Electroluminescence from suspended and on-substrate metallic single-walled carbon nanotubes.

Abstract

In this work, we carried out electroluminescence (EL) measurements on metallic single-walled carbon nanotubes (SWNTs) and compared the light emission from the suspended section with the on-substrate section along the same SWNT. In addition to the lowest excitonic emission for metallic SWNTs (M(11)), a side peak was observed at an energy of 0.17-0.20 eV lower than the M(11) peak, which is attributed to a phonon-assisted sideband. Interestingly, this side peak was only observed from on-substrate sections but not from suspended sections. This is likely due to the higher electric field used in the EL measurement of on-substrate sections and the asymmetric surroundings for on-substrate SWNT sections. When the drain voltage is increased, either a blue shift or a red shift of the M(11) emission (up to +/-20 meV) was observed in different suspended SWNTs. The red shift can be explained by the temperature-dependence of the M(11) transition energy, whereas the blue shift is surprising and has never been observed before. Some possible mechanisms for the blue shift are discussed.

Full text

Page 1

Electroluminescence fromSuspended
and On-Substrate Metallic Single-Walled
Carbon Nanotubes
Liming Xie,†,|Hootan Farhat,‡Hyungbin Son,†Jin Zhang,*,|
Mildred S. Dresselhaus,§,†Jing Kong,*,†and Zhongfan Liu|
Department of Electrical Engineering and Computer Science, Department of Materials
Science and Engineering, Department of Physics, Massachusetts Institute of
Technology, Cambridge, Massachusetts 02139, and Beijing National Laboratory for
Molecular Sciences, College of Chemistry and Molecular Engineering, Peking
UniVersity, Beijing 100871, People’s Republic of China
Received October 3, 2008; Revised Manuscript Received January 17, 2009
ABSTRACT
Inthiswork, wecarriedout electroluminescence(EL) measurementsonmetallicsingle-walledcarbonnanotubes(SWNTs) andcomparedthe
light emissionfromthe suspendedsectionwiththe on-substrate sectionalongthe same SWNT. Inadditiontothe lowest excitonic emission
for metallic SWNTs (M11), a side peak was observed at an energy of 0.17-0.20 eVlower than the M11peak, which is attributed to a phonon-
assistedsideband. Interestingly, thissidepeakwasonlyobservedfromon-substratesectionsbut not fromsuspendedsections. Thisislikely
duetothehigherelectricfieldusedintheELmeasurementofon-substratesectionsandtheasymmetricsurroundingsforon-substrateSWNT
sections. Whenthedrainvoltageisincreased, either ablueshift or aredshift of theM11emission(upto(20meV) wasobservedindifferent
suspended SWNTs. The red shift can be explained by the temperature-dependence of the M11transition energy, whereas the blue shift is
surprising and has never been observed before. Some possible mechanisms for the blue shift are discussed.
With their unique electronic properties and one-dimensional
structure, single-walled carbon nanotubes (SWNTs) have
potential applications in future nanoelectronics and nano-
photonics. Photoluminescence (PL) and electroluminescence
(EL) have been intensively investigated to understand the
fundamental photophysics in SWNTs.1-6In both PL and EL
studies, substrates have played an important role in influenc-
ing the properties of SWNTs.2,7-9Up to now, PL has not
been observed from semiconducting SWNTs directly grown
on substrates,2and suspended semiconducting SWNTs give
a much higher PL efficiency10than surfactant-wrapped
semiconducting SWNTs.5On the other hand, EL has been
observed from all SWNTs (semiconducting or metallic,
suspended or on-substrate).1,4,11Thus EL is especially useful
to investigate the fundamental photophysics of metallic
SWNTs as well as substrate effects. Since the EL properties
of different SWNT devices (either suspended or on-substrate)
are dramatically different,1,3,4comparing the EL spectra from
suspended and on-substrate sections from the same SWNT
will provide further insight into the fundamental physics of
SWNTs.
In the present work, we measured the EL emission from
suspended and on-substrate sections along the same metallic
SWNT (Figure 1). While the suspended section exhibits a
stronger EL signal that turns on at a lower drain voltage Vds,
the on-substrate section has an additional side peak at
0.17-0.20 eV below the lowest excitonic transition (M11).
This side peak is assigned to phonon-assisted emission.
Moreover, in suspended sections, we have observed an
interesting blue shift of the EL emission energy at higher
* To whom correspondence should be addressed. E-mail: (J.Z.)
jinzhang@pku.edu.cn; (J.K.) jingkong@mit.edu.
†Department of Electrical Engineering and Computer Science, Mas-
sachusetts Institute of Technology.
‡Department of Materials Science and Engineering, Massachusetts
Institute of Technology.
§Department of Physics, Massachusetts Institute of Technology.
|Peking University.
Figure 1. Schematic illustration (a) and scanning electron micros-
copy image (b) of suspended and on-substrate devices fabricated
along the same SWNT.
NANO
LETTERS
2009
Vol. 9, No. 5
1747-1751
10.1021/nl803004mCCC: $40.75
Published on Web 03/30/2009
 2009 American Chemical Society
Downloaded by BEIJING UNIV on October 23, 2009 | http://pubs.acs.org
Publication Date (Web): March 30, 2009 | doi: 10.1021/nl803004m

Page 2

drain voltages, which cannot be explained by the temperature
dependence of the transition energy M11.
Devices with on-substrate and suspended sections along
the same individual SWNTs were fabricated either by
transfering12or by directly growing ultralong SWNTs by
chemical vapor deposition (CVD)13onto substrates with
prefabricated trenches and electrode patterns. These sub-
strates are the same as those used in ref 11. Briefly, a Si3N4
film (thickness ∼ 38 nm) was deposited on a SiO2/Si wafer
(oxide thickness ) 500 nm). Trenches were then photolitho-
graphically defined and etched through the Si3N4on SiO2.
Finally, Pt electrodes (thickness ∼ 25 nm) were patterned
by lift-off. The CVD growth of ultralong SWNTs was carried
out by an Fe-assisted ethanol CVD method.13The SWNT
transfer was done by a PMMA-mediated transfer method.12
The EL measurement was achieved by a home-built setup.
A 50× aspheric lens was used to collect the EL signal, and
a 300 line/mm grating and a 512-pixel InGaAs detector
(Hamamatsu G9204-512D, spectrum range 900-1700 nm)
were used to disperse the signal and record the spectra. A
typical integration time of 60-120 s was used for collecting
the spectra. All EL measurements were conducted under an
argon atmosphere.
Figure 1a is a schematic illustration of a metallic SWNT
device with one section on the Si3N4substrate and the other
section suspended. Figure 1b is a scanning electron micro-
scope (SEM) image of a typical device. The highly doped
Si substrate was used as a back gate. The channel lengths of
the devices are about 1 µm. The trench depth is about 300
nm. Figure 2a shows a typical drain-source current Idsversus
drain-source voltage Vdsplot of our devices. Due to electron-
phonon scattering by optical phonons,9for the on-substrate
section, Ids saturates at high Vds, and for the suspended
section, a negative differential conductance occurs at Vds>
1.2 V. The inset in Figure 2a shows the drain current Ids
versus gate voltage (Vgs) curve (under constant bias voltage
Vds) 10 mV) for the same SWNT (the suspended section),
which is a typical Ids-Vgscurve for a metallic SWNT. In
order to measure EL from individual SWNTs, but not from
multiwalled carbon nanotubes or nanotube-bundles, we used
the nanotube devices in which the on-substrate sections have
saturation currents of about 25 µA.14,15The corresponding
EL spectra from this SWNT in Figure 2a are presented in
Figure 2b. For the suspended section, only one EL peak was
observed centered at 1.054 eV at Vds) 2.0 V and Vgs)
-20 V. Using the diameter information (∼2.3 nm) obtained
by atomic force microscopy (AFM) measurements, the 1.054
eV peak from the suspended section can be assigned to the
M11 excitonic transition.16There could be an emission
component from the M11continuum that is embedded in the
M11exciton emission since the exciton binding energy (∼30
meV for 2.3 nm diameter metallic SWNTs, estimated from
the 50 meV reported for a (10,10) SWNT17) is less than both
the thermal energy (>60 meV11) and the EL emission peak
width (∼100 meV).
In contrast, for the on-substrate section, two EL peaks
(0.865 and 1.036 eV) were observed. Measurements from
two other SWNT devices showed similar phenomena,
namely, the suspended sections only displayed the M11
emissions, whereas the on-substrate sections exhibited two
emission peaks, one dominant peak near M11of the sus-
pended section and the other one lying 0.17-0.20 eV lower
in energy. For the on-substrate sections, a higher Vds is
required for observable EL emission. Since the exciton
binding energy for a 2.3 nm diameter metallic SWNT is
about 30 meV, the two peaks in the emission spectrum of
the on-substrate section cannot together be identified as the
M11exciton transition and the M11continuum. Additionally,
the peak separation does not correspond to the trigonal-
warping induced splitting of the M11transition which is less
than 70 meV for a 2.3 nm diameter metallic SWNT.18
Considering that the energy difference between the two peaks
(∼0.17 eV) is close to a Γ-point optical phonon energy (∼0.2
eV) and a K-point phonon energy (0.16 eV), the lower energy
side peak is assigned to an optical phonon-assisted emission
(either from the Γ or K point. From our later discussion, it
appears that a Γ-point optical phonon-assisted emission is
more likely). Thus the higher energy peak is assigned to the
M11emission. Additionally, the M11emission from the on-
Figure 2. (a) Drain current (Ids) vs drain voltage (Vds) plot for
suspended (red line) and on-substrate (black line) sections along
the same SWNT. The inset shows the drain current vs gate voltage
(Vgs) plot for the suspended section. (b) Electroluminescence (EL)
from the suspended section (red circle) at Vds) 2.0 V and the on-
substrate section (black circle) at Vds) 5.6 V. A gate voltage Vgs
) -20 V is used in all EL measurements. Data are offset for clarity.
Red lines are a Lorentzian fit to the results and green lines are
fitting curves for the individual peaks.
1748
Nano Lett., Vol. 9, No. 5, 2009
Downloaded by BEIJING UNIV on October 23, 2009 | http://pubs.acs.org
Publication Date (Web): March 30, 2009 | doi: 10.1021/nl803004m

Page 3

substrate section was found to be redshifted with respect to
the suspended section as will be discussed later.
The drain current and emission intensity versus drain
voltage are plotted in Figure 3. For suspended sections,
similarly to the observation in ref 11, EL was observed in
the negative differential conductance region, where Ids
decreases with Vdsbut the EL intensity increases with Vds
(Figure 3a). In contrast to the suspended SWNTs, Idsfor the
on-substrate SWNT section used in Figure 2 continues to
increase with Vdsunder a large bias of up to 5.6 V, as shown
in Figure 3b. But the EL intensity dependence on Vdsfor
the on-substrate section is different from that observed for
the suspended sections. The emission intensity for both the
phonon-assisted transition (EP-A) and the M11emission for
the on-substrate section increased slowly in the range Vds)
5.0-5.4 V and then increased more rapidly in the range Vds
) 5.4-5.6 V.
Now we shift our attention to the origin of the phonon-
assisted emissions. This phonon-assisted sideband was only
observed from the on-substrate sections. Before discussing
this difference, the EL mechanisms for suspended and on-
substrate SWNTs should be discussed in more detail.
Previous works11,19have shown that the EL from suspended
metallic SWNTs occurs via a thermal light emission mech-
anism. In this case, hot electrons are thermally distributed
in higher energy bands. However, for on-substrate metallic
SWNTs, no EL mechanism has previously been reported.
Since suspended SWNTs give observable thermal light
emission above 800 K,11and on-substrate metallic SWNTs
can be heated up to 900 K in air before they burn20(in an
Ar atmosphere, the same on-substrate SWNT devices can
go up to an even higher Vds, indicating that on-substrate
SWNTs can go up to a higher temperature), thermal light
emission should also be possible for on-substrate metallic
SWNTs. Impact excitation, which occurs under a high
electrical field,1,21is another possible mechanism for EL
emission from on-substrate metallic SWNTs. The average
electrical field in our EL experiment is about 5 V/µm. The
electrical field in the regions of the contacts, defects, and
impurities may be higher.
For the energy difference between the sideband and the
M11emission, two other SWNTs give an energy difference
of 0.20 and 0.17 eV. Since different exciton levels in 1 nm
diameter semiconducting SWNTs only have an energy
difference of <40 meV22,23and the exciton binding energy
in metallic SWNTs is much smaller than that in semicon-
ducting SWNTs,17,24the energy difference for different
exciton levels is much smaller in 2 nm diameter metallic
SWNTs. Therefore, the K-point phonon-assisted E symmetry
exciton emission should appear at slightly less than 0.16 eV
lower energy than the bright exciton emission23and the
Γ-point phonon-assisted emission should appear at slightly
less than 0.20 eV lower than the bright exciton emission.
Therefore, the sideband observed in our EL measurement is
more likely to be the Γ-point phonon-assisted emission. The
smaller observed energy difference (0.17 eV) relative to 0.2
eV is attributed to the kinetic energy of excitons.
Figure 4 depicts the processes of the M11emission (afb)
and the phonon-assisted emission (cfdfb). The intensity
of the phonon-assisted emission is related to the number of
excitons distributed above the M11band edge with nonzero
momentum and the efficiency for these excitons to emit via
the phonon-assisted process. For the on-substrate sections,
there may be more higher-energy excitons with nonzero
Figure 3. (a) Drain current (Ids) and EL intensity plot vs drain
voltage (Vds) for two suspended SWNTs A and B. (b) Drain current
(Ids) and EL intensity plot vs drain voltage (Vds) for the same on-
substrate section of the SWNT used in Figure 2. M11and EP-Adenote
the lowest excitonic transition and phonon-assisted emission,
respectively.
Figure 4. Schematic diagram for direct M11emission (afb) and
phonon-assisted emission (cfdfb). Black dots denote excitons.
The phonon-assisted emission energy is M11+ ∆E - Eph, where
M11is the lowest excitonic transition energy for metallic SWNTs,
∆E is the average energy difference between the excitons at “c”
and the excitons at “a”, and Ephis the phonon energy.
Nano Lett., Vol. 9, No. 5, 20091749
Downloaded by BEIJING UNIV on October 23, 2009 | http://pubs.acs.org
Publication Date (Web): March 30, 2009 | doi: 10.1021/nl803004m

Page 4

momentum because a higher electric field was used in the
EL measurements of the on-substrate sections, and this higher
electric field will accelerate the carriers more and may create
more excitons with higher energy and nonzero momentum.21
Meanwhile, the suspended metallic SWNTs may have a very
low phonon-assisted emission efficiency judging from PL
studies of semiconducting SWNTs. Up to now there has been
no report of phonon-assisted PL from suspended semicon-
ducting SWNTs. But phonon-assisted PL emission has been
reported from surfactant-wrapped or DNA-wrapped semi-
conducting SWNTs.25,26It is possible that due to their
symmetric and unperturbed surroundings, suspended semi-
conducting SWNTs have a low phonon-assisted efficiency.
This may be the same case for our suspended metallic
SWNTs, which will account for the absence of phonon-
assisted EL emission.
The phonon-assisted emission energy EP-A can be ex-
pressed as M11+ ∆E - Eph, where ∆E is the average exciton
energy relative to the zero-momentum exciton at the M11
edge and Ephis the phonon energy. Because of the nonzero
value of ∆E, the energy difference between EP-Aand M11
should be less than Eph, which is consistent with the
experimental data. At a lower Vds, ∆E is expected to be
smaller since in this case there are fewer high-energy excitons
in the system. Therefore the energy difference between EP-A
and M11 is expected to be larger at lower Vds. This is
confirmed by our observations (Figure 5). At Vds) 5.0 V,
the energy difference between the EP-Aand the M11emission
is 0.19 eV. When Vdsis changed to 5.6 V, it can be seen
that the M11 emission is redshifted by 20 meV and the
difference between the two peaks becomes 0.17 eV. The red
shift of the M11emission could be due to the increase of the
lattice temperature at higher Vds.20
The energy difference between the M11emissions from
the suspended and the on-substrate sections, as shown in
Figure 2b, can be attributed to environmental effects (dif-
ferences in the dielectric constant of the SWNT surroundings,
substrate interaction, etc.), dark-bright exciton mixing and/
or the temperature difference between the suspended and the
on-substrate sections under the EL experimental conditions.
Because the dielectric constant of Si3N4 (ε ) 7.5) is
significantly larger than that of air (ε ) 1) and the transition
energy of SWNTs is smaller in higher dielectric constant
surroundings,7,27substrates can induce a red shift of M11.
Van der Waals interactions with the substrates can induce
an elastic deformation in the SWNTs28and make their
transition energies shift (tens of meV).29The asymmetric
surrounding and higher electric field required for the emission
can introduce a perturbation to the on-substrate sections. Such
perturbations can promote dark-bright exciton mixing and
also can induce a red shift of the transition energy (by several
meV).30,31The temperatures for the suspended and on-
substrate sections are unknown and the temperature differ-
ence may contribute up to a -10 meV shift in the M11
transition per 100 K temperature increase.32
For the suspended sections, some SWNTs show a red shift
and others show a blue shift of the EL peak starting from
the Vdsat the onset of the EL emission and going to the Vds
point just before the SWNTs are burnt out (Figure 6a,b).
All EL emission from suspended sections was measured in
the negative differential conductance region. Different
SWNTs give different maximum energy shifts (between (20
meV). For the suspended SWNTs, the lattice temperature
of the SWNTs increases with increasing Vds11and the
transition energies (such as M11) decrease with increasing
temperature.32Therefore, in some suspended SWNTs the
temperature dependence of the transition energies may
dominate the EL emission shift, which gives a red shift of
the EL emission with increasing Vds. In other SWNTs, other
factors may dominate the EL emission shift. Noticing that
the EL emission peaks from the suspended section are
Figure 5. EL from the on-substrate section of the SWNT used in
Figure 2 at different drain voltages. Red line for Vds) 5.6 V, black
line for Vds)5.0 V. (For both data sets Vgsis at -20 V.) Blue
lines are a Lorentzian fit to the emissions and green lines are fitting
curves for the individual peaks.
Figure 6. Red shift (a) and blue shift (b) of the EL emission position
with increasing drain voltage observed for suspended sections of
two different metallic SWNTs. The dashed red lines are guidance
for the eyes. From bottom to top in panel a, Vds) 2.0-2.4 V with
steps of 0.1 V; in panel b, Vds) 1.0-1.6 V with steps of 0.1 V.
For all data sets Vgsis at -20 V. (c) A two-Lorentzian peak fit to
the emission profiles from the suspended SWNT in panel b; from
bottom to up, Vds) 1.3-1.6 V with a step of 0.1 V.
1750
Nano Lett., Vol. 9, No. 5, 2009
Downloaded by BEIJING UNIV on October 23, 2009 | http://pubs.acs.org
Publication Date (Web): March 30, 2009 | doi: 10.1021/nl803004m

Page 5

asymmetric, the emission profile can be fit by two compo-
nents: a low-energy component and a high-energy component
(Figure 6c). As Vdsincreases, the intensity of both the low-
energy component and the high-energy component increase
but the energy of the peak positions of these two components
vary only slightly compared to the shift (Figure 6c). However
the intensity ratio for the high-energy component to the low-
energy component increases with Vds(from 1.3 V to 1.5 V),
which gives rise to a blue shift of the overall spectrum.
Therefore, this blue shift is due to the relatively higher
intensity of the high-energy component. Since suspended
SWNTs have a higher temperature at a higher Vds,11the
relatively higher intensity of the high-energy component may
correspond to relatively more thermally excited excitons
occupying the higher energy level. The higher energy level
could be identified as either the M11continuum and/or the
M11+exciton level (the upper component of the trigonal
warping-split M11transition). For the SWNT shown in Figure
6c, the M11+exciton emission is the most likely explanation
judging from the energy difference between these two
emission components (∼50 meV).
In conclusion, we have carried out EL measurements on
suspended and on-substrate sections of the same metallic
SWNTs. A phonon-assisted emission was found from the
on-substrate sections, which may be due to the higher electric
field and the asymmetric surroundings for the on-substrate
SWNT sections. In the suspended SWNT sections, some
SWNTs showed a red shift and others showed a blue shift
of the EL peak position with increasing Vds. The blue shift,
which cannot be understood by a temperature-dependent
transition energy, could be caused by the fine features of
the emission spectrum at a higher temperature, either by
emission from the M11continuum and/or from the upper
component M11+of the trigonal warping-split transition
energies.
Acknowledgment. This work was partially supported by
the Materials, Structure, and Device Center, one of the five
centers of the Focus Center Research Program (FCRP) and
NSF/DMR 07-04197. The authors acknowledge Mr. Xinran
Wang, Professor Hongjie Dai, Professor Tony Heinz, Profes-
sor Riichiro Saito, and Dr. Georgy Samsonidze for helpful
discussions, and Dr. Qian Wang and Professor Hongjie Dai
for providing substrates with prepatterned electrodes. L.X.
acknowledges a Scholarship from the China Scholarship
Council, the Peking University CDY Scholarship. J.K. and
J.Z. acknowledges NSFC 20828004.
References
(1) Chen, J.; Perebeinos, V.; Freitag, M.; Tsang, J.; Fu, Q.; Liu, J.; Avouris,
P. Science 2005, 310 (5751), 1171–1174.
(2) Lefebvre, J.; Homma, Y.; Finnie, P. Phys. ReV. Lett. 2003, 90 (21),
217401.
(3) Marty, L.; Adam, E.; Albert, L.; Doyon, R.; Menard, D.; Martel, R.
Phys. ReV. Lett. 2006, 96 (13), 136803.
(4) Misewich, J. A.; Martel, R.; Avouris, P.; Tsang, J. C.; Heinze, S.;
Tersoff, J. Science 2003, 300 (5620), 783–786.
(5) O’Connell, M. J.; Bachilo, S. M.; Huffman, C. B.; Moore, V. C.;
Strano, M. S.; Haroz, E. H.; Rialon, K. L.; Boul, P. J.; Noon, W. H.;
Kittrell, C.; Ma, J. P.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E.
Science 2002, 297 (5581), 593–596.
(6) Xie, L. M.; Liu, C.; Zhang, J.; Zhang, Y. Y.; Jiao, L. Y.; Jiang, L.;
Dai, L.; Liu, Z. F. J. Am. Chem. Soc. 2007, 129 (41), 12382–12383.
(7) Walsh, A. G.; Vamivakas, A. N.; Yin, Y.; Cronin, S. B.; Unlu, M. S.;
Goldberg, B. B.; Swan, A. K. Nano Lett. 2007, 7 (6), 1485–1488.
(8) Pop, E.; Mann, D.; Cao, J.; Wang, Q.; Goodson, K.; Dai, H. J. Phys.
ReV. Lett. 2005, 95 (15), 155505.
(9) Yao, Z.; Kane, C. L.; Dekker, C. Phys. ReV. Lett. 2000, 84 (13), 2941–
2944.
(10) Lefebvre, J.; Austing, D. G.; Bond, J.; Finnie, P. Nano Lett. 2006, 6
(8), 1603–1608.
(11) Mann, D.; Kato, Y. K.; Kinkhabwala, A.; Pop, E.; Cao, J.; Wang,
X. R.; Zhang, L.; Wang, Q.; Guo, J.; Dai, H. J. Nat. Nanotechnol.
2007, 2 (1), 33–38.
(12) Jiao, L.; Fan, B.; Xian, X.; Wu, Z.; Zhang, J.; Liu, Z. J. Am. Chem.
Soc. 2008, 130 (38), 12612–12613.
(13) Zheng, L. X.; O’Connell, M. J.; Doorn, S. K.; Liao, X. Z.; Zhao, Y. H.;
Akhadov, E. A.; Hoffbauer, M. A.; Roop, B. J.; Jia, Q. X.; Dye, R. C.;
Peterson, D. E.; Huang, S. M.; Liu, J.; Zhu, Y. T. Nat. Mater. 2004,
3 (10), 673–676.
(14) Collins, P. C.; Arnold, M. S.; Avouris, P. Science 2001, 292 (5517),
706–709.
(15) Javey, A.; Guo, J.; Paulsson, M.; Wang, Q.; Mann, D.; Lundstrom,
M.; Dai, H. J. Phys. ReV. Lett. 2004, 92 (10), 106804.
(16) Strano, M. S. J. Am. Chem. Soc. 2003, 125 (51), 16148–16153.
(17) Deslippe, J.; Spataru, C. D.; Prendergast, D.; Louie, S. G. Nano Lett.
2007, 7 (6), 1626–1630.
(18) Saito, R.; Dresselhaus, G.; Dresselhaus, M. S. Phys. ReV. B 2000, 61
(4), 2981–2990.
(19) Wang, X.; Zhang, L.; Lu, Y.; Dai, H.; Kato, Y. K.; Pop, E. Appl.
Phys. Lett. 2007, 91 (26), 261102.
(20) Pop, E.; Mann, D. A.; Goodson, K. E.; Dai, H. J. J. Appl. Phys. 2007,
101 (9), 093710.
(21) Perebeinos, V.; Avouris, P. Phys. ReV. B 2006, 74 (12), 121410.
(22) Srivastava, A.; Htoon, H.; Klimov, V. I.; Kono, J. Phys. ReV. Lett.
2008, 101 (8), 087402.
(23) Torrens, O. N.; Zheng, M.; Kikkawa, J. M. Phys. ReV. Lett. 2008,
101 (15), 157401.
(24) Wang, F.; Dukovic, G.; Brus, L. E.; Heinz, T. F. Science 2005, 308
(5723), 838–841.
(25) Lebedkin, S.; Hennrich, F.; Kiowski, O.; Kappes, M. M. Phys. ReV.
B 2008, 77 (16), 165429.
(26) Chou, S. G.; Plentz, F.; Jiang, J.; Saito, R.; Nezich, D.; Ribeiro, H. B.;
Jorio, A.; Pimenta, M. A.; Samsonidze, G. G.; Santos, A. P.; Zheng,
M.; Onoa, G. B.; Semke, E. D.; Dresselhaus, G.; Dresselhaus, M. S.
Phys. ReV. Lett. 2005, 94 (12), 127402.
(27) Miyauchi, Y.; Saito, R.; Sato, K.; Ohno, Y.; Iwasaki, S.; Mizutani,
T.; Jiang, J.; Maruyama, S. Chem. Phys. Lett. 2007, 442 (4-6), 394–
399.
(28) Hertel, T.; Walkup, R. E.; Avouris, P. Phys. ReV. B 1998, 58 (20),
13870–13873.
(29) Shan, B.; Lakatos, G. W.; Peng, S.; Cho, K. J. Appl. Phys. Lett. 2005,
87 (17), 173109.
(30) Shaver, J.; Kono, J.; Portugall, O.; Krstic, V.; Rikken, G. L. J. A.;
Miyauchi, Y.; Maruyama, S.; Perebeinos, V. Nano Lett. 2007, 7 (7),
1851–1855.
(31) Mortimer, I. B.; Nicholas, R. J. Phys. ReV. Lett. 2007, 98 (2), 027404.
(32) Cronin, S. B.; Yin, Y.; Walsh, A.; Capaz, R. B.; Stolyarov, A.;
Tangney, P.; Cohen, M. L.; Louie, S. G.; Swan, A. K.; Unlu, M. S.;
Goldberg, B. B.; Tinkham, M. Phys. ReV. Lett. 2006, 96 (12), 127403.
NL803004M
Nano Lett., Vol. 9, No. 5, 20091751
Downloaded by BEIJING UNIV on October 23, 2009 | http://pubs.acs.org
Publication Date (Web): March 30, 2009 | doi: 10.1021/nl803004m