Plasma Fluorination of Highly Ordered Pyrolytic Graphite and Single Walled
Carbon Nanotube Surfaces
Anders Barlow, Anthony Birch, Alec Deslandes and Jamie S. Quinton*
Smart Surface Structures Group
School of Chemistry, Physics, and Earth Sciences, Flinders University, GPO Box 2100 Adelaide 5001, SA, Australia
Corresponding Author, Email: [email protected]
Abstract–A simple room temperature fluorination process for Highly Ordered Pyrolytic Graphite (HOPG) and Single Walled Carbon Nanotube (SWCNT) surfaces is presented and discussed.
Graphite and SWCNT surfaces are functionalised using radio frequency SF6 plasma. Samples were characterised before and after treatment using XPS, EDX, SEM, and STM techniques. It is found that functionalisation of SWCNT surfaces occurs rapidly indicating an efficient fluorination process with minimal damage to the carbon surfaces, and furthermore, without presence of sulphur-containing functional groups.
Keywords–carbon nanotubes; fluorination; graphite; CNT; HOPG;
plasma surface modification I. INTRODUCTION
The effect of fluorination on the properties of carbon surfaces have been well investigated with F2, HF, and CF4 as the source species [1-4]. However, treatment of carbon nanotubes (CNTs) with chemical species containing a higher fraction of fluorine such as sulphur hexafluoride (SF6) do not appear to have been widely investigated and have only recently been reported by Plank et al. [1].
Carbon surfaces are of interest for a number of applications. In particular, Highly Ordered Pyrolytic Graphite (HOPG) has proven to be extremely versatile. The surfaces are easy to prepare, chemically and thermally stable, and have atomically flat surfaces with large terraces. HOPG serves as an ideal, model substrate for glassy carbon, which is of interest for novel biosensor supports, and carbon nanotubes.
CNTs have developed much interest in the nanotechnological realm since their discovery in 1991 by Sumio Iijima [5]. They display remarkable electronic and mechanical properties, and have applications in nanoelectronics, molecular wires and diodes. However, for useful devices to be fabricated, the ability to handle and control nanotube properties are paramount. Functionalisation of CNTs enhances their solubility (through reduction in their tendency to aggregate into bundles), field emission (through modification of their electronic band structure), and capability for surface attachment (through the presence of active chemical groups that can subsequently bond to active surface species).
Many potential applications have been foreshadowed for fluorinated carbon nanotubes (F-CNTs). F-CNTs are known to increase resistivity with fluorination, due to C-F bond formation resulting in a reduction in electron occupation in states just below the Fermi level [6]. The ability to control the resistivity of carbon nanotubes will enhance their versatility for future-generation electronic devices. F-CNTs may, for example, replace the fluorographite cathodes that are currently in use in lithium batteries. Lithium cells with F-CNTs for cathodes have been shown to have a greater cell potential than a conventional lithium cell with a fluorographite cathode [7].
II. EXPERIMENTAL
A. Preparation of Carbon Surfaces
HOPG (10 x 10 mm, ZYH grade) samples were obtained from Coherent Scientific. Substrate preparation for plasma functionalisation involved simply cleaving the graphite using adhesive tape to expose a fresh layer.
Carbon nanotubes (purified and –COOH functionalised P3- SWCNTs; and purified but unmodified RFP-SWCNTs, with 80-90% purity) were obtained from Carbon Solutions.
The aluminium substrates used were 99.999% pure and polished using SiC abrasive paper, and then etched in an ultrasonic bath of acetic acid (pH ~ 3) for 2 hours at room temperature to remove residual amorphous carbon impurities.
The nanotubes were deposited onto the aluminium substrates by suspending 0.010g of the sample in 3mL of dimethylformamide (DMF), which was then sonicated for 2 hours at 50°C to ensure maximum dispersion. To produce a consistent coverage, 4 drops of the solution was deposited onto the aluminium, and baked at 150°C under a vacuum of 10-2 Torr to remove DMF solvent from the sample.
B. Fluorination of CNTs and HOPG
The HOPG and CNT surfaces were functionalised using high vacuum plasma functionalisation techniques. Initially, radio frequency (13.56 MHz) SF6 plasma (of power 25.5 W and gas pressure of 1.0x10-2 Torr) was used to achieve functionalisation of HOPG, with the surface being freshly cleaved prior to plasma treatment. An O2 plasma (40.2 W, 2.9x10-2 Torr) pre-treatment (prior to SF6 plasma treatment)
ICONN 2006 103
1-4244-0453-3/06/$20.00 2006 IEEE
was also performed to determine any influence it had on surface functionalisation. Each of the plasma treatments were performed for 15 minutes with the samples being subsequently analysed with X-ray photoelectron spectroscopy (XPS).
Plasma functionalisation was performed on P3-SWCNTs at a gas pressure of 1.0x10-2 Torr and power of 23.1W for times of 3 and 15 minutes. To examine the difference between plasma fluorination and molecular adsorption, a third control sample was produced (ie exposed to SF6 gas at the same pressure as the plasma exposed samples). Determination of the influence of exposure time on the extent of fluorination was investigated using the lowest possible RF power and pressure achievable to sustain an SF6plasma, and then increasing the exposure times stepwise. HOPG surfaces were fluorinated under a gas pressure of 2.5x10-3 Torr and plasma power of 10W for 3, 5, 10, 15, and 30 minute exposures. As prepared RFP-SWCNTs (non-functionalised) were also used for this study and were fluorinated under identical conditions for 3, 5, 10, and 15 minutes only.
200 300
400 500
600 700
800
Binding Energy (eV)
C 1s F 1s O 1s
Clean 3 min 5 min 10 min 15 min 30 min
Figure 2. XPS spectra of HOPG exposed to SF6 plasma for varying times.
C. Analysis
XPS spectra were obtained using a Leybold LHS-10 instrument in our laboratory and a Kratos Axis Ultra instrument at the Ian Wark Research Institute, Adelaide. All scanning tunnelling microscopy (STM) images were obtained using a Digital Instruments MultiMode STM/AFM setup, and NanoScope IV software. Scanning electron microscopy (SEM) images and energy dispersive X-ray (EDX) analysis spectra were obtained from Adelaide Microscopy.
III. RESULTS
A. Graphite Studies
The XPS spectra obtained for pristine HOPG, SF6 plasma treated HOPG, and two-step O2 plasma followed by SF6
plasma treatment, are shown below in Figure 1.
Inspection of the figure clearly demonstrates that pristine HOPG is fluorine free, whereas a marked presence of fluorine (via the F1s photopeak at 683eV) occurs subsequent to SF6
plasma treatment. Interestingly, when the sample was pre-
treated with oxygen plasma prior to SF6 plasma treatment, the resultant fluorine presence is ~30% greater and the surface oxygen content is less than that present when treated with SF6
alone. Moreover, for both cases the absence of sulphur in fluorinated HOPG is noteworthy and somewhat unexpected, given the sulphur content of the gas.
Time dependent fluorination studies have also been performed on HOPG. Figure 2 shows a comparison of the spectra obtained for each of the HOPG samples. It is clear from these surveys that the HOPG is not easily fluorinated, and the extent of functionalisation is only minimally dependent on exposure time at such low plasma power and gas pressure. The maximum fluorine to carbon (F:C) ratio obtained was 0.12 for the two stage O2 and SF6 plasma treatments. For straight SF6
treatments, the maximum F:C ratio obtained was 0.09, resulting from the 3 minute exposure at 2.5x10-3 Torr and forward power of 10W.
Figure 3. STM height image of fluorinated HOPG showing disruptions formed within the graphite lattice. The image is a 10x10nm scan.
200 300
400 500
600 700
800
Binding Energy (eV) O 1s C 1s F 1s
O 1s C 1s F 1s
O 1s C 1s
Figure 1. XPS spectra obtained for the SF6 sample, O2 pre-treated SF6 sample, and pristine graphite.
Clean SF6 Treated O2/SF6 Treated
104
Figure 4. STM current image of fluorinated graphite surface. The image is a 10x10nm scan showing disruption of the periodic graphite lattice.
STM studies were also undertaken on clean graphite, and a graphite sample exposed to SF6 plasma for 10 minutes at a pressure of 2.5x10-3 Torr and power 10W. Figures 3 and 4 show the STM images obtained for HOPG before and after fluorination respectively. The images clearly highlight regions where the fluorination has disrupted the graphene sheet. Fig. 4 shows that the surface still retains some periodicity arising from the graphite, indicating that the fluorination is not occurring evenly across the surface, most likely first taking place at higher energy defect sites within the lattice.
Figure 5. SEM images of (top) pristine CNTs on aluminium substrate and (bottom) after fluorination.
0 200 400
600 800
1000
Binding Energy (eV)
Intensity (counts/sec)
15min
3min
Control
O1s C1s
C1s
C1s O1s
O1s F1s
F1s
F1s
Figure 6. XPS spectra obtained for CNTs exposed to SF6 plasma for 3min and 15min along with a control sample.
B. Carbon Nanotube Studies
CNT fluorination was performed using the same SF6
plasma conditions as those adopted earlier with the HOPG surfaces. SEM measurements of CNTs were undertaken to examine the extent of damage by the plasma fluorination process.
In figure 5, the top image shows untreated CNTs on an aluminium substrate, whereas the bottom image shows the CNTs subsequent to plasma treatment. At points X and Y on this image, the flat surface is thought to be a mat of bundled CNTs, in which the SEM could not resolve individual nanotubes. This image does show, however, that the CNTs have not been significantly damaged by the disruption of carbon-carbon bonds by fluorine during the process. It is suspected that fluorination may cause the nanotube mat or ropes to unwind, but is difficult to discern with the SEM technique as ex-situ characterisation means that one is not able to find identical points on the surface and thus definitively show such changes in morphology.
EDX analysis and XPS techniques were utilised to determine whether the CNTs on the aluminium substrate had been successfully functionalised. A small peak was observed in the EDX spectrum at the expected energy for fluorine, but it
was barely above that of the background and not taken as a definitive result.
XPS characterisation of the same samples was performed and the spectra obtained are shown in figure 6. These results give much clearer evidence that the CNTs had in fact been fluorinated. The control sample referred to in figure 6 is for CNTs exposed only to SF6 gas at the same pressure as that used for plasma treatment, for comparison. While some fluorine was present after this experiment, it was significantly less than that obtained after plasma treatment. The spectra of the 3 and 15 minute exposed samples show that under the
105
0 2000 4000 6000 8000 10000 12000
282 283 284 285 286 287 288 289 290 291 292 293
Binding Energy (eV)
Intensity (counts/sec)
C-O C-C
S-S* C=O
the
with further fluo
The results of this w e show hat fluorination of carbon surfaces can be ed via SF plasma treat
nts looking at the influence of plasma power, gas pres
ation of carbon nanotubes for molecular electronics”, M ineering, 73-74, pp578-582,
ation of
ation of
[5]
[7]
”.
adopted conditions, the fluorination process appears to have almost occurred to completion within the first 3 minutes of plasma treatment. Furthermore, the absence of sulphur is consistent with observations made earlier with HOPG. From high-resolution spectra shown in Figure 7, the maximum fluorine to carbon ratio (after correction for elemental sensitivity) obtained through this procedure was measured to be 0.25r0.03.
It has been previously reported that F-CNTs can be functionalised with an F:C ratio of up to 0.51 [8], which suggests that further fluorination can be theoretically achieved.
presence of fluorination via C-F bond formation, at a binding energy of 289.4 eV, as shown in figure 7.
From figure 6 it is again evident that a majority of the fluorination occurs within the first few minutes,
rination slowed beyond this time. This indicates that the extent of fluorination may have a maximum which is dependent on the experimental protocols of exposure time, gas pressure, and plasma power. Time dependent fluorination studies on the CNT surfaces were further explored, with XPS results presented in figure 8, and the spectra clearly displaying that the CNT surfaces are more easily fluorinated than the HOPG surfaces. This is an expected result as the sp2 hybridised bonding arrangement in graphite is far more stable than the distorted sp2 bonding seen in nanotubes. The significant oxygen photo peaks are thought to result from the DMF used to solvate the CNTs, hence indicating longer drying times are necessary. The highest F:C ratio obtained for these studies was 0.16 for the 5 minute exposure.
IV. CONCLUSION
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
282 283 284 285 286 287 288 289 290 291 292 293
Binding Energy (eV)
Intensity (counts/sec)
C-O C-C
C=O C-F C-CF CF2
ork hav n t
successfully achiev 6
ment, without significant sulphur contamination. It has been shown that there is a clear impact on the extent of fluorination by the exposure time of the samples. CNTs have been shown to more readily functionalise with fluorine plasma than HOPG, and have been fluorinated up to a fluorine-carbon ratio of 0.25.
Further studies on the kinetics of the fluorination process, with experime
Figure 7. High Resolution C1s XPS spectra obtained for control CNTs (top) and exposed to SF6 plasma for 3min (bottom).
200 300
400 500
600 700
800
Binding Energy (eV) O 1s C 1s F 1s
Clean 3 min 5 min 10 min 15 min
Figure 8. XPS survey scans of CNTs exposed to SF6 plasma for varying treatment times.
sures and exposure times on the resultant fluorination are currently underway our laboratory.
REFERENCES [1] N.O.V. Plank and R. Cheung, “Fluorin
icroelectronic Eng 2004.
[2] N.O.V. Plank, L. Jiang, and R. Cheung, “Fluorination of carbon nanotubes in CF4 plasma” Applied Physics Letters, 83, 12, p. 2426(2003) [3] Bishun N Khare, Patrik Wilhite, and M Meyyappan, “The fluorin
single wall carbon nanotubes using microwave plasma”, Nanotechnology, 15, p. 1650 (2004).
[4] E.T. Mickelson, I.W. Chiang, J.L. Zimmerman, P.J. Boul, J. Lozano, J.
Liu, R.E. Smalley, R.H. Hauge, and J.L. Margrave, “Solv
fluorination single-wall carbon nanotubes in alcohol solvents”, J. Phys.
Chem. B, 103, p. 4318 (1999).
S. Iijima, “Helical microtubules of graphitic carbon”, Nature, 354, p. 56- 58 (1991).
[6] P. J. Boul, J. Liu, E. T. Mickelson, C. B. Huffman, L. M. Ericson, I. W.
Chiang, K. A. Smith, D. T. Colbert, R. H. Hauge, J. L. Margrave and R.
E. Smalley, “Reversible sidewall functionalization of buckytubes”, Chem. Phys. Lett., 310(3-4), p. 367-372 (1999).
Kyung Ah Park, Yong Soo Choi, and Young Hee Lee, “Atomic and electronic structures of fluorinated single-walled carbon nanotubes”, Physical Review B, 68(4), (2003).
[8] S. Kawasaki, K.K., F. Okino, H. Touharaa and H. Kataurab,
“Fluorination of open- and closed-end single-walled carbon nanotubes Phys. Chem. Chem. Phys., 6, p. 1769-1772 (2004).
In addition to the elemental fluorine to carbon ratio determination, high resolution spectra of the C1s peak reveal
106