Journal of Magnetism and Magnetic Materials 349 (2014) 3544

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials

journal homepage: www.elsevier.com/locate/jmmm

Synthesis and characterization of iron, iron oxide

and iron carbide nanostructures
Ron Snovski a, Judith Grinblat a, Moulay-Tahar Sougrati b, Jean-Claude Jumas b,
Shlomo Margel a,n
a
b

Institute of Nanotechnology & Advanced Materials, Department of Chemistry, Bar Ilan University, 52900 Ramat Gan, Israel
Institut Charles Gerhardt (UMR 5253 CNRS), Universit Montpellier 2, CC 1502, Place E. Bataillon, 34095 Montpellier Cedex 5, France

art ic l e i nf o

a b s t r a c t

Article history:
Received 22 May 2013
Received in revised form
15 August 2013
Available online 28 August 2013

Magnetic iron oxide (Fe3O4 and -Fe2O3) and iron carbide (Fe3C) nanoparticles of different geometrical
shapes: cubes, spheres, rods and plates, have been prepared by thermal decomposition of a mixture
containing the metal precursor Fe(CO)5 and the stabilizer polyvinylpyrrolidone (PVP) at 300 1C in a
sealed cell under inert atmosphere. The thermal decomposition process was performed for 4 or 24 h at
([PVP]/[Fe(CO)5]) (w/v) ratio of 1:1 or 1:5. Elemental iron nanospheres embedded within a mixture of
amorphous and graphitic carbon coating were obtained by hydrogen reduction of the prepared iron
oxide and iron carbide nanoparticles at 450 1C. The formation of the graphitic carbon phase at such a low
temperature is unique and probably obtained by catalysis of the elemental iron nanoparticles. Changing
the annealing time period and the ([PVP]/[Fe(CO)5]) ratio allowed control of the composition, size, size
distribution, crystallinity, geometrical shape and magnetic properties of the different magnetic
nanoparticles.
& 2013 Elsevier B.V. All rights reserved.

Keywords:
Magnetic nanoparticle
Graphitic carbon
Iron oxide nanoparticle
Iron nanoparticle

1. Introduction
Coated magnetic nanoparticles have numerous potential applications, particularly as recoverable catalysts [1,2] and in the
biomedical sciences [3,4]. For example, functionalized magnetic
nanoparticles have been employed for site-specic drug delivery
[5,6]. The shape, size and size distribution of the magnetic
materials are the key factors that determine their chemical and
physical properties [7]. Thus, the development of size and shapecontrolled magnetic materials has become very important for
different uses. So far, the most widely used and studied magnetic
material is iron oxide, in the form of magnetite (Fe3O4) and
maghemite (-Fe2O3). The applications of these iron oxide particles rely upon the biodegradability and biocompatibility of the
iron oxide phase [8,9]. Elemental iron on the other hand has a
signicantly higher magnetic moment than its oxides. Moreover,
elemental iron is the most useful among the ferromagnetic
elements; it has the highest magnetic moment at room temperature, and a curie temperature which is high enough for the vast
majority of practical applications. In addition, iron is a widespread

Corresponding author. Tel.: 972 3 5318861; fax: 972 3 7384053.

E-mail addresses: ron.snovski@gmail.com (R. Snovski),
grinby@mail.biu.ac.il (J. Grinblat), Moulay-Tahar.Sougrati@univ-montp2.fr
(M.-T. Sougrati), jumas@univ-montp2.fr (J.-C. Jumas),
shlomo.margel@mail.biu.ac.il (S. Margel).
0304-8853/$ - see front matter & 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jmmm.2013.08.043

element, and therefore signicantly cheaper than other ferromagnetic elements such as nickel and cobalt. However, Fe particles are
easily oxidized leading to a signicant reduction in their magnetic
moment. In order to prevent this phenomenon, Fe particles should
be protected by a protective layer, such as carbon, [1013] silica
[14,15] or alumina [16]. Recently, a few publications described
the synthesis of Fe magnetic silica particles, by entrapment of
iron nitrate within the mesopores of silica particles, followed
by impregnation with ethylene glycol and then annealing at
450 1C [14].
An attractive alternative to elemental iron, offering both high
magnetic saturation and chemical stability, is iron carbide (Fe3C) [17].
As with many metal carbide materials, the synthesis of Fe3C nanoparticles presents a considerable challenge. Many of the wellestablished routes to metal or metal oxide nanoparticles, such as
microemulsion or coprecipitation synthesis, are unsuitable for
metal carbide formation. The synthesis of phase-pure Fe3C is
particularly difcult, since the relatively high-temperature and
reducing conditions characteristic of many carbide syntheses often
result in the formation of metallic iron or mixed Fe and Fe3C
products [18,19]. While some routes, such as ame spray pyrolysis,
[20] have been used to synthesize phase-pure Fe3C; most current
methods involve multiple-step or high-energy procedures using
hazardous or costly chemical precursors [21,22].
Magnetic nanoparticles tend to aggregate due to strong magnetic dipoledipole and van der Waals attraction. Thus, the main

36

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

challenge in the synthesis is to overcome these aggregation

phenomena. This is usually done by coating the nanoparticles'
surface with desired functional polymers or surfactants. The
surfactant also plays a role in the nucleation process and in
limiting particle growth. The functional groups belonging to the
organic coating also allow the binding of biological molecules,
such as proteins, oligonucleotides, etc., to the particle's surface for
biomedical applications [23]. Iron and iron oxide nanoparticles are
typically prepared by the decomposition of soluble iron precursors
in solutions containing an appropriate stabilizer. The decomposition of the iron precursors is accomplished by means of processes
such as sonochemistry, [2427] thermal decomposition, [28,29]
electrochemical [30] and laser decomposition [31]. Among the iron
precursors, iron carbonyl compounds are the most useful ones,
since they decompose and release CO molecules, which can easily
be removed from the reaction mixture. Recently, several ways to
prepare iron oxide and elemental iron nanoparticles using different iron precursors and surfactants have been reported [3235].
The present manuscript describes a simple one-step route to
synthesize iron oxide and iron carbide nanoparticles of different
shapes. Elemental iron nanospheres embedded within graphitic
carbon matrix are obtained by hydrogen reduction of the prepared
iron oxide and iron carbide nanoparticles at 450 1C.
2. Materials and methods
2.1. Materials
The following analytical-grade chemicals were purchased from
Aldrich (Israel) and were used without further purication: Fe
(CO)5 ( 499%) and polyvinylpyrrolidone (PVP, MW 360,000).
Water was puried by passing deionized water through the
Elgastat Spectrum reverse osmosis system (Elga, High Wycombe,
UK). The following gases were purchased from Israel Oxygen and
Argon Factories Inc.: Ar and H2, 99.999% purity.
2.2. Synthesis of uniform iron oxide and iron carbide nanoparticles
of different shapes
200 mg of PVP were weighed into a letlock followed by the
addition of 0.2 or 1.0 ml of Fe(CO)5 (1:1 or 1:5 w/v ratio,
respectively). The letlock was carefully ushed for 2 min by Ar
(10 ml/min ow rate) and then sealed in order to prevent Fe(CO)5
evaporation. The different mixtures were thermally decomposed
in a pre-heated tube furnace at 300 1C for 4 or 24 h. The mixtures
were then left to cool to room temperature.
2.3. Synthesis of uniform elemental iron nanoparticles embedded
in carbon
Elemental iron nanospheres of 290.1 7247.1 nm size embedded
in a mixture of amorphous and graphitic carbon were formed by
reducing the iron oxide and iron carbide nanoparticles, prepared
by thermal decomposition for 4 h of the mixture of PVP and Fe
(CO)5 at ([PVP]/[Fe(CO)5]) ratio of 1:1 (w/v) with H2 (20 ml/min
ow rate) at 450 1C for 2.5 h. The H2 reduction reaction was
performed in a combustion boat containing the mentioned nanoparticles. The obtained carbon-coated nanospheres were then left
to cool to room temperature.
2.4. Characterization of the different magnetic nanoparticles
High-resolution transmission electron microscope (HRTEM)
images were obtained by employing a 200 kV, JEOL-2100 device
with integrated digital scanning transmission electron microscope

(STEM) (annular dark and bright eld detectors). Samples for

HRTEM were prepared by placing a drop of the diluted sample
on a 400-mesh carbon-coated copper grid. The particles' average
size and size distribution of the electronic images were determined by measuring the diameter of approximately 200 particles
with an image analysis software, AnalySIS Auto (Soft Imaging
System GmbH, Germany). The sizes of the differently shaped
nanoparticles were measured as follows: nanospheres were measured by diameter, nanocubes by the length of their diagonal and
nanorods and nanoplates according to their width. Raman spectroscopy (HORIBA Jobin Yvon) measurements were carried out by
using an Ar ion laser (power10 mW and wavelength632.8 nm).
The exposure and data acquisition times were 20 and 2 s, respectively. C, H and O weight % of the various particles were measured
using an elemental analysis instrument, model FlashEA1112
Instruments, Thermoquast. Fe weight % was calculated by subtracting the sum of the other elements from 100.
Powder X-ray diffraction (XRD) patterns were recorded using
an X-ray diffractometer (model D8 Advance, Bruker AXS) with
Cu K radiation. Isothermal magnetization measurements at room
temperature were performed in a commercial (Quantum Design)
super-conducting quantum interference device (SQUID) magnetometer. The Mssbauer spectroscopy measurements were performed by putting the samples' powders (100 mg each) on a
PMMA holder of 16 mm diameter without any binder. The powders could be recovered if necessary. 57Fe Mssbauer spectra were
then recorded in the constant acceleration mode and in transmission geometry on a standard Mssbauer spectrometer. A 57Co(Rh)
source with a nominal activity of 130 MBq was used. The source
and the absorber were always kept at room temperature. The
isomer shift is given with respect to -Fe standard at room
temperature. Data renement has been performed using the
Lorentzian lines. The thermal behavior of the particles was measured
with a TC15 system equipped with TGA (thermal gravimetric analysis), model TG-50, from Mettler Toledo. C, H, N and
O analyses of the various particles was performed using an
elemental analysis instrument, model FlashEA1112 Instruments,
Thermoquast.

3. Results and discussion

In this work, we use the term nanospheres for truly spherical
particles, whereas the term nanoparticles is used for the particles that, while they are rounded in shape, are not perfectly
spherical. The size and size distribution and the different geometrical shapes of the nanoparticles are summarized in Table 1. The
different shapes are presented in Figs. 15. The sizes of the
nanocubes and nanoparticles, obtained after thermal decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) (w/v)
ratio of 1:1 at 300 1C for 4 h under Ar atmosphere, are 54.5 76.8
and 43.1 713.0 nm, respectively. Following H2 reduction of iron
oxide, a change in shape to nanospheres and a signicant increase
in particle size to 290.1 7247.1 nm was observed. This increase in
size probably occurred due to agglomeration. Thermal decomposition of a mixture of PVP and Fe(CO)5 was also performed at a ratio
of 1:5 for 24 h, also at 300 1C. The increase in Fe(CO)5 relative
concentration and duration of thermal decomposition resulted in
formation of iron oxide nanoplates with a width of 120.0 7
27.4 nm. Previous Fourier-transform infrared (FTIR) spectrometry
studies have shown that PVP molecules may coordinate with
metal ions to form stable metalPVP complex molecules [36,37].
This complex formation may also explain the changes in the
different geometrical shapes. Note that rather than obtaining a
mixture of different shapes, each composite obtained in each
experiment was produced with its own characteristic shape.

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

37

Table 1
Size and size distribution, shape, Ms and coercivity values of the different nanoparticles obtained by annealing a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) (w/v) ratios
of 1:1 and 1:5 for 4 and 24 h.
[PVP]/[Fe(CO)5] (w/v)

Annealing time (h)

Temperature (1C)

1:1

300

1:1

24

300

1:5

300

1:5

24

300

H2 reduction of 1:1 for 4 h

2.5

450

Particles obtained

Shape

Diameter (mm)

Fe3O4/-Fe2O3
Fe3C
Fe3O4/-Fe2O3
Fe3C
Fe3O4/-Fe2O3
Fe3C
Fe3O4/-Fe2O3
Fe3C

Nanocubes
Nanoparticles
Nanoparticles
Nanorods
Nanoparticles
Nanorods
Nanoplates
Nanorods

54.5 7 6.8
43.17 13.0
8.75 7 1.7
10.5 7 3.7
40.17 15.6
14.0 7 1.8
120.0 7 27.4
39.8 7 6.8

Fe

Nanospheres

290.17 247.1

Ms (emu g  1)

Coercivity (Oe)

300 K

300 K

5K

5K

23.4

20.8

10.6

17.3

38.8

43.4

15.6

35.0

52.3

60.7

24.3

54.5

92.2

96.0

37.0

77.9

176.1

179.0

21.2

26.7

Fig. 1. Bright eld image of typical iron oxide nanocubes and Fe3C nanoparticles obtained after thermal decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5])
(w/v) ratio of 1:1 for 4 h as described in the experimental section (A); HRTEM images of a nanocube (B) and of a nanoparticle (D), displaying the well resolved lattice-fringes
of -Fe2O3 or Fe3O4; and (C) and (E) are the FFTs of the portion of the image outlined by the white square which, like a diffraction pattern, were indexed on the basis of the
unit cell of -Fe2O3 or Fe3O4 (B) and the unit cell of Fe3C (D).

HRTEM studies combined with the transmission electron

microscope (TEM) mode were performed in order to provide
evidence for the structure of the differently obtained nanoparticles. The TEM mode was performed using the conventional
selected-area electron diffraction (SAED) technique, nano-beam
diffraction (NBD) and fast Fourier transform (FFT) analysis of the
high resolution images. The resulting FFT of a high-resolution
image is geometrically equivalent to a diffraction pattern; hence

the FFT pattern contains information about the internal crystalline

structure of the analyzed area of the sample. The reections in a
FFT pattern can be analyzed in the same manner as the reections
in a diffraction pattern. Finally, the different structures of the
compounds were indexed and identied on the basis of the unit
cell of the FFC structure of -Fe2O3 (maghemite) or Fe3O4 (magnetite) (pdf # 04-0755 for -Fe2O3 a 8.350 and pdf # 11-0614
for Fe3O4 a 8.396 ) and that of the orthorhombic structure of

38

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

Fig. 2. Bright eld images of representative -Fe2O3 or Fe3O4 nanoparticles (A) and typical nanorods of Fe3C (E) obtained after thermal decomposition of a mixture of PVP
and Fe(CO)5 at ([PVP]/[Fe(CO)5]) (w/v) ratio of 1:5 for 4 h as described in the experimental section; typical bright eld image comprising both nanoparticles of -Fe2O3 or
Fe3O4 and nanorods of Fe3C (B); inset (C) is the DF image of (B), the bright spots are the crystalline -Fe2O3 or Fe3O4; NBD taken from a 4 nm area of a nanoparticle of iron
oxide (D). Inset (F) is the DF image of (B) taken in the (200 and 210) directions of the iron carbide nanorods, showing the spatial distribution of the iron carbide phase in the
image; HRTEM image of a Fe3C nanorod (G); and inset (H) is the SAED taken from image (E).

Fig. 3. Bright eld image of representative nanoparticles of -Fe2O3 or Fe3O4 and nanorods of Fe3C obtained after thermal decomposition of a mixture of PVP and Fe(CO)5 at
([PVP]/[Fe(CO)5]) (w/v) ratio of 1:1 for 24 h as described in the experimental section (A); inset (C) is the SAED taken from an area with a 200 nm diameter of image (B).
HRTEM image of a Fe3C nanorod (D); and inset (E) is the FFT of the portion of the image outlined by white square in (D).

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

39

Fig. 4. Bright eld image of representative -Fe2O3 or Fe3O4 nanoplates and Fe3C nanorods obtained by thermal decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe
(CO)5]) ratio of 1:5 for 24 h as described in the experimental section (A); inset (B) is the FFT of the portion of the image outlined by the white square in the HRTEM image
(C) of the Fe3C; HRTEM image of a Fe3C nanorod (D); and inset (E) is the FFT of the portion of the image outlined by white square in the HRTEM inset (F) of a iron oxide
nanoplate.

Fe3C (pdf # 35-0772, a5.091 , b6.743 , c4.526 , S.G. Pnma

(62)). As both oxides are structurally similar, they cannot be
distinguished according to their electron diffraction pattern or
their FFT images. Hence we may presume that either one or both
of the oxides are present in the different samples. However,
Mssbauer spectroscopy was performed in order to distinguish
between the two oxides, as shown in Fig. 10.
Fig. 1(A) is a typical TEM micrograph of the iron oxide
nanocubes with size and size distribution of 54.5 7 6.8 nm, and
Fe3C nanoparticles with size and size distribution of 43.1 7
13.0 nm (see Table 1), obtained after thermal decomposition of a
mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) (w/v) ratio of 1:1
at 300 1C for 4 h under Ar atmosphere, as described in the
experimental section. A thin organic coating layer, probably of
PVP, can be seen to be coating the nanocubes and nanoparticles.
The set of reections in the FFT pattern (C), taken from the marked
area of HRTEM image of a nanocube (B), refers to the FCC structure
of both -Fe2O3 and Fe3O4. The d111 and d200 family of planes are
marked in the image. In the same manner, the sets of reection
present in the FFT pattern (E) of the HRTEM image of the
nanoparticle (D) were found to match the d020, d220 and d200
planes in the orthorhombic structure of Fe3C.
Fig. 2(B) is a typical bright eld (BF) image showing the
presence of the -Fe2O3 or Fe3O4 nanoparticles with size and
size distribution of 8.757 1.7 nm, and Fe3C nanorods with size
and size distribution of 10.5 73.7 nm, obtained after thermal

decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe

(CO)5]) (w/v) ratio of 1:5 at 300 1C for 4 h under Ar atmosphere,
as described in the experimental section. A thin organic coating
layer can be seen to be coating the nanoparticles and nanorods.
Inset (A) exhibits the PVP coating on these nanoparticles. Probing
the local structure of these nanoparticles using the NBD technique
indicates the presence of iron oxide. Example of NBD pattern
(D) taken with a 4 nm spot-size from the marked area of the
nanoparticle from inset (A) shows that it contains sets of reections which can be uniquely indexed on the basis of the FFC
structure of -Fe2O3 or Fe3O4. The dendrite, the typical multibranching tree-like form in inset (E), was found to be Fe3C as
supported by SAED (H). (C) and (F) are the dark-eld (DF) images
of iron oxide (B) and iron carbide (E), respectively. In the DF mode
of imaging technique, only specic sets of reections (certain
diffracted beams caused by a specic crystalline element) make
up the image, so that the diffracting areas will appear as bright
spots in the DF image. In each case, the DF images were taken
using the reections unique to the iron oxide phase or to the iron
carbide phase. DF image (C) (associated with the bright-eld (BF)
image (B)) was taken from the (220) direction of the cubic iron
oxide, so that it displays the spatial distribution of iron oxide in
inset (B). Similarly the DF image (F) (associated with the BF image (E))
was taken in the (200 and 210) directions of the orthorhombic
structure of Fe3C, and shows the spatial distribution of the iron
carbide in inset (E). The SAED (H), taken with a 300 nm diameter

40

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

Fig. 5. Bright eld image of Fe nanospheres embedded into a thin layer of a mixture of amorphous and graphitic carbon formed according to the experimental section (A);
HRTEM image of a single nanosphere in a shell of the carbon layer (B); and FFT of the portion of the image outlined by white square in (B) showing the d110 family of
planes (C). Inset (D) is the line prole displaying the measurements for four d-spacings (1.36 nm) from the area marked by the white rectangular in (B).

Fig. 6. Raman spectrum of the mixture of the amorphous and graphitic carbon
layer containing the Fe nanospheres formed according to the experimental section.

Fig. 7. TGA thermogram of PVP.

aperture in image (E), shows the reections (200), (210), (031) and
(113) that were indexed in terms of the orthorhombic structure of
Fe3C. Inset (G) is a HRTEM image of a single Fe3C nanorod.
Fig. 3(A) is a typical BF image exhibiting the multi-structure
nature of the iron oxide (-Fe2O3 or Fe3O4) nanoparticles with size
and size distribution of 40.1 715.6 nm, and Fe3C nanorods with
size and size distribution of 14.0 71.8 nm, obtained after thermal
decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe
(CO)5]) (w/v) ratio of 1:5 at 300 1C for 4 h under Ar atmosphere,

as described in the experimental section. A thin organic coating

layer can be seen to be coating the nanoparticles and nanorods.
The SAED (C) taken with a 200 nm diameter aperture in inset
(B) demonstrates ring-pattern electron diffraction. The calculated
planes (111), (220), and (311) were indexed on the basis of the FCC
structure of both -Fe2O3 and Fe3O4. HRTEM image of a Fe3C
nanorod is seen in inset (D). The FFT (E) of the portion of the image
outlined by the white square in (D) contains a set of reections
which can be uniquely indexed on the basis of the unit cell of the
orthorhombic structure of Fe3C.

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

The bright eld image in Fig. 4(A) exhibits some typical iron
oxide (-Fe2O3 or Fe3O4) nanoplates with size and size distribution
of 120.0 727.4 nm, and Fe3C nanorods with size and size distribution of 39.8 7 6.8 nm, obtained after thermal decomposition of a
mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) (w/v) ratio of 1:5
at 300 1C for 24 h, under Ar atmosphere as described in the
experimental section. A thin organic coating layer can be seen to
be coating the nanoplates and nanorods. The structures of the
compounds were identied using FFT analysis of the HRTEM
images of a Fe3C nanorod (C) and of an iron oxide nanoplate (F).
The set of reections that appear in the FFT in inset (B) (d111, d121
and d210) were indexed on the basis of the unit cell of the
orthorhombic structure of Fe3C. The FFT (E) related to the
nanoplate (F) was uniquely indexed as iron oxide (family of planes
d222 and d400). As can be seen in the HRTEM image of a single
nanorod (D), lattice fringes with interplanar distances of 6.6
were measured and correspond to (010) plane of Fe3C.
Fig. 5(A) shows the BF image of typical Fe nanospheres
embedded into a thin layer of a mixture of amorphous and
graphitic carbon. The Fe nanospheres obtained by H2 reduction
of the previously-mentioned iron oxide nanocubes and iron
carbide nanoparticles (produced by thermal decomposition of a
mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) ratio of 1:1 at
300 1C for 4 h under inert atmosphere as described in the experimental section) at 450 1C for 2.5 h as described in the experimental section. In the HRTEM image (B) the d110 lattice of Fe can
be seen (d 0.203 nm), in spite of the 4 nm carbon shell coating.
In order to measure this lattice, FFT analysis (C) was applied to
the area marked by the white square and shows the d110 (d
0.203 nm) lattice of the unit cell for elemental iron, (pdf # 011262,
a 2.857, S.G. Im3m). The line prole (D) displays the measurements for four d-spacings (1.36 nm) from the area marked by the
white rectangle in inset (B). The calculated d-spacing from this
line prole is 0.34 nm, matching well the expected spacing for the
(00.2) family of planes in the hexagonal structure of graphitic
carbon (pdf # 26-1079, a 2.456, c 10.04 S.G. R3). The obtained
graphitic carbon phase is quite unique due to the low temperature
of 450 1C that was maintained in this experiment.
Other evidence of the existence of mixture of amorphous and
graphitic carbon coating can be found in the Raman spectrum of
Fig. 6. The D-band peak (disordered layer) can be seen at
1329 cm  1 and the G-band (graphitic layer) can be seen at
1580 cm  1. It should be noted that amorphous and graphitic
carbon coatings were not observed by Raman spectroscopy for
iron oxide and iron carbide different nanoparticles obtained after
thermal decomposition of PVP and Fe(CO)5 mixtures at 300 1C.
Fig. 7 shows the TGA curve of PVP. This curve exhibits a steep
slope between 350 and 440 1C, indicating 89% weight loss due to
PVP decomposition, leaving residual carbon. This TGA behavior
indeed indicates that at 300 1C where the different mixtures of
PVP and Fe(CO)5 annealed, the PVP is stable while Fe(CO)5
decomposes to iron oxides and Fe3C. Therefore, it is reasonable
to assume that the organic thin coatings, surrounding the different
nanoparticles shown in the HRTEM pictures of Figs. 14, are
composed of a PVP layer. On the other hand, at 450 1C where the
H2 reduction was performed, iron oxides and Fe3C were reduced to
elemental iron protected by a carbon layer produced by the
decomposition of PVP. Elemental analysis measurements conrmed that PVP is stable at 300 1C for 24 h under Ar atmosphere
as was demonstrated by C, H, N and O analyses.
Fig. 8 illustrates the XRD pattern of iron oxide, iron carbide and
the elemental iron nanoparticles obtained by thermal decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) ratio of
1:1 at 300 1C for 4 h, followed by H2 reduction at 450 1C for 2.5 h
as described in the experimental section. Similar XRD patterns
were observed for the other decomposed mixtures of PVP and Fe

41

Fig. 8. XRD patterns of the iron oxide and iron carbide nanoparticles prepared by
thermal decomposition of a mixture of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) ratio of
1:1 at 300 1C for 4 h under inert atmosphere according to the experimental section
(A) and of the elemental Fe nanospheres embedded in a mixture of amorphous and
graphitic carbon prepared by H2 reduction of these nanoparticles at 450 1C as
described in the experimental section (B).

(CO)5. Pattern (A) represents a typical spectrum of the iron oxide

and iron carbide nanoparticles. The peaks in this spectrum at
2 35.4 (311), 2 43 (400) and 2 62.5 (440) correspond to
the crystal plane of iron oxide (Fe3O4 and -Fe2O3). A representative peak of iron carbide can be seen at 2 54.582 (230). This iron
carbide peak is quite weak and wide, probably due to the relatively
low crystallinity of the particles obtained at the low temperature
of 300 1C. The relatively wide peaks of the iron oxide also indicate
the semi-crystalline structure of these particles. In order to
distinguish between Fe3O4 and -Fe2O3 Mssbauer spectroscopy
was performed, as presented in Fig. 10. Pattern (B) refers to zerovalent iron peaks obtained by the H2 reduction of the former iron
oxide and iron carbide nanoparticles. The peaks demonstrate the
formation of a body-centered cubic (bcc) Fe and are at 2 44.6
(110) and 2 65 (200). All the iron oxide and iron-carbide phases
have been transformed completely into a zero-valent iron phase,
and carbide has been transformed to amorphous and graphitic
carbon within which the elemental iron is embedded. This can be
explained by the following reactions:
Fe3O4 4H2-3Fe4H2O
Fe2O3 3H2-2Fe3H2O
2Fe3C 2H2-6FeC CH4
Fig. 9 presents the magnetization versus magnetic eld at
300 K (I) and 5 K (II) and magnetization versus temperature (III)
at the applied external magnetic eld (10 Oe) of the different
nanoparticles, obtained by thermal decomposition of the various
mixtures of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) ratio and annealing duration time, respectively, of: 1:1 for 4 h (A), 1:1 for 24 h (B),
1:5 for 4 h (C), and 1:5 for 24 h (D), at 300 1C under Ar atmosphere

42

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

Fig. 9. Magnetization versus magnetic eld at 300 K (I) and 5 K (II), and magnetization versus temperature (III) at the applied external magnetic eld (100 Oe) of the
nanoparticles obtained by thermal decomposition of the various mixtures of PVP and Fe(CO)5 at ([PVP]/[Fe(CO)5]) ratio and annealing duration time, respectively of: 1:1 for
4 h (A), 1:1 for 24 h (B), 1:5 for 4 h (C) and 1:5 for 24 h (D) at 300 1C under Ar atmosphere according to the experimental section. Curve (E) represents the magnetization
curve of the elemental iron nanoparticles obtained by H2 reduction of the nanoparticles belonging to curve (A) (1:1 for 4 h), prepared as described in the experimental
section.

according to the experimental section. Curve (E) represents the

magnetization curve of the elemental iron nanoparticles obtained
by H2 reduction of the nanoparticles belonging to curve (A) (1:1
for 4 h), prepared as described in the experimental section. The
nanoparticles show ferromagnetic-type curves due to the presence
of the hysteresis loop. The magnetic susceptibility (Ms) and
coercivity values obtained at 300 K (I) and 5 K (II) are summarized
in Table 1. The Ms value increases as the ratio [PVP]/[Fe(CO)5]
decreases and the duration of the thermal decomposition
increases. For example, decreasing the [PVP]/[Fe(CO)5] ratio from
1:1 and annealing time of 4 h (A) and 24 h (B) to a ratio of 1:5 and
annealing time of 4 h (C) and 24 h (D) increases the Ms value at
300 K (I) from 23.4 emu g  1 to 38.8, 52.3 and 92.2 emu g  1,
respectively and at 5 K (II) from 20.8 emu g  1 to 43.4, 60.7 and
96.0 emu g  1, respectively. As expected, the coercivity values of
the different nanoparticles measured at 5 K (II) are higher than
that measured at 300 K (I). Therefore, we can say for sure once
again that these nanoparticles are ferromagnetic. For example,
decreasing the magnetization measurement temperature of the
nanoparticles from 300 K (I) to 5 K (II) results in increasing the
coercivity values from 10.6 (A), 15.6 (B), 24.3 (C) and 37.0 (D) Oe to
17.3 (A), 35.0 (B), 54.5 (C) and 77.9 (D) Oe, respectively. It is clearly
shown in graph (III) which represents the magnetization versus
temperature of the different nanoparticles that the two branches
(ZFC and FC) of each curve do not merge at any point below room
temperature. This result proves that the nanoparticles are ferromagnetic and not superparamagnetic, since superparamagnetism
is a phenomenon in which hysteresis does not occur at temperatures in the range of room temperature and above [7]. An additional

Table 2
Percentage of phases according to Mssbauer spectroscopy of the different
nanoparticles obtained by annealing mixtures of PVP and Fe(CO)5 at [PVP]/[Fe
(CO)5] ratios of 1:1 and 1:5 for 4 and 24 h as described in the experimental section.
[PVP]/[Fe(CO)5] (w/v)

1:1
1:1
1:5
1:5

for
for
for
for

4h
24 h
4h
24 h

H2 reduction of 1:1 for 2.5 h

% Phases
-Fe2O3

Fe3O4

Fe3C

Fe

25.0
15.3
29.9
27.8

47.3
43.8
20.7
15.1

27.7
40.9
49.4
57.1

1.4

7.3

91.3

evidence for the ferromagnetic behavior of these nanoparticles is

due to the fact that the gap between the ZFC branch to the FC
branch of each curve (A)(D) increases in the same manner as the
coercivity values of the nanoparticles increases. Magnetization
curves (A)(D) represent iron oxide and iron carbide nanoparticles. The literature-reported Ms bulk values of Fe3O4, -Fe2O3 and
Fe3C are 92, 87 and 140 emu g  1, respectively [38]. Therefore,
as the amount of Fe3C phase increases (summarized in Table 2) the
Ms value increases. The relatively low magnetic moment of these
samples can be explained by the PVP coating and relatively low
crystallinity of the different nanoparticles as observed in the XRD
spectrum in Fig. 8(A). Curve (E) (I) represents Ms value of
176.1 emu g  1 for the obtained elemental iron nanospheres
embedded within amorphous and graphitic carbon. This value is
equivalent to the elemental iron bulk value [38] of 222 emu g  1

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

43

represent different phases as seen in the gure key. The dotted

line represents the experimental results. The black curve represents the total t. -Fe2O3 phase is represented by the red curve,
the Fe3O4 phase by the blue and brown curves, the Fe3C phase by
the green curve, [3941] and elemental Fe by the pink curve.
Fig. 10(A)(D) reveals two sextets that are characteristic of Fe3O4
(blue and brown curves) [42,43]. The red doublet have an isomer
shift of IS 0.32 mm s  1 and a quadrupole of QS 0.76 mm s  1
which are characteristic of paramagnetic or superparamagnetic
Fe3 as reported for nanoparticles of -Fe2O3. The blue sextet
(higher eld) is assigned to the Fe3 sites on the tetrahedral sites
whereas the brown sextet is assigned to the Fe2.5 sites. These
sextets have approximately IS0.29 and 0.69 mm s  1, and
Heff 500 and 469 kOe, respectively, which are related to the two
crystallographic sites of Fe3O4. Spectrum (E) is composed mainly of
well resolved sextet (pink) which is unambiguously assigned to
elemental Fe (more than 90%) with 7% of Fe3O4 and some traces of
-Fe2O3. These iron oxides are probably obtained due to slight
oxidation of Fe on the outermost surface. In a rst approximation
by assuming the LambMssbauer factors for the involved phases,
the amounts of the different phases of all the samples are
summarized in Table 2. This table indicates that at an annealing
temperature of 300 1C, decreasing the ([PVP]/[Fe(CO)5]) ratio and
increasing the duration time of the decomposition, did not
signicantly affect the content of the -Fe2O3 phase, and leads to
a decrease in the content of the Fe3O4 phase and an increase in the
content of the Fe3C phase. Table 2 also indicates that H2 annealing
of the iron oxide and iron carbide phases at 450 1C for 2.5 h leads
to the formation of almost pure elemental Fe phase (91.3% Fe, 1.4%
-Fe2O3 and 7.3% Fe3O4) of a spherical shape.

4. Conclusions

Fig. 10. Mssbauer spectra at room temperature of the different nanoparticles

obtained by thermal decomposition of the various mixtures of PVP and Fe(CO)5 at
([PVP]/[Fe(CO)5]) ratio and annealing duration time, respectively, of: 1:1 for 4 h (A),
1:1 for 24 h (B), 1:5 for 4 h (C), and 1:5 for 24 h (D), at 300 1C under Ar atmosphere
according to the experimental section. Curve (E) represents the magnetization
curve of the elemental iron nanoparticles obtained by H2 reduction of the
nanoparticles belonging to curve A (1:1 for 4 h), prepared as described in the
experimental section. The different phases are represented by different colors as
seen in the key gure. (For interpretation of the references to color in this gure,
the reader is referred to the web version of this article.)

since these elemental iron nanoparticles represents 79.2% of the

sample and the carbon layer represents 19.7% of the sample, as
measured by elemental analysis.
Additional quantitative information related to the different Fe
phases obtained in the present work was obtained by Mssbauer
spectroscopy. Fig. 10 illustrates the Mssbauer spectroscopy spectra at room temperature of the different nanoparticles obtained at
([PVP]/[Fe(CO)5]) ratios of 1:1 and 1:5 for 4 and 24 h according to
the experimental section. The different colors of the curves

This study demonstrates a simple thermal decomposition of a

mixture of PVP and Fe(CO)5 at 300 1C under inert atmosphere in a
sealed cell, for the preparation of magnetic iron oxide (Fe3O4 and
-Fe2O3 ) and iron carbide (Fe3C) nanoparticles of different shapes:
cubes, spheres, rods and plates. Elemental iron nanospheres
embedded within a mixture of amorphous and graphitic carbon
were formed by H2 reduction at 450 1C of the former nanoparticles. The different nanoparticles' size and size distribution, composition, crystallinity, geometrical shape and magnetic properties
were controlled by changing the ratio between the reactants and
the thermal decomposition duration. The formation of the graphitic carbon phase at such a low temperature as 450 1C is a unique
phenomenon [44,45] and probably occurs by catalysis [46] of the
elemental iron nanoparticles. The different nanoparticles preserve
their magnetic properties for at least 1 year. In a future work,
attempts to synthesize nanoparticles of pure phases and shapes
will be accomplished. Attempts to synthesize various other magnetic nanoparticles, using different stabilizers and metal precursors mixtures, will also be done. These magnetic composite
nanoparticles may then be used for various medical and nonmedical applications [1,3,5,47,48].

Acknowledgments
This study was partially supported by a Minerva Grant (Microscale & nanoscale particles & thin lms for medical applications).
References
[1] D.K. Yi, S.S. Lee, J.Y. Ying, Synthesis and applications of magnetic nanocomposite catalysts, Chemistry of Materials 18 (2006) 24592461.

44

R. Snovski et al. / Journal of Magnetism and Magnetic Materials 349 (2014) 3544

[2] S. Wei, Z. Dong, Z. Ma, J. Sun, J. Ma, Palladium supported on magnetic

nanoparticles as recoverable catalyst for one-pot reductive amination of
aldehydes with nitroarenes under ambient conditions, Catalysis Communications 30 (2013) 4044.
[3] A.H. Lu, E.L. Salabas, F. Schuth, Magnetic nanoparticles: synthesis, protection,
functionalization, and application, Angewandte Chemie-International Edition
46 (2007) 12221244.
[4] Y.H. Liu, J. Deng, J.W. Gao, Z.H. Zhan, Triic acid-functionalized silica-coated
magnetic nanoparticles as a magnetically separable catalyst for synthesis of
gem-dihydroperoxides, Advanced Synthesis and Catalysis 354 (2012) 441447.
[5] J.H. Lee, Y.M. Huh, Y. Jun, J. Seo, J. Jang, H.T. Song, S. Kim, E.J. Cho, H.G. Yoon,
J.S. Suh, J. Cheon, Articially engineered magnetic nanoparticles for ultrasensitive molecular imaging, Nature Medicine 13 (2007) 9599.
[6] A.K.A. Silva, C. Wilhelm, J. Kolosnjaj-Tabi, N. Luciani, F. Gazeau, Cellular transfer
of magnetic nanoparticles via cell microvesicles: impact on cell tracking by
magnetic resonance imaging, Pharmaceutical Research 29 (2012) 13921403.
[7] D.L. Huber, Synthesis, properties, and applications of iron nanoparticles, Small
1 (2005) 482501.
[8] J.W.M. Bulte, D.L. Kraitchman, Iron oxide MR contrast agents for molecular and
cellular imaging, NMR in Biomedicine 17 (2004) 484499.
[9] R.R.T. Ragheb, D. Kim, A. Bandyopadhyay, H. Chahboune, B. Bulutoglu,
H. Ezeldein, J.M. Criscione, T.M. Fahmy, Induced clustered nanoconnement
of superparamagnetic iron oxide in biodegradable nanoparticles enhances
transverse relaxivity for targeted theranostics, Magnetic Resonance in Medicine (2013) 10.1002/mrm.24622.
[10] N. Shpaisman, S. Margel, Synthesis and characterization of air-stable iron
nanocrystalline particles based on a single-step swelling process of uniform
polystyrene template microspheres, Chemistry of Materials 18 (2006)
396402.
[11] R. Snovski, J. Grinblat, S. Margel, Synthesis and characterization of magnetic
poly(divinyl benzene)/Fe3O4, C/Fe3O4/Fe, and C/Fe onionlike fullerene
micrometer-sized particles with a narrow size distribution, Langmuir 27
(2011) 1107111080.
[12] R. Snovski, J. Grinblat, S. Margel, Novel magnetic Fe onion-like fullerene
micrometer-sized particles of narrow size distribution, Journal of Magnetism
and Magnetic Materials 324 (2012) 9094.
[13] A. Schaetz, M. Zeltner, T.D. Michl, M. Rossier, R. Fuhrer, W.J. Stark, Magnetic
silyl scaffold enables efcient recycling of protecting groups, ChemistryA
European Journal 17 (2011) 1056510572.
[14] T. Valdes-Solis, A.F. Rebolledo, M. Sevilla, P. Valle-Vigon, O. Bomati-Miguel, A.
B. Fuertes, P. Tartaj, Preparation, characterization, and enzyme immobilization
capacities of superparamagnetic silica/iron oxide nanocomposites with
mesostructured porosity, Chemistry of Materials 21 (2009) 18061814.
[15] P.-L. Liu, Y.-P. Xu, P. Zheng, H.-W. Tong, Y.-X. Liu, Z.-G. Zha, Q.-D. Su, S.-M. Liu,
Mesoporous Silica-coated magnetic nanoparticles for mixed hemimicelles
solid-phase extraction of phthalate esters in environmental water samples
with liquid chromatographic analysis, Journal of the Chinese Chemical Society
60 (2013) 5362.
[16] F. Jay, V. Gauthier, S. Dubois, Iron particles coated with alumina: synthesis by a
mechanofusion process and study of the high-temperature oxidation resistance, Journal of the American Ceramic Society 89 (2006) 35223528.
[17] Z. Schnepp, S.C. Wimbush, M. Antonietti, C. Giordano, Synthesis of highly
magnetic iron carbide nanoparticles via a biopolymer route, Chemistry of
Materials 22 (2010) 53405344.
[18] X. Cheng, B. Wu, Y. Yang, Y. Li, Synthesis of iron nanoparticles in water-in-oil
microemulsions for liquid-phase FischerTropsch synthesis in polyethylene
glycol, Catalysis Communications 12 (2011) 431435.
[19] T. Iwasaki, K. Kosaka, N. Mizutani, S. Watano, T. Yanagida, H. Tanaka, T. Kawai,
Mechanochemical preparation of magnetite nanoparticles by coprecipitation,
Materials Letters 62 (2008) 41554157.
[20] J.N. Wang, L. Zhang, F. Yu, Z.M. Sheng, Synthesis of carbon encapsulated
magnetic nanoparticles with giant coercivity by a spray pyrolysis approach,
Journal of Physical Chemistry B 111 (2007) 21192124.
[21] S.I. Nikitenko, Y. Koltypin, L. Felner, I. Yeshurun, A.I. Shames, J.Z. Jiang,
V. Markovich, G. Gorodetsky, A. Gedanken, Tailoring the properties of FeFe3C nanocrystalline particles prepared by sonochemistry, Journal of Physical
Chemistry B 108 (2004) 76207626.
[22] I.K. Herrmann, R.N. Grass, D. Mazunin, W.J. Stark, Synthesis and covalent
surface functionalization of nonoxidic iron coreshell nanomagnets, Chemistry of Materials 21 (2009) 32753281.
[23] M.V. Yigit, D. Mazumdar, Y. Lu, MRI detection of thrombin with aptamer
functionalized superparamagnetic iron oxide nanoparticles, Bioconjugate
Chemistry 19 (2008) 412417.

[24] K.S. Suslick, T.W. Hyeon, M.M. Fang, Nanostructured materials generated by
high-intensity ultrasound: sonochemical synthesis and catalytic studies,
Chemistry of Materials 8 (1996) 21722179.
[25] K.V.P.M. Sha, A. Ulman, X.Z. Yan, N.L. Yang, C. Estournes, H. White,
M. Rafailovich, Sonochemical synthesis of functionalized amorphous iron
oxide nanoparticles, Langmuir 17 (2001) 50935097.
[26] S. Wizel, S. Margel, A. Gedanken, The preparation of a polystyrene-iron
composite by using ultrasound radiation, Polymer International 49 (2000)
445448.
[27] S. Zhang, Y. Zhang, Y. Wang, S. Liu, Y. Deng, Sonochemical formation of iron
oxide nanoparticles in ionic liquids for magnetic liquid marble, Physical
Chemistry Chemical Physics 14 (2012) 51325138.
[28] J.I. Lai, K.V.P.M. Sha, A. Ulman, K. Loos, Y.J. Lee, T. Vogt, W.L. Lee, N.P. Ong,
Controlling the size of magnetic nanoparticles using pluronic block copolymer
surfactants, Journal of Physical Chemistry B 109 (2005) 1518.
[29] S. Belaid, S. Laurent, M. Vermeech, L. Vander Elst, D. Perez-Morga, R.N. Muller,
A new approach to follow the formation of iron oxide nanoparticles synthesized by thermal decomposition, Nanotechnology 24 (2013) 055705.
[30] D.E. Weisshaar, T. Kuwana, Electrodeposition of metal microparticles in a
polymer lm on a glassy-carbon electrode, Journal of Electroanalytical
Chemistry 163 (1984) 395399.
[31] E.Y. Ye, B.H. Liu, W.Y. Fan, Preparation of graphite-coated iron nanoparticles
using pulsed laser decomposition of Fe-3(CO)(12) and PPh3 in hexane,
Chemistry of Materials 19 (2007) 38453849.
[32] Y.C. Han, H.G. Cha, C.W. Kim, Y.H. Kim, Y.S. Kang, Synthesis of highly
magnetized iron nanoparticles by a solventless thermal decomposition
method, Journal of Physical Chemistry C 111 (2007) 62756280.
[33] T.C. Monson, Q. Ma, T.E. Stevens, J.M. Lavin, J.L. Leger, P.V. Klimov, D.L. Huber,
Implication of ligand choice on surface properties, crystal structure, and
magnetic properties of iron nanoparticles, Particle and Particle Systems
Characterization 30 (2013) 205.
[34] D. Amara, I. Felner, I. Nowik, S. Margel, Synthesis and characterization of Fe
and Fe3O4 nanoparticles by thermal decomposition of triiron dodecacarbonyl,
Colloids and Surfaces APhysicochemical and Engineering Aspects 339 (2009)
106110.
[35] D. Amara, S. Margel, Solventless thermal decomposition of ferrocene as a new
approach for the synthesis of porous superparamagnetic and ferromagnetic
composite microspheres of narrow size distribution, Journal of Materials
Chemistry 21 (2011) 1576415772.
[36] Z.T. Zhang, B. Zhao, L.M. Hu, PVP protective mechanism of ultrane silver
powder synthesized by chemical reduction processes, Journal of Solid State
Chemistry 121 (1996) 105110.
[37] B.L. Rivas, E.D. Pereira, I. Moreno-Villoslada, Water-soluble polymermetal ion
interactions, Progress in Polymer Science 28 (2003) 173208.
[38] B.D. Cullity, Introduction to Magnetic Materials, Addison-Wesley Pub. Co.,
London, 1972.
[39] P. Matteazzi, F. Miani, G. Caer, Kinetics of cementite mechanosynthesis,
Hyperne Interactions 68 (1992) 173176.
[40] M. Ron, H. Shechter, A.A. Hirsch, S. Niedzwie, Mossbauer study cementite,
Physics Letters 20 (1966) 481483.
[41] Y. Yamada, H. Yoshida, Y. Kobayashi, Laser deposition of iron on graphite
substrates, Hyperne Interactions 198 (2010) 5559.
[42] T.C. Gibb, N.N. Greenwood, Mossbauer Spectroscopy, Chapman and Hall,
London, UK (1971) 659.
[43] R.M. Soyer, Cornell, Characterization in the iron oxides, Wiley-VCH Verlag
GmbH & Co. KGaA, London (2004) 152160.
[44] K.Y. Amsharov, M. Jansen, Formation of fullerenes by pyrolysis of perchlorofulvalene and its derivatives, Carbon 45 (2007) 117123.
[45] X.M. Wang, B.S. Xu, H.S. Jia, X.G. Liu, I. Hideki, HRTEM and Raman study of
onion-like fullerenes encapsulated Fe, Journal of Physics and Chemistry of
Solids 67 (2006) 871874.
[46] S. Chandra Kishore, A. Pandurangan, Synthesis and characterization of
Y-shaped carbon nanotubes using Fe/AlPO4 catalyst by CVD, Chemical
Engineering Journal 222 (2013) 472477.
[47] J.E. Lee, N. Lee, H. Kim, J. Kim, S.H. Choi, J.H. Kim, T. Kim, I.C. Song, S.P. Park, W.
K. Moon, T. Hyeon, Uniform mesoporous dye-doped silica nanoparticles
decorated with multiple magnetite nanocrystals for simultaneous enhanced
magnetic resonance imaging, uorescence imaging, and drug delivery, Journal
of the American Chemical Society 132 (2009) 552557.
[48] S. Mornet, S. Vasseur, F. Grasset, E. Duguet, Magnetic nanoparticle design for
medical diagnosis and therapy, Journal of Materials Chemistry 14 (2004)
21612175.