6114
J. Phys. Chem. C 2009, 113, 6114–6117
An X-ray Absorption, Photoemission, and Raman Study of the Interaction between SnO2 Nanoparticle and Carbon Nanotube J. G. Zhou,¶,†,§ H. T. Fang,¶,‡ J. M. Maley,| J. Y. P. Ko,† M. Murphy,† Y. Chu,‡ R. Sammynaiken,| and T. K. Sham*,† Department of Chemistry, The UniVersity of Western Ontario, London, Canada, School of Materials Science and Engineering, Harbin Institute of Technology, Harbin, People’s Republic of China, Canadian Light Source Inc., UniVersity of Saskatchewan, Saskatoon, Canada, and Saskatchewan Structural Sciences Centre, UniVersity of Saskatchewan, Saskatoon, Canada ReceiVed: NoVember 27, 2008; ReVised Manuscript ReceiVed: January 19, 2009
The interaction between SnO2 and substrate in SnO2 nanoparicles (NPs)-carbon nanotubes (CNTs) composite has been studied by X-ray absorption near-edge structures (XANES) at Sn M5,4, O K-edge and C K-edge. SnO2 NPs in the composite have a rutile crystal structure with abundant surface states. The variation in resonance features of the XANES strongly supports that the crystalline SnO2 NPs interact with CNTs through synergic bonding involving charge redistribution between C 2p-derived states and the valence and conduction bands in SnO2 NPs via interaction at the interface facilitated by oxidation treatment of the CNT prior to composite formation. Raman and ultraviolet photoelectron spectroscopy (UPS) results support the synergic bonding interaction. Such interaction is expected not only to immobilize SnO2 NPs on CNT but also to improve the conductivity of SnO2 NPs. Introduction Tin oxide (SnO2) nanoparticles (NPs) coated carbon nanotubes (CNTs) are useful functional nanocomposite in many applications including gas sensors,1 fuel cells,2 batteries,3 and supercapacitors.4 The special configuration in this nanocomposite is expected to prevent the SnO2 NPs from aggregation and to increase its conductivity, hence the performance. The interaction between SnO2 NP and CNT is crucial in immobilizing SnO2 NP. Interactions between particles and support in conventional solid catalyst support5 and Pt NPs-CNTs systems6,7 leading to desirable functionality have been recognized. Although information about the interaction in nanocomposite is important to the understanding of its functionality, the nature of the interactions remains poorly understood because of the complexity of composites and relatively limited analysis methods. X-ray absorption near-edge structures (XANES) measure the modulation of the absorption coefficient at a particular core level of an atom in a chemical environment, therefore it is element specific and also very sensitive to the local chemistry. XANES provides information about the symmetry and occupancy of the low-lying unoccupied electronic states specific to the absorbing atom (LUMO and LUMO+ in molecules) and has been successfully applied to investigate chemical bonding, electronic structure, and surface chemistry of many nanomaterials.8-11 For example, the C K-edge and Pt M3-edge XANES have been successfully applied to elucidate the synergic interaction between * To whom correspondence should be addressed. Fax: +1-519-661-3022. E-mail:
[email protected]. ¶ Contributed equally. † The University of Western Ontario. § Canadian Light Source Inc., University of Saskatchewan. ‡ Harbin Institute of Technology. | Saskatchewan Structural Sciences Centre, University of Saskatchewan.
Pt NPs and CNTs in a Pt NPs-CNTs composite system.6 In this work, we report a study of SnO2 NP-CNT interaction using XANES. Experimental Section In this study, CNTs before SnO2 NPs deposition were used as a reference and are henceforth denoted CNTs. SnO2 NPs coated CNTs were prepared by a simple SnCl2 solution method.12,13 Purified multiwalled CNTs with a diameter of 25-40 nm and a purity of 95% were purchased from Shenzhen Nanotech Port Co. Ltd. and treated by refluxing in nitric acid (40%) at 110 °C for 2 h. SnO2 NPs coating was performed in SnCl2 · 2H2O solution. Using a desired amount of HCl acid in the SnCl2 · 2H2O solution is the key in obtaining uniformly dispersed and well-separated SnO2 NPs coated on CNT (0.7 mL of 38% HCl in 40 mL of H2O) rather than SnO NP (no HCl in H2O) or continuous SnO2 thin layer coating (0.13 mL of HCl in 40 mL of H2O). The SnO2 NPs coated CNTs obtained from a solution with 0.7 mL of 38% HCl in 40 mL of H2O are henceforth denoted SnO2 NPs-CNTs. The NPs-CNTs composite thus obtained has been characterized by transmission electron microscopy (TEM) using a Philips CM20 EM microscope operated at 80 KeV. The details about the synthesis and conventional characterization can be found in a previous report, which showed that the SnO phase was absent in the SnO2 NPsCNTs by XRD.13 Free-standing SnO2 NPs were obtained by “burning” the SnO2 NPs-CNT composite in an oven at 650 °C. The products thus obtained are henceforth denoted SnO2 NPs. The XANES were obtained on the spherical grating monochromator (SGM) beamline (∆E/E: ∼10-4) at the Canadian Light Source (CLS) in a surface-sensitive, total electron yield (TEY) mode with use of specimen current. Data were normalized to the incident photon flux, which were collected with a refreshed gold mesh (evaporation of a fresh Au layer prior to the measurements). The spectra at the C K-edge and the O K-edge (including Sn M5,4-edge) are normalized to the edge
10.1021/jp810639y CCC: $40.75 2009 American Chemical Society Published on Web 03/20/2009
Interaction of SnO2 Nanoparticle and Carbon Nanotube
J. Phys. Chem. C, Vol. 113, No. 15, 2009 6115
Figure 2. (a) O K-edge and (b) M5,4-edge XANES of SnO2 NPs-CNTs composite and free-standing SnO2 NPs. Figure 1. (a) TEM image of SnO2 NPs coated CNTs; (b) TEM image of the free-standing SnO2 NPs after burning off the carbon in the SnO2 NPs-CNTs composite.
jump, the difference in absorption coefficient just below and at a flat region above the edge (300 and 570 eV for C and O, respectively). The Raman spectroscopy (Renishaw System 2000) was conducted under ambient conditions with a 514 nm laser. The ultraviolet photoelectron spectroscopy (UPS) measurements were performed at Surface Science Western (SSW) with a Kratos Axis Ultra spectrometer with He I (21.22 eV) light. Work functions were determined from the separation between the secondary electron cutoff and the Fermi level of the UPS spectra. Results and Discussions Figure 1a shows that the size of the SnO2 NPs coated on CNTs (tube diameter of 20-40 nm) is 2 to 4 nm in diameter and the coating is uniform. TEM image of SnO2 NPs in Figure 1b shows that SnO2 NPs have a diameter of ∼10 nm, indicating that the removal of CNT from SnO2 NPs-CNTs results in the congregation of SnO2 NPs. The O K-edge and Sn M5,4-edge XANES of the SnO2 NPsCNTs and SnO2 NPs are shown in Figure 2, panels a and b, respectively. The spectral features arise from dipole transitions (∆l ) (1).14 Thus, the O K-edge and Sn M5,4-edge involves s to p and d to p or f transitions, respectively. Since the transition maps the oxygen p (O K-edge) or Sn p and f projected (Sn M5,4-edge) electronic states in the conduction band, the area under these peaks (also known as whitelines), is proportional to the unoccupied partial densities of states, which compare favorably with band theory calculations.15 The resonances beyond the whiteline are transitions to the higher energy bands/ multiple scattering states, which are characteristic of the immediate vicinity of the absorbing atom in the crystal.14 The
absorption features in these nanomaterials are similar to those of microsized rutile SnO2 powders,15 in agreement with the previous XRD charaterization,13 confirming the rutile characteristic of SnO2 NPs. Pre-edge resonances, peak a, at Sn M5edge, and peak d, at Sn M4-edge, which overlap with a bulk feature in Figure 2b, reflect the existence of surface states.10,16 More intensive pre-edge peaks in SnO2 NPs-CNTs compared to SnO2 powder are due to size effect15 since SnO2 NP (2-4 nm) in the SnO2 NPs-CNTs composite are smaller than the freestanding SnO2 NPs (∼10 nm); the fraction of surface Sn atoms increases markedly in the former. A close inspection of Figure 2 reveals a reduction in intensity in the SnO2 NPs-CNTs XANES relative to that of the freestanding SnO2 NPs. The reduction is more apparent at the Sn M5,4-edge than at the O K-edge. This variation in intensity, hence unoccupied densities of states, suggests that charge redistribution between the NPs and the CNT substrate has occurredsthat is that the weaker the intensity, the lower the unoccupied densities of states. We also note that the weaker resonances at the Sn edges in SnO2-CNT are accompanied by a slight broadening, suggesting that the surface and the bulk atoms are in a slightly different environment. Charge redistribution has been seen in Pt NPs on porous silicon due to Pt NP-Si oxide interactions, which withdraw d charges from Pt17 and in Pt NPs supported on CNTs due to the synergic bonding.6 The reduced unoccupied density of states of the conduction band in SnO2 NPs-CNTs inferred from the resonance intensity suggests charge transfer from CNT to SnO2 NP via interaction at the SnO2-CNT interface. Another interesting observation in the Sn M5,4-edge XANES is that the positions of peak b to peak f shift to higher energy in SnO2 NPs-CNTs. We propose that the observed peak shifts arise from quantum confinement (opening up of the band gap). Figure 3 shows the C K-edge XANES in TEY for CNTs and SnO2 NPs-CNTs. The photon energy is calibrated to the C 1s
6116 J. Phys. Chem. C, Vol. 113, No. 15, 2009
Zhou et al.
Figure 3. C K-edge XANES of CNTs and the SnO2 NPs-CNTs composite.
to π* transition of CNTs at 285 eV. Two main peaks are clearly displayed at ∼285 and ∼291 eV for both samples. These can be attributed to the transitions from C 1s to unoccupied states of C-C π* and C-C σ* characters,18 respectively (XANES from fluorescence yield, not shown here, exhibit similar but much broader features due to the thickness effect). A small peak is observed at ∼288 eV that can be attributed to the CdO group (e.g., carboxylates) resulting from the oxidation.9 A weak preedge transition at 283 eV is also observed in CNTs, which is associated with amorphous carbon.19,20 The existence of C-O bonds and amorphous carbon in CNTs is expected since the CNTs have been oxidized in HNO3 acid13 to produce surface defects as the nucleation sites to anchor SnO2 NPs. The 290.5 eV feature can be attributed to SnO2 NPs-CNT interaction, which elongates the C-C bond of CNT at the interface, resulting in a shape resonance closer to the threshold. The existence of surface defects has also been confirmed by HRTEM and Raman spectroscopy as reported previously.13 In SnO2 NPs-CNTs, the C-C π* transition intensity is reduced and down shifted by ∼0.1 eV and this is accompanied by an increase in the intensity of the absorption at 283 eV. Also the peaks are slightly broader relative to that of CNTs. This observation indicates that in SnO2 NPs-CNTs, the defect abundant surface of the CNT interacts with SnO2 strongly, most likely via the formation of Sn-O or Sn-C bonds, shifting the surface contributions for both π* and σ* transitions to lower energy while the electronic structure of the bulk CNT remains unchanged.8 The similarity in the rest of the features in the C K-edge XANES for SnO2 NPs-CNTs and CNTs suggests that the NPs coating does not change the graphene structure inside CNT significantly albeit the interface interaction is considerable; therefore an improved electric conductivity in this nanocomposite is expected. Thus, the intensity increase at 283.5 and 290.5 eV (attributed to SnO2 NPs-C interaction) at the expense of the intensity decrease in π* and σ* transitions in SnO2 NPsCNTs is associated with strong interaction between SnO2 and CNTs at the interface. Since the C K-edge XANES investigates the unoccupied molecular orbital in CNTs and SnO2 NPs-CNTs with carbon p characters, the area under the resonance, pending no countervailing symmetry arguments, is proportional to the unoccupied densities of states (DOS). The reduction of unoccupied DOS of carbon p character in SnO2 NPs-CNTs relative to that of CNTs and the accompanying intensity increase of the defect peak in Figure 3 provide spectroscopic evidence that charge redistribution involving the SnO2 NPs, and C 2p-derived states has occurred. It is plausible that the conduction band of SnO2 accepts electron (decrease intensity at the Sn edge in Figure 2) via surface Sn from defect sites in CNTs (increase
Figure 4. Raman spectra of CNTs and the SnO2 NPs-CNTs composite.
intensity of the 283.5 eV feature in Figure 3) to form the σ bond and back-donates electron to the π* of the CNTs. This synergic bonding is common in the alloying of noble metals21 and metal carbonyl chemistry. Thus the chemical interaction will prevent lateral diffusion of SnO2 NPs, especially under practical operation conditions. It should be noted that the resonance at ∼283.5 eV in the XANES for SnO2 NPs can also originate from an increase in amorphous carbon resulting from SnO2 NPs interaction. We used Raman spectroscopy to track the amorphous carbon before and after SnO2 deposition since amorphous carbon exhibits characteristic Raman shift. The Raman spectra of CNTs and SnO2 NPs-CNTs are compared in Figure 4. It is apparent from Figure 4 that upon deposition, the D band (1345 cm-1) intensity characteristic of amorphous carbon is reduced relative to that of the G band (1571 cm-1).13 This supports that the intensity increase at 283.5 eV in NPs-CNTs is due to the SnO2 NPs-C interaction rather than the increase of amorphous carbon. The electronic structures of the valence band of SnO2 NPsCNTs, SnO2 NPs, and CNT were investigated with UPS (Figure 5). Relevantly, the work function of SnO2 NPs-CNTs, 4.68 eV, is similar to that of SnO2 NPs (4.72 eV), but significantly smaller than the value of 5.42 eV in CNTs.22,23 This result indicates that CNTs are coated and in good electric contact with SnO2 NPs. It is also interesting to note that the top of the valence band measured at the rising edge (see Figure 5) of the valence band is ∼3.5 eV below the Fermi level. This value is comparable to the band gap of ∼3.6 eV for a heavily n-doped SnO2 (top of the VB to Fermi level is 1.8 eV in an intrinsic semiconductor). Thus there may be some charging in the measurement but it would not be too significant since SnO2 is always n-type due to O vacancies (the Fermi level is much closer to the bottom of the conduction band than an intrinsic semiconductor). In addition, two interesting features are observed in the valence band: (a) there are tailing densities of states for CNT all the way to the Fermi level (zero binding energy) as expected while for both SnO2 NP and SnO2 NPs-CNT these tailing states begin to appear at ∼2.4 eV binding energy and can be attributed to surface states of SnO2 NP, hence the CNT contribution to the VB in SnO2 NP-CNTs is strongly suppressed by the surface sensitivity of UPS, and (b) the valence band for the smaller SnO2 NPs (∼2 to 4 nm) in SnO2 NPs-CNT is narrower than that of SnO2 NP (∼10 nm). Valence band narrowing is consistent with the reduction of coordination number on average due to increasing surface atom to bulk atom ratio, as the
Interaction of SnO2 Nanoparticle and Carbon Nanotube
J. Phys. Chem. C, Vol. 113, No. 15, 2009 6117 CRC (T.K.S.). Research at HIT was supported by the National Science Foundation of China (Grant Nos. 50472084 and 5060211) and Science Foundation of Heilongjiang Province (E200517). H.T.F. also acknowledges the support of a Visiting Fellowship at the Centre for Chemical Physics, The University of Western Ontario. References and Notes
Figure 5. UPS of CNTs, SnO2 NPs, and the SnO2 NPs-CNTs composite; the inset shows the top of the valence band (approximately the same for both SnO2 samples at ∼3.7 eV) marked with a vertical line; tailing DOS toward the Fermi level is also noticeable (see text).
nanoparticles become smaller.24 These results are consistent with the previously observed behavior of SnO2 nanostructures.16 Conclusions Using a number of spectroscopic techniques, we have investigated the interaction between SnO2 NP and preoxidized CNT. We find that SnO2 NP can indeed be immobilized on CNT surface via a strong SnO2 NP-CNT substrate synergic interaction via surface defects on the CNT. The above results and analyses indicate that XANES at the C-, O K-edges and Sn M5,4-edge is a powerful tool to elucidate the interactions between crystalline SnO2 NPs and CNTs in SnO2 NPs-CNTs composite. The advantage of element and site specific XANES in clarifying the interactions between different components in the complex nanocomposites is demonstrated. Acknowledgment. . We thank Dr. R. Blyth and T. Regier of the Canadian Light Source and Mark Biesinger of Surface Science Western for their technical assistance. CLS is supported by NSERC, NRC, CIHR, and the University of Saskatchewan. The work at the University of Western Ontario was supported by NSERC, CFI, OIT, and
(1) Liang, Y. X.; Chen, Y. J.; Wang, T. H. Appl. Phys. Lett. 2004, 85, 666–668. (2) Ke, K.; Waki, K. J. Electrochem. Soc. 2007, 154, A207-A212. (3) Xie, J.; Varadan, V. K. Mater. Chem. Phys. 2005, 91, 274–280. (4) Reddy, A. L. M.; Ramaprabhu, S. J. Phys. Chem. C 2007, 111, 16138–16146. (5) Baker, R. T. K.; Tauster, S. J.; Dumesic, J. A. Strong Metal Support Interactions;American Chemical Society: Washington, D.C., 1986. (6) Zhou, J. G.; Zhou, X. T.; Sun, X. H.; Li, R. Y.; Murphy, M.; Ding, Z. F.; Sun, X. L.; Sham, T. K. Chem. Phys. Lett. 2007, 437, 229–232. (7) Hull, R. V.; Li, L.; Xing, Y. C.; Chusuci, C. C. Chem. Mater. 2006, 18, 1780–1788. (8) Tang, Y. H.; Sham, T. K.; Hu, Y. F.; Lee, C. S.; Lee, S. T. Chem. Phys. Lett. 2002, 366, 636–641. (9) Kuznetsova, A.; Popova, I.; Yates, J. T.; Bronikowski, M.; Huffman, C.; Liu, J.; Smalley, R. E.; Hwu, H. H.; Chen, J. G. J. Am. Chem. Soc. 2001, 123, 10699–10704. (10) Zhou, X. T.; Heigl, F.; Murphy, M. W.; Sham, T. K.; Regier, T.; Coulthard, I.; Blyth, R. I. R. Appl. Phys. Lett. 2006, 89, 213109-1–2131093. (11) Zhou, J. G.; Zhou, X. T.; Sun, X. H.; Murphy, M.; Heigl, F.; Sham, T. K.; Ding, Z. F. Can. J. Chem. 2007, 85, 756–760. (12) Han, W. Q.; Zettl, A. Nano Lett. 2003, 3, 681–683. (13) Fang, H. T.; Sun, X.; Qian, L. H.; Wang, D. W.; Li, F.; Chu, Y.; Wang, F. P.; Cheng, H. M. J. Phys. Chem. C 2008, 112, 5790–5794. (14) Sham, T. K.; Naftel, S. J.; Coulthard, I. J. Appl. Phys. 1996, 79, 7134–7138. (15) Kucheyev, S. O.; Baumann, T. F.; Sterne, P. A.; Wang, Y. M.; Van Buuren, T.; Hamaza, A. V.; Terminello, L. J.; Willey, T. M. Phys. ReV. B 2005, 72, 035404-1–035404-5. (16) Zhou, X. T.; Zhou, J. G.; Murphy, M. W.; Ko, J. Y. P.; Heigl, F.; Regier, T.; Blyth, R. I. R.; Sham, T. K. J. Chem. Phys. 2008, 128, 1447031–144703-5. (17) Coulthard, I.; Sham, T. K. Solid State Commun. 1998, 105, 751– 754. (18) Banerjee, S.; Hemraj-Benny, T.; Balasubramanian, M.; Fischer, D. A.; Misewich, J. A.; Wong, S. S. Chem. Commun. 2004, 772–773. (19) Moror, J. F.; Himpsel, F. J.; Hollinger, G.; Jordon, J. L.; Hughes, G.; McFeely, F. R. Phys. ReV. B 1986, 33, 1346–1349. (20) Ray, S. C.; Chiou, J. W.; Pong, W. F.; Tsai, M. H. Crit. ReV. Solid State Mater. Sci. 2006, 31, 91–110. (21) Sham, T. K.; Perlman, M. L.; Watson, R. E. Phys. ReV. B 1979, 19, 539–545. (22) Ago, H.; Kugler, T.; Cacialli, F.; Salaneck, W. R.; Shaffer, M. S. P.; Windle, A. H.; Friend, R. H. J. Phys. Chem. B 1999, 103, 8116–8121. (23) Suzuki, S.; Bower, C.; Watanabe, Y.; Zhou, O. Appl. Phys. Lett. 2000, 76, 4007–4009. (24) Zhang, P.; Sham, T. K. Phys. ReV. Lett. 2003, 90, 2455021245502-4() .
JP810639Y