Synthesis, characterization, and catalytic use of acicular iron particles

Article abstract of DOI:10.1006/jcat.2002.3543

The synthesis of granular iron particles via hydrotriorganoborate reduction of FeCl₃ has been reported previously by several research groups. Notably, during the course of studying some of the reaction parameters involved in FeCl₃ reduction, we discovered the experimental conditions necessary to synthesize acicular iron particles, at low temperature and low pressure, in the absence of a constraining medium. Specifically, the effects of the reaction temperature, the stirring rate, and the FeCl₃ solution dropping rate on the size and morphology of the iron particles were examined. Under optimal conditions acicular iron particles were produced having diameters of ～300 nm, lengths on the order of tens to hundreds of micrometers, and surface areas of approximately 20 m²/g. Both the acicular and the granular iron particles were found to be active catalysts for the formation of carbon nanofibers from the reaction of carbon monoxide and hydrogen gases. The graphitic natures of the carbon nanofibers formed from either acicular or granular iron catalysts were characterized by HRTEM and XRD. Also, while both types of iron catalysts produced straight carbon nanofiber morphologies, the granular iron catalysts also formed carbon nanofibers with spiral gross morphologies. Finally, the diameters of the carbon nanofibers were comparable to the diameters of the acicular or granular iron catalysts.

Full text of DOI:10.1006/jcat.2002.3543

Journal of Catalysis 208, 150–157 (2002)
doi:10.1006/jcat.2002.3543, available online at http://www.idealibrary.com on
Synthesis, Characterization, and Catalytic Use of Acicular Iron Particles
∗
,1
∗
,1
Kenneth J. Takeuchi, Amy C. Marschilok, Carol A. Bessel,† and Norman R. Dollahon‡
∗
Department of Chemistry, State University of New York at Buffalo, Buffalo, New York 14260; and †Department of Chemistry,
and ‡Department of Biology, Villanova University, 800 Lancaster Ave., Villanova, Pennsylvania 19085
Received October 2, 2001; revised January 17, 2002; accepted January 17, 2002
Two extremes in the morphology of finely divided iron
The synthesis of granular iron particles via hydrotriorganoborate powders are the granular and the acicular forms. Granular
reduction of FeCl has been reported previously by several research particles are approximately spherical in shape with an as-
groups. Notably, during the course of studying some of the reaction
parameters involved in FeCl reduction, we discovered the experi-
3
pect ratio (length/width) near 1. Granular powders of iron,
as well as a number of other metals, bimetallics, and metal
alloys, have been generated from the hydrotriorganoborate
reduction of a metal salt solution. These reactions generally
proceed according to the following equation (written for
iron)(2, 17):
3
mental conditions necessary to synthesize acicular iron particles, at
low temperature and low pressure, in the absence of a constraining
medium. Specifically, the effects of the reaction temperature, the
3
stirring rate, and the FeCl solution dropping rate on the size and
morphology of the iron particles were examined. Under optimal
conditions acicular iron particles were produced having diameters
of ∼300 nm, lengths on the order of tens to hundreds of microme-
FeCl3 +3LiBEt3H → Fe(s) +3LiCl+3/2H2 +3BEt3. [1]
2
ters, and surface areas of approximately 20 m /g. Both the acicular
In this technique, a metal salt solution is gradually mixed
with the reductant to precipitate a finely divided, granu-
lar metallic powder under low pressures and temperatures.
Notably, acicular iron particles have not produced via the
and the granular iron particles were found to be active catalysts
for the formation of carbon nanofibers from the reaction of carbon
monoxide and hydrogen gases. The graphitic natures of the carbon
nanofibers formed from either acicular or granular iron catalysts
were characterized by HRTEM and XRD. Also, while both types of chemistry described in Eq. [1].
iron catalysts produced straight carbon nanofiber morphologies, the
Acicular or fibrous particles are rodlike in nature, with
granular iron catalysts also formed carbon nanofibers with spiral a length significantly greater than their cross-sectional
gross morphologies. Finally, the diameters of the carbon nanofibers
were comparable to the diameters of the acicular or granular iron
area (5). The synthesis of acicular iron particles generally
requires additional synthetic considerations, such as the use
catalysts.
ꢀc 2002 Elsevier Science (USA)
of acicular starting materials, high pressures, high temper-
atures, constraining media, or magnetic fields (18–24). For
examples of various synthetic strategies for the generation
of acicular iron particles, consider the following reports.
The use of acicular starting materials was investigated by
NaonoandalsobyBoennemann. Naonoandco-workersre-
Key Words: acicular iron particles; granular iron particles; carbon
nanofibers.
1
. INTRODUCTION
The preparation of submicrometer-sized metal particles ported the synthesis of acicular iron particles via reduction
has been the topic of much recent research in areas as of silica-coated acicular hematite particles in air to form
diverse as catalysis (1, 2), powder technology (3–5), nano- acicular iron products ∼110 nm in length, with a surface
2
electronics (6–9), and environmental chemistry (10). In area of 80 m /g and an aspect ratio of 10 (18). Boenne-
particular, finely divided metal particles containing iron are mann et al. suspended acicular α-FeOOH particles in or-
of interest due to their use as catalysts in the preparation ganic solvents and then chemically reduced the particles
of an important family of carbon solids, namely, carbon with a combination of hydrotriorganoborates and H2 (20–
◦
nanofibers (vide infra) (11–16). Since the composition 100 bar) at 80 C to yield acicular iron products ∼20 nm in
and the morphology of the metal catalysts can affect the diameter and 250 nm long (aspect ratio ca. 13) (17). The
structure of the carbon nanofibers, the study of novel forms synthesis of acicular iron particles in constraining media
of iron particles as catalysts may yield interesting carbon was demonstrated by Nikles and co-workers in several re-
products.
ports (19–23). First, lecithin-encapsulated ferromagnetic α-
iron particles were generated from a sodium borohydride
reduction of ferrous chloride solutions containing soybean
lecithin in the presence of an applied magnetic field (19–22).
This preparation yielded acicular iron products that were
1
To whom correspondence should be addressed. For KJT: Fax: (716)
6
7
45-6963. E-mail: takeuchi@nsm.buffalo.edu; For CAB: Fax: (610) 519-
167. E-mail: carol.bessel@villanova.edu.
150
0
ꢀ
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.

SYNTHESIS OF ACICULAR IRON PARTICLES
.25–5 micrometers in length with an aspect ratio of 6. In a 2. EXPERIMENTAL AND METHODS
151
0
separatereport, ellipsoid-shapedβ-FeOOHparticleswhich
had been prepared from the “forced hydrolysis” of ferric
chloride were suspended in a lamellar liquid crystal phase
and then reduced in a sodium borohydride solution. The
resulting acicular iron products had aspect ratios greater
than 5 and consisted of “chains of cubes,” 20–25 nm in size
2
.1. Materials
The syntheses and isolation of iron particles were car-
ried out under inert atmospheres of nitrogen or argon.
Iron trichloride (98%, assay) was purchased from Alfa
AESAR. Lithium triethylhydridoborate, 1.0 M in tetrahy-
drofuran (THF), was purchased from Aldrich Chemi-
cal Company. Ethanol was purchased from Pharmco and
distilled before use. Ultralow water tetrahydrofuran was
purchased from JT Baker and used as received. Argon
(
23). The use of magnetic fields for the formation of acicu-
lar iron particles was studied by Gedanken and co-workers
24). Acicularironparticles, 50nmlongand5nmwide, were
(
synthesized via the sonochemical irradiation of a Fe(CO)5
solution in decalin in the presence of a 7-kG external mag-
netic field. Iron products produced using similar reaction
conditions in the absence of a magnetic field were found
to be granular in nature, with a mean particle size of about
(
99.998%), helium (99.995%), nitrogen (99.998%), and
hydrogen (99.95%) were purchased from Irish Welding,
and ultrahigh purity (UHP) carbon monoxide was pur-
chased from Matheson Gas. All of the gases were used
as received, except the nitrogen was purified before use
2
5 nm.
As mentioned above, the formation of nanostructured
3
by passage through a 500-cm scrubbing column contain-
carbon materials from the interaction of metal catalyst
particles with carbon-containing gases remains a subject
of much interest, since carbon particles, fibers, and tubes
have potential uses in electron emission devices (25–27),
nanodevices (28, 29), hydrogen storage materials (30, 31),
and fuel cells (1), as well as in other applications (15,
ing a 1 : 1 Ridox oxygen scavenging agent (Fisher Scien-
tific Co.) and 13X molecular sieves (Davidson Chemi-
cals).
2
.2. Equipment
3
2). Rodriguez, Baker, and co-workers (11–13, 15) re-
ported the formation of carbon nanofibers of various struc-
Kontes constant addition dropping funnels were used to
tures by controlling the nature of the metal catalyst sur- achieve constant rates of addition of the FeCl solution to
3
face, the composition of the reactant gases, the reaction the reductant solution. VWR Series 400S magnetic stirrers
temperature, and the presence of gas-phase or solid ad- and VWR Spinplus stir bars were used to maintain con-
ditives. The interaction of an iron-based catalyst with a stant and controlled stirring rates throughout the course of
◦
4
: 1 CO/H2 mixture at 600 C resulted in a generation of the reaction. The reaction temperature was held constant
carbon nanofibers having a “platelet” morphology con- throughout the reaction using a Haake A80 constant tem-
sisting of graphite layers stacked parallel to the base of perature bath. A Thermolyne type F79300 tube furnace was
the metal catalyst and perpendicular to the fiber axis. used for preparation of carbon nanofiber samples.
The spacing between adjacent graphite layers compared
A Hitachi S-570 scanning electron microscope was used
well with that of single crystal graphite (0.335 nm). Gen- for scanning electron microscopy (SEM) analysis. An accel-
erally, it has been noted that the diameters of carbon erating voltage of 15 keV and a working distance of 25 mm
nanofibers are typically dictated by the size of the cata- were used. SEM samples were prepared on carbon adhesive
lyst particle, with metal powder catalysts producing 5- to tabs (Electron Microscopy Sciences). Scanning electron mi-
5
00-nm diameter carbon fibers (15).
crographs were taken with Polaroid Polar Pan 400 film.
We report here the specific reaction conditions necessary EDS was performed using a PGT IMIX-XD X-ray analysis
to produce either granular or acicular iron particles via the system equipped with PGT IMIX microanalysis software,
hydrotriorganoborate reduction of FeCl3. In particular, we Version 8. The detector was a PGT Omega X-ray detector
discovered the experimental conditions which produce aci- fitted with a barium light element window. The EDS sam-
cular iron particles with unusually large aspect ratios (>50). ples were also prepared on carbon adhesive tabs. A JEOL
Notably, our synthetic method affords a low-temperature, JEM 2010 electron microscope was used for transmission
low-pressure synthesis of acicular iron that does not require electron microscopy (TEM) and high resolution transmis-
either acicular starting materials or a constraining medium. sion electron microscopy (HRTEM). TEM and HRTEM
Also, using modifications to literature procedures (12), we samples were mounted onto 300-mesh copper grids with a
report here the use of our acicular and granular iron par- lacy SiO on Formvar support film (Ted Pella, Inc.). Surface
ticles as catalysts for the formation of carbon nanofibers areaanalyseswereperformedwithanautomaticvolumetric
from the reaction of carbon monoxide with hydrogen gas. sorption analyzer (Quantachrome NOVA 1200 Version 3.5
◦
The carbon nanofiber samples formed from the iron cata- surface area analyzer) using N as an adsorbate at −196 C.
2
lysts were characterized using BET, HRTEM, EDS, and The surface areas were determined from the Brunauer–
XRD analyses.
Emmett–Teller (BET) equation.

1
52
TAKEUCHI ET AL.
2
.3. Synthesis of Iron Particles
ples. (Caution: some of the iron samples were found to be
pyrophoric.) Carbon nanofiber samples were also prepared
on double-sided sticky carbon tape, but the stubs were not
prepared or transferred under an inert atmosphere. Iron
samples for BET analyses were weighed into glass cells in
an inert atmosphere glovebox and then removed from the
glovebox and inserted quickly into the BET instrument to
minimize oxidation of the sample. Carbon samples were not
transferred to the BET analyzer under an inert atmosphere.
The iron samples were analyzed immediately, while the car-
UsingmodificationstoasynthesisreportedbyHampden-
Smith and co-workers (2), 0.580 g (3.58 mmol) of FeCl3
was stirred in 75 ml of THF for 8 h and then allowed to
age without stirring for at least 48 h to allow complete for-
mation of a red insoluble solid. The resulting pale yellow
solution was filtered and then placed in a constant addition
dropping funnel. These steps were performed in an inert
atmosphere glove box. The FeCl3 solution was then added
dropwise to a stirring solution of 12 mL LiBEt3H in THF
bon nanofiber samples were vacuum outgassed for 2 h at
(
1.0 M) and held at a constant temperature under an inert
◦
1
25 C prior to analysis. Selected iron and carbon samples
N2 atmosphere. Returning to the glove box, the resulting
black solid was separated from the reaction solution by vac-
uum filtration, washed first with 10 mL of a 1 : 1 mixture of
THF/EtOHandthenwith15mLofTHF, and finallydriedin
vacuo. Typical yields were ∼90%, based on the initial FeCl3
masses. The reaction conditions, which were carefully con-
trolled and investigated, include (i) the dropping rate of the
FeCl3 solution, (ii) the stirring rate of the solution of 1.0 M
LiBEt3H in THF, and (iii) the temperature of the reaction
vessel containing the solution of 1.0 M LiBEt3H in THF.
were also studied using TEM and HRTEM. Iron samples
were prepared by dispersing <1 mg of the iron in approx-
imately 1 ml of dry ethanol and allowing two drops of this
suspension to dry on the TEM grid in an inert atmosphere
glovebox. Carbon samples were prepared by sonicating
<1 mg of the carbon sample in approximately 1 ml tert-
butanol and then the allowing 1 or 2 drops of this suspension
to air-dry on the TEM grids.
In preparation for x-ray powder diffraction (XRD) anal-
ysis, our carbon nanofibers were stirred in 1 M hydrochloric
acid solution for at least 2 days, then washed thoroughly
2
.4. Preparation of Carbon Nanofibers Catalyzed
by Iron Particles
◦
with water, and finally dried under vacuum at 60 C. XRD
data were collected with a Shimadzu Lab X-6000 X-ray
diffractometerusingCuKα radiation. Thecarbonnanofiber
samples were placed in an aluminum “well holder” with a
sample chamber 25 mm in diameter × 1-mm deep. Fixed
The iron particles were tested for their catalytic activity
toward carbon formation using modifications of procedures
previously reported by Rodriguez et al. (12). In an inert at-
mosphere glovebox, 6 mg of iron was weighed into a porce-
lain combustion boat which was then placed in the center
of a quartz reaction tube. The reaction tube was sealed and
then transferred from the glove box to a horizontal tube
furnace, and the tube was purged for several minutes under
a flow of helium gas. A 10% H2/He mixture was introduced
into the reaction tube as the oven was heated from room
◦
divergence and scatter slits of 1.00 and a receiving slit of
0.30 mm were used. The spacing between adjacent car-
bon layers (d002) was determined using the Bragg equation,
while the dimensions of crystallites with graphite-like order
(
L002, L101, L100, L110) were calculated using the Scherrer
equation (L = Kλ/B cos ϑ) and the values of the (002),
101), (100), and (110) reflections, respectively. For the 002
(
◦
◦
temperature of 600 C at a rate of 20 C/min. Following this
heating and reduction step, a 4 : 1 CO/H2 mixture was intro-
duced into the reaction tube and carbon was formed for 6 h
reflection, K = 0.89, while for the 101, 100, and 110 reflec-
tions, K = 1.84 was used (33).
◦
at a constant temperature of 600 C. Finally, the carbon pro-
3
. RESULTS AND DISCUSSION
duced was cooled to room temperature under He flow and
passivated for 45 min using a 2% CO/He mixture before
the carbon was removed from the reaction tube. In typical
preparations, ∼1.3 g of carbon was produced from 6 mg of
iron catalyst.
3
.1. Iron Nanoparticle Synthesis and Characterization
In the course of preparing iron particles using reaction
conditions described previously by Hampden-Smith and
co-workers (2) and Bonnemann et al. (17), we found that
the morphologies of the iron particles varied as a function
of several reaction conditions, namely, the dropping rate of
the FeCl3/THF solution, the stir rate of the 1.0 M LiBEt3H
2
.5. Characterization of Iron Particles
and Carbon Nanofibers
In an inert atmosphere glovebox, iron samples were pre- in THF, and the reaction temperature. To quantitatively
pared for SEM and EDS analyses by sprinkling the iron control the dropping rate, we employed a constant addi-
onto one side of double-sided sticky carbon tape (the other tion dropping funnel (Kontes). Also, a specialty magnetic
side adhered to an aluminum stub) and sealing the stubs stirrer (VWR series 400S) and a stir bar (VWR Spinplus)
in a capped vial. The vials were opened under a nitrogen were required to control the stir rate of the reaction solu-
gas flow, and the stubs were transferred quickly into the tion from 200 to 1600 rpm. Finally, the reaction temperature
SEM sample chamber to minimize oxidation of the sam- was controlled with a constant temperature bath.

SYNTHESIS OF ACICULAR IRON PARTICLES
153
First, a study was conducted to determine the effect of
the FeCl3/THF solution addition rate on iron particle mor-
phology. Three experiments were conducted at a constant
◦
temperature (0 C) and at a constant stirring rate (200 rpm)
of the LiBEt3H solution, where the addition rates of the
FeCl3/THF solution were 1 drop/10 s, 1 drop/s, or 2 drops/s.
The low and high dropping rates were the minimum and
maximum rates which the dropping funnel could maintain.
At the fastest addition rate (2 drops/s), the morphologies
of the resulting iron particles were mixed. Some acicular
and dendritic materials were present, but granular mate-
rial was the primary product. The predominant granular
particles had diameters of approximately 200–300 nm and
2
specific surface areas of ca. 46 m /g. Simple geometric calcu-
lations based on the measured BET surface areas suggested
that the granular iron particles might be agglomerates of
smaller primary particles, with diameters of approximately
1
–2 nm (assuming spherical primary particles with smooth,
nonporous surfaces). This is consistent with the assertion by
Hampden-Smith and co-workers that, based on BET mea-
surements, 200-nm iron samples produced from the lithium
triethylhydridoborate reduction of iron trichloride could be
agglomerates of 2–4 nm primary particles (2). At the slow-
est addition rate (1 drop/10 s), the large majority of the
product had an acicular morphology. The acicular particles
had diameters of ∼200 nm, lengths on the order of tens
to hundreds of micrometers, and a specific surface area of
FIG. 1. SEM micrograph of iron particles (predominantly granular)
formed at a reaction temperature of 0 C, a FeCl3 solution addition rate of
.1 drop/s, and a LiBEt3H solution stirring rate of 1600 rpm.
◦
0
2
ca. 17 m /g. Notably, our acicular iron particles display the
highest aspect ratios ever reported (18–24).
From simple geometric considerations, spherical samples
should display a larger surface area than that of acicular character. These granular particles had diameters of ∼200–
2
samples, given that both samples have approximately the 300 nm and specific surface areas of ∼40 m /g (see Fig. 1).
same diameters and similar smooth, nonporous surfaces. Simplegeometric calculations (assumingspherical, smooth,
This trend was observed within our addition rate exper- nonporous primary particles) based on the measured BET
2
iments; specific surface areas ranged from 17 m /g at an (N ) surface areas suggested that our granular iron particles
2
2
addition rate of 1 drop/10 s (primarily acicular) to 46 m /g might be agglomerates of smaller particles (vide supra). At
at an addition rate of 2 drops/s (primarily granular). Fi- theloweststirringrate(200rpm), theresultingironparticles
nally, thesynthesisinvolvingtheintermediatedroppingrate were primarily acicular, with little to no granular character.
2
(
1 drop/s) showed a distribution of granular and acicular
Specific surface areas ranged from 17 m /g at 200 rpm
2
iron particles between the distributions discussed above, (primarilyacicular)to40m /gat1600rpm(primarilygranu-
reinforcing the above relationship between the slow addi- lar), which is consistent with the surface area measurements
tion rate and the high percentage of acicular iron products. for the dropping rate experiments (vide supra). Finally, the
A second study was conducted to determine the effect synthesis involving the intermediate stirring rate (400 rpm)
of the stirring rate of the LiBEt3H solution on iron par- showed a distribution of granular and acicular iron particles
ticle morphology. While a constant temperature of the between the distributions discussed above, reinforcing the
◦
LiBEt3H solution (0 C) and a constant addition rate of above relationship between the slow stirring rate and the
the FeCl3 solution (1 drop/10 s), were maintained, three high percentage of acicular iron products.
stirring rates of the LiBEt3H solution were employed: 200,
The final study on iron synthesis parameters was con-
4
00, and 1600 rpm. The low and high values were chosen be- ducted to determine the effect of temperature on iron parti-
cause they represented the minimum and maximum stirring cle morphology. At a constant stirring rate of the LiBEt H
3
rates which the magnetic stirrer and stir bar could maintain solution (200 rpm) and at a constant addition rate of the
throughout the course of the reaction. At the highest stir- FeCl /THF solution (0.1 drop/s), the temperature of the
3
◦
ring rate (1600 rpm), the iron particles that were formed LiBEt H reaction solution was varied from 0 to 60 C. As
3
were almost exclusively granular, with little to no acicular anticipated from the selected stirring rate and addition

1
54
TAKEUCHI ET AL.
rate experiments, all iron particles formed in the tempera-
ture study were highly acicular. At lower temperatures (0–
◦
2
0 C), the iron products formed were almost exclusively
◦
single strands. At higher temperatures (40 C), the acicular
strands became thinner and showed evidence of increased
branching. The iron product formed at 60 C had a highly
◦
branched or weblike character. Specific surface areas in this
2
◦
2
◦
series ranged from 17 m /g at 0 C to 42 m /g at 40 C. As
◦
a result of this study, it was determined that 20 C was an
optimal temperature for the formation of single-stranded,
acicular iron products.
Notably, our optimized acicular iron particles have un-
usual dimensions (see Fig. 2). Previously reported acicular
iron particles display lengths of 10–100 µm and diameters
of ∼200–300 nm, with aspect ratios ranging from ∼ 5 to
∼
12 (18–24). Our iron particles have aspect ratios ranging
from ∼50 to ∼500, which represent a notable increase in as-
pect ratios from those previously reported in the literature.
Also, since HRTEM images of our acicular iron samples do
not show individual spheres or cubes within a single acicu-
lar iron fiber (see Fig. 3), we propose that our acicular iron
samples do not display a “chain of cubes” (23) morphology.
Finally, electron diffraction analysis of our highly acicu-
lar and highly granular iron samples showed both samples
to be microcrystalline, with several phases of iron present,
FIG. 3. HRTEM micrograph of an individual acicular iron particle
from the collection of iron particles shown in Fig. 2.
although it was not possible to conclusively identify each
phase. Energy dispersive spectral analyses (EDS) showed
the samples to be homogeneous, with iron as the only ele-
ment detected.
3.2. Carbon Nanofiber Preparation and Characterization
Carbon nanofibers were synthesized using our acicular
or granular iron particles as catalysts, in a modification of
a procedure previously reported by Rodriguez et al. (12).
In typical preparations, ∼1.3 g of carbon was produced
from 6 mg of iron catalyst after 6 h of carbon deposition.
BET surface area analyses showed that the carbon samples
FIG. 2. SEM micrograph of iron particles (predominantly acicular)
formed at a reaction temperature of 20 C, a FeCl3 solution addition rate
of 0.1 drop/s, and a LiBEt3H solution stirring rate of 200 rpm.
◦

SYNTHESIS OF ACICULAR IRON PARTICLES
155
produced from granular and acicular iron catalysts had spe-
2
cific surface areas of ca. 130 and 160 m /g, respectively. EDS
of the carbon samples produced from granular and acicular
iron catalysts were similar, showing large peaks for carbon
and small peaks corresponding to the residual catalyst on
exposure to air (i.e., iron and oxygen) which were just above
the level of the background noise.
The carbon nanofiber samples were imaged using SEM,
TEM, and HRTEM. For the carbon fibers prepared from
either granular or acicular iron catalysts, the catalyst parti-
cle was typically observed at one end of the fiber. No ev-
idence was seen for the existence of branched fibers from
either sample. Carbon nanofibers produced from either aci-
cular or granular iron catalysts showed similar fine struc-
tures under high magnification, namely, regions of ordered
graphitic planes along the edges of the fibers (see HRTEM
image, Fig. 4). Additionally, the spacings between the
graphitic planes in the carbon nanofibers were found to be
∼
0.34 nm, reminiscent of the spacings found in single-
crystal graphite.
XRD analysis was used to gain information about the di-
mensions of the ordered regions in the two types of carbon
nanofiber samples. The spacing between adjacent carbon
layers d002 was found to be similar, with carbon nanofibers
produced from acicular and granular iron catalysts having
˚
spacings of 3.40 and 3.41 A, respectively. The dimensions
FIG. 5. TEM micrograph of carbon nanofibers grown from the gran-
ular iron catalyst shown in Fig. 1.
of the parallel carbon layers (L101, L100, L110) were deter-
mined from the (101), (100), and (110) reflections, respec-
tively, and the average stacking height of parallel layers
(
L002) was determined from the (002) reflection. The car-
bon nanofibers produced from acicular iron catalysts had
˚
˚
˚
an L002 of 57 A, an L101 of 162 A, an L110 of 211 A, and an
˚
L100 of 92 A. Interestingly, the carbon nanofibers produced
from granular iron catalysts had larger dimensions, with an
˚
˚
˚
L002 of 72 A, an L101 of 270 A, an L110 of 290 A, and an L100
˚
of 100 A. This result suggests that the morphology of the
iron catalyst does have an effect on crystallite dimensions
of the resulting carbon nanofibers.
Interestingly, some of the carbon nanofibers grown from
the granular iron catalyst displayed a spiral gross morphol-
ogy under lower magnification (see TEM image, Fig. 5).
This morphology was not observed in the carbon grown
from the acicular iron catalyst, where the carbon nanofibers
were straight or gently curved (see TEM image, Fig. 6). It
has been previously shown that the symmetrical diffusion of
the carbonaceous starting materials through the metal cata-
lysts results in the generation of relatively straight fibers,
while unsymmetrical diffusion results in the generation of
twisted or helical fibers. It has also been noted that the
presence of additives in the metal particle can promote for-
mation of these twisted structures, as they may alter the
diffusion pathway of the catalyst (34–38). Our results indi-
cate that the morphology of the iron catalyst may be a third
influence on the structure of the carbon nanofibers.
FIG. 4. HRTEM micrograph showing graphitic platelets of an indi-
vidual carbon nanofiber from the collection of carbon nanofibers shown
in Fig. 6.

1
56
TAKEUCHI ET AL.
4
. CONCLUSIONS
With precise control of key reaction conditions, it is pos-
sible to utilize a single-step chemical reduction method to
produce iron particles that are either primarily acicular or
granular in morphology. The acicular iron particles formed
herein are notable in that their aspect ratios are unusu-
ally large (50–500). Both the acicular and the granular iron
particles functioned well as catalysts for carbon nanofiber
formation. The carbon nanofibers were found to be at least
partially graphitic in nature, and the d002 spacings were very
similar, but the crystallite dimensions differed, depending
on whether the iron catalyst was acicular or granular. In
addition, the diameters of the iron catalysts corresponded
to the diameters of the carbon nanofibers, and the forma-
tion of carbon nanofibers having a straight or spiral gross
morphology may be related in part to the iron catalyst mor-
phology. Thus, future studies are warranted, where further
modification of either the iron catalyst structure or the com-
position may provide some tuning of the graphitic proper-
ties and the gross morphologies of carbon nanofibers and
ultimately some control of their chemical properties.
ACKNOWLEDGMENTS
FIG. 6. TEM micrograph of carbon nanofibers grown from the acic-
ular iron catalyst shown in Figs. 2 and 3.
This work was supported in part by grants from the Research Corpora-
tion (ROA for KJT, and CC4252 for CAB) and from the Office of Naval
Research (N00014-96-1-0047 for CAB).
Finally, while the granular iron catalyst particles were
structurallyrobustundercarbonfiberformationconditions,
some fragmentation did occur with the acicular catalyst par-
ticles. Analysis of the electron micrographs indicates that
the acicular particles undergo transverse fragmentation to
form more granular particles of a diameter similar to that
of the starting acicular particles, but of aspect ratios closer
to 1. Other authors have noted that fragmentation of the
powdered metal starting materials is possible during the
carbon deposition process and that the diameter of the ac-
tive metal catalyst may be smaller than the diameter of
the powdered metal starting materials (15). As a result, the
diameter of the carbon nanofiber can be smaller than the
diameter of the powdered metal starting material. How-
ever, in our observation, the cross-sectional diameter of the
carbon nanofibers matched the diameter of the powdered
metal starting materials from which they were produced,
suggesting that the diameter of our active metal catalysts
matches the diameter of our powdered metal starting ma-
terials. The diameters of the starting acicular and granular
REFERENCES
1
. Bessel, C. A., Laubernds, N. M., Rodriguez, N. M., and Baker,
R. T. K., J. Phys. Chem. B 105(6), 1115 (2001).
2. Duan, Z., Hampden-Smith, M. J., Datye, A., Nigrey, P., Quintana, C.,
and Sylwester, A. P., J. Catal. 139, 504 (1993).
3
. Kieffer, R., and Benesovsky, F., Ullmanns Encykl. Tech. Chem. 4, 563
(
1980).
4
5
. Roll, K. H., Kirk-Othmer Encycl. Chem. Tech. 3, 28 (1982).
. Lee, P. W., Trudel, Y., Iacocca, R., German, R. M., Ferguson, B. L.,
Eisen, W. B., Moyer, K., Madan, D., and Sanderow, H., Eds., “ASM
Handbook, Vol 7: Powder Metal Technologies and Applications,”
ASM International, Materials Park, OH, 1998.
. Rao, C. N. R., Kulkarni, G. U., Govindaraj, A., Satishkumar, B. C., and
John Thomas, P., Pure Appl. Chem. 72(1–2), 21 (2000).
. Fendler, J. H., Ed., “Nanoparticles and Nanostructured Films.” Wiley–
VCH, Weinheim, 1998.
. Braunstein, P., Oro, G., and Raithby, P. R., Eds., “Metal Clusters in
Chemistry, Nanomaterials and Solid State Chemistry,” Vol. 3. Wiley–
VCH, Weinheim, 1999.
. Markovich, G., Collier, C. P., Henrichs, S. E., Remacle, F., Levine,
R. D., and Heath, J. R., Acc. Chem. Res. 32, 397 (1999).
6
7
8
9
iron powders were similar, on the order of 200–300 nm, and 10. For recent examples of environmental chemistry involving iron par-
ticles, see (a) Farrell, J., Melitas, N, Kason, M., and Li, T., Envi-
ron. Sci. Technol. 34, 2549 (2001); (b) Schuetz, C., Winther, K., and
Kjeldsen, P., Environ. Sci. Technol. 34, 2557 (2001); (c) Ponder, S.
M., Darab, J. G., and Mallouk, T. E., Environ. Sci. Technol. 34, 2564
the diameters of the carbon nanofibers formed from both
powders were also similar, approximately 200–300 nm. In
carbon nanofiber micrographs in which intact metal cata-
lysts were visible, the catalyst particle was typically at one
(
2001).
end of the nanofiber.
11. Baker, R. T. K., Carbon 27, 315 (1989).

SYNTHESIS OF ACICULAR IRON PARTICLES
157
1
1
2. Rodriguez, N. M., Chambers, A., and Baker, R. T. K., Langmuir 11,
862 (1995).
26. Rinzler, A. G., Hafner, J. H., Nikolaev, P., Lou, L., Kim, S. G.,
Tomanek, D., Nordlander, P., Colbert, D. T., and Smalley, R. E.,
Science 269, 1550 (1995).
27. Wang, Q. H., Setlur, A. A., Laverhaas, J. M., Dai, J. Y., Seelig, E. W.,
and Chang, R. P. H., Appl. Phys. Lett. 272, 2912 (1998).
28. Collins, P. G., Zettl, A., Bando, H., Thess, A., and Smalley, R. E.,
Science 278, 100 (1997).
3
3. Baker, R. T. K., and Rodriguez, N. M., in Catalytic growth of car-
bon nanofibers and nanotubes. “Novel Forms of Carbon II.” Vol. 349,
pp. 251–256. Materials Research Society, Pittsburgh, PA, 1994.
4. Hernadi, K., Fonseca, A., Nagy, J. B., Bernaerts, D., and Lucas, A. A.,
Carbon 34(10), 1249 (1996).
1
1
1
1
5. Rodriguez, N. M., J. Mater. Res. 8, 3233 (1993) and references therein.
6. Yang, R. T., and Chen, J. P., J. Catal. 115, 52 (1989).
7. Boennemann, H., Brijoux, W., Brinkmann, R., Fretzen, R., Joussen,
T., Koeppler, R., Korall, B., Neitler, P., and Richter, J., J. Mol. Catal.
29. Frank, S., Poncharal, P., Wang, Z. L., and de Heer, W. A., Science 280,
1744 (1998).
30. Chambers, A., Park, C., Baker, R. T. K., and Rodriguez, N. M., J. Phys.
Chem. B 102, 4253 (1998).
8
6, 129 (1994).
8. Sueyoshi, T., Naono, H., and Amemiya, M., IEEE Trans. Magn. 23, 80
1987).
31. Park, C., Anderson, P. E., Chambers, A., Tan, C. D., Hidalgo, R., and
Rodriguez, N. M., J. Phys. Chem. B 103, 10,572 (1999).
32. Ebbesen, T. W., “Carbon Nanotubes: Preparation and Properties.”
CRC Press, Boca Raton, FL, 1997.
33. Kinoshita, Kim, “Carbon: Electrochemical and Physicochemical
Properties.” Wiley, New York, 1988.
1
(
1
2
2
9. Cain, J. L., and Nikles, D. E., J. Appl. Phys. 79, 4860 (1996).
0. Cain, J. L., and Nikles, D. E., IEEE Trans. Magn. 32, 4490 (1996).
1. Cain, J. L., and Nikles, D. E., IEEE Trans. Magn. 33, 3718
(
1997).
34. Anderson, P. E., and Rodriguez, N. M., Chem. Mater. 12, 823 (2000).
35. Nishiyama, Y., and Tamai, Y., J. Catal. 33, 98 (1974).
36. De Bokx, P. K., Kock, A. J. H. M., Boellaard, E., Klop, W., and Geus,
J. W., J. Catal. 96, 454 (1985).
2
2
2. Chen, M., and Nikles, D. E., J. Appl. Phys. 85, 5504 (1999).
3. Chen, M., Tang, B., and Nikles, D. E., IEEE Trans. Magn. 34, 1141
(
1998).
2
4. Prozorov, T., Prozorov, R., Koltypin, Y., Felner, I., and Gedanken, A.,
J. Phys. Chem. B 102, 10,165 (1998).
37. Kim, M. S., Rodriguez, N. M., and Baker, R. T. K., J. Catal. 131, 60
(1991).
2
5. de Heer, W. A., Chatelain, A., and Ugarte, D., Science. 270, 1179
38. Krishnankutty, N., Park, C., Rodriguez, N. M., and Baker, R. T. K.,
Catal. Today 37, 295 (1997).
(1995).