Effect of Catalyst Support on Carbon Growth by CVD
J. Phys. Chem. B, Vol. 109, No. 13, 2005 6101
4. Conclusions
CNTs with high quality, provided that a “milder” atmosphere
such as low active hydrocarbons are used for carbon growth.
The carbon yield and nanostructure have been studied on
powder and alumina-supported Fe catalysts from the decom-
position of different carbon-containing gases. Opposite reactivity
was demonstrated for CO or hydrocarbon decomposition over
the two catalysts. Completely different or similar structures were
synthesized depending on the nature of the gas precursors. It is
reasoned that the reactivity and the structure of carbon deposits
are dictated by the size and crystallographic faces of the catalyst
particles, which are in turn a direct consequence of the strength
of metal-support interaction, and are further influenced by the
reactants during carbon growth. Hydrogen plays an essential
role in the processes by surface reconstruction, keeping the
catalyst surface clean of carbon, and satisfying dangling bonds.
Based on the experimental results and existing CNT growth
models, a mechanism is proposed to illustrate the growth of all
carbon nanostructures. Earlier studies have suggested that
powder catalysts generate larger diameter fibers than the
supported ones. The findings from this study point out that it is
not always true, because the final diameter of the fibers is also
reactant dependent.
It is further interesting to correlate the carbon growth rate
with the final diameter of the carbon filaments (Tables 1 and
2). It turns out that the higher the carbon growth rate, the larger
the diameter of the carbon filaments, thus the resulting catalyst
particles (CO vs hydrocarbons). This can be understood by the
dissolution of carbon in Fe particles, which will lead to a
semiliquid state of the system, at a temperature far below the
melting point of Fe.45 It is reasonable to say that carbon dissolves
in the metal with a higher concentration when the growth is
faster. This leads to the higher mobility of the metal particles
to form larger filaments.
3.4. Growth Mechanism. Various models for filamentous
carbon formation have been proposed.39,46,47 It is normally
accepted that carbon growth involves surface carbon decom-
position, carbon diffusion, and then precipitation. However,
different views exist on how different carbon nanostructures
nucleate, and especially on how carbon fibers grow. Based on
previous available models and the experimental observations
in this study, a mechanism can be proposed to explain the
growth of all carbon structures.
The Fe particles initially decompose the carbon-containing
molecules on the catalyst surfaces and lead to the formation of
surface carbon. The formation of surface carbon could probably
go through an intermediate iron carbide phase.45,48 Then carbon
dissolves and diffuses on the surface or through the bulk of the
metal particles. When carbon supersaturation is reached, carbon
segregation will occur. Carbon atoms will move towards the
surface and combine to form an initial graphene layer, as sug-
gested by B.Nagy et al.47 This early stage of carbon nucleation
on the particle surface is probably common for both CNTs and
CNFs. Carbon will deposit in various structures afterward,
depending on the crystallographic faces of the resulting catalyst
particles, which is a function of the metal-support interaction,
the gas precursors, and particularly hydrogen.
Finally, a major factor to emerge from this investigation is
the development of processes to synthesize relatively large
quantities of CNTs, fishbone-tubular fibers, platelet fibers, and
onionlike carbon encapsulating magnetic particles, all with high
quality and selectivity. The findings here are important for
realizing controlled synthesis of different carbon nanostructures
for their applications in various fields.
Acknowledgment. We thank the Norwegian Research
Council for financial support.
References and Notes
(1) Kroto, H. W.; Heath, J. R.; Obrien, S. C.; Curl, R. F.; Smalley, R.
E. Nature 1985, 318, 162.
(2) Oberlin, A.; Endo, M.; Koyama, T. J. Cryst. Growth 1976, 32,
335.
For CNT growth, the initial nucleated graphene layer may
detach from the particle surface and form a fullerene-like
cap.39,47 This cap is always chemisorbed to the catalyst particles
during the growth. This will avoid the growth of any open edges,
which would expose energetically costly dangling bonds.39
Newly arrived carbon atoms will nucleate additional graphite
layers beneath the initial cap and in the inward direction,39,45
or move toward the graphene-metal interface and be incorpo-
rated into the graphite sheets. This contributes to the lengthening
of the tube.
The growth of CNFs is possible in the presence of hydro-
gen: the initial nucleated graphene layer will be directly
segregated as graphite sheets instead of the formation of a
fullerene-like cap. Hydrogen probably initiates the segregation
of the graphene layer.28 The exposure of energetically costly
dangling bonds of the released graphite sheets is possible
because hydrogen can saturate the dangling bonds. Newly
arrived carbon atoms will nucleate additional graphene layers,
and CNFs then grow in a layer-by-layer manner, which
contributes to the lengthening of the fibers. This accounts for
the growth of both platelet and fishbone-tubular structures.
(3) Iijima, S. Nature 1991, 354, 56.
(4) Iijima, S.; Ichihashi, T. Nature 1993, 363, 603.
(5) de Jong, K. P.; Geus, J. W. Catal. ReV.sSci. Eng. 2000, 42, 481.
(6) Nikolaev, P.; Bronikowski, M. J.; Bradley, R. K.; Rohmund, F.;
Colbert, D. T.; Smith, K. A.; Smalley, R. E. Chem. Phys. Lett. 1999, 313,
91.
(7) Toebes, M. L.; Bitter, J. H.; van Dillen, A. J.; de Jong, K. P. Catal.
Today 2002, 76, 33.
(8) Kukovecz, A.; Konya, Z.; Nagaraju, N.; Willems, I.; Tamasi, A.;
Fonseca, A.; B.Nagy, J.; Kiricsi, I. Phys. Chem. Chem. Phys. 2000, 2, 3071.
(9) Huang, Z. P.; Wang, D. Z.; Wen, J. G.; Sennett, M.; Gibson, H.;
Ren, Z. F. Appl. Phys. A 2002, 74, 387.
(10) Takenaka, S.; Serizawa, M.; Otsuka, K. J. Catal. 2004, 222, 520.
(11) Pham-Huu, C.; Keller, N.; Roddatis, V. V.; Mestl, G.; Schlogl, R.;
Ledoux, M. J. Phys. Chem. Chem. Phys. 2002, 4, 514.
(12) Park, C.; Keane, M. A. J. Catal. 2004, 221, 386.
(13) Anderson, P. E.; Rodriguez, N. M. J. Mater. Res. 1999, 14, 2912.
(14) Liao, X. Z.; Serquis, A.; Jia, Q. X.; Peterson, D. E.; Zhu, Y. T.;
Xu, H. F. Appl. Phys. Lett. 2003, 82, 2694.
(15) Li, W. Z.; Wen, J. G.; Ren, Z. F. Appl. Phys. A 2002, 74, 397.
(16) Li, W. Z.; Wen, J. G.; Tu, Y.; Ren, Z. F. Appl. Phys. A 2001, 73,
259.
(17) Hernadi, K.; Konya, Z.; Siska, A.; Kiss, J.; Oszko, A.; B.Nagy, J.;
Kiricsi, I. Mater. Chem. Phys. 2002, 77, 536.
(18) Hernadi, K. Chem. Phys. Lett. 2002, 363, 169.
(19) Anderson, P. E.; Rodriguez, N. M. Chem. Mater. 2000, 12, 823.
(20) Ermakova, M. A.; Ermakov, D. Y.; Chuvilin, A. L.; Kuvshinov,
G. G. J. Catal. 2001, 201, 183.
(21) Kuznetsov, V. L.; Usoltseva, A. N.; Chuvilin, A. L.; Obraztsova,
E. D.; Bonard, J. M. Phys. ReV. B 2001, 64.
(22) Park, C.; Baker, R. T. K. J. Catal. 2000, 190, 104.
(23) Bhattacharyya, A.; Chang, W.; Kleefisch, M. S.; Udovich, C. A.
U.S. Patent 5,767,040, 1998.
(24) Kang, Y. S.; Risbud, S.; Rabolt, J. F.; Stroeve, P. Chem. Mater.
1996, 8, 2209.
If there is not enough energy to nucleate CNTs or CNFs, for
example over large particles, carbon will deposit as a film
encapsulating the particles and form an onionlike structure. The
size of the catalyst particle is also influenced by the metal-
support interaction and the reactants. In all cases surface carbon
encapsulation surely occurs at some rate, which finally deac-
tivates the catalysts.