An induction heater device for studies of magnetic hyperthermia and specific

absorption ratio measurements

M. E. Cano, A. Barrera, J. C. Estrada, A. Hernandez, and T. Cordova
Citation: Review of Scientific Instruments 82, 114904 (2011); doi: 10.1063/1.3658818
View online: http://dx.doi.org/10.1063/1.3658818
View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/82/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Controlling temperature in magnetic hyperthermia with low Curie temperature particles
J. Appl. Phys. 115, 17B531 (2014); 10.1063/1.4868709
Quasi-static magnetic measurements to predict specific absorption rates in magnetic fluid hyperthermia
experiments
J. Appl. Phys. 115, 043907 (2014); 10.1063/1.4862647
Characterization of tissue magnetic susceptibility-induced distortions for MRIgRT
Med. Phys. 39, 7185 (2012); 10.1118/1.4764481
Fe Fe oxide nanocomposite particles with large specific absorption rate for hyperthermia
Appl. Phys. Lett. 90, 233112 (2007); 10.1063/1.2746064
Experimental study on magnetic and heating characteristics of magnetic wood
J. Appl. Phys. 85, 5732 (1999); 10.1063/1.370267

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24

REVIEW OF SCIENTIFIC INSTRUMENTS 82, 114904 (2011)

An induction heater device for studies of magnetic hyperthermia

and specific absorption ratio measurements
M. E. Cano,1,a) A. Barrera,1 J. C. Estrada,1 A. Hernandez,2 and T. Cordova2
1
2

Centro Universitario de la Cinega, Universidad de Guadalajara, 47810 Ocotln, JAL, Mexico

Divisin de Ciencias e Ingenieras, Universidad de Guanajuato, 37150 Len, GTO, Mexico

(Received 10 September 2011; accepted 16 October 2011; published online 30 November 2011)
The development of a device for generating ac magnetic fields based on a resonant inverter is presented, which has been specially designed to carry out experiments of magnetic hyperthermia. By
determining the electric current in the LC resonant circuit, a maximum intensity of magnetic field
around of 15 mT is calculated, with a frequency around of 206 kHz. This ac magnetic field is able to
heat powdered magnetic materials embedded in biological systems to be used in biomedical applications. Indeed, in order to evaluate the sensitivity of the device we also present the measurements of
the specific absorption rate in phantoms performed with commercially prepared Fe3 O4 and distilled
water at different concentrations. 2011 American Institute of Physics. [doi:10.1063/1.3658818]
I. INTRODUCTION

Magnetic hyperthermia is a biomedical technique which

recently has been used in order to confirm the therapeutic
effect in the elimination of tumors in animals.1, 2 Unlike radiotherapy, this novel technique is based on the use of nonionizing radiation and the concept is to heat a set of magnetic nanoparticles functionalized and non-toxics, which are
embedded in malignant cells to reach 46 C in order to produce necrosis.15 For many years, it has stated that the heating by power losses on superparamagnetic or ferromagnetic
particles by using ac magnetic fields, is mainly due to the
Neel-Brownian relaxation, magnetic hysteresis loop, and the
eddy currents or a mixture of thereof.3, 6, 7 Nevertheless, very
recent works8, 9 have reported that the induction heating is always due at the hysteresis losses in both regimes superparamagnetic or ferromagnetic. The kind of superparamagnetic or
ferromagnetic materials commonly used are iron oxides, with
diameters from 10 to 100 nm.1013 Also, the frequencies and
amplitudes of the ac magnetic fields are from 100 to 500 kHz
and from 10 to 1300 Oe, respectively.1, 2, 7, 1016
On the other hand, the specific absorption ratio (SAR)
gives a measure of the power absorbed by the material in
the presence of an ac magnetic field, i.e., the quantity of energy dissipated in heat per unit of mass. The SAR is given by
Eq. (1),
dT 1
,
(1)
dt m Fe
where c is the total specific heat capacity of the material and
mFe is the iron concentration per ml. Currently, several kinds
of nanoparticles and coatings are performed in laboratories
of chemistry or biochemistry with the purpose of to produce
functional and selective magnetic labels with high SAR values, to ensure the efficiency of the thermotherapy. Likewise,
before experimenting on animals, these magnetic markers are
very used in experiments in vitro in cellular biology laboratories, for carry out the necrosis or apoptosis evaluation directly
SAR = c

a) Electronic mail: eduardo.cano@cuci.udg.mx.

0034-6748/2011/82(11)/114904/6/$30.00

on cells cultures. To all these laboratories, a compact experimental setup such as our device represents a good alternative
to perform their fast trials and precise measurements, in other
words, for them this is a powerful tool.
Devices for the generation of ac magnetic fields at high
frequency and power have been reported in the last decade4, 16
for applications of magnetic hyperthermia in whole body.
These devices are bulky and sometimes inappropriate to perform preliminary tests in laboratories of chemistry. Meanwhile, Fumiko et al.15 and Ivkov et al.17 have employed
experimental setups for thermotherapy applications on rodents, by plugging coils specially designed on commercial
switching devices; but they do not reported information about
the efficiency of the ac magnetic field generators. Moreover, all the devices are based on the use of resonant inverters; specifically the switching of the dc current by arrangements of transistors of Insulated Gate Bipolar Transistor
(IGBT) or Metal Oxide Semiconductor Field Effect Transistor
(MOSFET) technology, which feeds a resonant tank circuit.
Commonly some electronic topologies related with the coupling in series or parallel of the passives elements L and C
for the resonant circuit (hybrids) are employed such as the
capacitor-capacitor-inductor (CLC) and inductor-capacitorinductor (LCL), with the object of incrementing the efficiency
of the resonant signal by filtering all the unwanted frequencies
and reducing its total harmonic distortion (THD).1820 ChengChi Tai et al.21, 22 and Goya et al.23 reported recently the
implementation of similar devices of smaller dimensions for
induction heating nanomaterials to biomedical applications.
These devices have been built employing resonant inverters
using only a half H-bridge in the switching stage and showing a good efficiency.
The main aim of this work is to show the implementation, assembly, and characterization of a compact device
able to produce ac magnetic fields with amplitude adjustable,
which has been performed with a current inverter in full
H-bridge configuration coupled to one resonant CL-CL stage;
in agreement with24 the use of a full H-bridge increase
the output power four times by comparing with the devices

82, 114904-1

2011 American Institute of Physics

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24

114904-2

Cano et al.

FIG. 1. (a and b) The signals measured of the PWM driver only for two
terminals (Pulse 1 and Pulse 2) of the H-bridge and the electronic diagram of
the resonant inverter, respectively.

employing only the half one, then in experiments where is

needed a higher amplitude of magnetic field our device is a
better alternative. This device was particularly designed for
experiments in vitro of magnetic hyperthermia. Furthermore,
for testing the capability of heating, this device was also used
for determining the temperature rise and the SAR from phantoms, which were prepared with a mixture of commercial
magnetite in non-functionalized powder and distillated water.
II. MATERIALS AND METHODS

The assembly of the induction heater includes three

stages of construction which are summarized as follows: (1)
control for pulse width modulation (PWM) of low power;
(2) switching the dc current at high frequency and power; and
(3) resonance of the electric current in the circuit CL seriesCL parallel.
The PWM control is carried out using an integrated circuit of rectangular oscillation with adjustable frequency and
duty cycle. In particular we have chosen a rate frequency of
206 2 kHz with a duty cycle at 40% in high, 60% in low
and including a small dead time. From this output and using a
logical gate, two rectangular signals with a phase shift of 180
are produced. Afterwards they are amplified by using a field
effect transistor of low power. Subsequently, each signal obtained is transformed into two identical signals by using a ferrite core transformer; at the output we have one rectangularlike pulse with amplitude adjustable for each terminal
of the H-bridge. The signals measured with the oscilloscope
of the PWM driver only for two terminals of the H-bridge and
the duty cycle is shown in Figure 1(a). These rectangular-like
pulses present a shift of 180o between the phases; the negative values of the rectangles ensure the cut-off state, the positive ones release the saturation of the power transistors, and
the dead time small avoids the activation of two transistors in
the same branch.
The stage of current switching (inverter) includes four
power transistors in configuration of full H-bridge fed with
a dc power supply. The dc source was built by rectifying the
output of a voltage transformer form 110 V to 170 V using
four rectifier diodes plugged to a high capacitance bank of
five condensers in parallel and a smoothing LC filter in series.
Cheng et al.21, 22 studied this configuration of the source and
they reported a high reduction of interference in the inverter,
indeed we also have adapted a commercial dimmer regulator
of high power to control the amplitude; and the maximum reg-

Rev. Sci. Instrum. 82, 114904 (2011)

ulated voltage is VDC = 240 V. Each power transistor (MOSFET) has its own antiparallel rectifier diode (D1, D2, D3, and
D4) between the drain and source, also a RC parallel circuit
in order to ensure the correct draining of the remaining current when the transistor falls in the cut-off condition. The output of the inverter is coupled to a resonant electrical circuit
Lr Cr (with Cr = 0.1 f and Lr plugged in parallel) through
an electrical matching circuit CM LM (with CM = 0.39 f and
LM = 6 H plugged in series), see Figure 1(b). The task of
the matching circuit is impedance coupling between the inverter and the tank circuit, this kind of electronic topology
CL-CL has been also reported in Ref. 23; for its part the CM
blocks a possible flow of dc into the tank, avoiding a short
circuit with the power supply. The Lm filters the higher frequency components of the square-like output of the inverter.
The magnitude of the total impedance of our CL-CL circuit
|ZT | = 7.04  can be calculated by Eq. (2), which depends
of the parameters mentioned above, plus the resistances R L m
= 5 m, R L r = 1 m, and the working frequency f,

2

1
2
|Z T | = R L m + 2 f L m
2 f Cm

2
R Lr
+ f 2 L r2
+
.
(2)
2
R Lr
f 2 Cr2 + |1 f 2 L r Cr |2
The resonant sample is under study. This coil is specially designed to generate a uniform magnetic field region of
Eq. (4) 0.6 cm 0.5 cm, it was arranged with N = 25 turns
of cooper distributed on three layers; also it has h = 6 cm
of length, inner diameter id = 5 cm and inductance Lr = 6.0
0.5 H. It is important to note that the maximum current
and frequency of the resonant circuit, in ideal conditions, satisfies Eqs. (3) and (4), respectively,
fR =

1
,

2 L R C R

Vdc C R
.
IR =
LR

(3)

(4)

Finally, the maximum intensity of the magnetic field in the

center of the coils is given by Eq. (5),
0 IRMax N
.
(5)
h
Due to the high current in circulation on Lr , it was immersed in a water chiller with a flux of wf = 1.0 l/s to maintain it at room temperature. A block diagram of the setup
and the schematic representation of the cooled work coil
with the phantom inside are shown in Figures 2(a) and 2(b),
respectively. More details about the setup are presented in
Figure 2(b). There is a water container shown with a cylindrical cavity of 4 cm of diameter where an adiabatic cavity
of 2 cm is also conditioned to hold the sample under study.
Inside of the tube with the sample is fixed a Fluoroptic transducer IPITEK 500 for recording the temperature and which
does not interact with the ac magnetic field.
In order to show the ability of heating magnetic nanomaterials and obtain an evaluation of the SAR, a set of five
BMax =

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24

114904-3

Cano et al.

Rev. Sci. Instrum. 82, 114904 (2011)

FIG. 2. (a and b) Block diagram of the setup and the schematic representation of the device, respectively.

phantoms of each fixed concentration were prepared using

non-functionalized powder of commercially prepared magnetite Fe3 O4 (biferrox) with distillated water, hence a total of
15 samples were studied, see Table I. The average diameter of
the magnetic multidomains is from 0.1 m to 3 m, according to the specifications of the manufacturers.
The biferrox specific heat capacity c was measured using a scanning differential calorimeter model DSC6, Perkin
Elmer. In the interval from 30 C to 65 C the specific heat exhibits a slightly linear decreasing rate and the average value is
c = 0.37 0.01 JK1 /g, this value is around 11 times lower
than the c of the water.25
Finally, to determine the amplitude of the magnetic field
generated in the work coil and its THD, the current in the
tank circuit was measured using a Hall effect current probe;
model Tektronix P6303 with its own amplifier AM503, which
was directly coupled to the input of a digital oscilloscope Tektronix DPO7104. Data file was stored in a flash memory.
III. RESULTS AND DISCUSSIONS

An inspection on the resonant inverter operation free of

phantoms was carried out beginning with the matching stage,
by analyzing the current flow on the Lm Cm series circuit with
the H-bridge powered with 240 V. Figure 3(a) shows the output voltage of the PWM (rectangular-like shapes with a short
dead time) and the current on the matching circuit, theoretically calculated in the LT-Spice platform. It is possible to observe that each rectangular trace coincide on two points of
the current one, and one of them is the maximum or minimum peak value around of 10 A. These facts are also exhibited in the measured signals displayed in Figure 3(c) where
the output of our PWM circuit and the corresponding current
trace obtained from the oscilloscope are shown. Additionally,
in Figure 3(b) the theoretical current signals on Lm Cm and
Lr Cr , respectively, are overlapped. There, the maximum value

of the current in Lr (sinusoidal) results three times higher than

the corresponding value on Lm , besides the phase shift between the peak values is negligible. This behavior is also in
very good concordance with the measured signals shown in
Figure 3(d), where too it can see the overlapped signals from
the oscilloscope. In general all these findings observed in the
matching stage are in agreement with.24, 26
Moreover, to analyze the working of the parallel resonant circuit Lr Cr is displayed Figure 4(a). In this figure we
have the image stored form the oscilloscope, which contains
the sinusoidal graphs of the voltage with a maxim amplitude
around of 100 V (Chanel2 Ch2), and maxim current amplitude of 13 A (Chanel3 Ch3) as well as the FFT of the
current respectively (thin line M), these measurements are
performed without phantom inside of the work coil. Analyzing the maximum values of the voltage and the null value of
the current we found a shift in the phase of 90 , which indicates an efficient exchange between the capacitive and inductive energy of our electronic topology, this expected behavior is in agreement with the simulation of the circuit obtained with the LT-Spice (data not shown) and with Refs. 24,
26, and 27. Meanwhile in the FFT appears the main frequency, the second harmonic with a very small amplitude
and a zoom in the signal is needed to determine the amplitude of the following harmonics, this fact is an evidence
of a very low distortion of the signal, around the 0.9% (see
Eq. (6)). In Figure 4(b) the maximum peak to peak current
amplitude Imax = 58 2A recorded for the resonant circuit with an input dc voltage around 240 V is shown, those
values are in quite good agreement to Eqs. (3) and (4). With
this current amplitude, the above characteristic of the work
coil and Eq. (4), the maximum intensity of magnetic field is
around Bmax = 15.2 0.5 mT, whereas it is required 156 W
of total power to run the setup, this is 2.4 higher than the
used by Lacroix et al.28 to generate an ac magnetic field of
19.3 mT and 100 kHz. By continuing the inspection of this

TABLE I. SAR experimental results.

Phantom type
P. type 1
P. type 2
P. type 3

Amount
(Fe/ml)

Heat capacity C
(JK1 /g)

Initial slope
(dT/dt) (K/s)

Concentration
(Fe/ml)

SAR
(W/g)

Fe3 O4 (0.01 ml) + Water (0.5 ml)

Fe3 O4 (0.015 ml) + Water (0.5 ml)
Fe3 O4 (0.02 ml) + Water (0.5 ml)

3.718
3.527
3.356

0.007 0.001
0.011 0.001
0.014 0.002

0.167
0.231
0.285

0.163 0.022
0.168 0.015
0.165 0.024

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24

114904-4

Cano et al.

Rev. Sci. Instrum. 82, 114904 (2011)

FIG. 3. (a) Theoretical signals simulated of the output voltage of the PWM overlaped with the signal of the current flow on the Lm Cm circuit; (b) the current
traces on Lm Cm and Lr Cr ; (C) and (D) the corresponding signals measured with the oscilloscope.

signal and taking into account the maximum and minimum

points, a small superposition of other frequencies is clearly
caused for the changing of the inductance due at the phantom;
nevertheless from the FFT of the current the THD determined
is very close to 1%, considering the amplitudes to the third
harmonic of the main frequency. The maximum increasing of
the THD registered in our experiments is around 10%.

THD =

Y22 + Y32 + + Yn2

Y1

(6)

where Yi : i = 1, 2, . . . , n is the average amplitude of the harmonics of the fundamental frequency Y1 .

On the other hand to perform the measurement of the
temperature vs. time and SAR from each phantom, we have
closely followed the procedure described for Ming et al. and
Cheng21, 29 in these works, the sample is placed inside of the
work coil and the heating start from room temperature TO
(around 30 C) during a while, to reach a top of temperature Tf
(in our case is 40 C), with amplitude and frequency constant.
In Figures 5(a)5(c) the temperature curves of the heated samples are displayed, each graph showing the five temperature

FIG. 4. (a and b) Image obtained from the oscilloscope of the graphs of the voltage, current and the FFT of the current with a dc voltage input of 110 V. Also
the maximum current measured with a dc voltage input of 240 V, respectively.

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24

114904-5

Cano et al.

Rev. Sci. Instrum. 82, 114904 (2011)

FIG. 5. (ac) Temperature curves of the heated samples in each graph are shown the five curves for each concentration.

curves for each concentration used. Each set of curves has

similar initial room temperature of around 30.5 o C, 28.5 o C,
and 27 o C, respectively. Indeed, in Figure 6 the average temperature curve of each kind of phantom is shown, from which
it is possible to observe that the timing for reaching the 40 o C
increase varies markedly in the samples with major concentra-

tion of magnetite. From the lineal region of this curve the initial slope dT/dt was determined, and in Table I the outcomes
of the measurement performed is summarized. A simple analysis of the obtained values of SAR indicates a gradual increasing of the ferromagnetic powder concentration, which implies
a decreasing of the total specific heat in the same proportion,
but a velocity of heating proportionally growing. Moreover,
there are not significant differences between the SAR values
recorded and at the same time the measurements are qualitatively in agreement with those reported by Cheng21 that has
used similar smaller particles and different amplitude of magnetic field.
IV. CONCLUSIONS

FIG. 6. The average curve of the temperature measured for the phantoms of
each concentration.

The implementation, assembly and characterization of a

device developed in order to perform magnetic hyperthermia
in vitro have been presented. This device has shown a low
harmonic distortion of the resonant current, a magnetic field
amplitude and frequency similar at the employed for several
authors for thermotherapy in whole body or rodents of laboratory; furthermore, it has been capable of heating phantoms
arranged with commercial magnetite, which were characterized through the SAR. The option of frequency adjust can be
easily suitable at our device to generate magnetic field in the
range from 100 kHz to 300 kHz with steps of 50 kHz, by

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24

114904-6

Cano et al.

adjusting a single resistance in the PWM stage and simultaneously replacing the appropriate work coil. Nevertheless,
subsequent studies to increase the efficiency of the work coil
must to be realized, in order to enhance the magnetic field amplitude and avoid the heating of the wire due at the skin effect,
such as is recommended in Ref. 28. Another remarkable feature of our induction heater in addition at the compact design,
is the low price involved to its construction because we just
have utilized very commercial components. This heater can
be reproduced with approximately of US $600, this is about
15% of the Fluoroptic system, which are the most expensive
module of our full setup. Indeed the optic detection of temperature to determine the SAR in samples can be replaced for
a single system based in an alcohol thermometer and a videocamera system. This device could be easily implemented in
any cell laboratory in order to study the necrosis in tissues
previously marked magnetically for biomedical applications.
ACKNOWLEDGMENTS

The authors wish to thank Tomas M. Trent for reviewing the language of the paper, also at the Mexican institutions for the partial support in this study: CONACYT
under Grant No. Apoyo-Compl-2009-118168 and 119058;
PROMEP: 103.5-10-1796, COECYTJAL: 06-2009-954 and
06-2009-732; DAIP 017/2010 and CONCYTEG 074568-87.
1 M.

Johannsen, U. Gneveckow, K. Taymoorian, C. H. Cho, B. Thiesen,

R. Scholz, N. Waldfner, S. A. Loening, P. Wust, and A. Jordan, Actas
Urol. Esp. 31(6), 660 (2007).
2 M. Yanase, M. Shinkai, H. Honda, T. Wakabayashi, J. Yoshida, and
T. Kobayashi, Jpn. J. Cancer Res. 89, 463 (1998).
3 R. Hergt, S. Dutz, R. Mller, and M Zeisberger, J Phys.: Condens. Matter
18, s2919 (2006).
4 A. Jordan, R. Scholz, K. Maier-Hau, M. Johannsen, P. Wust, J. Nadobny,
H. Schirra, H. Schmidt, S. Deger, S. Loening, W. Lanksch, and R. Felix, J.
Magn. Magn. Mater. 225, 118 (2001).
5 C. C Berry, J. Phys. D: Appl. Phys. 42, 224003 (2009).

Rev. Sci. Instrum. 82, 114904 (2011)

6 G.

Nedelcu, Dig. J. Nanomater. Biostruct. 3(3), 103 (2008).

Hellstern, K. Schulze, B. Schpf, A. Petri-Fink, B. Steitz, S. Kamau,
M. Hilbe, S. Koch-Schneidemann, L. Vaughan, M. Hottiger, M. Hofmann, H. Hofmann, and B. von Rechenberg, J. Nanosci. Nanotechnol. 6, 1
(2006).
8 J. Carrey, B. Mehdaoui, and M. Respaud, J. Appl. Phys. 109, 083921
(2011).
9 I. S. Poperechny, Yu. L. Rainker, and V. I. Stepanov, Phys. Rev. B 82,
174423 (2010).
10 A. A. Kuznetsov, O. A. Shlyakhtin, N. A. Brusentsov, and O. A.
Kuznetsov, Eur. Cells Mater. 3(2), 75 (2002).
11 G. F. Goya, E. Lima, A. D. Arelaro, T. Torres, H. R. Rechenberg, L. Rossi,
C. Marquina, and M. R. Ibarra, IEEE Trans. Magn. 44(11), 4444 (2008).
12 D. Bahadur and J. Giri, Sadhana 28, 639 (2003).
13 Q. A. Pankhurst, N. K. Thanh, S. K. Jones, and J. Dobson, J. Phys. D: Appl.
Phys. 42, 224001 (2009).
14 M. Bettge, J. Chatterjee, and Y. Haik, Biomagn. Res. Technol. 2(4), 1
(2004).
15 F. Matsuoka, M. Shinkai, H. Honda, T. Kubo, T. Sugita, and T. Kobayashi,
Biomagn. Res. Technol. 2(3), 1 (2004).
16 A. Jordan, R. Scholz, P. Wust, and H. Schirra, J. Magn. Magn. Mater. 194,
185 (1999).
17 R. Ivkov, S. J. DeNardo, W. Daum, A. R. Foreman, R. C. Goldstein,
V. S. Nemkov, and G. L. DeNardo, Clin. Cancer Res. 11(19), 7093s (2005).
18 S. Dieckerhoff, M. J. Ryan, and R. W. De Doncker, IEEE-IAS Annual Meeting. Conference on Digital Object Identifier Record 1999, San Francisco,
CA, 3, 1999, pp 20392045.
19 S. V. Mollov, M. Theodoridis, and A. J. Forsyth, IEE Proc.: Electr. Power
Appl. 151(1), 12 (2004).
20 A. Kelemen, I. Szekely, and N. Kutasi, Pollack Periodica 2(1), 27 (2007).
21 Cheng-Chi Tai and Chien-Chang Chen, PIERS Online 4(2), 276 (2008).
22 Cheng-Chi Tai and Ming-Kun Cheng, PIERS Online 4(2), 417 (2008).
23 G. F. Goya, N. Cassinelli, and M. R. Ibarra, Patent W0 2009/135970A1 (12
November 2009).
24 M. H. Rashid, Power Electronics Handbook Academic (Academic,
San Diego, 2001), p. 231.
25 J. L Fortier, P. A. Leduc, and J. E. Desnoyers, J. Solution Chem. 3(4), 323
(1974).
26 K. S. Ashoka, IEEE Trans. Power Electron. 8(1), 1 (1993).
27 Zhongming Ye, K. J. Praveen, and C. Paresh, IEEE Trans. Ind. Electron.
54(5), 2831 (2007).
28 L. M. Lacroix, J. Carrey, and M. Respaud, Rev. Sci. Instrum. 79, 093909
(2008).
29 Ming Ma, Ya Wu, Jie Zhou, Yongkang Sun, Yu Zhang, and Ning Gu, J.
Magn. Magn. Mater. 268, 33 (2004).
7 D.

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 200.23.6.133 On: Mon, 08 Feb 2016 16:40:24