ARTICLE pubs.acs.org/Langmuir
Raman Characterization of Thermal Conduction in Transparent Carbon Nanotube Films Duckjong Kim,*,† Lijing Zhu,†,‡ Chang-Soo Han,§ Jae-Hyun Kim,† and Seunghyun Baik‡ †
Department of Nano Mechanics, Korea Institute of Machinery and Materials Department of Mechanical Engineering, Sungkyunkwan University § Department of Mechanical Engineering, Korea University, Seoul, 136-701, South Korea ‡
bS Supporting Information ABSTRACT: Using materials with high thermal conductivity is a matter of great concern in the field of thermal management. In this study, we present our experimental results on two-dimensional thermal conductivity of carbon nanotube (CNT) films obtained by using an optical method based on Raman spectroscopy. We use four kinds of CNTs in film preparation to investigate the effect of CNT type on heat spreading performance of CNT films. This first comparative study using the optical method shows that the arc-discharge single-walled carbon nanotubes yield the best heat spreading film. We also show that the Raman method renders reasonable thermal conductivity value as long as the sample is a transparent film by testing CNT films with various transmittance. This study provides useful information on characterization of thermal conduction in transparent CNT films and could be an important step toward high-performance carbonbased heat spreading films.
’ INTRODUCTION Carbon nanotubes (CNTs) are in the spotlight as a good candidate for next generation heat spreading material due to their high thermal conductivity, flexibility in fabrication processes, and abundance of raw carbon materials. Many research works on measurement of thermal conductivity of CNTs have been reported as shown in Table 1.114 Several researchers reported that the thermal conductivity of individual single-walled carbon nanotube (SWCNT) exceeds that of diamond (9002320 W/mK).13,15 Previous works mainly covered the thermal property of individual CNTs and bundled CNTs. Since CNTs would be primarily used in the form of networks, clarification of heat transfer characteristics of CNT films is important, but just a few measurement results have been reported.6,7 In addition, since there are several kinds of CNTs which are categorized according to the number of walls and synthesis method, identification of the most appropriate type of CNTs yielding the best heat spreading films would be very important in the field of thermal management. Table 1 shows that a lot of methods have been used for determining thermal conductivity of CNT samples, such as thermal bridge,24,6 3ω method,5,8,1113 laser-flash,9,14 etc. There are wide variations in the reported data even for similar CNT samples. That would be due to the uncertainty in the thermal resistance of the CNTCNT junctions and the filling factor of CNTs in the bundles. Moreover, the thermal contact resistance would be an important and difficult problem in determining the intrinsic thermal conductivity in almost all the steady-state heat flow methods using contact temperature readings.1 Using noncontact method can avoid large-heat-capacity objects contacting r 2011 American Chemical Society
with the CNT for its local temperature measurement. Raman spectroscopy is a noninvasive technique for characterizing physical properties of carbon-based materials. The Raman method has been applied to measurement of thermal conductivity of various materials, such as individual CNT,1 monolayer graphene,1620 and porous silicon layers.21,22 However, the Raman method has not been used for thermal characterization of CNT films, and required specifications on samples to be measured have not been clarified. Here, we first measured two-dimensional thermal conductivity of CNT films by using the Raman spectroscopy. We used four types of CNTs in film preparation with variations in film transmittance and size. Based on the results, we discussed the possibility of CNT films as a heat spreader and applicable range of the optical method.
’ EXPERIMENTAL SECTION We made CNT films by using vacuum filtration method.23 We filtered CNTs by using an anodic aluminum oxide filter membrane (Anodisc 47, Whatman) with 0.2 μm pore and removed the membrane by using 3 mol/L sodium hydroxide solution (Samchun Pure Chemical). Then, we transferred the CNT film on a perforated substrate to form suspended CNT films as shown in Figure 1. We used four types of CNTs in the present work: single-walled carbon nanotubes (SWCNTs) synthesized by arc-discharge method or high-pressure carbon monoxide process Received: August 2, 2011 Revised: October 7, 2011 Published: October 17, 2011 14532
dx.doi.org/10.1021/la203007y | Langmuir 2011, 27, 14532–14538
Langmuir
ARTICLE
Table 1. Thermal Conductivity Values of CNTs Reported in Previous Works thermal CNT
synthesis
form of
measurement
conductivity
type
method
sample
method
(W/mK)
SWCNT
CVD
individual
Raman
2400 ( 4001
thermal bridge
3000, 5000, 100002 35003
arcdischarge MWCNT
CVD
MPCVD
bundles
thermal bridge
1504
bundles
3ω
3.585
random
thermal bridge
35, 2.36
network individual
bolometric Raman
30, 75, 837 1400 ( 2501
3ω
300 ( 208
bundles
laser-flash
2509
bundles
pulsed
121710
photothermal reflectance PECVD
bundles
3ω
748311
CCVD
individual
3ω
600 ( 10012 650, 83013
bundles
3ω
15060012
aligned
3ω
50 ( 512,14
film
(HiPco) and multiwalled carbon nanotubes (MWCNTs) synthesized by catalytic chemical vapor deposition (CCVD) or thermal CVD (TCVD). We obtained arc-discharge SWCNTs, HiPco SWCNTs, CCVD MWCNTs (NC3100, 95% purity), and TCVD MWCNTs (CM-100, 95% purity) from Topnanosys Co., Carbon Nanotechnologies Inc., Nanocyl Inc., and Hanwha Nanotech Co., respectively. Raman spectroscopy relies on Raman scattering of monochromatic light, usually from a laser. The laser light interacts with molecular vibrations, phonons, or other excitations in a system, and the energy of the laser is shifted up or down, giving information about the system. It is well-known that G-peak in Raman spectra is shifted due to the temperature change.2430 The power of the laser used in the Raman spectroscopy is high enough to heat the CNT film. Hence, in this work, we used the shift of the G-peak and the laser as a temperature sensing probe and a way of heating, respectively, to determine the two-dimensional thermal conductivity of the CNT films. Detailed procedures are shown in Figure 2. The transmittance of CNT films were measured by using a laser power meter (Nova, Ophir) with 514 nm laser excitation (Spectra Physics). The ratio of transmitted laser power, which was collected by laser power meter, to output laser power was defined to be the transmittance of the freestanding CNT films. We checked the transmittance of the CNT films for various output laser powers and found that there is no saturated absorption in our measurement. The thicknesses of CNT films were measured by using atomic force microscopy (AFM) (XE-100, Park Systems) as shown in Figure 1b. Raman spectra were captured by using a Raman microscope (inVia Raman Microscope, Renishaw) operated by a software (WiRE 3.2, Renishaw). In our measurement, grating scan type, spectrum center, and exposure time are set to static, 1590 cm1, and 30 s, respectively. The G-peak shift according to various CNT film temperatures was measured, and then the G-peak shift for various laser heating powers which varied from 1 to 6 mW was obtained. From the G-peak shift data, we obtained the film temperature according to the laser power.1620 Finally, we calculated the two-dimensional thermal conductivity of CNT films based on the experimental results.
’ RESULTS AND DISCUSSION The G-peak shift according to temperature change has been reported by several researchers. The origin of the frequency shifts in Raman spectra of CNTs with increasing temperature has been reported to be thermal expansion2427 and softening of CC bonds.28,29 Figure 3a shows the G-peak shift according to the temperature change for HiPco sample. The temperature coefficient of HiPco sample obtained in this study is close to the value reported in the previous work as shown in Table 2, and so are those of other CNT films.24,25,28,30 All the G-peak shift data show similar temperature dependence regardless of the CNT type. The Raman shift of the G-peak shows universal temperature dependence. Thus, the temperature of CNT film could be measured from the Raman shift of the G-peak. Figure 3b shows the G-peak shift according to the laser power change for HiPco samples. From Figure 3a and b, we could obtain the temperature change of the CNT film according to the laser power change as shown in Figure 3c. It is clear that the temperature of the film center linearly increases as the laser power increases. We obtained similar results for films made of other types of CNTs (see Supporting Information Figures S1S3). When the suspended CNT film is tested in the air, there would be three modes of heat transfer: conduction, natural convection, and radiation. In this study, the heat from the laser is assumed to be dissipated only by the thermal conduction through the suspended CNT film. From the energy balance between the heat from the laser and the conduction heat transfer, the film temperature profile, T(r), and the film temperature captured by the Raman spectroscopy, Tm, could be predicted as follows16 TðrÞ ¼ T1 þ
Q 1 lnðR=rÞ fEið R2 =r02 Þ Eið r 2 =r02 Þg 2πkt 2
ð1Þ Z Tm ¼
Q πkt
R 0
1 lnðR=rÞ fEið R2 =r02 Þ Eið r 2 =r02 Þg expð r 2 =r02 Þr dr 2 r02 f1 expð R 2 =r02 Þg
þ T1
ð2Þ
where k, t, Q, T1, R, and r0 represent thermal conductivity and thickness of the CNT film, the heat from the laser, the edge temperature and the radius of the suspended CNT film, and the radius of the laser beam, respectively. Ei(x) is the exponential integral. The absorbed heat was estimated from the output laser power and the optical transmittance of the CNT film. The size of the suspended CNT film was measured from the micrograph of the CNT film. The laser beam size was estimated to be 1.36 μm from the numerical aperture of the object lens of the Raman microscope.16 From the rate of temperature rise according to absorbed heat increase and eq 2, we could determine the thermal conductivity. To validate the assumption of negligible convection and radiation, we estimated heat transfer rate from the CNT film dissipated by the natural convection and the radiation. The pattern of the natural convection changes according to the Rayleigh number, Ra, which is the ratio of buoyancy forces and thermal and momentum diffusivities.31 Ra ¼
8βΔTgR 3 να
ð3Þ
β, ν, α, ΔT, and g represent the thermal coefficient of volume expansion, the kinematic viscosity and the thermal diffusivity of 14533
dx.doi.org/10.1021/la203007y |Langmuir 2011, 27, 14532–14538
Langmuir
ARTICLE
Figure 1. Prepared CNT film: (a) CNT film on a perforated substrate; (b) AFM result for CNT film (arc-discharge SWCNT, 51.5% transmittance).
from eq 1 as follows: 2Z R TðrÞr dr T1 R2 0 Q Z R 1 2 2 2 2 fEið R ¼ lnðR=rÞ =r Þ Eið r =r Þg r dr 0 0 πktR 2 0 2
ΔT ¼
ð4Þ For the present system, Ra is below 0.001. When the Rayleigh number is small, the Nusselt number for the upper surface and that for the lower surface of the CNT film are almost same.32 For this case, Martorell et al. give the average Nusselt number as follows33 ! 2hR Nu ¼ ð5Þ ¼ 0:78 þ 0:82Ra1=5 kf Figure 2. Procedures for measurement of CNT film thermal conductivity.
the air, the mean temperature difference between the CNT film and the air, and the gravitational acceleration, respectively. If the edge temperature of the CNT film, T1, is assumed to be independent of the heat from the laser, Q, and if T1 is assumed to be close to the ambient temperature, ΔT could be obtained
where h and kf represent the average heat transfer coefficient and the thermal conductivity of the air, respectively. From eq 5, we could estimate the average heat transfer coefficient. Convective heat transfer rate from each surface of the circular suspended CNT film is πR2hΔT. If the CNT film is assumed to be black surface, the radiant heat transfer from the CNT film would be determined by the following equation31 4 4 Qrad ¼ 2πR 2 σðTfilm Tamb Þ
14534
ð6Þ
dx.doi.org/10.1021/la203007y |Langmuir 2011, 27, 14532–14538
Langmuir
ARTICLE
Figure 3. Experimental results from optical method (HiPco SWCNT films): (a) G-peak shift according to film temperature; (b) G-peak shift according to absorbed heat; (c) film temperature according to absorbed heat.
Table 2. Temperature Coefficient for G-Mode of CNTs temperature coefficient for G-mode (cm1/K)
a
CNT type
present study
SWCNT arc-discharge
0.0388
0.023, 0.042,25 0.04430
HiPco
0.0268
0.018928
MWCNT CCVD
0.0467
0.02824
TCVD
0.0379
a
No information could be found.
previous reports 24
where σ, Tfilm, and Tamb represent the StefanBoltzmann constant, the film temperature, and the ambient temperature, respectively. By using the above equations and the experimental results on the CNT film temperatures according to various laser powers shown in Figure 3c, we calculated heat transfer rates driven by the natural convection, the radiation, and the conduction for a HiPco SWCNT film. Figure 4 is the result of the calculation, and it clearly shows that more than 96% of the heat from the laser is dissipated by the thermal conduction. We could confirm that the thermal conduction is the predominant heat 14535
dx.doi.org/10.1021/la203007y |Langmuir 2011, 27, 14532–14538
Langmuir
ARTICLE
Figure 4. Heat transfer rate dissipated by each heat transfer mode (HiPco SWCNT film with 50.5% transmittance).
Table 3. Thermal Conductivity of CNT Films Obtained in This Study thermal CNT
transmittance
thickness
R
conductivity
type
(%)
(nm)
(μm)
(W/mK)
78.6
55
23
65.9
59.5 51.2
106 128
29 19
58.6 58.5
46.0
142
500
70.9
66.0
69
19
32.9
59.0
91
21
31.3
50.5
108
23
26.8
70.0
84
26
15.9
60.5
102
30
19.8
53.5 72.0
123 72
28 23
18.8 23.9
62.0
98
25
22.5
48.5
132
23
21.5
34.0
170
27
24.2
16.0
216
22
20.3
12.5
225
23
24.5
8.5
235
27
24.6
0.0
93000
2500
SWCNT arcdischarge
HiPco
MWCNT
CCVD
TCVD
Figure 5. (a) Raman spectra and (b) TEM images of tested CNTs.
Table 4. Effect of ID/IG ratio and CNT Diameter on Thermal Conductivity of CNT Films
0.203
transfer mode also for films made of other types of CNTs (see Supporting Information Figure S4). This finding is in good agreement with the well-known fact that heat transfer is primarily in the form of conduction for the low Rayleigh number regime.32,33 Therefore, in the present system, the main heat transfer mode is the conduction. Table 3 summarizes thermal conductivity values of the CNT films obtained in this study. Arc-discharge SWCNT films show the best heat spreading capability, and their thermal conductivity is in the range of reported values for random network of arcdischarge SWCNTs.7 However, even the best thermal conductivity is lower than that of aluminum and copper which are widely used materials in the field of thermal management. Even though the as-prepared CNT films are not excellent heat spreaders, several postprocesses developed to improve electrical conductance of the CNT films could be helpful in enhancing the thermal conductivity. For that issue, intensive research work would be required. To clarify important parameters affecting the thermal conductivity, we checked crystallinity and diameter of CNTs
CNT ID/IGa
type SWCNT
MWCNT a
arc-
0
CNT
thermal
diameter
conductivity
(nm)a
(W/mK)a
1.4 ( 0.2
63.5 ( 9.7
discharge HiPco
0.059 ( 0.009
1.2 ( 0.1
30.4 ( 5.1
CCVD
1.594 ( 0.318
11.6 ( 0.1
18.2 ( 7.9
TCVD
0.822 ( 0.062
17.1 ( 0.1
24.1 ( 2.8
With uncertainty of measurements for a 95% level of confidence.
tested in this work as shown in Figure 5 and Table 4. The ratio of the disorder-induced D-band to the Raman tangential G-band intensity, ID/IG, is often used to compare the graphitic ordering in carbon materials. The decrease of the ID/IG ratio has been shown to correlate with higher crystallinity and higher thermal conductivity for carbon.34,35 Figure 5a and Table 4 show that, in the order of arc-discharge SWCNT, HiPco SWCNT, TCVD MWCNT, and CCVD MWCNT, the ID/IG ratio increases and 14536
dx.doi.org/10.1021/la203007y |Langmuir 2011, 27, 14532–14538
Langmuir
ARTICLE
Table 5. Uncertainty in the Thermal Conductivity Measurement uncertainty in uncertainty in
G-peak shift
uncertainty
measurement
in thermal
G-peak shift measurement for absorbed conductivity for temperature variation heat variation evaluation sample
(%)
(%)
(%)
9.7
10.4
20.1
HiPco SWCNT
7.6
4.6
12.2
film (50.5%) CCVD
9.0
6.1
15.1
11.4
14.4
25.8
arc-discharge SWCNT film (51.2%)
MWCNT film (53.5%) TCVD MWCNT film (48.5%)
the thermal spreading capability decreases. According to Yu et al., phonon mean free path decreases as the diameter of CNT increases.2 However, Table 4 shows that the larger the diameter of CNTs, the higher the thermal conductivity of the CNT film when the CNTs are categorized only by the number of CNT walls. This means that the crystallinity of CNTs is a predominant parameter affecting the thermal conductivity of CNTs in comparison with the CNT diameter. In this study, the edge temperature of the suspended CNT film, T1 , was assumed to be independent of the absorbed heat from the laser, Q. To validate the assumption, we checked the thermal conductivity result for a much larger CNT film for which T1 is almost independent of Q as shown in Table 3. For the tested ratio of the CNT film radius to the laser beam size exceeding 13, the size of the CNT film did not affect the thermal conductivity value. In addition, we measured the thermal conductivity of CNT films with various transmittances and found that the effect of transmittance of films on the value of thermal conductivity is negligible when the light can penetrate the CNT film. However, when the CNT film is not transparent, the determined thermal conductivity seriously deviates from the thermal conductivity of transparent CNT films. Therefore, the present method is applicable, as long as the sample film is transparent. In this study, the whole cross-sectional area is assumed to be used as heat transfer path. However, when the film is not transparent, laser could not heat the whole cross-sectional area of the film and, due to the overestimated thickness of the heat transfer path, the thermal conductivity is underestimated as shown in Table 3. From the underestimated thermal conductivity value and eq 2, the real thickness of the heat transfer path of the opaque CNT buckypaper could be obtained. The estimated thickness is 783 nm, and this means that the 514 nm laser could penetrate the TCVD MWCNT film up to 783 nm. This example shows that, if we know the thermal conductivity of an opaque film, we can measure the penetration depth of a laser for the tested film. We calculated the uncertainty in the thermal conductivity evaluation connected to the uncertainty in the G-peak shift measurement. We first obtained the uncertainty in the G-peak shift data because the error in the thermal conductivity estimation
is affected by both the uncertainty in the G-peak shift measurement for temperature variation and that for absorbed heat change. The uncertainty values are summarized in Table 5. The errors in the thermal conductivity caused by the uncertainty in the G-peak shift measurement are comparable to the uncertainty values in the thermal conductivity measurement shown in Table 4. This means that the uncertainty in the G-peak shift measurement is the main source of the error. To reduce the uncertainty, we need to widen the test range of the film temperature and the heating laser power or we may increase the number of repeated measurement. Finally, it is worth mentioning that the conclusion on the best CNT type yielding highest film thermal conductivity is still valid even if we consider the uncertainty in this study.
’ CONCLUSIONS In this study, we have reported two-dimensional thermal conductivity of CNT films obtained by using the Raman method. We found that the arc-discharge SWCNT film shows the best heat spreading performance, and we ascribed it to the best crystallinity of the arc-discharge SWCNTs. The optical method based on the Raman spectroscopy renders reasonable thermal conductivity value as long as the sample is a transparent film. The optical method could be a useful thermal characterization tool for transparent films. Although the heat spreading performance of the as-prepared CNT films is not satisfactory, there is still room for improvement. It would not be absurd to anticipate that carbon-based heat spreaders would replace metal spreaders in the future. ’ ASSOCIATED CONTENT
bS
Supporting Information. G-peak shift according to film temperature; G-peak shift according to absorbed heat; film temperature according to absorbed heat and heat transfer rate dissipated by each heat transfer mode for films made of three kinds of CNTs (arc-discharge SWCNTs, CCVD MWCNTs, and TCVD MWCNTs). This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail
[email protected]; phone +82-42-868-7121; fax +82-42868-7884.
’ ACKNOWLEDGMENT This research was jointly supported by Center for Nanoscale Mechatronics & Manufacturing, one of the 21st Century Frontier Research Programs supported by Ministry of Education, Science and Technology, Korea and Platform Technology Program supported by Ministry of Knowledge Economy, Korea. ’ REFERENCES (1) Li, Q. W.; Liu, C. H.; Wang, X. S; Fan, S. S. Nanotechnology 2009, 20, 145702. (2) Yu, C. H.; Shi, L.; Yao, Z.; Li, D. Y.; Majumdar, A. Nano Lett. 2005, 5, 1842–1846. (3) Pop, E.; Mann, D.; Wang, Q.; Goodson, K.; Dai, H. J. Nano Lett. 2006, 6, 96–100. 14537
dx.doi.org/10.1021/la203007y |Langmuir 2011, 27, 14532–14538
Langmuir
ARTICLE
(4) Shi, L.; Li, D. Y.; Yu, C. H.; Jang, W. Y.; Kim, D. H.; Yao, Z.; Kim, P.; Majumdar, A. J. Heat Transfer 2003, 125, 881–888. (5) Hou, J. B.; Wang, X. W.; Vellelacheruvu, P.; Guo, J. Q. J. Appl. Phys. 2006, 100, 124314. (6) Hone, J.; Whitney, M.; Piskoti, C.; Zettl, A. Phys. Rev. B 1999, 59, R2514–R2516. (7) Itkis, M. E.; Borondics, F.; Yu, A. P.; Haddon, R. C. Nano Lett. 2007, 7, 900–904. (8) Choi, T. Y.; Poulikakos, D.; Tharian, J.; Sennhauser, U. Nano Lett. 2006, 6, 1589–1593. (9) Sihn, S. W.; Ganguli, S.; Roy, A. K.; Qu, L. T.; Dai, L. M. Compos. Sci. Technol. 2008, 68, 658–665. (10) Yang, D. J.; Zhang, Q.; Chen, G.; Yoon, S. F.; Ahn, J.; Wang, S. G.; Zhou, Q.; Wang, Q.; Li, J. Q. Phys. Rev. B 2002, 66, 165440. (11) Hu, X. J.; Padilla, A. A.; Xu, J.; Fisher, T. S.; Goodson, K. E. J. Heat Transfer 2006, 128, 1109–1113. (12) Aliev, A. E.; Lima, M. H.; Silverman, E. M.; Baughman, R. H. Nanotechnology 2010, 21, 035709. (13) Choi, T. Y.; Poulikakos, D.; Tharian, J.; Sennhauser, U. Appl. Phys. Lett. 2005, 87, 013108. (14) Aliev, A. E.; Guthy, C.; Zhang, M.; Fang, S. L.; Zakhidov, A. A.; Fischer, J. E.; Baughman, R. H. Carbon 2007, 45, 2880–2888. (15) Wei, L.; Kuo, P. K.; Thomas, R. L.; Anthony, T.; Banholzer, W. Phys. Rev. Lett. 1993, 70, 3764–3767. (16) Cai, W. W.; Moore, A. L.; Zhu, Y. W.; Li, X. S.; Chen, S. S.; Shi, L.; Ruoff, R. S. Nano Lett. 2010, 10, 1645–1651. (17) Chen, S. S.; Moore, A. L.; Cai, W. W.; Ji, W. S.; An, J. H.; Mishra, C.; Amos, C.; Magnuson, C. W.; Kang, J. Y.; Shi, L.; Ruoff, R. S. ACS Nano 2011, 5, 321–328. (18) Balandin, A. A.; Ghosh, S.; Bao, W. Z.; Calizo, I.; Teweldebrhan, D.; Miao., F.; Lau, C. N. Nano Lett. 2008, 8, 902–907. (19) Calizo, I.; Balandin, A. A.; Bao, W. Z.; Miao., F.; Lau, C. N. Nano Lett. 2007, 7, 2645–2649. (20) Ghosh, S.; Nika, D. L.; Pokatilov, E. P.; Balandin, A. A. New J. Phys. 2009, 11, 095012. (21) Perichon, S.; Lysenko, V.; Remaki, B.; Barbier, D.; Champagnon, B. J. Appl. Phys. 1999, 86, 4700–4702. (22) Lysenko, V.; Perichon, S.; Remaki, B.; Barbier, D.; Champagnon, B. J. Appl. Phys. 1999, 86, 6841–6846. (23) Wu, Z. C.; Chen, Z. H.; Du, X.; Logan, J. M.; Sippel, J.; Nikolou, M.; Kamaras, K.; Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.; Rinzler, A. G. Science 2004, 305, 1273–1276. (24) Huang, F. M.; Yue, K. T.; Tang, P. H.; Zhang, S. L.; Shi, Z. J.; Zhou, X. H.; Gu, Z. N. J. Appl. Phys. 1998, 84, 4022–4024. (25) Li, H. D.; Yue, K. T.; Lian, Z. L.; Zhan, Y.; Zhou, L. X.; Zhang, S. L.; Shi, Z. J.; Gu, Z. N.; Liu, B. B.; Yang, R. S.; Yang, H. B.; Zou, G. T.; Zhang, Y.; Iijima, S. Appl. Phys. Lett. 2000, 76, 2053–2055. (26) Zhang, Q.; Yang, D. J.; Wang, S. G.; Yoon, S. F.; Ahn, J. Smart Mater. Struct. 2006, 15, S1–S4. (27) Huong, P. V.; Cavagnat, R.; Ajayan, P. M.; Stephan, O. Phys. Rev. B 1995, 51, 10048–10051. (28) Raravikar, N. R.; Keblinski, P.; Rao, A. M.; Dresselhaus, M. S.; Schadler, L. S.; Ajayan, P. M. Phys. Rev. B 2002, 66, 235424. (29) Ci, L. J.; Zhou, Z. P.; Song, L.; Yan, X. Q.; Liu, D. F.; Yuan, H. J.; Gao, Y.; Wang, J. X.; Liu, L. F.; Zhou, W. Y.; Wang, G.; Xie, S. S. Appl. Phys. Lett. 2003, 82, 3098–3011. (30) Atashbar, M. Z.; Singamaneni, S. Appl. Phys. Lett. 2005, 86, 123112. (31) Mills, A. F. Heat and Mass Transfer; Irwin: Chicago, IL, 1995. (32) Suriano, F. J.; Yang, K.-T. Int. J. Heat Mass Transfer 1968, 11, 473–490. (33) Martorell, I.; Herrero, J.; Grau, F. X. Int. J. Heat Mass Transfer 2003, 46, 2389–2402. (34) Ivanov, I.; Puretzky, A.; Eres, G.; Wang, H.; Pan, Z. W.; Cui, H. T.; Jin, R. Y.; Howe, J.; Geohegan, D. B. Appl. Phys. Lett. 2006, 89, 223110. (35) Heremans, J.; Rahim, I.; Dresselhaus, M. S. Phys. Rev. B 1985, 32, 6742–6747.
14538
dx.doi.org/10.1021/la203007y |Langmuir 2011, 27, 14532–14538
