nanomaterials
Article
Improving Powder Magnetic Core Properties via
Application of Thin, Insulating Silica-Nanosheet
Layers on Iron Powder Particles
Toshitaka Ishizaki *, Hideyuki Nakano, Shin Tajima and Naoko Takahashi
Toyota Central R&D Labs., Inc., 41-1 Nagakute, Aichi 480-1192, Japan; hnakano@mosk.tytlabs.co.jp (H.N.);
tajima-angler@mosk.tytlabs.co.jp (S.T.); nao-t@mosk.tytlabs.co.jp (N.T.)
* Correspondence: ishizaki@mosk.tytlabs.co.jp; Tel.: +81-561-71-7654; Fax: +81-561-63-6135
Academic Editor: Thomas Nann
Received: 19 October 2016; Accepted: 15 December 2016; Published: 23 December 2016
Abstract:
A thin, insulating layer with high electrical resistivity is vital to achieving high performance
of powder magnetic cores. Using layer-by-layer deposition of silica nanosheets or colloidal silica over
insulating layers composed of strontium phosphate and boron oxide, we succeeded in fabricating
insulating layers with high electrical resistivity on iron powder particles, which were subsequently
used to prepare toroidal cores. The compact density of these cores decreased after coating with
colloidal silica due to the substantial increase in the volume, causing the magnetic flux density
to deteriorate. Coating with silica nanosheets, on the other hand, resulted in a higher electrical
resistivity and a good balance between high magnetic flux density and low iron loss due to the
thinner silica layers. Transmission electron microscopy images showed that the thickness of the
colloidal silica coating was about 700 nm, while that of the silica nanosheet coating was 30 nm.
There was one drawback to using silica nanosheets, namely a deterioration in the core mechanical
strength. Nevertheless, the silica nanosheet coating resulted in nanoscale-thick silica layers that are
favorable for enhancing the electrical resistivity.
Keywords:
powder magnetic core; silica nanosheet; insulating layer; electrical resistivity; iron loss;
magnetic flux density
1. Introduction
Powder magnetic cores have been extensively studied in recent years due to the many advantages
they offer over electromagnetic steel sheets with respect to their isotropic magnetic properties, high
electrical resistivity, flexible design, potential for size reduction, and high design flexibility [
1
]. In order
to use powder magnetic cores in AC magnetic field applications, however, it is important to reduce
the iron loss, i.e., the sum of eddy current loss and hysteresis loss. Pure iron powders, in particular,
are well known to offer a high magnetic flux density at low cost, but with the drawback of low
electrical resistance and an associated high eddy current loss [
2
]. Powder magnetic cores are, therefore,
usually fabricated by compacting magnetic powder particles and coating them with insulating layers
to prevent an eddy current from forming. This technique of coating powder particles with appropriate
insulation is crucial to improving the magnetic properties of a core.
Powder magnetic cores generally need to be annealed after press forming to reduce plastic strain,
as this accumulation of strain causes a high hysteresis loss. However, if the insulating layers used
do not have sufficient thermal resistance they may break after annealing, resulting in a decrease in
electrical resistivity and a very high eddy current loss. Although an increase in the thickness of the
insulating layers would enhance their thermal resistance, this would also reduce the magnetic flux
density due to the increased volume of non-magnetic phase. Consequently, the powder particles need
Nanomaterials 2017, 7, 1; doi:10.3390/nano7010001 www.mdpi.com/journal/nanomaterials
Nanomaterials 2017, 7, 1 2 of 13
to be coated with insulating layers that have high thermal and electrical resistance, and which are also
thin and homogeneous.
Several insulating materials have been used to coat powder particles, including silicone resins [
3
],
phosphates [
4
–
7
], silicate glasses [
8
,
9
], magnesium oxide [
10
], aluminum oxide [
11
–
13
], ferrites [
14
–
16
],
sodium salts [
17
,
18
], and silica [
19
–
23
]. A homogeneous and thin layer can be realized by using
silicone resin or phosphate, because these materials undergo chemical reactions with the particle
surface. However, the thermal resistance of these materials is not very high, and so they are prone
to breakage during annealing. An insulating layer composed of silicate glass offers a high thermal
resistance, but the thermal stress generated by the difference in the coefficients of thermal expansion
between it and iron causes the magnetic properties to deteriorate. Magnesium and aluminum oxides
also have high thermal resistance, but as the deposition of thin, homogenous layers is challenging,
the magnetic flux density generally deteriorates due to the extra volume of non-magnetic layers.
Ferrites have the advantage of being magnetic and insulating, but their nano-scale deposition with
high coverage is still difficult to achieve by ordinary methods involving the mixing of ferrite particles
or chemical precipitation. An insulating layer composed of sodium salts has a high thermal resistance
and can be formed homogeneously, but pure iron is prone to corrosion by sodium salts. Of the
various insulating materials used for magnetic materials, silica provides excellent stability against
corrosion, oxidation, and a high thermal-resistance [
22
–
26
]. Silica coating of magnetic powder particles
is often carried out via the hydrolyzation of tetraethoxysilane (TEOS), but this method makes it
difficult to achieve thin, homogeneous layers of silica [
19
–
23
]. An appropriate technique for creating
the homogenous, nanoscale silica layers needed for magnetic powder cores has, to the best of our
knowledge, not yet been reported.
Tajima et al. developed a new type of insulating layer composed of strontium phosphate and
boron, and showed that this has a higher electrical resistivity than conventional insulating layers
composed of phosphates, even after annealing at 673 K (this layer is hereafter referred to as a Sr-B-P-O
insulating layer) [
27
]. However, the thermal resistance of this Sr-B-P-O insulating layer was still
insufficient to remove strain by annealing at high temperature, and so an attempt was made to enhance
its thermal resistance by mixing silica compounds into iron powders coated with Sr-B-P-O insulating
layers. This resulted in a high electrical resistivity, but at the cost of a substantial decrease in the
compact and magnetic flux densities. The thickness of the silica coating, therefore, needs to be reduced,
and its homogeneity increased, by adopting an effective technique for precise deposition on iron
surfaces. There have, however, been very few studies in which the deposition of thin insulating layers
of silica have been deposited on iron powder particles for fabricating a magnetic core, as there has
been no effective method of achieving thin silica layers with precise control over their deposition.
A new inorganic-layer coating technique that involves the alternating deposition of cationic
polymers and negatively-charged colloids on a solid surface has attracted much attention of
late because it allows the layer thickness to be controlled by simply changing the number of
coatings. This coating technique is based on the layer-by-layer (lbl) assembly of oppositely-charged
polyelectrolyte and colloidal inorganic nanoparticles to create a multilayer film [
28
,
29
]. Using this
technique, oxide nanosheets were coated onto inorganic substrates in a colloidal suspension, from
which a variety of unique optical and electronic properties were obtained by precisely controlling
the coating thickness [
30
–
33
]. It can, therefore, be hypothesized that thin insulating layers with high
thermal and electrical resistivities could be realized through the precise deposition of nanoscale silica
onto magnetic powder particles.
Although commercial suspensions of colloidal silica are currently available, their particle diameter
is normally more than 10 nm. Thus, in order to reduce the size of silicon-based materials,
Nakano et al.
synthesized nanosheets with a nanometer scale thickness composed of silicon and amorphous
silica by exfoliating layered polysilane to create colloidal monolayers [
34
–
36
]. If these very thin
silica nanosheets could be coated onto magnetic powder particles, they should provide an effective
insulating layer. To test this idea, silica was coated onto iron powder particles via the lbl method using
Nanomaterials 2017, 7, 1 3 of 13
poly(diallyldimethylammonium chloride) (PDADMAC) as the cationic polymer, and the magnetic
properties of the magnetic cores produced from this material were examined.
2. Results and Discussion
2.1. Colloidal Silica Coating of Iron Powder Particles with and without Sr-B-P-O Insulating Layers
In order to assess the adhesiveness of silica, pure iron powder particles with a diameter of 20–160
µ
m
were used with or without Sr-B-P-O insulating layers, which was produced using a previously
described method [
27
]. The colloidal silica with a diameter of 10–20 nm was coated onto pure iron
powder particles with or without Sr-B-P-O insulating layers after coating of PDADMAC, and this
process was repeated 1–5 times. Toroidal cores with outer and inner diameters of 39 and 30 mm and
a thickness of 5 mm were fabricated by warm compaction of the coated iron powders at 423 K with
a pressure of 1176 MPa using the die wall lubrication method [
37
,
38
]. These toroidal cores were then
annealed for 30 min at 773 K in a nitrogen atmosphere to remove plastic strain.
Figure 1 provides the comparison of the electrical resistivities and compact densities of the
annealed toroidal cores fabricated from iron powder particles coated with colloidal silica with and
without Sr-B-P-O insulating layers increasing number of silica coatings from 1 to 5. Note that when
the iron powder particles were directly coated with colloidal silica, the electrical resistivity did not
increase appreciably with an increasing number of silica coatings, but the compact density decreased
proportionally. When Sr-B-P-O insulating layers were applied prior to coating with colloidal silica
the electrical resistivity increased by a much greater extent, though the decrease in compact density
was much the same in both cases. This indicates that amount of colloidal silica deposited was almost
the same regardless of whether Sr-B-P-O insulating layers were used or not, but those particles with
Sr-B-P-O insulating layers could be coated with silica more effectively. A higher resistivity could,
therefore, be obtained by coating silica over Sr-B-P-O insulating layers, and after every cycle of coating,
the compact density was found to be slightly lower than the particles without Sr-B-P-O insulating layers.
We can conclude from this that Sr-B-P-O insulating layers can accelerate the adsorption of colloidal
silica, which should result in a core with a higher electrical resistivity and lower compact density.
–
μ
–
–
(a)
(b)
–
–
0
40
80
120
160
200
1 3 5
Electrical resistivity (μΩ m)
Number of Coats
colloidal silica
Sr-B-P-O + colloidal silica
7.4
7.5
7.6
7.7
7.8
7.9
1 3 5
Compact density (Mg/m
3
)
Number of Coats
colloidal silica
Sr-B-P-O + colloidal silica
Figure 1.
(
a
) Electrical resistivities; and (
b
) compact densities of the annealed toroidal cores fabricated
from iron powder particles coated with colloidal silica with and without Sr-B-P-O insulating layers
increasing number of silica coatings from 1–5.
The obtained toroidal cores were used for evaluation of the magnetic properties because they
can prevent the influence of a demagnetizing field. The magnetic properties were measured by DC
and AC B-H curve tracers. Magnetic flux densities and maximum permeabilities were estimated at
a magnetic field strength of 10 kA/m from the DC B-H curves. Coercivities were also estimated from
the DC B-H curves applying a magnetic field strength to be 2 kA/m. Hysteresis and eddy current
losses were estimated at the maximum magnetic flux density of 1 T and frequency of 400 Hz by an AC
B-H curve tracer. An iron loss was the sum of hysteresis and eddy current losses.
Nanomaterials 2017, 7, 1 4 of 13
Figure 2 shows the comparison of the iron losses, magnetic flux densities, coercivities and
maximum permeabilities of the annealed toroidal cores fabricated from iron powder particles coated
with colloidal silica with and without Sr-B-P-O insulating layers increasing number of silica coatings
from 1 to 5. The magnetic hysteresis curves measured by a DC B-H curve tracer are shown in
Figures S1 and S2
for the particles with and without Sr-B-P-O insulating layers after coating of colloidal
silica, respectively. The decreasing trend in magnetic flux density with increasing number of silica
coatings was almost the same whether Sr-B-P-O insulating layers were used or not, but as both coatings
are non-magnetic, there was greater degradation in magnetic flux density when Sr-B-P-O insulating
layers were used in addition to silica. This agrees well with the decreasing trend in the compact
densities, which is closely related to the magnetic flux density. The eddy current losses also decreased
with increasing number of silica coatings, regardless of whether Sr-B-P-O insulating layers were used
or not, and the hysteresis losses were also almost the same. However, when the particles with Sr-B-P-O
insulating layers were used, the eddy current loss after every coating cycle was substantially less than
with the particles without Sr-B-P-O insulating layers. This can be explained by the fact that the eddy
current loss decreases in inverse proportion to the electrical resistivity, which was higher when silica
was coated over Sr-B-P-O insulating layers. The coercivities of the cores decreased with increasing
number of silica coatings, as did the compact and magnetic flux densities. This suggests that silica
layers provide a barrier against plastic strain during fabrication, and as coercivity is dependent on
plastic strain, it is gradually reduced with an increase in the thickness of the silica coating applied.
The permeabilities of the cores also decreased with the increasing number of silica coatings, because
a permeability is generally reduced owing to increase of an amount of silica.
(a)
(b)
(c)
–
1.2
1.3
1.4
1.5
1.6
1.7
1.8
0
500
1000
1500
2000
2500
3000
1 3 5 1 3 5
Magnetic flux density (T)
Iron loss (kW/m
3
)
Number of Coats
Eddy current loss
Hysteresis loss
Magnetic flux
density
colloidal silica
Sr-B-P-O + colloidal silica
0
50
100
150
200
250
300
1 3 5 1 3 5
Coercivity (A/m)
Number of Coats
Coercivity
colloidal silica
Sr-B-P-O + colloidal silica
0
200
400
600
800
1000
1 3 5 1 3 5
Maximum permeability
Number of Coats
Maximum
permeability
colloidal silica
Sr-B-P-O + colloidal silica
Figure 2.
(
a
) Iron losses, magnetic flux densities; (
b
) coercivities; and (
c
) maximum permeabilities of
the annealed toroidal cores fabricated from iron powder particles coated with colloidal silica with and
without Sr-B-P-O insulating layers increasing number of silica coatings from 1 to 5.
The above results suggest that irrespective of the surface condition, the similar amount of silica is
deposited onto surfaces with each increase in the number of silica coatings. XPS (X-ray photoelectron
spectroscopy) analysis was conducted on uncoated iron powder particles and iron powder particles
coated five times by colloidal silica with and without Sr-B-P-O insulating layers in order to analyze
Nanomaterials 2017, 7, 1 5 of 13
the surface chemical states. The Si2p and Fe2p3/2 core level spectra in Figure 3 show a strong peak
corresponding to silica (SiO
2
) at around 103 eV for the particles with and without Sr-B-P-O after coating
with colloidal silica, which indicates that silica can be deposited regardless of whether an insulating
layer exists or not. A strong peak corresponding to iron oxides (Fe
2
O
3
, Fe
3
O
4
) was observed at around
710–712 eV in the case of the uncoated particles, and this is attributed to naturally-formed surface
oxides. A small peak corresponding to iron oxides was also observed with the particles coated only
with colloidal silica, but this was weaker than that of the uncoated particles. This peak corresponding
to iron oxides was barely visible for the particles coated with colloidal silica over Sr-B-P-O insulating
layers. As the analysis depth of XPS is only about 2 or 3 nm, iron oxides cannot be detected if the
surface is thoroughly coated with colloidal silica. This means that the particles without Sr-B-P-O
insulating layers were not completely covered by silica, as the uncoated surface was detected by XPS,
whereas no surface was exposed on particles with Sr-B-P-O insulating layers. This difference agrees
well with the experimental results for the electrical resistivities and eddy current losses between the
cores prepared from particles coated with colloidal silica with and without Sr-B-P-O insulating layers.
(a)
(b)
100101102103104105106107108
Binding Energy (eV)
no coat
colloidal silica
Sr-B-P-O + colloidal silica
707708709710711712713714
Binding Energy (eV)
no coat
colloidal silica
Sr-B-P-O +
colloidal silica
Figure 3.
X-ray photoelectron spectroscopy (XPS) (
a
) Si2p; and (
b
) Fe2p3/2 core level spectra of
uncoated iron powder particles and iron power particles coated five times with colloidal silica with
and without Sr-B-P-O insulating layers.
Schematic models for coating powder particles via the lbl method, with and without Sr-B-P-O
insulating layers, are shown in Figure
4. Note that when no Sr-B-P-O insulating layers are used,
PDADMAC is not adequately attracted to the overall surface during the first coating step. However,
those areas that are covered by PDADMAC succeed in attracting silica, and become repeatedly coated
with silica with each coating cycle. This coating process results in islands of insulating layers, which
do not increase the electrical resistivity, even though the density does decrease. Powder particles with
Sr-B-P-O insulating layers, on the other hand, attract much more PDADMAC during the first step,
with silica being subsequently deposited on these PDADMAC-covered surfaces to produce robust
insulating layers.
(a)
(b)
(+)
Fe
PDADMAC
Silica (-)
Silica (-)
Fe
Sr-B-P-O
PDADMAC
(+)
Figure 4.
Schematic models for the silica coating of iron powder particles by the lbl method (
a
) without;
and (b) with Sr-B-P-O insulating layers.