54
EXPERIMENTAL STUDY OF FERROMAGNETIC CHAINS . . .
3463
of Fe2 with potassium borohydride and then a subsequent
heat treatment. Since the borohydride reduction is a fast exo-
thermic reaction, it was performed in a T-form reactor in
order to achieve a homogeneous composition and size distri-
bution. The reaction was controlled carefully to allow the
growth of particle chains and suppress self-nucleated par-
ticles by mixing specific concentrations of ferrous salt and
borohydride. Equal volumes of aqueous solutions of 0.1 M
FeCl •4H O and 0.5 M KBH were allowed to mix in the
ϩ
2
2
4
T-form reactor in the presence of an applied magnetic field
of 1500 Oe. The magnetic field was used to achieve the
growth in one direction. Although the previous studies of
borohydride reduction5 lack the control of particle size, we
found that the particle size can be controlled and varied
within the desired nanometer range by the addition of vary-
ing amounts of ethanol to the aqueous solutions. The Fe
particle chains so obtained was filtered, washed first with
water and then with acetone, and allowed to dry overnight in
a glove box filled with nitrogen. The samples were then
–7
FIG. 2. ͑a͒ TEM photograph, and ͑b͒ electron-diffraction pattern
of randomly oriented chains composed of Fe spheres with particle
size of 35 nm.
x-ray data. High-resolution TEM with a convergent beam of
Ͻ1 nm indicates that each spherical particle in the chain is a
single crystal.
ICPA data show that the boron content in the samples is
small ͑Ͻ8 at. %͒. Also, the boron content was found to de-
crease as the particle size increases. For example, the boron
content is about 8 at. % at a particle size of 20 nm and Ͻ3
at. % at 70 nm. The relatively lower boron contents ͑Ͻ8
at. %͒ in our samples seem to help to achieve crystalline
samples.
heated in a furnace in a flowing 90% Ar–10% H gas mix-
2
ture first at 150 °C for 30 min to remove the moisture, and
then at 500 °C for 1 h to clean the oxidized surface and to
achieve better crystallinity of the particles.
The magnetic measurements were carried out with a su-
perconducting quantum interference device magnetometer at
298 K. In order to compare the experimental results with the
chain of spheres model, random and oriented particle chains
were dispersed in a polyurethane film matrix and cured in the
presence and absence of an applied magnetic field of 1500
Oe. As the polyurethane matrix, the plastic substrate, and the
drinking straw used to hold the sample during the magnetic
measurements produce a significant diamagnetic moment at
higher fields, the observed magnetizations were corrected for
this diamagnetic background. The data presented are all after
this correction. The samples were also characterized by x-ray
diffraction and transmission electron microscopy ͑TEM͒
equipped with an energy dispersive spectroscopic analysis. A
high-resolution TEM with a beam point resolution of Ͻ1 nm
was employed to study the crystal orientation of each particle
in the chain. The boron contents in the samples were ob-
tained by inductively coupled plasma analysis ͑ICPA͒.
Since the Fe particles are crystalline in nature as indicated
by x-ray and electron diffractions, one would expect the
magnetic moments of the spheres to point along the easy axis
of magnetization. However, no preferred crystal direction
could be noticed to lie along the chain direction with conver-
gent electron-beam diffraction. This could be due to a fan-
ning configuration of the easy axis of magnetization as
shown in Fig. 1. As the spheres grow on the adjacent spheres
to give chains, the energy minimum may play a role in align-
ing the easy axis of magnetization on the adjacent spheres.
1
Jacobs and Bean treated each sphere as a dipole of moment
and diameter a in the chain of spheres model and showed
their energy to be
3
ij
Wϭ͑ /r ͓͒cos͑ Ϫ ͒Ϫ3 cos cos ͔,
͑1͒
i
j
i
j
i
j
III. RESULTS AND DISCUSSION
where and are the adjacent dipoles separated by a
i
j
A. X-ray and microscopic characterization
distance of rij and making angles of i and j with the vector
joining them. Equation ͑1͒ suggests that a fanning configu-
ration can keep the energy minimum.
X-ray powder diffraction recorded with the randomly ori-
ented chains shows reflections corresponding to crystalline
Fe having the bcc structure. The TEM photograph and elec-
tron diffraction recorded with randomly oriented chains are
shown in Fig. 2. The photograph shows that the chains are
composed of spherical Fe particles with a clear interface be-
tween the particles. The interface is more or less perpendicu-
lar to the chain growth direction. As the borohydride-
reduction process is a fast exothermic process, we believe the
interface between the spheres is free from oxidized layers,
which assures magnetostatic interactions between the adja-
cent particles. Also, the spherical particles in the chains are
found to be uniform in size with a variation of about Ϯ3 nm.
The electron-diffraction pattern recorded in a selected area
shows polycrystalline bcc Fe, which is consistent with the
We find that the one-dimensional chains are formed only
when the borohydride-reduction reaction is carried out in the
presence of an applied magnetic field. Also, the chains are
formed only with lower borohydride concentrations. Higher
borohydride concentrations lead to higher boron contents and
amorphous nature and to an increase in self-nucleation.
These observations suggest that crystalline anisotropy, mag-
netostatic interactions, and suppression of self-nucleation
may all play a role in the formation of chains.
B. Magnetization and coercivity
The variation of saturation magnetization Ms with the
particle size ͑diameter of Fe spheres͒ is shown in Fig. 3. Ms
Zhang
