Catalytic behavior of graphite nanofiber supported nickel particles. 3. The effect of chemical blocking on the performance of the system

Article abstract of DOI:10.1021/jp983802d

Graphite nanofibers are a newly developed type of material produced by the catalytic decomposition of carbon containing gases at high temperatures. The individual components of these conformations, small-sized graphite crystallites, are arranged in such a manner that only the edge regions are exposed. The carbon atoms at these sites that are arranged in two conformations, `armchair' or `zigzag', act as templates for the nucleation of metal crystallites. Treatment of graphite with certain phosphorus compounds is a process that is known to result in preferential blocking of the `armchair' faces, whereas boron oxide selectively substitutes into the `zigzag' faces. In the current investigation pretreatment in phosphorus oxide was found to exert little or no effect on the subsequent catalytic performance of graphite nanofiber supported nickel with respect to hydrogenation of ethylene and 1-butene. In contrast, incorporation of boron into the carbonaceous support, which resulted in blockage of the `zigzag' sites of the graphite nanofibers rendered the supported metal system virtually inactive toward hydrogenation of either of the olefins. These results suggest that the active state of nickel is one where the particles are preferentially located on the `zigzag' faces of the nanofiber structures. Under these conditions the metal particles adopt a crystallographic arrangement that is favorable toward reaction with both reactant molecules. It is evident that one can control the catalytic behavior of a given metal by careful tailoring the support structure at the atomic level.

Full text of DOI:10.1021/jp983802d

J. Phys. Chem. B 1999, 103, 2453-2459
2453
Catalytic Behavior of Graphite Nanofiber Supported Nickel Particles. 3. The Effect of
Chemical Blocking on the Performance of the System
Colin Park and R. Terry K. Baker*
Department of Chemistry, Northeastern UniVersity, Boston, Massachusetts 02115
ReceiVed: September 21, 1998; In Final Form: December 31, 1998
Graphite nanofibers are a newly developed type of material produced by the catalytic decomposition of carbon
containing gases at high temperatures. The individual components of these conformations, small-sized graphite
crystallites, are arranged in such a manner that only the edge regions are exposed. The carbon atoms at these
sites that are arranged in two conformations, “armchair” or “zigzag”, act as templates for the nucleation of
metal crystallites. Treatment of graphite with certain phosphorus compounds is a process that is known to
result in preferential blocking of the “armchair” faces, whereas boron oxide selectively substitutes into the
“zigzag” faces. In the current investigation pretreatment in phosphorus oxide was found to exert little or no
effect on the subsequent catalytic performance of graphite nanofiber supported nickel with respect to
hydrogenation of ethylene and 1-butene. In contrast, incorporation of boron into the carbonaceous support,
which resulted in blockage of the “zigzag” sites of the graphite nanofibers rendered the supported metal
system virtually inactive toward hydrogenation of either of the olefins. These results suggest that the active
state of nickel is one where the particles are preferentially located on the “zigzag” faces of the nanofiber
structures. Under these conditions the metal particles adopt a crystallographic arrangement that is favorable
toward reaction with both reactant molecules. It is evident that one can control the catalytic behavior of a
given metal by careful tailoring the support structure at the atomic level.
Introduction
hydrogenated than mono-olefins. In a reactant mixture contain-
ing both sets of compounds, however, it was found that the
The use of additives in catalysis to selectively promote or
inhibit specific reactions in hydrocarbon decomposition reactions
is widely practiced.1-3 In this context, the most commonly
studied nonmetal additives are sulfur and chlorine.4-13 The
addition of sulfur in large amounts poisons a metal catalyst;
however, trace amounts of the element can enhance the
selectivity to aromatic products and decrease the hydrogenolysis
activity of a reforming catalyst.4-8 Chlorine is a well-known
promoter of isomerization reactions and is frequently added to
catalysts to enhance the acidic nature of a support material such
as γ-Al2O3. The acidity of the catalyst has been used to control
the relative rates of hydrocracking and isomerization reactions
and also inhibit the deposition of carbon.9,10 The use of additives
must, however, be carried out under the correct conditions, with
the appropriate loading in order to achieve the desired effect.
di-olefins were so strongly adsorbed on the catalyst surface that
inhibition of the hydrogenation of the mono-olefins occurred.
The use of Ni-P catalysts has also been found to exhibit a high
activity for the reduction of nitrobenzene to aniline.21
Graphite nanofibers are produced by the thermal decomposi-
tion of carbon-containing gases over selected metal catalysts.
These solids consist of extremely small graphite crystallites
stacked in various directions with respect to the fiber axis, with
only edges exposed. It is this feature that makes the material
different from any other form of carbon because, since there is
no basal plane exposed, interaction of the solid with other phases
occurs solely at these edges. Research in the area of carbon
nanofibers that possess unique tailored structures has gained
momentum in the last 10 years as potential applications for these
novel materials are identified.24-30 In previous studies, both
nanofibers and nanotubes have been utilized as novel catalyst
support media with exciting results.30-34 Park and co-workers33
found that the structural characteristics adopted by particular
types of graphite nanofibers exerted a significant impact on the
catalytic behavior of supported nickel particles. This difference
in performance was attributed to the ability of the nickel to adopt
specific orientations following nucleation on either the “zigzag”
(101h0) or the “armchair” (112h0) faces of the graphitic nano-
fibers. In this respect it is important to consider the work of
Yang and Chen35 who reported that a given crystallographic
orientation of nickel would preferentially decompose carbon
monoxide while a different set of faces favored the precipitation
of dissolved carbon in the form of graphite.
The use of phosphorus as an additive to various metals has
been used in several catalytic studies involving hydrode-
sulfurization reactions.14-17 Eijsbouts and co-workers17 reported
that the hydrodesulfurization activity of a carbon-supported
nickel catalyst was enhanced, whereas the performance of other
transition metal catalysts for this reaction were inhibited by the
presence of phosphorus. The effect of introducing phosphorus
into supported metal particles for various hydrogenation reac-
tions has been examined by a number of research groups.18-23
Ko and Chou18 demonstrated that P-Ni/Al2O3 catalysts were
highly selective for the hydrogenation of R-pinene to cis-pinene
when compared to the performance of a Ni/Al2O3 catalyst. In
another investigation, Yoshida and co-workers,19,20 using pre-
treated amorphous Ni-P and Ni-B alloys, claimed that these
catalysts were highly active for the hydrogenation of olefins.
These workers also concluded that di-olefins were less readily
A similar pattern of behavior was found by Cooper and
Trimm36 during a study of the decomposition of propylene over
10.1021/jp983802d CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/12/1999

2454 J. Phys. Chem. B, Vol. 103, No. 13, 1999
Park and Baker
The graphitic nanofibers were subsequently impregnated with
an aqueous solution of either methylphosphonic acid or am-
monium pentaborate to realize a 5 wt % additive loading. The
aqueous solution was added slowly to the carbonaceous solid
with constant stirring at 90 °C until the consistency attained
that of a thick paste. This doped nanofiber mixture was dried
overnight in an oven at 110 °C. The treated nanofibers were
then subjected to an oxidation treatment in a 50% air/helium
mixture where the temperature of the sample was progressively
raised to 850 °C at a rate of 5 °C/min and held at this level for
3 h before being cooled to ambient conditions. This oxidation
treatment of the doped nanofiber precursors creates the condi-
tions necessary to chemically bond phosphorus species onto the
“armchair” faces of the graphite nanofibers and create the
conditions necessary for boron to spread and interact with the
“zigzag” regions. Phosphorus oxides not bound to the graphite
face would undergo decomposition at much lower temperatures
(ca. 350 °C). Prior to the introduction of nickel onto the
nanofiber support, any unbound phosphorus and boron oxides
were eliminated by thorough washing in dilute acid solution.
Figure 1. Schematic representation showing the proposed structural
arrangements involved in the bonding of (a) phosphorus and (b) boron
oxide species to the respective “armchair” and “zigzag” faces of graphite
nanofibers. Functionality has been omitted from many of the layers
for the sake of clarity.
iron catalysts possessing different crystallographic orientations.
These workers found that only certain crystal faces were capable
of decomposing propylene, while others exhibited a limited
activity toward the reaction. It is reasonable to assume therefore,
that if one could impregnate a metal in a desired orientation
onto a support material it would be possible to tailor both the
activity and selectivity of the system to a desired level.
In this regard, it is well established that phosphorus species
will preferentially bond to the graphite “armchair” (112h0) faces
while boron species tend to favor attachment to the “zigzag”
(101h0) faces.37-40 A schematic rendition of the manner by which
these additives bond to the respective faces of the graphite
structure is shown in Figure 1a,b. It should be emphasized that
these observations have provided the rudiments for the develop-
ment of surface treatments designed to protect graphite against
oxidation at high temperatures.37
The 5 wt % Ni graphite nanofiber supported catalysts used
in this study were prepared by a standard incipient wetness
technique, as described in detail in a previous paper.33 The
catalyst precursor was initially calcined in air at 250 °C for 4
h to convert the metal nitrate to the thermally stable oxide state.
The calcined catalyst was then flushed in He at ambient
temperature before being reduced at 350 °C in a 10% H2/He
mixture for 20 h. The reduced catalyst was cooled in He,
passivated at ambient temperature in a 2% O2/He mixture for 1
h before the catalyst was removed from the reactor and finally
stored in a sealed container for future use.
The gases used in this work, helium (99.999%), hydrogen
(99.999%), carbon monoxide (99.99%), 1-butene (99.95%), and
ethylene (99.95%) were obtained from Medical Technical Gases
and used without any further purification. The dopants, methyl-
phosphonic acid (98%) and ammonium pentaborate (99.99%),
were obtained from Alpha Products and Aldrich Chemical Co.,
respectively; reagent grade nickel nitrate [Ni(NO3)2‚6H2O] was
purchased from Fisher Scientific for the catalyst preparation.
In the current investigation, we have endeavored to exploit
this concept by depositing nickel onto graphite nanofibers that
have been pretreated with either phosphorus- or boron-contain-
ing compounds. From such studies it should be feasible to
discriminate between the behavior of nickel on the pristine and
modified nanofiber surfaces. One might argue that this procedure
would merely result in a moderation in the catalytic behavior
of the metal crystallites located on the capped regions of the
surface. Previous studies, however, have indicated that when a
metal is introduced onto such treated graphite surfaces then
complete suppression of catalytic activity of the crystallites that
form on the phosphorus- and boron-substituted regions of the
substrate occurs, whereas those that accumulate on the unpro-
tected faces exhibited normal behavior.38,39
Apparatus and Procedures. The catalysis unit used through-
out this study consisted of a vertical quartz flow reactor, fitted
with a quartz frit located in the central region of the reactor
tube, which was heated with 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. Catalyst samples (100
mg) were placed on the quartz frit and the tube was positioned
in such a manner that the frit was always in approximately the
same point in the reactor. After reduction in a 10% H2/He
mixture for 2 h at 350 °C the system was cooled to the desired
reaction temperature while maintaining this gas flow. Once the
desired reaction temperature was attained, the reactant hydro-
carbon gas, or a predetermined hydrocarbon/H2/He mixture, was
introduced to the reduced catalyst sample at a flow rate of 100
sccm and the reaction was allowed to proceed for periods of
up to 3 h. The product distribution was followed as a function
of time by sampling the inlet and outlet gas streams at regular
intervals. The concentrations of reactants and products were
analyzed by gas chromatography using a 30 m megabore (GS-
Q) capillary column for reactions involving ethylene and a 30
m megabore (GS-AL) capillary column for reactions involving
1-butene in a Varian 3400 GC unit.
Experimental Section
Materials. The graphite nanofibers chosen for this present
study were the “platelet” variety, in which the graphite platelets
constituting the material were aligned in a direction perpen-
dicular to the fiber axis. In this particular arrangement, the
nanofibers expose an equivalent number of “armchair” and
“zigzag” faces to the gas phase. These structures were grown
according to the protocol described in a previous paper41 by
passing a desired mixture of ethylene, carbon monoxide, and
hydrogen over iron powder at 600 °C for periods of up to 3 h.
The resultant carbon nanofibers were demineralized in dilute
HNO3 acid for a period of 7 days to remove any of the original
metal catalyst particles. The effectiveness of this step was
checked by X-ray diffraction analysis, which showed a complete
absence of any metallic components at this stage. The nanofibers
were then thoroughly washed in deionized water prior to being
dried overnight in air at 110 °C and stored until required for
use.
The nature and characteristics of all the graphite nanofiber
supported nickel catalysts were established using a combination

Catalytic Behavior of Nickel Particles
J. Phys. Chem. B, Vol. 103, No. 13, 1999 2455
Figure 2. Comparison of the gasification characteristics of doped and untreated “platelet” graphite nanofibers in air.
of techniques, including high-resolution transmission electron
microscopy (HRTEM), electron diffraction, and temperature-
programmed oxidation (TPO) studies. The nickel catalysts were
studied before and after use by HRTEM to establish any changes
in the support structure and in the morphological characteristics
of the impregnated metal particles. These examinations were
carried out in a JEOL 2000EX II transmission electron
microscope fitted with a high-resolution pole piece capable of
giving a lattice resolution of 0.18 nm. Suitable transmission
specimens were prepared by the ultrasonic dispersion of catalyst
samples in isobutanol and application of a drop of the resultant
suspension to a holey carbon support grid. Using this approach,
it was possible to locate sections of the catalyst that protruded
over the edge of the carbon, thereby permitting an examination
of the sample without interference from the substrate. Images
of very many different fields were displayed on the TV monitor,
and the high-magnification images were transferred to a
Mitsubishi printer unit for subsequent analysis. The size
distribution plots were constructed from the measurements of
over 500 particles in both the pristine and phosphorus-doped
systems. In the case of the boron-doped material, far fewer metal
particles were present and in this system the average size was
estimated from about 150 particles. When operated in the
diffraction mode, it was possible to determine the chemical state
of the catalyst before and after reaction in the respective olefin/
hydrogen environments.
Figure 3. Transmission electron micrograph showing the typical
appearance of nickel particles supported on “platelet” GNF following
reaction in 1-butene/hydrogen (1:1) mixture at 125 °C.
2. From these three profiles it can be seen that the introduction
of phosphorus or boron onto the support surfaces exerts an
appreciable retarding effect on the gasification of this type of
nanofibers. Untreated graphitic nanofibers start to gasify under
these conditions at 725 °C while the corresponding doped
materials do not exhibit any significant weight loss until the
temperature is raised to 940 °C. One may therefore assume that
the objective of selectively blocking of the “armchair” faces of
these nanofibers by phosphorus and the “zigzag” faces by boron
has been successfully achieved in this study.
1.2. Transmission Electron Microscopy Examination. A
representative example of the appearance of nickel particles on
the untreated platelet type of graphite nanofiber (GNF) support
following reaction in a 1-butene/hydrogen (1:1) mixture at 125
°C is presented in Figure 3. High-resolution electron microscopy
examination of these specimens has enabled one to determine
the location and morphology of the metal particles on these
materials. Nickel particles were found to adopt a well-defined
faceted shape, as shown in Figure 4. These morphological
Temperature-programmed oxidation studies were carried out
using a Cahn 2000 microbalance on demineralized and both
doped graphite nanofiber samples in an air/Ar (1:1) mixture at
a constant heating rate of 5 deg/min, as outlined in a previous
paper from this laboratory.11 This method was used to establish
the effectiveness of the doping procedures by comparing the
oxidation profiles of the phosphorus- and boron-treated nano-
fibers against that of an ostensibly clean demineralized sample.
Results
1. Catalyst Characterization Studies. 1.1. Temperature-
Programmed Oxidation. A comparison of the reactivity of
phosphorus- and boron-doped platelet graphite nanofibers with
that of the pristine material toward air is presented in Figure

2456 J. Phys. Chem. B, Vol. 103, No. 13, 1999
Park and Baker
Electron diffraction analysis was carried out on phosphorus-
and boron-doped platelet GNF samples that had been impreg-
nated with 5 wt % nickel. These experiments showed that nickel
particles were in the metallic state in both systems, and no
evidence was found for the existence of either Ni-P or Ni-B
species in either the fresh or reacted catalyst samples.
2. Flow Reactor Studies. 2.1. Catalytic Hydrogenation of
Ethylene by Graphite Nanofiber Supported Nickel Particles. In
this set of experiments both untreated 5 wt % nickel/platelet
GNF and the analogous phosphorus- and boron-doped catalyst
systems were tested over a period of 3.0 h at 120 °C to compare
their ability to hydrogenate a simple probe molecule, ethylene.
The results of these experiments are presented in Table 2, from
which it can be seen that almost identical activities and
selectivities were obtained with the former two catalyst systems,
whereas the 5 wt % Ni/B-platelet GNF system was virtually
inactive for this reaction. This overall pattern of behavior was
maintained when the reaction was carried out at the slightly
higher temperature of 140 °C over the same time period.
2.2. Catalytic Hydrogenation of 1-Butene by Graphite
Nanofiber Supported Nickel Particles. When the degree of
complexity of the reactant probe molecule was increased,
significant differences were observed in the reactivity patterns
of the three catalyst systems. Table 3 shows the product
distributions obtained when the pristine and treated catalysts
were treated in a 1-butene/hydrogen (1:1) mixture at 125 °C
for 3.0 h. Once again, the Ni supported on boron-doped GNF
samples exhibited an extremely low activity, which did not
improve when the temperature was changed. Examination of
the data from the other catalyst systems indicates that in both
cases the major product is n-butane, which is formed from the
complete hydrogenation of the reactant molecule. Under these
conditions, however, the phosphorus-doped catalyst exhibited
a significantly higher activity and a modification in the
selectivity pattern that is reflected in lower yields of the
isomerization products, cis- and trans-2-butenes.
Figure 4. High-resolution transmission electron micrograph showing
the details of the nickel particle morphology when dispersed on
“platelet” GNF.
characteristics are consistent with the existence of a relatively
strong interaction between the metal particles and the support
medium, whereas, in the case of a weak metal-support
interaction, the metal particles would tend to acquire a more
dense globular geometry42 and readily undergo sintering with
a concomitant loss of catalytic activity.
Particle size distribution measurements of the fresh and
reacted Ni/platelet GNF and the corresponding samples of Ni/
phosphorus-doped GNF are displayed in Figures 5 and 6,
respectively. These plots are based on the measurements of over
500 particles in each sample. Inspection of the two sets of data
reveal the existence of major differences in the distribution
profiles of the untreated and phosphorus-doped GNF supported
metal catalysts. While the profiles of the two fresh catalysts
are very similar, after reaction in a 1-butene/hydrogen (1:1)
mixture at 125 °C, a significantly narrower size range is attained
when nickel was impregnated on the phosphorus-treated nano-
fiber surface, indicating that particle sintering is impeded to an
appreciable extent under these circumstances. Table 1 highlights
the differences in average particle size when one edge of the
graphite nanofiber has been excluded from the impregnation
process and the effect of a 3.0 h treatment in a 1-butene/
hydrogen reactant mixture at 125 °C.
In a final series of experiments the catalytic activity of the
pristine, phosphorus- and boron-doped GNF materials for the
hydrogenation of both olefins was investigated over the tem-
perature range 100 to 140 °C. No evidence for decomposition
of the respective reactants was found, indicating that in the
absence of nickel, none of these supports was exerting any
catalytic action on the hydrocarbon conversion reactions.
Discussion
Examination of Ni/boron-doped GNF revealed the existence
of a very sparse collection of metal particles. It was evident
that the few particles that were present adopted a globular
morphology and exhibited a nonwetting behavior with respect
to the interaction with the support. These are characteristics
generally associated with the existence of a weak metal-support
interaction. Since boron substitutes into the “zigzag” edges of
graphite it is highly likely that these regions of the nanofibers
will be masked and unavailable for the nucleation and growth
of nickel crystallites following the metal nitrate impregnation
step. Under such circumstances the only available unadulterated
locations of the nanofibers will be the “armchair” faces. In this
configuration the interfacial properties between the components
may not favor the establishment of a strong metal-support
interaction, and as a consequence, the growth characteristics of
nickel crystallites on such surfaces will be quite different from
those prevailing on the “zigzag” faces. On the other hand, in
the case of the phosphorus-doped system, the location for the
collection of metal species will be restricted to the “zigzag”
faces of the graphite nanofiber support, where conditions exist
for the creation of a strong metal-support interaction.
In the current study, the impregnation of a platelet type of
graphitic nanofiber with either phosphorus or boron species has
been shown by TPO studies to offer a certain degree of
protection of the material toward gasification by oxygen. These
results are consistent with those obtained by Oh and Rodriguez,38
who used a variety of techniques, including controlled atmo-
sphere electron microscopy, to directly observe the effect of
phosphorus species on the graphite-oxygen reaction and
concluded that the additive preferentially bonded to the “arm-
chair” faces, leaving the “zigzag” regions vulnerable to oxida-
tion. When these specimens were simultaneously examined by
in-situ electron diffraction techniques, the d spacings were
attributed to the establishment of a chemical bond between
phosphorus oxide species and the coated carbon edge atoms.
Others workers43-45 have reported on the ability of phosphorus
to act as an inhibitor for graphite oxidation and suggested that
physical blockage of edge sites was responsible for the action
of the additive.
If one considers the action of the phosphorus dopant, which
effectively blocks the access to the “armchair” faces of the

Catalytic Behavior of Nickel Particles
J. Phys. Chem. B, Vol. 103, No. 13, 1999 2457
Figure 5. Particle size distribution profiles of metal particles supported on both fresh and used 5 wt % Ni/platelet GNF catalysts.
Figure 6. Particle size distribution profiles of Ni particles supported on both fresh and used 5 wt % Ni/phosphorus treated-platelet GNF catalysts.
TABLE 1: Comparison of the Average Nickel Particle Size
Supported on Pristine and Phosphorus- and Boron-Treated
GNF before and after Reaction in 1-Butene/Hydrogen (1:1)
at 125 °C for 3.0 h
TABLE 2: Selectivity of Graphite Nanofiber Supported 5 wt
% Nickel Catalysts for a C2H4/H2 (1:1) Reactant Mixture at
Various Temperatures
% conversion of
av particle size
ethylene to selected products
5 wt % Ni/GNF
platelet GNF
platelet GNF(used)
phosphorus-platelet GNF
phosphorus-platelet GNF (used)
boron-platelet GNF
(nm)
solid % con-
methane ethane carbon version
7.92
13.61
7.51
10.23
29.82
36.70
catalyst support
“platelet” GNF (120 °C)
0.03
0.03
0.00
0.02
0.01
0.00
51.4
8.50 59.6
61.8
0.01 1.41
P-doped “platelet” GNF (120 °C)
B-doped “platelet” GNF (120 °C)
“platelet” GNF (140 °C)
P-doped “platelet” GNF (140 °C)
B-doped “platelet” GNF (140 °C)
51.0 10.1
1.2
45.9
46.1
1.0
7.0
9.0
51.8
54.2
boron-platelet GNF (used)
0.01
1.22
nanofiber support, then it may be assumed that under these
circumstances only nickel particles that collect on the “zigzag”
edges will be active in the olefin hydrogenation reactions. It
has been previously demonstrated that in a reducing environ-
ment, nickel preferentially wets and spreads along the “zigzag”
faces of single-crystal graphite, leading to an epitaxial relation-
ship metal and the substrate medium.46-48 These results were
consistent with the theoretical studies of the surface energetics
of the nickel/graphite system presented by Abrahamson,49 who
concluded that metal particles could adopt different morphol-
ogies on the two prismatic faces.
Since the “zigzag” edges are the preferred location for nickel
particles on a graphite substrate, one would not expect that the
presence of phosphorus on the adjacent edge sites would
interfere with the catalytic activity of the system to any
appreciable degree. On the other hand, since boron selectively
interacts with the “zigzag” faces of graphite, access to these
regions will be blocked and nickel will be forced to accumulate
on the “armchair” faces. Under the conditions used in the current
experiments, it is not entirely unexpected that any beneficial
effects afforded by the interaction of certain edge sites with

2458 J. Phys. Chem. B, Vol. 103, No. 13, 1999
Park and Baker
TABLE 3: Selectivity of Graphite Nanofiber Supported 5 wt
% Nickel Catalysts for a 1-C4H8/H2 (1:1) Reactant Mixture
at 125 °C
crystallographic arrangement adopted by the metal particles,
which will ultimately be manifested in a different catalytic
reactivity pattern. The results of the present investigation clearly
demonstrate that the catalytic behavior of a given metal can be
manipulated by careful tailoring of the support structure at the
atomic level.
% conversion of
1-butene to selected products
trans-
cis-
%
catalyst support
“platelet” GNF
P-doped “platelet” GNF
B-doped “platelet” GNF
n-butane 2-butene 2-butene conversion
Conclusions
39.2
79.6
2.4
12.6
6.2
0.6
14.5
5.5
0.8
67.8
92.0
3.8
The use of phosphorus to selectively block the “armchair”
faces of graphite nanofibers has been shown to exert little or
no effect on the catalytic behavior of nickel with respect to
hydrogenation of ethylene and 1-butene but did appear to result
in a narrower particle size distribution following reaction. In
contrast, addition of boron onto the carbonaceous support
resulted in blocking of the “zigzag” faces, a step that appeared
to subsequently render the supported metal system virtually
inactive toward hydrogenation of either of the olefins. Based
on these findings it is concluded that the active state of nickel
is one where the particles are preferentially located on the
“zigzag” faces of the graphite nanofibers. Under these conditions
the metal particles adopt a crystallographic arrangement that is
conducive toward reaction with both olefin molecules.
Transmission electron microscopy examination of the various
nanofiber supported metal catalyst samples provides a picture
that is consistent with these arguments. Nickel particles sup-
ported on pristine and phosphorus-treated platelet GNF exhibited
a similar appearance, being relatively thin, flat, and hexagonal
shaped, morphological characteristics that are associated with
the existence of a strong metal-support interaction. On the other
hand, those present on the corresponding boron-treated nano-
fibers were relatively large and globular in outline, features that
are indicative of a weak metal-support interaction.
metal crystallites will be eliminated in the presence of this
additive. These types of studies are being extended to other
metals, such as copper, palladium, and platinum, to ascertain
the preferential wetting behavior of these respective metals on
the prismatic faces of graphite. This information will then enable
us to develop more profound arguments with regard to the
controlling factors on the metal particle morphology.
One might argue that impregnation of either a phosphorus-
or boron-treated nanofiber material with nickel would lead to
the formation of the respective Ni-P or Ni-B compounds.
Electron diffraction analysis failed to reveal the presence of such
compounds either before or after reaction. It should be stressed,
however, that this technique is limited to the determination of
the bulk state and does not allow one to ascertain the chemical
characteristics of the surfaces of the metal particles. Studies
using a nickel phosphide as a catalyst have demonstrated that
significant changes in the catalytic activity could be induced
when compared to a similar untreated nickel catalyst.18-23
In the current investigation, the catalytic hydrogenation of
ethylene and 1-butene over both Ni/platelet GNF and the
corresponding Ni/phosphorus-platelet GNF system were identi-
cal above 120 °C. This result would therefore tend to rule out
the possibility of such a Ni-P species being the active catalyst
in this study. In contrast, the use of a 5 wt % Ni/B-platelet
GNF catalyst brought about significant changes in the catalytic
behavior of the metal with respect to the hydrogenation of both
ethylene and 1-butene. At all temperatures investigated in this
study the boron-doped catalyst exhibited an extremely low
activity toward the hydrogenation of both olefins. One possible
explanation for this behavior is that the presence of boron in
the catalyst acts as a poison to nickel, thereby inhibiting the
hydrogenation of olefins. This explanation is somewhat tenuous
since nickel-boron has been shown to exert a high activity for
the hydrogenation of olefins. Indeed, the presence of a partially
oxidized boron species on the catalyst surface was postulated
to be responsible for the observed high selectivity and activity
of the catalyst system.19,20,23
A more plausible rationale may lie in the notion that boron
forms a strong interaction with the “zigzag” edges and ef-
fectively blocks the interaction of nickel with these sites. As a
consequence, any beneficial metal-support interaction will not
be realized and nickel will only exist in a weakly bound state
to the “armchair” faces. Under these circumstances, the potential
decrease in catalytic activity due to sintering of nickel crystallites
will be greatly enhanced. Transmission electron microscopic
examination of these specimens showed that the metal particles
were relatively large and acquired a globular shape, features
that are consistent with the above arguments. This behavior will
result in a loss of nickel surface area and concomitant decrease
in the catalytic performance of the system.
Acknowledgment. The work was supported by a grant from
the National Science Foundation, Division of Chemical and
Transport Systems, No. CTS-9634266. We are indebted to Prof.
N. M. Rodriguez for many suggestions and stimulating discus-
sions.
References and Notes
(1) Pines, H. The Chemistry of Catalytic Hydrocarbon ConVersion;
Academic Press: New York, 1981.
(2) Butt, J. B.; Petersen, E. E. ActiVation, DeactiVation and Poisoning
of Catalysts; Academic Press: San Diego, 1988.
(3) Gates, B. C.; Katzer, J. R.; Schuit, G. C. Chemistry of Catalytic
Processess; McGraw-Hill: New York, 1979.
(4) Biswas, J.; Bickle, G. M.; Gray, P. G.; Do, D. D.; Barbier, J. Catal.
ReV. Sci. Eng. 1988, 30, 161.
(5) Apestiguia, C. R.; Barbier, J. J. Catal. 1982, 78, 352.
(6) Cosyns, J.; Verna, F. AdV. Catal. 1990, 37, 279.
(7) van Trimpont, P. A.; Marin, G. B.; Froment, C. F. Appl. Catal.
1985, 17, 161.
(8) Verderone, R. J.; Pieck, C. L.; Sad, M. R.; Parera, J. M. Appl. Catal.
1986, 21, 239.
(9) Bishara, A.; Murad, K. M.; Stanislaus, A.; Ismail, M.; Hussain, S.
S. Appl. Catal. 1988, 7, 337.
(10) Figoli, N. S.; Sad, M. R.; Beltramini, J. N.; Jablonski, E. L.; Parera,
J. M. Ind. Eng. Chem. Prod. Res. DeV. 1980, 19, 545.
(11) Chambers, A.; Baker, R. T. K. J. Phys. Chem. 1997, 101, 1621.
(12) Kim, M. S.; Rodriguez, N. M.; Baker, R. T. K. J. Catal. 1993,
143, 449.
(13) Owens, W. T.; Rodriguez, N. M.; Baker, R. T. K. Catal. Today
1994, 21, 3.
(14) Atanasova, P.; Halachev, V. Appl. Catal. 1989, 48, 295.
(15) Gishti, K.; Iannibello, A.; Marengo, S.; Morelli, G.; Tittarelli, P.
Appl. Catal. 1984, 12, 381.
(16) Okamoto, Y.; Gomi, T.; Mori, Y.; Imanaka, T.; Teranishi, S. React.
Kinet. Catal. Lett. 1983, 22, 417.
(17) Eijsbouts, S.; van Gestel, J. N. M.; de Beer, V. H. J. Appl. Catal.
1994, 119, 293.
It is apparent that graphite nanofibers represent a new type
of support media, where two crystal faces are available for
subsequent impregnation and nucleation of metal particles. Each
of the faces will exert a distinct effect on the eventual
(18) Ko, S. H.; Chou, T. C. Can. J. Chem. Eng. 1994, 72, 862.

Catalytic Behavior of Nickel Particles
J. Phys. Chem. B, Vol. 103, No. 13, 1999 2459
(19) Yoshida, S.; Yamashita, H.; Funabiki, T.; Yonezawa, T. J. Chem.
Soc., Chem. Commun. 1982, 964.
(20) Yoshida, S.; Yamashita, H.; Funabiki, T.; Yonezawa, T. J. Chem.
Soc., Faraday Trans. 1 1984, 80, 1435.
(21) Sweeny, N. P.; Rohrer, C. S.; Brown, O. W. J. Am. Chem. Soc.
1958, 80, 799.
(22) Nozaki, F., Adachi, R. J. Catal. 1975, 40, 166.
(23) Okamoto, Y.; Nitta, Y.; Imanaka, T.; Teranishi, S. J. Chem. Soc.,
Faraday Trans. 1 1979, 75, 2027.
(32) Chambers, A.; Nemes, T.; Rodriguez, N. M.; Baker, R. T. K. J.
Phys. Chem. 1998, 102, 2251.
(33) Park, C.; Baker, R. T. K. J. Phys. Chem. 1998, 102, 5168.
(34) Salman, F.; Park, C.; Baker, R. T. K. Submitted to Catal. Today.
(35) Yang, R. T.; Chen, J. P. J. Catal. 1989, 115, 52.
(36) Cooper, B. J.; Trimm, D. L. J. Catal. 1980, 62, 35.
(37) McKee, D. W.; Spiro, C. L.; Lamby, E. J. Carbon 1984, 22, 507.
(38) Oh, S. G.; Rodriguez, N. M. J. Mater. Res. 1993, 8, 2879.
(39) Rodriguez, N. M.; Baker, R. T. K. J. Mater. Res. 1993, 8, 1886.
(40) Allardice, D. J.; Walker, P. L., Jr. Carbon 1970, 8, 375.
(41) Rodriguez, N. M.; Kim, M. S.; Baker, R. T. K. J. Catal. 1993,
144, 93.
(24) Baker, R. T. K.; Terry, S.; Harris, P. S. Nature 1975, 253, 37.
(25) Kim, M. S.; Rodriguez, N. M.; Baker, R. T. K. J. Catal. 1990,
131, 60.
(26) Vijayakrishnan, V.; Santra, A. K.; Seshadri, R.; Nagarajan, R.;
Pradeep, T.; Rao, C. N. R. Surf. Sci. Lett. 1992, 262, L87.
(27) Rodriguez, N. M. J. Mater. Res. 1993, 8, 3233.
(28) Rodriguez, N. M.; Chambers, A.; Baker, R. T. K. Langmuir 1995,
11, 3862.
(29) Downs, W. B.; Baker, R. T. K. J. Mater. Res. 1995, 10, 625.
(30) Rodriguez, N. M.; Kim, M. S.; Baker, R. T. K. J. Phys. Chem.
1994, 98, 13108.
(42) Baker, R. T. K.; Prestridge, E. B.; Garten, R. L. J. Catal. 1979,
59, 293.
(43) Rakszawski, J. F.; Parker, W. E. Carbon 1964, 2, 53.
(44) McKee, D. W.; Spiro, C. L.; Lamby, E. J. Carbon 1984, 22, 285.
(45) McKee, D. W. Carbon 1987, 25, 551.
(46) Baker, R. T. K.; Sherwood, R. D. J. Catal. 1981, 70, 198.
(47) Baker, R. T. K.; Sherwood, R. D.; Derouane, E. G. J. Catal. 1982,
75, 382.
(31) 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.
(48) Simoens, A. J.; Derouane, E. G.; Baker, R. T. K. J. Catal. 1982,
75, 184.
(49) Abrahamson, J. Carbon 1973, 11, 337.