Investigation of magnetic properties from a manganese–zinc

Abstract

This paper presents a new type of a polymer bonded soft magnetic material (PBSMM) which can be used to create complex magnetic core geometries. The PBSMM consists of two separate materials, manganese zinc ferrite powder and polydimethylsiloxane (PDMS) as matrix material. The magnetic core losses of the PBSMM where analyzed at 10 kHz and 1 MHz with self-fabricated toroidal core samples by using a custom-made measurement setup in accordance with the IEC-62044-3 norm. The complex permeability was determined by evaluating impedance analyzer measurements between 10 kHz and 1 MHz. It is pointed out that the investigated material can be used for power electronic filter applications and opens the opportunity to create geometrical complex inductive components.

Keywords

Soft magnetic-polymer manganese–zinc–ferrite high frequency

1. Introduction

In most power electronic applications, the construction space for passive components, especially for inductive components is limited [1,2]. Therefore, innovative ideas are needed for creating new core shapes for example by 3D printing [3] or other manufacturing techniques to fill out the available space. This paper characterizes the magnetic properties (core losses and complex permeability) of a new polymer bonded soft magnetic materials (PBSMM) by measurements with a hysteresis loop measurement setup and impedance analyzer. The investigated toroidal cores were manufactured by casting mold technique offering the opportunity to be used in printing applications and other manufacturing processes.

2. Description of the inductors

To manufacture test inductors with polymer bonded soft magnetic materials (PBSMM), manganese- zinc-ferrite powder, purchased by TRIDELTA Weichferrite GmbH, was mixed with polydimethylsiloxane (PDMS) from Dow Corning Corporation with a weight ratio of 75% powder and 25% PDMS. The particles inside the commercial powder are nearly round-shaped and exhibit a wide diameter range. Prior to mixing the powder was sieved to receive powders with different particle diameter fractions. The obtained powders were labeled by the desired main particle size of 20 μm, 50 μm, 80 μm, 100 μm and 200 μm. The sieved powders were analyzed by the manufacturer with a “CILAS 1090 Naß” particle size analyzer. Figure 1 shows a nearly Gaussian density distribution in dependence of the particle size diameter for most powders. Only the 200 μm powder shows a higher fraction of large particles rearwards the 200 μm particle size peak.

Fig. 1.

Particle density distribution in % of all used manganese zinc ferrite powders.

The mixed composite was filled in a toroidal casting mold with an inner diameter a of 8 mm, an outer diameter b of 16 mm, a height h of 10 mm and cured for 24 hours at room temperature till the toroidal cores were removed. With each powder five core samples were fabricated. All cores were shaped to adjust the desired height h of 5 mm. Figure 2 shows examples of manufactured toroidal cores ((2) and (3)) and a commercial pressed an sintered manganese-zinc-ferrite core by TRIDELTA Weichferrite GmbH made of the used powder material (1).

Fig. 2.

Sintered and pressed manganese-zinc-ferrite toroidal core (1) and self-manufactured PBSMM example (2), (3).

Prior to electrical characterization the volume filler fraction V _f of the formed test devices was determined because it is known that the magnetic properties are affected by V _f [4]. The volume filler fraction V _f can be calculated by [5,6]: $\begin{eqnarray}\displaystyle V_{\text{f}}=({\rho}_{\text{c}}-{\rho}_{\text{m}})/({\rho}_{\text{f}}-{\rho}_{\text{m}}), & & \displaystyle\end{eqnarray}$ (1) were 𝜌_f is the density of the powder particles, 𝜌_m the density of the polymer, and 𝜌_c the density of the core material. According to the data sheets of the manufacturers 𝜌_f and 𝜌_m are 4.8 g/cm³ and 0.98 g/cm³, respectively. To determine 𝜌_c the toroidal cores were weighed with a high precision scale and the geometrical parameters (a, b and h) were measured precisely with a caliper gauge to calculate the volume of each core. Figure 3 illustrates the achieved volume filler fractions V _f of the constructed ring core devices depending of the used powders. The filler fraction varies between 32% (20 μm powder) and 47% (100 μm powder). The commercial pressed and sintered core exhibit a density of approximately 4.8 g/cm³ which corresponds to a filler fraction of nearly 100% (not shown in Fig. 3). As described in [7], the particle size and volume filler fraction influences the magnetic properties. The influence of these parameters onto the core loss density and complex permeability is investigated in the next sections.

3. Characterization of magnetic properties

3.1. Measuring setup

The magnetic properties were measured with a custom-made hysteresis loop measurement setup in accordance with the IEC-62044-3 norm. The setup (Fig. 4) consists of signal generator, power amplifier, current sensor, probe, oscilloscope, commercial software MATLAB^®. It offers the opportunity to generate sinusoidal currents in a frequency range between 10 kHz and 1 MHz. To determine the complex permeability in a frequency range of 10 kHz and 10 MHz, a high precision impedance analyzer (Agilent Technologies 4294A) was used.

3.2. Core loss density

The core loss density p _l of a ring core sample can be calculated by (2) as showed in [8]: $\begin{eqnarray}\displaystyle \overline{p_{l}}=\frac{1}{T}\int _{0}^{T}H(t)\cdot \frac{\partial B(t)}{\partial t}dt. & & \displaystyle\end{eqnarray}$ (2)

For this calculation, the period T of the measured signal is needed as well as the time depending signals of the magnetic field strength H(t) and the magnetic flux density B(t). These data were evaluated by measuring the time depending current and voltage with the described measurement setup (Fig. 4).

3.3. Complex permeability

One of the most important properties is the complex permeability. As described in [8], the complex permeability for toroidal cores can be calculated by (3): $\begin{eqnarray}\displaystyle \text{}\underline{{\mu}}={\mu}\prime -\text{j}{\mu}\prime \prime =L_{\text{S}}\cdot \left(N_{\text{P}}^{2}\cdot \frac{h}{2{\pi}}\cdot \text{ln}\left(\frac{b}{a}\right)\right)^{-1}-\text{j}\cdot R_{\text{S}}\cdot \left({\omega}\cdot N_{\text{P}}^{2}\cdot \frac{h}{2{\pi}}\cdot \text{ln}\left(\frac{b}{a}\right)\right)^{-1}, & & \displaystyle\end{eqnarray}$ (3) with the real part 𝜇′, imaginary part 𝜇′′, angular frequency ω and number of turns of the primary winding N _P. Based on the equivalent circuit model for coils L _S is the series inductance and R _S is the series resistance. L _S and R _S can be determined by impedance analyzer measurements. Further the parallel capacitance of the equivalent circuit model can be neglected for frequencies below 10 MHz due to the low number of turns of the test devices.

Fig. 3.

Volume filler fraction of the test devices.

Fig. 4.

Schematic drawing of the hysteresis loop measurement equipment.

4. Experimental results

4.1. Characterization of core losses

To determine the core losses of the test devices hysteresis loop measurements were performed. Figure 5 shows the hysteresis loop measurements at 10 kHz and a peak flux density of 15 mT of the PBSMM toroidal cores with different particle sizes. For each particle size the sample with the highest filler fraction was measured. It can be seen that the slope of the hysteresis loop correlates with the filler fraction. The core with the highest filler fraction of 47% (100 μm-powder) exhibit the highest slope and the core with the filler fraction of 41% (20 μm-powder) the lowest. This observation can also be seen for a frequency of 1 MHz in Fig. 6, whereas no systematic correlation of the slopes of the B-H-curves with the particle size can be noticed.

Fig. 5.

Hysteresis loops of the PBSMM (highest filler fraction) at 10 kHz.

Fig. 6.

Hysteresis loops of the PBSMM (highest filler fraction) at 1 MHz.

Further hysteresis loop measurements with different peak flux densities were performed to identify the core losses for different saturation degrees. Figure 7 shows the calculated core losses at 10 kHz and 1 MHz in dependence of the peak flux density. Besides the molded PBSMM cores, a commercial pressed and sintered toroidal manganese-zinc-ferrite-core by TRIDELTA Weichferrite GmbH of the used powder material was evaluated. The solid lines show the 10 kHz and the dot-stroke-lines the 1 MHz results for the measured PBSMM test devices. The losses of the industrially pressed toroidal core device at a frequency of 10 kHz are not plotted because the losses are lower than 0.1 mW/cm³.

Fig. 7.

Core losses p _l of test devices at different frequencies f and magnetic flux densities B _Peak.

Fig. 8.

Complex permeability calculation results for all 50 μm-powder test devices.

It can be noticed that the losses of the PBSMM devices are much higher compared to those of the industrial device. This result was expected, as for the PBSMM devices the soft magnetic material does not fill out the complete core volume and therefore the expected core losses are higher than the losses of the raw material pressed and sintered toroidal core [4,9]. Considering the core losses of the PBSMM devices it can noticed that there is no clear dependence of the losses on the powder particle size. However, it was expected in [1] that with increasing main particle size the core loss density decreases due to the fact that smaller particles saturate earlier than larger particles.

4.2. Characterization of complex permeability

The complex permeability is considered in a frequency range from 10 kHz up to 10 MHz. Figure 8 shows the determined complex permeability for the five test devices manufactured of the 50 μm-powder. The maximum real part of the permeability at 10 kHz is 10 (V _f of 42%) and the lowest is 8.8 (V _f of 33%). Due to [4], Fig. 8 points out that the devices with the highest filler fraction exhibit the highest real part of the complex permeability (𝜇′∕𝜇₀). It can also be seen that the real part of the permeability is nearly constant up to a frequency of 1 MHz. The results show that there is no direct context between the real and imaginary part of the test devices. The device with a filler fraction of 40% shows the highest imaginary part with 𝜇′′∕𝜇₀ = 11.5. Figure 9 shows the results of the devices with the highest filler fraction from all used powders. The device with 100 μm powder (filler fraction of 47%) shows the highest permeability with 12 at a frequency of 10 kHz. The device with 20 μm-powder exhibits the lowest real part at 10 kHz with a value of 9 and a filler fraction of 41%. As shown in [10], it is expected that the real part of the permeability of the PBSMM is much less than the real part of the industrially pressed toroidal core (𝜇′∕𝜇₀ > 2000). Furthermore, in the considered frequency range, the real and imaginary part of the complex permeability decreases [11,12]. These results agree with [13] where a permeability of 12 and 5 for a 60 μm and 30 μm manganese-zinc-ferrite powder fraction were shown.

Fig. 9.

Complex permeability calculation results for test devices with highest filler fraction.

5. Conclusions

This paper shows an analysis of a new manganese-zinc-ferrite polymer bonded soft magnetic material. The used soft magnetic powder was sieved in five separate powder fractions from 20 μm up to 200 μm. The powders were mixed with PDMS to manufacture toroidal cores with different filler fractions. It can be seen that the manufacturing process enables realization of complex core geometries by using casting mold technology. The filler fraction of the toroidal cores varies between 32% and 47%. Furthermore, the core losses for the PBSMM and a commercially available sintered pressed toroidal core from the same powder were investigated at 10 kHz and 1 MHz. The measurements showed that the losses of the PBSMM are significant higher than of the pressed material at both frequencies. In addition to that there is no significant relation between the particle size distribution and core loss density. The investigation of the complex permeability indicates that the real part of the permeability is much lower than for industrially pressed toroidal core. It is also pointed out that for volume filler fractions less than 50% only the filler fraction determines the magnetic properties and not the particle size distribution. Further, the manufacturing process must be optimized to achieve constant volume filler fractions. The experimental results showed that the investigated material can be used for filter applications in power electronic devices where a constant permeability over a wide frequency range is required. Furthermore it was shown that PBSMM offers the opportunity to build up magnetic core geometries for inductive components apart of standard magnetic core geometries.

References

Egelkraut

, Charakterisierung und Applikation weichmagnetisch gefüllter Kunststoffe für induktive Komponenten leistungselektronischer Baugruppen, Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, 2011.

Egelkraut

Frey

Rauch

Schletz

and März

, Polymer bonded soft magnetics for EMI filter applications in power electronics, in: Twenty-Fifth Annual IEEE Applied Power Electronics Conference and Exposition (APEC), 2010, pp. 231–238.

Huber

, 3D print of polymer bonded rare-earth magnets, and 3D magnetic field scanning with an end-user 3D printer, Appl. Phys. Lett. 109(16) (2016)162401.

Anhalt

and Weidenfeller

, Permeability of soft magnetic FeCoV-composites for varying filler fractions, IEEE Trans. Magn. 46(2) (2010), 440–442.

Mamunya

Davydenko

V.V.

Pissis

and Lebedev

E.V.

, Electrical and thermal conductivity of polymers filled with metal powders, European Polymer Journal 38(9) (2002), 1887–1897.

German

R.M.

and Park

S.J.

, eds. Handbook of Mathematical Relations in Particulate Materials Processing: Ceramics, Powder Metals, Cermets, Carbides, Hard Materials, and Minerals, John Wiley & Sons, Inc., Canada, 2008.

Anhalt

, Systematic investigation of particle size dependence of magnetic properties in soft magnetic composites, Journal of Magnetism and Magnetic Materials 320(14) (2008), e366–e369.

Stadler

, Messtechnische Bestimmung und Simulation der Kernverluste in weichmagnetische Materialien, Thesis, Friedrich–Alexander–Universität Erlangen-Nürnberg (FAU), Erlangen, 2009.

Svensson

Frogner

Jeppsson

Cedell

and Andersson

, Soft magnetic moldable composites: Properties and applications, Journal of Magnetism and Magnetic Materials 324(18) (2012), 2717–2722.

10.

Hultmann

L.O.

and Jack

A.G.

, Soft magnetic composites - Materials and applications, IEEE International Electric Machines and Drives Conference IEMDC’03 (1) (2003), 516–522.

11.

Rekosova

Dosoudil

Usakova

Usak

and Hudec

, Magnetopolymer composites with soft magnetic ferrite filler, IEEE Trans. Magn. 49(1) (2013), 38–41.

12.

Verduzco

J.A.

, Magnetic properties of metal – Polymer composites or Ac applications at high frequencies, J. Mex. Chem. Soc. 2016 60(2) (2016), 58–61.

13.

Kazantseva

N.E.

, Magnetic behaviour of composites containing polyaniline-coated manganese–zinc ferrite, Journal of Magnetism and Magnetic Materials 269(1) (2004), 30–37.

Investigation of magnetic properties from a manganese–zinc–ferrite polymer bonded material