The thermal decomposition of three magnetic acetates at their autogenic pressure yields different products. Why?

Article abstract of DOI:10.1002/ejic.200700146

We report on the one-stage, reproducible, solvent-free, competent and straightforward approach for the synthesis of fullerene-like Ni@C, Co@C, and Fe₃O₄@C core-shell nanostructures that can be scaled up. The single precursor reactions of low cost acetates of Fe, Co and Ni are separately conducted at a relatively low temperature (700°C) in a closed Swagelok reactor, as compared to other methods for the formation of graphitic layers. It is worth mentioning that although identical reaction parameters are employed, using the three acetate precursors, the graphitic carbon is coated on nanosized metallic Ni and Co cores, while Fe tends to form Fe₃O ₄, maintaining the same core-shell morphology. The systematic morphological, compositional, structural characterization and the room temperature magnetic susceptibility measurements of the as-made particles are carried out on a vibrating sample magnetometer. The plausible mechanism is based on the comparison between the dissociation products of three acetate precursors, their obtained experimental data, and calculations on the enthalpy and free energy changes. Wiley-VCH Verlag GmbH & Co. KGaA, 2007.

Full text of DOI:10.1002/ejic.200700146

FULL PAPER
DOI: 10.1002/ejic.200700146
The Thermal Decomposition of Three Magnetic Acetates at Their Autogenic
Pressure Yields Different Products. Why?
Swati V. Pol,[a] Vilas G. Pol,[a] Israel Felner,[b] and Aharon Gedanken*[a]
Keywords: Decomposition / Acetates / Magnetic properties / Core-shell nanostructures
We report on the one-stage, reproducible, solvent-free, com- maintaining the same core-shell morphology. The systematic
petent and straightforward approach for the synthesis of ful-
lerene-like Ni@C, Co@C, and Fe3O4@C core-shell nanostruc-
tures that can be scaled up. The single precursor reactions of
low cost acetates of Fe, Co and Ni are separately conducted
at a relatively low temperature (700 °C) in a closed Swagelok
reactor, as compared to other methods for the formation of
graphitic layers. It is worth mentioning that although iden-
tical reaction parameters are employed, using the three ace-
tate precursors, the graphitic carbon is coated on nanosized
metallic Ni and Co cores, while Fe tends to form Fe3O4,
morphological, compositional, structural characterization
and the room temperature magnetic susceptibility measure-
ments of the as-made particles are carried out on a vibrating
sample magnetometer. The plausible mechanism is based on
the comparison between the dissociation products of three
acetate precursors, their obtained experimental data, and
calculations on the enthalpy and free energy changes.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
composition, particle size and morphology, a narrow
monodisperse size distribution, and stabilization against
oxidation in the case of reactive metals.
Introduction
The studies of magnetic nanoparticles and nanostruc-
tures combine a broad range of synthetic and investigative
techniques from physics, chemistry, and materials science.
Ultra fine magnetic particles have been the subject of in-
tense scientific research in recent years with respect to their
novel properties and commercially-potential valuable appli-
cations.[1] The application of the encapsulated magnetic
nanomaterials ranges from magnetic inks, magnetic re-
cording media, toners for xerography, and ferrofluids for
biomedical applications, such as contrast agents in magnetic
resonance imaging.[2] Single- and multi-walled carbon
nanotubes filled with ferromagnetic materials (Fe, Co, Ni)
showed significant potential in data storage.[3] The ferro-
magnetic nanosized Fe, Ni, or Co coated with carbon con-
stitute a group of especially interesting magnetic nanomat-
erials. This is because of their fascinating magnetic proper-
ties, and the carbon coating that protects the magnetic
nanoparticles against environmental degradation.[3] The
discovery of fullerenes and carbon-coated materials ac-
celerated extensive investigation on nanoparticles, carbon-
coated nanomaterials, and a wide range of carbon-based
materials.[4] Some of the standard problems encountered in
the synthesis of nanoparticles are those of good control of
Various synthetic methods of magnetic nanoparticles
have been reported in recent times, e.g., wet chemical pro-
cesses,[5–6] inert gas condensation,[7] ultra-rapid quenching
and vacuum-based deposition.[8] Carbon nanotubes and
filled carbon nanostructures such as onions, wires, and
rods, have been prepared by the pyrolysis of metallo-
cene.[9–10] Using Ni(acac) and Cu(acac) as catalyst precur-
sors, carbon nanotubes and onions from carbon monoxide
were produced previously.[11] An arc-discharge process[12]
was employed for the preparation of ferromagnetic carbon-
coated Fe, Co, and Ni nanoparticles. The major problem
with the synthesis of graphite-encapsulated nanocrystals by
using the standard arc method has been the production of
large amounts of unwanted carbonaceous debris.[13] This
causes the reduction of particle-exchange coupling, which
is an important issue in the ongoing attempt to synthesize
high density magnetic recording media.[14a] Vivekchand et
al. recently developed a new method for the preparation of
metal nanowires by the nebulized spray pyrolysis of precur-
sors.[14b] The arc evaporation method is also successfully
demonstrated for the encapsulation of ferromagnetic metals
(iron, cobalt, and nickel) nanoparticles in graphitic car-
bon.[14c] Chemical vapor synthesis (CVS) is becoming in-
creasingly important because of its potential for producing
nanomaterials in a controlled and scalable way with a low
degree of agglomeration.[15] Each technique has its own ad-
vantages and disadvantages. However, most of the tech-
niques require a high energy input and are not suitable for
[a] Department of Chemistry and Kanbar Laboratory for Nanom-
aterials at the Bar-Ilan University Center for Advanced Materi-
als and Nanotechnology, Bar-Ilan University,
Ramat-Gan, 52900, Israel
E-mail: gedanken@mail.biu.ac.il
[b] Rakah Institute of Physics, Hebrew University,
Jerusalem, 91904, Israel
Eur. J. Inorg. Chem. 2007, 2089–2096
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2089

S. V. Pol, V. G. Pol, I. Felner, A. Gedanken
FULL PAPER
mass production because of the low yield and the compli- Ni@C, Co@C, and Fe3O4@C products, respectively. The
cated process control. The research community is seeking calculated weight of metal Ni, Co, and Fe is 332 mg,
methods for producing carbon nanomaterials with specific 333 mg, and 321 mg in 1 g of nickel(II) acetate, cobalt(II)
properties and on an industrial scale. Thus, further experi- acetate, and iron(II) acetate precursors, while 337 mg,
mental research on their synthesis is needed.
339 mg, and 315 mg account for the 382 mg, 400 mg, and
Herein, we report on a one-stage, reproducible, solvent- 560 mg of the Ni@C, Co@C, and Fe3O4@C products,
free, competent, and straightforward approach for the syn- respectively. It is obvious that the products show less carbon
thesis of fullerene-like Ni@C, Co@C, and Fe3O4@C core- than the reactants, while the percent of hydrogen is almost
shell nanostructures that can be scaled up. The single pre- negligible. This means that some of the carbon is bonded
cursor reactions of low-cost acetates of Fe, Co and Ni are to the hydrogen in the form of hydrocarbons and/or to oxy-
conducted separately at a relatively low temperature gen to form carbon oxides. Furthermore, the reduction in
(700 °C) in a closed Swagelok reactor, as compared to other the transition metal is observed in both Ni@C and Co@C.
methods for the formation of graphitic layers. It is worth This reduction might have occurred due to carbon mon-
mentioning that although identical reaction parameters are oxide. Jesus et al. reported on the thermal decomposition[17]
employed, using the three acetate precursors, graphitic car- process of the nickel acetate precursor under inert
bon-coated nanosize metals are obtained for Ni and Co, atmosphere. Our obtained coupled TGA-Mass spec-
while Fe tends to form Fe3O4, maintaining the same core- troscopy analysis data for Ni acetate salt are in agreement
shell morphology. Early results on a similar reaction of with this. The TGA spectra of all the three acetates are
Ni(AcAc)2 have been reported.[16] This publication extends shown in Figure 1.
this report and also includes a comparison between FeII,
NiII, and CoII acetates, which yield different products. This
manuscript also attempts to account for and explain the
formation of different products. The systematic morpho-
logical, compositional, structural characterization and
room temperature magnetic susceptibility measurements of
the as-made particles are measured by using a vibrating
sample magnetometer. The suggested mechanism is sup-
ported by the obtained experimental data and calculations
on enthalpy, as well as free energy changes.
Results and Discussion
1. C, H, N, S, and O Analysis, X-ray Diffraction,
Mössbauer, and BET Surface Area Measurements
Figure 1. The TGA curve of Fe, Co, and Ni acetates.
Here we present and compare the results related to the
The complete decomposition of all the acetates is ob-
three metal acetates that produce core/shell structures of served at about 400 °C. The higher weight loss is observed
metal/metal oxide surrounded by carbon. The carbon and for Ni and Co acetates (68%), while early (ca. 300 °C) and
hydrogen content in the product/carbonaceous material is less weight loss (60%) accounts for Fe acetate. The residual
determined by elemental analysis measurements (Table 1). weight of about 32% corresponds to the metallic Ni and
The calculated element percentages of carbon in Ni, Co and Co for the related acetates, while the 40% for Fe acetate
Fe acetates are 27.2, 27.1, and 27.6%, while the element corresponds to FeO. The gaseous content of the dissoci-
percentages of hydrogen are 3.4, 3.4, and 3.5%, respectively. ation products of these acetates at 350 °C is monitored by
The measured percentage of carbon in Ni@C, Co@C, and an MS. The gases detected by the instrument are acetone,
Fe3O4@C products are 11.78, 15.25, and 22.3 wt.-%, while CH3COCH3 (m/z = 58), acetic acid, CH3COOH (60), CO2
the percentages of hydrogen are 0.18, 0.1, and 0.12%, (44), CO (28), CH4 (16), C2H6 (30), and C2H2 (26).
respectively. The calculated oxygen percentage for Ni, Co,
Conducting the RAPET reaction of all three acetates at
and Fe acetates is almost the same (ca. 36 wt.-%). The mea- 650 °C does not lead to any crystalline products, and only
sured oxygen percentages are 0.2, 0.3, and 21.43 wt.-% for amorphous products are obtained, based on XRD mea-
Table 1. The obtained products, % yield, morphology, crystal structure, percent of carbon, hydrogen and oxygen and BET surface area
are illustrated.
Sample name
% Yield
XRD
Morphology (TEM) wt.-% C
wt.-% H
wt.-% O
BET Surface
area m2/g
Ni@C
Co@C
Fe3O4@C
38.2
40
56
Ni
Co
Fe3O4
core-shell
core-shell
core-shell
11.78
15.25
22.3
0.18
0.10
0.12
0.2
0.3
21.43
1
3
2
2090
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 2089–2096

Nanosized Carbon-Coated Metallic Cores
FULL PAPER
surements. The first crystalline products for the RAPET of
the acetates appear at 700 °C. The RAPET of Ni and Co
acetates in a closed cell at 700 °C yields 38.2% and 40% of
products containing metallic Ni and Co, respectively coated
with carbon. The dissociation of Fe acetate in a closed RA-
PET system confirms the weight loss of 44%, keeping about
56% of a solid product consisting of Fe3O4 and carbon.
The rest of the weight loss is due to the formation of hydro-
carbons and carbon oxides. In the case of the RAPET of
Fe acetate, the reactant possesses a +2 oxidation state, while
the product has a 2.66 oxidation state. This means that un-
like Co and Ni ions, which undergo a reduction to the cor-
responding metals, the Fe+2 is partially oxidized to Fe3O4
before being coated by carbon. Since the particles are
densely coated with graphitic carbon, the oxidation of the
product due to handling under air atmosphere is imposs-
ible. It is noticeable that the Fe is more facile to oxidation
than Ni and Co. The measured surface areas for the Ni@C,
Co@C and Fe3O4@C products are 1, 3, and 2 m2/g, respec-
tively.
Figure 2. XRD patterns of as-prepared Ni@C (top), Co@C
(middle), and Fe3O4@C (bottom) samples.
two isomer shift values, 0.67 and 0.37 mm/s, correspond to
site A (Fe3+) ions and to site B (Fe2+, Fe3+) ions, respec-
tively. The high isomer shift for the B site is typical for Fe2+
The resultant parameters are given in Table 2.
.
The XRD pattern of the as-prepared Ni@C sample is
presented in Figure 2 (top). The diffraction peaks, observed
at 2θ = 44.49, 51.85, and 76.38, are assigned as (111), (200),
and (220) reflection lines of the face-centered cubic phase
of Ni (space group: Fm-3m). These values are in good
agreement with the diffraction peaks, peak intensities and
cell parameters of crystalline Ni (PDF No. 87-712). The
diffraction patterns of the as-prepared Co@C sample (mid-
dle) can be indexed to fcc (space group: Fm-3m) Co, and
matches well with PDF No. 89-4307. The diffraction peaks
observed at 2θ = 44.22, 51. 35, and 75.86 are assigned as
(111), (200), and (220) reflection lines. The XRD pattern of
the as-prepared Fe3O4@C sample is presented at the bot-
tom of Figure 2. The major diffraction peaks are observed
at 2θ = 30.10, 35.45, 37.08, 43.08, 53.45, 56.98, 62.57, and
74.02 and are assigned as (220), (311), (222), (400), (422),
(511), (440) and (533) reflection lines. These values are in
good agreement with the diffraction peaks, peak intensities
and cell parameters of crystalline Fe3O4 (PDF No. 65-
3107). However, since the XRD of γ-Fe2O3 is very similar
to that of magnetite, we also used Mössbauer spectroscopy
to further substantiate our assignment. In none of the three
cases is a carbide phase detected. The small diffraction peak
at 2θ = 26.6°, assigned to graphitic carbon, is only detected
for Co@C, and is suppressed in Ni and Fe3O4.
Figure 3. The Mössbauer spectrum of a Fe3O4@C sample at room
temperature.
Table 2. The Mössbauer parameters of Fe3O4@C sample: isomeric
shift (mm/s), quadrupole moment (mm/s) and effective field (kOe)
for site A and site B.
I.S. (mm/s)
eQ (mm/s)
Heff (kOe)
Site A
Site B
0.67
0.37
0.028
–
455
487
The nature of the coated carbon for all of three samples
is also studied by employing Raman spectroscopy and HR-
TEM measurements.
The Mössbauer spectra of Fe3O4 at 298 K are shown in
Figure 3. The spectra show two hyperfine magnetic splits,
which are clear evidence for the existence of Fe3O4 (magne-
tite). In the transmission spectrum, the hyperfine split arises
2. TEM, SEM, EDS, Raman, and HR-TEM
Measurements
The morphology of all the samples is studied by TEM
due to a nuclear quadrupole interaction. This indicates that measurements. The transmission electron micrograph of
there is some departure from the cubic symmetry of the Ni@C core-shell structures is shown in part a of Figure 4.
electric field due to the existence of an electric field gradient The thermal dissociation of Ni acetate favors the formation
3
at the nucleus. It means that the 57Fe excited level (I = /2) of the dark particles of Ni, which are coated with carbon
1
has split, whereas the ground state (I = /2) remains whole/ layers. We assume that the thermal decomposition of
intact. The difference in energy levels is only due to dif- Ni(CH3COO)2, leads to dissociation into carbon, hydrogen,
ferent electronic densities of the Fe atoms at the source, and oxygen and nickel atoms at 700 °C. The Ni atoms forms the
the absorber. Fe3O4 can be written as [(Fe3+)(Fe2+Fe3+)]O4. Ni core. The particle diameters range from 20 nm to
The magnetic hyperfine fields of 455 and 487 kOe and the 160 nm. The high resolution image (Figure 4b) taken on a
Eur. J. Inorg. Chem. 2007, 2089–2096
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2091

S. V. Pol, V. G. Pol, I. Felner, A. Gedanken
FULL PAPER
single particle revealed that the 25 nm dark core of Ni is 160 nm. The EDS of the Ni@C particles is presented in
encapsulated in 5–10 nm carbon layers. Using the SAEDS Figure 5b, and confirms the presence of Ni and C. No ad-
technique, the identification of the dark cores as Ni, Co ditional peak is detected, which verifies the purity of the
and iron oxide, which are covered by carbon, is confirmed Ni@C sample.
for all the three products of (Ni, Co, and Fe) acetates. The
TEM image of the Co@C sample demonstrates (Figure 4c)
that the particles, whose diameters range from 30 to
230 nm, are covered with in situ formed carbon layers after
the dissociation of cobalt acetate. The central 110 nm dia-
meter cobalt particle is uniformly surrounded by a 20 nm
carbon layer (Figure 4d). Thus, a similar trend is observed
for the thermal decomposition of Co(CH3COO)2 and
Ni(CH3COO)2, resulting in a metallic core and a carbon
shell structure. Part e of Figure 4 shows that the 5 to 20 nm
particles of Fe3O4 are densely covered with carbon. The
HR-TEM image (Figure 4f) of a 10 nm Fe3O4 particle is
surrounded by a thick layer of carbon. The thermal dissoci-
ation of Fe acetate favors the formation of a Fe3O4 core,
which is coated with carbon layers after the decomposition
of Fe(CH3COO)2. This means that the produced carbon,
hydrogen, oxygen and FeO facilitate the FeO to react with
oxygen to form Fe3O4, thus hindering the reduction.
Figure 5. Scanning electron micrographs of (a) Ni@C sample, (b)
EDS of Ni@C sample, (c) Co@C sample, (d) EDS of Co@C sam-
ple, (e) Fe3O4@C sample [arrow indicates egg-like carbon entities],
and (f) EDS of Fe3O4@C sample.
The SEM image of the Co@C sample demonstrates (Fig-
ure 5c) that the particle diameter ranges from 50 to 200 nm.
In addition, the obvious aggregating tendency of magnetic
Co@C particles is observed. The EDS analysis of the
Co@C sample (Figure 5d) detected only the presence of co-
balt and carbon, without additional impurities. Negligible
amounts of oxygen are detected in Ni@C and Co@C sam-
ples.
The Fe3O4@C sample possesses an oval shaped mixture
of about 400 nm and aggregates of about 100 nm particles
(Figure 5e). The SAEDS confirm that the oval shaped par-
ticles, indicated by an arrow, are composed solely of carbon,
Figure 4. Transmission electron micrographs of (a) Ni@C sample,
(b) single Ni@C particle at high resolution, (c) Co@C sample, (d)
while the smaller particles are composed of Fe3O4 and car-
bon. The core-shell nature of Fe3O4 surrounded by C is
single Co@C particle at high resolution, (e) Fe3O4@C sample, and
substantiated by TEM measurements. The excess of carbon
might have lead to the formation of egg-like carbon entities
in the Fe3O4@C sample. The bulk EDS analysis of the
(f) single Fe3O4@C particle at high resolution.
The SEM image shown in Figure 5a indicates the spheri- Fe3O4@C sample is shown in part f of Figure 5, presented
cal morphology of the Ni@C. The Ni-C particles have a Fe, O, and carbon occurrence. The formation of Fe3O4 and
tendency to form a few agglomerates, which are observed carbon without additional impurities is confirmed by the
in addition to the individual particles. The aggregation is EDS spectrum shown in Figure 5f.
attributed to their magnetic nature. The diameters of the
The nature of the carbon shell is investigated by Raman
separated individual particles are in the range of 40– spectroscopy. The Raman spectra (Figure 6) are the charac-
2092
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 2089–2096

Nanosized Carbon-Coated Metallic Cores
FULL PAPER
teristics of disordered graphitic carbon with nanosized gra-
phitic crystals. The first D peak, located at ca. 1330 cm–1,
originates from the in-plane lattice vibrations. The second
G peak is located at ca. 1595 cm–1, and only occurs when
small, crystal-size graphitic particles are present.[18] The D
band is associated with graphitic edge-planes and defects in
the graphitic lattice. The relative intensity of the D and G
bands is generally used to estimate the size of graphitic clus-
ters in the hexagonal plane, La. Although all the samples
are prepared at 700 °C, the intensity of graphitic carbon is
much higher than the disordered carbon for the Ni@C sam-
ple. Almost the same intensity is observed for the Co@C
and Fe3O4@C samples. The results indicate that Ni is the
best catalyst for the preparation of high order graphitic pla-
nes. This is also manifested in the relatively low temperature
that the graphitic layers reach.
Figure 7. HR-TEM images of a) carbon layer coated on the Ni
core, b) carbon coated on the Co core, and c) carbon layer on the
surface of Fe3O4 core.
To summarize, the few points that need further clarifica-
tion and discussion on the results reported for the decom-
position of all three acetates (Co, Ni, and Fe) are: 1) The
cores of Co, Ni, and iron oxide are surrounded by graphitic
carbon (interlayer spacing ≈ 0.34 nm). 2) A wide particle
size distribution is observed. 3) The thickness of the carbon
coating around the cores is proportional to the percentage
of carbon present in the product (see Table 1). 4) Although
the graphitic carbon encapsulates the cores, nanosize Ni
and Co are obtained in metallic form, while Fe tends to
form Fe3O4 (confirmed by XRD, EDS, and SAEDS), main-
taining the core-shell morphology. The three-dimensional-
ity of the onion-like structures of Ni@C, Co@C, and
Fe3O4@C products is confirmed by tilting the TEM grid
during HR-TEM measurements, when the pictures re-
mained unchanged.
3. Magnetic Properties
A room temperature magnetic susceptibility measure-
ment of Ni@C, Co@C, and Fe3O4@C samples is conducted
by employing a vibrating sample magnetometer. A small
amount of material, ca. 20 mg, is inserted into a plastic cap-
sule. Cotton wool is placed on the top of the sample and
the capsule is sealed. This prevents the particles from any
movement caused by the vibration of a VSM sample holder
rod, or by a variation in the applied magnetic field during
measurements. The magnetization vs. magnetic field curves
are measured for all the samples (Figure 8a–c) in the range
of –15800 to 16000 Oe.
The Ni@C sample exhibits typical hysteresis with a satu-
ration magnetization of 21.7 emu/g (Figure 8a). Taking into
account the fact that the sample contains 11.78% carbon
and the remaining 88.22% Ni, this yields a value of
24.6 emu/g for 100% Ni. A 151 Oe coercive field is obtained
Figure 6. Raman spectra of a) Ni@C, b) Co@C, and c) Fe3O4@C
samples.
In order to understand the graphitization of formed car- for the Ni@C sample. The Co@C sample shows a satura-
bon layers around the cores, HR-TEM measurements of tion magnetization of 21.87 emu/g (Figure 8b). When con-
the carbon-shells are carried out for the Ni@C, Co@C, and sidering that the sample contains 84.75% of Co, this gives
Fe3O4@C samples. The measured interlayer spacings be- a value of 25.81 emu/g for 100% Co. This sample shows
tween these graphitic planes are 0.345, 0.337, and 0.341 nm the coercivity of 97 Oe with a remanent magnetization of
for Ni@C (Figure 7a), Co@C (Figure 7b), and Fe3O4@C 2.6 emu/g. Figure 8c shows the magnetization curve of the
(Figure 7c) samples, respectively, which are very close to Fe3O4@C sample. Unlike the ferromagnetic behavior de-
that of the graphitic layers. It should be noted that the gra- tected for Ni@C and the Co@C, the characteristic behavior
phitic ordering is much better in Ni than in the Co and of superparamagnetic particles is detected for Fe3O4@C.
Fe3O4.
The magnetization curve does not show hysteresis and does
Eur. J. Inorg. Chem. 2007, 2089–2096
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2093

S. V. Pol, V. G. Pol, I. Felner, A. Gedanken
FULL PAPER
further to carbon. In our previous RAPET reactions of
various organic or organometallic materials, we observed
the appearance of carbon above 550 °C.[22–24]
The decomposition of Co(ac)2 and Ni(ac)2 results in a
metallic core and carbon shell structures, while the thermal
dissociation of Fe acetate favors the formation of a Fe3O4
core, which is coated with graphitic and disordered carbon
layers. It should be noted that the only source of oxygen is
the acetate ion, and its quantity in all three acetates is al-
most the same (36 wt.-%).
Based on the obtained MS data, we propose the dissoci-
ation of acetates as follows:
M(CH3COO)2 Ǟ MCO3 + CH3COCH3
MCO3 Ǟ MO + CO2
(1)
(2)
(3)
(4)
CH3COCH3 Ǟ CO + C2H6
MO + CO Ǟ M + CO2
where M = Ni, Co, Fe.
The formation of the first oxygenated product, acetone,
occurs by Equation (1). The metal carbonate,[17,25] which is
a short-lived intermediate, immediately decarboxylates to
give a metal oxide at 350 °C. This metal oxide, in the case
of Ni and Co acetates, gets reduced by carbon monoxide,
while the Fe atoms prefer to stay as FeO. It further is oxid-
ized to form Fe3O4 and then coated by carbon to form the
core-shell structure during solidification. This can be ex-
plained by the reduction potential values of Ni, Co, and Fe.
The well-known standard reduction potentials[26] at 25 °C
for Ni, Co and Fe are –0.25, –0.27, and –0.44 V, respec-
tively, for the metal–metal ion equilibrium. These numbers
indicate that iron is less favorably reduced than cobalt and
nickel.
Thus, all three acetates decompose into their correspond-
ing oxides, which further get reduced by carbon monoxide
to a metallic form. The values for the enthalpy change,
∆H0, and free energy change, ∆G0, are calculated to form
the respective metals from the corresponding oxide[27] in the
presence of carbon monoxide at 700 K. Since the cases of
Ni and Co are similar, the values are calculated for the rep-
resentative Co@C case and, for simplicity, are compared
with the case of Fe3O4@C.
Figure 8. Magnetization vs. magnetic field curves for a) Ni@C,
b) Co@C, and c) Fe3O4@C samples.
not saturate (a small increase in magnetization is measured
between 12–16 kOe), even at 16 kOe. This is due to the
Fe3O4 core, which is comprised of small 5–20 nm crystal-
lites (shown in the TEM measurements). The magnetization
of ferromagnetic materials is very sensitive to the micro-
structure of a sample. If a specimen consists of small par-
ticles, its total magnetization decreases with decreasing par-
ticle size due to the increased dispersion in the exchange
integral,[19] finally reaching the superparamagnetic state,
when each particle acts as a big ‘spin’ with suppressed ex-
change interaction between the particles, which is more hin-
dered because of the presence of the carbon shell. Nanom-
agnetic particles are expected to exhibit reduced magnetiza-
tion due to the large percentage of surface spins with disor-
dered magnetization orientation.[20] In fact, the magnetiza-
tion is smaller than the bare particles, indicating that there
is no magnetic proximity effect between the magnetic (Ni,
Co, or Fe3O4) particles and carbon, which is consistent with
recent experiments.[21]
FeO(s) + CO(s) Ǟ Fe(s) + CO2(g)
∆H0 = –14.7 kJ/mol at 700 K
∆G0 = +3.55 kJ/mol at 700 K
CoO(s) + CO(s) Ǟ Co(s) + CO2(g)
∆H0 = –49.89 kJ/mol at 700 K
∆G0 = –37.61 kJ/mol at 700 K
The main question remaining now is why are Ni and Co
obtained in the metallic form, while Fe tends to form
Fe3O4, maintaining a carbon shell in all three cases. The
The positive value obtained for the ∆G0 of formation of
thermal decomposition of Ni(CH3COO)2 at 350 °C[17] leads Fe from FeII acetate, unlike Ni and Co, indicates that ther-
to dissociation into nickel oxide and oxygenates like acetone modynamically, the reduction to the metallic state is not
and acetic acid, as perceived by TGA coupled with MS. It favorable. Therefore, the alternative is its partial oxidation
is worth mentioning that both, Jesus et al.[17] and our own and the kinetically-controlled coating by the carbon sur-
TGA+MS results show that CO is one of the intermediates rounding it, forming the Fe3O4@C. On the other hand, the
in the decomposition of the Ni acetate. We attribute the ∆G0 value is negative for the formation of metallic cobalt
reduction of nickel oxide to carbon monoxide. The product and Ni, and we assume that the reaction is also fast. Thus
of the reduction is the Ni atom, forming the Ni core, while the core of metallic Co or Ni is formed in this way in Co@C
the oxygenate molecule decomposes to hydrocarbon, and and Ni@C particles.
2094
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 2089–2096

Nanosized Carbon-Coated Metallic Cores
FULL PAPER
area of the samples. An Olympus BX41 (Jobin–Yvon–Horiba) Ra-
man spectrometer is employed, using the 514.5 nm line of an Ar
ion laser as the excitation source to analyze the nature of the car-
bon present in the samples. A vibrating sample magnetometer
(VSM-Oxford-3001) is used for the magnetic susceptibility mea-
surements for Ni@C, Co@C, and Fe3O4@C samples at room tem-
perature.
Conclusions
In summary, using low cost, transition metal acetate pre-
cursors and employing identical reaction parameters, gra-
phitic carbon-coated nanosize Ni, Co are obtained in the
metallic form. Metallic Fe is not obtained, and instead it
forms Fe3O4, maintaining the core-shell morphology. A
broad particle size distribution is observed. Graphite-en-
capsulated Ni, Co, and Fe3O4 fullerene-like core-shell nano-
structures could be obtained at a relatively low temperature
(700 °C) in a closed Swagelok reactor, as compared to other
methods for the formation of graphitic layers. The one-
stage, reproducible, solvent-free, competent and straightfor-
ward approach for the synthesis of Ni@C, Co@C, and
Fe3O4@C core-shell nanostructures can be scaled up.
[1] W. K r ätschmer, L. D. Lamb, K. Fostiropoulos, D. R. Huffman,
Nature 1990, 347, 354.
[2] T. Hayashi, S. Hirono, M. Tomita, S. Umemura, Nature 1996,
381, 772.
[3] A. Leonhardt, M. Ritschel, R. Kozhuharova, A. Graff, T.
Muhl, R. Huhle, I. Monch, D. Elefant, C. M. Schneider, Dia-
mond Relat. Mater. 2003, 12, 790–793.
[4] H. W. Kroto, J. R. Keath, S. C. O’Brien, R. F. Curl, R. E.
Smalley, Nature 1985, 318, 162–165.
[5] L. Guo, Q. J. Huang, X. Y. Li, S. H. Yang, Phys. Chem. Chem.
Phys. 2001, 3, 1661.
[6] T. Hyeon, Chem. Commun. 2003, 8, 927.
[7] S. Gangopadhyay, G. C. Hadjipanayis, B. Dale, C. M. Sor-
ensen, K. J. Klabunde, V. Papaefthymiou, A. Kostikas, Phys.
Rev. B 1992, 45, 9778.
Experimental Section
Synthesis of Fullerene-like Ni@C, Co@C and Fe3O4@C Core-Shell
Nanostructures: Acetates of FeII (Aldrich), CoII (Aldrich) and NiII
(Alfa Aesar), were purchased with Ͼ 99% purity and used as re-
ceived. In the fabrication of Ni@C, Co@C, and Fe3O4@C core-
shell nanostructures, 1 g of the acetate of Ni, Co and Fe are intro-
duced separately into a 3 mL stainless steel reactor at room tem-
perature in a nitrogen-filled glove box. The filled Swagelok reactors
are closed tightly with the other plug.[22] The temperature of the
tube furnace is raised to 700 °C at a rate of 10 °C/min, and the
temperature is maintained at 700 °C for 3 h. The Swagelok-reactors
heated at 700 °C are gradually cooled (ca. 6 h) to room temperature
and opened. A black powders of Ni@C (0.382 g), Co@C (0.400 g)
and Fe3O4@C (0.560 g) is obtained. The obtained products, their
morphology, crystal structure, carbon percent are shown in Table 1.
Earlier, such a one-stage, efficient, and the simplest economic
RAPET (Reactions under Autogenic Pressure at Elevated Tem-
perature) technique was employed for the fabrication of various
interesting nanostructures.[22–24]
[8] R. Sen, A. Govindaraj, C. N. R. Rao, Chem. Phys. Lett. 1997,
267, 276.
[9] C. N. R. Rao, R. Sen, B. C. Satishkumar, A. Govindaraj,
Chem. Commun. 1998, 1525.
[10] B. K. Pradhan, T. Kyotani, A. Tomita, Chem. Commun. 1999,
1317.
[11] A. G. Nasibulin, A. Moisala, D. P. Brown, E. I. Kauppinen,
Carbon 2003, 41, 2711–2724.
[12] J. Jiao, S. Seraphin, X. Wang, J. C. Withers, J. Appl. Phys. 1996,
1, 80.
[13] J. J. Host, J. A. Block, K. Parvin, V. P. Dravid, J. L. Alpers, T.
Sezen, R. LaDuca, J. Appl. Phys. 1998, 2, 83.
[14] a) T. Yogi, T. Nguyen, IEEE Trans. Magn. 1993, 29, 307; b)
S. R. C. Vivekchand, G. Gundiah, A. Govindaraj, C. N. R.
Rao, Adv. Mater. 2004, 16, 1842; c) S. R. Sen, R. Subbanna,
G. N. Kannan, C. N. R. Rao, Chem. Phys. Lett. 1995, 31, 308.
[15] S. Klein, M. Winterer, H. Hahn, Chem. Vap. Deposition 1998,
4, 143.
[16]
S. V. Pol, V. G. Pol, A. Frydman, G. N. Churilov, A. Gedanken,
Characterization: Various analytical techniques are used to identify
and characterize the as-prepared RAPET products of the Ni, Co
and Fe acetate reactions. To understand the purity of the crystal
structure (the interatomic distance, and angle) of the as-prepared
products, the X-ray diffraction patterns are measured with a Bruker
AXS D* Advance Powder X-ray diffractometer [λ(Cu-Kα) =
1.5418 Å]. Since all the acetate contains C, H, and O, in addition
to metal, C, H, N, S and O analyses are carried out for the as-
prepared products on an Eager 200 C,H,N,S analyzer and an
EA1110, oxygen analyzer, respectively. The morphology, grain size,
and chemical composition in a selected area are measured for the
fullerene-like Ni@C, Co@C, and Fe3O4@C core-shell nanostruc-
tures by employing transmission electron microscopy (TEM)
(JEM-1200EX) and high-resolution transmission electron micro-
scopy (HRTEM) (JOEL 2010), working at acceleration voltages of
80 kV and 200 kV, respectively. The morphology and energy disper-
sive X-ray analysis (EDX) of the materials are investigated by high
resolution scanning electron microscopy (HR-SEM, JSM, 7000 F).
The Mössbauer spectroscopy (MS) studies are carried out at room
temp. using a conventional constant acceleration spectrometer. The
57Fe MS are measured with a 50 mCi 57Co:Rh source and the spec-
tra were least square fitted with two sub spectra. The isomer shift
(IS) values are relative to Fe metal at 300 K. A Micromeritics
(Gemini 2375) surface area analyzer is used to measure the surface
J. Phys. Chem. B 2005, 109, 9495.
J. C. De Jesus, I. Gonzalez, A. Quevedo, T. Puerta, J. Mol. Ca-
tal. A 2005, 228, 283.
Y. Lifshitz, Diamond Relat. Mater. 1999, 8, 1659.
S. R. Elliott, Physics of Amorphous Materials, Longman, Lon-
don, New York, 1984, p. 350.
H. R. Kodoma, A. E. Berkovitz, E. J. Mcniff Jr, S. Foner, Phys.
Rev. Lett. 1996, 77, 394.
[17]
[18]
[19]
[20]
[21]
[22]
R. Hohne, M. Ziese, P. Esquinazi, Carbon 2004, 42, 3109.
S. V. Pol, V. G. Pol, A. Gedanken, Chem. Eur. J. 2004, 10, 4467–
4473.
[23]
[24]
a) V. G. Pol, S. V. Pol, A. Gedanken, J. Phys. Chem. B 2005,
109, 6121; b) S. V. Pol, V. G. Pol, V. G. Kessler, G. A. Seisen-
baeva, M. Sung, S. Asai, A. Gedanken, J. Phys. Chem. B 2004,
108, 6322; c) V. G. Pol, S. V. Pol, B. Markovsky, J. M. Calderon-
Moreno, A. Gedanken, Chem. Mater. 2006, 18, 1512; d) S. V.
Pol, V. G. Pol, A. Gedanken, Adv. Mater. 2006, 18, 2023.
a) S. V. Pol, V. G. Pol, G. Seisenbaeva, V. G. Kessler, A. Ged-
anken, Chem. Mater. 2004, 16, 1793; b) V. G. Pol, S. V. Pol, A.
Gedanken, Chem. Mater. 2005, 17, 1797; c) S. V. Pol, V. G. Pol,
V. G. Kessler, G. A. Seisenbaeva, L. A. Solovyov, A. Gedanken,
Inorg. Chem. 2005, 44, 9938; d) V. G. Pol, S. V. Pol, A. Ged-
anken, S. H. Lim, Z. Zhong, J. Lin, J. Phys. Chem. B 2006,
110, 11237.
[25]
G. A. M. Hussein, A. K. H. Nohman, K. A. M. Attiya, J.
Therm. Anal. 1994, 42, 1155.
Eur. J. Inorg. Chem. 2007, 2089–2096
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2095

S. V. Pol, V. G. Pol, I. Felner, A. Gedanken
FULL PAPER
[26] A. J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aque-
ous Solutions, IUPAC, Marcel Dekker, New York, USA, 1985.
[27] M. W. Chase Jr, C. A. Davies, J. R. Downey Jr, D. J. Frurip,
R. A. McDonald, A. N. Syverud, JANAF Thermochemical
Tables, 3rd ed., part I–III, ACS and Chemical Institute of Phys-
ics for the National Bureau of Standards, J. Phys. and Chem.
Reference data, 14, 1985, supplement no. 1.
Received: December 8, 2006
Published Online: April 10, 2007
2096
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 2089–2096