Time-Dependent Study of the Exfoliation Process of Carbon Nanotubes in Aqueous Dispersions by Using UV-Visible Spectroscopy

Transcription

1 Anal. Chem. 2005, 77, Time-Dependent Study of the Exfoliation Process of Carbon Nanotubes in Aqueous Dispersions by Using UV-Visible Spectroscopy Nadia Grossiord,, Oren Regev, Joachim Loos,, Jan Meuldijk,, and Cor E. Koning*,, Laboratory of Polymer Chemistry, Laboratories of Polymer Technology and Materials and Interface Chemistry, Process Development Group, Technical University of Eindhoven, Postbus 513, 5600 MB Eindhoven, The Netherlands, and Laboratory of Chemical Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel, and Dutch Polymer Institute, P.O. Box 902, NL-5600 AX Eindhoven In this paper we demonstrate that the sonication-driven exfoliation of aggregates and bundles of single-wall carbon nanotubes (SWNTs) in an aqueous surfactant solution can be easily monitored by UV-vis spectroscopy. The different stages of the exfoliation process were directly visualized by cryogenic temperature transmission electron microscopy, showing an excellent correspondence with the spectroscopic data: the maximum achievable exfoliation (which does not mean that 100% of the NTs are effectively exfoliated) corresponds to the maximum UV-vis absorbance of the NT solution. Moreover, it has been observed that NTs produced by the arc-discharge technology (Carbolex NTs) require less energy to achieve maximum exfoliation than NTs produced by chemical vapor deposition (HiPCO NTs). This difference is attributed to weaker van der Waals attraction between Carbolex NTs in the bundles and aggregates. Both single-wall nanotubes (SWNTs) and multiwall nanotubes (MWNTs) have outstanding mechanical, thermal, and electrical properties 1 that make them suitable for a wide range of applications. 2,3 SWNTs behave like rolled-up cylinders of graphene sheets of carbon atoms bonded by sp 2 hybrid orbitals, closed at both ends by semispherical caps that could be joined together to form a fullerene. MWNTs consist of sets of concentric shells, each of which resembles an SWNT. Thanks to the delocalization of the π-electrons, NTs can theoretically be described as one-dimensional conductors. Their structure is uniquely defined by the socalled chiral vector, which determines the orientation of the graphene sheet compared to the NT axis, and thus governs the properties of the NTs, i.e., distinction between metallic and * To whom correspondence should be addressed. Phone: Fax: c.e.koning@tue.nl. Laboratory of Polymer Chemistry, Technical University of Eindhoven. Dutch Polymer Institute, Eindhoven. Ben-Gurion University of the Negev. Laboratories of Polymer Technology and Materials and Interface Chemistry, Technical University of Eindhoven. Process Development Group, Technical University of Eindhoven. (1) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. C. Science of Fullerenes and Carbon Nanotubes; Academic Press: London, (2) Rouse, J. H.; Lillelei, P. T. Nano Lett. 2003, 3, (3) Vigolo, B.; Poulin, P.; Lucas, M.; Launois, P.; Bernier, P. J. Appl. Phys. Lett. 2002, 81, semiconductive NTs. 4 One important application of nanotubes (NTs) is their incorporation into a matrix, e.g., a polymer, to obtain conductive nanocomposites. 5,6 Since lower loadings of SWNTs are required to form a conductive network in a composite, dispersions of SWNTs in an electrically insulating matrix are often preferred over MWNTs to obtain conductive materials. The main bottleneck for the incorporation of NTs into nanocomposites is that as-produced NTs are held together in bundles of individual tubes by very strong van der Waals interactions, estimated at 500 ev/µm of tube length. 7 Therefore, the individual tubes tend to remain bundled even if attempts are made to disperse them. The achievement of stable dispersions of SWNTs in water is a significant challenge and can be a prerequisite for their further application, such as the incorporation into a polymer matrix. 8 The aim is then to obtain conductive plastics, which are cheaper, lighter, and easier to process compared to metals. In this case, the degree of NT aggregation significantly influences the properties of the composite obtained: the more exfoliated the NTs, the lower the NT concentration required to get conductive composites, on the basis of the formation of a percolation network. One way to achieve the exfoliation of SWNTs is to sonicate them in an aqueous surfactant solution. 9 During sonication, the provided mechanical energy can indeed overcome the van der Waals interactions in the SWNT bundles and lead to NT exfoliation. Simultaneously, surfactant molecules can adsorb on the surface of the SWNTs. As a result, a dispersion of NTs with adsorbed surfactant molecules is obtained. Colloidal stability originates from electrostatic and/or steric repulsion. Islam et al. 9 observed that the colloidal stability of these dispersions is maintained for several months. (4) Dresselhaus, M. S.; Dresselhaus, G.; Saito, R. Carbon 1995, 33, (5) Sandler, J.; Shaffer, M. S. P.; Prasse, T.; Bauhofer, W.; Schulte, K.; Windle, A. H. Polymer 1999, 40, (6) McCarthy, B.; Coleman, J. N.; Czerw, R.; Dalton, A. B.; in Het Panhuis, M.; Maiti, A.; Drury, A.; Bernier, P.; Nagy, J. B.; Lahr, B.; Byrne, H. J.; Carroll, D. L.; Blau, W. J. J. Phys. Chem. B 2002, 106, (7) Thess, A.; Lee, R.; Nikolaev, P.; Dai, H.; 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, (8) Regev, O.; El Katy, P. N. B.; Loos, J.; Koning, C. E. Adv. Mater. 2004, 16, (9) Islam, M. F.; Rojers, E.; Bergey, D. M.; Johnson, A. T.; Yodh, A. G. Nano Lett. 2003, 3, /ac050358j CCC: $ American Chemical Society Analytical Chemistry, Vol. 77, No. 16, August 15, Published on Web 07/20/2005

2 Although sonication of SWNTs in aqueous surfactant solutions is widely used, the mechanism of the dispersing process is poorly understood. Strano et al. 10 studied the dispersion of SWNTs in aqueous solutions containing sodium dodecyl sulfate (SDS) as surfactant. These authors claim that the dispersion mechanism consists of the formation of gaps at the bundle ends induced by the ultrasonic processing. These vacancies then allow surfactant adsorption and propagate along the bundle length. This unzippering mechanism finally leads to the release of an isolated, surfactant-coated NT into the continuous phase, e.g., water. Strano et al. also reported that, during the ultrasonic processing, there is dynamic equilibrium between free individual SWNTs and SWNTs aggregated in bundles. In other words, the exfoliation of NTs during sonication is never complete, i.e., solutions containing exclusively individual NTs do not exist. Jiang et al. 11 studied the same system and determined that, below ph values of 8, the hydrophibic hydrocarbon chains of SDS interact with the NT walls. The negatively charged sulfate groups provide electrostatic repulsion, and thus prevent aggregation. The exact way in which surfactant molecules organize on the NT surface is still unclear. Three main points of view about the role of surfactants in dispersing SWNTs have been reported: structureless random adsorption on the NTs with no preferential arrangement of the head and tail, 12 hemimicellar adsorption on the SWNT surface, 13 and encapsulation of the NTs in a cylindrical surfactant micelle. 14 All kinds of NTs are active in the UV-vis region and exhibit characteristic bands corresponding to additional absorption due to 1D van Hove singularities. Spectra obtained are characterized by a fine structure of bands caused by superimposed peaks for NTs of different diameters and chiral vectors. 15 Contrary to completely dispersed NTs, bundled NTs are hardly active in the wavelength region between 200 and 1200 nm. Therefore, it is possible to detect individual NTs via this spectroscopic technique, which implies that there is a relationship between the concentration of NTs individually suspended in solution and the intensity of the corresponding absorption spectrum. 11 However, this UVvis method has only been used to quantitatively characterize the colloidal stability of NT dispersions. Jiang et al. 11 have indeed used this tool to check the colloidal stability of SWNTs dispersed in aqueous SDS solution, whereas Lou et al. 16 demonstrated, by using this spectroscopic technique, that dispersions of polystyrenegrafted MWNTs in toluene can be stable for several weeks. The new contribution of the current paper is the demonstration that this simple technique is a promising tool to monitor the dynamics of the exfoliation process, allowing the determination of the optimal exposure time to ultrasound. (10) Strano, M. S.; Moore, C. M.; Miller, M. K.; Allen, M. J.; Haroz, E. H.; Kittrel, C.; Hauge, R. H.; Smalley, R. E. J. Nanosci. Nanotechnol. 2003, 3, (11) Jiang, L.; Gao, L.; Sun, J. J. Colloid Interface Sci. 2003, 260, (12) Yurekli, K.; Mitchell, C. A.; Krishnamoorti, R. J. Am. Chem. Soc. 2004, 126, (13) Richard, C.; Balavoine, F.; Schultz, P.; Ebbesen, T. W.; Mioskowski, C. Science 2003, 300, (14) Matarredona, O.; Rhoads, H.; Li, Z.; Harwell, J. H.; Balzano, L.; Resasco, D. E. J. Phys. Chem. B 2003, 107, (15) Ryabenko, A. G.; Dorofeeva, T. V.; Zvereva, G. I. Carbon 2004, 42, (16) Lou, X.; Detrembleur, C.; Sciannamea, V.; Pagnoulle, C.; Jérôme, R. Polymer 2004, 45, In this paper, the dispersion process of several types of NTs is monitored and quantified by using UV-vis spectroscopy. The results are complemented by the cryogenic temperature transmission electron microscopy (cryo-tem) direct imaging method. EXPERIMENTAL SECTION Materials. Two types of SWNTs have been studied, namely, Carbolex NTs provided by Carbolex Inc. and HiPCO NTs from Carbon Nanotechnology Inc. Carbolex NTs (batch number CLAP 8510) are produced by arc-discharge technology and contain about 30 wt % impurities composed of both carbon impurities and nickel/ yttrium catalyst particles. HiPCO NTs have been produced by a modified gas process based on chemical vapor deposition. The batch (PO 257) contains around wt % impurities, composed, according to the manufacturer, of 5 wt % small iron catalyst particles encased in carbon shells and, in some cases, also in the NTs themselves. The surfactant used for the dispersion of the NTs is sodium dodecyl sulfate (SDS; 90%) provided by Merck Chemical Co. All experiments were carried out with demineralized water. Instrumentation. All sonication processes have been carried out with a horn sonicator (Sonic Vibracell VC750) with a cylindrical tip (10 mm end cap diameter). The frequency is fixed at 20 khz ( 200 Hz. UV-vis absorption spectra were recorded by a Hewlett-Packard 8453 spectrometer operating between 200 and 1100 nm. For cryogenic temperature transmission electron microscopy, the samples were prepared by using a vitrification robot (Vitrobot, FEI) in which the relative humidity was kept close to saturation. A3µL drop of the NT solution was placed on a carbon-coated lacey substrate supported by a TEM 300 mesh copper grid (Ted Pella). After automatic blotting, the grid was rapidly plunged into liquid ethane at its melting temperature. This resulted in a vitrified film. The vitrified specimen was then transferred under a liquid nitrogen environment to a cryoholder (model 626, Gatan Inc., Warrendale, PA) into the electron microscope, Tecnai 12 (FEI), operating at 120 kv by a nominal under focus of 2-4 µm. The working temperature was kept below -175 C, and the images were recorded on a Gatan 794 MultiScan digital camera and processed with Digital Micrograph, version 3.6. Procedure. For each experiment, 0.5 wt % SWNTs was mixed with 20 ml of an aqueous solution containing 1 wt % SDS. The resulting mixture was then sonicated for different times under mild conditions, i.e., at a power of 20 W. Samples were taken regularly during the sonicating process and diluted by a factor of 30, resulting in an NT content of wt %, and UV spectra were recorded. Quartz cuvettes were employed, and the blank used as reference for the measurements was a 1 wt%sds solution diluted by a factor of 30, under the same conditions as the samples themselves. RESULTS AND DISCUSSION First of all, it had to be verified whether NT exfoliation effectively and gradually occurs when dispersions of SWNTs in aqueous surfactant solutions are sonicated. The first observation was made with the bare eye: a longer sonication time resulted in a darker NT solution (cf. Figure 1), which is an indication that more and more NTs are exfoliated and dispersed in the aqueous 5136 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

3 Figure 1. Evolution of the color of the wt % Carbolex NT solutions as a function of the sonication energy/time of sonication: sample a, without sonication; b, after 20 s of sonication, which corresponds to an energy input of about 360 J; c, 40 s; d, 60 s; e, 80 s; f, 100 s; g, 120 s; h, 140 s; i, 160 s; j, 200 s; k, 240 s; l, 300 s; m, 360 s; n, 420 s. The evolution of the color of aqueous solutions of HiPCO NTs was similar, but the time scale is different and the evolution slower. In this case, samples had to be taken on average every 10 min (about every J of energy delivered) of sonication to get a comparable evolution of color. Figure 2. Aqueous SDS solution of HiPCO SWNTs after 40 min of sonication at 20 W, corresponding to a total energy input of ca J. The white arrow indicates the location of a high degree of aggregation and bundling. The color of the solution is light gray. Figure 3. Aqueous SDS solution of HiPCO NTs after 130 min of sonication at 20 W, corresponding to a total energy input of ca J. Mostly individual NTs are observed. The color of the solution is dark gray, almost black. For this image the same SWNT dispersion was used as for Figure 2. phase. After a certain sonication time, that is to say, after about 5 min for Carbolex NTs (sample l shown in Figure 1) and after ca. 90 min for HiPCO NTs, no significant change of color of the solution could be observed anymore. Cryo-TEM was employed to obtain direct imaging of the NT clusters in solution. Cryo-TEM indeed provides information about the NTs being bundled or exfoliated for different sonication times. Two samples of HiPCO NTs (0.5 wt % concentration) were taken after 40 and 130 min of sonication (corresponding to energy inputs of and J, respectively). The first sample, see Figure 2, was taken when the color of the solution was light gray, i.e., when the color was still evolving. The second sample, see Figure 3, is darker and was taken after completion of the experiment. The cryo-tem micrograph in Figure 2 still shows clear aggregates of SWNTs. Most of the NTs are still bundled. Some of the NTs are already exfoliated (thin lines). The small black dots are catalyst particles. However, after 130 min of sonication, such aggregates and bundles had almost completely disappeared, leaving mostly individual SWNTs; see Figure 3. Realizing that cryo- TEM images only show a very small part of the entire sample, we performed an additional combined scanning electron microscopy (SEM) and atomic force microscopy (AFM) study (to be published in a separate paper), for which we developed a special sample preparation technique leaving the exfoliation state of the tubes unaffected. These images, providing information on a larger part of the SWNT dispersion, convinced us that the presented cryo-tem images are representative for the entire sample. However, cryo-tem data can only be interpreted qualitatively by comparing images. Once it had been verified that the SWNT exfoliation really occurs when the mixture of SWNTs and surfactant molecules are sonicated in water, it had to be checked whether UV-vis spectroscopy can effectively be used to monitor the exfoliation process in time. The same Carbolex and HiPCO NT solutions as those studied previously were examined with UV-vis spectroscopy by following the procedure described earlier in this paper. The UV-vis spectra recorded for aqueous HiPCO SWNT dispersions, obtained after different energy inputs and sonication times, are given in Figure 4. The corresponding spectra for Carbolex SWNTs show a similar development, but exhibit one maximum instead of two around nm (not shown). This indicates that the UV-vis spectra obtained are specific for the NT type studied. This could be speculated since the Carbolex and HiPCO NTs studied have different characteristics in terms of chirality and diameter. For both types of SWNTs, the absorbance gradually decreases from UV to near-ir, similarly to the absorption spectrum reported by Analytical Chemistry, Vol. 77, No. 16, August 15,

4 Figure 4. Evolution of the UV-vis spectra of aqueous SDS/HiPCO NT solutions as a function of sonication time, i.e., as a function of the energy delivered to the aqueous NT suspension ( wt % HiPCO solution, continuous power of sonication of 20 W). The first measurement (spectrum a, 0 J) was carried out before the beginning of the sonication. The second spectrum (b) corresponds to a sample sonicated for 10 min at 20 W. The third one (c) is for a sample sonicated for 40 min, which has been imaged by cryo-tem (cf. Figure 2). The fourth spectrum (d) corresponds to a 130 min sonicated NT solution, which has also been studied with cryo-tem (cf. Figure 3). Jiang et al. 11 This is partly due to scattering, especially in the lower wavelength range. During sonication, the increasing amount of exfoliated NTs results in an increasing area below the lines representing the absorbance (cf. Figure 4). This trend ceases at some point during the sonication process, that is to say, after approximately 5 min for Carbolex NTs and after min for HiPCO NTs. The determination of the area under the UV-vis spectra recorded is time-consuming and not very easy, since our quartz cuvettes do not allow absorbance measurements in wavelength ranges below 190 nm. As a consequence, the missing part of the spectrum at the lowest wavelengths would have to be estimated to enable the construction of a baseline and finally to allow the calculation of an approximate value of the area. Since the relative evolution of the area under the spectrum during the sonication process is proportional to the relative evolution of the absorbance value at a specific wavelength, it was decided to determine the absorbance at several wavelengths and to plot these as a function of the total energy supplied to the solution. Jiang et al., 11 as well as Lou et al., 16 also reported that the value of the UV absorbance at a certain wavelength is a proper quantity to study the colloidal stability of NT suspensions. Please note that the exact and absolute value of the absorbance at a certain wavelength corresponds to the superposition of different electronic transitions of several kinds of NTs, and since scattering takes place, it is not possible to attribute one peak of the spectrum to a specific kind of NT in the range of wavelengths considered. Since the power of sonication is kept constant throughout the experiments, there is a direct relationship between a specific sonication time and the energy delivered to the sample during this same time interval. Therefore, it is equivalent to plot the absorbance at a certain wavelength versus the time of sonication or versus the energy supplied to the solution. Figure 5 shows the absorbance at different wavelengths as a function of the energy supplied to the aqueous HiPCO/SDS solution. Figure 5. Evolution of the value of the absorbance at different wavelengths for an aqueous 0.5 wt % HiPCO NT solution, containing 1 wt % SDS, diluted 30 times. Please note that the dilution factor (30 times) has been chosen to ensure that all the UV absorption values stay below 1 so that the error implied by the measurement itself is reduced. The error induced by the sampling and the accuracy of the measurements of the UV spectra is about 0.03 au (value of the standard deviation). The general trend of the UV absorbance versus total energy curves obtained for aqueous Carbolex/SDS solutions (not shown) is very similar to that of the HiPCO curve, given in Figure 5. For both SWNT dispersions, after an increase at the beginning of the sonication process, the value of the absorbance reaches a plateau value after energy inputs of ca J for the HiPCO SWNTs and ca J for the Carbolex SWNTs, irrespective of the chosen wavelength. This is in agreement with what can be expected: the absorbance increases at the beginning of the sonication process, when the exfoliation is occurring. The leveling off and the ultimate highest limit of the absorbance, which follows the initial increase, correspond to the maximum achievable degree of exfoliation of the NTs. Although the shapes of the UV absorbance versus total energy input curves, as well as the plateau value, for the Carbolex and HiPCO systems are very similar, there is an important and interesting difference: Carbolex NTs demonstrate to exfoliate at a much higher rate than HiPCO NTs. Obviously, only 5000 J is necessary to reach the maximum degree of exfoliation, whereas almost J is required for HiPCO NTs to achieve the same. This result implies that HiPCO NTs exhibit stronger van der Waals attraction when bundled than Carbolex NTs. This different behavior might partly stem from the fact that Carbolex NTs contain more impurities. Moreover, Carbolex NTs are interconnected via catalyst particles, in a spiderlike structure. These particles, as well as other carbon impurities, can be present between the NTs in the bundles, decreasing the contact area between NTs in comparison with cleaner NTs, i.e., HiPCO. This, in turn, results in a weaker interaction between the NTs in the bundles, 15 and should lead to a faster exfoliation. CONCLUSIONS To conclude, in this paper we report a simple and quick UVvis spectroscopic technique to monitor the sonication-driven exfoliation of SWNTs in aqueous surfactant solutions. The technique is reproducible, and semiquantitative. It has been 5138 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5 demonstrated with cryo-tem that the maximum achievable exfoliation (which does not mean that 100% of the NTs are effectively exfoliated) corresponds to the maximum UV-vis absorbance of the NT solution. Moreover, it has been observed that NTs produced by the arc-discharge technology (Carbolex NTs) require less energy to achieve maximum exfoliation than NTs produced by chemical vapor deposition (HiPCO NTs). This difference is attributed to weaker van der Waals attraction between Carbolex NTs in the bundles and aggregates. Furthermore, UV-vis spectroscopy enables one to determine when the exfoliation of the NTs is complete, i.e., at its equilibrium value, and thus when the sonicating process can be stopped. In experiments, it is indeed crucial to sonicate to bring the NTs to a maximum degree of exfoliation. However, since an excessive energy input will damage the SWNTs, 17 the treatment must be stopped at that point. That is why it is also important to verify whether the NTs are not significantly damaged by the sonication (17) Lu, K.; Lago, L., M. R.; Chen, Y. K.; Green, M. L. H.; Harris, P. J. F.; Tsang, S. C. Carbon 1996, 34, treatment when the maximum exfoliation is achieved. Preliminary results with Raman spectroscopy indicate that the exfoliated NTs studied here are not significantly damaged by the sonication treatment applied. This study is still in progress. ACKNOWLEDGMENT We are thankful to the Dutch Polymer Institute (DPI) for the financial support of DPI Project No We thank Professor Satish Kumar of the Georgia Institute of Technology for his help and useful suggestions. We acknowledge Johan van de Sande for useful advice related to the experimental part, as well as Professor Rene Janssen, Dr. Stefan Meskers, and Dr. Peter Bobbert for helpful discussions. Special thanks are also due to Jurriën Mans for carrying out part of the experiments. Received for review March 1, Accepted June 8, AC050358J Analytical Chemistry, Vol. 77, No. 16, August 15,