J. Phys. Chem. B 2006, 110, 12017-12021
12017
Sodium Chloride-Catalyzed Oxidation of Multiwalled Carbon Nanotubes for Environmental Benefit Morinobu Endo,*,† Kenji Takeuchi,† Takeyuki Tajiri,† Ki Chul Park,† Feng Wang,† Yoong-Ahm Kim,† Takuya Hayashi,† Mauricio Terrones,‡ and Mildred S. Dresselhaus§ †Faculty
of Engineering, Shinshu UniVersity, 4-17-1 Wakasato, Nagano 380-8553, Japan, AdVanced Materials Department, IPICyT, AVenida Venustiano Carranza 2425-A, San Luis Porosi 78210, Mexico, and Department of Physics and Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139-4307 ReceiVed: February 19, 2006; In Final Form: April 10, 2006
A sodium chloride (NaCl) catalyst (0.1 w/w %) lowers the oxidation temperature of graphitized multiwalled carbon nanotubes: MWCNT-20 (diameter: 20-70 nm) and MWCNT-80 (diameter: 80-150 nm). The analysis of the reaction kinetics indicates that the oxidation of MWCNT-20 and MWCNT-80 mixed with no NaCl exhibits single reaction processes with activation energies of Ea ) 159 and 152 kJ mol-1, respectively. The oxidation reaction in the presence of NaCl is shown to consist of two different reaction processes, that is, a first reaction and a second reaction process. The first reaction process is dominant at a low temperature of around 600 °C, while the second reaction process becomes more dominant than the first one in a higher temperature region. The activation energies of the first reaction processes (MWCNT-20: Ea1 ) 35.7 kJ mol-1; MWCNT-80: Ea1 ) 43.5 kJ mol-1) are much smaller than those of the second reaction processes (MWCNT20: Ea2 ) 170 kJ mol-1; MWCNT-80: Ea2 ) 171 kJ mol-1). The comparison of the kinetic parameters and the results of the spectroscopic and microscopic analyses imply that the lowering of the oxidation temperature in the presence of NaCl results from the introduction of disorder into the graphitized MWCNTs (during the first reaction process), thus increasing the facility of the oxidation reaction of the disorder-induced nanotubes (in the second reaction process). It is found that the larger nanopits and cracks on the outer graphitic layers are caused by the catalytic effect of NaCl. Therefore, the NaCl-mixed samples showed more rapid and stronger oxidation compared with that of the nonmixed samples at the same residual quantity.
Introduction The easy availability of carbon nanotubes (CNTs) in a large scale ( ∼100 ton/year) through the development of a catalytic chemical vapor deposition (CCVD) method has established new areas of chemistry and physics for nanometer-sized carbon materials.1-6 The geometrical feature of the nanocarbon architectures leads to fascinating physicochemical properties, which make CNTs potentially useful for a wide range of applications.7-11 In fact, CNTs have been applied to technological and bio/ medical fields, where multiwalled carbon nanotubes (MWCNTs) available at the lowest production cost are mainly employed for use as various types of CNTs. For example, a lithium ion battery using MWCNTs as electrode fillers has already been commercialized,12 and a functional catheter13 made from a polymer-MWCNT composite is under in vivo investigation for clinical applications. Furthermore, in recent years, CNTs have begun to be used as supporting materials for noble metal catalysts in fuel-cell electrodes.14,15 Before too long, MWCNTbased energy/electronic devices, polymer/metal composite materials, and bio/medical products are expected to increasingly come into practical use. The expected near-future widespread use of MWCNT-based products will likely create the necessity for an economical and ecologically acceptable waste-treatment * Corresponding author. Tel: +81-26-269-5201. Fax: +81-26-269-5208. E-mail:
[email protected]. † Shinshu University. ‡ IPICyT. § Massachusetts Institute of Technology.
of MWCNTs, so that they can be safely discarded in large quantities. Particularly, the recovery of noble metal catalysts by the removal of the CNT support has become a crucial issue in the field of energy/electronic devices. Incineration is one of the most promising approaches for the waste-treatment of MWCNTs. The complete combustion of MWCNTs can be achieved at a modest temperature of 700800 °C. From the viewpoint of energy costs, however, it is more desirable to further lower the combustion temperature. Thus far, little investigation has been made on any catalytic oxidation procedure to lower the combustion temperature of MWCNTs. We report here the use of sodium chloride (NaCl) as a catalyst for facilitating the oxidation reaction of heat-treated (graphitized) MWCNTs, which normally exhibit a relatively high oxidation resistance. The results of kinetic, spectroscopic, and microscopic analyses are here discussed to clarify the reaction mechanism of the NaCl-catalyzed oxidation of MWCNTs. An accelerated oxidation process using the NaCl catalyst is an important contribution to facilitating the disposal of nanotubes for environmental considerations. Experimental Section In this work, two major kinds of MWCNTs were employed, that is, MWCNT-2016 and MWCNT-8012 prepared by a CCVD method, and were heat-treated at 2600 and 2800 °C under argon (Ar), respectively, to reduce the catalytic metal particles and to increase the crystallinity of the MWCNTs. The diameter, length
10.1021/jp061058o CCC: $33.50 © 2006 American Chemical Society Published on Web 05/26/2006
12018 J. Phys. Chem. B, Vol. 110, No. 24, 2006
Endo et al.
TABLE 1: Basic Properties and Activation Energies of Nonmixed and NaCl-Mixed MWCNTs (MWCNT-20 and MWCNT-80) NaCl (0 w/w %)
NaCl (0.1 w/w %)
sample
diameter (nm)
length (µm)
volume density (g/cm3)
Ea (kJ mol-1)
Ea1 (kJ mol-1)
Ea2 (kJ mol-1)
MWCNT-20 MWCNT-80
20-70 80-150
aspect ratio > 100 length 10-20
0.005 0.04
159 152
35.7 43.5
170 171
or aspect ratio, and volume density of the MWCNT-20 and or MWCNT-80 samples are listed in Table 1. Typically, a given amount of NaCl (0.1-1.0 w/w % for MWCNTs) was dissolved in deionized water. The heat-treated MWCNT-20 or MWCNT80 was added to the NaCl solution. Then, the solvent (water) of the mixture was evaporated on a hot plate at 100 °C for 2 h while stirring, and the materials were then dried in an oven at 90 °C for 24 h. The resulting solid mixture was ground up by an agate mortar and a pestle to give a fine powder. The oxidation behavior of the powdered samples was investigated by thermogravimetric and differential thermal analysis (TG/DTA). The TG/DTA analysis was performed on a Shimadzu DTG-60 measurement system (sample cup: a 60 µL platinum cup; reference sample: R-alumina; measuring accuracy: (1%) over the temperature range from room temperature to 950 °C (heating rate: 10 °C/min) under a flow of 1 vol % oxygen (O2)/argon (Ar) mixed gas (gas-flow rate: 125 mL/min). The structural and morphological variations of the oxidized samples were analyzed by high-resolution transmission electron microscopy (HR-TEM) and Raman spectroscopy. The oxidized samples were obtained in the course of the oxidation process by a rapid cooling in an Ar flow. The HR-TEM images were recorded on a JEOL JEM-2100F instrument operated at an acceleration voltage of 120 kV. The Raman measurements were recorded on a Kaiser Optical Systems Hololab 5000 apparatus (incident excitation: 532 nm (or 2.33 eV) from a Nd:YAG laser).
at a fixed temperature. In this work, the TGA measurements were performed in a non-isothermal process. At a given time t′, the differential coefficient of ∆m/m0 as t f t′ can be regarded as a reaction rate constant k at the temperature corresponding to the time t′. Thus, the reaction rate of the eq 1 can be written as
k ) d(∆m/m0)/dt
(2)
The dependence of the reaction rate constant k on temperature follows the Arrhenius equation:
ln k ) ln A - (Ea/R)/T
(3)
where A is a preexponential factor, Ea is the apparent activation
Results and Discussion Figure 1 shows the TGA profiles of MWCNT-20 and MWCNT-80 samples mixed with NaCl at the indicated MWCNT/ NaCl ratios. The results show that the oxidation process of MWCNT-20 and MWCNT-80 mixed with no NaCl started from around 600 °C, and finished at around 820 °C. In contrast, the MWCNT-20 and MWCNT-80 mixed with NaCl were oxidized at a lower temperature of around 550 °C, and were fully burned off at 700 °C. The oxidation regions of the NaCl-mixed samples were clearly narrower than those of the nonmixed samples. In addition, the decreases observed for the initiation temperatures of the oxidation process dropped off more sharply in the NaClmixed samples than in the nonmixed samples. These results clearly indicate that the presence of NaCl increases the efficiency of the oxidation reaction. Furthermore, the variation in the NaCl ratios caused little difference in the TGA profiles, which means that the oxidation efficiency is not dependent on the concentration of NaCl. Therefore, NaCl is regarded as a catalyst for the oxidation reaction of MWCNTs. The small difference in the catalytic effect that was observed between the MWCNT-20 and MWCNT-80 samples is attributed to their different tubediameters. The oxidation-weight loss of carbons is proportional to time t, which is expressed by the following equation:
∆m/m0 ) kt
(1)
where m0 is the initial weight of the MWCNT-20 (or MWCNT80), ∆m is the weight loss of the MWCNT-20 (or MWCNT80) at time t, and k is a reaction rate constant that is not changed
Figure 1. TGA curves of nonmixed and NaCl-mixed MWCNT-20 (a) and MWCNT-80 (b) for various mixing ratios of MWCNTs/NaCl.
NaCl-Catalyzed Oxidation of MWCNTs
J. Phys. Chem. B, Vol. 110, No. 24, 2006 12019
Figure 2. Arrhenius curves of nonmixed and NaCl-mixed MWCNT-20 (a) and MWCNT-80 (b) calculated from the TGA curves in Figure 1.
(4)
Figure 3. (a) Raman spectra of nonmixed and NaCl-mixed MWCNT80 (HTT: 2800 °C) for various temperatures during the oxidative process, and (b) the variation of R values dependent on the residual quantity of CNTs. The weight per weight percent values in parentheses represent the relative quantity of residues ([residual CNT (g)/initial CNT (g)] × 100).
In the limit of t f t′, the variation in the temperature approaches zero infinitesimally, so that ln A can be regarded as a constant. Thus, an Arrhenius curve is obtained by plotting ln[d(∆m/m0)/dt] versus 1/T. The values of ln[d(∆m/m0)/dt] were calculated by differentiating the time-evolution curves of the weight loss at several arbitrary points in the section between the initiation and the termination of the oxidation process (the time-evolution curves of the weight loss were obtained from the TGA curves and the measuring time). Figure 2 shows the Arrhenius curves of the MWCNT-20 and MWCNT-80 mixed with no NaCl and with 0.1 w/w % of NaCl. The Arrhenius curves of the nonmixed samples showed single slope lines. In contrast, the NaCl-mixed samples provided Arrhenius curves consisting of two different slope lines, resulting in two different values of activation energy, Ea1 and Ea2 (Table 1). This means that the oxidation of the MWCNTs in the presence of NaCl consists of two different reaction processes. The first reaction process with the smaller activation energy (Ea1) is dominant at the lower temperature. Above the isokinetic temperature (Figure 2), the rate of the second reaction process with the larger activation energy (Ea2) exceeds that of the first
one, so that the second reaction process becomes dominant in the higher temperature region. It should be noted that the activation energies of the second reaction processes are nearly comparable to those of the oxidation reactions in the nonmixed samples (Ea2 ≈ Ea). This suggests that the second reaction process proceeds through a reaction pathway analogous to the oxidation of the MWCNTs themselves (i.e., the oxidation of graphitic carbons). Therefore, the lowering of the oxidation temperature would be attributed to the modification of the graphitized MWCNTs to a more easily oxidized form by the first reaction process. Figure 3 shows the variation in the relative integrated intensities, R ) ID/IG, of the D and G bands obtained from the Raman spectra of the MWCNT-80 mixed with no NaCl and with 0.1 w/w % of NaCl. All the comparisons between the nonmixed and NaCl-mixed samples were made for those recovered at the temperature giving the nearly same weight loss. The change of the R value observed for the nonmixed sample was small in the low-temperature region, and then the R value became nearly constant in the higher temperature region. In contrast, the R value of the NaCl-mixed sample started to
energy of MWCNT oxidation in kJ mol-1, T is the absolute temperature in K, and R is the gas constant in J mol-1 K-1. Substitution of eq 2 into eq 3 gives the following eq 4:
ln[d(∆m/m0)/dt] ) ln A - (Ea/R)/T
12020 J. Phys. Chem. B, Vol. 110, No. 24, 2006
Endo et al.
Figure 4. (a) HR-TEM images (low magnification) of nonmixed and NaCl-mixed MWCNT-80 at the oxidative process (residual quantity of CNTs at TG/DTA measurement: (a,d) 92.7, 95.7%; (b,e) 80.7, 82.2%, (c,f) 24.8, 16.3%). (b) HR-TEM images (high magnification) of nonmixed and NaCl-mixed MWCNT-80 at the oxidative process.
drastically increase in the low-temperature region and was dominated by the first reaction process. This implies that the first reaction process induces disorder in the graphitic crystallites of the MWCNTs, and this disorder leads to the facile oxidation of MWCNTs in the second reaction process. As can be seen in Figure 1, the temperature regions of the oxidation process for the NaCl-mixed samples are very close to those of the as-grown (nongraphitized) MWCNT samples. This fact also supports our interpretation of an introduction of disorder into the graphitized MWCNT samples by the first reaction process. Figure 4 shows HR-TEM images of the nonmixed and NaClmixed MWCNTs-80 oxidized at a temperature giving nearly the same values of the relative residual quantities (a,d: approximately 92.7, 95.7 w/w %; b,e: 80.7, 82.2 w/w %; c,f: 24.8, 16.3 w/w %). In the nonmixed samples, the observed structural change in the graphitic layers becomes prominent with the progress of the oxidation reaction (at a higher temperature region). In contrast, the NaCl-mixed samples show clear pits and even cracks in the lower temperature region. The structural change of the multigraphitic layers in the low-temperature region ensures that the introduction of the disorder is caused by the first reaction process. Figure 5 shows the DTA profiles of the MWCNT-20 and MWCNT-80 mixed with NaCl at various ratios. The nonmixed samples exhibited a single exothermic peak in the temperature range of 760-780 °C, which originates from the single reaction
mechanism of graphitized-carbon oxidation. In contrast, the NaCl-mixed samples show two exothermic peaks that vary according to the ratios of NaCl. An increase in the NaCl ratio causes an increase in the exothermic peaks at around 690 °C, accompanied by a decrease in the exothermic peaks at around 660 °C. The sharp exothermic peaks at the higher temperature are identified as originating from the fast reaction rate of the second oxidation process, the increasing intensity of which would be attributed to the easier oxidation of (disorder-induced) MWCNTs with an increase in the NaCl ratios. The exothermic peaks in the lower temperature region have shifted to an endothermic direction with increasing NaCl ratios. As seen in Figure 2, the reaction rate of the second oxidation process in the lower temperature region is much slower than that of the first reaction process, which is the induction process of disordered carbons by NaCl. Therefore, the endothermic shift with the increase in NaCl ratios would result from the increasing consumption of the oxidation reaction heat (of the second oxidation process) by the disorder-induction process. The oxidation temperature of the graphitized MWCNT-20 and MWCNT-80 has been lowered by the NaCl catalyst. The reaction kinetic analysis indicates that the oxidation process of MWCNT-20 and MWCNT-80 mixed with no NaCl exhibits a single reaction process with activation energies of 159 kJ mol-1 (MWCNT-20) and 152 kJ mol-1 (MWCNT-80). In contrast, the oxidation reaction of the MWCNT-20 and MWCNT-80
NaCl-Catalyzed Oxidation of MWCNTs
J. Phys. Chem. B, Vol. 110, No. 24, 2006 12021 spectroscopic analyses indicate that the first reaction process induces disorder into the graphitized MWCNT-20 and MWCNT80, whereby more easily oxidized MWCNTs could be formed in the lower temperature region dominated by the first reaction process. It is therefore proposed that the lowering of the oxidation temperature in the presence of NaCl results from the introduction of disorder into the graphitized MWCNTs (in the first reaction process) and the facilitation of the oxidation of disorder-induced tubes (in the second reaction process). The difference in the NaCl-catalytic effect observed between the MWCNT-20 and MWCNT-80 samples is attributed to their different tube diameters. The more effective catalysis of NaCl for the MWCNT-20 with smaller diameters would be based on the more facile induction of disorder in the first reaction process. It is found that the larger nanopits and greater number of cracks on the outer graphitic layers are caused by the catalytic effect of NaCl. Therefore, the NaCl-mixed samples showed more rapid and stronger oxidation compared with that of the nonmixed samples at the same residual quantity. Conclusion This work has clearly demonstrated the effectiveness of NaCl as a catalyst for the oxidation of well-graphitized MWCNTs that exhibit a relatively high oxidation resistance. From the viewpoint of material costs, this approach has the advantage that the NaCl catalyst required for the large-scale treatment of MWCNT-based products could be available from abundant seawater, making the practical use of this approach very appealing. Acknowledgment. This work was supported by the CLUSTER of the Ministry of Education, Culture, Sports, Science and Technology of Japan. M.S.D. acknowledges support from NSF DMR-04-05534. References and Notes
Figure 5. DTA curves for nonmixed and NaCl-mixed MWCNT-20 (a) and MWCNT-80 (b) showing a dependence on the mixing ratio of MWCNTs and NaCl.
mixed with NaCl consists of two different reaction processes (the first and the second reaction processes). The activation energies of the first reaction processes (MWCNT-20: Ea1 ) 35.7 kJ mol-1; MWCNT-80: Ea1 ) 43.5 kJ mol-1) are much smaller than those of the second reaction processes (MWCNT20: Ea2 ) 170 kJ mol-1; MWCNT-80: Ea2 ) 171 kJ mol-1). The first reaction process is dominant at a low temperature of around 600 °C because of the lower activation energy and the faster reaction rate compared to those of the second reaction process. In the higher temperature region, however, the second reaction process was dominant, and the reaction rates become much faster than those of the first reaction processes. The comparable activation-energy values of the second reaction process and the nonmixed sample oxidation (i.e., Ea2 ≈ Ea) imply that the second reaction process proceeds through a single reaction mechanism analogous to the oxidation of the graphitized MWCNTs themselves. The results of the TEM and Raman
(1) Oberlin, A.; Endo, M.; Koyama, T. J. Cryst. Growth 1976, 32, 335. (2) Tibbetts, G. G. J. Cryst. Growth 1984, 6, 632. (3) Endo, M. CHEMTECH 1988, 18, 568. (4) Terrones, M.; Grobert, N.; Olivares, J.; Zhang, J. P.; Terrones, H.; Kordatos, K., Hsu, W. K.; Hare, J. P.; Townsend, P. D.; Prassides, K.; Cheetham, A.; Kroto, H. W.; Walton, D. R. Nature 1997, 388, 52. (5) Endo, M.; Muramatsu, H.; Hayashi, T.; Kim, Y. A.; Terrones, M.; Dresselhaus, M. S. Nature 2005, 433, 476. (6) Maruyama, S.; Kojima, R.; Miyauchi, Y.; Chiashi, S.; Kohno, M. Chem. Phys. Lett. 2002, 360, 229. (7) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. Science of Fullerenes and Carbon Nanotubes; Academic Press: San Diego, CA, 1996; p 756. (8) Tans, S. J.; Verschueren, A. R. M.; Dekker, C. Nature 1998, 393, 49. (9) Modi, A.; Koratkar, N.; Lass, E.; Wei, B.; Ajayan, P. M. Nature 2003, 424, 171. (10) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Science 2002, 297, 787. (11) Zhang, M.; Fang, S.; Zakhidov, A. A.; Lee, S. B.; Aliev, A. E.; Williams, C. D.; Atkinson, K. R.; Baughman, R. H. Science 2005, 309, 1215. (12) Endo, M.; Kim, Y. A.; Hayashi, T.; Nishimura, K.; Matsuda, T.; Miyashita, K.; Dresselhaus, M. S. Carbon 2001, 39, 1287. (13) Endo, M.; Koyama, S.; Matsuda, Y.; Hayashi, T.; Kim, Y. A. Nano Lett. 2005, 5, 101. (14) Kim, C.; Kim, Y. J.; Kim, Y. A.; Yanagisawa, T.; Park, K. C.; Endo, M.; Dresselhaus, M. S. J. Appl. Phys. 2004, 96 (10), 5903. (15) Serp, P.; Corrias, M.; Kalck, P. Appl. Catal., A 2003, 253, 337. (16) Kim, Y. A.; Hayashi, T.; Endo, M.; Kaburagi, Y.; Tsukada, T.; Shan, J.; Osato, K.; Tsuruoka, S. Carbon 2005, 43, 2243.
