As shown in Fig. , the SEM (Fig. a) and TEM (Fig. ) images show that the pristine MWCNTs are highly tangled with diameters ranging from 15 to 40 nm. The TEM image (Fig. ) also shows that the nanotubes are with densely distributed defects. After the ECR plasma treatment with a H2/O2ratio of 40/10 (sccm/sccm), as depicted in Fig. , the morphologies, structural and the diameter distributions of the sample do not show any observable difference in comparison with those of the pristine sample. In contrast to Fig. , the TEM image in Fig. shows that the dispersion ability of the nanotubes is increased after the plasma treatment so that the nanotube bundle can be significantly separated. Meanwhile, the insignificant change in morphology and structure can also be observed when sample is treated with any gas condition.
a SEM and b TEM images of the pristine MWCNTs,c SEM, and d TEM images of ECR-plasma-treated MWCNTs with 5 min exposure under H2/O2 gas compositions of 40/10 (sccm/sccm)
Figure shows the XPS survey spectra of the untreated and the ECR-plasma-treated MWCNTs. It is noted that the spectra showing the presence of carbon and oxygen on the treated and untreated samples are normalized with respect to C 1s intensity for comparison purposes. In contrast to the spectrum of the pristine MWCNTs, a higher concentration of oxygen is introduced to the surface of the nanotubes treated by the ECR plasma using any gas composition.
XPS survey spectra of the pristine MWCNTs and the ECR-plasma-treated MWCNTs under various H2/O2 (sccm/sccm) gas compositions
As a reference spectrum, the XPS C 1s spectrum of the pristine MWCNTs was recorded and is shown in Fig. . Based on the previous studies [19
], the spectra are deconvoluted into five Gaussian peaks centered at 284.5, 285.1, 286.2, 287.2, and 288.9 eV. Here, the main peak at 284.5 eV originates from a graphite signal. The peak at 285.1 eV is attributed to sp3
]. The peaks at 286.2, 287.2, and 288.9 eV correspond to hydroxyl, carbonyl (or ether), and carboxyl (or ester) groups, respectively. A peak attributed to π–π* shake-up bonds is observed at 290.4 eV [20
]; and the peak at 283.2 eV originates from carbidic carbon [27
]. Meanwhile, the deconvolution results also show that the oxygen functional groups have been grafted onto the surface of the pristine MWCNTs with a concentration of approximately 11.2%. This is consistent with the description provided by the vendor that the raw materials were treated using mild acids prior to shipment. Furthermore, the results show that the MWCNTs are still composed of approximately 40% amorphous carbon.
XPS C 1s spectra of the pristine MWCNTs and the five chemical species: (1) graphite; (2) sp3 carbons; (3) hydroxyl groups; (4) carbonyl groups; and (5) carboxyl groups
For a detailed comparison, all C 1s spectra of the MWCNT samples are presented in Fig. and the calculated results are summarized in Table . Note that the [O]/[C] ratio given in Table is based on the relative percentage of three carboxyl groups to the total number of carbon atoms detected. As shown in Fig. and Table , after the samples are treated by the plasma treatment, the XPS measurements show that the concentrations of the graphite, sp3carbons, and oxygen-containing functional groups are different according to the gas mixture composition. Also, it is clear that when the samples of the MWCNTs are treated with a H2/O2ratio of 25/25 (sccm/sccm), the highest concentration of oxygenated functional groups is achieved whilst the concentration of amorphous carbon is minimized.
XPS C 1s spectra of the pristine MWCNTs and the ECR-plasma-treated MWCNTs after 5 min of exposure under various H2/O2(sccm/sccm) gas compositions
The MWCNT specimen treated under various H2/O2gas compositions; the characterization results of XPS; and theID/IGratio of Raman spectra
Note that the amount of surface defects is important for functional groups to form on the nanotube surface [10
]. In order to evaluate the formation of defects on the nanotube surface by the ECR plasma treatment using different gas compositions, the Raman spectra of the pristine and the plasma-treated MWCNTs are presented in Fig. . Two characteristic peaks are observed and attributed to the D and G bands. The spectra have been normalized with respect to the G band for comparison. The intensity of the D band, at frequencies around 1,344 cm−1
, is correlated with structural disorder of CNTs, which originates from the defects including disordered materials, poor graphitization, functionalized carbon, and the amorphous carbon on the sidewall of nanotubes [28
]. The G band at frequencies around 1,572 cm−1
is activated by the graphite signal [30
]. It is suggested that the ID
ratio is closely associated with the defect density on the walls of the MWCNTs [30
]. Therefore, the intensity ratio can be used to evaluate the formation of defects which are preferential sites for functionalization.
Raman spectra of the pristine MWCNTs and the ECR-plasma-treated MWCNTs after 5 min of exposure under various H2/O2(sccm/sccm) gas conditions
As expected, allID/IGratios are increased after plasma treatment with any gas composition. As shown in the inset of Fig. , theID/IGratio increases from 1.03 to 1.27 when the H2concentration increases from 0 to 80 vol.%. Note that the ion density is very important for surface etching. By comparing the ionic current, it is found that current increases from 0.12 to 0.47 A while the H2concentration increases from 0 to 100%. This shows that the ion density of the cation stream increases as H2concentration increases. This leads to higherID/IGratio when H2concentration of the etching gas is higher. However, it is also shown that the ratio decreases to 1.08 when the nanotubes are treated with pure H2gas (H2/O2of 50/0 (sccm/sccm)). This could be because the ion bombardment under this gas condition can heavily etch the surface of the nanotubes to the extent that the concentration of amorphous carbon coated on the surface could thus be raised from 40 to 54% as shown in Table . The thick amorphous carbon can prevent defects from further forming on the surface during ion bombardment so that theID/IGratio is significantly smaller than that of the nanotubes treated with the etching gas containing 20 vol.% oxygen (ID/IG = 1.27, H2/O2of 40/10 (sccm/sccm)). Apart from their involvement in ion bombardment, the generated oxygen cations can also act as highly reactive chemical species which form covalent bonds with the amorphous carbon and then nanotube surface. More specifically, the amorphous carbon layer is more reactive than the cylindrical walls to form volatile products with the oxygen cations. The products are then pumped out by the vacuum system. Thus, as shown in Table , treatment using a H2/O2mixture can increase the concentration of oxygenated functional groups whilst also reducing the concentration of amorphous carbon. On the other hand, treatment with pure O2gas (with the exception of increasing theID/IGratio) does not yield any other obvious effects in regard to the purity of CNTs and the concentration of functional groups when compared with the pristine MWCNTs. This indicates that a H2/O2mixture not only facilitates defect formation but also promotes covalent bonding in this case. Therefore, even with the addition of 20 vol.% H2(H2/O2of 10/40 (sccm/sccm)) in gas, there is still a significant removal of amorphous carbon and formation of oxygen-containing groups on the nanotube surface.
Figure shows the weight-derivative curves of TGA analysis on the pristine MWCNT samples and the ECR-plasma-treated MWCNTs. The results show that with a decomposition temperature of 600 °C, the pristine MWCNT samples are the most thermally stable with respect to oxidative degradation. Correspondingly, the MWCNTs treated by the ECR plasma with a gas composition of 40/10 (sccm/sccm) have the lowest decomposition temperature (594 °C). It should be noted that the oxidation stability is a function of the combined effect of defects and the diameter of the nanotubes [12
]. With the same diameter distribution, the results are in agreement with the hypothesis that the decrease of the oxidation reaction temperature is mainly a result of the defects produced by the plasma treatment. The marginal differences between untreated and treated samples reflect the fact that the effect of gas composition on structural integrity of the nanotubes is insignificant in this case.
Weight-derivative curves of TGA on the pristine MWCNTs and the ECR-plasma-treated MWCNTs under various H2/O2(sccm/sccm) gas compositions