–347. [29] J. Gong, J. Liu, D. Wan, X.C. Chen, X. Wen, E. Mijowska, Z.W. Jiang, Y.H. Wang, T. Tang, Appl. Catal. A: Gen. 449 (2012) 112–120. [30] M. Ignat, C.J. Van Oers, J. Vernimmen, M. Mertens, S. Potgieter-Vermaak, V. Meynen, E. Popovici, P. Cool, Carbon 48 (2010) 1609–1618. [31] R. Demir-Cakan, N. Baccile, M. Antonietti, M.-M. Titirici, Chem. Mater. 21 (2009) 484–490. [32] X.H. Song, Y.B. Wang, K.A. Wang, R. Xu, Ind. Eng. Chem. Res. 51 (2012) 13438– 13444. [33] Y.Z. Jin, C. Gao, W.K. Hsu, Y.Q. Zhu, A. Huczko, M. Bystrzejewski, M. Roe, C.Y. Lee, S. Acquah, H. Kroto, D.R.M. Walton, Carbon 43 (2005) 1944–1953. [34] V.G. Pol, M. Motiei, A. Gedanken, J. Calderon-Moreno, M. Yoshimura, Carbon 42 (2004) 111–116. [35] W.Z. Qian, T. Liu, F. Wei, Z.W. Wang, Y.D. Li, Appl. Catal. A: Gen. 258 (2004) 121–124. [36] S. Amelinckx, X.B. Zhang, D. Bernaerts, X.F. Zhang, V. Ivanov, J.B. Nagy, Science 265 (1994) 635–639. [37] C.P. Deck, K. Vecchio, Carbon 44 (2006) 267–275.