Effect of oxygen adsorption on the electrochemical oxidative corrosion of single-walled carbon nanotubes

Abstract

The effect of adsorbed molecular oxygen on the oxidative corrosion of single-walled carbon nanotubes in aqueous solution was investigated by Raman spectroscopy. Adsorbed molecular oxygen affected nucleation and growth in the electrochemical oxidative corrosion of single-walled carbon nanotubes in aqueous electrolyte. Nucleation and growth began at defect sites in the presence of adsorbed oxygen, but occurred randomly in the absence of adsorbed oxygen. This insight furthers our understanding of the oxidative corrosion of sp²-hybridized carbon, and may enhance the development of fuel cells and sensors based on carbon nanotube electrodes.

---

Oxidation is one of the main routes for functionalizing and manipulating sp²-hybridized carbon (sp²-carbon). Graphene membranes contain sub-nm-sized pores formed by the oxidative corrosion of carbon, and tuning their size allows selective ionic transport.¹ The electronic properties of sp²-carbon can be tuned by controlled defect formation through oxidative corrosion.² The oxidative corrosion of carbon gradually decreases the output power of polymer electrolyte fuel cells because sp²-carbon is used as a platform electrode to immobilize the catalyst.³ Understanding the mechanism of nucleation and growth in oxidative corrosion at the surface of sp²-carbon is important for advancing its application. Many electrochemical studies have focused on elucidating the oxidation mechanism of highly oriented pyrolytic graphite (HOPG), carbon nanotubes and graphene as model sp²-carbon systems.⁴

The reactivity at basal-plane sites of sp²-carbon is considerably lower than that at the edges or defect sites.⁵ The reactivity of a carbon material is closely related to its surface defect sites, such as edges, steps, kinks and surface vacancies. Oxygen plays an important role in oxidative corrosion. Theoretical studies suggest that molecular oxygen readily adsorbs at surface defect sites of sp²-carbon, thus inducing oxidative corrosion.^5–8 Mechanisms for the nucleation and growth of defects induced by oxidative corrosion have also been theoretically investigated. There is currently little experimental evidence supporting this mechanism because of limited suitable analysis methods and conditions.⁴ We recently reported the oxidative corrosion potential vs. pH diagram for single-walled carbon nanotubes (SWCNTs).⁹ The current Raman spectroscopy studies based on the encapsulation of β-carotene into SWCNTs revealed that there are three types of oxidative corrosion of SWCNTs: non-oxidized, end-cap oxidized (end-cap eliminated), and side-wall oxidized SWCNTs. SWCNTs are good models for experimentally analyzing the oxidative corrosion of sp²-carbon.⁹ In the current study, the effect of adsorbed molecular oxygen on the oxidative corrosion of SWCNTs in aqueous solution has been investigated using Raman spectroscopy. We found that adsorbed molecular oxygen affects nucleation and growth in the electrochemical oxidative corrosion of SWCNTs.

SWCNTs were synthesized on a gold wire surface by chemical vapor deposition as described previously (see ESI†).¹⁰ The G-band/D-band intensity ratio indicated that the SWCNTs had highly crystalline sp²-carbon structures. The fresh SWCNT electrode was immediately immersed in an electrolyte solution (pH 11) containing HClO₄ and NaOH. Controlled-potential electrolysis (CPE) for 30 min using a conventional three-electrode system was used to oxidize the SWCNTs. The oxidation of the SWCNTs was probed by Raman spectroscopy and the encapsulation of β-carotene by the SWCNTs. β-Carotene was selected as an organic molecular encapsulation agent because it is well known that the encapsulation behavior is sensitive to the hydrophobic cavities of SWCNTs.^9a,11 Specifically, the SWCNTs were immersed in hexane containing 2 mmol dm⁻³ β-carotene, and refluxed for 10 h under Ar.¹⁰ The SWCNTs were characterized by Raman spectroscopy using a Horiba (Jobin Yvon) LabRAM HR-800 instrument, with 514 nm (2.41 eV) laser excitation. The diameter of the SWCNTs observed from transmission electron microscopy images and estimated from the radial breathing mode (RBM) in Raman spectra was 0.9–1.6 nm, which was sufficiently large for β-carotene to be encapsulated.¹¹

Fig. 1a shows cyclic voltammograms the SWCNTs in electrolyte solution (pH 11) under Ar. The first potential scan showed a weak cathodic current at ca. −0.15 V, which was not observed during the second potential cycle. The cathodic current was due to the reduction of adsorbed oxygen on the SWCNT surface.⁹ A similar voltammetric behavior was observed at an HOPG electrode, but the cathodic current began at ca. −0.25 V (Fig. 1b). This indicated that adsorbed oxygen was more readily electrochemically reduced at the SWCNT electrode than at HOPG, as reported previously.¹² Following the reduction of adsorbed oxygen, a similar reduction current was observed when the SWCNTs were exposed to aerobic conditions. Molecular oxygen readily adsorbed to the sp²-carbon, and was not readily removed by inert gas purging. Thus, −0.6 V (vs. Ag/AgCl/saturated KCl) was applied to the SWCNTs prior to CPE to remove adsorbed oxygen on the SWCNT surface.⁹ The amount of adsorbed oxygen on the SWCNTs was estimated to be ca. 5 × 10⁻⁵ per carbon atom, based on the reduction charge from adsorbed oxygen in the first cyclic voltammogram scan (see ESI†). Ulbricht et al. estimated the equilibrium oxygen coverage of SWCNTs (average diameter ca. 1.2 nm) to be ca. 10⁻⁵ molecules per carbon atom (SWNCTs fabricated by heating and annealing to remove functionality, measured at room temperature and atmospheric conditions).¹³ The minor difference to the current result may reflect the different measurement conditions (atmospheric vs. electrolyte solution) and/or SWCNT quality.

Fig. 1 Typical cyclic voltammograms of (a) single-walled carbon nanotubes (SWCNTs) and (b) highly oriented pyrolytic graphite in aqueous electrolyte (pH 11) containing HClO₄ and NaOH under Ar. Electrode surface areas: (a) 0.25 and (b) 0.50 cm². Potential sweep rate: 20 mV s⁻¹.

Fig. 2 shows Raman spectra of SWCNTs that were subjected to CPE at different potentials in alkaline solution (pH 11), followed by encapsulation treatment. The encapsulation of β-carotene within SWCNTs was apparent when the SWCNTs were subjected to CPE at 0.8–1.2 V (vs. Ag/AgCl) because the v₄ out-of-plane C–H wagging mode was observed at ca. 945 cm⁻¹ (see ESI†).¹¹ This behavior was observed for the as-prepared SWCNTs and those subjected to oxygen reduction, though the peak intensity of the latter decreased because of their oxidative corrosion. β-carotene is stabilized within hydrophobic cavities, resulting in the v₄ mode. Thus, the hydrophobic tubular spacing within the SWCNTs was maintained. For CPE at 1.4 V, the v₄ mode was observed for the as-prepared SWCNTs, but not for the SWCNTs subjected to oxygen reduction. This indicated that the hydrophobic tubular spacing of the oxygen-free SWCNTs was not maintained at 1.4 V. This result can be explained as such: the oxygen-containing hydrophilic functional groups (such as –C=O, –COOH, and –C–OH), generated upon oxidation of SWCNTs, were preferentially widely distributed on oxygen-free SWCNTs rather than on the as-prepared SWCNTs. A differing D-band/G-band intensity ratio was observed for different CPE potentials. At 1.2 and 1.4 V, the D-band/G-band intensity ratio of the SWCNTs subjected to oxygen reduction was significantly higher than that of the as-prepared SWCNTs. The D-band results from an imperfect lattice structure, particularly the edges and defects of sp²-carbon.¹⁴ These results led to the following two hypotheses: (1) the oxidative corrosion of oxygen-free SWCNTs occurred more randomly than that of oxygen-adsorbed SWCNTs, and decreased the perfect lattice structure area of the SWCNTs, or (2) the oxidative corrosion of the oxygen-free SWCNTs occurred at lower potential and/or faster than that in oxygen-adsorbed SWCNTs. Theoretical studies have suggested that molecular oxygen enhances the oxidative corrosion of carbon;^5–7 so the second hypothesis was discounted. Molecular oxygen has been theoretically proposed to readily adsorb onto defect sites of sp²-carbon,^5–7 and supports the first hypothesis. Nucleation and growth of the oxidative corrosion of oxygen-adsorbed SWCNTs was expected to begin at defect sites. Thus, nucleation and growth occurred randomly when adsorbed oxygen was removed, which decreased the perfect lattice structure area, and increased the intensity of the D-band. This was supported by the observed encapsulation behavior of β-carotene. The random oxidative corrosion decreased the hydrophobic tubular spacing within the SWCNTs. It also generated oxygenated-carbon functionality, which increased the hydrophilic character and inhibited the encapsulation of β-carotene within the SWCNTs. Nucleation and growth of the oxidative corrosion is shown schematically in Fig. 3. This is the first significant evidence of differing nucleation and growth in the oxidative corrosion of sp²-carbon under aerobic and anaerobic conditions.

Fig. 2 Raman spectra of SWCNTs subjected to controlled-potential electrolysis (CPE) followed by β-carotene encapsulation. CPE was carried out at various potentials (a) before and (b) after the reduction of adsorbed oxygen.

Fig. 3 Schematic of the oxidative corrosion of SWCNTs (a) with and (b) without adsorbed oxygen.

Fig. 4 shows the RBM region of the Raman spectra of SWCNTs subjected to CPE at different potentials in alkaline solution (pH 11). The RBM peak intensity decreased with increasing positive potential from 0.9 V. This indicated that oxidative corrosion occurred and was enhanced at higher potential. The peak intensity for the oxygen-adsorbed SWCNTs was slightly higher than that for the oxygen-free SWCNTs under the same potential, at 1.1 and 1.4 V. This suggested that nucleation and growth of oxidative corrosion began at defect sites in the oxygen-adsorbed SWCNTs, but occurred randomly in the oxygen-free SWCNTs. Fig. 5 shows the normalized RBM peak intensity at ca. 185 and 270 cm⁻¹ as a function of CPE potential. The peaks at ca. 185 and 270 cm⁻¹ originated from semiconductor-type SWCNTs (s-SWCNTs) (diameter ca. 1.3 nm) and metallic-type SWCNTs (m-SWCNTs) (diameter ca. 0.9 nm), respectively.^11,14 The trends at ca. 1.0 V during the initial oxidative corrosion are explained as follows. First, the peak intensity decrease with potential for the oxygen-adsorbed SWCNTs was much lower than that for the oxygen-free SWCNTs, for both s-SWCNTs and m-SWCNTs. Second, m-SWCNTs were more readily oxidized than s-SWCNTs under the same conditions. Third, adsorbed oxygen had a greater effect on m-SWCNTs than on s-SWCNTs. These results suggested that the differing oxidized corrosion behavior between s-SWCNTs and m-SWCNTs was due to adsorbed oxygen. This also accounts for reported observations (though under different conditions) suggesting that m-SWCNTs are more readily electrochemically etched in aqueous electrolyte under aerobic conditions than s-SWCNTs of similar diameter.¹⁵

Fig. 4 RBM region of Raman spectra of SWCNTs subjected to CPE followed by β-carotene encapsulation. CPE was carried out at various potentials (a) before and (b) after the reduction of adsorbed oxygen.

Fig. 5 Normalized RBM peak intensity as a function of CPE potential. Closed and open squares show the peak intensity (185 cm⁻¹) corresponding to semiconductor-type SWCNTs with and without adsorbed oxygen, respectively. Closed and open circles show the peak intensity (270 cm⁻¹) corresponding to metallic-type SWCNTs with and without adsorbed oxygen, respectively.

The amount of adsorbed oxygen was estimated to be ca. 5 × 10⁻⁵ per carbon atom, as described above. The initial adsorbed oxygen was insufficient to sustain further oxidative corrosion of the SWCNTs at ca. >1.0 V. Thus, oxygen reacting with the SWCNTs was supplied by the electrolysis of water. Heller et al. suggested that the electrochemical reaction rate is related to the overlap of the unoccupied states of SWCNTs and reduced states in solution.¹⁶ The differing oxidized corrosion behavior between s-SWCNTs and m-SWCNTs at high positive potential may have reflected these differing electrochemical reaction rates. The diameter selectivity was attributed to the stability of the SWCNTs. Small-diameter SWCNTs are preferentially oxidized because of the higher C–C bonding strain.^4,15,17 The oxidative corrosion behavior at up to ca. 1.0 V differed between the oxygen-adsorbed and oxygen-free m-SWCNTs in the current study. Thus, the effect of adsorbed oxygen was much larger than the effect of diameter selectivity during initial oxidative corrosion, even for small diameter (0.9 nm) SWCNTs.

In summary, the electrochemical oxidative corrosion of SWCNTs in aqueous solution was affected by adsorbed oxygen. Nucleation and growth of the oxidative corrosion began at defect sites in the presence of adsorbed oxygen, but occurred randomly in the absence of adsorbed oxygen. Similar behavior was observed for m-SWCNTs and s-SWCNTs. The effect of adsorbed oxygen was significant for m-SWCNTs during initial oxidative corrosion, and was much larger than the effect of diameter selectivity. These insights further our understanding of the oxidative corrosion of sp²-carbon, and may enhance the development of fuel cells and sensors based on carbon nanotube electrodes.

Acknowledgements

This work was supported by a Grant-in-Aid for Scientific Research (no. 24550159) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Notes and references

1. (a) H. W. Kim, H. W. Yoon, S.-M. Yoon, B. M. Yoo, B. K. Ahn, Y. J. Cho, H. J. Shin, H. Yang, U. Paik, S. Kwon, J.-Y. Choi and H. B. Park, Science, 2013, 342, 91; (b) H. Li, Z. Song, X. Zhang, Y. Huang, S. Li, Y. Mao, H. J. Ploehn, Y. Bao and M. Yu, Science, 2013, 342, 95; (c) R. K. Joshi, P. Carbone, F. C. Wang, V. G. Kravets, Y. Su, I. V. Grigorieva, H. A. Wu, A. K. Geim and R. R. Nair, Science, 2014, 343, 752; (d) S. C. O'Hern, M. S. H. Boutilier, J.-C. Idrobo, Y. Song, J. Kong, T. Laoui, M. Atieh and R. Karnik, Nano Lett., 2014, 14, 1234.
2. S. Ghosh, S. M. Bachilo, R. A. Simonette, K. M. Beckingham and R. B. Weisman, Science, 2010, 330, 1656.
3. R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan, N. Garland, D. Myers, M. Wilson, F. Garzon and D. Woode, et al., Chem. Rev., 2007, 107, 3904.
4. (a) H. Chang and A. J. Bard, J. Am. Chem. Soc., 1991, 113, 5588; (b) M. Nose, T. Kinumoto, H.-S. Choo, K. Miyazai, T. Abe and Z. Ogumi, Fuel Cells, 2009, 9, 284; (c) M. Matsumoto, T. Manako and H. Imai, J. Electrochem. Soc., 2009, 156, B1208; (d) R. L. McCreery and M. T. McDermott, Anal. Chem., 2012, 84, 2602; (e) S. Ohmori and T. Saito, Carbon, 2012, 50, 4932; (f) J. Kang, J. Wen, S. H. Jayaram, X. Wang and S.-K. Chen, J. Power Sources, 2012, 234, 208.
5. (a) K. Kinoshita, Carbon: Electrochemical and Physicochemical Properties, John Wiley & Sons, New York, 1988; (b) M. Noel and P. N. Anantharaman, Surf. Coat. Technol., 1986, 28, 161.
6. (a) R. Backreedy, J. M. Jones, M. Pourkashanian and A. Williams, Faraday Discuss., 2001, 119, 385; (b) B. Sanyal, O. Eriksson, U. Jansson and H. Grennberg, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 113409; (c) S. Ghosh, S. M. Bachilo, R. A. Simonette, K. M. Beckingham and R. B. Weisman, Science, 2010, 330, 1656; (d) X. Qi, X. Guo and C. Zheng, Appl. Surf. Sci., 2012, 259, 195; (e) D. J. Mann and W. L. Hase, Phys. Chem. Chem. Phys., 2001, 3, 4376.
7. (a) M. Grujicic, G. Cao and R. Singh, Appl. Surf. Sci., 2003, 211, 166.
8. (a) S.-H. Jhi, S. G. Louie and M. L. Cohen, Phys. Rev. Lett., 2000, 85, 1710; (b) P. Giannozzi, R. Car and G. Scoles, J. Chem. Phys., 2003, 118, 1003; (c) H. Ulbricht, G. Moos and T. Hertel, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 66, 075404.
9. (a) M. Tominaga, Y. Yatsugi and N. Watanabe, RSC Adv., 2014, 4, 27224; (b) S. Sakamoto and M. Tominaga, Chem.–Asian J., 2013, 8, 2680.
10. (a) M. Tominaga, S. Sakamoto and H. Yamaguchi, J. Phys. Chem. C, 2012, 116, 9498; (b) M. Tominaga, A. Iwaoka, D. Kawai and S. Sakamoto, Electrochem. Commun., 2013, 31, 76; (c) M. Tominaga, M. Togami, M. Tsushida and D. Kawai, Anal. Chem., 2014, 86, 5053.
11. (a) K. Yanagi, Y. Miyata and H. Kataura, Adv. Mater., 2006, 18, 437; (b) P. Horn and M. Kertesz, J. Phys. Chem. C, 2010, 114, 12139; (c) K. Yanagi, K. Iakoubovskii, S. Kazaoui, N. Minami, Y. Maniwa, Y. Miyata and H. Kataura, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 155420.
12. (a) M. Zhang, Y. Yan, K. Gong, L. Mao, Z. Guo and Y. Chen, Langmuir, 2004, 20, 8781; (b) I. Kruusenberg, N. Alexeyeva and K. Tammeveski, Carbon, 2009, 47, 651; (c) A. Dumitru, M. Mamlouk and K. Scott, Electrochim. Acta, 2014, 135, 428.
13. H. Ulbricht, G. Moos and T. Hertel, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 66, 075404.
14. (a) S. Iijima, Nature, 1991, 354, 56; (b) P. M. Ajayan, Chem. Rev., 1999, 99, 1787; (c) P. R. Wallace, Phys. Rev., 1947, 71, 622; (d) R. Saito, G. Dresselhaus and M. S. Dresselhaus, Physical Properties of Carbon Nanotubes, Imperial College Press, London, 1998.
15. D. Wei, Y. Liu, L. Cao, H. Zhang, L. Huang, G. Yu, H. Kajiura and Y. Li, Adv. Funct. Mater., 2009, 19, 3618.
16. (a) I. Heller, J. Kong, K. A. Williams, C. Dekker and S. G. Lemay, J. Am. Chem. Soc., 2006, 128, 7353; (b) I. Heller, J. Kong, H. A. Heering, K. A. Williams, S. G. Lemay and C. Dekker, Nano Lett., 2005, 5, 137.
17. (a) D. H. Robertson, D. W. Brenner and J. W. Mintmire, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 45, 12592; (b) J. W. Mintmire and C. T. White, Carbon, 1995, 33, 893.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra10521d

---