Selective matching of catalyst element and carbon source in single-walled carbon nanotube synthesis on silicon substrates

Article abstract of DOI:10.1021/jp0454117

We have studied the compatibility of various catalysts for ethylene and ethanol chemical vapor deposition (CVD) syntheses of single-walled carbon nanotubes (SWNTs) on Si substrates. A strong selectivity between the catalyst elemental species and carbon source was found; SWNT yield for Fe (Co) catalysts was much higher for ethylene (ethanol) CVD than for ethanol (ethylene) CVD. This strong and completely opposite selectivity implies significantly different SWNT growth mechanisms for ethanol and ethylene CVD on Si substrates. ? 2005 American Chemical Society.

Full text of DOI:10.1021/jp0454117

2632
J. Phys. Chem. B 2005, 109, 2632-2637
Selective Matching of Catalyst Element and Carbon Source in Single-Walled Carbon
Nanotube Synthesis on Silicon Substrates
Kohei Mizuno, Kenji Hata,* Takeshi Saito, Satoshi Ohshima, Motoo Yumura, and
Sumio Iijima
Research Center for AdVanced Carbon Materials, National Institute of AdVanced Industrial Science and
Technology, Tsukuba, 305-8565, Japan
ReceiVed: October 7, 2004; In Final Form: NoVember 29, 2004
We have studied the compatibility of various catalysts for ethylene and ethanol chemical vapor deposition
(CVD) syntheses of single-walled carbon nanotubes (SWNTs) on Si substrates. A strong selectivity between
the catalyst elemental species and carbon source was found; SWNT yield for Fe (Co) catalysts was much
higher for ethylene (ethanol) CVD than for ethanol (ethylene) CVD. This strong and completely opposite
selectivity implies significantly different SWNT growth mechanisms for ethanol and ethylene CVD on Si
substrates.
Introduction
regarding the compatibility among various catalysts and carbon
sources, specifically ethylene and ethanol. We prepared most
of the catalysts that had been reported in the past to provide a
reasonable SWNT yield and carried out a comparative series
of both ethylene and ethanol CVD growths with these catalysts
on Si substrates. From our survey, while the catalyst-preparation
method proved to be a less significant issue, we discovered a
significant importance in the matching of the elemental species
of the catalyst and carbon source for high-yield SWNT synthesis.
Specifically, Fe (Co) catalysts showed high SWNT yields for
ethylene (ethanol) CVD while showing very low yield for
ethanol (ethylene) CVD. This strong and completely opposite
selectivity implies that the growth mechanism of SWNTs
substantially differs for ethanol and ethylene CVD when grown
on Si substrates.
Since their discovery by S. Iijima in 1991,1 carbon nanotubes,
especially SWNTs, have garnered an extremely high level of
attention from both the academic and industrial societies due
to their fascinating and promising physical and chemical
characteristics. Even though it was anticipated that SWNTs
would serve as an ideal material for applications spanning
electronic devices,2-8 high strength composites,9 to conductive
composites,10 real products have not yet been realized because
of the difficulties surrounding the synthesis of SWNTs. There-
fore, it is of critical importance to accumulate knowledge
regarding the growth of SWNTs through fundamental research.
The importance of SWNT synthesis on Si substrates by CVD
stems from the use of standard lithography for the fabrication
of SWNT electronic devices. However, SWNT synthesis is
commonly difficult on Si substrates which results in low yields,
therefore, improvement in CVD growth is eagerly demanded.
In CVD, the catalyst and the carbon source are the two most
important factors. On Si substrates, hydrocarbons (methane/
ethylene)2,7,8,11-16 and ethanol6,17-20 represent two classes of
carbon sources where successful SWNT synthesis has been
reported by many groups. Additionally, a number of different
catalysts have been previously reported using a variety of
preparation methods and compositions.2,4-8,11-15,17-20 Catalysts
have been prepared in the form of thin metal layers,6,8,11 thin
metal salt layers,2,5,7,15,17,18 and nanoparticles,4,12-14,19,20 and their
composition varied from Fe, Co, to their alloys with Mo.
Furthermore, a number of catalyst/carbon source combinations
exist. However, a general understanding concerning the compat-
ibility of the catalysts and the carbon sources does not exist
because only one, or at the most two, of the combinations have
been investigated in each individual research.
Experimental Section
Catalyst Preparation. Fe, Co, and their alloys with Mo are
often used as catalysts for SWNT synthesis. We prepared most
of the Fe and Co-Mo catalysts (Table 1) that have been
previously reported to provide a high yield of SWNT growth
in various forms, such as metal thin layers, metal salt layers,
and nanoparticles, synthesized by wet chemistry, metal deposi-
tion by sputtering and ultrahigh-vacuum e-beam deposition, and
the reverse-micelle method, respectively. Each catalyst was
synthesized following a recipe adopted from the literature that
has been adjusted locally to optimize the SWNT yield.2,6,8,11,13,15,17
Fe Salt. A solution of Fe(NO3)3‚9H2O in isopropyl alcohol
with a concentration of 0.25 mM was prepared. Si substrates
were dipped into the solution for 10 s and rinsed 10 times in
hexane and dried with argon gas.2
Nanoparticles by FeCl3. Fe-containing nanoparticles were
formed by immersing Si substrates into a solution of 10 mL of
0.01 mM FeCl3‚6H2O(aq), followed by the immediate addition
of 100 µL of 40 mM NH2OH‚HCl(aq). After soaking in the
solution for 2 min, the substrate was removed, rinsed with
isopropyl alcohol, and dried.15
Our research is motivated to establish a general understanding
of CVD synthesis of SWNTs on Si substrates particularly
* To whom correspondence may be addressed. Fax: +81-29-861-4654.
E-mail: kenji-hata@aist.go.jp.
10.1021/jp0454117 CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/22/2005

Matching of Catalyst Element and Carbon Source
J. Phys. Chem. B, Vol. 109, No. 7, 2005 2633
TABLE 1: Summary of the Compatibility of the Catalyst
(Element and Preparation Method) and Carbon Sources
(Ethanol or Ethylene) with the Yield of SWNTsa
into the prepared solution for 10 min. After drying, the substrate
was calcined at 400 °C for 5 min in air.17
Fe and Co-Mo Nanoparticles Synthesized by the Reverse-
Micelle Method. Iron chloride, FeCl3, was dissolved into
toluene with a cationic surfactant, didecyldimethylammonium
bromide, to prepare the Fe nanoparticles. Similarly, a solution
of cobalt chloride hexahydrate (CoCl2‚6H2O) and molybdenum
chloride MoCl5 was prepared for the Co-Mo nanoparticles.
Reduction of the metals was carried out by adding aqueous
solution of sodium tetrahydroborate NaBH4 into each of the Fe
and Co-Mo solutions to form colloid of nanoparticles. Finally
the prepared nanoparticles were dropped onto the Si substrate
and left to dry.13
yields
preparation
methods
hydrocarbon
CVD
alcohol
CVD
catalyst
Fe Catalysts
wet chemistry
nanoparticles
thin film
1
Fe(NO3)3
0.27
2.28
1.08
15
FeCl3
Fe thin film8
Fe nanoparticles13
nanoparticles
Co Catalysts
wet chemistry
thin film
Co-Mo acetates17
Co/Mo thin film8
0.92
0.14
0.31
Co-Mo nanoparticles13
nanoparticles
Fe or Co Catalysts with an Al2O3 Underlayer
Chemical Vapor Deposition. Ethylene and ethanol CVD
syntheses were implemented at specific experimental conditions
that were determined from our previous CVD experiments
covering a wide range of growth conditions and were very
similar to those used previously by other groups to synthesize
SWNTs.2,14,16,17,19 CVD at these growth conditions produced a
fairly good yield of SWNTs for all of the catalysts we had
prepared. For ethylene CVD, samples were annealed to 750 °C
in Ar and H2 (total 1000 sccm), and growth was carried out at
750 °C for 2 min with 10 sccm of ethylene.2,14,16 For ethanol
CVD, samples were annealed to 800 °C in Ar gas with 3% H2
at 300 Torr, and upon reaching 800 °C, the gas supply was
stopped and the furnace was evacuated to 1-2 Torr. Conse-
quently CVD growth was initiated by supplying ethanol vapor
from a thermal bath at 10 Torr for 15 min.17,19
Fe/Al2O3 thin film
Co/Al2O3 thin film
thin film
thin film
1086.00
0.82
a The yield of SWNTs was analyzed from Raman spectra where it
was assumed to be propotional to the ratio of the peak heights of the
tangential-mode G band due to SWNTs against the Si Raman peak at
520 cm-1 due to the Si substrate. The value for Fe/Al2O3 thin film is
very high due to the fact that the laser power was substantically
attenuated by the thick SWNT mat.
Fe and Co/Mo Thin Films. An iron thin film of 1-2 nm on
a Si wafer was prepared by magnetron sputtering. Similarly,
Co/Mo thin films were fabricated by successive sputtering of
molybdenum (1-2 nm) and cobalt (1-2 nm).6,8,11
Co-Mo Salt. Molybdenum acetate ((CH3COO)2Mo) and
cobalt acetate ((CH3COO)2Co‚4H2O) were dissolved into etha-
nol by 4 hours of sonication to make a metal acetate solution
with a concentration of 0.01 wt %. Si substrates were immersed
Figure 1. AFM images of the SWNTs on Si substrates grown by ethylene CVD with various catalysts: (a) Fe(NO3)3 salt; (b) nanoparticles by
FeCl3; (c) Fe thin film; (d) Co-Mo salt; (e) Co/Mo thin film; (f) Co-Mo nanoparticles synthesized by the reverse-micelle method; (g) Fe nanoparticles
synthesized by the reverse-micelle method.

2634 J. Phys. Chem. B, Vol. 109, No. 7, 2005
Mizuno et al.
Figure 3. A TEM image of SWNTs grown by ethylene CVD.
We confirmed that the nanotubes observed in Figure 1 were
SWNTs by Raman analysis and transmission electron micro-
scope (TEM) observations. Figure 2 shows a series of Raman
(excited by Ar ion laser at 514 nm) spectra of the samples where
nanotubes were observed. The radial breathing mode (RBM)
peaks that are characteristic to SWNTs were observed in the
low-frequency range of 150-260 cm-1, confirming the existence
of SWNTs. The diameter of SWNTs was calculated to be in
the range of 1.0 to 1.7 nm, from the expression ω ) 248/d21,
where ω is the RBM frequency (cm-1), and d is the diameter
of SWNTs (nm). The tangential-mode G band showed narrow,
strong peaks in the region of 1590 cm-1. In contrast, the so-
called “D band” around 1350 cm-1 related to defects, and the
presence of amorphous carbon was small. These results make
evident that the nanotubes are well crystallized high-quality
SWNTs without an amorphous component. A typical TEM
image of the nanotubes from the FeCl3 catalyst (Figure 1c) is
presented in Figure 3. The image clearly shows that essentially
all of the nanotubes are long, clean, and high-quality SWNTs.
TEM observations of nanotubes grown from other catalysts
showed similar results, and no evidence for multiwalled carbon
nanotube formation was found, demonstrating that our catalysts
and CVD growth were well-suited to selectively grow SWNTs.
Surprisingly, vastly different trends were observed for ethanol
CVD on Si substrates. Figure 4 shows a set of AFM images of
the surface of as-grown ethanol CVD samples with the different
catalysts. While a strong selectivity was found again, this time,
the behavior was completely reversed in terms of the activity
of the catalysts when compared to ethylene CVD. For Co-Mo
salt (Figure 1b), Co-Mo thin film (Figure 1d), and Co-Mo
nanoparticles (Figure 1f) catalysts, a high density of nanotubes
entangled into spaghetti forms was observed, but for Fe salt
(Figure 1a), nanoparticles by FeCl3 (Figure 1c), Fe thin film
(Figure 1e), and Fe nanoparticles synthesized by the reverse-
micelle method (Figure 1g), the yield of SWNTs was very low.
In parts b and d of Figure 1, nanotube yields were very high
and the Si substrates cannot be imaged, indicating that the
catalyst preparation and CVD growth procedures were well
optimized. We confirmed that the nanotubes in Figure 4 are
high-quality SWNTs by Raman analysis and TEM observations.
No evidence for multiwalled carbon nanotubes formation was
found by TEM observations, and Raman spectra showed RBM
peaks at 150-190 cm-1, confirming that the nanotubes are
SWNTs (Figure 5).
Figure 2. Raman spectra showing the presence of the SWNTs on the
Si substrates synthesized by the hydrocarbon CVD with iron catalysts
formed from various catalysts: (i) Fe thin film; (ii) nanoparticles by
FeCl3. The laser excitation wavelength was 514.5 nm. The frequency
ranges were 1800-100 cm-1 (top) and 350-100 cm-1 (bottom). For
the samples where Fe salt and Fe nanopatricles synthesized by the
reverse-micelle method were used as the catalyst, the yields of the
SWNTs were not sufficient to implement reliable Raman experiments.
However, from AFM images, the tube sizes were estimated to be in
the range of 1-3 nm; thus we conclude that the synthesized nanotubes
are SWNTs.
Results and Discussion
A series of ethylene and ethanol CVD syntheses on Si
substrates was carried out at appropriate growth conditions for
each of the seven types of catalysts. Figure 1 shows atomic
force microscopy (AFM) images of the surface of the as-grown
ethylene CVD samples with different catalysts. For Fe salt
(Figure 1a), nanoparticles by FeCl3 (Figure 1c), Fe thin film
(Figure 1e), and Fe nanoparticles synthesized by the reverse-
micelle method (Figure 1g), a high density of nanotubes
entangled into a spaghetti-like form was observed. It is worthy
to note that in parts c and e of Figure 1, the Si substrates could
not be imaged and the nanotubes formed a thick mat (thickness
estimated up to 100 nm) covering the entire substrate. The yield
of SWNTs reported here is comparable to the highest reported
yield of SWNTs synthesized on Si substrates with a hydrocarbon
source, a fact that highlights our well-optimized catalyst
preparation and CVD growth procedures. In sharp contrast
however, for Co-Mo salt (Figure 1b), Co-Mo thin film (Figure
1d), and Co-Mo nanoparticles (Figure 1f) catalysts, the yield
of SWNTs was very low, despite the surface being covered with
numerous catalysts in the form of nanoparticles whose sizes
fall in the range suitable for SWNT growth. As clearly shown
later, failure to synthesize nanotubes from these catalysts was
due to an extremely strong selectivity of SWNT growth on the
combination of the catalyst and carbon source rather than the
ability of the catalysts to produce SWNTs.
The results are summarized in Table 1 where the catalyst
element, preparation method, and carbon source are listed with
their associated SWNT yields. The yield of the SWNTs was

Matching of Catalyst Element and Carbon Source
J. Phys. Chem. B, Vol. 109, No. 7, 2005 2635
Figure 4. AFM images of the SWNTs on the Si substrates grown by ethanol CVD with various catalysts: (a) Fe(NO3)3 salt; (b) nanoparticles by
FeCl3; (c) Fe thin film; (d) Co-Mo salt; (e) Co/Mo thin film; (f) Co-Mo nanoparticles synthesized by the reverse-micelle method; (g) Fe nanoparticles
synthesized by the reverse-micelle method.
approximated to be propotional to the ratio of the tangential-
mode G band peak heights from SWNTs and the Si Raman
peak at 520 cm-1 from the Si substrate. This method cannot be
regarded as an accurate quantitative analysis because a thick
SWNT mat would attenuate the incident laser energy on the Si
substrate and other graphitic carbon materials can also contribute
to the G-band; however, we assume that it provides a reasonable
first-order approximation of the yield of SWNTs on Si
substrates. From Table 1, it is clear that there exists a strong
selectivity between the combination of the atomic species of
the catalyst and carbon source. Matching of the catalyst element
and carbon source was found to be of extreme importance, while
the catalyst-preparation method proved to be a less significant
issue. Fe catalysts produced a high yield of SWNTs growth for
ethylene CVD, while the yield was very low for ethanol CVD.
Interestingly, for Co catalysts, the selectivity was completely
reversed where the SWNT yield was high for ethanol CVD but
very low for ethylene CVD. This strong and completly reversed
selectivity involving the catalytic element and carbon source
observed on SWNT synthesis implies that the SWNT growth
mechanism is considerably different for ethylene and ethanol
CVD.
catalysts (e.g., Fe-Mo nanoparticles, Fe-Mo salt, Co thin film,
and Co nanoparticles). Generally speaking, for both Fe and Co,
alloying with Mo increased the SWNT yield; however, this did
not change the selectivity.
The specific CVD growth conditions for Figures 1-5 were
carefully chosen based on preliminary CVD experiments results
covering a wide range of growth conditions. For ethylene CVD,
the growth temperature was varied from 700 to 800 °C, growth
time was varied from a few seconds to 30 min, ethylene flow
rate was varied from 0.5 to 100 sccm, hydrogen concentration
was varied from 20% to 100%, and He was used as the carrier
gas. For ethanol CVD, the growth temperature was varied from
700 to 900 °C, growth time was varied from 2 to 60 min, the
growth pressure was varied from 5 to 30 Torr, and the H2
concentration was varied 3 to 6%. Among the growths
conducted on this wide range of experimental conditions, the
specific growth conditions of Figures 1-5 showed fairly good,
if not the best, yield of SWNTs for all of the catalysts we have
prepared. If carefully tuned, the best growth conditions differ
with each catalyst. Nonetheless, as long as the same carbon
source is used, the optimized growth conditions for each
different catalyst seems to fall into a similar range of experi-
mental conditions.
We reached this conclusion regarding selectivity through a
wide range of CVD experiments of which Figures 1-5 are only
a part. The results of the whole body of experiments are divided
into two aspects: (1) elemental composition of the catalyst and
(2) CVD growth conditions. The catalysts used in experiments
of Figures 1-5 were made of Fe and Co-Mo, compositions of
catalysts that are most commonly used in SWNTs syntheses.
We have also prepared and tested several Fe-Mo and Co
By a change of the catalyst-substrate interaction, it might
be possible to relax the strong selectivity of the catalyst and
carbon source on Si substrates and grow SWNTs. Some recent
research11,22,23 has pointed out that SWNT growth can be
significantly improved by placing a thin intermediate layer
(Al2O3, Ti) between the catalyst and Si substrates. To investigate
how this factor influences the selectivity, Fe and Co metal thin

2636 J. Phys. Chem. B, Vol. 109, No. 7, 2005
Mizuno et al.
Figure 5. Raman spectra showing the presence of the SWNTs on the
Si substrates synthesized by ethanol CVD with Co-Mo catalysts of
various forms: (i) Co-Mo salt; (ii) Co/Mo thin film; (iii) Co-Mo
nanoparticles. The laser excitation wavelength was 514.5 nm. The
frequency ranges were 1800-100 cm-1 (top) and 350-100 cm-1
(bottom).
Figure 6. AFM images showing SWNTs grown by catalysts prepared
on Si substrate with an Al2O3 underlayer: (a) SiO2 (500 nm)/Al2O3
(10 nm)/Fe (1 nm); (b) SiO2 (500 nm)/Al2O3 (10 nm)/Co (1 nm). In
both cases, the yields of SWNTs were high. However, the yield was
much higher when Fe was used as the catalyst than when Co was used.
This point could be easily confirmed by the color of the sample; the
sample with Fe catalysts became totally black, while the color of the
sample with Co catalyst remained similar to that of the Si wafer.
film catalysts with thin Al2O3 (10 nm) layers underneath were
fabricated on Si, and ethylene CVD was performed. The double-
layered catalyst was fabricated by standard magnetron sputtering
where Al2O3 was first deposited from a Al2O3 sputter target
followed by Fe deposition. Interestingly, high yields of SWNTs
could be grown with ethylene even with Co catalysts (Figure
6b and Table 1) with the Al2O3 intermediate layer underneath
the catalyst. It is worthy to note that growth with Fe catalysts
(Figure 6a and Table 1) with an Al2O3 sublayer showed a even
higher yield growth. It is reported that the Al2O3 thin layer
promotes the formation of sufficiently small catalyst particles
that are able to catalyze SWNT growth and also inhibit further
coalescence of the catalysts.22,24 In addition, the Al2O3 thin layer
could keep the catalysts from interacting from the silica surface
or change the atomic strain of the intermediate layer.
From experiments alone, it is difficult to understand why such
a strong selectivity exists between the catalyst and carbon source.
To understand the mechanism of the selectivity, it is crucial to
simultaneously observe the atomic structure of the catalysts and
SWNTs and obtain information about their interface. Currently
such a work is in progress.
In conclusion, we have investigated the matching of various
catalysts and two representive carbon sources (ethylene and
ethanol) for SWNT CVD on Si substrates. We found an
extremely strong selectivity; Fe (Co) catalysts showed high
SWNT yields for ethylene (ethanol) CVD while producing very
low SWNT yields for ethanol (ethylene) CVD. The method by
which the catalyst was synthesized proved to be less of a factor.
The observed strong and completely reversed selectivity suggests
that ethanol and ethylene CVD have a significantly different
growth mechanism of SWNTs on Si substrates. Our results
clearly demonstrate that it is of critical importance to choose
the appropriate combination of substrate, catalyst, and carbon
source for efficient SWNTs synthesis.
Acknowledgment. This work was supported by the New
Energy and Industrial Technology Development Organization
(NEDO) as a part of the Nano Carbon Technology (NCT)
project.
References and Notes
(1) Iijima, S. Nature 1991, 354, 56.
(2) Hafner, J. H.; Cheung, C.-L.; Oosterkamp, T. H.; Lieber, C. M. J.
Phys. Chem. B 2001, 105, 743.
(3) Postma, H. W. Ch.; Teepen, T.; Yao, Z.; Grifoni, M.; Dekker, C.
Science 2001, 293, 76.
(4) Kim, W.; Choi, H. C.; Shim, M.; Li, Y.; Wang, D.; Dai, H. Nano
Lett. 2002, 2, 703.
(5) Javey, A.; Guo, J.; Wang, Q.; Lundstrom, M.; Dai, H. Nature 2003,
424, 654.
(6) Ohno, Y.; Iwatsuki, S.; Hiraoka, T.; Okazaki, T.; Kishimoto, S.;
Maezawa, K.; Shinohara, H.; Mizutani, T. Jpn. J. Appl. Phys. 2003, 42,
4116.
(7) Durkop, T.; Getty, S. A.; Cobas, E.; Fuhrer, M. S. Nano Lett. 2004,
4, 35.
(8) Li, Y.; Mann, D.; Rolandi, M.; Kim, W.; Ural, A.; Hung, S.; Javey,
A.; Cao, J.; Wang, D.; Yenilmez, E.; Wang, Q.; Gibbons, F.; Nishi, Y.;
Dai, H. Nano Lett. 2004, 4, 317.
(9) Kumar, S.; Dang, T. D.; Arnold, F. E.; Bhattacharyya, A. R.; Min,
B. G.; Zhang, X. F.; Vaia, R. A.; Park, C.; Adams, W. W.; Hauge, R. H.;
Smalley, R. E.; Ramesh, S.; Willis, P. A. Macromolecules 2002, 35, 9039.
(10) Ramasubramaniam, R.; Chen, J.; Liu, H. Appl. Phys. Lett. 2003,
83, 2928.
(11) Delzeit, L.; Chen, B.; Cassell, A.; Stevens, R.; Nguyen, C.;
Meyyappan, M. Chem. Phys. Lett. 2001, 348, 368.

Matching of Catalyst Element and Carbon Source
J. Phys. Chem. B, Vol. 109, No. 7, 2005 2637
(12) Li, Y.; Kim, W.; Zhang, Y.; Rolandi, M.; Wang, D.; Dai, H. J.
Phys. Chem. B 2001, 105, 11424.
(13) Ago, H.; Ohshima, S.; Uchida, K.; Yumura, M. J. Phys. Chem. B
2001, 105, 10453.
(19) Kohno, M.; Orii, T.; Hirasawa, M.; Seto, T.; Murakami, Y.; Chiashi,
S.; Miyauchi, Y.; Maruyama, S. Appl. Phys. A 2004, 79, 787.
(20) Ishida, M.; Hongo, H.; Nihey, F.; Ochiai, Y. Jpn. J. Appl. Phys.
2004, 43, L1356.
(14) Cheung, C. L.; Kurtz, A.; Park, H.; Lieber, C. M. J. Phys. Chem.
B 2002, 106, 2429.
(15) Choi, H. C.; Kundaria, S.; Wang, D.; Javey, A.; Wang, Q.; Rolandi,
M.; Dai, H. Nano Lett. 2003, 3, 157.
(16) Liao, H.; Hafner, J. H. J. Phys. Chem. B 2004, 108, 6941.
(17) Murakami, Y.; Miyauchi, Y.; Chiashi, S.; Maruyama, S. Chem.
Phys. Lett. 2003, 377, 49.
(21) Jorio, A.; Saito, R.; Hafner, J. H.; Lieber, C. M.; Hunter, M.;
McClure, T.; Dresselhaus, G.; Dresselhaus, M. S. Phys. ReV. Lett. 2001,
86, 1118.
(22) Seidel, R.; Liebau, M.; Deusberg, G. S.; Kreupl, F.; Unger, E.;
Graham, A. P.; Hoenlein, W.; Pompe, W. Nano Lett. 2003, 3, 965.
(23) Zhang, R. Y.; Amlani, I.; Baker, J.; Tresek, J.; Tsui, R. K.; Fejes,
P. Nano Lett. 2003, 3, 731.
(18) Murakami, Y.; Chiashi, S.; Miyauchi, Y.; Maruyama, S. Jpn. J.
Appl. Phys. 2004, 43, 1221.
(24) Hongo, H.; Nihey, F.; Ichihashi, T.; Ochiai, Y.; Yudasaka, M.;
Iijima, S. Chem. Phys. Lett. 2003, 380, 158.