You are on page 1of 10

Journal of

Materials Chemistry C
View Article Online

PAPER

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Cite this: J. Mater. Chem. C, 2013, 1,


6325

View Journal | View Issue

Percolative NZFO/BTO ceramic composite with


magnetism threshold
Bin Xiao, Ning Ma and Piyi Du*
NZFO/BTO ceramic composites with dierent ferrite contents were successfully prepared by a solgel in situ
method. X-Ray diractometry (XRD), scanning electron microscopy (SEM), vibrating sample magnetometry
(VSM) and impedance analysis were used to reveal the microstructure and magnetic properties of the asprepared composites. The inuence of the non-magnetic phase on the magnetic properties of the
composite was analyzed based on the composition dependence of the microstructure, eective
permeability, saturation magnetization and coercivity. It reveals that a transition of topological structure
appears when the volume fraction of the NZFO phase approaches the percolation threshold fc 0.55.
In the composites below fc, the magnetic particles are surrounded by a non-magnetic phase, and the

Received 29th April 2013


Accepted 1st August 2013
DOI: 10.1039/c3tc30807c
www.rsc.org/MaterialsC

magnetic properties are controlled by both the demagnetizing eld and domain wall motion in the
ferrite phase. However, in the composites above fc, the magnetic particles are completely coalescent
with each other, thus the magnetic behaviors are closely dependent on the anisotropic eld of the
ferrite phase. The magnetic properties of the NZFO/BTO ceramic composite synchronously exhibit a
critical behavior at the percolation threshold.

Introduction
Multi-functional materials are important members of the
materials family because of their excellent performance and
potential applications in advanced electronic devices such as
magnetic nano-oscillators, printed circuit boards, passive
lters, electromagnetic sensors etc. in modern technologies.16
Due to the scarcity of single-phased multi-functional materials
in nature, articial composite materials such as ferromagnetic/
ferroelectric composites, have attracted much interest owing to
their combinational characteristics of electrical and magnetic
responses under external elds.711 Since they are expected to
improve the functionality of future electronic devices, especially
where both controllable dielectric and magnetic properties are
needed, it is of fundamental interest and practical signicance
to develop such high-property composite materials.
As revealed by many pioneering researches, the percolation
eect could appear in several composites with one of the
constituent phases being conductive, substantially inuencing
the dielectric properties of these composites.1215 In percolative
composites, the permittivity will present an abnormally giant
and nonlinear increase with the volume fraction of the
conductive phase approaching the percolation threshold fc,
which is theoretically predicted to be only 0.16 for threedimensional systems.16 However, subsequent investigations
pointed out that the value of fc is closely associated with the

State Key Laboratory of Silicon Materials, Department of Materials Science and


Engineering, Zhejiang University, Hangzhou, 310027, China. E-mail: dupy@zju.edu.cn

This journal is The Royal Society of Chemistry 2013

microstructure and varies with the preparation method,


meaning it could be much higher in real materials.17,18 For real
percolative composites in which one phase is generally more
conductive (or semi-conductive) than another, a high value of fc
is signicantly important in elevating the magnetic performance of the composites. Under such circumstances, the
permittivity still keeps a large value ascribed to the percolation
eect, meanwhile the permeability increases greatly due to the
avoidance of the negative eect of the composite law.18,26 These
discoveries give access to the obtainment of composites with
both super high dielectric properties and excellent magnetic
properties appearing simultaneously.
Although many eorts have been devoted to modifying
the properties of newly debuted percolative materials, basic
theories and principles involving the interactions between the
constituent phases are still far from perfect on account of the
complexity of the physics and chemistry in these composites.
Over the last decades, a physical picture of dielectric properties
based on percolation theory has become well-established and
accepted for interpreting the appearance of high permittivity in
percolative composites, but the aspect of magnetic properties
has not been studied so profoundly, seriously restricting the
development of multi-functional devices. In fact, several
magnetic anomalies correlated with the percolation process
have been observed in a variety of percolative systems. For
example, the eective permeability would diminish to zero near
the percolation point in iron-containing insulatorconductor
composite materials,19 and the coercivity reaches a giant
maximum value at the percolation threshold in FeSiO2

J. Mater. Chem. C, 2013, 1, 63256334 | 6325

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Journal of Materials Chemistry C


system.20,21 In some other materials, an interesting magnetic
transition from superparamagnetism to ferromagnetism could
be triggered by the percolation eect.22,23 These abnormal
phenomena enlighten us on the inherent essence of percolative
composites: is there any new unpredicted magnetic behavior in
similar systems? In fact, a new magnetic behavior controlled by
the percolation eect would appear possible in magnetic/nonmagnetic composites. Thus seeking novel magnetic
phenomena in such composites is desirable for designing new
electronic devices.
For this purpose, the Ni0.5Zn0.5Fe2O4 (NZFO)/BaTiO3 (BTO)
composite is unquestionably a suitable choice. It is a typical
percolative system composed of a general non-magnetic BTO
phase and a ferromagnetic NZFO phase.18,2427 Moreover, a
universal solgel process is employed for specimen preparation,
by which the constituent phases can mutually contact in
molecular level and easily form a uniform structure. We believe
this work is of special importance in establishing a complete
view of the magnetic behaviors of percolative composites. It may
also provide new guidance for the development of high-functional composite materials and related electromagnetic devices.

Experimental
xNi0.5Zn0.5Fe2O4/(1  x)BaTiO3 composite ceramics with the
molar fraction of NZFO ferrite x varying from 0 to 1 were
prepared by a solgel in situ method through the following
steps. AR-grade acetic acid (CH3COOH), barium acetate
(Ba(CH3COO)2), nickel acetate (Ni(CH3COO)2$4H2O), iron
nitrate (Fe(NO3)3$9H2O), zinc nitrate (Zn(CH3COO)2$4H2O),
tetrabutyl titanate (Ti(OC4H9)4) and ethylene glycol monomethyl ether (CH3OCH2CH2OH) were used as raw materials to
prepare the sol precursor. The obtained sol was dried at 120  C
in an oven for 3648 h. During the heating process, the sol
changed into a gel and further transformed into composite
powder, which was ground in an agate mortar to obtain a ner
state for subsequent treatment. In order to remove the residual
organic solvent thoroughly, the composite powder was presintered at 750  C for 1.5 h. Aerward, it was added 5% PVA,
pressed into toroidal rings under a non-axial pressure of 200
MPa, and sintered in air at 1310  C for 12 h and then at 1320  C
for 20 min. The step-sintering process will enhance the calcination of the NZFO/BTO composite, ensuring the constituent
phases to grow in a saturated state. The sintering process was
controlled precisely to avoid possible occurrence of big pores at
high temperature. The sample rings average 20 mm in outer
diameter, 10 mm in inner diameter and 12 mm in height. The
surfaces were polished with abrasive paper before XRD testing
and SEM observation. A detailed description of the preparation
procedures of solgel in situ derived NZFO/BTO composite can
be found in ref. 28.
The constituent phases were identied by a RIGAKUD/MAXC type X-ray diractometer (XRD, Cu Ka target) produced in
Tokyo, Japan. The wavelength of the X-ray source is 0.1540562
nm, and the step width is 0.02 with a scanning speed of 4 per
minute between 10 and 80 . The morphology and microstructure were observed by a SIRION-type eld emitting
6326 | J. Mater. Chem. C, 2013, 1, 63256334

Paper
scanning electron microscope (SEM) produced by Japan FEI
Corporation. The permeability spectra were measured by an
Agilent 16451B precision impedance analyzer (Palo Alto, CA)
within the frequency range of 100 kHz to 110 MHz. The
magnetic hysteresis loops and zero-eld cooled (ZFC)-eld
cooled (FC) curves were measured by a Vibrating Sample
Magnetometer (VSM) and a Physical Property Measurement
System (PPMS-9T) produced by Quantum Design Cooperation
(USA).

Results and discussion


Fig. 1 shows the X-ray diraction (XRD) patterns of as-prepared
NZFO/BTO composite with the molar fraction of ferrite x varying
from 0.1 to 1.0. It is conrmed that the NZFO/BTO composites
were successfully prepared by solgel in situ method, characterized by the well-dened diraction peaks assigned to the
BTO phase and NZFO phase. The peak intensity of the NZFO
phase increases notably in accordance with its increased molar
fraction in the composite. No additional peaks were observed
within the resolution limit of our equipment.
Fig. 2 demonstrates the SEM images of the fractures of asprepared NZFO/BTO composite with x 0.1, 0.4, 0.5, 0.8, 0.9
and 1.0, respectively. It can be seen that the constituent phases
exist in dierent morphology and geometric shapes in the
composite. The irregular polyhedral particles represent the BTO
phase, while the particles exhibiting staircase growth represent
the ferromagnetic NZFO phase. When the NZFO content is as
low as 10%, the microstructure is dominated by the nonmagnetic BTO phase as shown in Fig. 2(a). With the increment
of NZFO content, the NZFO phase that was formed in a
completely dierent shape can be distinguished clearly as
shown in Fig. 2(b) and (c). The volume content of the NZFO
phase increases distinctly with its content, gradually occupying
the whole space of the composite as shown in Fig. 2(d) and (e).
The dimension of the NZFO particles increases with increasing
NZFO content, from about 1 mm at x 0.1 to approximately 20
mm at x 0.9, while the dimension of the BTO particles
decreases. The dimension is estimated by measuring the
average grain size of the particles from SEM images. The
morphology of a single NZFO ferrite is also shown in Fig. 2(f). It
is noticeable that the grain size of the NZFO particles in singlephased ferrite is smaller than that in the composite.

Fig. 1 Typical XRD patterns of as-prepared NZFO/BTO composite with dierent


molar fraction of NZFO ferrite.

This journal is The Royal Society of Chemistry 2013

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Paper

Journal of Materials Chemistry C

Fig. 2 SEM images of the fractures of as-prepared NZFO/BTO composite


with dierent molar fractions of NZFO ferrite: (a) x 0.1; (b) x 0.4; (c) x 0.5; (d)
x 0.8; (e) x 0.9; (f) x 1.0.

Fig. 3 illustrates the Energy Dispersive Spectroscopy (EDS)


results on samples x 0.3, x 0.6 and x 0.9, respectively. The
white areas represent the distribution of constituent phases in
the composite. By EDS mapping, the evolution of the matrix
from the non-magnetic BTO phase to the magnetic NZFO phase
is directly presented. It is seen that with increasing ferrite
content, the NZFO particles tend to grow to larger size, form

Fig. 3

percolation paths, and nally agglomerate into large clusters


with a layered BTO phase embedded among the interfaces as
shown in Fig. 3.
Generally, the melting point and the calcination temperature
of either constituent phase in a composite system will be lower
than that of their single ones. The same case also happens in
the as-prepared NZFO/BTO composite. In other words, the
introduction of the BTO phase may cause a decrement in the
calcination temperature of the NZFO phase to some extent, thus
the NZFO particles in the composite grew more to a greater
extent and attained a much larger grain size as shown in
Fig. 2(d) and (e) compared with those in single ferrite as shown
in Fig. 2(f). The same eect has also been observed over a wide
range of sintering temperatures for the NZFO/BTO composite.28
Moreover, it reveals that both constituent phases are prone to
agglomerate into clusters in the composite to achieve a dense
state, that is, the majority of NZFO particles tend to contact
tightly with each other to form a magnetic matrix at high ferrite
content, while the majority of BTO particles try to connect
together to form a non-magnetic matrix at low ferrite content.
When the ferrite content increases to a certain value, percolation paths begin to form, leading to a transition of topological
microstructure in the composite.
The frequency dependence of the eective permeability of
as-prepared NZFO/BTO composite with various volume fractions of NZFO ferrite (fNZFO) is shown in Fig. 4(a), and the
composition dependence of the initial permeability is plotted in
Fig. 4(b). The inset in Fig. 4(b) shows the variation of the grain
size of the NZFO particles as a function of fNZFO. As can be
seen, the eective permeability exhibits a stable plateau over a
wide frequency range initially till several million Hertz, then
decreases fast at a particular frequency (called the cut-o
frequency) in high frequency region. The cut-o frequency

EDS mapping results on samples x 0.3, x 0.6 and x 0.9. The white areas represent the distribution of constituent phases.

This journal is The Royal Society of Chemistry 2013

J. Mater. Chem. C, 2013, 1, 63256334 | 6327

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Journal of Materials Chemistry C

Paper

Fig. 4 (a) The frequency dependence of the eective permeability of as-prepared NZFO/BTO composite with various volume fractions of NZFO ferrite fNZFO; (b) plots of
the initial permeability as a function of fNZFO. The inset shows the change of the grain size of NZFO particles. (c) Three-dimensional diagrams depicting the formation of
percolation paths and the change of topological structure with increasing fNZFO as described in the text.

decreases with increasing fNZFO, which can be well explained by


the famous Snoek's law.29 It states that for magnetic materials
the product of the initial permeability and the cut-o frequency
is a constant, which means if the initial permeability of the
composite increases, the cut-o frequency would move to lower
frequencies, and vice versa. As shown in Fig. 4(b), the initial
permeability shows a nonlinear increment with increasing
fNZFO. When fNZFO is smaller than 55%, the initial permeability
exhibits only a slight increase and maintains a low value of no
larger than 10, but it increases precipitately when fNZFO exceeds
55% and jumps up explosively in the high ferrite content region.
It should be noted that the nonlinear change of initial
permeability is quite dierent from the almost linear increment
of grain size as shown in Fig. 4(b). This is an indication that the
grain size is not the solely factor that determines the value of
initial permeability. For binary composites, the well-known
MaxwellGarnett model (eqn (1)) and BruggemanHanai equation (eqn (2)) can be adopted to analyze the composition
dependence of the initial permeability.30,31 The mathematical
expressions can be written as below:


3fNZFO b
m
 mBT
m mBT 1
(1)
; b NZFO
mNZFO 2mBT
1  fNZFO b
L

m  mNZFO mBT
1  fNZFO
mBT  mNZFO m

(2)

where mBT 1, mNZFO 156, fNZFO is the volume fraction of


NZFO phase, and L 1/3 for spherical particles. The MG model
is established on the premise that the interactions between the
ferrites could be neglected when calculating the eective
permeability of the composite; while the BH equation takes the
contact eect of the ferrite llers into consideration. In Fig. 4(b),
the dashdot line represents the theoretical predictions of the

6328 | J. Mater. Chem. C, 2013, 1, 63256334

MG model, and the dashed line represents the theoretical


predictions of the BH equation. The MG model matches
acceptably with the experimental results (hollow circles) below
fNZFO 0.55, but the deviation becomes very severe above 55%.
Comparatively, the predictions of the BH equation give satisfactory match above 55% in spite of some slight deviation.
Obviously, ignoring the interactions between the NZFO
particles is not applicable for the NZFO/BTO composite, especially when fNZFO is above 55%. The success of the BH equation to describe the variation of initial permeability above 55%
indicates that there are strong interactions between the
magnetic NZFO particles. In fact, with increasing fNZFO, the
NZFO particles grow larger and tend to contact naturally with
each other, i.e., they are inclined to percolate. When fNZFO reaches a certain value, such as 55% in this case, the magnetic
particles start to undergo partial coalescence in the composite,
and such coalescence will become enhanced with increasing
space-occupation of the NZFO phase. Therefore, the composite
experiences a transition of topological microstructure within
x 0.50.8. During this process, the fully contacting phase
changes from a non-magnetic BTO phase to a magnetic NZFO
phase as shown in Fig. 2 and 3. The schematic diagrams illustrating such topological microstructure transition are depicted
in Fig. 4(c). The whole percolation process can be divided into
three stages: before percolation, the NZFO particles (green
spheres) are dispersed and separated in a non-magnetic BTO
matrix; while at the percolation point the NZFO particles begin
to contact one another; nally evolving into another structure
where the BTO particles (blue spheres) are embedded in
magnetic NZFO matrix. At fNZFO 0.55, several percolation
paths start to come into being, viz. the topological structure
transition starts to appear. If the contact eect of the NZFO
phase is considered, satisfactory matches with the experimental

This journal is The Royal Society of Chemistry 2013

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Paper

Journal of Materials Chemistry C

values can be expected as is the case in the BH equation. It is


worth noting that the theoretical values of the BH equation
above fNZFO 0.55 are slightly larger than experimental data
although this gives better predictions than the MG model. This
small deviation probably reects the limited inuence of the
non-magnetic BTO phase, because the initial permeability of
the ferrite is intrinsically no longer a constant in the composite,
but varies with BTO content due to impurities engendered in
the NZFO phase.32
Referring to Takanori's classical theories on ferrites and
their composites,33 the magnetic response below 1 GHz is
mainly contributed by domain wall motion and spin rotation
for multi-domain magnetic materials. In this work, the ferrite
particles are several microns in size, thus they have a multidomain structure. As a result, two processes could give rise to
magnetization under an external magnetic eld: (1) the spins
inside the magnetic domains would turn and align parallel to
the external eld (spin rotation); (2) the domain walls would
move in order to produce larger magnetization, if the externally
applied magnetic eld is strong enough (domain wall motion).
Since the mobility of domain walls is easily aected by the
change of microstructure, the contribution of domain wall
motion is of particular emphasis to analyze the variation of
eective permeability. The eective permeability based on the
two magnetization processes can be mathematically described
by the following equations, which are helpful to analyze the
critical behavior correlated with the contact eect of the
magnetic phase in NZFO/BTO composite:
m m0  jm00 1 + cs + cd
cs

cd

ud

Ks
1i

(3)
(4)

u
us

Kd ud 2
 u2 ibu

(5)

where m is the eective permeability of the composites, m0 and


m00 the real and imaginary part of the eective permeability, cs
the magnetic susceptibility of spin rotation, cd the magnetic
susceptibility of domain wall motion, Ks the static susceptibility
with respect to spin rotation, Kd the static susceptibility with
respect to domain wall motion, u the externally applied
frequency, us the spin resonance frequency, ud the domain wall
resonance frequency, and b the damping factor of domain wall
motion.
The real and imaginary part of the complex eective
permeability can be written as below aer merging eqn (3)(5):

Kd ud 2 ud 2  u2
Ks us 2
m0 1 2

(6)

2
us u2
ud 2  u2 b2 u2
m00

Ks uus
Kd ud bu

2
2
us 2 u2
ud  u2 b2 u2
2

(7)

Approximately, the eective permeability of the NZFO/BTO


composites with various fNZFO is considered to merely stem
from the ferrite phase, herein the possible inuence of

This journal is The Royal Society of Chemistry 2013

non-magnetic BTO phase is neglected. By numerically tting


with experimental data, the theoretical curves of the eective
permeability could be obtained from eqn (6) and (7) with the
parameters listed in Table 1. One representative calculation
result for fNZFO 0.916 is displayed in Fig. 5 with the solid and
hollow circles representing the measured real and imaginary
parts of the complex eective permeability, all the solid lines
representing the results of theoretical calculations, respectively.
The analogical results of other samples are not shown for the
sake of brevity. As can be seen, the theoretical curves t very well
with measured values. It shows that both spin rotation and
domain wall motion should account for the magnetic response,
i.e. eective permeability when an external electromagnetic
eld is applied on the composite. As calculated in Table 1, both
the static spin susceptibility Ks and the static susceptibility of
domain wall motion Kd increase with increasing fNZFO. This
fact is well matched with the relations33 of
Ks 4pMs/(Ha + Hd) z 4pMs/Hd

(8)

Kd 3pMs2D/4g

(9)

and

where Ms is the saturation magnetization of the composites, Ha


the anisotropy eld, Hd the demagnetizing eld, D the grain size
of NZFO particles, and g the wall energy. The demagnetizing
eld usually reduces the total magnetic moment during the
magnetization of ferrite, and its strength is decided by the
geometric shape of the magnetic particles, which usually varies
with composition in the composite.33 The approximation in
Relation (8) is allowed given the fact that the magnetic anisotropy in NiZn ferrite can be neglected compared with the
demagnetizing eld Hd because the latter is much greater.34
Apropos for spin rotation, on one hand, the demagnetizing led
Hd decreases with increasing fNZFO because the grain size of
the NZFO phase increases as shown in Fig. 2. On the other
hand, the saturation magnetization of the composite increases
with increasing fNZFO. The combinational eect of these two
factors therefore leads to an increase in Ks with increasing
fNZFO according to eqn (8). As regards domain wall motion, both
saturation magnetization and grain size increase when fNZFO
increases, thus the increase in Kd can be expected. The increase
in both Ks and Kd ultimately result in signicant enhancement
to the eective permeability with increasing fNZFO as shown in
Fig. 4(b).

Table 1 Parameters obtained from numerical tting of experimental results in


eqn (6) and (7)

Spin rotation

Domain wall motion

fNZFO

Ks

us ( 10 )

Kd

ud ( 107)

b ( 107)

0.447
0.548
0.829
0.916
1.0

4.09
6.20
34.82
52.76
89.47

13.80
11.90
8.31
7.11
6.45

1.24
1.51
17.93
33.86
68.21

3.12
2.85
2.11
1.75
1.71

3.84
2.66
2.63
2.45
2.45

J. Mater. Chem. C, 2013, 1, 63256334 | 6329

View Article Online

Journal of Materials Chemistry C

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Fig. 5 The frequency dependence of the complex eective permeability of the


NZFO/BTO composite with fNZFO 0.916: the comparison between theoretical
calculations and experimental results. The results of other samples are omitted for
brevity.

It is worth mentioning that from the parameters in Table 1,


the rate of increase of the susceptibilities of both spin rotation
and domain wall in the composite below fc are very dierent
from that above fc. Below fc, the contributed proportion of
domain wall motion in total susceptibility is only about 25%
while that of spin rotation approaches 75%. It is apparently that
the surrounding non-magnetic phase aects seriously the
characteristics of ferrite in the composite, for the mobility of
domain walls is remarkably restrained. As described in eqn (8)
and (9), the susceptibility of spin rotation is controlled by
demagnetizing eld Hd and that of the domain wall is reciprocally dependent on the wall energy g. The demagnetizing eld is
largest when the ferrite particles are surrounded by a nonmagnetic phase.35 The result is that the response of spin
rotation would mostly be shielded, and the domain wall motion
becomes very dicult. However, the obstruction on spin rotation and domain wall motion will be dramatically reduced when
the ferrite starts to percolate. Aer percolation, a sharp reduction of the shielding eect brings about great enhancement in
the value of Ks.
The damping factor b scales the degree of impediment to the
movement of domain wall in the composite. It jumps down
from a high value of 38.4 MHz below 55% to 26.6 MHz at the
percolation point due to the formation of percolation paths,
viz., the contact eect of NZFO particles. Aerward, it exhibits
only a small decrease toward 24.5 MHz, which is a characteristic
value of a single ferrite phase as shown in Table 1. It is evident
that the composite below fc shows an extremely high pinning
eect on domain wall motion caused by the non-magnetic
phase and the pinning eect above fc is vastly relaxed, not only
owing to the dense cluster structure with fewer defects28 but
essentially beneting from the generation of percolation
networks. In addition, the wall energy g is generally dependent
on the pinning eect of the non-magnetic phase on domain wall
motion.36 Obviously, both the demagnetizing eld Hd and the
wall energy g are much smaller in the composite above fc than
that below fc as shown in Table 1. Furthermore, according to the
calculation results, the contributed proportion of the domain
wall motion in total susceptibility is about 0.25 below fc, much
smaller than that of spin rotation. This fact indicates that the

6330 | J. Mater. Chem. C, 2013, 1, 63256334

Paper
permeability component contributed by domain wall motion in
the composite below fc is much smaller than that contributed by
spin rotation. Therefore, domain wall motion is an important
factor governing the behavior of eective permeability below fc,
but such governability gets seriously undermined above the
percolation point when the mobility of domain walls becomes
quite easy through the percolation networks. Consequently, the
magnetic susceptibilities of the composite are typically dierent
before and aer the transition point of the topological microstructure, during which the shielding eect on the response of
magnetic moments and the pinning eect on domain wall
motion compete with each other in controlling the value of
permeability. At low NZFO content, the ferrites are surrounded
by a non-magnetic phase, and the permeability is primarily
determined by domain wall motion and the demagnetizing
eld, whereas at high content when the ferrite particles form
percolation networks, the inuence of the two factors becomes
much smaller.
The magnetic hysteresis loops of as-prepared NZFO/BTO
composites with dierent fNZFO measured at room temperature
are shown in Fig. 6. The arrow denotes the increased volume
fraction of NZFO ferrite. The details near zero magnetic eld are
shown as insets in Fig. 6(a), and the variation of saturation
magnetization (Ms) and coercivity (Hc) as a function of fNZFO are
summarized in Fig. 6(b). All the samples are saturated in
magnetization under an externally applied eld of 6 kOe,
exhibiting novel so magnetic properties that could be sensitively tuned by external magnetic eld. It is clearly illustrated in
Fig. 6(b) that the saturation magnetization Ms increases almost
linearly with increasing fNZFO, from 15.0 emu g1 at fNZFO
0.342, 72.8 emu g1 at fNZFO 0.916 to 76.6 emu g1 at fNZFO
1.0. The saturation magnetization is insensitive to microstructure because it is determined by the total amount of magnetic
moments in the system. However, the coercivity shows an
interesting behavior which is completely dierent from that of
saturation magnetization. As plotted in Fig. 6(b), it increases
slightly at rst, then decreases dramatically until it reaches an
impressively low value of 0.788 Oe near the percolation
threshold. Aerward, it increases with increasing fNZFO, and
nally approaches a stable value of about 6 Oe. This behavior is
a reection of its intrinsic structure-sensitive feature in
response to the change of NZFO content in the composite.
In this context, the composition dependence of coercivity
could be divided into three regions over the whole volume
fraction range. In region I, where fNZFO is far below the percolation threshold, the magnetic NZFO islands are uniformly
dispersed in a non-magnetic matrix, being separated by BTO
clusters. In such situation, the spin rotation is signicantly
blocked by the demagnetizing eld exerted on the ferrite phase,
and the domain wall motion is seriously pinned by the
surrounding BTO phase. Hence, the coercivity of the composite
might be relatively large and controlled mainly by shielding of
spin rotation and pinning of domain wall motion in this region.
In region III, where fNZFO is much higher than the percolation threshold, the NZFO particles fully contact with each other
and tend to act as a single ferrite cluster with a small quantity of
non-magnetic BTO islands dispersed among them. The

This journal is The Royal Society of Chemistry 2013

View Article Online

Paper

Journal of Materials Chemistry C

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Fig. 6 Magnetic hysteresis loops measured at room temperature: (a) dierent volume fraction of NZFO ferrite. The insets show the details of the loops near zero
magnetic eld. (b) The variation of saturation magnetization and coercivity as a function of fNZFO.

parameters in Table 1 suggest that the damping factor is


roughly 36% smaller for the composite in region I than that in
region III. This fact implies that the coercivity may no longer be
contributed by the pinning eect on domain wall motion,
although various kinds of defects exist inevitably in the
composite. In Fig. 8, the zero eld cooled (ZFC) and eld cooled
(FC) curves of the magnetization of the composite with fNZFO
0.447, fNZFO 0.645 and fNZFO 0.916 are depicted. The irreversible temperature Tirr, at which the ZFC and FC curves
separate with each other, is much higher for fNZFO 0.916
(300 K) than that for fNZFO 0.447 (230 K). As is known, the
irreversible temperature is closely related to the anisotropy eld
of the composite,37 namely, a stronger anisotropy eld is usually
characterized by a higher irreversible temperature. Generally,
there are two sources of magnetic anisotropy involved in NiZn
ferrite: (1) magnetocrystalline anisotropy determined by the
atomic structure of the spinel crystal that introduces preferential directions for magnetization; (2) shape anisotropy when the
particle is not perfectly spherical so that the demagnetizing eld
is not equal in all directions. In the case of NZFO/BTO
composite, high fNZFO ensures a more perfect formation of
crystalline NZFO phase with larger grain size in region III, while
those in region I have small grain size.28 Therefore, the strength
of the anisotropy eld is signicantly intensied with increased
NZFO content and crystallinity. It is also known that the coercivity is dependent on the anisotropy eld of ferrite (if it is the
main factor deciding the value of Hc),38 thus the higher the
anisotropy eld is, the larger the coercivity of the ferrites will be.
Hence, the coercivity of the ferrite phase becomes considerably
high and increases with increasing fNZFO in this composition
region. Obviously, the anisotropy eld takes the place of regulating the magnetic response in this region.
Now in region II, fNZFO is just in the vicinity of the percolation
threshold, and the NZFO particles begin to coalesce with neighboring ferrite particles and percolation paths start to form. On
one hand, with the increase in fNZFO, the interactions between the
magnetic moments in adjacent ferrite particles are so strong
enough that they could orientate themselves along the same
direction, considerably reducing the contribution of the pinning
eect imposed by a circumambient non-magnetic phase. On the
other hand, in the vicinity of the percolation threshold, the grain
size of the NZFO phase is still not so large and the non-magnetic

This journal is The Royal Society of Chemistry 2013

phase particles still exert an eect around NZFO, hence the


contribution of the anisotropy eld may still be suppressed. As a
result, an extremely small value of coercivity appears near the
percolation point. A similar phenomenon has also been observed
in percolative thin lms that exhibit a suppression of coercivity
due to the percolation eect.39 A theoretical model was also
proposed to explain the suppressed behavior of coercivity near fc,
which further broadened the understanding of the percolation
eect on the magnetic performance of percolative systems.40 In a
word, both a small anisotropy eld and suddenly weakened
pinning eect on the domain walls result in the smallest value of
coercivity in this region.
The thermal dependence of the magnetic hysteresis loops of
as-prepared NZFO/BTO composites between 2 and 300 K is
shown in Fig. 7, in which Fig. 7(a) demonstrates the loops of
fNZFO 0.447 (below fc), Fig. 7(c) the loops of fNZFO 0.916
(above fc), and Fig. 7(e) the loops of fNZFO 0.645 (critical
region). The enlarged loop areas near zero magnetic eld are
also shown as insets for clarity. The variation of saturation
magnetization and coercivity of the composite as a function of
temperature are plotted in Fig. 7(b), (d) and (f), respectively. For
all samples, the saturation magnetization decreases monotonously with increasing temperature, namely, from 41.9 emu g1
at 2 K to 22.0 emu g1 at 300 K for fNZFO 0.447, from 68.7 emu
g1 at 2 K to 39.8 emu g1 at 300 K for fNZFO 0.645, and from
111.0 emu g1 at 2 K to 69.4 emu g1 at 300 K for fNZFO 0.916.
However, the coercivity decreases with increasing temperature
from 15.2 Oe at 2 K to 1.1 Oe at 75 K initially, then increases to a
value of 7.9 Oe at 300 K for fNZFO 0.447. Contrastingly, it
increases monotonously with increasing temperature from
3.9 Oe at 2 K to 7.5 Oe at 300 K for fNZFO 0.916.
The thermal uctuation of magnetic moments is apt to make
the spin orientation depart from the direction of external
magnetic eld, and this is why the saturation magnetization
decreases with increasing temperature. Since the saturation
magnetization is mainly contributed by magnetons and their
orientation in the system and is not sensitive to the change of
microstructure, it shows the same variation with temperature
for all samples, regardless of its position below, near or above
the percolation threshold as shown in Fig. 7(b), (d) and (f).
Concerning the coercivity of the composite, dierent
tendencies of thermal dependence between 2 K and 300 K have

J. Mater. Chem. C, 2013, 1, 63256334 | 6331

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Journal of Materials Chemistry C

Paper

Fig. 7 Magnetic hysteresis loops of as-prepared NZFO/BTO composite with (a) fNZFO 0.447; (c) fNZFO 0.916; (e) fNZFO 0.645 measured between 2 and 300 K. The
plots of saturation magnetization and coercivity as a function of temperature are shown in (b), (d) and (f), respectively.

been observed. For the system above fc as shown in Fig. 7(d), the
coercivity increases with increasing temperature from about
3.5 Oe at 2 K to 78 Oe at 300 K. As discussed, the dominant
factor controlling the magnetic response in this case is the
anisotropy eld of the ferrite phase, so the coercivity behavior is
evidence of increased anisotropy with rising temperature in the
composite. A similar dependence has also been reported within
the same temperature range.41 For the system below fc, the
coercivity decreases with decreasing temperature from 78 Oe at
300 K to about 1 Oe at 70 K and then increases sharply to about
15 Oe at 2 K as shown in Fig. 7(b). Its blocking temperature Tb,
shown in Fig. 8, also appears near 70 K. The blocking temperature can be dened as the temperature at which thermal
activation begins to exceed the energy barrier created by
anisotropy and a broad maximum of magnetization in ZFC
process may occur around it. Below this temperature, the spins
are locked randomly in all possible directions and this state is
usually accompanied with restrained magnetization.42 Since the
coercivity is controlled by domain wall motion and spin rotation
for the NZFO/BTO composite below fc, if the temperature is low

6332 | J. Mater. Chem. C, 2013, 1, 63256334

enough below Tb, the spins will be seriously locked and the
coercivity may increase with decreasing temperature due to a
strong eective blocking eect on spin rotation and the domain
wall motion. Obviously, the increment of the coercivity with

Fig. 8 ZFCFC curves of the NZFO/BTO composite with fNZFO 0.447, fNZFO
0.645 and fNZFO 0.916 measured between 2 and 300 K under an applied
magnetic eld of 200 Oe.

This journal is The Royal Society of Chemistry 2013

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Paper
decreasing temperature below Tb conrms the essential role of
domain wall motion. When the temperature increases up to the
blocking temperature, the domain walls will be unblocked
completely and the coercivity will reach the smallest value of
only 1 Oe. Nevertheless with further increase in temperature,
the anisotropy eld increases gradually and plays a overwhelming role in controlling coercivity, thus the coercivity
increases with rising temperature from its smallest value
around 70 K to 78 Oe at 300 K as shown in Fig. 7(b).
Comparatively, the experimental results of fNZFO 0.645 (critical region) shown in Fig. 7(e) and (f) reveals that the behavior of
coercivity resembles the system below fc (it is dicult to locate
the precise position of fc, thus the measured sample may lie
only below or above the critical point), however, from the ZFC
FC curves in Fig. 8, we can see that the magnetization of fNZFO
0.645 is the largest, corresponding to its smallest coercivity
value of only 0.788 Oe. Likewise, the magnetization of fNZFO
0.447 is the smallest as it has the largest coercivity (8.03 Oe). A
smaller coercivity usually implies an easier magnetization of the
sample under external magnetic eld, therefore, these results
are in good agreement with the experimental data shown in
Fig. 6(b), providing solid evidence for the existence of a
magnetism threshold in the NZFO/BTO composite.
In a word, the presented experimental evidence reveals that
there does exist a magnetism threshold in the NZFO/BTO
composite. At the percolation point, the functional properties
synchronously exhibit a critical behavior following the transition of topological structure. Below the percolation threshold,
the magnetic response is mainly correlated with inhibited spin
rotation and pinned domain wall motion, whereas above the
percolation threshold, it is signicantly inuenced by the
anisotropy eld of the ferrite phase. The percolation threshold
controls the transition of magnetic behavior in the percolative
magnetic/non-magnetic composite.

Conclusions
The NZFO/BTO ceramic composites were successfully prepared
through a modied sintering process by a solgel in situ
method. The composition dependence of the magnetic properties reveals that a transition of topological microstructure
appears at the percolation threshold fc 0.55. The matrix
changes from a non-magnetic BTO phase to a magnetic NZFO
phase near the percolation point, resulting in a critical behavior
of magnetic responses. The static spin susceptibility and
domain wall motion contribute directly to the eective permeability of the composite, but the behavior of magnetic susceptibility is typically dierent below and above the threshold. The
demagnetizing eld imposed on spin rotation and the pinning
of non-magnetic phase on domain walls cooperatively control
the susceptibility of the composite below fc 0.55. The
anisotropy eld of the ferrite phase plays an important role in
determining the values of the eective permeability and the
coercivity of the composite above fc 0.55. The ndings may be
of interest and importance for establishing a clear physical
picture of percolative composites and may be useful to improve
the performance of related multifunctional devices.

This journal is The Royal Society of Chemistry 2013

Journal of Materials Chemistry C

Acknowledgements
This work was supported by the Natural Science Foundation of
China under grant no. 51272230 and no. 50872120, Zhejiang
Provincial Natural Science Foundation (Grant no. Z4110040)
and the National Key Scientic and Technological Project of
China (Grant no. 2009CB623302).

References
1 V. E. Demidov, S. Urazhdin, H. Ulrichs, V. Tiberkevich,
A. Slavin, D. Baither, G. Schmitz and S. O. Demokritov,
Nat. Mater., 2012, 11, 1028.
2 H. Zheng, J. Wang, S. E. Loand, Z. Ma, L. MohaddesArdabili, T. Zhao, L. Salamanca-Riba, S. R. Shinde,
S. B. Ogale, F. Bai, D. Viehland, Y. Jia, D. G. Schlom,
M. Wuttig, A. Roytburd and R. Ramesh, Science, 2004, 303,
661.
3 W. Eerenstein, N. D. Mathur and J. F. Scott, Nature, 2006,
442, 759.
4 O. D. Jayakumar, B. P. Mandal, J. Majeed, G. Lawes, R. Naik
and A. K. Tyagi, J. Mater. Chem. C, 2013, 1, 3710, DOI:
10.1039/c3tc30216d.
5 D. S. Rana, I. Kawayama, K. Mavani, K. Takahashi,
H. Murakami and M. Tonouchi, Adv. Mater., 2009, 21, 2881.
6 C. N. R. Rao and C. R. Serrao, J. Mater. Chem., 2007, 17, 4931.
7 X. Qi, J. Zhou, Z. Yue, Z. Gui, L. Li and S. Buddhudu, Adv.
Funct. Mater., 2004, 14, 920.
8 H. C. He, J. Wang, J. P. Zhou and C. W. Nan, Adv. Funct.
Mater., 2007, 17, 1333.
9 C. Deng, Y. Zhang, J. Ma, Y. Lin and C. W. Nan, Acta Mater.,
2008, 56, 405.
10 X. Qi, J. Zhou, B. Li, Y. Zhang, Z. Yue, Z. Gui and L. Li, J. Am.
Ceram. Soc., 2005, 87, 1848.
11 M. Stingaciu, R. K. Kremer, P. Lemmens and M. Johnsson,
J. Appl. Phys., 2011, 110, 044903.
12 D. Wang, T. Zhou, J. W. Zha, J. Zhao, C. Y. Shi and
Z. M. Dang, J. Mater. Chem. A, 2013, 1, 6162, DOI: 10.1039/
c3ta10460e.
13 Z. Wang, T. Hu, L. Wen, N. Ma, C. Song, G. Han, W. Weng
and P. Du, Appl. Phys. Lett., 2008, 93, 222901.
14 J. Huang, H. Zheng, Z. Chen, Q. Gao, N. Ma and P. Du,
J. Mater. Chem., 2009, 19, 3909.
15 Z. Chen, J. Huang, Q. Chen, C. Song, G. Han, W. Weng and
P. Du, Scr. Mater., 2007, 57, 921.
16 S. Kirkpatrick, Rev. Mod. Phys., 1973, 45, 574.
17 A. Malliaris and D. T. Turner, J. Appl. Phys., 1971, 42, 614.
18 H. Zheng, Y. Dong, X. Wang, W. Weng, G. Han, N. Ma and
P. Du, Angew. Chem., Int. Ed., 2009, 48, 8927.
19 L. V. Panina, A. S. Antonov, A. K. Sarychev, V. P. Paramonov,
E. V. Timasheva and A. N. Lagarikov, J. Appl. Phys., 1994, 76,
6365.
20 G. Xiao and C. L. Chien, Appl. Phys. Lett., 1987, 51, 1280.
21 C. L. Chien, J. Appl. Phys., 1991, 69, 5267.
22 H. J. Elmers, J. Hauschild, H. Hoche, U. Gradmann,
H. Bethge, D. Heuer and U. Kohler, Phys. Rev. Lett., 1994,
73, 898.
J. Mater. Chem. C, 2013, 1, 63256334 | 6333

View Article Online

Published on 01 August 2013. Downloaded on 24/01/2014 13:54:08.

Journal of Materials Chemistry C


23 L. P. Gor'kov and V. Z. Kresin, JETP Lett., 1998, 67, 985.
24 O. M. Hemeda, A. Tawk, A. A. Sharif, M. A. Amer,
B. M. Kamal, D. E. E. Reay and M. Bououdina, J. Magn.
Magn. Mater., 2012, 324, 4118.
25 L. P. Curecheriu, M. T. Buscaglia, V. Buscaglia, L. Mitoseriu,
P. Postolache, A. Lanculescu and P. Nanni, J. Appl. Phys.,
2010, 107, 104106.
26 J. Q. Huang, P. Y. Du, L. X. Hong, Y. L. Dong and M. C. Hong,
Adv. Mater., 2007, 19, 437.
27 L. Mitoseriu, I. Pallecchi, V. Buscaglia, A. Testino,
C. E. Ciomaga and A. Stancu, J. Magn. Magn. Mater., 2007,
316, e603.
28 B. Xiao, Y. L. Dong, N. Ma and P. Y. Du, J. Am. Ceram. Soc.,
2013, 96, 1240.
29 J. L. Snoek, Physica, 1948, 14, 207.
30 D. A. G. Bruggeman, Ann. Phys., 1935, 416, 636.
31 J. V. Mantese, A. L. Micheli, D. F. Dungan, R. G. Geyer,
J. B. Jarvis and J. Grosvenor, J. Appl. Phys., 1996, 79,
1655.

6334 | J. Mater. Chem. C, 2013, 1, 63256334

Paper
32 X. H. Zhang, L. Zhu, Y. L. Dong, W. J. Weng, G. R. Han, N. Ma
and P. Y. Du, J. Mater. Chem., 2010, 20, 10856.
33 H. Zheng, L. Li, Z. J. Xu, W. J. Weng, G. R. Han, N. Ma and
P. Y. Du, J. Phys. D: Appl. Phys., 2013, 46, 185002.
34 T. Tsutaoka, J. Appl. Phys., 2003, 93, 2789.
35 H. Igarashi and K. Okazaki, J. Am. Ceram. Soc., 1977, 60, 51.
36 P. Gaunt, Philos. Mag. B, 1983, 48, 261.
37 P. S. A. Kumar, P. A. Joy and S. K. Date, Bull. Mater. Sci., 2000,
23, 97.
38 C. R. Vestal and Z. J. Zhang, Nano Lett., 2003, 3, 1739.
39 J. M. Choi, S. Kim, I. K. Schuller, S. M. Parik and
C. N. Whang, J. Magn. Magn. Mater., 1999, 191, 54.
40 J. Meja-Lopez, R. Ramrez, M. Kiwi, M. J. Pechan, J. Z. Hilt,
S. Kim, H. Suhl and I. K. Shuller, Phys. Rev. B: Condens.
Matter, 2000, 63, 060401.
41 O. Kubo, T. Nomura and H. Yokoyama, IEEE Trans. Magn.,
1988, 24, 2859.
42 A. J. Rondinone, A. C. S. Samia and Z. J. Zhang, J. Phys. Chem.
B, 1999, 103, 6876.

This journal is The Royal Society of Chemistry 2013

You might also like