Catalytic behavior of graphite nanofiber supported nickel particles. 2. The influence of the nanofiber structure

Article abstract of DOI:10.1021/jp981210p

The hydrogenation of ethylene, 1-butene, and 1,3-butadiene has been used as a probe reaction in an attempt to monitor any possible changes in catalytic behavior induced by supporting nickel on different types of graphite nanofiber support materials. This study is designed to compare the catalytic behavior of the metal particles when dispersed on three types of nanofibers, where the orientation of the graphite platelets within the structures is significantly different in each case. The metal crystallites are located in such a manner that the majority of particles are in direct contact with graphite edge regions. It should be emphasized, however, that there are subtle differences in the spacing between adjacent exposed carbon atoms in the various nanofiber structures. As a consequence, it is highly probable that the atomic arrangement of the surfaces of nickel particles that nucleate on these different graphite edges will be dictated to a large degree by the interaction with the atoms in nanofiber supports. Under such circumstances one might reasonably expect that different crystallographic faces of nickel will be exposed to the reactant gas depending on which type of nanofiber structure is used as the supporting medium. For comparison purposes, the same set of hydrogenation reactions were carried out under similar conditions over γ-alumina supported nickel particles.

Full text of DOI:10.1021/jp981210p

5168
J. Phys. Chem. B 1998, 102, 5168-5177
Catalytic Behavior of Graphite Nanofiber Supported Nickel Particles. 2. The Influence of
the Nanofiber Structure
Colin Park and R. Terry K. Baker*
Department of Chemistry, Northeastern UniVersity, Boston, Massachusetts 02115
ReceiVed: February 19, 1998; In Final Form: April 22, 1998
The hydrogenation of ethylene, 1-butene, and 1,3-butadiene has been used as a probe reaction in an attempt
to monitor any possible changes in catalytic behavior induced by supporting nickel on different types of
graphite nanofiber support materials. This study is designed to compare the catalytic behavior of the metal
particles when dispersed on three types of nanofibers, where the orientation of the graphite platelets within
the structures is significantly different in each case. The metal crystallites are located in such a manner that
the majority of particles are in direct contact with graphite edge regions. It should be emphasized, however,
that there are subtle differences in the spacing between adjacent exposed carbon atoms in the various nanofiber
structures. As a consequence, it is highly probable that the atomic arrangement of the surfaces of nickel
particles that nucleate on these different graphite edges will be dictated to a large degree by the interaction
with the atoms in nanofiber supports. Under such circumstances one might reasonably expect that different
crystallographic faces of nickel will be exposed to the reactant gas depending on which type of nanofiber
structure is used as the supporting medium. For comparison purposes, the same set of hydrogenation reactions
were carried out under similar conditions over γ-alumina supported nickel particles.
Introduction
to this finding, they observed that at low nickel loadings a
significantly greater increase in the nickel(2p) binding energy
occurred on C60 than on graphite surfaces, indicating the
existence of a much stronger metal-support interaction in the
former system. In a further aspect relating to carbonaceous
support materials, Schlogl and co-workers6 have recently
demonstrated that ruthenium particles appear to be extremely
stable with respect to sintering when dispersed on fullerene-
based support media. It was suggested that this behavior was
due to the interaction of the metal clusters with the non-six-
membered carbon rings in the support structure. It was also
claimed that the ability of the support to anchor the ruthenium
moieties was responsible for the observed high catalytic
hydrogenation activity exhibited by this system.7
The discovery of buckminsterfullerene, the so-called “bucky
ball” molecule,1 along with carbon nanotubes2 and the develop-
ment of catalytically grown carbon nanofibers3 have led to a
subsequent avalanche of papers dealing with the characterization
and potential uses of these carbonaceous structures. These
unique materials all possess chemical and physical properties
that could be of significant importance in a number of areas,
including gas storage and separation, reinforcement agents in
polymeric matrixes, electronic devices, and novel support
materials for metal catalyst particles.
The more conventional form of graphite, while offering sev-
eral desirable features as a catalyst support, does have major
limitations. The most serious of these is the fact that graphite
8
Planeix and co-workers investigated the hydrogenation of
has a relatively low surface area, approximately 1.0 m2 g-1
.
cinnamaldehyde and reported that 0.2 wt % ruthenium supported
on carbon nanotubes was a highly selective catalyst for this
reaction, up to 92% cinnamyl alcohol being produced at an
overall high conversion of cinnamaldehyde. While these results
are extremely encouraging there are several drawbacks associ-
ated with the use of carbon nanotubes, not least being the fact
that they are expensive to generate in large quantities and
possess a relatively low surface area, typically of the order of
25 m2 g-1. The majority of any impregnated metal species is
located on the outer carbon shell, a condition that would be
expected if the proposed tubular model for the structure of the
material is correct. This factor will tend to limit both the number
and the different types of interaction that can be created between
the metal particles and the carbon nanotube surfaces.
Amorphous carbon offers substantial advantages as a catalyst
support when compared with the traditional oxide support
materials. This form of carbon is relatively inexpensive to pro-
duce and exhibits a range of surface areas, up to 1000 m2 g-1
in its activated form. Catalyst precursors supported upon carbon
media are readily reduced to the metallic state, a condition that
generally requires prolonged treatment in hydrogen when the
same metal salts are dispersed on an oxide support.4
It is now being recognized that the recently developed types
of carbon materials, which possess a variety of extraordinary
structures, afford some exciting opportunities for exploiting the
potential of nanofibers as novel catalyst supports for metal
particles. The interaction of transition metals with buckmin-
sterfullerene has been investigated by Vijayakrishnan and co-
workers5 using photoemission spectroscopy. These workers
established that, in contrast to the situation encountered with
graphite, nickel strongly interacted with C60, causing significant
changes in the C(1s) and C(2s) binding energies. In addition
The development of methods to tailor both the degree of
crystalline perfection and the orientation of graphite platelets
that constitute the nanofibers has overcome many of the
limitations encountered with other more traditional forms of
carbon.9,10 One of the most outstanding features of these
structures is the presence of a large number of edges, which in
* To whom correspondence should be addressed.
S1089-5647(98)01210-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/05/1998

Graphite Nanofiber Supported Ni Particles
J. Phys. Chem. B, Vol. 102, No. 26, 1998 5169
turn constitute sites that are readily available for chemical and
physical interactions. Perhaps one of the most unexpected
findings is that such ordered crystalline solids can exhibit high
surface areas (50-500 m2 g-1), where the totality of the surface
area is chemically active. It this these characteristics that we
have attempted to take advantage of in the present study.
The growth and characterization of carbon nanofibers pro-
duced from the interaction of a variety of pure metals and
bimetallic catalysts with hydrocarbons have been extensively
investigated by Baker and co-workers.11-14 By judicious choice
of the catalyst and careful control of the reaction conditions, it
is possible to tailor the growth of graphite nanofibers to generate
structures of a desired conformation where the platelets are
oriented in a particular direction with respect to the fiber axis.
Figure 1. Schematic representation of the three unique conformations
of graphite nanofibers used as a catalyst support material.
While the electron microscopy studies demonstrate that
graphite nanofiber supported metal particles exhibit structural
features that are generally associated with the existence of a
strong interfacial reaction,15 the data acquired up to this point
in time do not allow one to reach any definitive conclusions
with regard to the possibility of preferred growth locations. If
the metal establishes a very strong interaction with carbon atoms
located at the “armchair” face, then one could speculate that
the particles would adopt a different morphology with these
sites compared to that found at either the “zigzag” or basal plane
sites, and as a consequence, one might expect to find concomi-
tant variations in catalytic activity and selectivity. It has been
well documented in the literature that specific faces of a metal
particle are required to catalyze certain reactions.16-18 Yang
and Chen19 demonstrated by selected area electron diffraction
that Ni(100) and Ni(110) faces were present at the metal/gas
interface, whereas Ni(111) and Ni(311) faces were in contact with
a graphite substrate during the formation of carbon nanofibers
from the metal-catalyzed decomposition of hydrocarbons.
Initial studies by Rodriguez and co-workers20 showed that
Fe-Cu particles supported on graphite nanofibers exhibited a
significantly higher catalytic activity for the conversion of
hydrocarbons than when the same loadings of the bimetallic
were supported on either activated carbon or γ-alumina. A
subsequent study by Chambers and co-workers15 established that
the performance of nickel particles supported on carbonaceous
materials was extremely sensitive to the degree of crystalline
perfection of the substrate; the highest activity for the hydro-
genation of 1-butene and 1,3-butadiene was displayed by a
system in which the metal was dispersed on nanofibers
possessing a high graphitic content. These results were in
accord with those from earlier work by Brownlie and co-
workers,21 who reported that, for the hydrogenation of selected
hydrocarbons, catalysts based on palladium decorated graphite
samples exhibited a significantly higher activity and selectivity
compared to that where the metal was impregnated on other
less ordered carbonaceous solids.
The use of 1,3-butadiene has long been recognized as being
a convenient and proficient method to probe the surface
characteristics of supported metal catalysts.22-24 For a series
of supported metal catalysts the hydrogenation of 1,3-butadiene
offers an ideal means of simultaneously comparing the activity
and selectivity of these systems. The selective hydrogenation
of 1,3-butadiene has been studied extensively by Wells and co-
workers25-27 for a range of catalyst systems.
Experimental Section
Materials. The three types of graphite nanofibers used in
this work, which are shown in the schematic diagram, Figure
1, were produced by passing a desired mixture of a carbon-
containing gassethylene and/or carbon monoxidesand hydro-
gen over a specific mono- or bimetallic catalyst powder for
periods of up to 3 h. The nanofibers are classified as possessing
a “platelet” structure where the graphite platelets in the carbon
material are oriented perpendicular to the fiber axis, (Figure
1a), a “spiral” form where the platelets in the helical nanofibers
are oriented parallel to the fiber axis, (Figure 1c), and a “ribbon”
structure where the graphite platelets are aligned parallel to the
fiber axis (Figure 1b). The exact conditions used to grow these
nanofibers have been described in detail in previous pa-
pers.9,13,14,28 The resultant structures were demineralized in
dilute nitric acid over a period of 7 days to remove the original
metal catalyst. The nanofibers were then thoroughly washed
in deionized water before being dried overnight at 110 °C and
then stored in sealed vessels.
The 5 wt % supported nickel catalysts used in this study were
prepared by a standard incipient wetness technique. Each
support medium was impregnated with a solution of nickel
nitrate containing the appropriate amount of the precursor salt
necessary to achieve the desired 5 wt % metal loading. The
nickel nitrate was dissolved in 25 mL of ethanol, since in the
“as-prepared condition” the graphite nanofibers are hydrophobic
in nature and poor metal dispersion was achieved when an
aqueous solvent was used. After drying overnight at 110 °C,
the impregnated materials were calcined in air at 250 °C for 4
h to convert the metal nitrate to the oxide and then reduced at
350 °C in a 10% H2/He mixture for 20 h. The reduced catalysts
were cooled in He before being passivated at ambient temper-
ature in a 2% O2/He mixture for 1 h prior to removal from the
reactor. The catalyst precursor was deliberately calcined at a
lower temperature to avoid gasification of the graphite nanofiber
support material. With this safeguard no loss of support material
was recorded during the catalyst preparation. In the case of
the γ-alumina supported metal catalyst samples the time of the
hydrogen treatment at 350 °C was varied from 20 to 60 h in
order to check that complete reduction of the nickel oxide was
In the current investigation we have used the hydrogenation
of simple alkenes and dienes including, ethylene, 1-butene, and
1,3-butadiene as probe reactions in an attempt to monitor any
possible changes in catalytic behavior induced by supporting
nickel on three different types of nanofibers, where the
orientation of the graphite platelets within the structures is
significantly different in each case. These experiments have
enabled us to evaluate not only the potential of the graphite
nanofibers as a support medium but also the impact of the
orientation of the graphite platelets on the catalytic activity and
selectivity of the metal. For comparison purposes, the same
set of hydrogenation reactions were carried out under similar
conditions over γ-alumina supported nickel particles.

5170 J. Phys. Chem. B, Vol. 102, No. 26, 1998
Park and Baker
being achieved. It was significant to find that the subsequent
performance of these catalysts appeared to be the same, within
experimental error, indicating that in all cases the supported
particles were in the metallic state.
The gases used in this work, helium (99.999%), hydrogen
(99.999%), carbon monoxide (99.99%), ethylene (99.95%),
1-butene (99.95%), and 1,3-butadiene (99.95%), were obtained
from MG Industries and used without any further purification.
Reagent grade nickel nitrate [Ni(NO3)2‚6H2O] was obtained
from Fisher Scientific for the catalyst preparation. The γ-alu-
mina was supplied by the Degussa Corporation and was used
as received.
Apparatus and Procedures. The apparatus used for the
catalyst studies consisted of a quartz flow reactor, fitted with a
quartz frit at the midpoint of the tube, and this assembly was
located within a split vertical tube furnace. The gas flow to
the reactor was precisely 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
(100 mg) were placed on the quartz frit, and the tube was aligned
in such a manner that this region of the reactor was always in
approximately the same position in the furnace. After reduction
in a 10% H2/He mixture for 2 h at 400 °C, the system was
cooled to the desired reaction temperature, and the reactant
hydrocarbon gas, or a predetermined hydrocarbon/H2/He mix-
ture, was introduced to the catalyst sample for periods of up to
3 h. The reaction was monitored as a function of time by
sampling the inlet and outlet gas streams at regular intervals.
The reactants and products were analyzed by gas chromatog-
raphy (Varian 3400 unit) using a 30 m megabore (GS-Q)
capillary column for reactions involving ethylene and a 30 m
megabore (GS-AL) capillary column for those where 1-butene
and 1,3-butadiene were the reactants. Turnover frequencies
(TOF) for all reactions were estimated using the average metal
particle sizes as determined from TEM measurements and
calculated according to the accepted procedure.29
Figure 2. Comparison of the CO2 temperature-programmed oxidation
profiles of the graphite nanofibers.
90° and compared with the known X-ray powder files to
determine the phase identities.
Efforts were made to determine metal particle sizes by
standard hydrogen chemisorption procedures; however, serious
difficulties were encountered with the graphite nanofiber
supported nickel systems, since copious amounts of hydrogen
were absorbed by these types of carbon solids, and as a
consequence, meaningful data were only obtained for the metal
particles dispersed on γ-alumina.
Results
1. Characterization Studies. 1.1. XRD Analysis. X-ray
diffraction studies of the nanofibers failed to detect the presence
of any of the original catalyst in the demineralized sample,
indicating that removal of metal by treatment in acid was a
satisfactory method. The predominant peak was identified as
being the (002) peak of graphite, giving a clear indication of
the highly crystalline nature of these nanofiber structures. The
average interlayer spacing of each graphitic nanofiber structure
was also determined using XRD. The “spiral-like” nanofibers
had a d spacing of 0.337 nm, 0.002 nm larger than that of either
the “platelet” or “ribbon” types of nanofiber structures.
The characteristics of all the support materials and catalyst
samples were established using a combination of techniques,
including high-resolution transmission electron microscopy
(HRTEM), temperature-programmed oxidation (TPO), X-ray
diffraction (XRD), and nitrogen BET surface area measure-
ments. In addition, the nickel catalysts were examined after
use by HRTEM to ascertain any changes in either the support
structure or the morphology of the metal particles induced by
exposure to a hydrocarbon/hydrogen environment. These
studies were carried out in a JEOL 2000EXII electron micro-
scope fitted with a high-resolution pole piece capable of a lattice
resolution of 0.14 nm. Suitable transmission specimens were
prepared by ultrasonic dispersion of catalyst samples in isobu-
tanol and application of a drop of the resultant suspension to a
holey carbon support grid.
1.2. Temperature-Programmed Oxidation. CO2 temperature-
programmed oxidation profiles indicate the degree of crystalline
perfection, that is the graphitic nature, of each type of nanofiber
support material. A carbonaceous material that has a high
degree of crystalline perfection will start to gasify under these
conditions at approximately 850 °C. In the other extreme, if
the material is amorphous in nature, the material will start to
undergo gasification at approximately 600 °C. It is evident from
Figure 2 that the nanofibers used in this study are highly
graphitic materials with an extremely low amorphous carbon
content. The graphite nanofibers that were comprised of a
“ribbon” structure were identified as being the most graphitic
in nature followed by the “platelet” and the “spiral” conforma-
tions.
Temperature-programmed oxidation studies were carried out
on the demineralized graphite nanofibers in the presence of CO2
at a constant heating rate using a Cahn 2000 microbalance as
outlined previously.14 This procedure has been shown to
provide an ideal means of establishing the graphitic content of
a carbonaceous solid. BET surface areas, calculated from
nitrogen adsorption isotherms at -196 °C, were carried out on
the various support media using a Coulter Omnisorp 100CX
surface analysis unit. Powder XRD patterns of the graphite
nanofibers and all catalyst samples were obtained with a Scintag
diffractometer using nickel filtered Cu KR radiation. Diffraction
patterns were recorded over a range of 2θ angles from 10° to
1.3. Surface Area Measurements. The N2 BET surface area
of demineralized samples of each catalyst support, measured at
-196 °C, are summarized in Table 1. The surface area of the
demineralized graphite nanofibers varied according to the type
of conformation, but in all cases these values were substantially
higher than that of a conventional graphite sample, typically
<1 m2/g, due to the extremely large number of edge regions
that are exposed in the former materials. The surface area of
the nanofibers can be significantly enhanced by careful activa-
tion in CO2 at 850 °C to achieve values of up to 700 m2 g-1
.

Graphite Nanofiber Supported Ni Particles
J. Phys. Chem. B, Vol. 102, No. 26, 1998 5171
TABLE 1: Calculated Nitrogen BET Surface Areas of the
Catalyst Support Materials
ments a larger probe molecule, 1-butene, was used. This par-
ticular reactant molecule was chosen since it offers the pos-
sibility of isomerization in addition to hydrogenation pathways,
thus highlighting changes in the selectivity pattern due to the
influence of the support medium. From Figure 4 and Table 4
it is apparent that when a 1-butene/hydrogen (1:1) reactant
mixture was passed over nickel at 90 °C, the observed catalytic
activity and the selectivity patterns were strongly dependent
upon the nature of the supporting medium. Under these con-
ditions a much higher activity was attained with nickel deposited
on the three types of graphite nanofibers compared to that where
the metal particles were supported on γ-alumina. In all cases,
however, the primary product of this reaction was n-butane,
resulting from the complete hydrogenation, and this was fol-
lowed by isomerization of the reactant molecules to cis- and
trans-2-butene in varying ratios, which were highly dependent
upon the particular catalyst system employed. Inspection of
the reaction profiles shows that when nickel was supported on
either the “platelet” or “spiral” forms of graphitic nanofibers,
an induction period of between 35 and 60 min was required in
order to achieve a high conversion rate, a criterion that was not
observed when the metal was dispersed on either the “ribbon-
like” nanofibers.
support material
surface area (m2 g-1
)
“ribbon-like” graphite nanofibers
“spiral-like” graphite nanofibers
“platelet-like” graphite nanofibers
γ-alumina
85
45
120
115
TABLE 2: Average Size of Nickel Crystallites Dispersed on
the Various Types of Support Material Measured from
Transmission Electron Micrographs
support material
av particle size (nm)
“ribbon-like” graphite nanofibers
“spiral-like” graphite nanofibers
“platelet-like” graphite nanofibers
γ-alumina
8.1
7.2
6.4
1.4
The surface area of the more traditional catalyst support,
γ-alumina, was found to be similar to that of the graphite
nanofibers.
1.4. Transmission Electron Microscopy Examination. High-
resolution transmission electron microscopy studies indicated
that the graphite nanofibers were between 100 and 150 nm in
width and varied from 5 to 50 µm in length. Examination of
the nanofiber supported nickel catalyst samples revealed that,
in all cases, the metal particles were located along the edges of
each particular graphite nanofiber structure. When nickel was
supported on the graphite nanofibers, the average crystallite size,
as determined from measurements of over 400 particles in each
sample, was found to be significantly greater than that where
the metal was dispersed on γ-alumina, as outlined in Table 2.
It was an arduous task to discern the appearance of metal
particles on γ-alumina specimens due to the interference of the
support texture; however, it was possible to establish that in
this case the metal particles did not adopt any preferred
geometrical shape. In sharp contrast, metal particles supported
on the three types of graphitic nanofibers acquired definite
hexagonal shapes, an example of which is presented in Figure
3. The morphological characteristics exhibited by these particles
are generally those associated with a system that exhibits a
strong metal-support interaction.
2. Flow Reactor Studies. 2.1. Catalytic Hydrogenation of
Ethylene by Supported Nickel Particles. From the data pre-
sented in Table 3 it is evident that the ability of nickel to
catalytically hydrogenate ethylene at a relatively low temper-
ature, 100 °C, does not appear to exhibit any substantial
variations when the metal is dispersed on the different support
media. An overall high conversion of ethylene, primarily to
ethane and to a lesser extent solid carbon, was achieved when
a C2H4/H2 (1:1) mixture was passed over each supported nickel
catalyst system. The introduction of nickel onto graphite
nanofibers possessing a “ribbon-like” structure did, however
result in a marginally higher activity for this reaction than that
of the other systems. It is evident that the turnover frequencies
(TOF) for the hydrogenation of ethylene are similar for all these
supported nickel systems.
On close scrutiny of the data in Table 4 showing the percent
product distribution from the hydrogenation of 1-butene at 90
°C over these supported nickel catalysts, several salient features
can be identified. It is evident that the particular orientation of
the graphite platelets within the nanofiber structures exerts a
strong influence on the selectivity of the supported nickel
particles for this reaction. The selectivity toward the formation
of 2-butenes, S1, is similar for all three types of graphite
nanofiber supported nickel systems and appreciably higher than
that realized from the nickel/γ-alumina catalyst. It is also
noticeable that the trans/cis-2-butene ratio is extremely sensitive
to orientation of the graphite platelets in the nanofiber structures,
decreasing according to the sequence “ribbon-like” >“spiral”
> “platelet” forms. Clearly therefore, a change in orientation
of the graphite platelets from being aligned parallel to being
perpendicular to the fiber axis modifies the structure of the
supported nickel crystallites in such a manner that favors the
formation of cis-2-butene over that of trans-2-butene.
The turnover frequencies for the hydrogenation of 1-butene
over the various nickel catalysts exhibit a dramatic change as a
function of the support media. Nickel deposited on the “ribbon”
type of nanofibers is seen to exhibit the highest TOF, more than
twice that observed when the same metal loading is dispersed
on the “platelet” nanofibers. An intermediate value is found
for nickel supported on the “spiral” form of nanofibers, which
is similar to that calculated for the hydrogenation of ethylene
over this catalyst system.
2.3. Catalytic Hydrogenation Studies of 1,3-Butadiene by
Nickel. In this series of experiments an attempt was made to
ascertain the impact that the arrangement of the graphite platelets
in the nanofiber support materials exerted on the hydrogenation
activity and selectivity of nickel for 1,3-butadiene, a more
demanding and sensitive probe molecule than 1-butene. The
hydrogenation of the diene can occur by several different
reaction pathways giving rise to numerous products, thereby
allowing for the simultaneous comparison of the activity and
selectivity.
An interesting feature was observed with a catalyst consisting
of nickel supported on the “platelet” type of graphite nanofibers.
At 100 °C an induction period of 35 min was required for this
system to achieve an overall conversion of ethylene similar to
that displayed by the other supported nickel catalysts. This
reactivity sequence was found to persist when the temperature
was increased up to a maximum level of 200 °C.
2.2. Catalytic Hydrogenation of 1-Butene by Nickel. In an
attempt to gain further insights into the catalytic behavior of
these supported nickel catalysts, in a second series of experi-
Figure 5 shows the overall conversion of a 1,3-butadiene/
hydrogen (1:1) mixture at 90 °C over the various supported Ni
catalyst systems. From these plots it is apparent that the overall
conversion of the diene is very limited over all these catalyst

5172 J. Phys. Chem. B, Vol. 102, No. 26, 1998
Park and Baker
Figure 3. High-resolution transmission electron micrograph highlighting the morphology and size of Ni particles deposited on the “platelet” form
of graphite nanofibers. The direction of the lattice fringes of the nanofibers are indicated.
TABLE 3: Selectivity of Supported Nickel Catalysts for a
C2H4/H2 (1:1) Reactant Mixture at 100 °C
% conversion of ethylene to selected products
%
TOF
catalyst support
methane ethane solid C conversn (s-1
)
“ribbon-like” GNF
“spiral-like” GNF
“platelet-like” GNF
γ-alumina
0.02
0.04
0.01
0.03
63.69
71.64
52.16
68.78
8.60
11.61
7.37
72.31
83.29
59.54
79.39
0.376
0.384
0.242
0.301
10.58
systems, being approximately 7% for all graphite nanofiber
supported metal systems and less than 4% when nickel was
deposited on γ-alumina. When nickel was supported on the
“spiral” type of nanofibers, the system exhibited a significant
decrease in activity for this reaction over the first 35 min on
stream settling down to level comparable to that obtained with
the other nickel/graphite nanofiber catalysts. Despite the low
activity achieved with these supported nickel catalysts, a
comparison of the product selectivities raises several very key
points. One general feature to emerge was that butenes, partially
hydrogenated molecules, were the major products. In this series
of experiments conditions were deliberately selected so as to
minimize the complete hydrogenation of 1,3-butadiene to
butane, and this situation was achieved by limiting the concen-
tration of hydrogen in the reactant mixture.
Figure 4. Effect of the support on the catalytic activity of 5 wt % Ni
in the hydrogenation of a C4H8/H2 (1:1) reactant mixture at 75 °C.
data presented in Figure 6. Under these conditions, the catalytic
activity of nickel supported on graphite nanofibers possessing
either “platelet” or “spiral” conformations exhibited a dramatic
increase over that where the reaction was performed in a
equimolar reactant mixture; however, in both cases an induction
period of about 35 min is required to reach the optimum
conversion level. On the other hand, the activity of systems
where the metal particles were supported on γ-alumina and or
on nanofibers with a “ribbon” conformation did not change to
any significant degree upon exposure to the reactant feed
containing a larger fraction of hydrogen.
When the reaction was conducted with a 1,3-butadiene/
hydrogen (1:2) mixture at 90 °C over the same series of
supported nickel, a dramatic enhancement in catalytic activity
of some of the systems was observed as evidenced from the

Graphite Nanofiber Supported Ni Particles
J. Phys. Chem. B, Vol. 102, No. 26, 1998 5173
TABLE 4: Selectivity of Supported Nickel Catalysts for a C4H8/H2 (1:1) Reactant Mixture at 90 °Ca
% conversion of 1-butene to selected products
catalyst support
i-butane
n-butane
trans-2-butene
cis-2-butene
S1
T/C
TOF (s-1
)
“ribbon-like” GNF
“spiral-like” GNF
“platelet-like”GNF
γ-alumina
0.11
0.10
0.09
0.08
67.81
61.52
42.48
0.38
13.88
11.90
4.98
0
10.34
10.26
9.01
0
0.26
0.26
0.25
0
1.34
1.09
0.55
0
0.479
0.386
0.230
2.0 × 10-3
a S1 ) (trans-2-butene + cis-2-butene)/(trans-2-butene + cis-2-butene + butane). T/C ) trans-2-butene/cis-2-butene.
butene is favored when nickel was supported on “spiral”
nanofibers, whereas cis-2-butene was the preferred isomer when
the metal was dispersed on the “platelet” form. In contrast,
the 1-butene to 2-butene ratio is substantially higher when the
“platelet” nanofibers are used as the catalyst support. Although
nickel supported on the “ribbon-like” graphite nanofibers
exhibited an extremely low overall activity for the hydrogenation
of 1,3-butadiene, it was found to display the highest selectivity
toward the formation of butenes, and in particular to trans-2-
butene.
Finally, the turnover frequency for the hydrogenation of 1,3-
butadiene over nickel supported on the “ribbon-like” nanofibers
decreased dramatically when compared to the corresponding
values calculated for the hydrogenation of ethylene and 1-butene.
In sharp contrast, this trend was reversed, to a somewhat lesser
degree, when the metal was supported on either the “platelet”
or “spiral” graphite nanofiber forms.
Figure 5. Effect of the support on the catalytic activity of 5 wt % Ni
in the hydrogenation of a C4H6/H2 (1:1) reactant mixture at 90 °C.
Discussion
The most important feature to emerge from this investigation
is the finding that when nickel is supported on graphite
nanofibers, the system can exhibit some unusual patterns of
behavior with respect to the catalytic hydrogenation of olefins
and dienes. Moreover, the performance of the supported metal
catalyst for a given reaction is extremely sensitive to the
particular geometric arrangement adopted by the graphite
platelets constituting the structure of the nanofibers. We have
demonstrated that it is possible to create a highly selective and
active catalyst for the hydrogenation of both ethylene and
1-butene to the corresponding alkanes by supporting 5 wt %
nickel on certain types of graphite nanofibers. The optimum
performance for these reactions was achieved with a catalyst
in which the nickel was dispersed on graphite nanofibers
possessing a “ribbon-like” conformation. A comparison of
graphite nanofiber supported nickel catalysts with a more
conventional oxide supported metal system, nickel/γ-alumina,
showed that in all cases the former combinations exhibited the
superior performance.
Conventional wisdom might suggest that the observed be-
havior of the latter system could be rationalized according to
the notion that when nickel was dispersed on the oxide support,
complete reduction to the active metallic state was not achieved
during the preparation step. There are a number of factors that
argue against this possibility; hydrogen chemisorption measure-
ments were consistent with a dispersion corresponding to an
average metal particle size of about 1.5 nm. The catalytic
behavior of alumina supported nickel particles did not appear
to exhibit any significant changes as a result of increasing the
time of reduction in the preparative step from 20 to 60 h. The
5 wt % nickel/alumina catalyst exhibited a relatively high
activity for the hydrogenation of ethylene, and portions of this
same sample were subsequently used for the hydrogenation
reactions of 1-butene and 1,3-butadiene.
Figure 6. Effect of the support on the catalytic activity of 5 wt % Ni
in the hydrogenation of a C4H6/H2 (1:2) reactant mixture at 90 °C.
Examination of the product distributions, given in Table 5,
shows that the major product from all catalysts is n-butane,
which produced as a result of the complete hydrogenation of
1,3-butadiene. Although nickel particles supported on both the
“spiral” and “platelet” types of graphite nanofiber conformations
exhibited a similar overall activity, substantial differences are
apparent in the product distributions derived from these two
systems. It can be seen that when the metal was supported on
graphite nanofibers possessing a “spiral” conformation, the
major route was the complete hydrogenation of the diene to
n-butane, whereas when nickel was dispersed on “platelet-like”
nanofibers, the pathways leading to partial hydrogenation to
1-butene and complete hydrogenation to n-butane were equally
favored. This aspect is highlighted in Table 5 where S2, the
selectivity to the partially hydrogenated products, is observed
to be substantially higher when nickel was supported upon
“platelet” nanofibers rather than on “spiral” nanofibers. Sig-
nificant differences between these two catalysts can also be seen
from a comparison of the ratios of trans/cis-2-butene and
1-butene/2-butene, respectively. The formation of trans-2-
Perhaps the most intriguing aspect of these reactions is the
manner by which the selectivity patterns exhibited by nickel

5174 J. Phys. Chem. B, Vol. 102, No. 26, 1998
Park and Baker
TABLE 5: Conversion of a C4H6/H2 (1:2) Reactant Mixture at 90 °C to Selected Products for a Series of Supported Nickel
Catalystsa
% conversion of 1,3-butadiene to selected products
catalyst support
n-butane
1-butene
cis-2-butene
trans-2-butene
S2
B1/B2
T/C
TOF (s-1
)
“ribbon-like” GNF
“spiral-like” GNF
“platelet-like” GNF
γ-alumina
1.10
62.77
39.26
0
3.68
10.40
34.89
0.04
0.92
11.58
14.83
0.20
1.51
15.28
14.09
0.03
0.85
0.37
0.60
1
1.51
0.39
1.43
0.17
1.64
1.32
0.95
0.15
0.038
0.461
0.401
5.1 × 10-3
a S2 ) (1-butene + trans-2-butene + cis-2-butene)/(1-butene + trans-2-butene + cis-2-butene + butane). B1/B2 ) 1-butene/(trans-2-butene +
cis-2-butene). T/C ) trans-2-butene/cis-2-butene.
are modified by the structural characteristics of the supporting
medium. For the most part, all the catalyst systems behaved
in a similar fashion during reaction in an ethylene/hydrogen
(1:1) mixture. As the complexity of the hydrocarbon reactant
molecule was increased, so differences in the overall catalytic
activity and selectivity of the various systems became apparent.
The predominant product from the 1-butene/hydrogen reaction
was n-butane; however, when graphite nanofibers were used
as the support, an appreciable fraction of the reactant underwent
isomerization to either trans- or cis-2-butene. The observed
increase in the formation of isomerization products when
nanofibers were used as the support is believed to be directly
related to the modification in morphological characteristics of
the nickel crystallites induced by the interaction with these types
of novel graphite materials.
Figure 7. Schematic representation of the “armchair” and “zigzag”
faces of a graphite platelet.
The nickel-catalyzed hydrogenation of 1,3-butadiene is a more
demanding reaction that presents the opportunity to probe the
existence of any further nuances contained within the various
product distributions that might be associated with the structural
characteristics of the supporting medium. In this series of
experiments nickel supported on both the “platelet” and “ribbon-
like” forms of graphitic nanofibers was found to be highly
selective toward the partial hydrogenation of 1,3-butadiene to
1-butene rather than proceeding to undergo complete hydroge-
nation to n-butane. The relatively low activity of the metal when
supported on the “ribbon” type of nanofibers was unexpected,
especially since this particular catalyst system had displayed
the highest activity for the hydrogenation of both ethylene and
1-butene.
1. Influence of the Support on the Morphological Char-
acteristics of Nickel Particles. High-resolution transmission
electron microscopy studies provided some key insights regard-
ing the morphological characteristics adopted by the nickel
crystallites on the graphite nanofibers. It was apparent that
many of the metal particle characteristics were similar on all
three types of graphite nanofibers and quite different to those
observed when alumina was used as the catalyst support. On
the graphite supports the crystallites were found to adopt very
thin, hexagonal morphologies, and since the electron density
was uniform across a given particle, one may tentatively
conclude that they were relatively flat. These growth features
are generally believed to originate from the creation of a strong
metal-support interaction that induces a spreading action of
the metal on a carrier surface. In contrast, a more globular
particle geometry was prevalent when nickel was supported on
γ-alumina, consistent with the existence of a somewhat weaker
metal-support interaction. For comparison purposes it would
have been desirable to generate similar sized particles on all
the support media; however, even when the metal loading was
increased by an order of magnitude with the nickel/γ-alumina
system, the crystallites were still significantly smaller than those
formed from the lighter loading on the carbon materials.
In this regard it is instructive to examine the data obtained
from controlled atmosphere electron microscopy studies of the
nickel/graphite-hydrogen system performed by Baker and co-
workers.30-33 They observed that catalytic hydrogenation of
the graphite proceeded by a channeling mode and that the tracks
propagated by active nickel particles tended to be straight with
occasional 60° and 120° bends and aligned along the 1120
crystallographic orientations of the graphite. These findings
indicated that in the presence of hydrogen the metal exhibited
a preferential wetting action with respect to the “armchair” faces
of graphite and that the particles moved in such a manner so as
to maintain contact with carbon atoms in this arrangement. It
is therefore not unreasonable to assume that the interaction of
nickel with the “zigzag” faces of graphite will be somewhat
weaker. These experimental observations were found to be in
complete agreement with the theoretical studies reported by
Abrahamson,34 who calculated the surface energies of a number
of metals on the different faces of graphite. On the basis of
these conclusions, one can confidently predict that subtle
differences are likely to exist in both the morphological and
chemical characteristics of metal particles that are formed on
the “armchair” and “zigzag” faces of graphite, respectively. In
the present circumstances the graphite nanofiber surfaces are
in essence acting as templates for the generation of metal
particles in a desired orientation. Upon consideration of the
arrangement of carbon atoms in the prismatic faces of graphite,
the so-called “armchair” {1120} and “zigzag” {1010} faces, it
is very easy to visualize from Figure 7 that the former face
presents a much more open structure for the deposition of metal
when compared to that of the “zigzag” face. Since the ratio of
“armchair” to “zigzag” sites will vary from one type of graphite
nanofiber structure to another, dramatic changes in both the
catalytic activity and selectivity of nickel may be explained in
terms of the variations in crystallographic orientations of the
particles associated with the differing structures of the supporting
media.
2. Ramifications of Graphite Nanofiber Structure on the
Catalytic Behavior of Nickel. It is an interesting exercise to
examine the catalytic data obtained for nickel dispersed on the

Graphite Nanofiber Supported Ni Particles
J. Phys. Chem. B, Vol. 102, No. 26, 1998 5175
three types of graphite nanofibers in light of the possible
differences in morphological characteristics of the metal particles
that can exist on these particular support structures. For this
purpose it is necessary to refer to the schematic structures
presented in Figure 1. It can be seen that the surfaces exposed
by the “ribbon-like” nanofibers consist of two relatively large
basal plane regions, each of which are bounded by two long
and two short edge regions. At this time it is not possible to
establish the precise orientation of these two edges; however,
since they are aligned in directions that are perpendicular to
one another, one can state with certainty that the grouping of
the carbon atoms along these two regions will be different, i.e.,
one edge will adopt the “armchair” configuration and the other
will be in the “zigzag” arrangement. Preliminary theoretical
calculations and modeling studies have indicated that the long
edge region of this type of nanofiber is most likely to be in the
“zigzag” orientation. On this type of material the majority of
the metal species will tend to accumulate on the long edge
region, and those particles that collect on other locations are
not likely to make a significant contribution to the overall
catalytic performance of the system. By comparison, graphite
nanofibers that possess a “platelet” structure have two edges of
approximately the same length that are oriented to an equal
degree in either the “armchair” or “zigzag” configurations.
Finally, the “spiral-like” nanofibers are comprised of graphitic
platelets that are oriented in a manner similar to the “ribbon-
like” material; however, in this case the geometrical require-
ments associated with these structures place a great deal of strain
on the platelets, resulting in an average d spacing of 0.337 nm,
a value that is 0.002 nm larger than that of single-crystal
graphite. This expansion of the layers will result in a modi-
fication of the interaction between the nanofiber surface and
the impregnated metal atoms. We believe that this feature
accounts for the observed difference in behavior of this system
and where the nickel is dispersed on the “ribbon-like” nanofi-
bers. Indeed, a significant difference in the hydrogenation
behavior of nickel was observed when the metal was supported
on the these two types of carbon structures. Nickel was found
to be highly active for the hydrogenation of 1,3-butadiene when
supported on the “spiral” type of nanofibers, but in contrast,
the metal was virtually inactive when dispersed on the “ribbon-
like” material.
was observed. As a consequence, it is probable that under these
circumstances the formation of 2-butenes would appear to
proceed via both pathways to similar extents.
It was significant to find that when nickel was supported on
γ-alumina, the trans/cis ratios, calculated from the hydrogenation
of 1-butene and 1,3-butadiene, became increasingly different,
indicating that on this particular catalyst system 2-butenes appear
to be produced directly by the 1,4 addition of hydrogen to diene.
Based on these data, it is evident that there is a divergence in
the isomerization ability of these catalysts. Nickel itself has
been reported to be a highly effective catalyst for the isomer-
ization of alkenes, but when supported on various materials,
the ability of the metal to perform this function can be radically
changed.35 Once again it becomes apparent that the orientation
of the graphite platelets in the graphite nanofiber supports and
the associated surface atomic arrangement of the dispersed
nickel crystallites play a critical role in determining the activity
and selectivity of the catalyst system. Moreover, it is clear that
major differences exist in the catalytic performance of the
supported nickel particles depending on whether the graphite
platelets constituting the nanofiber supports are oriented parallel
or perpendicular to the fiber axis.
Catalytic hydrogenation was originally considered to be a
structure-insensitive reaction, but it has become increasingly
evident from experimental data that this is not the case.36
Brownlie and co-workers17 claimed that the observed enhance-
ment in the catalytic activity of palladium particles that were
located at graphite edges compared to that exhibited by those
particles in contact with the basal plane region was directly
related to the strong metal-support interaction that existed at
the former regions, which was responsible for maintaining a
fine dispersion of metal crystallites. These conclusions were
later confirmed by Masson and co-workers,37-39 who demon-
strated that the rate of hydrogenation of ethylene occurred much
faster over small rather than large platinum particles when the
metal was dispersed on silica. This particle size effect has been
also observed with nickel38 and palladium40,41 supported on
alumina, silica, and carbon media.18,39
It is clear that the data obtained in present investigation are
at variance with the previous conclusions, since the hydrogena-
tion activity of the relatively large nickel particles dispersed on
the graphite nanofibers is significantly higher than that obtained
from an equal loading of very small crystallites of the same
metal supported on a traditional carrier, such as γ-alumina. In
this context it is important to emphasize the work of Boitiaux
and co-workers,42,43 who challenged the previously held notions
regarding the relationship between particle size and catalytic
activity. These workers investigated the performance of sup-
ported palladium catalysts for the hydrogenation of 1-butyne,
1,3-butadiene, and isoprene and demonstrated that there was
indeed a strong particle size dependency; however, they ob-
served a dramatic decrease in the catalytic activity when ex-
tremely small metal particles (<3 nm) were used for these
reactions. This trend was explained in terms of a stronger ad-
sorption of a these unsaturated, electron-rich molecules on the
smaller metal particles.42-44 It was somewhat surprising that
the hydrogenation of 1-butene was unaffected by the size of
the supported palladium particles. These results have since been
reproduced for a variety of metals and support combinations.44-46
A comparison of the trans/cis-2-butene (T/C) ratios for both
the hydrogenation of 1-butene and 1,3-butadiene obtained in
this investigation allows one to make a number of important
deductions. The formation of 2-butenes from the hydrogenation
of 1,3-butadiene has been purported to occur via two separate
pathways: isomerization of the partially hydrogenated product
1-butene prior to its desorption from the catalyst surface and
by a 1,4 addition of hydrogen to 1,3-butadiene.10-12 Both
pathways are equally valid, but if the trans/cis ratio is similar
for both hydrogenation reactions, then the formation of 2-butenes
has been postulated to proceed mainly via the isomerization of
1-butene. On the other hand, if significant differences in the
trans/cis ratios are found for the two systems, the reaction will
proceed via a 1,4 addition of hydrogen. Inspection of the data
presented in Tables 4 and 5 shows that the trans/cis-2-butene
ratios for these two hydrogenation reactions were similar when
nickel was supported on either the “spiral” or “ribbon-like”
forms of graphite nanofibers, indicating that in these cases the
formation of 2-butenes occurs primarily via the isomerization
of 1-butene. It is evident, however, that when the metal was
supported on the “platelet” type of nanofibers, a significant
difference in the trans/cis ratio for the same set of reactions
The inactivity displayed by nickel in the current investigation
when supported on γ-alumina cannot be readily explained in
similar terms as those proposed by Boitiaux and co-workers,42,43
since the system displayed a low activity for both the hydro-
genation of 1-butene and 1,3-butadiene but, on the other hand,

5176 J. Phys. Chem. B, Vol. 102, No. 26, 1998
Park and Baker
exhibited a high activity for the hydrogenation of ethylene. The
extremely small nickel particles, 1.4 nm in diameter, may
strongly adsorb both 1-butene and 1,3-butadiene but do not
appear to exert the same effect on ethylene.
crystals and found that the (100) face was virtually inactive
while the activity of the (111) face was about twice that of the
(100) face. Gallezot and co-workers50 reported significant
differences in the activity and selectivity of platinum for the
hydrogenation of cinnamaldehyde when the metal particles were
supported on graphite and charcoal. These variations were
attributed to the existence of a strong interaction between
platinum crystallites and the graphite support, a condition that
was not observed with charcoal. Goetz and co-workers51
stressed the importance of metal particle morphology in their
studies of the palladium-catalyzed hydrogenation of 1-butene
and 1,3-butadiene. They claimed that when palladium under-
went a strong interaction with γ-alumina to form thin, flat
particles, isomerization reactions tended to be highly favored.
In contrast, when the metal particles did not interact strongly
with the oxide support, hydrogenation was found to be the
predominant reaction.
The induction period required to achieve a steady-state con-
version observed when nickel was supported on the various
graphitic nanofibers might be an indication that a surface re-
construction phenomenon is operative during the initial stages
of many of these reactions. The reconstruction of a solid surface
has been attributed to the enhanced mobility of surface atoms,
resulting from a weakening of the metal-metal bond brought
about by the chemisorption of gaseous molecules.52-55 This
phenomenon has, however, been shown to be reversible in
nature, the surface eventually relaxing back to its original
arrangement upon desorption of the chemisorbed species. A
further factor that cannot be ignored is the possibility that the
delocalized π-electrons present within these graphitic nanofiber
structures might perturb the electronic structure of the metal
particles and, as a consequence, exert an impact on their catalytic
behavior. Finally, one should not overlook the possibility that
differences in the architectural features of the various types of
graphite nanofibers may have a profound effect on the manner
by which the hydrocarbon reactant molecules can be adsorbed
on the carbon structure in the vicinity of the metal particles. If
such behavior was operative, then one would expect it to be
most pronounced with the larger molecules, 1-butene and 1,3-
butadiene, rather than the less demanding matching requirements
encountered with ethylene.
Based on these criteria, it is apparent that one cannot account
for the observed higher activity displayed by nickel when
supported on graphite nanofibers for specific hydrogenation
reactions by merely presenting arguments based on average
metal particle sizes. While a reasonably high metal dispersion
on a given support material is desirable, it may not be the most
critical factor in determining the catalytic behavior of the system
in certain hydrogenation reactions. Clearly, the reasons for the
unexpected performance of the graphite nanofiber supported
metal systems require a more profound understanding. We
therefore offer an alternative explanation that focuses attention
on the morphological characteristics of the metal particles, a
property in which the support location site plays a pivotal role.
When one considers the structural features of the various types
of graphite nanofibers, these materials are unique in that for
the most part they present an abundance of edge sites to the
gas phase. With the choice of two distinct adsorption sites on
the graphitic platelets, the “armchair” and “zigzag” faces, metal
particles may be expected to adopt unusual geometries when
they nucleate and grow on these template regions. It is this
factor that we believe is largely responsible for dictating the
arrangement of metal atoms in the exposed faces of the particles
and, as a consequence, is responsible for controlling the catalytic
behavior of the system toward specific hydrogenation reactions.
An analogous set of arguments were proposed by Massardier
and co-workers47 to account for the observations that in the
hydrogenation of 1,3-butadiene over palladium, where the Pd(110)
face was found to be significantly more active than the Pd(111)
face.
Taking into consideration the data obtained here for the
hydrogenation of 1-butene and 1,3-butadiene along with that
derived from previous CAEM studies of the nickel/graphite-
hydrogen system,30-33 and if one assumes that the metal
crystallites adopt different orientations on the “armchair” and
“zigzag” faces, respectively, then it is possible to develop a
rationale to account for the catalytic behavior of the nickel/
graphite nanofiber systems. We suggest that the hydrogenation
of alkenes may be favored via adsorption at certain crystal-
lographic faces of the nickel particles, which are generated from
the interaction of the metal atoms with one particular face of
the graphitic nanofibers. This argument is predicated on the
notion that the graphite platelets act as templates for the creation
of nickel particles that possess a more accessible structure with
a smaller coordination number and, as such, generates conditions
for the formation of a more stable adsorbate. Further detailed
high-resolution transmission electron microscopy studies de-
signed to examine the relationship between the crystallographic
features of the metal particles and the orientation of the graphite
platelets in the various nanofiber structures are currently in
progress.
Conclusions
We have shown that the catalytic performance of nickel for
the hydrogenation of alkenes and dienes at moderate temper-
atures is extremely sensitive to the nature of the support medium.
When nickel was dispersed on three unique forms of graphite
nanofibers, significant improvements in the catalytic activity
and selectivity to specific products were attained over that using
the same metal loading on a more traditional support material,
γ-alumina. The observed variations in catalytic activity and
selectivity are attributed to the different orientations adopted
by the nickel particles on the support materials. It is suggested
that the graphite platelets in the nanofiber support media act as
templates for the dispersed nickel crystallites, which adopt a
specific geometry that is determined by their specific site
location.
In all cases studied in this work the metal crystallites are
located in such a manner that the majority of particles are in
direct contact with graphite edge regions. It should be
emphasized, however, that there are subtle differences in the
spacing between adjacent exposed carbon atoms in the various
nanofiber structures. As a consequence, it is highly probable
that the atomic arrangement of the surfaces of nickel particles
Corroboration for some of the arguments outlined above can
be obtained from the work of other groups. Dalmai-Imelik and
co-workers48 prepared nickel/silica catalysts where the crystal-
lites were oriented with either the Ni(111) or the Ni(110) planes
parallel to the support surface, a procedure that resulted in a
higher catalytic activity for the hydrogenation of ethylene than
that achieved from a system in which the metal particles were
distributed in a random fashion. In a further investigation
Dalmai-Imelik and Massardier49 examined the hydrogenation
of ethylene at 25 °C over three low index planes of nickel single

Graphite Nanofiber Supported Ni Particles
J. Phys. Chem. B, Vol. 102, No. 26, 1998 5177
(21) Brownlie, I. C.; Fryer, J. R.; Webb, G. J. Catal. 1969, 64, 263.
(22) Bond, G. C. Chem. Soc. ReV. 1991, 20, 441.
(23) Bond, G. C.; Wells, P. B. AdV. Catal. 1964, 15, 92.
(24) Bond, G. C.; Webb, G.; Wells, P. B.; Winterbottom, J. M. J. Chem.
Soc. 1965, 3128.
(25) Phillipson, J. J.; Wells, P. B.; Wilson, G. R. J. Chem. Soc. (A)
1969, 1351.
(26) Grant, J.; Moyes, R. B.; Oliver, R. G.; Wells, P. B. J. Catal. 1976,
42, 213.
(27) George, M.; Moyes, R. B.; Ramanarao, D.; Wells, P. B. J. Catal.
that nucleate on these different graphite edges will be dictated
to a large degree by the interaction with the atoms in nanofiber
supports. Under such circumstances one might reasonably
expect that different crystallographic faces of nickel will be
exposed to the reactant gas depending on which type of
nanofiber structure is used as the supporting medium. The
ramifications of this remarkable example of a metal-support
interaction is that it opens up the possibility of a tailoring the
morphological characteristics of metal particles in such a fashion
so to achieve a desired catalytic performance.
1978, 52, 486.
(28) Rodriguez, N. M.; Chambers, A.; Baker, R. T. K. Langmuir 1995,
11, 3862.
(29) Schlatter, J. C.; Boudart, M. J. Catal. 1972, 24, 482.
(30) Baker, R. T. K.; Sherwood, R. D. J. Catal. 1981, 70, 198.
(31) Baker, R. T. K.; Sherwood, R. D.; Derouane, E. G. J. Catal. 1982,
75, 382.
(32) Simoens, A. J.; Derouane, E. G.; Baker, R. T. K. J. Catal. 1982,
75, 184.
Acknowledgment. This work was supported by a grant from
the National Science Foundation, Division of Chemical and
Transport Systems, No CTS-9634266. The authors would like
to thank Prof. N. M. Rodriguez for fruitful discussions.
(33) Baker, R. T. K. J. Adhesion 1995, 52, 13.
(34) Abrahamson, J. Carbon 1973, 11, 337.
(35) Aria, M.; Nishiyama, Y. J. Catal. 1985, 104, 175.
(36) Boudart, M. AdV. Catal. 1969, 20, 153.
(37) Masson, A.; Bellamy, B.; Colomer, G.; M’Bedi, M.; Rabette, P.;
Che, M., Proc. 8th Int. Congr. Catal. 1984, 4, 333.
(38) Hadj Romdhane, Y.; Bellamy, B.; Degouveia, V.; Masson, A.; Che,
M. Appl. Surf. Sci. 1988, 31, 383.
(39) Masson, A.; Bellamy, B.; Hadj Romdhane, Y.; Che, M.; Dufour,
G. Surf. Sci. 1986, 173, 479.
(40) Nosova, L. V.; Stenin, M. V.; Nogin, Y. N.; Ryndin, Y. A. Appl.
Surf. Sci. 1992, 55, 43.
(41) Deganello, G.; Duca, D.; Martorana, A.; Fagherazzi, G.; Benedetti,
References and Notes
(1) Kroto, H. W.; Heath, J. R.; O’Brien, S. C.; Curl, R. F.; Smalley,
R. E. Nature 1985, 318, 162.
(2) Iijima, S. Nature 1991, 354, 56.
(3) Baker, R. T. K. Carbon 1989, 27, 315.
(4) Baker, R. T. K. Catal. ReV.sSci. Eng. 1979, 19, 161.
(5) Vijayakrishnan, V.; Santra, A. K.; Seshadri, R.; Nagarajan, R.;
Pradeep, T.; Rao, C. N. R. Surf. Sci. Lett. 1992, 262, L87.
(6) Braun, Th.; Wohlers, M.; Belz, T.; Nowitzke, G.; Wortmann, G.;
Uchida, Y.; Pfander, N.; Schlogl, R. Catal. Lett. 1997, 43, 167.
(7) Braun, Th.; Wohlers, M.; Belz, T.; Schlogl, R. Catal. Lett. 1997,
43, 175.
(8) Planeix, J. M.; Coustel, N.; Coq, B.; Brotons, V.; Kumbhar, P. S.;
Dutartre, R.; Geneste, P.; Bernier, P.; Ajayan, P. M. J. Am. Chem. Soc.
1994, 116, 7935.
(9) Rodriguez, N. M. J. Mater. Res. 1993, 8, 3233.
(10) Baker, R. T. K. In Carbon Fibers, Filaments and Composites;
Figueiredo, J. L., Bernado, C. A., Baker, R. T. K., Huttinger, K. J.,Eds.;
Kluwer Academic Publishers: Netherlands, 1990; p 405.
(11) Baker, R. T. K.; Terry, S.; Harris, P. S. Nature 1975, 253, 37.
(12) Baker, R. T. K.; Harris, P. S. In Chemistry and Physics of Carbon;
Walker, P. L., Jr., Thrower, P. A., Eds.; Marcel Dekker: New York, 1978;
Vol. 14, p 83.
(13) Kim, M. S.; Rodriguez, N. M.; Baker, R. T. K. J. Catal. 1990,
131, 60.
(14) Kim, M. S.; Rodriguez, N. M.; Baker, R. T. K. J. Catal. 1992,
134, 253.
(15) Chambers, A.; Nemes, T.; Rodriguez, N. M.; Baker, R. T. K. J.
Phys. Chem. B 1998, 102, 2251.
(16) Cooper, B. J.; Trimm, D. L. J. Catal. 1980, 62, 35.
(17) Santiesteban, J.; Fuentes, S.; Yacaman, M. J. J. Vac. Sci. Technol.
A 1983, 1, 1198.
A. J. Catal. 1994, 150, 127.
(42) Boitiaux, J. P.; Cosyns, J.; Vasudevan, S. Appl. Catal. 1983, 6, 4.
(43) Boitiaux, J. P.; Cosyns, J.; Vasudevan, S. In Preparation of
Catalysts III; Poncelot, C., Grange, P., Jacobs, P. A., Eds.; Elseiver:
Amsterdam, 1983; p 123.
(44) Hub, S.; Hilaire, L.; Touroude, R. Appl. Catal. 1988, 36, 307.
(45) Tardy, B.; Noupa, C.; Leclerq, C.; Bertolini, J. C.; Hoareau, A.;
Treilleux, M.; Faure, J. P.; Nihoul, G. J. Catal. 1991, 129, 1.
(46) Otero Schipper, P. H.; Wachter, W. A.; Butt, J. B.; Burwell, R. L.;
Cohen, J. B. J. Catal. 1977, 50, 494.
(47) Massardier, J.; Bertolini, J. C.; Renouprez, A. Proc. 9th Int. Congr.
Catal. 1988, 1222.
(48) Dalmai-Imelik, G.; Leclercq, C.; Maubert-Maubert, A. J. Solid State
Chem. 1976, 16, 129.
(49) Dalmai-Imelik, G.; Massardier, J. Proc. Int. Congr. Catal. 1976,
90.
(50) Gallezot, P.; Richard, D.; Bergeret, G. NoVel Materials in Hetero-
geneous Catalysis; Baker, R. T. K., Murrell, L. L., Eds.; ACS Symp. Ser.
1990, 437, 150.
(51) Goetz, J.; Volpe, M. A.; Touroude, R. J. Catal. 1996, 164, 369.
(52) Van Hove, M. A.; Koestner, R. J.; Stair, P. C.; Biberian, J. P.;
Kesmodel, L. I.; Bartos, J.; Somorjai, G. A. Surf. Sci. 1991, 103, 218.
(53) Somorjai, G. A.; Van Hove, M. A. Catal. Lett. 1988, 1, 433.
(54) Muller, K. Prog. Surf. Sci. 1993, 42, 245.
(55) Anger, G.; Berger, H. F.; Luger, M.; Feistritzer, S.; Winkler, A.;
Renduloc, K. O. Surf. Sci. 1989, 219, L583.
(18) Poppa, H.; Rumpf, F.; Moorhead, R. D.; Henry, C. In Microstruc-
ture and Properties of Catalysts; Treacy, M. M. J., Thomas, J. M.,White,
J. M., Eds.; MRS Symp. Ser. 1987, 111, 1.
(19) Yang, R. T.; Chen, J. P. J. Catal. 1989, 115, 52.
(20) Rodriguez, N. M.; Kim, M. S.; Baker, R. T. K. J. Phys. Chem.
1994, 98, 13108.