Characterizations of MWCNTs of various experiment conditions and

Figure 4-5 is a low magnification TEM image, which shows that raw samples contain multi-walled carbon nanotubes and metal catalysts. In this image, metal catalysts of various sizes can be found in the tip or tube wall of MWCNTs. MWCNTs with diameters from 40 to 60 nm can also be displayed by TEM. In Figure 4-6, the image shows that most of metal particles were removed from MWCNTs after purification. As shown in Figure 4-7, catalysts embedded in the tip of MWCNTs were removed. Obviously, the closed tube tip was opened by acid treatment in microwave digestion. It seems that closed tip was opened first and then the metal catalyst was eliminated by nitric acid. This high resolution TEM image also shows that the tube wall structure was not damaged by acid treatment.

Fig. 4-5 Low magnification TEM image of raw MWCNTs.

Fig. 4-6 Low magnification TEM image of purified MWCNTs.

Fig. 4-7 Catalysts embedded in the tip of MWCNTs were removed.

Figure 4-8 shows the effect of acid concentration on purification for various processing time. The TGA measurement of the amount of residual catalyst is effective to evaluate the key features of purification. Behavior of the weight-temperature curve reveals that by-products during acid treatment such as carbonaceous appear in as-purified samples [126]. Curves a and b show that the purification for shorter microwave-digestion time is not strong enough no matter how high the acid concentration is. However, curve c shows that catalyst elimination increases with increasing acid concentration for 90 min treatment. This result could be that it takes time for the tube-open process to take place. In excess of the critical time, metal could be etched away quickly. In this experiment, 5 M nitric acid and 90 min might be appropriate parameters to purify carbon nanotubes.

Fig. 4-8 Effect of acid concentration on purification ability for different processing time.

Figure 4-9 shows the effect of ultra-sonication time on purification efficiency with 5 M nitric acid for various treatment time. In curves d and e, increases of sonication time result in the decrease of residual metal amount. In samples with 60 min purification treatments, catalyst amount measured by TGA decreases from 0.49% to 0.31% with increasing treatment time from 1 h to 3 h. It seems that raw samples are more dispersed in ultrasonic treatments for longer acid treatments. However, acid treatment time of 90 min (Curve f) reveals almost the same amount of residual catalyst. The results could be that acid treatment time might dominate purification efficiency and long sonication time, which might induce the defect of tubes wall [125], is unnecessary.

Fig. 4-9 Effect of ultra-sonication time on purification efficiency of 5 M nitric acid.

Figure 4-10 shows the results of TGA analysis of raw and purified samples. In Figure 6(a), the TGA analysis shows that metal catalyst content in the raw sample was 1.34%. The early slight weight gain in the raw sample might be due to the oxidation of metal catalysts. Raw sample weight starts to reduce near 515°C and MWCNTs completely evaporate above 712°C. The decomposition temperature at 673°C is defined as the inflection point during the oxidation of tubes [126]. Different from the raw sample, purified MWCNTs start to lose weight by burn-off from 550°C and completely burn-out near 758°C, as shown in Figure 6(b). The amount of residual catalyst dropped to 0.1 wt% in purified sample of group C for 90 min purification.

The decomposition temperature in this purified sample was 703°C, which was slightly higher than that in the raw sample. This result reveals that residual metal embedded in raw samples acted as oxidation site and initiated oxidations [179]. It was also reported that metal impurities in the sample would lower the decomposition temperature and increase the decomposition rate [179]. The purified sample is thermally more stable towards oxidative destruction than raw MWCNTs. In Figures 6(a) and 6(b), there are

evidently no other separate regimes except the main combustion regime. This points out that microwave digestion does not introduce carboxyl, aldehyde and other oxygen-containing functional groups on the surface of the non-nanotubes carbonaceous fractions. These groups, by report, were extremely hygroscopic and reactive towards oxidation.

Fig. 4-10 Results of TGA analysis of (a) raw and (b) purified samples.

Figure 4-11 reported the catalyst contents after purification with 5 M acid for various treatment time. The amount of catalyst decreases form 1.34% to 0.6% at the processing time of 30 minutes and to 0.49% at 60 min treatment. After treated for 90 min, the total amount of catalyst dropped to 0.1%. This result clearly indicated that nitric acid could eliminate more catalysts under sufficient treatment time. The purity of purified MWCNTs could reach 99.9% only in 90 min acid processing time without damaging the wall structure.

Fig. 4-11 Catalyst contents after purification with 5 M acid for various treatment time.

The Raman spectroscopies of raw and purified samples in Fig. 4-12 are composed of two characteristic peaks for nanotubes. The G band near 1580 cm^-1 is related to the graphite E2g symmetry of the interlayer mode, which reflects the structural intensity of the sp²-hybridized carbon atoms of the nanotubes. Another peak at 1350 cm^-1, the D band, indicates disordered carbon atoms. The extent of defect in carbon nanotubes can be evaluated with these two peaks. Raman analysis shows that ratios of IG/ID before and after purification are almost the same. This means that the bonding structure of

tube walls was not damaged by microwave-assisted purification. Some studies [120,179] have suggested that the IG/ID ratio would increase after purification, because of the improvement in nanotubes content by eliminating amorphous carbon.

In this study the amount of carbonaceous materials was low (<3%) and the purification process did not damage bonds, so Raman curves of raw and purified sample are almost the same.

Fig. 4-12 Raman spectrums of raw and purified samples.

Research has been reported to obtain purified carbon nanotubes. Ando et al. [132]

reported that MWCNTs were ground and boiled with 20% H2O2 in a reflux condenser for 45 h. Then the residual material was refluxed for 24h in a mixture of sulfuric acid (96%) and nitric acid (61%) with the ratio of 3:1. Chen et al. [129] proposed that MWCNTs were stirred in 3 M nitric acid and refluxed in 5 M HCl solution for 6 h at 120°C. After acid treatment, samples were calcined in static air at 510°C for about 60 min. Moon et al. reported a two step process of thermal annealing in air with acid treatment to purify single walled carbon nanotubes. In this study, the acid treatment with HCl for 24h was to etch away the metal catalyst [126]. Evidently most acid

treatments removed metal catalysts for more than 24 h. It was reported that carbon nanotubes would break into short pieces for too long acid treatments [179]. This is because oxygen containing mineral acid, HNO3 for example, is very efficient in dissolving metal particles and polyaromatic solids such as graphites or amorphous carbons. However, in the microwave digestion system, nitric acid can rapidly absorb microwave energy without agitation in solution. Owing to the high efficiency of heat adsorption and no agitation, the processing time of microwave digestion to dissolve catalyst in MWCNTs can be reduced to 2 hours and metals could be eliminated from MWCNTs rapidly without destruction in wall structure.

In addition to the acid treatment, Shelimov et al. [125] also proposed a method of ultrasonically assisted filtration. In this method, sample sonication during filtration prevents filter contamination and provides a fine nanotube-nanoparticle suspension through purification. The process generates SWNTs with purity higher than 90%.

Although this method could separate coexisting carbon nanospheres, metal nanoparticles, polyaromatic carbons and fullerenes from carbon nanotube fractions, metal catalysts embedded in the tip and wall could not be eliminated by this method.

One advantage of microwave digestion method was that the embedded metal catalyst would be eliminated and the purity of carbon nanotubes could be higher.

Another purification method to eliminate metal catalysts was proposed by Chiang et al. [133]. This method suggests a purification strategy based on oxidizing Fe and then dissolving the oxide. Raw material was heated in static air at 200°C for 24 h and followed by sonication in concentrated HCl (37%) in 80°C water bath for 15 min.

Although treated in HCl for only 15 min, the total purification time was obviously above 24 h. In this report, the total acid treatment time was below two hours. It was apparent that microwave digestion could effectively eliminate catalysts from carbon nanotubes and would not introduce structure defects.

4.2.3 Summary

Microwave digestion method was developed in our previous work with the advantages of high efficiency, easy operation, short time, damages free on CNTs and little consumption in reagents. However, residual catalyst amount of ECR-synthesized MWCNTs after purification was still high, because of the high catalyst content in raw samples. Investigated in this work is the purification efficiency of MWCNTs synthesized by thermal chemical vapor deposition with different parameters by using TGA, SEM, TEM and Raman spectroscopy and MWCNTs of high purity are expected.

The results show that the purification efficiency increases with increasing acid treatment time. The amount of residual catalysts in purified samples was reduced to 0.1% after digestion for 90 min at 210°C. In conclusion, microwave digestion may have great potential in mass purification. High quality and large amount of purified CNTs would be applied to more intrinsic studies and industrial applications.

4.3 Purification efficiency of multi-walled carbon nanotubes