Investigation on the temperature-dependent growth rate of carbon nanotubes using chemical vapor deposition of ferrocene and acetylene

Article abstract of DOI:10.1016/j.cplett.2004.11.113

Carbon nanotubes (CNTs) were synthesized using chemical vapor deposition of ferrocene and acetylene in the temperature range 600-800°C. The growth rate increases exponentially with the temperature, providing an activation energy 30 kcal/mol which is close to the diffusion energy of carbons in the bulk iron metal. We compared the growth rate with those of the CNTs grown using other thermal methods reported previously and suggest a simple kinetic model to explain the growth rate depending on the method. The supply rate of C atoms in gas phase would influence the growth rate of CNTs.

Full text of DOI:10.1016/j.cplett.2004.11.113

Chemical Physics Letters 401 (2005) 459–464
www.elsevier.com/locate/cplett
Investigation on the temperature-dependent growth rate of
carbon nanotubes using chemical vapor deposition of
ferrocene and acetylene
a
a,
b
b
Ki-Eun Kim , Kang-Jin Kim *, Woo Sung Jung , Seung Yong Bae ,
b,
c
c
Jeunghee Park *, Junghyun Choi , Jaebum Choo
a
Department of Chemistry, Korea University, 1 Anam Dong Seoul 136-701, Republic of Korea
b
Department of Chemistry, Korea University, Jochiwon 339-700, Republic of Korea
c
Department of Chemistry, Hanyang University, Ansan 425-791, Republic of Korea
Received 29 October 2004; in final form 19 November 2004
Available online 15 December 2004
Abstract
Carbon nanotubes (CNTs) were synthesized using chemical vapor deposition of ferrocene and acetylene in the temperature range
600–800 ꢀC. The growth rate increases exponentially with the temperature, providing an activation energy 30 kcal/mol which is close
to the diffusion energy of carbons in the bulk iron metal. We compared the growth rate with those of the CNTs grown using other
thermal methods reported previously and suggest a simple kinetic model to explain the growth rate depending on the method. The
supply rate of C atoms in gas phase would influence the growth rate of CNTs.
ꢁ 2004 Elsevier B.V. All rights reserved.
1. Introduction
multiwalled CNTs (MWNTs) are typically grown with
cylindrical or bamboo-like structures. The efficient syn-
thesis of MWNTs poses an important question whether
it is possible to manipulate the structure, length, and de-
gree of crystalline perfection using the growth condition.
However, this issue would be hardly solved unless com-
plete understanding of the growth mechanism.
In the present work, the vertically aligned CNTs were
successfully synthesized on the silicon oxide substrates
using a chemical vapor deposition (CVD) of ferrocene/
acetylene (C2H2) over the temperature range 600–800
ꢀC. Although the synthesis of CNTs using ferrocence/
acetylene mixture has been extensively performed
by a number research groups [23–27], the temperature-
dependent growth behaviors have not been much con-
sidered yet. We observe that the growth rate of CNTs
exhibits an exponential increase with the temperature,
which is consistent with our previous works using
CVD and pyrolysis methods [27–31]. However, the
growth rate usually varies with the growth method;
Since the first observation of carbon nanotubes
(CNTs) by Iijima [1], the CNTs have attracted intensive
research due to their unique physical and chemical prop-
erties. The promising application of CNTs have already
been demonstrated as electron field emitter of displays
[2–4], nanoscale electronic devices [5–9], biosensors
[10–12], hydrogen storage cell [13,14], and reinforcing
agents in polymer composites [15,16]. Various methods
were developed to synthesize the high-yield CNTs,
which are prerequisite for many potential applications
[17–21]. It has established that the use of catalysts (metal
nanoparticles) is crucial for the synthesis and the size of
catalysts usually determines the diameter of CNTs [22].
As the size of catalytic nanoparticles becomes larger, the
*
Corresponding authors. Fax: +2 3290 3973.
E-mail addresses: kjkim@korea.ac.kr (K.-J. Kim), parkjh@
korea.ac.kr (J. Park).
0009-2614/$ - see front matter ꢁ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2004.11.113

460
K.-E. Kim et al. / Chemical Physics Letters 401 (2005) 459–464
one of examples is vapor phase growth using ferrocene/
C2H2 without the use of the substrates in the wide tem-
perature range 700–1000 ꢀC; the growth rate is about
five times larger than that of the present growth at the
same temperature [27]. Here we speculate the main fac-
tors to determine the growth rate of CNTs, using com-
prehensive comparison of the present result with other
growth methods [27–31]. We propose a kinetic model
for the synthesis of multiwalled CNTs that can be ap-
plied irrespective of the growth method. These provide
a remarkable insight for the growth mechanism which
would be essential to synthesize the CNTs in a con-
trolled manner. To our best knowledge, systematic com-
parison between the CNTs synthesized via various
methods has been first made.
(Fig. 1a). The length is approximately 4 lm. Fig. 1b
illustrates the CNT arrays grown at 700 ꢀC, with an
average length of 30 lm. The CNT arrays grown at
800 ꢀC exhibit a length of 100 lm (Fig. 1c). From the
result of several experimental runs, the average length
of CNTs grown for 10 min has been measured to be
4
1, 12 2, 30 10, 50 20, and 100 40 lm, respec-
tively, at 600, 650, 700, 750, and 800 ꢀC. The respective
average growth rate of CNTs is 0.4 0.1, 1.2 0.2,
3
1, 5 2, and 10 4 lm/min. The growth rate is con-
stant during 10 min. As the growth temperature in-
creases from 600 to 800 ꢀC, the average growth rate
increases nonlinearly by a factor of 25.
Fig. 2a–c demonstrate the TEM images for the CNTs
grown at 600, 700, and 800 ꢀC, respectively. The CNTs
were dispersed in ethanol to load on the TEM grid. The
diameter of CNTs is uniformly 10–20 nm at all temper-
atures, with an average value of 15 nm. The CNTs exhi-
bit exclusively a cylindrical structure. The closed end of
the CNTs frequently encapsulates the catalytic particles,
which are presumably the tip. The other end is usually
open or corked with the catalytic particles. The catalytic
particles frequently encapsulated even in the middle of
the tube. High-resolution TEM (HRTEM) image of
CNTs grown at 600 ꢀC is displayed in Fig. 2d. The
diameter is about 10 nm and the crystalline graphitic
sheets are aligned parallel to the tube axis. The CNT
grown at the higher temperature usually have a higher
degree of crystalline perfection.
We employed first-order Raman spectroscopy in or-
der to obtain the information about the crystallinity of
entire CNTs (Fig. 3). Two Raman bands at ꢀ1580
cmꢁ1 (G band) and ꢀ1350 cmꢁ1 (D band) are originated
from the Raman active in-plane atomic displacement
E2g mode and disorder-induced features due to the finite
particle size effect or lattice distortion, respectively [32].
The intensity ratio of D band to G band (ID/IG) has a
linear relation with the inverse of the in-plane crystallite
dimension. As the temperature increases from 600 to
800 ꢀC, the value of ID/IG decreases from 0.87 to 0.44.
It indicates that the degree of long-range ordered crys-
talline perfection of the CNTs increases with the temper-
ature. The average length, growth rate, and ID/IG value
at all temperatures are listed in Table 1.
2. Experimental section
Ferrocene (98+%, Aldrich) was used after vacuum dry
at 100 ꢀC for 1 h. A CVD reactor (inner diameter of the
quartz tube 2.5 cm and effective heating length 55 cm)
was heated in a dual furnace fitted with independent tem-
perature controllers. About 0.1 g of ferrocene was placed
in the first heating stage. Silicon substrate (1 cm · 3 cm)
was thermally oxidized to form 300 nm thick oxide layers
and placed in the second heating stage. Ferrocene was
vaporized at 200 ꢀC and carried by the Ar flow (200 sccm)
into the second heating stage of the quartz tube where the
pyrolysis was performed at 600–800 ꢀC. C2H2 (Kyungin
Gas, 99.99%) flowed at a rate of 10 sccm. Then the reac-
tor was cooled down to room temperature under Ar
ambient. The CNTs were grown on the substrates in
the second stage. They were collected and analyzed by
scanning electron microscopy (SEM, Hitachi S-4300),
transmission electron microscopy (TEM, JEOL JEM-
4010), and Raman spectroscopy (Renishaw micro-
Raman 2000) using the 514.5 nm line ofan argon ion laser.
3. Results
SEM micrograph shows vertically aligned CNTs
grown on a large area of the substrates at 600 ꢀC
Fig. 1. SEM micrograph of the CNTs grown via CVD at: (a) 600, (b) 700, and (c) 800 ꢀC. The length is 4, 30 and 100 lm, respectively.

K.-E. Kim et al. / Chemical Physics Letters 401 (2005) 459–464
461
Fig. 2. TEM images of CNTs grown at: (a) 600, (b) 700, and (c) 800 ꢀC. The average diameter is 15 nm. (d) HRTEM image for a cylindrical CNT
grown at 600 ꢀC.
Table 1
The length, diameter, and degree of crystallinity of the CNTs grown
via CVD of ferrocene and C2H2
1580 (G)
1350 (D)
a
Temperature Average
(ꢀC)
Growth
Average
ID/IG
ID/IG
0.44
length (lm) rate (lm/min) diameter (nm)
800 oC
700 oC
600
650
700
750
800
a
4
12
1
2
0.4 0.1
1.2 0.2
15
15
15
15
15
0.87
0.68
0.53
0.50
0.44
30 10
50 20
100 40
3
5
1
2
4
10
0.53
0.87
Intensity ratio of D to G Raman bands.
600 oC
The temperature-dependent growth rate using the Fe
catalysts has been also observed from our previous
works [27–31]. We designate the methods as I–V. The
growth condition and the results are summarized in
Table 2. Method I is so called Ôvapor phase growthÕ or
Ôfloating CVDÕ because of no substrates used; ferrocene
was pyrolyzed as a catalyst (Fe) source and C2H2 was
used as carbon source at 700–1000 ꢀC [27]. The aligned
CNTs were grown vertically on the substrates using
methods II–V. Method II is a CVD method; the Fe
catalytic nanoparticles were deposited on the silicon
800 1000 1200 1400 1600 1800 2000
Raman Shift(cm-1)
Fig. 3. Raman spectrum of CNTs grown at 600, 700, and 800 ꢀC.
4. Discussion
As the temperature increases from 600 to 800 ꢀC, the
growth rate of CNTs increases from 0.4 to 10 lm/min.

462
K.-E. Kim et al. / Chemical Physics Letters 401 (2005) 459–464
Table 2
Growth conditions, morphology, and the content of N atoms of CNTs synthesized by various CVD, vapor phase growth, and pyrolysis methods
Present work
I
II
III
IV
V
Method
C source
CVD
C2H2
Ferrocene
Silicon oxide
·
600–800
4–100
10
Vapor phase
C2H2
CVD
C2H2
FeCl2
Silicon
ꢃ
CVD
CH4
FeCl2
Silicon
ꢃ
CVD
C2H2
FeCl2
Alumina
ꢃ
Pyrolysis
FePc
Catalyst source
Substrates
NH3pretreatmenta
Temperature (ꢀC)
Length (lm)
Ferrocene
·
·
700–1000
50–3000
Silicon oxide
·
700–1000
1.5–70
20
800–1100
8–140
5
900–1100
5–150
5
900–1000
30–65
5
Time (min)
5
10–600
Growth rate (lm/min)
Avg. diameter (nm)b
At.% of Nc
0.4–10
15
1.6–28
20–150
1–2
1–30
50–150
1–2
6–13
50–80
1–2
0.075–3.5
10–80
2–4
20
0
0
Reference
27
28
29
30
31
The symbols ÔxÕ and ÔꢃÕ stand for ÕnoÕ and ÕyesÕ, respectively.
a
NH3 was pretreated before the growth of CNTs and used as N source during the growth.
Average diameter.
Atomic % of N measured by XPS.
b
c
6
substrates by coating the substrates with iron chloride
(FeCl2) ethanol solution following the ammonia (NH3)
pretreatment at the growth temperature 800–1100 ꢀC.
C2H2 was used as carbon source [28]. Method III is an-
other CVD method; the Fe catalytic particles were
deposited on the silicon substrates by the same procedure
as that of method II, but methane (CH4) instead of
C2H2 was used as carbon source over the temperature
range 900–1100 ꢀC [29]. Method IV is also a CVD
method; Fe catalytic particles were deposited on the alu-
mina substrates by the same procedure as that of method
II and C2H2 was used in the range 900–1000 ꢀC [30].
Method V is pyrolysis method; Fe phthalocyanine (FePc)
molecules were pyrolyzed in the temperature range 700–
1000 ꢀC, using two-temperature zone system; FePc subli-
mated in the first temperature zone and delivered into the
second temperature zone by the H2/Ar flow [31].
Present work
Method I
Method II &IV
Method III
Method V
4
2
0
-2
-4
Ea=30 5 kcal/mol
8
10
12
1/T (X10-4)
Fig. 4. Arrhenius plot for growth rate of CNTs, providing an
activation energy (Ea) of 30 5 kcal/mol, for the present method,
vapor phase growth (method I), CVD (method II, III, and IV), and
pyrolysis of FePc (method V).
Method III is only exceptional case that the growth rate
does not fit to the Arrhenius plot. The growth rate at
900 and 1000 ꢀC is about the same as that of method
V, but that at 1100 ꢀC is close to that of method II
and IV. Nevertheless, the activation energy of CNT
growth suggests that the bulk diffusion rate of C would
be a main factor in determining the growth rate of
CNTs. The growth process of CNTs can be pictured
as follows; the C atoms dissolve into the metal nanopar-
ticles deposited on the substrates (or formed in gas
phase) and diffuse through them. As the C atoms satu-
rate into the nanoparticles, they precipitate and form
the graphitic sheets. The diffusion of C atoms in the me-
tal nanoparticles would be a rate-determining step for
whole growth process of CNTs.
Firstly, let us assume that the growth rate (m) of
CNTs is simply first-order reaction expressed by follow-
ing equation:
d½CNTꢂ
m ¼
¼ kd½Cꢂ ;
sat
dt
where kd is rate constant proportional to the diffusion
coefficient of carbons (C), and [C]sat is the saturated con-
centration of C in bulk metal. The C atoms are saturated
into bulk c-Fe by an extent of 3–7% in the temperature
range 600–1000 ꢀC [33]. Using the saturated C concentra-
ꢀ
ꢁ
m
tion [C]sat, the Arrhenius plot of in
vs: 1=T ðKÞ
½Cꢂsat
can be used to derive the activation energy (Ea) of the
present CNT, as displayed in Fig. 4. The linear fit gives
a value of Ea, 30 kcal/mol, which is very close to that
the diffusion energy of C in bulk c-Fe, 35 kcal/mol
[22,34]. The data of other four methods (I, II, IV, and
V) also fit well to a linear function, yielding Ea in the
The assumption that the growth rate is simply first-or-
der reaction m = kd[C]sat cannot explain properly the
growth rate of CNTs depending on the growth method.
The method I shows about five times higher growth rate
than the present method at the same growth tempera-
ture, while the method V shows about 25 times lower
range of 25–35 kcal/mol, with an average value 30
5
kcal/mol. The growth rate of methods II and IV is
the same, showing an independence of the substrates.

K.-E. Kim et al. / Chemical Physics Letters 401 (2005) 459–464
463
growth rate. Moreover, the growth rate of method III
apparently does not follow the first-order reaction.
Now let us propose a slightly complicated mechanism
in which the C atoms dissolved in the metal nanoparticles
that are saturated with C atoms and form an intermedi-
ate before the precipitation producing the CNTs. The
reaction can be separated into three steps as follows:
The C supply rate of method I is much higher than that
of the present methods due to the absence of the sub-
strates. The decomposition of phthalocyanine molecules
supplies the C source less efficiently than that of C2H2.
In method III, thermal decomposition of CH4 pro-
gresses significantly with the temperature, so K[C]free
would increase and reach to the same value of method
II or IV at 1100 ꢀC. This may explain why the Arrhenuis
plot of method III is non-linear.
k1
Cfree þ Csat ! Cexcess
k
ꢁ1
Cexcess ! Cfree þ Csat
The present CNTs exhibit the cylindrical structure
and the diameter is uniformly 15 nm. The multiwalled
CNTs are usually either cylindrical or bamboo-like
structured. We found that the N content and diameter
influence the structure. When the growth temperature
is higher than 900 ꢀC in the case of method II–V, the
diameter is usually larger than 50 nm and the bam-
boo-like structure is preferentially grown. The compart-
ment layers are formed by the C atoms precipitated
inside the tubes and must overcome a stress in order
to join with the graphitic sheets of the wall. Thus the
vertical growth of graphitic sheets would be disturbed
by the joints with compartment layers, producing the
bamboo-like structure. The incorporation of N atoms
induces the flexibility of the compartment layers, so
the connection between wall and compartment layers
will take place under a reduced stress. Therefore the in-
crease of N content would help the production of the
bamboo-like structure. However, as the diameter is
smaller than 50 nm and no N-doping occurs, the struc-
ture remains cylindrical due to the tremendous stress in
connecting the compartments with the wall, as in the
case of present work and method I.
The degree of crystalline perfection increases with the
growth temperature. As the diffusion rate of C increases,
the graphitic sheets would build with a less defect. The
CNTs grown via pyrolysis of FePc (method V) at 700
and 800 ꢀC contain the N content 3.9 and 3.2 at.%.
The ID/IG value is 0.87 and 0.70, which is larger than
that of the present CNTs. It is because the pyridine-type
N atoms reduce the crystalline perfection of the gra-
phitic sheets [35].The present CNTs have the higher de-
gree of crystalline perfection due to pure graphite
structures.
kd
Cexcess ! Csat þ CNTs
The Cfree and Csat correspond to the C atoms sup-
plied from the C source in the gas phase and the C
atoms saturated in the nanoparticles, respectively. When
C
free dissolves in the nanoparticles saturated with Csat, it
forms the C excess states (we call as Cexcess). These C ex-
cess states are unstable, so that they decay back into
Cfree and Csat, and also precipitate to produce the CNTs,
accompanying the diffusion through the nanoparticles.
The rate constants are k1 and k for the forward and
ꢁ1
reverse reaction of the equilibrium, respectively, and kd
for the CNT growth step. The kd is proportional to
the diffusion rate of C atoms in the metal nanoparticles.
This scheme involves a pre-equilibrium that the interme-
diate Cexcess is in equilibrium with the reactants Cfree and
C
sat. If the rate of formation of the intermediate Cexcess
and its decay back into reactants were much faster
than the rate of formation of CNTs (k1 ꢄ kd), the rate-
determining step of CNT growth would be the reaction
step (iii). Because we assume that the reactants Cfree
and Csat are in equilibrium with the intermediate Cexcess
the equilibrium constant is
,
½Cꢂ
k1
excess
K ¼
¼
:
½Cꢂ ½Cꢂ
Then the growth rate of CNTs is
k
ꢁ1
free
sat
d½CNTꢂ
¼ kd½Cꢂ
¼ kdK½Cꢂ ½Cꢂ
excess
free
sat
dt
¼ ðK½Cꢂ Þkd½Cꢂ ¼ Akd½Cꢂ :
This rate law has the form of second-order with a
composite rate constant. However, if k1 and k have
free
sat
sat
ꢁ1
no activation energy and [C]free is constant for all tem-
peratures, the first-order assumption we made previ-
ously is still valid. The growth rate can depend on the
[C]free and equilibrium constant K that are determined
by the supply rate of C atoms in gas phase. As the length
increases, the C supply rate becomes slower down due to
5. Conclusion
The CNTs were synthesized using CVD of ferrocene
and C2H2 in the temperature range 600–800 ꢀC. As the
growth temperature increases from 600 to 800 ꢀC, the
growth rate enhances by a factor of 25 and the degree
of crystalline perfection increases. They have a cylindri-
cal structure with uniform diameter 15 nm for all tem-
peratures. We compared the temperature-dependent
growth rate with those of the CNTs grown using other
thermal methods. The activation energy (Ea) of the
the steric hindrance of high-density CNTs (>1010 cmꢁ2
)
and the growth would be finally stopped (k1 ! 0). Un-
der the time regime of constant K[C]free, K[C]free would
follow the order; method I > present method > method
II and IV > method V. When C2H2 is used, the CNTs
would be grown under a complete thermal decomposi-
tion of C2H2 over all temperature range 600–1100 ꢀC.

464
K.-E. Kim et al. / Chemical Physics Letters 401 (2005) 459–464
[12] J.J. Gooding, R. Wibowo, J. Liu, W. Yang, D. Losic, S. Orbons,
F.J. Mearns, J.G. Shapter, D.B. Hibbert, J. Am. Chem. Soc. 125
(2003) 9006.
CNT growth was obtained by a simple assumption that
the growth reaction is first-order reaction whose rate is
defined as m = kd[C]sat, where kd is proportional to the
diffusion coefficient of C atoms and [C]sat is the satu-
rated concentration of C in bulk metal. The Arrhenius
plot gives Ea = 30 kcal/mol, which is close to the diffu-
sion energy of C into the bulk c-Fe. The dependence
of the growth rate on the method has been explained
by a kinetic model that a fast equilibrium exists between
the free C atoms in gas phase and the dissolved C atoms,
and the diffusion of dissolved C atom is the rate-deter-
mining step in overall growth process. The supply rate
of C atoms in gas phase would influence the growth rate
of CNTs. The cylindrical structure rather than bamboo-
like structure is explained by the tremendous stress in
building the compartment layers. Finally we suggest
that the growth rate of CNTs could be controllable by
the growth condition such as C source and temperature.
[13] C. Liu, Y.Y. Fan, M. Liu, H.T. Cong, H.M. Cheng, M.S.
Dresselhaus, Science 286 (1999) 1127.
[14] G.-P. Dai, C. Liu, M. Liu, M.-Z. Wang, H.-M. Cheng, Nano
Lett. 2 (2002) 503.
[15] C. Velasco-Santos, A.L. Martinez-Hernandez, F.T. Fisher, R.
Ruoff, V.M. Castano, Chem. Mater. 15 (2003) 4470.
[16] X. Zhang, T. Liu, T.V. Sreekumar, S. Kumar, V.C. Moore, R.H.
Hauge, R.E. Smalley, Nano Lett. 3 (2003) 1285.
[17] D.S. Bethune, C.H. Kiang, M.S. deVries, G. Gorman, R. Savoy,
J. Vazquez, R. Beyers, Nature 363 (1993) 605.
[18] C. Journet, W.K. Maser, P. Bernier, A. Loiseau, M. Lamy de la
Chapelle, S. Lefrant, P. Deniard, R. Lee, J.E. Fischer, Nature
388 (1997) 756.
[19] A. Thess, R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C.
Xu, Y.H. Lee, S.G. Kim, A.G. Rinzler, D.T. Colbert, G.E.
Scuseria, D. Toma´nek, J.E. Fisher, R.E. Smalley, Science 273
(1996) 483.
[20] M. Terrones, N. Grobert, J. Olivares, J.P. Zhang, H. Terrones,
K. Kordatos, W.K. Hsu, J.P. Hare, P.D. Townsend, K. Prassides,
A.K. Cheetham, H.W. Kroto, D.R.M. Walton, Nature 388
(1997) 52.
[21] Z.F. Ren, Z.P. Huang, J.W. Xu, J.H. Wang, P. Bush, M.P.
Siegal, P.N. Provencio, Science 282 (1998) 1105.
[22] R.T.K. Baker, Carbon 27 (1989) 315.
Acknowledgements
The KOSEF (Project No. R14-2004-033-0100-0; R02-
2004-000-10025-0) and KRF (Project No. 2004-015-
C00265) supported the present work. SEM analysis was
performed at Korea Basic Research Institute in Seoul.
[23] C.N.R. Rao, R. Sen, B.C. Satishkumar, A. Govindaraj, Chem.
Commun. (1998) 1525.
[24] B.C. Satishkumar, A. Govindaraj, C.N.R. Rao, Chem. Phys.
Lett. 307 (1999) 158.
[25] R.L. Vander Wal, T.M. Ticich, V.E. Curtis, J. Phys. Chem. A 104
(2000) 7209.
[26] C.N.R. Rao, A. Govindaraj, Acc. Chem. Res. 35 (2002) 998.
[27] Y.T. Lee, N.S. Kim, J. Park, J.B. Han, Y.S. Choi, H. Ryu, H.J.
Lee, Chem. Phys. Lett. 372 (2003) 853.
References
[1] S. Iijima, Nature 354 (1991) 56.
[28] Y.T. Lee, J. Park, Y.S. Choi, H. Ryu, H.J. Lee, J. Phys. Chem. B
106 (2002) 7614.
ˆ
[2] W.A. De Heer, A. Chatelain, D. Ugarte, Science 270 (1995) 1179.
[3] S. Fan, M.G. Chapine, N.R. Franklin, T.W. Tombler, A.M.
Cassell, H. Dai, Science 283 (1999) 512.
[29] N.S. Kim, Y.T. Lee, J. Park, H. Ryu, H.J. Lee, S.Y. Choi, J.
Choo, J. Phys. Chem. B 106 (2002) 9286.
[4] N. de Jonge, Y. Lamy, K. Schoots, T.H. Oosterkamp, Nature 420
(2002) 393.
[30] Y.T. Lee, N.S. Kim, S.Y. Bae, J. Park, S. Yu, H. Ryu, H.J. Lee,
J. Phys. Chem. B 107 (2003) 12958.
[5] P. Collins, A. Zettl, H. Bando, A. Thess, R.E. Smalley, Science
278 (1997) 100.
[31] N.S. Kim, Y.T. Lee, J. Park, J.B. Han, Y.S. Choi, S.Y. Choi, J.
Choo, G.H. Lee, J. Phys. Chem. B 107 (2003) 9249.
[32] F. Tuinstra, J.L. Koenig, J. Chem. Phys. 53 (1970) 1126.
[33] T.B. Massalski, Binary Alloy Phase Diagrams, American Society
of Materials, Ohio, 1986.
[6] S.J. Tans, A.R.M. Verschueren, C. Dekker, Nature 393 (1998) 49.
[7] P. Kim, C.M. Lieber, Science 286 (1999) 2148.
[8] A. Bachtold, P. Hadley, T. Nakanishi, C. Dekker, Science 294
(2001) 1317.
[34] C.J. Smithells, in: E.A. Brandes, G.B. Book (Eds.), SmithellsÕ
Metals Reference Book, seventh ed., Butterworth-Heinemann
Ltd., London, 1992.
[9] A.M. Fennimore, T.D. Yuzvinsky, W. Han, M.S. Fuhrer, J.
Cumings, A. Zettl, Nature 424 (2003) 408.
[10] R.J. Chen, Y. Zhang, D. Wang, H. Dai, J. Am. Chem. Soc. 123
(2001) 3838.
[35] M. Terrones, P.M. Ajayan, F. Banhart, X. Blase, D.L. Carroll,
J.C. Charlier, R. Czerw, B. Foley, N. Grobert, R. Kamalakaran,
P. Kohler-Redlich, M. Ruhle, T. Seeger, H. Terrones, Appl.
¨
[11] W. Huang, S. Taylor, K. Fu, Y. Lin, D. Zhang, T.W. Hanks,
A.M. Rao, Y.-P. Sun, Nano Lett. 2 (2002) 311.
Phys. A 74 (2002) 355.