LETTER pubs.acs.org/NanoLett
Gas Dwell Time Control for Rapid and Long Lifetime Growth of Single-Walled Carbon Nanotube Forests Satoshi Yasuda,† Don N. Futaba,† Takeo Yamada,† Motoo Yumura,† and Kenji Hata*,†,‡ †
Nanotube Research Center, National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi, Tsukuba 305-8565, Japan ‡ Japan Science and Technology Agency (JST), Kawaguchi, 332-0012, Japan
bS Supporting Information ABSTRACT: The heat history (i.e., “dwell time”) of the carbon source gas was demonstrated as a vital parameter for very rapid single-walled carbon nanotube (SWNT) forest growth with long lifetime. When the dwell time was raised to 7 s from the 4 s used for standard growth, the growth rate increased to 620 μm/min: a benchmark for SWNT forest growth on substrates. Importantly, the increase in growth rate was achieved without decreasing either the growth lifetime or the quality of the SWNTs. We interpret that the conversion rate of the carbon feedstock into CNTs was selectively increased (versus catalyst deactivation) by delivering a thermally decomposed carbon source with the optimum thermal history to the catalyst site. KEYWORDS: Single-walled carbon nanotube, SWNT forest, water-assisted chemical vapor deposition, dwell time, heat history
T
he low synthesis efficiency of single-walled carbon nanotubes (SWNTs) has always been the limiting factor to explore their full potential. Much research has been conducted over the last two decades to develop synthetic approaches to improve carbon nanotube (CNT) growth. Examples include, alloy catalysts with tailored composition, Al2O3 support, substrates with porous surface morphology, carbon sources containing oxygen, plasma, hot-filament, and the introduction of a growth enhancer, such as water.17 As a result, the synthesis of CNTs has improved considerably to a stage where millimeterscale, vertically aligned SWNTs forests can be routinely synthesized. SWNTs within a forest possess excellent properties, such as high carbon purity, alignment, high surface area, and long length.7,8 These properties have provided new avenues for CNT applications, exemplified by strain sensors,9 temperature invariant viscoelastic material,10 aerogel muscles,11 strong adhesive tapes,12 electrocatalysts for fuel cells,13 stretchable conductors,14 and CNT membranes with ultrafast water transport.15 A literature survey revealed that these highly efficient growths of SWNT forests could be classified into two categories. The first category is featured by long lifetime and tall forests. For example, Kawarada and Robertson have reported the growth of 5 mm SWNT forests by point-arc microwave plasma chemical vapor deposition (CVD) (∼2.6 μm/min growth rate with a ∼32 h growth time).16 Maruyama et al. have produced SWNT forests by water-assisted CVD and reported a linear increase in height with a 1.5 μm/min growth rate for up to 6 h.17 In addition, Einarsson and Maruyama have synthesized SWNT forests by alcohol CVD where the initial growth rate of a few μm/min r 2011 American Chemical Society
exponentially decreased and terminated after ∼30 min.18 Although characterized with long catalyst lifetimes, the growth rates within this category were low and, thus, usually the production rate was not high. In contrast, the second category is featured by fast growth rate but short lifetime. For example, Futaba et al. reported a growth rate of ∼200 μm/min for SWNT forests by water-assisted CVD, a technique denoted as “supergrowth”.19 The growth rate gradually decreased over 20 min, and finally growth terminated at a height of 970 μm. Maruyama et al. reported an acetylene-accelerated alcohol CVD with enhanced growth rate (initial growth rate of ∼100 μm/min), yet the lifetime was short (a few minutes).20 Noda has produced millimeter SWNT forest by water-assisted CVD with an initial growth rate of ∼300 μm/min, but the growth terminated within ∼7 min.21 As shown, much research in this category utilizes a combination of a carbon source and an oxygen source, represented by water. Oxygen sources were proposed to act as growth enhancers that improved the resilience of the catalyst to bear rapid growth. To illustrate this aspect more clearly, we plotted growth rate versus lifetime to map these representative SWNT forest growths (Figure 1). The plot revealed the two growth categories in the leftupper and right-lower side that clearly demonstrated an inverse relationship between the catalyst lifetime and growth rate.22 This inverse relationship demonstrated the difficulty in simultaneously achieving a growth with fast growth rate and long lifetime. Received: April 28, 2011 Revised: July 31, 2011 Published: August 08, 2011 3617
dx.doi.org/10.1021/nl201416c | Nano Lett. 2011, 11, 3617–3623
Nano Letters
LETTER
In this Letter, we report dwell time control of the carbon feedstock as an approach to bridge the two categories by achieving both rapid growth and long lifetime. Dwell time is the time duration of the carbon feedstock in the gas heating zone and is a growth parameter that has been rather overlooked thus far. We developed a CVD furnace designed to control the dwell time that enabled this investigation of the unexplored long dwell time CNT growth region and found that the growth rate and yield significantly increased, up to 620 μm/min and 7.6 mg/cm2 for a 10 min growth, by lengthening the dwell time (from 4 to 7 s) without sacrificing the growth lifetime and SWNT quality. By delivery of the thermally decomposed carbon source with an optimum thermal history to the catalyst sites, the conversion rate to CNTs could be selectively increased over other reaction pathways. The central concept of this work was to introduce the dwell time and carbon flux as two critical parameters to control and improve CNT forest growth (Figure 2a). Carbon flux was defined as the amount of carbon reaching the catalyst per unit time per unit area. By this approach, we could maximize the conversion rate of the carbon source to CNTs at the catalyst sites by optimizing the pyrolysis of the carbon source. To regulate the dwell time and carbon flux over a wide range, we constructed an chemcial vapor reactor composed of a 3 in. infrared-heated vertical furnace (Figure 2a) with a long heating length (265 mm) equipped with a gas shower head with the face oriented parallel to the substrate and connected to the gas inlet tube.23 The large volume inside the furnace enabled a uniform increase in the carbon feedstock dwell time. The dwell time could be adjusted by modifying the length (315 cm) of the gas heating zone by effectively moving the shower position within the furnace. In addition, the total gas flow (10009000 sccm) of the carrier gas (with 10% ethylene) was tuned to modify the carbon flux. The dwell time and the carbon flux were calculated as t ¼ 273
Fc ¼
Ad FTGT
1 f 11:2 A
ð1Þ
ð2Þ
where A is the cross sectional area of the shower head, d is the length of the gas heating zone, F is the total inlet gas flow, TGT is the furnace temperature, and f is the ethylene flow, respectively. From eqs 1 and 2, we calculated the dwell time and carbon flux for each heating zone length and total flow and plotted them into two-dimensional maps (Figure 2bd) to express the region investigated in this work. Standard water-assisted (supergrowth) condition was marked as a black circle (at dwell time of 4.4 s and carbon flux of 1.7 103 mol cm2 min1) for comparison.19 Our experiments covered a wide and unexplored growth region, particularly in the long dwell time zone (yellow and red plots). SWNT forests were synthesized on 2 2 cm silicon substrates with Al2O3 (10 nm)/Fe (1 nm) catalysts at 800 °C by water-assisted CVD. The catalyst was fixed as to not influence the wall number and diameter of the grown CNTs. We used He with H2 as the carrier gas (total gas flow from 1000 to 9000 sccm) and ethylene as the carbon source (10% of the total gas flow). Growth characterization revealed the sensitivity of the SWNT growth to the dwell time. As a function of the dwell time and
Figure 1. Growth rate versus lifetime from representative SWNT forest growths from ref 22. Red and blue diamonds indicate growth results obtained from this work and previous reported water-assisted growth, respectively. Blue circles show other previous SWNT growth results.
carbon flux, the SWNT forests grown in 10 min were characterized by the yield (CNT weight per substrate area) (Figure 2b), height (Figure 2c), and the G/D ratio (Figure 2d). For each growth condition, the water level was individually tuned and optimized, and only the results obtained from the optimum water level are presented here. First, and central to our work, was the observation of a significant increase in both the yield and height with increasing dwell time that maximized at ∼7 s. The maximum yield (7.6 mg/cm2) and height (1892 μm) represent 3.0 and 2.2 times improvement, respectively, from standard supergrowth. Transmission electron microscopy (TEM) observations confirmed that the CNTs were SWNTs. Raman spectroscopy showed that the relative intensities (G/D ratio) of the graphitic G-band (1590 cm1) and the disorder D-band (1340 cm1) decreased only marginally with increased dwell time. Brunauer EmmettTeller (BET) surface areas of the SWNT forest at maximum yield was estimated from nitrogen adsorption isotherms as 1100 m2/g, which was identical with that of growths at standard growth conditions. The measured BET surface area of the SWNT forests corresponded to ∼90% absolute purity, i.e., the carbon nanotube weight percent of a CNT sample.24 This meant that the amount of carbonaceous impurities was below ∼10%. We think this result is significant because long dwell time would be expected to cause excessive gas-phase decomposition that would contaminate the growth environment with soot. Such environments would inevitably lead to an increase in carbonaceous impurities; however, that was not observed. It is important to mention that the carbon flux was also concurrently tuned with the dwell time, where the carbon flux was decreased when the dwell time was increased. Therefore, the growth yield increased despite a decrease in the carbon source, which meant that this approach demonstrated an increased conversion rate efficiency of the carbon source gas into CNTs. We interpreted that the decreased carbon flux suppressed contamination of the growth environment by soot. These results demonstrated that the growth yield of water-assisted CVD could be significantly improved 3618
dx.doi.org/10.1021/nl201416c |Nano Lett. 2011, 11, 3617–3623
Nano Letters
LETTER
Figure 2. (a) A homemade infrared-heated 3 in. vertical furnace equipped with a gas shower head with adjustable height (315 cm) to control the dwell time. Two-dimensional mapping of (b) yield, (c) height, and (d) G/D ratio as functions of dwell time and carbon flux. Growth results are divided into the three categories of low (blue), medium (red), and long (yellow) dwell time by colors. Standard water-assisted (supergrowth) condition is marked as a black circle.
by tailoring the dwell time and carbon flux without sacrificing the quality of the SWNTs. Our approach was compared to relevant studies. Previous studies have reported preheating of the incoming gases to decouple the substrate temperature from the gas heating temperature to grow CNT forests at low temperatures25 or to study the growth kinetics.26 The dwell time in this research was less than 0.1 s and more than an order of magnitude smaller than the dwell time of this research (1 s). While these studies have demonstrated their distinctive advantages, when the gases were preheated at high temperature, the growth lifetime and CNT quality suffered from the excessive gas decomposition. Conceptually, our approach differs from these studies in that we chose the combination of dwell time and carbon flux as the key parameters rather than the gas heating temperature. The benefit of our approach was the ability to enhance the growth yield and rate while not causing excessive gas decomposition. This is important to preserve the quality and purity of the grown CNTs, important factors particularly for SWNT synthesis. With an in situ telecentric height monitoring system,27 we monitored the growth curves (Figure 3) of forests grown with diverse range of dwell times and carbon fluxes each at their respective optimum water level. While the grown conditions
covered very different dwell times and carbon fluxes, all of the growth curves showed a similar overall trend; i.e., the growth rate was highest at the onset of growth, gradually decreased, and finally terminated, as reported previously.19 From these similarities, we inferred an identical growth mechanism for all growths. Although the trend was universal, the lifetime (960 min), growth rate (45620 μm/min), and height grown in 10 min (511892 μm) showed diverse behavior and significantly depended on the dwell time and carbon flux, demonstrating the importance of these two parameters for growth tuning. To investigate the dependence of the growth dynamics on the dwell time and carbon flux, we chose to investigate the growth curves by first-order kinetics analysis. We did not use the diffusion-limited model,28 where the forest acts as a diffusion barrier of the reactant gases, as this model predicted continuous growth without a termination. According to first-order kinetics, the growth-equation of the height (yield) of the forest was described as, H(t) = βτo(1et/τo), where H was the forest height and the two fitting parameters, β and τo, represent the initial growth rate (IGR) and lifetime, respectively.19 By fitting each growth curve in Figure 3 to this growth equation, the IGR (Figure 4a) and the catalyst lifetime (Figure 4b), which characterized the evolution of the growth, were determined and 3619
dx.doi.org/10.1021/nl201416c |Nano Lett. 2011, 11, 3617–3623
Nano Letters
LETTER
Figure 3. Growths curves measured for different dwell time and carbon flux combinations at their respective optimum water levels. Inset figures show time evolution of the forest growth observed in situ telecentric height monitoring system, and vertical and horizontal axes are height (03 mm) and growth time (010 min).
Figure 4. Two-dimensional mapping of (a) initial growth rate and (b) saturated lifetime as functions of dwell time and carbon flux. Plots of (c) initial growth rate and (d) saturated lifetime as a function of the dwell time. The green dotted line is drawn to guide the eye.
plotted as functions of dwell time and carbon flux, respectively. To enable direct comparison with other studies, in this work, the lifetime was defined as the time at growth termination and was chosen as 5τo. Significantly, we found a sharp increase in the growth rate with dwell time (Figure 4a,c), and this observation constitutes one of our main findings. Indeed, the growth rate climaxed at 620 μm/ min at a dwell time of 7 s and then decreased with increasing dwell time. The growth rate of 620 μm/min was a 2.5-fold improvement from previous water-assisted growth results19 and represents a benchmark for the growth rate of SWNT forest growth on substrates (refer to Figure 1 for comparison with literature). In fact, this growth rate is on par with some of the highest growth rates reported for multiwalled carbon nanotubes (MWNTs) (Table 1S, Supporting Information) and approaches the fastest reported growth rate (∼1000 μm/min, lifetime ∼1 min), but with a much longer lifetime.26 As described in the introduction, an increase in the growth rate has historically been tied to a decrease in lifetime. In contrast, as displayed in Figure 4b, in our approach, the lifetime varied only
slightly with the dwell time. In addition, we rearranged the data of Figure 4b into a plot of the lifetime as a function of dwell time (Figure 4d). To organize the somewhat scattered data, we added the average lifetimes of the short, medium, and long dwell time regions as green circles and found that they were nearly identical. We believe it is significant that our approach could improve the growth rate while preserving the lifetime, since this would have a great impact on the mass production of SWNT forests. For example, a pilot plant for the continuous synthesis of SWNT forests has recently been constructed where substrates are continuously fed into a furnace by a belt conveyer.29 In this system, the growth yield (CNT weight per substrate area), Y, could be estimated as Y = W L D R, where W and L are the width and length of the furnace, D is the density of the forest, and R is the growth rate. The direct proportionality of the growth yield to the growth rate demonstrates its importance for mass production. To gain insight into the underlying mechanism of the effect of dwell time, we investigated the optimum water level (Figure 5a) required to maximize growth efficiency as functions of the dwell 3620
dx.doi.org/10.1021/nl201416c |Nano Lett. 2011, 11, 3617–3623
Nano Letters
LETTER
Figure 5. Plot of (a) optimum water level maximized the forest height as functions of dwell time and carbon flux. (b) Carbon efficiency as a function of the dwell time. (c) Schematic of highly efficient SWNT growth. (d) Arrhenius plots for initial growth rate for the growths at maximum yield with long dwell time (∼7 s) and standard supergrowth. Black dotted lines are drawn to guide the eye.
time and carbon flux. Interestingly, the optimum water level was nearly unvarying and was equivalent to the level of standard supergrowth. This differed from the trend of the growth rate that greatly varied with dwell time and carbon flux. The carbon flux (0.6 103 mol cm2 min1) at the maximum yield was smaller than that (1.7 103 mol cm2 min1) of standard supergrowth, albeit resulting in a higher growth yield. This meant an increase in the carbon efficiency with dwell time. Carbon efficiency is defined as the percentage of the input carbon feedstock converted into the CNT product and is important for economical CNT production as well as synthesizing highpurity CNTs. As demonstrated in Figure 5b, the carbon efficiency increased monotonically and almost linearly with the dwell time and reached a value exceeding 10% at the point of maximum yield. This value is 8.3 times higher than that of standard supergrowth and is very high for SWNT growth on substrates. The underlying mechanism for highly efficient growth was used to explain the experimental results. The basic chemical reactions (Figure 5c) of water-assisted CVD have been described as follows: (i) carbon feedstock decomposes and either deactivates the catalysts by amorphous carbon coating or converts into CNTs; (ii) water reactivates the catalysts by removing the amorphous carbon coating.30 Within this model, the conversion and deactivation rates are coupled to the growth rate and lifetime, respectively. Therefore, as conversion into CNTs and deactivation of the catalysts are competing pathways, this explains the inverse relationship between the growth rate and lifetime. We believe that an appropriate increase in dwell time invoked a shift among the competing pathways where the conversion rate into CNTs was selectively increased while the deactivation rate did not. Such preferential increase in the conversion rate occurred because the dwell time was controlled to deliver thermally decomposed carbon source with the best thermal history to the catalyst site. This would result in an increase of both the growth rate and carbon efficiency, as observed experimentally. In contrast, the unvarying optimum water levels and unvarying lifetimes agreed with our assumption that the deactivation rate did not increase with the dwell time. This was because any variation in the catalyst deactivation rate would be expected to result in a change to the catalyst lifetime or require an adjustment in the optimum water level; however, neither was observed. The optimum water level was expected to increase/decrease with an increased deactivation rate to maintain the reactivation process
Figure 6. (a) Time evolution of the forest growth observed in situ telecentric optical monitoring system. (b) Photograph of SWNT forest with 30 min growth time.
brought about by water. We note that this explanation is only based upon the chemical reactions of the gases and does not rule out any of the other mechanisms, particularly those that consider the catalysts, such as inhibition of Ostwald ripening of the catalysts by water.17 Figure 5d shows an Arrhenius plots of the temperature dependence of the initial growth rate for the growths at maximum yield (dwell time ∼7 s) and standard supergrowth (dwell time ∼4 s). The activation energies, as calculated from their respective slopes, were found to be quite similar: 2.83 and 2.84 eV, respectively. Wirth et al. proposed that the growth of CNTs was limited by carbon diffusion in the catalyst.31 The fact that the activation energies were similar indicated that an increase in the dwell time did not influence the rate-limiting step of CNT growth, and the activation energy of ∼2.8 eV falls among the highest reported activation energies for CNT growth.26,3135 For example, an activation energy of 0.93 eV has been reported for acetylene growth31 and an even lower 0.23 eV for plasma enhanced CVD.32 We believe it is significant that dwell time 3621
dx.doi.org/10.1021/nl201416c |Nano Lett. 2011, 11, 3617–3623
Nano Letters control of water-assisted CVD achieved a very high growth rate albeit exhibiting a high activation energy, and we envision that further improvements might be possible by adopting this approach to growth systems with lower activation energy. Finally, to demonstrate the potential of our approach, a multimillimeter SWNT forest was synthesized in a short time. Figure 6 shows a picture and the corresponding growth curve of a SWNT forest with 5 mm height synthesized in 30 min. While not the tallest SWNT forest reported, it was grown in a much shorter time. In comparison, for example, Kawarada and Robertson have reported the growth of a 5.2 mm SWNT forest in 32 h.16 The BET specific surface area of the 5 mm SWNT forest was 912 m2/g which corresponds to ∼80% absolute purity,24 and thus we believe this forest represents the tallest SWNT forest with a high absolute purity. Such forests would be valuable for supercapacitor electrodes where the concentration of impurities must be minimized to avoid any parasitic chemical reactions that would degrade operational performance. In conclusion, we showed the dwell time of the carbon source gas as a crucial parameter for improving the growth rate and yield without sacrificing the quality of SWNTs. When the dwell time was tuned, the conversion rate of the carbon source to CNTs at the catalyst site was maximized by controlling the pyrolysis of the carbon source gas. When the dwell time was increased to 7 s, the growth rate reached 620 μm/min, a value that represents a benchmark for SWNT growth on substrates. The lifetime and CNT quality were preserved even at the fastest growth, a point that would have a great impact on economical mass production of SWNT forests. We believe that the present research demonstrates an approach to surmount the general problem of the inverse relationship between the lifetime and growth rate and provides a rational route to for rapid SWNT growth with long lifetime.
’ ASSOCIATED CONTENT
bS
Supporting Information. A figure showing growth rate vs lifetime from representative SWNT and MWNT forest growth and a table presenting a summary of growth rate and lifetime for MWNT growth reproted by recent studies. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT Support from the Nanotechnology Program “Carbon Nanotube Capacitor Development Project” (2006-2011) by the New Energy and Industrial Technology Development Organization (“NEDO”) is acknowledged. ’ REFERENCES (1) Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Ohtsuka, Y.; Sen, R.; Suzuki, S.; Achiba, Y. Carbon 2000, 38, 1691–1697. (2) Hongo, H.; Yudasaka, M.; Ichihashi, T.; Nihey, F.; Iijima, S. Chem. Phys. Lett. 2002, 361, 349–354. (3) Kong, J.; Soh, H. T.; Cassell., A. M.; Quante, C. F.; Dai, H. Nature 1998, 395, 878–881.
LETTER
(4) Futaba, D. N.; Goto, J.; Yasuda, S.; Yamada, T.; Yumura, M.; Hata, K. Adv. Mater. 2009, 21, 4811–4815. (5) Hofmann, S.; Ducati, C.; Kleinsorge, B.; Robertson, R. Appl. Phys. Lett. 2003, 83, 4661–4663. (6) Huang, Z. P.; Xu, J. W.; Ren, Z. F.; Wang, J. H.; Siegal, M. P.; Provencio, P. N. Appl. Phys. Lett. 1998, 73, 3845–3847. (7) Hata, K.; Futaba, D. N.; Mizuno, K.; Namai, T.; Yumura, M.; Iijima, S. Science 2004, 306, 1362–1364. (8) Futaba, D. N.; Hata, K.; Yamada, T.; Hiraoka, T.; Hayamizu, Y.; Kakudate, Y.; Tanaike, O.; Hatori, H.; Yumura, M.; Iijima, S. Nat. Mater. 2006, 5, 987–994. (9) Yamada, T.; Hayamizu, Y.; Yamamoto, Y.; Yomogida, Y.; LzadiNajafabadi, A.; Futaba, D. N.; Hata, K. Nat. Nanotechnol., in press. (10) Xu, M.; Futaba, D. N.; Yamada, T.; Yumura, M.; Hata, K. Science 2010, 330, 1364–1368. (11) Aliev, A. E.; Oh, J.; Kozlov, M. E.; Kuznetsov, A. A.; Fang, S.; Fonseca, A. F.; Ovalle, R.; Lima, M. D.; Haque, H. M.; Gartstein, Y. N.; Zhang, M.; Zakhidov, A. A.; Baughman Science 2009, 323, 1575–1578. (12) Qu, L.; Dai, L.; Stone, M.; Xia, Z.; Wang, Z. L. Science 2008, 322, 238–242. (13) Gong, K.; Du, F.; Xia, Z.; Durstock, M.; Dai, L. Science 2009, 323, 760–764. (14) Sekitani, T.; Noguchi, Y.; Hata, K.; Fukushima, T.; Aida, T.; Someya, T. Science 2008, 321, 1468–1472. (15) Hinds, B. J .; Chopra, N.; Rantell, T.; Andrews, R.; Gavalas, V.; Bachas, L. G. Science 2004, 303, 62–65. (16) Zhong, G.; Iwasaki, T.; Robertson, J.; Kawarada, H. J. Phys. Chem. B 2007, 111, 1907–1910. (17) Amama, P. B.; Pint, C. L.; McJilton, L.; Kim, S. M.; Stach, E. A.; Murray, P. T.; Hauge, R. H.; Maruyama, B. Nano Lett. 2009, 9, 44–49. (18) Einarsson, E.; Murakami, Y.; Kadowaki, M.; Maruyama, S. Chem. Phys. Lett. 2008, 46, 923–930. (19) Futaba, D. N.; Hata, K.; Yamada, T.; Mizuno, K.; Yumura, M.; Iijima, S. Phys. Rev. Lett. 2005, 95, 056104. (20) Xiang, R.; Einarsson, E.; Okawa, J.; Miyauchi, Y.; Maruyama, S. J. Phys. Chem. C 2009, 113, 7511–7515. (21) Hasegawa, K.; Noda, S. Jpn. J. Appl. Phys. 2010, 49, 085104. (22) (a) Zhong, G.; Iwasaki, T.; Robertson, J.; Kawarada, H. J. Phys. Chem. B 2007, 111, 1907–1910. (b) Eres, G.; Kinkhabwala, A. A.; Cui, H.; Geohegan, D. B.; Puretzky, A. A.; Lowndes, D. H. J. Phys. Chem. B 2005, 109, 16684–16694. (c) Amama, P. B.; Pint, C. L.; McJilton, L.; Kim, S. M.; Stach, E. A.; Murray, P. T.; Hauge, R. H.; Maruyama, B. Nano Lett. 2009, 9, 44–49. (d) Einarsson, E.; Murakami, Y.; Kadowaki, M.; Maruyama, S. Chem. Phys. Lett. 2008, 46, 923–930. (e) Xiang, R.; Einarsson, E.; Okawa, J.; Miyauchi, Y.; Maruyama, S. J. Phys. Chem. C 2009, 113, 7511–7515. (f) Futaba, D. N.; Hata, K.; Yamada, T.; Mizuno, K.; Yumura, M.; Iijima, S. Phys. Rev. Lett. 2005, 95, 056104. (g) Yasuda, S.; Futaba, D. N.; Yamada, T.; Satou, J.; Shibuya, A.; Takai, H.; Arakawa, K.; Yumura, M.; Hata, K. ACS Nano 2009, 12, 4164–4170. (h) Hasegawa, K.; Noda, S. Jpn. J. Appl. Phys. 2010, 49, 085104. (23) Yasuda, S.; Futaba, D. N.; Yamada, T.; Satou, J.; Shibuya, A.; Takai, H.; Arakawa, K.; Yumura, M.; Hata, K. ACS Nano 2009, 12, 4164– 4170. (24) Futaba, D. N.; Goto, J.; Yamada, T.; Yasuda, S.; Yumura, M.; Hata, K. Carbon 2010, 48, 4542–4546. (25) Nessim, G. D.; Seita, M.; O’Brien, K. P.; Hart, A. J.; Bonaparte, R. K.; Mitchell, R. R.; Thompson., C. V. Nano Lett. 2009, 9, 3398–3405. (26) Meshot, E. R.; Plata, D. L.; Tawfick, S.; Zhang, Y.; Verploegen, E. A.; Hart, A. J. ACS Nano 2009, 9, 2477–2486. (27) Yasuda, S.; Futaba, D. N.; Yumura, M.; Iijima, S.; Hata, K. Appl. Phys. Lett. 2008, 93, 143115. (28) Zhu, L.; Hess, D. W.; Wong, C. P. J. Phys. Chem. B 2006, 110, 5445–5449. (29) PCT/JP2009/056878US (30) Yamada, T.; Maigne, A.; Yudasaka, M.; Mizuno, K.; Futaba, D. N.; Yumura, M.; Iijima, S.; Hata, K. Nano Lett. 2008, 8, 4288–4292. (31) Wirth, C. T.; Zhang, C.; Zhong, G.; Hofmann, S.; Robertson, J. ACS Nano 2009, 3, 3560–3566. 3622
dx.doi.org/10.1021/nl201416c |Nano Lett. 2011, 11, 3617–3623
Nano Letters
LETTER
(32) Hofmann, S.; Csanyi, G.; Ferrari, A. C.; Payne, M. C.; Robertson, J. Phys. Rev. Lett. 2005, 95, 036101. (33) Kim, K. E.; Kim, K. J.; Jung, W. S.; Bae, S. Y.; Park, J.; Choi, J.; Choo, J. Chem. Phys. Lett. 2005, 401, 459–464. (34) Einarsson, E.; Murakami, Y.; Kadowaki, M.; Maruyama, S. Carbon 2008, 46, 923–930. (35) Bronikowski, M. J. J. Phys. Chem. C 2007, 111, 17705–17712.
3623
dx.doi.org/10.1021/nl201416c |Nano Lett. 2011, 11, 3617–3623
