ARTICLE IN PRESS

Physica E 37 (2007) 5861 www.elsevier.com/locate/physe

Purication of carbon nanotubes grown by thermal CVD

S. Porroa,, S. Mussoa, M. Vinanteb, L. Vanzettib, M. Anderleb, F. Trottac, A. Tagliaferroa
a

Dipartimento di Fisica, Politecnico di Torino, C.so Duca degli Abruzzi 24, 10129 Torino, Italy b ITC-Irst FCS Division, Via Sommarive 18, 38050, Povo Torino, Italy c ` Dipartimento di Chimica IFM, Universita di Torino, v. Pietro Giuria 7, 10125 Torino, Italy Available online 11 September 2006

Abstract We show the results of a set of purications on carbon nanotubes (CNT) by acid and basic treatments. CNTs were obtained by thermal decomposition of camphor at 850 1C in a CVD growth system, by means of a growth process catalyzed by iron clusters originating from the addition of ferrocene in the precursors mixture. The purication procedures involved HNO3, H2SO4, HSO3Cl and NaOH for different process temperatures. As-grown CNTs showed a consistent presence of metal catalyst (about 6 wt%), evidenced by TGA. The purication treatments led to a certain amount of opening of the CNT tips, with a consequent loss of metal catalyst encapsulated in tips. This is also conrmed by BET analysis, which showed an increase of the surface area density of CNT after the purication. FT-IR and XPS revealed the presence of carboxylic groups on the CNT surface chemically modied by the harsh environment of the purication process. Among the various treatments that have been tested, the 1:3 solution of nitric and sulphuric acid was the most effective in modifying the CNT surface and inducing the formation of functional groups. r 2006 Elsevier B.V. All rights reserved.
PACS: 81.07.De; 81.05.Uw; 79.60.Jv Keywords: Carbon nanotubes; Acid/basic treatments; Functionalization

1. Introduction Carbon nanotubes (CNTs) grown by arc discharge, laser ablation of graphite and chemical vapor deposition (CVD) contain impurities, most notably metal catalyst particles and in some cases unwanted carbonaceous byproducts (such as amorphous carbon, fullerene, etc.) [14]. A CNT grown by catalytic CVD usually keeps the catalyst metal particle that allowed the growth in its tip [5,6]. This could signicantly affect the CNT properties. Metal particles could also be observed in CNT bodies, as a result of the catalytic growth process itself. Hence, the purication is a needed and crucial step in order to open the CNT tips and remove the metal catalyst. This challenge is still far from an exhaustive solution, for the troubling step is to get rid of

the undesired byproducts without damaging the CNT structure [79]. The purpose of this work is to test the reaction of the CNT material grown by the catalytic thermal CVD described in Refs. [10,11], to a series of different purications by acid and basic treatments, in order to open the CNT tips and remove the metal catalyst particles, and lead to the formation of functional groups (such as carboxylic group) which are essential in the formation of covalently bonded adducts of biological molecules with CNTs [12]. 2. Experimental Multi-walled CNTs (MWNT) were grown by thermal CVD at 850 1C in a horizontal quartz tube housed in a cylindrical furnace. An inert gas ow carried the gas mixture of carbon precursor and metal catalyst towards the center of the furnace, where gas pyrolysis led to the deposition of a CNT layer on the substrate. The carbon

Corresponding author. Tel.: +39 011 564 7305; fax: +39 011 564 3799.

E-mail address: samuele.porro@polito.it (S. Porro). 1386-9477/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.physe.2006.07.014

ARTICLE IN PRESS
S. Porro et al. / Physica E 37 (2007) 5861 59

precursor used was camphor, a natural and inexpensive hydrocarbon [6]; iron was the metal catalyst provided by ferrocene precursor, added in solid mixture with camphor before its evaporation [10,11]. The as-prepared MWNT carpet contains iron particles that could signicantly affect the CNT properties. Therefore, metal catalyst particles encapsulated in the CNT closed tips and walls need to be removed, trying to avoid a concurrent damage of the tubular structures. Here we report a brief summary of the purications applied to different samples (the acid/base treatments are detailed in Table 1). 2.1. Acid treatment A nanotube sample was milled and subsequently soaked in acid solution (HSO3Cl or a mixture of HNO3:H2SO4 is 1:3). The CNT/acid mixture (0.15 g CNTs/10 ml acid solution) was stirred for 5 h at 0 1C, room temperature and 70 1C. The treatment was followed by quenching in ice and then neutralization by NaOH water solution. The resulting suspension was ltered and the black powder deposited on the lter was washed with distilled water, then with ethanol and acetone and eventually dried in an oven. The as-collected ltered powder was ready for characterization. 2.2. Basic treatment The base purication and opening reaction is performed by preparing a solid mixture of an excess of NaOH pellets with 1.4 g of MWNTs and heating the mixture up to 700 1C for 1 h in a tubular furnace using a nitrogen ow. After the reaction, the material was extensively washed in diluted HCl and water. It has to be noticed that NaOH is extremely reactive at the used temperature, which can create various drawbacks during the purication process, making this step quite difcult to control. The presence of catalyst and other elements derived from the purication steps was revealed by electron dispersion X-ray (EDX) analysis. Thermo-gravimetric analysis (TGA) gave information about the changes in thermal stability

with the purication treatments. Brunauer-emmet-teller analysis (BET) was used to monitor changes in surface area density and porosity. Fourier transformed infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS) were used to check for the presence of reactive groups on the CNT surface after the chemical modication induced by the harsh acid and basic treatments.

3. Results and discussion Among the various treatments that have been tested, every characterization showed that the 1:3 solution of nitric and sulphuric acid was the most effective in order to modify the CNT surface and induce the formation of functional groups (Fig. 1 reports a SEM image of a sulphonitrated CNT sample). As shown by TGA (Fig. 2a), the pristine MWNTs sample is thermally more stable towards oxidative destruction than the HNO3H2SO4 treated one. In addition, after the sulpho-nitration at room temperature and particularly at 70 1C the main combustion regime is split according to the introduction of oxygen-containing

Fig. 1. SEM image of carbon nanotubes treated with HNO3:H2SO4 is 1:3 at 70 1C for 5 h.

Table 1 Purications details and results from different characterizations (some data are not available) Treatment T (1C) TGA oxidation T (1C) 580 580 570 580 520 500 460 435 EDX (qualitative) BET (m2/g) XPS C (%) C, Fe C, Fe, O (few) 20 60 33 24 n.a. n.a. 48 400 92.6 O (%) 7.4 S (%) n.a. n.a. 0.4 0.6 0.8 0.9 n.a. Cl (%) Na (%) No features No features FT-IR (groups)

As-grown HSO3Cl

HNO3:H2SO4 1:3

NaOH

0 25 70 0 25 70 700

C, Fe, O, S, Na

93.6 90.5 88.3 72.0

5.1 8.7 10.2 20.5

0.9

0.2 0.6 5.1

Carboxylic

n.a.

n.a.

ARTICLE IN PRESS
60 S. Porro et al. / Physica E 37 (2007) 5861

Fig. 2. Weight-derivative curves of thermo-gravimetric analysis on: (a) asgrown and HNO3:H2SO4-treated CNTs; (b) NaOH- and HSO3Cl-treated CNTs.

Fig. 3. XPS plots: (a) samples with different treatments; (b) samples treated with HNO3:H2SO4 at different temperatures.

functional groups on the nanotube surface, such as carboxyl and sulphonic acid groups. On the other hand, in agreement with EDX investigation, thermo-oxidative analysis showed a negligible change between MWNTs before and after HSO3Cl treatment, at any temperature (Fig. 2b). In fact, like pristine MWNTs, the presence of carbon, iron and a little amount of oxygen on HSO3Cl puried samples was found by EDX analysis. Further EDX investigation on thermo-oxidized residuals from TGA showed the presence of iron and oxygen because of the production of Fe2O3. EDX on samples treated with nitric/sulphuric solutions still showed the presence of carbon and iron, but in addition we found a higher relative concentration of oxygen, and the presence of sodium and sulphur, proving that the treatment led to the formation of sodium salt derivatives of both carboxylic and sulphonic groups on the

CNT surface (sodium is due to the NaOH neutralization washing). BET analyses of the acid-treated samples show a slight increase of the surface area density (Table 1), conrming that the purications only partially opened CNT tips and removed the iron from them. Samples treated with NaOH exhibited TGA plots similar to the acid-treated ones (Fig. 2b), even though they started their oxidation at a lower temperature, showing that the reaction induced by the base was more effective. According to this fact, BET analysis showed a large increase of the surface area density (Table 1). In Table 1 is also reported the XPS semi-quantitative estimation of the elements present on the surface of the samples (it has to be noticed that the used experimental setup did not allow the detection of iron). We have used the atomic sensitivity factors for the calculation. Fig. 3a shows

ARTICLE IN PRESS
S. Porro et al. / Physica E 37 (2007) 5861 61

4. Conclusions Acid and basic purications on thermal CVD grown massive carbon nanotubes have been tested. TGA and EDX analyses showed the presence of oxygen-containing functional groups on the surface of nanotubes treated with a HNO3/H2SO4 mixture. FT-IR and XPS revealed the presence of carboxylic groups. As shown by several studies, this fact could allow the chemical adduction of nucleic acids, proteins and other biological molecules, with the aim of CNT application as nanobiological sensors. Acknowledgments We thank A. Chiodoni for SEM and EDX analyses, and C. Galletti and J. Vitillo for BET analyses. This work was accomplished in the framework of Laboratorio di Tecnologie Elettrobiochimiche Miniaturizzate per lAnalisi e la RicercaLaboratory of Miniaturized Electrobiochemical Technologies for Analysis and Research (LaTEMAR), Centre of Excellence funded by MIUR (Italian Ministry for Education, University and Research) Grants FIRB 20032004 for public/private structures involved in research elds characterized by strategic value. References
[1] A. Cao, L. Ci, G. Wu, B. Wei, C. Xu, J. Liang, D. Wu, Carbon 39 (2001) 152. [2] J. Li, Y. Zhang, Physica E 28 (2005) 309. [3] Z.E. Horvath, K. Kertesz, L. Petho, A.A. Koos, L. Tapaszto, Z. Vertesy, Z. Osvath, A. Darabont, P. Nemes-Incze, Z. Sarkozi, L.P. Biro, Curr. Appl. Phys. 6 (2006) 135. [4] M. Zhang, M. Yudasaka, S. Iijima, Chem. Phys. Lett. 336 (2001) 196. [5] M. Pinault, V. Pichot, H. Khodja, P. Launois, C. Reynaud, M. Mayne-LHermite, Nanoletters 5/12 (2005) 2394. [6] M. Kumar, Y. Ando, Chem. Phys. Lett. 374 (2003) 521. [7] P.X. Hou, S. Bai, Q.H. Yang, C. Liu, H.M. Cheng, Carbon 40 (2002) 81. [8] Y.S. Park, Y.C. Choi, K.S. Kim, D.C. Chung, D.J. Bae, K.H. An, S.C. Lim, X.Y. Zhu, Y.H. Lee, Carbon 39 (2001) 655. [9] L.P. Biro, N.Q. Khanh, Z. Vertesy, Z.E. Horvath, Z. Osvath, A. Koos, J. Gyulai, A. Kocsonya, Z. Konya, X.B. Zhang, G. Van Tendeloo, A. Fonseca, J.B. Nagy, Mater. Sci. Eng. C 19 (2002) 9. [10] S. Musso, G. Fanchini, A. Tagliaferro, Diamond Rel. Mater. 14 (2005) 784. [11] S. Porro, S. Musso, M. Giorcelli, A. Tagliaferro, S.H. Dalal, K.B.K. Teo, D.A. Jefferson, W.I. Milne, J. Non-Cryst. Solids 352 (2006) 1310. [12] S.E. Baker, W. Cai, T.L. Lasseter, K.P. Weidkamp, R.J. Hamers, Nanoletters 2/12 (2002) 1413. [13] T.I.T. Okpalugo, P. Papakonstantinou, H. Murphy, J. McLaughlin, N.M.D. Brown, Carbon 43 (2005) 153. [14] Y. Wang, Z. Iqbal, S.V. Malhotra, Chem. Phys. Lett. 402 (2005) 96.

Fig. 4. FT-IR spectrum of carbon nanotubes treated with HNO3:H2SO4 is 1:3 at 70 1C for 5 h.

the C 1s core levels for samples with different treatments. The spectra show a large peak at 284.5 eV (CC/CH bonds), and a smaller shoulder around 286 eV (CO bonds). In addition the sample treated with H2SO4HNO3 shows a peak at 288.5 eV attributed to the CQO group of carboxylic acid groups [12]. The C 1s core level looks exactly the same for the untreated sample and for the one treated with HSO3Cl. Fig. 3b shows the same core level for the samples treated with H2SO4HNO3 at three different temperatures. The peak at higher binding energy is present only in the sample treated at the highest temperature. These results are quite in agreement with the literature [12,13]. Although FT-IR spectra of MWNTs do not generally show strong features, modications in some characteristic peaks allowed us to conrm the presence of carboxylic groups in samples treated with the HNO3H2SO4 mixture [14]. The 1725 cm1 peak in Fig. 4 may correspond to the stretching vibration mode of CQO (carboxylic group); the 1585 cm1 peak to the asymmetric stretching vibration of the anion carboxylate (COO) and stretching of the CQC bond placed by the functional group; the 1415 cm1 peak to the symmetric stretching vibration of the anion carboxylate (COO), probably overlapped with the aromatic CQC signal. Features at 1195/1115 cm1 are not relevant, since they are probably due to some contamination during the sample preparation steps. HSO3Cl treated samples show FT-IR spectra totally similar to the as-grown samples (i.e., with no relevant features), proving that no observable CNT modication occurred.