Decomposition of CO-H2 over graphite nanofiber-supported iron and iron-copper catalysts

Article abstract of DOI:10.1021/jp031235a

We investigated the effect on the catalytic behavior of iron and iron-copper induced by supporting the particles on platelet graphite nanofibers. The decomposition of CO/H₂ was selected as a probe reaction to ascertain the manner by which this unique type of support medium influences the activity and selectivity of the metal catalysts. For comparison purposes a corresponding set of experiments was carried out where the metal and bimetallic particles were dispersed on silica, a traditional support material. It was found that the overall performance of the graphite nanofiber-supported metal catalysts was superior to that of the corresponding silica-supported systems. Furthermore, it was significant to find that the structures of the solid carbon products were entirely different from those produced by the respective unsupported metal catalyst systems. These modifications in the behavior of the iron-containing particles are discussed in terms of the impact of a metal-support interaction.

Full text of DOI:10.1021/jp031235a

J. Phys. Chem. B 2004, 108, 13307-13314
13307
Decomposition of CO-H2 over Graphite Nanofiber-Supported Iron and Iron-Copper
Catalysts
†
O. C. Carneiro, P. E. Anderson, N. M. Rodriguez, and R. T. K. Baker*
Catalytic Materials LLC, 1750 Washington Street, Holliston, Massachusetts 01746
ReceiVed: NoVember 10, 2003; In Final Form: May 4, 2004
We investigated the effect on the catalytic behavior of iron and iron-copper induced by supporting the particles
on “platelet” graphite nanofibers. The decomposition of CO/H was selected as a probe reaction to ascertain
2
the manner by which this unique type of support medium influences the activity and selectivity of the metal
catalysts. For comparison purposes a corresponding set of experiments was carried out where the metal and
bimetallic particles were dispersed on silica, a traditional support material. It was found that the overall
performance of the graphite nanofiber-supported metal catalysts was superior to that of the corresponding
silica-supported systems. Furthermore, it was significant to find that the structures of the solid carbon products
were entirely different from those produced by the respective unsupported metal catalyst systems. These
modifications in the behavior of the iron-containing particles are discussed in terms of the impact of a metal-
support interaction.
Introduction
structure of GNF can be such that only edge regions are
revealed. The ability to tailor the structural characteristics of
The performance of a supported catalyst system is governed
to a large degree by a combination of the chemical nature of
GNF so that the product has the blend of properties associated
with active carbons (high surface area) and graphite (high
electrical conductivity) makes these materials ideal for use as
the reactant gas and the strength of the metal-support
interaction.1-8 These factors together with the reaction condi-
2
2
support media.
tions can induce perturbations in the metal crystallites that
modify both the reactivity and the selectivity of the catalyst
and ultimately the pathway of the reaction.9,10 Further modifica-
tions can result from the incorporation of a second metal to the
formation of a bimetallic catalyst system. This procedure can
alter the adsorptive and chemical properties of the catalyst,
Previous work from this group has demonstrated the dramatic
improvements in catalytic performance of iron and iron-copper
for the conversion of ethylene achieved by supporting the
particles on “platelet” GNF compared to that found from similar
2
3
metal loading on active carbon and γ-alumina. An investiga-
2
4
making its overall performance quite different from that of either
tion by Salman and co-workers showed that when nickel was
supported on GNF, the gas-phase hydrogenation of crotonal-
dehyde to crotyl alcohol was significantly increased over that
where the metal was dispersed on γ-alumina. This unexpected
behavior was attributed to a combination of factors, the unusual
crystallographic orientation of nickel particles, and the possibility
that electronic perturbations were being induced in the metal
from interactions with the electrically conductive nanofiber
support media. In a recent study GNF-supported ruthenium-
barium catalysts were found to exhibit remarkably high activity
of the respective monometallic entities.1
1-14
It has been sug-
gested that the observed enhancement in the performance
exhibited by bimetallics can be attributed to the components
being in a state that is more easily reduced due to a weakening
of the strength of the metal-support interaction that ensues from
alloy formation.15
The electrical conductivity properties of graphite offer some
interesting possibilities when the material is used as a support
medium for small metal particles.16,17 Unfortunately, in its
conventional form, graphite has limitations for this type of
2
5
26
and stability for ammonia synthesis. French workers inves-
tigated the use of palladium dispersed on “herringbone” GNF
for the liquid-phase hydrogenation of cinnamaldehyde and
reported that with this catalyst system it was possible to achieve
2
application due to the low surface area (1 m /g) and restricted
number of reactive sites (edge regions). In this respect,
significant improvements could be achieved if the material
possessed a higher surface area coupled with a larger fraction
of exposed edge sites and a smaller cross-sectional area. These
conditions are met with certain types of graphite nanofibers
9
8% selectivity to hydrocinnamaldehyde at 80 °C. This result
was compared with that obtained with a commercial 5-wt %
Pd/charcoal catalyst where the selectivity to hydrocinnamalde-
hyde was only 40% and the major product was phenylpropanol.
They suggested that the enhancement in performance of the
GNF-supported palladium catalyst could be due to the elimina-
tion of mass-transfer processes that are generally encountered
with carbon-supported metals. They also invoked the existence
of an unusual metal-support interaction; a finding that had
previously been reported from XFAS studies of GNF-supported
(GNF), which are produced from the metal-catalyzed decom-
position of selected carbon-containing gases at temperatures
ranging from 500 to 900 °C.1
8-20
Individual GNF are typically
between 5 and 100 nm in width and 5 and 100 µm in length.
These materials consist of graphene sheets aligned in definite
directions that are dictated by the catalytic entity chosen for
the growth process.21 Unlike conventional graphite materials
where for the most part only the basal plane is exposed, the
27
metals.
The potential of GNF to function as an electrode for fuel-
cell applications was tested using the electrochemical oxidation
*
To whom correspondence should be addressed.
Chemistry Department, Northeastern University, Boston, MA 02115.
†
1
0.1021/jp031235a CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/07/2004

1
3308 J. Phys. Chem. B, Vol. 108, No. 35, 2004
Carneiro et al.
of methanol at 40 °C as a probe reaction.28 It was found that
catalysts comprised of 5-wt % Pt supported on either “platelet”
or “ribbon”-type GNFs exhibited activities comparable to those
observed with 25-wt % platinum on Vulcan carbon. In addition,
GNF-supported platinum particles were substantially more
resistant to CO poisoning. The observed improvement in
performance was attributed to the fact that the metal particles
adopted specific crystallographic orientations when dispersed
on the highly tailored GNF structures. Steigerwalt and co-
workers later carried out analogous studies in which they
investigated the behavior of platinum-ruthenium (1:1) alloy
clusters dispersed on “herringbone” GNF as the anode for a
working direct methanol fuel cell. They reported that the
performance of this electrode was 50% better than that of an
unsupported platinum-ruthenium colloid of similar surface area
and catalyst particle size.
The objective of the current investigation is to determine the
impact of the chemical and physical properties of the support
on the iron- and iron-copper-catalyzed decomposition of CO/
H2 mixtures. Specifically, we compared the catalytic activity
and selectivity of the iron-containing particles dispersed on
highly ordered “platelet” GNF with that observed for the same
metal loading on amorphous silica. In these experiments both
the gas-phase and solid-phase products have been examined.
regulated by the use of MKS mass flow controllers, allowing a
constant composition of a desired reactant feed to be delivered
to the system. Catalyst samples (200 mg) were placed in a
ceramic boat in the center of the reactor tube inside a Lindberg
split furnace. Initially, the catalyst was reduced in a 10%H2/He
mixture while the temperature was raised to 350 °C (600 °C
for silica-supported catalysts) and held at these respective levels
for 2 h. The difference in the respective reduction steps results
from the fact that one must avoid catalyzed hydrogenation of
the nanofiber support, which will occur at temperatures in excess
of 350 °C. On the other hand, it is necessary to carry out this
step at higher temperatures to ensure complete reduction of the
metals on the oxide support. Following the reduction step, the
system was brought to the desired reaction temperature while
the reactor was flushed with helium. At this stage a CO-H2
(4:1) mixture was introduced into the system at a total flow
2
9
3
rate of 100 cm /min, and then the reaction was allowed to
proceed for various periods up to 5 h at temperatures ranging
from 500 to 700 °C. The reaction was monitored as a function
of time by sampling the inlet and outlet gas streams at regular
intervals. The gas-phase products were analyzed by a Varian
3400 gas chromatography unit using a 30 m megabore (GS-Q)
capillary column. The amounts of solid carbon generated by
the catalyst were determined by weight difference and were
within 5% of the value calculated from carbon mass balances.
The characteristics of the solid carbon products were deter-
mined from a combination of transmission electron microscopy
Experimental Section
Materials. The “platelet” GNF employed as a catalyst support
was prepared from the decomposition of a CO-H2 gas mixture
over an iron-based catalyst at 600 °C following a previously
described protocol.30 The metal catalyst used in the growth
process of the nanofibers was removed by dissolution in a 1 M
HCl solution over a period of 7 days. The nanofibers were then
thoroughly washed in deionized water and dried in air at 110
(TEM). TEM examinations of the materials were performed in
a JEOL 2000EXII instrument (point-to-point resolution of 0.18
nm). Transmission specimens were prepared by ultrasonic
dispersion of a small quantity of a given sample in isobutanol
and then application of a drop of the supernate onto a holey
carbon film. Size distribution profiles were constructed from
measurements of over 300 particles in each system. BET surface
areas, calculated from nitrogen adsorption at -196 °C, were
carried out on various materials using a Coulter Omnisorp
°
C. Subsequent X-ray diffraction analysis together with trans-
mission electron microscopy failed to reveal the presence of
any metal particles following this treatment. BET surface area
1
00CX automated unit.
measurements of the purified material gave values ranging from
2
6
0 to 80 m /g. Amorphous fumed silica (M-5 Cab-O-Sil) used
Results
Flow Reactor Studies. The activity of the iron-based catalysts
2
as a support media had a BET surface area of 234 m /g.
The iron- and iron-copper-supported catalysts used in this
investigation were prepared by a standard incipient wetness
technique using an alcohol solution of iron and copper nitrates
mixed in the desired ratios to give a 5 wt % total metal loading
on the respective support media. The impregnates were dried
overnight in air at 110 °C, calcined in air at 250 °C for 4 h to
convert the metal nitrates to oxides, and then reduced at 350
toward the formation of solid carbon was determined gravi-
metrically as a function of composition and reaction temperature
for both support systems. Examination of the data for “platelet”
GNF-supported catalysts presented in Figure 1 shows that the
maximum yield of solid carbon occurs in the temperature range
550-600 °C. It is evident that the addition of copper to iron
does not enhance the formation of carbon. Indeed, as the fraction
of copper is progressively raised so the amount of solid carbon
generated is decreased. Inspection of the corresponding data
for the silica-supported iron-based particles shows the existence
of some major differences to the former systems, Figure 2. In
contrast to the “platelet” GNF-supported systems, in this case
maximum activity for solid carbon growth is observed at about
500 °C and decreases at higher temperatures. Furthermore, the
addition of 50% or less of copper to iron promotes the activity
of the catalyst toward carbon formation. Despite this trend,
however, it is clear that the amount of carbon produced from
the GNF-supported iron catalysts is significantly higher than
that obtained when silica is utilized as the carrier medium.
The gas-phase product distribution analysis was carried out
for the decomposition of CO-H2 (4:1) over the 5-wt % Fe
supported on “platelet” GNF and silica, and these data are shown
in Figures 3 and 4, respectively. Inspection of these plots reveals
the existence of some fundamental differences between the two
°
C in a 10% H2/He mixture for 24 h. For the silica-supported
metal catalysts this step was continued for a period of 72 h in
order to ensure complete reduction of the oxides to the metallic
state. After the reduction step, the samples were cooled to room
temperature under flowing helium and then passivated in a 2%
air-helium mixture for 1 h prior to removal from the reactor.
The gases used in this work, helium (99.999%), hydrogen
(99.99%), and carbon monoxide (99.99%), were obtained from
Air Products Inc. Traces of iron carbonyl were removed from
the CO by passing the gas through a heated coil maintained at
200 °C. All gases were passed through molecular sieve traps to
remove traces of water. Reagent-grade cupric nitrate [Cu(NO3)2‚
3
H2O] and ferric nitrate [Fe(NO3)3‚9H2O] were obtained from
Fisher Scientific for the catalyst preparations.
Apparatus and Procedures. The apparatus used for the
catalyst studies consisted of a horizontal quartz reactor tube,
which was connected to gas supply lines and operated at
atmospheric pressure. The gas flow to the reactor was precisely

Decomposition of CO-H2
J. Phys. Chem. B, Vol. 108, No. 35, 2004 13309
Figure 1. Amount of solid carbon formed as a function of reaction temperature from the interaction of “platelet” GNF-supported iron-containing
particles with CO/H (4:1).
2
Figure 2. Amount of solid carbon formed as a function of reaction temperature from the interaction of silica-supported iron-containing particles
with CO/H (4:1).
2
catalyst systems. The decomposition of CO is about an order
of magnitude higher over the Fe/GNF than the Fe/SiO2 system.
A further aspect concerns the amount of CO2 that is generated
in the two systems. When the reaction is conducted over Fe/
GNF, the amount of CO2 formed is close to that of solid carbon.
In sharp contrast, when the decomposition reaction is carried
out over Fe/SiO2, the amounts of CO2 are appreciably lower
than those of solid carbon. These product relationships were
found to persist for all the respective supported Fe-Cu
bimetallic systems studied in this work as shown in Table 1. It
is evident therefore that the nature of the support medium is
exerting an impact on the selectivity of the reaction for all these
systems.
GNF- and silica-supported Fe-Cu (8:2) with CO/H2 (4:1) at
600 °C are presented in Figures 5 and 6, respectively. Inspection
of these plots shows that when the iron-containing particles were
dispersed on “platelet” GNF the size distribution was much
broader than that obtained from the corresponding silica-
supported system. Even larger deviations in distributions are
observed when one examines the widths of nanofibers generated
from these systems.
A summary of the average sizes of the iron-containing
particles along with the average widths of the graphite nanofibers
generated from the respective supported systems after reaction
in CO/H2 (4:1) at 600 °C for 1 h are presented in Tables 2 and
3, respectively. Inspection of these data reveals that dispersion
of the metallic phase on “platelet” GNF always produced
particles that were on average twice as large as those observed
with the corresponding silica-supported systems. It is also
evident that in all cases the widths of the nanofibers formed
during the CO decomposition reaction was on average signifi-
Characterization Studies. Examination of the solid carbon
products by TEM showed that for the most part the deposits
consisted exclusively of carbon nanofibers. Examples of the
typical size distributions of metal particles and the corresponding
widths of the nanofibers formed after interaction of “platelet”

13310 J. Phys. Chem. B, Vol. 108, No. 35, 2004
Carneiro et al.
2
Figure 3. Product distribution as a function of reaction time for the decomposition of CO/H (4:1) over 5 wt % Fe/GNF at 600 °C.
2
Figure 4. Product distribution as a function of reaction time for the decomposition of CO/H (4:1) over 5 wt % Fe/silica at 600 °C.
TABLE 1: Product Distribution from Various wt % Fe-Cu Bimetallics Supported on “Platelet” GNF and Silica after Reaction
in CO/H
2
(4:1) at 600 °C for 1 h
SiO
2
“platelet” GNF
Fe-Cu (8:2)
Fe-Cu (5:5)
Fe-Cu (2:8)
Fe-Cu (8:2)
Fe-Cu (5:5)
Fe-Cu (2:8)
CO
97.2
0.3
2.5
96.6
0.2
3.1
93.3
0.2
6.5
74.1
12.3
13.6
74.8
12.0
13.2
78.9
9.4
10.4
CO
2
C
s
cantly wider than those of the initial metal catalyst particles.
This relationship was most pronounced for the series of
A further feature exhibited by the latter system is that as the
fraction of copper in the bimetallic is progressively increased,
there is a concomitant decrease in metal particle size and
associated nanofiber width. It is significant that this aspect is
not evident with the silica-supported iron-containing catalysts.
The morphology and overall structural characteristics of the
nanofibers were ascertained from high-resolution TEM studies.
The appearance of nanofibers grown from the interaction of 5-wt
“platelet” GNF-supported iron and iron-copper catalysts. These
findings would appear to indicate that the metal particles
undergo a change in morphology prior to nanofiber growth. It
should be stressed, however, that once the growth of the
nanofibers commences, the width of the structures is ap-
proximately equal to that of the reconstructed catalyst particles.

Decomposition of CO-H2
J. Phys. Chem. B, Vol. 108, No. 35, 2004 13311
Figure 5. Size distributions of metal particles and the corresponding widths of the nanofibers formed after interaction of “platelet” GNF-supported
Fe-Cu (8:2) with CO/H (4:1) at 600 °C.
2
Figure 6. Size distributions of metal particles and the corresponding widths of the nanofibers formed after interaction of silica-supported Fe-Cu
8:2) with CO/H (4:1) at 600 °C.
(
2
TABLE 2: Average Widths of Metal Particles Supported on
Platelet” GNF and GNF Structures Formed after Reaction
in CO/H (4:1) at 600 °C for 1 h
TABLE 3: Average Widths of Metal Particles Supported on
“
Silica and GNF Structures Formed after Reaction in CO/H
2
2
(4:1) at 600 °C for 1 h
average
average
average
average
catalyst
particle size (nm) GNF width (nm)
catalyst
particle size (nm)
GNF width (nm)
5
5
5
5
wt % Fe/pGNF
16.2
13.1
11.1
12.7
83.4
62.9
57.7
43.2
5 wt % Fe/silica
5.2
5.6
7.6
8.1
11.8
12.2
17.9
14.0
wt % Fe-Cu (8:2)/pGNF
wt % Fe-Cu (5:5)/pGNF
wt % Fe-Cu (2:8)/pGNF
5 wt % Fe-Cu (8:2)/silica
5 wt % Fe-Cu (5:5)/silica
5 wt % Fe-Cu (2:8)/silica
%
Fe supported on “platelet” GNF is shown in the high-
silica samples were treated in CO/H2 (4:1) at 600 °C, electron
microscopic examination revealed that the structures were quite
different from those of the previous system. In this case, the
nanofibers were very thin and exhibited a “beadlike” appearance
as shown in Figure 8. Other workers have reported the formation
of these types of nanofiber structures during the metal-catalyzed
resolution electron micrograph, Figure 7. The structure can be
described as a pseudo-“herringbone” edge arrangement sur-
rounding a disordered core region. This type of nanofiber
structure is to be contrasted with that of the “platelet” GNF
support, which was generated from the interaction of Fe powder
with CO/H2 (4:1) at 600 °C and consists of graphite sheets
3
1,32
cracking of ethane and methane.
When GNF- and silica-
21
aligned in direction perpendicular to the growth axis. One can
readily distinguish the secondary GNF growths from the parent
supported 5 wt % Fe-Cu (8:2) were reacted in CO/H2 (4:1) at
600 °C, the nanofibers were found to exhibit similar structural
features to those formed from the corresponding supported iron
systems when treated under the same conditions.
“platelet” structures by the fact that the former always contained
an associated catalyst particle at the growing end. When Fe/

13312 J. Phys. Chem. B, Vol. 108, No. 35, 2004
Carneiro et al.
Figure 9. High-resolution micrograph showing the detailed structural
features of “tubular” nanofibers produced from the interaction of
“
2
platelet” GNF-supported Fe-Cu (8:2) with CO/H (4:1) at 700 °C.
Figure 7. High-resolution electron micrograph showing the appearance
of nanofibers grown from the interaction of 5-wt % Fe supported on
“
platelet” GNF.
Figure 10. Transmission electron micrograph of the “shell-like”
deposits formed on some of the GNF-supported Fe-Cu (8:2) particles
Figure 8. Transmission electron micrograph showing the “beadlike”
2
during reaction with CO/H (4:1) at 700 °C.
nanofiber structures generated from the Fe/silica-CO/H
at 600 °C.
2
(4:1) system
the characteristics of the nanofibers tended to follow the same
trend as a function of temperature as those generated from the
iron-rich catalysts, under these conditions the GNF tended to
acquire a more disordered structure, an example is presented in
Figure 11.
When the supported 5 wt % Fe-Cu (8:2) samples were
heated at 700 °C and reacted in CO/H2 (4:1), dramatic changes
in the structural features of the nanofibers were observed. The
high-resolution micrograph, Figure 9, shows the formation of
“
tubular”-type structures generated from the “platelet” GNF-
Discussion
supported bimetallic system. Close scrutiny of the edge regions
reveals that the graphite sheets constituting these materials are
aligned at a shallow angle with respect to the nanofiber growth
axis. As a consequence, the arrangement allows for exposure
of a small fraction of edges to the surrounding medium and
potential to build chemical functionality on the surfaces. Under
these conditions one also observes the formation of a “shell-
like” deposit in which the metal catalyst particles appeared to
be surrounded by several layers of graphite, Figure 10. Both
types of structures were formed on the corresponding silica-
supported Fe-Cu (8:2) samples; however, the relative widths
and yields of the “tubular” structures were lower than on the
GNF-supported particles. An analogous set of experiments was
carried out with GNF- and silica-supported Fe-Cu (5:5)
samples. Examination of these specimens showed that while
It is interesting to observe that the silica-supported metal
catalysts tend to deactivate after about 3 h on stream, whereas
the corresponding “platelet” GNF-supported systems showed
no loss of activity after 5 h of reaction. These observations are
consistent with those reported by Rodriguez and co-workers,23
who studied the conversion of hydrocarbons over Fe-Cu (7:3)
and Fe particles supported on “platelet” GNF, active carbon,
and γ-alumina. The performance of the GNF-supported iron-
containing particles was far superior to that of either of the other
two systems. Furthermore, it was found that the alumina-
supported catalysts underwent severe deactivation after about
20 min on stream. In contrast, with the GNF-supported metal
system, maximum activity was maintained for a prolonged
period of time.

Decomposition of CO-H2
J. Phys. Chem. B, Vol. 108, No. 35, 2004 13313
interactions with a conductive support is likely to exert an
influence on the events occurring at both the metal-gas and
metal-carbon interfaces. As a consequence, one would expect
to find significant differences in the structural characteristics
of the nanofibers generated from the two supported metal
systems.
It is instructive to compare the structural characteristics of
the nanofibers produced from the supported iron-based catalyst
particles in the present study with those grown under the same
reaction conditions from the corresponding unsupported pow-
3
4
dered metal catalysts. In the latter case, the Fe-rich Cu
bimetallics all produced highly graphitic nanofibers that exhib-
ited a “platelet” structure at 600 °C. When the reaction was
carried out at 700 °C, the structural conformation changed from
a “platelet” to “tubular” conformation that correlated with the
R-Fe to γ-Fe phase change over this temperature range.
Comparison of these data highlights the manner by which a
support medium can exert an impact on catalytic behavior of
small metal particles toward nanofiber growth. In this regard,
3
5
it should be noted that Anderson and Rodriguez reported a
similar pattern of behavior during their study of nanofibers
formed from the interaction of supported and unsupported
nickel-iron particles with CO/H2 mixtures.
Conclusions
The results of this investigation have demonstrated that the
choice of the bimetallic catalyst, the nature of the support
medium, and the temperature at which the reaction is performed
are all factors that ultimately must be considered in the design
of the structural architecture of graphite nanofibers. Furthermore,
the dimensions of the nanofibers that are dictated by the sizes
of the metal catalyst particles responsible for their growth are
also dependent upon the nature of the interaction with the
support.
Figure 11. High-resolution micrograph showing the disordered ap-
pearance of nanofibers formed by silica-supported Fe-Cu (5:5) samples
following interaction with CO/H (4:1) at 600 °C.
2
References and Notes
This difference in catalyst behavior could not be rationalized
on the basis of metal particle sizes. Alternative explanations
were suggested that were predicated on the notion that the metal
particles adopted a preferred crystallographic orientation when
dispersed on the highly tailored nanofiber support.23 Under these
circumstances, particular metal faces would be exposed to the
reactant environment. One might tentatively attribute this
variation in catalyst behavior to differences in the morphological
characteristics adopted by the metal particles on the two
supports. A further factor that must be taken into consideration
concerns the electronic properties of the respective support
media. Since the “platelet” GNF is conductive in nature, the
possibility exists for electron transfer between the metal particles
and the support. Such a phenomenon is expected to exert an
impact on the catalytic performance of the system. This
condition would not be encountered with silica as it is an
insulator.
(1) Schwarb, G. M. AdV. Catal. 1978, 27, 1.
(2) Solymosi, F. Catal. ReV.-Sci. Eng. 1967, 1, 233.
(3) Ichikawa, M. J. Catal. 1979, 59, 67.
(4) Verykios, X. E.; Stein, F. P.; Coughlin, R. W. J. Catal. 1980, 66,
47.
1
(
5) Tauster, S. J.; Fung, S. C.; Baker, R. T. K.; Horsley, J. A. Science
981, 211, 1121.
6) Derouane, E. G.; Baker, R. T. K.; Dumesic, J. A.; Sherwood, R.
1
(
D. J. Catal. 1981, 69, 101.
(7) Ruckenstein, E.; Chu, Y. F. J. Catal. 1979, 59, 109.
(8) Stevenson, S. A.; Dumesic, J. A.; Baker, R. T. K.; Ruckenstein, E.
In Metal-Support Interactions in Catalysis, Sintering and Redispersion; Van
Nostrand Reinhold: New York, 1987.
(9) Satterfield, C. N. In Heterogeneous Catalysis in Practice; McGraw-
Hill: New York, 1980.
(10) Baker, R. T. K. J. Adhes. 1995, 52, 13.
(
11) Sinfelt, J. H. In Bimetallic Catalysts; Exxon Monograph, John Wiley
&
Sons: New York, 1983.
(
(
(
12) Sachtler, W. H. M.; van Santen, R. A. AdV. Catal. 1977, 26, 69.
13) Sachtler, W. H. M. Faraday Discuss. Chem. Soc. 1981, 72, 7.
14) Sinfelt, J. H.; Via, G.; Lytle, F. W. Catal. ReV.-Sci. Eng. 1984, 26,
Differences in the observed structural characteristics of the
solid carbon products can also be attributed to the effect of the
respective metal-support interactions. It is well established that
the geometric shape and crystallographic arrangement of the
exposed faces of the metal catalyst particles determines the
alignment and perfection of graphite sheets constituting the
81.
(15) Rao, C. N. R.; Kulkarni, G. U.; Kannan, K. R.; Chaturvedi, S. J.
Phys. Chem. 1992, 96, 7379.
(
(
16) Brownlie, I. C.; Fryer, J. R.; Webb, G. J. Catal. 1969, 64, 263.
17) Gallezot, P.; Richard, D.; Bergeret, G. In NoVel Materials in
2
1
nanofibers. In this regard, reference should be made to the
work of Yang and Chen,33 who reported that in a nickel
crystallite certain metal faces would preferentially decompose
CO whereas another set of faces would favor the precipitation
of dissolved carbon in the form of graphite. Clearly, the
existence of electronic perturbations in the metal resulting from
Heterogeneous Catalysis; Baker, R. T. K., Murrell, L. L., Eds.; ACS
Symposium Series 437; American Chemical Society: Washington, D.C.,
1990; p 150.
(18) Baker, R. T. K. In Carbon Fibers, Filaments and Composites;
Figueiredo, J. L., et al., Eds.; NATO ASI Series 177; Kluwer: Dordrecht,
1990; p 405.
(19) Rodriguez, N. M. J. Mater. Res. 1993, 8, 3233.

13314 J. Phys. Chem. B, Vol. 108, No. 35, 2004
Carneiro et al.
(
20) Laurent, C.; Flahaut, E.; Peigney, A.; Rousset, A. New J. Chem.
998, 1229.
21) Rodriguez, N. M.; Chambers, A.; Baker, R. T. K. Langmuir 1995,
1, 3862.
(28) Bessel, C. A.; Laubernds, K.; Rodriguez, N. M.; Baker, R. T. K. J.
Phys. Chem. B 2001, 105, 1115.
(29) Steigerwalt, E. S.; Deluga, G. A.; Lukehart, C. M. J. Phys. Chem.
B 2001, 105, 8097.
1
(
1
(
22) de Jong, K. P.; Geus, J. W. Catal. ReV.-Sci. Eng. 2000, 42, 481.
(30) Rodriguez, N. M.; Kim, M. S.; Baker, R. T. K. J. Catal. 1993,
144, 93.
(23) Rodriguez, N. M.; Kim, M. S.; Baker, R. T. K. J. Phys. Chem.
1
994, 98, 13108.
(31) Rodriguez, N. M.; Kim, M. S.; Fortin, F.; Mochida, I.; Baker, R.
T. K. Appl. Catal., A: Gen. 1997, 148, 265.
(32) Zhang, T.; Amiridis, M. D. Appl. Catal., A: Gen. 1998, 167, 161.
(33) Yang, R. T.; Chen, J. P. J. Catal. 1989, 115, 52.
(34) Carneiro, O. C.; Kim, M. S.; Yim, J. B.; Rodriguez, N. M.; Baker,
R. T. K. J. Phys. Chem. B 2003, 107, 4237.
(24) Salman, F.; Park, C.; Baker, R. T. K. Catal. Today 1999, 53, 3862.
(25) Liang, C.; Li, Z.; Qiu, J.; Li, C. J. Catal. 2002, 211, 278.
(26) Pham-Huu, C.; Keller, N.; Ehret, G.; Charbonniere, L. J.; Ziessel,
R.; Ledoux, M. J. J. Mol. Catal., A: Chem. 2001, 170, 155.
27) Mojet, B. L.; Hoogenraad, M. S.; van Dillen, A. J.; Geus, J. W.;
Koningsberger, D. C. J. Chem. Soc., Faraday Trans. 1997, 93, 4371.
(
(35) Anderson, P. E.; Rodriguez, N. M. J. Mater. Res. 1999, 14, 2912.