Electrochemical identification of metallic and semiconducting single-walled carbon nanotubes using the water gate effect

Abstract

By fabricating aPMMA control strip at the SWNTs-electrode contact area to screen off the water gate effect, metallic and semiconducting SWNTs were easily identified during the conventional electropolymerization process.

---

Single-walled carbon nanotubes (SWNTs) have demonstrated great potential in creating high-performance field effect transistors (FETs), CMOS circuits, thin film based nano-devices and flexible electronics because of their high carrier mobilities and novel properties.¹ One of the technological challenges for fabricating practical SWNTs-based electronic devices is the identification and separation of metallic and semiconducting SWNTs. It has been a hot topic to effectively identify and separate the semiconducting tubes from metallic mixtures. The widely used methods for identifying metallic and semiconducting SWNTs include spatially resolved resonant-Raman spectroscopy and direct conductivity measurement in a FET configuration. In the case of the Raman approach, only those tubes which are resonant with the incident light can be “seen” while for the conductivity measurement approach, it requires a tedious micro/nanofabrication process. In particular, both of these approaches are a time-consuming tube-by-tube identification. On the other hand, the typical techniques developed for separation purposes up to now include electrical breakdown,² selective etching by gas-phase reaction,³ light irradiation,^4,5dielectrophoresis,⁶DNA-assisted dispersion and separation,⁷ultracentrifugation-based separation,⁸ and selective chemical functionalization.⁹ From the point of view of large-scale device fabrication, of particular importance are the highly-oriented ultralong SWNT arrays directly grown on wafers by catalytic chemical vapor deposition (CVD).¹⁰ With such an ultralong SWNT array system, one needs to develop an effective method for the identification and further separation. We report herein an electrochemical approach for the identification purpose for such systems by using the water gate effect in an aqueous electrochemical cell. Our electrochemical identification is a batch-like operation, which can simultaneously recognize all the metallic and semiconducting tubes on the surface. This approach may also provide possibility for further separation.

Electrochemical techniques have often been used for synthesizing nanomaterials, and carbon nanotubes have been employed as unique nanoelectrodes for deposition of polypyrrole to form core–shell nanostructures.¹¹ In the electrochemical cell system, an electrical double layer is generated at the SWNTs–electrolyte interface. Thus the aqueous electrolyte solution can function as a water gate to tune the conductance of semiconducting SWNTs, leading to a water-gated high-performance SWNT field effect transistor (FET).¹² Because the electrodeposition of polypyrrole takes place at a very positive potential, the semiconducting SWNTs are turned on during the conventional electropolymerization process. It is thus not expected to make a reliable discrimination between metallic and semiconducting tubes. The key of our approach is to fabricate a poly(methyl methacrylate) (PMMA) control strip at the SWNTs-electrode contact area to screen off the water gate effect. The semiconducting tubes are therefore turned off at the PMMA-covered area, and become more resistive to the electrochemical reaction while metallic tubes are not affected. Our results demonstrate that the PMMA control strip has drastically magnified the difference of metallic and semiconducting tubes upon electropolymerization.

Fig. 1(a) illustrates the experimental approach. Pt or Au electrodes were fabricated on silicon wafers by electron-beam lithography (EBL), Ti/Pt or Ti/Au deposition, and lift-off process. The ultralong SWNTs were grown by CVD method¹³ directly on the wafer with Ti/Pt electrodes or were transferred onto the wafer with Ti/Au electrodes using the PMMA-mediated nanotransfer printing technique we developed recently.¹⁴ The samples with transferred SWNTs on Ti/Au electrodes were annealed to improve the contact and remove the contamination. I–V measurements were carried out to determine whether the nanotube was semiconducting or metallic. Raman spectroscopy was used to exclude the multiwalled tubes (Fig. S1, ESI† ). The PMMA control strip was spin-coated and EBL-patterned onto the nanotube–metal contact area with a typical strip width of 50 μm from the electrode edge. Electrodeposition of polypyrrole was performed by adding a drop of pyrrole solution (25 mM pyrrole and 0.1 M SDS in water) to the device, followed by applying a bias voltage on the nanotube (vs. Ag/AgCl) using a CHI-710 electrochemical workstation (Fig. S2, ESI† ). The deposition process was off-line monitored by SEM and AFM characterization. Fig. 1(b) illustrates the water gate effect during the electropolymerization process. Under a 0.6 V of potential bias, an electrical double layer is generated at the SWNT–electrolyte interface, which induces the band structure change of a semiconducting tube. Because the PMMA control strip can screen off this water gate effect by preventing the SWNT from contacting with electrolyte solution, it can control the band bending along the tube axis, as shown in Fig. 1(b), which is expected to enlarge the difference of metallic and semiconducting tubes upon electropolymerization. Fig. 2(a) shows the SEM image of the as-grown ultralong SWNT array. These ultralong SWNTs were transferred onto the wafer with Ti/Au electrodes as shown in Fig. 2(b) and the diameter of which were measured by tapping mode AFM. Fig. 2(c) and 2(d) show the typical transfer characteristics of the metallic and semiconducting ultralong SWNTs.

Fig. 1 (a) Schematic illustration of the experimental procedures. (b) Illustration of the water gate effect on the band structure of a semiconducting tube in aqueous electrochemical system with (middle) and without (bottom) a PMMA control strip.

Fig. 2 (a) SEM image of the as-grown ultralong SWNTs array. (b) SEM image of a typical sample with transferred SWNTs on Ti/Au electrodes (lower inset: AFM image of SWNT; upper inset: SEM of SWNT between electrodes). (c) Transfer characteristics of a typical metallic SWNT. (d) Transfer characteristics of a typical semiconducting SWNT. The bias voltage was 100 mV.

Fig. 3 compares the electrodeposition results of polypyrrole on SWNTs without (a) and with (b) a PMMA control strip obtained at a typical bias potential of +0.6 V (vs. Ag/AgCl) for 30 s. The deposited thickness of polypyrrole is plotted vs. the pristine tube diameter. Fig. 3(a) summarizes the results of four metallic and four semiconducting tubes without PMMA control strip (Table S1, ESI† ). The AFM images show the difference of typical metallic and semiconducting tubes after electropolymerization. It is found that metallic tubes always have more deposition of polypyrrole than semiconducting tubes though the difference is not very large. For semiconducting tubes, the deposition thickness of polypyrrole increases with the tube diameter, approaching to that for the metallic tubes. This performance can be well understood considering the reciprocal dependence of bandgap energy on semiconducting tube diameter. Fig. 3(b) summarizes the results of five metallic and five semiconducting tubes taken with a PMMA control strip (Table S2, ESI† ). As clearly seen from the inset AFM images, the PMMA control strip brings about a drastic change on the electrodeposition behavior of metallic and semiconducting tubes. For metallic tubes, no remarkable changes are observed under the same experimental conditions as compared with the case of no PMMA control strip. In contrast, the deposition thickness of polypyrrole on semiconducting tubes shows a sharp decrease with the use of a PMMA control strip. For instance, a 2.7 nm metallic tube electrodeposited 183.3 nm of polypyrrole while a 2.5 nm semiconducting tube only electrodeposited 18.5 nm of polypyrrole at the same condition. Different from the case of no PMMA control strip, there is no clear dependence of deposition thickness on semiconducting tube diameter. Although there exists a large deviation from tube to tube and from sample to sample, possibly arising from the less controllable electrical contact between nanotubes and electrodes, the semiconducting tubes are always much less active as nanoelectrodes for electropolymerization than metallic tubes when the PMMA control strip is employed. Shown in Fig. 4 is another example demonstrating the remarkable difference of metallic and semiconducting tubes upon electropolymerization. In this sample, the SWNTs array was transferred to the Au electrode patterns on the SiO₂/Si surface using the nanotransfer printing technique. After electropolymerization, three tubes showed great changes in diameter and two tubes showed almost no diameter changes. Subsequent electrical measurements indicated that the thick tubes correspond to metallic and the thin tubes to semiconducting. Similar conclusions have been drawn from repeating experiments using different batches of samples. As a consequence, we can easily discriminate metallic and semiconducting SWNTs from the arrays by electrodeposition of polypyrrole with the aid of a PMMA control strip. Under the same electropolymerization conditions, drastically thickened tubes are metallic and the others are semiconducting.

Fig. 3 Dependence of polypyrrole thickness electrodeposited on SWNTs on the tube diameter for metallic (blue) and semiconducting tubes (red) without (a) and with (b) a PMMA control strip. The thickness data were measured at a position of ca. 60 mm away from electrode edge. The blue and red guide lines show the changing tendency of deposition thickness on tube diameter. The bottom AFM images demonstrate the difference of metallic and semiconducting tubes after electrodeposition.

Fig. 4 Electrodeposition of polypyrrole on an ultralong SWNTs array (with five SWNTs) transferred on Au electrodes and covered with a PMMA control strip. The thick tubes correspond to metallic SWNTs and the thin tubes to semiconducting SWNTs as confirmed by electrical measurements.

The electrochemistry of SWNTs has been studied theoretically by Heller et al.¹⁵ According to the theory, the total capacitance of the electrical double layer formed at the SWNT–electrolyte interface includes two different contributions: the electrostatic double-layer capacitance C_dl and the chemical quantum capacitance C_q originating from the rather low density of electronic states per unit energy around the Fermi level of SWNT. Since C_q is an order of magnitude smaller than C_dl, C_q dominates the overall capacitance.¹² This quantum effect governed capacitance makes the electrolyte an excellent gate coupling to the SWNT. The electrochemical gating effect of SWNTs FETs has been intensively studied not only in aqueous NaCl or H₂SO₄ solution^12,16 but also in polymer electrolyte using LiClO₄/PEO or LiClO₄/PEI.^17–19 In our system, the SWNT under PMMA control strip is isolated from the electrolyte solution. Hence the water gate does not affect the band structure of the semiconducting tube and the Fermi level is still located in the bandgap of SWNT. As a result, the carrier concentration of the semiconducting tube under the PMMA control strip will be very low, leading to the high resistance. On the other hand, the Schottky barrier formed at the nanotube–metal contact plays an important role for carrier injection. A gate-induced electric field at the contact controls the barrier width and hence the current.²⁰ When the SWNT is covered with PMMA control strip, this part of the SWNT does not couple with the water gate so that the water gate can not modulate the Schottky barrier formed in this region. This thick barrier blocks the charge tunneling and contributes to the large contact resistance at the tube–metal interface. The above situation is schematically illustrated in Fig. 1(b). When a voltage is applied to the metal electrode in contact with the semiconducting SWNT, a remarkable potential drop occurs along the nanotube under the PMMA-covered region. This effectively retards the electrochemical reaction on the nanotube surface. In contrast, the metallic SWNTs do not have a bandgap in their band structure, and can form a good contact with metal electrode. In this case, the water gate effect does not tune the conductance of the metallic tube. As a result, the electrochemical reaction on a metallic SWNT is not affected by the PMMA control strip.

In summary, using the water gate effect in an electrochemical cell, we can easily identify metallic and semiconducting SWNTs with the aid of a PMMA control strip. This electrochemical identification technique is a batch-like operation and one can simultaneously identify all the SWNTs aligned on surfaces.

This work was supported by NSFC (grants no. 50821061, 20833001) and MOST (grants no. 2007CB936203).

Notes and references

1. (a) M. Ouyang, J. L. Huang and C. M. Lieber, Acc. Chem. Res., 2002, 35, 1018; (b) P. Avouris, Acc. Chem. Res., 2002, 35, 1026; (c) H. Dai, Acc. Chem. Res., 2002, 35, 1035; (d) H. Zhu and B. Wei, J. Mater. Sci. Technol., 2008, 24, 447; (e) X. Sun and Y. Sun, J. Mater. Sci. Technol., 2008, 24, 569.
2. P. G. Collins, M. S. Arnold and P. Avouris, Science, 2001, 292, 706.
3. G. Zhang, P. Qi, X. Wang, Y. Lu, X. Li, R. Tu, S. Bangsaruntip, D. Mann, L. Zhang and H. Dai, Science, 2006, 314, 974.
4. H. J. Huang, R. Maruyama, K. Noda, H. Kajiura and K. Kadono, J. Phys. Chem. B, 2006, 110, 7316.
5. Y. Y. Zhang, Y. Zhang, X. J. Xian, J. Zhang and Z. F. Liu, J. Phys. Chem. C, 2008, 112, 3849.
6. (a) R. Krupke, F. Hennrich, H. von Lohneysen and M. M. Kappes, Science, 2003, 301, 344; (b) Z. Chen, Z. Y. Wu, L. M. Tong, H. P. Pan and Z. F. Liu, Anal. Chem., 2006, 78, 8069.
7. M. Zeng, A. Jagota, E. D. Semke, B. A. Diner, R. S. Mclean, S. R. Rustig, R. E. Richardson and N. G. Tassi, Nat. Mater., 2003, 2, 338.
8. (a) M. S. Arnold, A. A. Green, J. F. Hulvat, S. I. Stupp and M. C. Hersam, Nat. Nanotechnol., 2006, 1, 60; (b) N. Izard, S. Kazaoui, K. Hata, T. Okazaki, T. Saito, S. Iijima and N. Minami, Appl. Phys. Lett., 2008, 92, 243112.
9. (a) M. S. Strano, C. A. Dyke, M. L. Usrey, P. W. Baron, M. J. Allen, H. Shan, C. Kittrell, R. H. Hauge, J. M. Tour and R. E. Smalley, Science, 2003, 301, 1519; (b) L. An, Q. A. Fu, C. G. Lu and J. Liu, J. Am. Chem. Soc., 2004, 126, 10520; (c) M. Kanungo, H. Lu, G. G. Malliaras and G. B. Blanchet, Science, 2009, 323, 234.
10. S. M. Huang, X. Y. Cai and J. Liu, J. Am. Chem. Soc., 2003, 125, 5636.
11. (a) Q. L. Zhao, Z. L. Zhang, B. H. Huang, J. Peng, M. Zhang and D. W. Pang, Chem. Commun., 2008, 5116; (b) Q. Shen, L. Jiang, J. Miao, W. Hou and J. J. Zhu, Chem. Commun., 2008, 1683; (c) M. Gao, S. M. Huang, L. M. Dai, G. Wallace, R. P. Gao and Z. L. Wang, Angew. Chem., Int. Ed., 2000, 39, 3664.
12. S. Rosenblatt, Y. Yaish, J. Park, J. Gore, V. Sazonova and P. L. McEuen, Nano Lett., 2002, 2, 869.
13. Y. G. Yao, Q. W. Li, J. Zhang, R. Liu, L. Y. Jiao, Y. T. Zhu and Z. F. Liu, Nat. Mater., 2007, 6, 283.
14. (a) L. Y. Jiao, B. Fan, X. J. Xian, Z. Y. Wu, J. Zhang and Z. F. Liu, J. Am. Chem. Soc., 2008, 130, 12612; (b) L. Y. Jiao, X. J. Xian and Z. F. Liu, J. Phys. Chem. C, 2008, 112, 9963.
15. I. Heller, J. Kong, K. A. Williams, C. Dekker and S. G. Lemay, J. Am. Chem. Soc., 2006, 128, 7353.
16. J. Mannik, B. R. Goldsmith, A. Kane and P. G. Collins, Phys. Rev. Lett., 2006, 97, 016601.
17. C. G. Lu, Q. Fu, S. M. Huang and J. Liu, Nano Lett., 2004, 4, 623.
18. G. P. Siddons, D. Merchin, J. H. Back, J. K. Jeong and M. Shim, Nano Lett., 2004, 4, 927.
19. T. Ozel, A. Gaur, J. A. Rogers and M. Shim, Nano Lett., 2005, 5, 905.
20. S. Heinze, J. Tersoff, R. Martel, V. Derycke, J. Appenzeller and P. Avouris, Phys. Rev. Lett., 2002, 89, 10681.

---

Footnote

† Electronic supplementary information (ESI) available: Raman spectra, experimental setup for electrodeposition, statistics of polypyrrole thickness and the corresponding metallic/semiconducting tube diameter with and without PMMA control strip. See DOI: 10.1039/b900379g

---