Controlling the Carbon Nanotube-to-Medium Conductivity Ratio for

The surface conductivity of colloidal nanotubes, induced by ionic surfactants, is known to affect alternating current dielectrophoresis, which has bee...
1 downloads 0 Views 1MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Controlling the Carbon Nanotube-to-Medium Conductivity Ratio for Dielectrophoretic Separation Junmo Kang,† Seunghyun Hong,† Youngjin Kim,†,‡ and Seunghyun Baik*,†,‡,§ †

SKKU Advanced Institute of Nanotechnology (SAINT) and ‡School of Mechanical Engineering and § Department of Energy Science, Sungkyunkwan University, Suwon, GyeongGi-Do 440-746, Korea Received September 9, 2009. Revised Manuscript Received October 6, 2009

The surface conductivity of colloidal nanotubes, induced by ionic surfactants, is known to affect alternating current dielectrophoresis, which has been actively investigated with regard to separating single-walled carbon nanotubes according to electronic type. The nanotube-to-suspending medium conductivity ratio is a primary factor for determining the dielectrophoretic behavior of semiconducting nanotubes. In this study, our theoretical and experimental analysis revealed that the suspending medium conductivity also plays an important role in controlling the conductivity ratio. This work elucidates the effects of several surfactant systems on the conductivity ratio and therefore the degree of separation between metallic and semiconducting nanotubes. The equimolar mixture of anionic and cationic surfactants was more effective than a nonionic polymer in reducing the conductivity ratio because the conductivity of colloidal nanotubes was decreased and that of the suspending medium was increased. Besides, the surfactant mixture provided a better dispersion of nanotubes. The dielectrophoretic separation was carried out using microelectrodes with a gap size of 4 μm at an electric field frequency of 10 MHz. The complete separation of nanotubes at the reduced conductivity ratio was confirmed by Raman spectroscopy and electrical transport measurements.

Introduction A variety of methodologies have been actively investigated with regard to separating single-walled carbon nanotubes (SWNTs) according to electronic type1-7 because heterogeneous mixtures of metallic, semimetallic, and semiconducting species are synthesized by currently available production techniques.8,9 The separated nanotubes can be used as major components in future electronic devices.10-12 Alternating current dielectrophoresis has received considerable attention as a nondestructive separation method.7,13,14 Recently, dielectrophoresis was successfully realized in a microfluidic channel to demonstrate a potential to scale up the separation of nanotubes.15 *Corresponding author. E-mail: [email protected]. (1) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C. Nat. Nanotechnol. 2006, 1, 60. (2) Tananka, T.; Jin, H.; Miyata, Y.; Fujii, S.; Suga, H.; Naitoh, Y.; Minari, T.; Miyadera, T.; Tsukagoshi, K.; Kataura, T. Nano Lett. 2009, 9, 1497. (3) Strano, M. S.; Dyke, C. A.; Usrey, M. L.; Barone, P. W.; Allen, M. J.; Shan, H. W.; Kittrell, C.; Hauge, R. H.; Tour, J. M.; Smalley, R. E. Science 2003, 301, 1519. (4) Chen, Z.; Du, X.; Du, M.; Rancken, C. D.; Cheng, H.; Rinzler, A. G. Nano. Lett. 2003, 3, 1245. (5) An, K. H.; Park, J. S.; Yang, C. M.; Jeong, S. Y.; Lim, S. C.; Kang, C.; Son, J. H.; Jeong, M. S.; Lee, Y. H. J. Am. Chem. Soc. 2005, 127, 5196. (6) Zhang, G.; Qi, P.; Wang, X.; Lu, Y.; Li, X.; Tu, R.; Bangsaruntip, S.; Mann, D.; Zhang, L.; Dai, H. J. Science 2006, 314, 974. (7) Krupke, R.; Hennrich, F.; Lohneysen, H.; Kappes, M. Science 2003, 301, 344. (8) Thess, A.; Lee, R.; Nikolaev, P.; Dai, H. J.; Petit, P.; Robert, J.; Xu, C.; Lee, Y. H.; Kim, S. G.; Rinzler, A. G.; Colbert, D. T.; Scuseria, G. E.; Tomanek, D.; Fischer, J. E.; Smalley, R. E. Science 1996, 273, 483. (9) Dresselhaus, M. S.; Dresselhaus, G.; Avouris, Ph. Carbon Nanotubes: Synthesis, Structure, Properties and Applications; Springer-Verlag: New York, 2001. (10) McEuen, P. Phys. World 2000, 13, 31. (11) Avouris, Ph. Acc. Chem. Res. 2002, 35, 1026. (12) Tans, S. J.; Verschueren, A. R. M.; Dekker, C. Nature 1998, 393, 49. (13) Krupke, R.; Hennrich, F.; Kappes, M.; Lohneysen, H. Nano Lett. 2004, 4, 1395. (14) Baik, S.; Usrey, M.; Rotkina, L.; Strano, M. S. J. Phys. Chem. B 2004, 108, 15560. (15) Shin, D. H.; Kim, J. E.; Shim, H. C.; Song, J. W.; Yoon, J. H.; Kim, J.; Jeong, S.; Kang, J.; Baik, S.; Han, C. S. Nano Lett. 2008, 8, 4380.

Langmuir 2009, 25(21), 12471–12474

A dipole moment is induced for a colloidal nanotube suspended in a dielectric solution medium under the influence of a nonuniform electric field.7,16,17 Metallic nanotubes always experience positive dielectrophoresis in moving toward regions of high electric field strength. However, semiconducting tubes exhibit either positive or negative dielectrophoresis depending on the electric field frequency and surface conductivity of nanotubes.7,14,18,19 In all cases, dielectrophoretic separation resulted in metal-enriched carbon nanotubes because the dielectrophoretic force for metallic species is significantly greater than that for semiconducting types.7,13-15,18,19 Separation could also be achieved when the induced positive force for semiconducting nanotubes was not greater than the competing forces caused by Brownian motion and convection.15 Nevertheless, it is important to induce negative dielectrophoresis of semiconducting species, pushing tubes toward regions of low electric field strength, to maintain more stable, complete separation according to the electronic type. Our previous work demonstrated that the surface conductivity of nanotubes can be induced by dispersing agents, such as ionic surfactants, resulting in positive dielectrophoresis of the semiconducting type.14,18,19 Mixtures of cationic and anionic surfactants were suggested to decrease the surface conductivity of nanotubes.18,19 White et al. systematically investigated the effects of diverse dispersants on the surface charge of nanotubes and suggested nonionic polymers to decrease the surface conductivity of nanotubes effectively.20 The nanotube-to-suspending medium conductivity ratio is a primary factor in determining the dielectrophoretic behavior of (16) Morgan, H. J.; Green, N. G. AC Electrokinetics: Colloids and Nanoparticles; Research Studies Press: Philadelphia, PA, 2003; Chapters 6 and 7. (17) Green, N. G.; Morgan, H. J. J. Phys. Chem. B 1999, 103, 41. (18) Kim, Y. J.; Hong, S.; Jung, S. H.; Strano, M. S.; Choi, J. B.; Baik, S. J. Phys. Chem. B 2006, 110, 1541. (19) Hong, S.; Jung, S. H.; Choi, J. B.; Kim, Y. J.; Baik, S. Langmuir 2007, 23, 4749. (20) White, B.; Banerjee, S.; O’Brien, S.; Turro, N. J.; Herman, I. P. J. Phys. Chem. C 2007, 111, 13684.

Published on Web 10/09/2009

DOI: 10.1021/la903382b

12471

Letter

Kang et al.

semiconducting nanotubes.18,19 In this study, our further theoretical and experimental analysis demonstrated that the suspending medium conductivity also plays an important role in controlling the conductivity ratio. The ratio of several surfactant systems was experimentally determined by the ζ potential of colloidal nanotubes and medium conductivity. The higher conductivity ratio leads to the stronger positive dielectrophoretic force of semiconducting species. The equimolar mixture of anionic sodium dodecyl sulfate (SDS) and cationic cetyltrimethylammonium bromide (CTAB) was more effective than nonionic Pluronic F-127 in reducing the conductivity ratio because the surface conductivity of nanotubes was decreased and that of the medium was increased. Moreover, the surfactant mixture provided a better dispersion of nanotubes, which was monitored using near-infrared fluorescence from semiconducting tubes. The surfactant mixing order was also important in the sample preparation process. Dielectrophoresis was carried out using microelectrodes with a gap size of 4 μm, and separated nanotubes were characterized by Raman spectroscopy and electrical transport measurements.

Experimental Section Five nanotube suspensions were prepared using different dispersing agents as summarized in Table 1, and detailed sample preparation procedures are described in Supporting Information. First, three samples were prepared by dispersing raw HiPco SWNTs (purchased from Carbon Nanotechnology Inc.) in aqueous solutions using anionic SDS (C16), cationic CTAB (C18), and nonionic Pluronic F-127, respectively. Nanotubes were suspended according to a previously reported protocol involving homogenization, ultrasonication, and ultracentrifugation.21 Then two more nanotube suspensions were prepared by mixing surfactants to modulate the surface charge of colloidal nanotubes. For the first mixture, a CTAB solution (w/o SWNTs) was added in a 1:1 molar ratio to the SDS-micellized nanotube suspension (CTAB|SDSSWNTs). The equimolar mixture of CTAB and SDS is expected to form an ion-paired complex, hexadecyltrimethylammonium dodecyl sulfate (CTADS).22,23 The second mixture was made by switching the mixing order. SDS solution (w/o SWNTs) was added to the CTAB-micellized SWNTs (SDS|CTAB-SWNTs) and was previously used in an attempt to tune the surface conductivity of colloidal nanotubes.18,19 We also tried to mix the two surfactants before suspending the SWNTs in the dualsurfactant system. However, SWNTs could not be well-dispersed by this approach (Supporting Information). The surface charge of colloidal nanotubes was measured by the ζ potential (Otsuka Electronics, ELS-8000), and the medium conductivity was measured using a conductivity meter (Hanna, HI2300). The Raman and near-infrared spectra were characterized at an excitation wavelength of 785 nm (Kaiser Optical Systems) to compare the debundling state of colloidal nanotubes. A drop of the SWNT suspension (20 μL) was placed on an electrode with a gap size of 4 μm after switching on the function generator (Agilent 33220A), which was operated at a frequency of 10 MHz and a peak-to-peak voltage (Vp-p) of 7 V. After 1 min, the drop was blown off using nitrogen gas and the function generator was switched off. The dielectrophoretically deposited nanotubes were rinsed before characterization using methanol and deionized water to remove remaining surfactants (Supporting Information). Deposited nanotube arrays were investigated with a Raman microscope (Renishaw, inVia Reflex) at an excitation wavelength of 632.8 nm (HeNe laser). Both metallic and semi(21) O’Connell, M. J.; Bachilo, S.; Huffman, C.; Moore, V. C.; Strano, M. S.; Haroz, E.; Rialon, K.; Boul, B.; Noon, W.; Kitrell, C.; Ma, J.; Hauge, R. H.; Weisman, R.; Smalley, R. E. Science 2002, 297, 593.  (22) Tomasic, V.; Stefani c, I.; Filipovic-Vincekovic, N. Colloid Polym. Sci. 1999, 277, 153. (23) Wang, C.; Lucy, C. A. Electrophoresis 2004, 25, 825.

12472 DOI: 10.1021/la903382b

Table 1. Five Nanotube Suspensions primary dispersant SDS SDS CTAB CTAB Pluronic F-127

secondary dispersant

final suspension SDS-SWNTs CTAB|SDS-SWNTs CTAB-SWNTs SDS|CTAB-SWNTs Pluronic-SWNTs

CTAB SDS

conducting SWNTs are resonant at 632.8 nm. Electrical transport measurements were conducted using a semiconductor parameter analyzer (Agilent, E5262A) and a probe station (CASCADE, RF-1) under ambient conditions.

Results and Discussion A colloidal particle suspended in a dielectric medium is subjected to a dielectrophoretic force that is proportional to the real part of the Clausius-Mossotti factor (CM factor).7,16,17,24 A detailed theoretical analysis to calculate the CM factor was provided previously,18,19 and a brief summary follows. The CM factor for a long rod with the principal axis parallel to the electric field can be written as εp - εm ð1Þ fCM ¼ εm þ ðεp - εm ÞL εp ¼ εp - i

Kp Km , εm ¼ εm - i ω ω

ð2Þ

where ε is the dielectric constant, ω is the applied field frequency, and K is the conductivity. Subscripts p and m indicate the suspended particle and medium fluid, respectively. L is the depolarization factor, defined as L=d2/l2[ln(2l/d) - 1], where d and l indicate the diameter and length of the nanotube, respectively.13,24 The particle moves toward or away from the strong electric field region when the real part of the CM factor is positive or negative, respectively.16 We rederived the CM factor using two parameters: R = εp/εm and β = Kp/Km.18,19 Metallic nanotubes have exceedingly high dielectric constants7,25 (i.e., R . 1) and always experience positive dielectrophoresis. Semiconducting types have a finite value of the dielectric constant smaller than 5 7 and exhibit either positive or negative dielectrophoresis depending on the particle-to-medium conductivity ratio, β.18,19 The particle conductivity, Kp, comprises the intrinsic conductivity and the surface conductivity 2λs ð3Þ Kp ¼ Kint þ a where Kint is the internal particle conductivity, λs is the surface conductance, and a is the radius of a spherical particle.16,26 We assumed that the radius of the double layer of micellized-SWNTs is 2.7 nm.19,21 The surface conductance, λs, consists of two components: diffuse layer conductance and Stern layer conductance.16 The diffuse layer conductance can be calculated using the ζ-potential values of nanotubes and electrolytes.16,19,27 The Stern layer conductance was estimated using the typical ratio of the Stern layer conductance to the diffuse layer conductance of 0.56.19,28 Figure 1a shows the ζ potentials and medium conductivities of the five nanotube suspensions. The surface conductivity of a colloidal nanotube is proportional to the surface charge, which (24) Jones, T. B. Electromechanisms of Particles; Cambridge University Press: Cambridge, England, 1995; Chapter 3. (25) Benedict, L. X.; Louie, S. G.; Cohen, M. L. Phys. Rev. B 1995, 52, 8541. (26) O’Lonski, C. T. J. Phys. Chem. 1960, 64, 605. (27) Lyklema, J. Fundamentals of Interface and Colloid Science; Academic Press: London, 1995; Vol. 2, Chapter 4. (28) Ermolina, I.; Morgan, H. J. Colloid Interface Sci. 2005, 285, 419.

Langmuir 2009, 25(21), 12471–12474

Kang et al.

Letter

Figure 1. (a) ζ potential and solution medium conductivities of five nanotube suspensions. (b) Comparison of the fluorescence emissions of the nanotube suspensions, normalized by the G mode of the Raman spectra.

can be determined by the ζ potential.19 The anionic SDS- and cationic CTAB-micellized SWNTs displayed high negative (-49.8 mV) and positive (56.8 mV) potentials, respectively. The surfactant molecules spontaneously adsorbed on the tube surface via the hydrophobic interaction between the hydrocarbon alkyl chain and the surface atoms of the nanotubes. The surface charge of the nanotubes was determined by the adsorbed ionic surfactants.20 The nonionic Pluronic-suspended nanotubes had a smaller negative charge (-11.5 mV). The ζ potentials for the equimolar surfactant mixture-suspended nanotubes were dependent on the mixing order. The absolute value of the ζ potential was reduced from 49.8 to 34.8 mV when CTAB was added to SDS-micellized nanotubes (CTAB|SDS-SWNTs). However, there was only a slight decrease in the ζ potential, from 56.8 to 52.9 mV, when SDS was added to CTAB-micellized nanotubes (SDS|CTAB-SWNTs). This is contradictory to the previous observation where the ζ potential for the equimolar mixture of SDS and CTAB, without nanotubes, was decreased to almost zero.22 The difference could come from the different affinities of surfactants to nanotubes.20 It is also plausible that the different alkyl chain length could affect the conductivity of surfactant-coated nanotubes. CTAB has a C18 hydrocarbon chain whereas SDS has a C16 hydrocarbon chain. The primary surfactant forms cylindrical micelles around nanotubes.21 The addition of the secondary surfactant was expected to decrease the surface charge because of the electrostatic attraction between anionic and cationic headgroups.22 However, it is evident that the secondary surfactant could not fully penetrate into the micelle structure of the primary surfactant, resulting in incomplete electrical neutralization. The primary surfactant had better affinity to the nanotubes. Also, the CTAB molecule with a longer alkyl chain had better affinity than the SDS molecule because there was a larger decrease in the ζ potential when CTAB was added to the SDS-micellized nanotubes. This may indicate that the affinity of surfactants to nanotubes can be controlled by tailoring the hydrocarbon chain length of surfactants. The solution medium conductivities of the nanotube suspensions are affected by the migration of charged ions under an electric field. The medium conductivities of SDS-SWNTs and CTAB-SWNTs were 0.0862 and 0.0561 S/m, respectively. These values were increased by 58.6 and 99.3% after mixing with the secondary surfactants as a result of the supply of additional counterion carries (Naþ or Br-). CTAB|SDS-SWNTs had their highest conductivity of 0.1367 S/m. The nonionic Pluronicmicellized SWNTs had their lowest conductivity of 0.0063 S/m because they do not have counterion carries. The debundling state of colloidal nanotubes was investigated by near-infrared fluorescence emission (785 nm excitation) as shown in Figure 1b. All of the fluorescence emissions were normaLangmuir 2009, 25(21), 12471–12474

Figure 2. Theoretical plot showing the variation in the real part of the Clausius-Mossotti factor as a function of the permittivity ratio, R, and conductivity ratio, β. The real part of the CM factor is zero at β = 1.28.

lized by the G mode of Raman spectra. The dispersion quality can be estimated by near-infrared fluorescence because it is observed for individually dispersed nanotubes in solution.21,29-31 The fluorescence of Pluronic-SWNTs was significantly lower than those of other surfactant-micellized nanotubes, indicating the existence of more nanotube bundles. Charged surfactants disperse nanotubes by electrostatic repulsion21,32 whereas nonionic polymers provide steric dispersion only. The individual dispersion of nanotubes is important for electronic-type dependent separation because bundles contain both metallic and semiconducting species.7 Figure 2 shows the calculated CM factor, at a frequency of 10 MHz, as a function of the permittivity ratio, R, and the conductivity ratio, β. Experimentally measured ζ-potential values were used for the calculation, and electrolyte properties were obtained from the literature.19,33 The major components of the electrical double layer of Pluronic-micellized nanotubes were assumed to be Hþ and OH-.34 The permittivity of SWNTs is inversely proportional to the square of (29) Strano, M. S.; Moore, V. C.; Miller, M. K.; Allen, M. J.; Haroz, E. H.; Kittrell, C.; Hauge, R. H.; Smalley, R. E. J. Nanosci. Nanotechnol. 2003, 3, 81. (30) Moore, V. C.; Strano, M. S.; Haroz, E. H.; Hauge, R. H.; Smalley, R. E. Nano Lett. 2003, 3, 1379. (31) Yoon, D.; Kang, S. J.; Choi, J. B.; Kim, Y. J.; Baik, S. J. Nanosci. Nanotechnol. 2007, 7, 3727. (32) Sun, Z.; Nicolosi, V.; Rickard, R.; Bergin, S. D.; Aherne, D.; Coleman, J. N. J. Phys. Chem. C 2008, 112, 10692. (33) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 85th ed.; CRC Press: London, 2004. (34) Han, F.; Li, S.; Yin, R.; Liu, H.; Xu, L. Colloids Surf., A 2008, 315, 210.

DOI: 10.1021/la903382b

12473

Letter

Kang et al.

Figure 3. Characterization of the dielectrophoretically deposited nanotubes from SDS-SWNTs, Pluronic-SWNTs, and CTAB|SDSSWNTs. (a) Radial breathing mode of Raman spectra. (b) Electrical transport data from the same samples investigated by Raman spectroscopy. The source-drain current was measured as a function of the gate voltage.

the band gap energy.7,19 A large value of εp,m = 4000 was assumed for metallic tubes, whereas that of semiconducting nanotubes was found to be small, εp,s =5.7,19 The permittivity of solution, εm, was assumed to be 80. As shown in Figure 2, metallic tubes (R = 50) always have positive dielectrophoresis whereas semiconducting species are subject to either positive or negative dielectrophoresis depending on β, which is consistent with the previous analysis.18,19 For semiconducting types, the real part of the CM factor increased with increasing β. The ionic-surfactant-micellized nanotubes, SDSSWNTs and CTAB-SWNTs, displayed large β values because of the large ζ potentials. Surprisingly, the β value for the nonionic-polymermicellized nanotubes, Pluronic-SWNTs, was 15.7 although the ζ potential was small. This is due to the small medium conductivity as shown in Figure 1a. Equimolar mixtures of anionic and cationic surfactants were more effective than the nonionic polymer in reducing β because the ζ potential of colloidal nanotubes was decreased and the conductivity of the suspending medium was increased. A minimum β value of 1.9 was observed for CTAB| SDS-SWNTs. The dielectrophoretic force is expected to be negligible at this β value, although a negative force cannot be induced for CTAB|SDS-SWNTs. To validate the theoretical analysis, the dielectrophoretic deposition of micellized nanotubes was carried out using a microelectrode. Three samples with high, intermediate, and low β values were selected (i.e., SDS-SWNTs, Pluronic-SWNTs, and CTAB|SDS-SWNTs). Figure 3a shows the radial breathing mode of the Raman spectra for the dielectrophoretically deposited nanotube arrays. The spectra were normalized by the metallic (13, 4) peak, and chiralities for metallic species are displayed in bold italics. As predicted by the CM factor analysis, metallic nanotubes were enriched for all three cases. However, complete separation was observed for CTAB|SDS-SWNTs only. Semiconducting nanotubes were removed when the droplet on the electrode was blown off because the dielectrophoretic force was negligible. Control samples were also prepared by air drying drops of the nanotube suspension on a glass substrate without applying an electric field. Both metallic and semiconducting modes were clearly observed on control samples (Supporting Information). It is interesting that the semiconducting peaks of Pluronic-SWNTs were slightly larger than those of SDS-SWNTs although Pluronic-SWNTs had a smaller β value. This could be due to SWNT bundles remaining in the solution as evidenced by the weak fluorescence. The bundles containing both metallic and semiconducting species would increase the Raman modes for semiconducting nanotubes when deposited. 12474 DOI: 10.1021/la903382b

Electrical transport measurement was also carried out using the identical samples characterized by Raman spectroscopy. The source-drain current was measured as a function of the back gate voltage. CTAB|SDS-SWNTs displayed no gate dependence, indicating that only metallic nanotubes were preferentially deposited from the nanotube suspension. SDS-SWNTs and Pluronic-SWNTs demonstrated typical characteristics of the mixture of metallic and p-type semiconducting species.35 There was current flow at Vg = 30 V, and the current was increased with decreasing gate voltage. The on-off ratio of Pluronic-SWNTs was greater than that of SDS-SWNTs, indicating that more semiconducting nanotubes were deposited for Pluronic-SWNTs. This is consistent with the Raman observation.

Conclusions Dispersing agents such as ionic surfactants and nonionic polymers affect the characteristics of both colloidal nanotubes and suspending solution media. The theoretical analysis of the Clausius-Mossotti factor indicated that the nanotube-to-medium conductivity ratio is a primary factor in determining the dielectrophoretic behavior of semiconducting nanotubes. Positive dielectrophoretic force is induced with the increasing conductivity ratio. Five samples were investigated to control the conductivity ratio. The equimolar surfactant mixture was more effective than the nonionic polymer in reducing the conductivity ratio because the conductivity of colloidal tubes was decreased and that of the medium was increased. Furthermore, the surfactant mixture led to better dispersion of nanotubes. The mixing order was also important because of the different affinity of surfactants to nanotubes. Complete dielectrophoretic separation could be achieved by CTAB|SDS-SWNTs. The theoretical analysis was also supported by the experimental observation. Acknowledgment. This work was supported by a grant (2009K000160) from the Center for Nanoscale Mechatronics & Manufacturing (one of the 21st Century Frontier Research programs), a grant (no. KRF-2008-313-D00085) from National Research Foundation of Korea, and the WCU (World Class University) program (R31-2008-000-10029-0) funded by the Ministry of Education, Science and Technology, Korea. Supporting Information Available: Detailed experimental procedures and additional data including figures. This material is available free of charge via the Internet at http://pubs.acs.org. (35) Martel, R.; Schmidt, T.; Shea, H. R.; Hertel, T.; Avouris, P. Appl. Phys. Lett. 1998, 73, 2447.

Langmuir 2009, 25(21), 12471–12474