3394 IEEE TRANSACTIONS ON MAGNETICS, VOL. 48, NO.

11, NOVEMBER 2012

Microwave Behavior of Polymer Bonded Iron Oxide Nanoparticles

A. Caprile , M. Coïsson , F. Fiorillo , P. Kabos , Fellow, IEEE, O. M. Manu , E. S. Olivetti , M. A. Olariu ,
M. Pasquale , and V. A. Scarlatache
Istituto Nazionale di Ricerca Metrologica, Torino, Italy
Dipartimento di Fisica, Politecnico di Torino, Torino, Italy
National Institute of Standards and Technology, Boulder, CO 80305 USA
Universitatea “Stefan cel Mare”, Suceava, Romania
Gheorghe Asachi Technical University of Iasi, Faculty of Electrical Engineering, Romania

Samples composed of a polymer matrix were loaded with different fractions from 0% to 30% of Fe oxide magnetic nanoparticles
with an average size ranging from 5 to 25 nm. The permittivity and permeability of the composites were determined upon a very wide
frequency range, spanning from DC to 30 GHz using different methods with overlapping frequency intervals. The nanoparticle content
was observed to significantly modify the permeability and permittivity behavior of the polymer. The frequency and intensity of both
dielectric and ferromagnetic resonances of the composites are clearly affected by the presence of the magnetic nanoparticles, and the
composites can be optimized for specific applications requiring absorption of electromagnetic radiation from the MHz to the GHz range.
Index Terms—Ferromagnetic resonance, magnetic composites, magnetic nanoparticles.

I. INTRODUCTION were selected as polymer base for the inclusion of the MNPs.

T HE DESIGN and use of polymer components containing Several samples were prepared with different types of MNPs
magnetic nanoparticles [1] is in rapid development due to and different volume content, ranging from 0 to 30%. A wide-
the wide variety of existing and potential applications in elec- band electromagnetic characterization was performed using dif-
tronics and especially in transportation (rail, automotive, and ferent experimental methods:
aerospace) where low-cost, lightweight and high strength ma- — dielectric measurements were made using the parallel
terials are required more and more for energy saving purposes. plates method and an RLC meter in the frequency range
Embedding nanoparticles in a well known polymer-based com- from 40 Hz to 1 MHz;
ponent allows one to include additional electromagnetic func- — magnetic properties were obtained in DC with a vibrating
tionalities at low cost without inducing a significant decrease sample magnetometer and in AC up to 1 MHz with a
of mechanical and thermal properties. Polymer-based nanocom- wattmeter hysteresisgraph setup;
posites with high dieletric constant are already well known for — the electromagnetic behavior of the samples was further
applications in high speed electronics and antennas, which allow characterized up to 30 GHz using a vector network an-
improvements in the bandwidth and size reduction of compo- alyzer (VNA), through the analysis of the scattering pa-
nents [2], [3]. Here the polymers containing magnetic nanopar- rameters S in open and short circuit configuration using
ticles (MNPs) have been studied for thermoplastic bonding of an N-type coaxial line (up to 6 GHz) and a coplanar wave-
different materials such as glass-metal and plastic-metal in in- guide (CPW) (up to 30 GHz) inserted in the gap of an elec-
dustrial processes based on RF inductive heating. A general tromagnet, reaching a 0.5 T static magnetic field.
improvement of the bonding process of windshields, bumpers The results show that the ferromagnetic resonance (FMR) fre-
and other accessories was obtained by inclusion of MNPs to the quency of the polymers containing MNPs can be increased from
polymer matrix [4], leading to a higher heating rate and a more to 6 GHz by increasing the average size of the
uniform temperature distribution of the polymer. In particular it MNPs. It was also verified that the FMR frequency tends to in-
is possible to exploit their absorption properties in order to re- crease when the MNPs content increases. Due to the large width
alize hot melt adhesives or microwave absorbers, by the proper of the absorption peak , the polymer-MNPs
combination of the MNPs and the polymer matrix. composites can be optimized for energy absorption in a specific
In our work, samples composed of polymer and Fe oxide frequency range, starting from the RF regime up to several GHz,
nanoparticles were analyzed and optimized for electromagnetic by choosing the size and content of MNPs. Dielectric losses,
absorption up to the microwave frequency range. Iron oxide which were also analyzed in this study, appear to be present in
MNPs of average size from 5 to 25 nm were prepared with dif- the MHz region. Additional studies may be required in the fu-
ferent methods and thermoplastic polyolefin elastomers (TPO) ture to optimize the interplay of MNPs and different polymers
for different applications.

Manuscript received March 02, 2012; revised May 14, 2012; accepted May II. SAMPLES AND EXPERIMENTAL SETUPS
15, 2012. Date of current version October 19, 2012. Corresponding author:
A. Caprile (e-mail: [email protected]). A. Samples
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. A variety of MNPs consisting of a mixture of magnetite/
Digital Object Identifier 10.1109/TMAG.2012.2200462 maghemite was produced with methods based on the

0018-9464/$31.00 © 2012 IEEE

CAPRILE et al.: MICROWAVE BEHAVIOR OF POLYMER BONDED IRON OXIDE NANOPARTICLES 3395

TABLE I
EXPERIMENTAL TECHNIQUES USED VERSUS FREQUENCY RANGE

Fig. 1. VNA setup with a shorted termination.

coprecipitation of Fe (II) and Fe (III) oxides from an alkaline B. Experimental Setups

solution, obtaining uncoated hydrophilic particles which were A wideband characterization of the magnetic and dielectric
then dispersed in different concentration in a common polymer properties of the samples was performed using different exper-
matrix. MNPs were produced in the size regime where single imental setups, in order to obtain consistent results in overlap-
MNPs present a single magnetic domain, exhibiting a super- ping frequency ranges, as described in the following and de-
paramagnetic behavior [5]. picted in Table I:
Two families of MNPs were studied: “N type” composed in 1) Hysteresisgraph-Wattmeter: In the quasi-static regime
prevalence by maghemite particles having an average size and up to a few MHz, the magnetic characterization of the
and “C type” composed of almost pure magnetite samples was performed through a hysteresisgraph-wattmeter
with , as determined from the analysis of the setup previously developed at INRIM [6], composed of a
waveform generator, a 50 W DC-10 MHz power amplifier and
relevant magnetization curves, X-ray diffraction and electron
a two-channel 500 MSample/s digital oscilloscope. Primary
microscopy imaging TEM/FESEM. and secondary windings and connecting cables were adjusted to
The adhesive matrix chosen to include the MNPs was a TPO, minimize stray capacitances. The primary current, detected on
which showed an excellent stability when subjected to thermal a calibrated anti-inductive resistor, and the secondary voltage
cycles and to radiofrequency (RF) fields, as verified by means of are transferred and integrated off-line to obtain magnetization
differential scanning calorimetry measurements, performed be- curves and hysteresis loops.
fore and after a thermal aging process and also after subjecting 2) Coaxial VNA Measurements (Short Circuit): A VNA was
the samples to RF fields (not shown here). used to collect the scattering matrix in different conditions. The
Composite samples were thus produced by heating and analysis of scattering parameters , using an N-type coaxial
mixing the polymer matrix with different volume contents, line and shorted termination allowed us to obtain reliable mea-
ranging from 0% to 30% of the different types of Fe oxide surements of the complex magnetic permeability up to 6 GHz.
A schematic view of the experimental setup is depicted in Fig. 1.
MNPs. The TPO mixture was first rolled in a ribbon shape
The reflection parameter is converted into the real and
and then punched in the shape of discs, with diameters of 26 imaginary part of the magnetic permeability using the following
mm for the electric permittivity measurements, and as rings for relations [6]:
the insertion in the N-type connector, with an inner diameter
of 3 mm and external diameter of 7 mm, for permeability and (1)
permittivity measurements up to 6 GHz.
3396 IEEE TRANSACTIONS ON MAGNETICS, VOL. 48, NO. 11, NOVEMBER 2012

(2)

where are respectively the thickness, the outer di-

ameter and the inner diameter of the ring sample and
. In fact, the input impedance at the sample
plane (see Fig. 1) is

(3)

3) CPW VNA FMR Measurements: Two-ports VNA fer-

romagnetic resonance measurements using a CPW were
performed in order to evaluate the FMR response of our
samples, using a 65 GHz VNA with 2.4 mm connectors and
cables. The standard inductive setup uses a commercial CPW
Fig. 2. FMR peak positions as a function of the applied magnetic field for dif-
with a 50 GHz bandwidth. The microwave measurements were ferent MNPs powders (solid squares) and with MNPs dispersed in a Polyolefine
performed under in-plane DC magnetic fields up to 0.5 T. A matrix with different concentration. The solid lines are the linear fit of the pure
flat face of the disk sample was positioned on the CPW and powder responses. (Measurement type: HF-b).
parameter were recorded and analyzed to obtain the
FMR frequency and sample normalized permeability [7]. This
setup has been widely used for characterization of samples
up to 30 GHz. This experimental technique, being based on a
“through” configuration, produces a signal which is a mixture
of dielectric and magnetic response of the sample. The dielec-
tric part can be neglected when differential tests are made at
different values of the applied magnetic field, and the dielectric
components cancel out.
The RF and microwave dielectric properties were determined
using the following experimental setups:
1) Impedance analyzer Dielectric properties were measured
in the frequency range by means of
a parallel plate capacitor, with a plate diameter of 26 mm,
connected to an Impedance Analyzer. Sample permittivity
was also measured in the frequency range from 1 MHz to
3 GHz with an RF Impedance Analyzer and an RF sample
cell.
2) Coaxial VNA measurements (open circuit) The VNA was Fig. 3. Real and the imaginary part of the magnetic permeability for the N type
samples (4–6 nm). A wide peak in the imaginary part of is clearly visible,
also used to characterize the complex dielectric behavior consistently with the FMR measurements. (Measurement type: HF-a).
by using an open circuit N-type coaxial termination. The
relations used to convert the reflection parameter into
the real and imaginary dielectric constant are
the data to zero applied field), a result which allows an opti-
and (4) mization of the sample preparation to fit specific applications
and frequency ranges through particle control.
Typical examples of the real and the imaginary part of the
where
magnetic permeability obtained using the coaxial line short
circuit method are shown in Fig. 3. A wide peak in the imagi-
nary part of is clearly visible, a result which is consistent with
the FMR results of Fig. 2 obtained using the CPW VNA FMR
method (see Table I). On samples C it was not possible to ob-
and and is given by (3). tain a similar FMR peak due to the shift to higher frequencies.
However, an increase in MNPs content leads to an increase in
the real part of the permeability, consistently with [8].
III. RESULTS AND DISCUSSION
The real part of the dielectric constant of the C type sam-
The FMR absorption spectra of various samples with dif- ples obtained with the three different experimental techniques
ferent MNPs volume content and of pure MNPs in response is shown in Fig. 4. The overlap between data points obtained
to an applied magnetic field are shown in Fig. 2. The effect of from different methods is excellent over a wide range of fre-
prevalent composition and particle size is quite apparent in the quencies. The role of MNPs results in an increase of the
different zero-field resonances (see the linear extrapolation of with increasing content.
CAPRILE et al.: MICROWAVE BEHAVIOR OF POLYMER BONDED IRON OXIDE NANOPARTICLES 3397

6 GHz, a result well supported by a large set of additional mea-

surements on other MNPs of different size and similar compo-
sition (not shown here) and also by [9] and [10]. Both the mag-
netic permeability and the dielectric constant are found to
be dependent on the MNPs concentration.
In the imaginary part of a peak is clearly visible, consis-
tently with the FMR measurements shown in Fig. 2, and an in-
crease of magnetic permeability due to higher content of mag-
netic material in the samples is observed.
The real part of of the composites shows an increase with
increasing MNPs content, due to the higher permittivity of the
MNPs with respect to the pure TPO. For what concerns the
imaginary part of , an increase in the dielectric loss is also ob-
served in the MHz range.
It can be noted that the superposition of experimental data
Fig. 4. Real part of the dielectric constant for the C type samples (20–30 nm).
from the different techniques exploited is excellent, providing a
An increase in the MNPs content causes an increase of . (Measurement good reliability of the values of the permittivity and also of the
types: stars: HF-a, triangles: HF-b, circles: LF). permeability, which are quite small.

ACKNOWLEDGMENT
This work was partially funded by Regione Piemonte through
the “Polimag” project, POR.FESR07-13-I.1.3, by the Roma-
nian CNCSIS-UEFISCSU Project number RU-107/2010, and
also supported by PERFORM-ERA “Postdoctoral Performance
for Integration in the European Research Area” (ID-57649), fi-
nanced by the European Social Fund and the Romanian Govern-
ment, and by EURODOC “Doctoral Scholarships for research
performance at European level” project, financed by the Euro-
pean Social Found and Romanian Government.

REFERENCES
[1] H. Ma, A. K.-Y. Jen, and L. R. Dalton, “Polymer-based optical waveg-
uides: Materials, processing, and devices,” Adv. Mater., vol. 14, pp.
1339–1365, 2002.
[2] H. Mosallaei and K. Sarabandi, “Magneto-dielectrics in electromag-
Fig. 5. Imaginary part of the dielectric constant of the pure polymer matrix, netics: Concept and applications,” IEEE Trans. Antenn. Propag., vol.
and of the three C-samples, showing some resonances in the MHz range. (Mea- 52, no. 6, pp. 1558–1567, Jun 2004.
surement type: HF-b). [3] R. C. Hansen and M. Burke, “Antennas with magneto-dielectrics,” Mi-
crow. Opt. Technol. Lett., vol. 26, pp. 75–78, 2000.
[4] J. Gass, P. Poddar, J. Almand, S. Srinath, and H. Srikanth, “Superpara-
magnetic polymer nanocomposites with uniform Fe3O4 nanoparticle
The imaginary part of the permittivity of C type samples and dispersions,” Adv. Funct. Mater., vol. 16, pp. 71–75, 2006.
the polimer matrix alone (solid line), is shown in Fig. 5, where [5] M. Pasquale, E. S. Olivetti, M. Coïsson, P. Rizzi, and G. Bertotti, “Fer-
romagnetic resonance and superparamagnetic behavior of iron oxide
an increasing content of MNPs causes an increase of the dielec- nanoparticles injected in porous anodic alumina,” J. Appl. Phys., vol.
tric losses in the MHz range, which may be attributed to the 103, p. 07D527, 2008.
orientation polarization and to a slight shift toward lower fre- [6] F. Fiorillo, “Measurements of magnetic materials,” Metrologia, vol.
47, pp. S114–S142, 2010.
quency of the resonance peak. [7] S. S. Kalarickal, P. Krivosik, M. Wu, C. E. Patton, M. L. Schneider,
P. Kabos, T. J. Silva, and J. P. Nibarger, “Ferromagnetic reso-
nance linewidth in metallic thin films: Comparison of measurement
IV. SUMMARY methods,” J. Appl. Phys., vol. 99, p. 093909, 2006.
[8] C. Morales, J. Dewdney, S. Pal, K. Stojak, S. Skidmore, H. Srikanth, J.
The analysis of FMR response at zero-field shows that it pos- Wang, and T. Weller, “Tunable magneto-dielectric polymer nanocom-
sible to control the FMR peak position through particle size con- posites for microwave applications,” IEEE Trans. Microw. Theory
Tech., vol. 59, no. 2, pp. 302–305, Feb. 2011.
trol. In particular, we found that MNPs with a smaller size show [9] F. J. Owens, “Ferromagnetic resonance of magnetic field oriented
a resonance in a lower frequency range: maghemite Fe3O4 nanoparticles in frozen ferrofluids,” J. Phys. Chem. Solids, vol.
N type MNPs, of average size of 4–6 nm, have the zero-field 64, pp. 2289–2292, 2003.
[10] J. M. Vargas, E. Lima, Jr, R. D. Zysler, J. G. S. Duque, E. De Biasi, and
peak at about 1.5 GHz, while magnetite C type MNPs, M. Knobel, “Effective anisotropy field variation of magnetite nanopar-
with mean diameter of 25–30 nm resonate at zero-field around ticles with size reduction,” Eur. Phys. J. B, vol. 64, pp. 211–218, 2008.