IEEE TRANSACTIONS ON MAGNETICS, VOL. 46, NO. 2, FEBRUARY 2010
467
AC Magnetic Properties of Fe-Based Composite Materials Peter Kollár1 , Já Füzer1 , Radovan Bureˇs2 , and Mária Fáberová2 ˇ Department of Condensed Matter Physics, Faculty of Science, P. J. Safárik University, Koˇsice 04180, Slovakia Institute of Materials Research, Slovak Academy of Sciences, Koˇsice 040 01, Slovakia We investigated the influence of resin content on ac magnetic properties of Fe-based composite material to extend possibilities for application of this kind of material at higher frequency. The samples of composite material with different content of phenol-formaldehyde resin acting as insulator were prepared by conventional powder metallurgy. The samples in the form of the ring were used for ac magnetic measurements and in the form of cylinder for specific resistivity measurements. The ac magnetic properties (sample in the form of ring) were measured at maximum flux density up to 0.4 T by MATS-2010SA loop tracer in a frequency range 1–150 kHz and obtained total power losses. The dc hysteresis loops were measured by a fluxmeter-based hysteresisgraph. The values of the specific resistivity measured by the van der Pauw method (sample in the form of cylinder) were used for calculation of eddy current losses. All magnetic properties of the composite rings were compared with the properties of the material prepared from the powder provided by Höganäs AB Sweden by the technology usually used by the producer. Index Terms—Composite material, eddy current, hysteresis loop, losses. TABLE I TECHNOLOGIC PARAMETERS OF THE SAMPLES
I. INTRODUCTION
OFT magnetic composites is a well-known group of materials suitable for various ac (in an alternating magnetic field, alternating current) and dc (in a quasistatic magnetic field, direct current) electromagnetic applications as cores with three-dimensional isotropic ferromagnetic behavior for transformers, electromotors, and electromagnetic circuits, sensors, electromagnetic actuation devices, low-frequency filters, induction field coils, magnetic seal systems, and magnetic field shielding [1]. Many scientists have an ambition to replace electrical steel sheets or ferrites in some applications by these composites [1]. The composite materials consist of small particles of ferromagnetic material (pure elements of Fe, Co, Ni, or alloys based on them) isolated by an electrical insulating film and exhibit soft magnetic properties as low hysteresis losses, low eddy current losses, high permeability at low magnetic field, and high saturation flux density [2], [3]. The aim of this work was to investigate the influence of resin content on ac magnetic properties of Fe-based composite material to extend possibilities for application of this kind of material at higher frequency.
S
II. SAMPLES PREPARATION We used conventional powder metallurgy for preparation of soft magnetic composite samples. We choose pure iron powder ASC 100.29, Höganäs AB, Sweden [4], as ferromagnetic material and phenol-formaldehyde resin (Bakelite ATM) as insulator. The particles of the iron powder exhibit (according to material producer) following particle size distribution: 50 wt. % particles have the diameter below 75 m and 0.9 wt. % above 180 m.
Manuscript received June 20, 2009; revised September 14, 2009; accepted September 21, 2009. Current version published January 20, 2010. Corresponding author: P. Kollár (e-mail:
[email protected]). Digital Object Identifier 10.1109/TMAG.2009.2033338
We milled the bakelite for 3 30 s in a knife mill to obtain the particle size below 100 m. We prepared the experimental samples by mixing of 90 vol. % of iron (sample A), 80 vol. % of iron (sample B), and 70 vol. % of iron (sample C) and resin in a Turbula mixer. The mixed powder was pressed in a cylindrical die at a uniaxial pressure to obtain compacts in the form of cylinder or ring. After that, we cured the resulting product at a temperature for 60 min in an electric furnace in air atmosphere. of 165 Parameters of the samples are in Table I. The thermoset phenol-formaldehyde resin covers the iron powder particles; see Fig. 1. The resin electrically insulates conductive iron particles and creates the mechanical bond after curing; see Fig. 2. Sample S is a reference material, prepared from Somaloy 700 powder (covered by insulating film) provided by Höganäs AB, Sweden, by producer-developed technology (compaction at 800 MPa and heat treatment at 530 for 30 min in air). Parameters of the sample are in Table I. III. EXPERIMENTAL We investigated the structure of the samples by scanning electron microscopy (JEOL JSM-7000F, TESLA BS 340). AC hysteresis loops were measured by an ac hysteresisgraph Model MATS-2010SA at maximum flux density of 0.05, 0.1, and 0.2 T in the frequency range 1–150 kHz. The total losses were calculated directly by the hysteresisgraph. The specific electrical resistivity of the composite material in the form of the cylinder with diameter of about 24 mm and from 2 to 4 mm height was
0018-9464/$26.00 © 2010 IEEE
468
IEEE TRANSACTIONS ON MAGNETICS, VOL. 46, NO. 2, FEBRUARY 2010
Fig. 3. DC hysteresis loops of samples A, B, C, and S measured at maximum flux density B : T.
= 02
Fig. 1. The view of the fracture area of the bulk sample C shows Fe particle coated by resin, observed by SEM.
Hysteresis losses are assumed to be frequency-independent and can be obtained by extrapolating the total losses fre. quency dependence to The eddy current losses can be expressed according to [6], [7] by (3) where is effective dimension (for the material prepared by casting, is thickness of the sample, for laminated material is maximum flux density, frethickness of the sheet), quency, specific resistivity, density of the material, and geometrical coefficient. The rectangular cross section perpendicular to the direction of magnetic flux is (4) where
Fig. 2. The view of the polished area of the bulk sample C shows iron powder particles (dark) and resin (white) in interparticle spaces, observed by SEM.
measured by the van der Pauw method (assuming to be isotropic in the plane parallel to the cylinder base) [5].
IV. POWER LOSSES The total power losses per cycle (in units J/kg) consist of , eddy current , and anomalous [6], [7] hysteresis
is width and
height of the rectangle.
V. RESULTS The dc hysteresis loops (flux density dependencies on the external magnetic field) measured at maximum flux density T are plotted in Fig. 3. The tilt of the loops increases with the increase of the resin content caused by the increase of the inner demagnetization factor [11]–[13]. The iron powder particles create a demagnetization field (which is ruled by the mean distance between a pair of ferromagnetic particles), oriented opposite to the external magnetic field , causing the decrease of the internal magnetic [6] field
(1) (5) According to this theory, the experimental value of total losses of some material can be separated to the components [8]–[10]. Multifying (1) by frequency, we obtain total power losses in units W/kg (2)
is magnetization of the powder particle. where The hysteresis losses of all samples calculated from hysteresis loops are in Table II. The total losses dependencies as a function of maximum flux measured at the frequency kHz of all samdensity ples are depicted in Fig. 4.
KOLLÁR et al.: AC MAGNETIC PROPERTIES OF Fe-BASED COMPOSITE MATERIALS
Fig. 4. Total losses P as a function of maximum flux density B kHz. for samples A, B, C, and S at the frequency f
= 10
measured
469
Fig. 5. Total losses P as a function of the frequency f for samples A, B, C, and S measured at maximum flux density B : T.
=02
TABLE II ELECTROMAGNETIC PARAMETERS OF THE SAMPLES
In (4). Experimental value measured at the frequency of 10 kHz and maximum flux density of 0.2 T. Value measured at quasi-static magnetic field at maximum flux density of 0.2 T Value calculated by (3) for frequency of 10 kHz and maximum flux density of 0.2 T
Fig. 6. Total losses P as a function of the frequency f for samples A and S : , 0.1, and 0.2 T. measured at maximum flux density B
= 0 05
The values of total losses as a function of maximum flux density increase similarly for samples A, B, and C and less steeply than that for reference sample S. We assume that lower values of losses of samples A, B, and C are caused by lower contribution of the eddy current losses to the total losses due to higher specific resistivity of these samples in comparison to the specific resistivity of sample S; see Table II. The difference between losses versus flux density dependencies of samples A, B, and C is negligible. We explain this effect by low contribution of the eddy current losses, which decrease with the increase of volume content of the resin in the composite material. The largest component of the total power losses of these samples are the hysteresis losses. For calculation of eddy current losses according to (3), it is necessary to know the effective dimension , which is very difficult to obtain for composite material. The value of the height of the cross section of the area in which eddy currents flow is un, we obtain known. If we take the largest value of , i.e., the largest theoretically possible value of eddy current losses, which is still very small in comparison to total losses and moreover decreases with the increase of the resin content due to the increase of specific resistivity. The total losses dependencies as a function of frequency meaT of all samples sured at the maximum flux density are depicted in Fig. 5. The slope of total losses dependencies versus frequency of samples A, B, and C is smaller than that for sample S. It is
caused by a lower contribution of eddy current losses of samples A, B, and C due to higher specific resistivity of the composite material. Similar tendencies have the frequency dependencies of total losses of the sample A measured at maximum flux densities , 0.1, and 0.2 T is shown in Fig. 6. We assume that sample A containing 90% of iron is more suitable for application at frequencies above 1 kHz than sample S. The absolute values of total losses and the slope frequency dependence of total losses of composite material are lower than that for metallic compacted powder Fe-Ni-based material [14]. The hysteresis loops measured at maximum flux density of 0.1 T for frequencies of 2 and 50 kHz of samples A and S are compared in Figs. 7 and 8. The form of the hysteresis loops of both samples measured at 2 kHz is similar (the values of total losses and coercivity differ negligibly), only the loop of sample A is more tilted (because of larger demagnetization field) than the loop of sample S, confirming the opinion that sample S is more suitable for application at low frequencies. The peak permeability of sample S is higher than that for sample A. The slope (reflecting the peak permeability) slightly decreases in the frequency range 2–50 kHz for both materials. The form of the hysteresis loop of sample S indicates that total losses and coercivity increase significantly with the frequency than that for sample A.
470
IEEE TRANSACTIONS ON MAGNETICS, VOL. 46, NO. 2, FEBRUARY 2010
rial causes the increase of the demagnetization field and consequently the increase of the tilt of the hysteresis loops. ACKNOWLEDGMENT
Fig. 7. Hysteresis loops of samples A and S measured at maximum flux density B : T and frequency f kHz.
=01
=2
This work was realized within the frame of the project “Centre of Excellence of Advanced Materials with Nano- and Submicron-Structure,” which is supported by the Operational Program “Research and Development” financed through the European Regional Development Fund. This work was also supported by the Slovak Research and Development Agency under the Contract No. APVV-0490-07 MICOMAT and by the Scientific Grant Agency of the Ministry of Education of Slovak Republic and the Slovak Academy of Sciences, Project No. VEGA 1/4020/07 and No. VEGA 2/0149/09. Special thanks to Höganäs AB, Sweden, for providing Somaloy powder. REFERENCES
Fig. 8. Hysteresis loops of samples A and S measured at maximum flux density : T and frequency f kHz. B
=01
= 50
VI. CONCLUSION In the investigation of ac magnetic properties of composite material based on iron powder (90 vol. %) and phenol-formaldehyde resin (10 vol. %) as insulator, we found that the frequency dependence of total losses increases more slowly than that for reference Somaloy 700 powder-based sample. It is caused by the decrease of eddy current losses of composite material in which hysteresis losses become dominant. Further increase of the specific resistivity of the composite material (sample with higher resin content) does not have significant influence on the value of total losses, although it lowers eddy current losses’ contribution to the total losses. The slope of the dc and ac hysteresis loops is strongly influenced by inner demagnetization factor. The decrease of the iron powder content in the composite mate-
[1] T. Maeda, H. Tyoda, N. Igarashi, K. Hirose, K. Mimura, T. Nishioka, and A. Ikegaya, “Development of super low iron-loss P/M soft magnetic materials,” SEI Tech. Rev., vol. 60, pp. 3–9, Jun. 2005. [2] H. Shokrollahi and K. Janghorban, “Soft magnetic composite materials (SMCs),” J. Mater. Process. Technol., vol. 189, pp. 1–12, Jul. 2007. [3] A. H. Taghvaei, H. Shokrollahi, and K. Janghorbana, “Properties of iron-based soft magnetic composite with iron phosphate-silane insulation coating,” J. Alloys Compounds, vol. 481, pp. 681–686, Jul. 2009. [4] “Product catalog: Somaloy metal powders for SMC components,” Höganäs AB, Sweden [Online]. Available: http://www.hoganas.se, [5] L. J. van der Pauw, “A method of measuring the resistivity and Hall coefficient on lamellae of arbitrary shape,” Philips Techn. Rev., vol. 20, pp. 220–224, 1958. [6] R. M. Bozorth, Ferromagnetism, 3rd ed. Piscataway, NJ: IEEE Press, 1993. [7] G. Bertotti, Hysteresis in Magnetism. San Diego, CA: Academic, May 1998. [8] T. D. Shen, U. Harms, and R. B. Schwarz, “Bulk Fe-based metallic glass with extremely soft ferromagnetic properties,” Mater. Sci. Forum, vol. 386–388, pp. 441–446, 2002. [9] H. Chiriac, N. Lupu, and M. Tibu, “Design and preparation of new Fe-based bulk amorphous alloys torroids,” IEEE Trans. Magn., vol. 39, no. 5, pt. 2, pp. 2040–2042, Sep. 2003. [10] R. Piccin, P. Tiberto, H. Chiriac, and M. Baricco, “Magnetic properties and power losses in Fe-Co-based, bulk metallic glasses,” J. Magn. Magn. Mater., vol. 320, pp. e806–e809, Oct. 2008. [11] J.-L. Mattei and M. Le Floc’h, “Percolative behaviour and demagnetization effects in disordered heterostructures,” J. Magn. Magn. Mater., vol. 257, pp. 335–345, Feb. 2003. [12] J.-L. Mattei and M. Le Floc’h, “Effects of the magnetic dilution on the ferrimagnetic resonance of disordered heterostructures,” J. Magn. Magn. Mater., vol. 264, pp. 86–94, Aug. 2003. [13] M. Anhalt, B. Weidenfeller, and J.-L. Mattei, “Inner demagnetization factor in polymer bonded soft magnetic materials,” J. Magn. Magn. Mater., vol. 320, pp. e844–e348, Oct. 2008. [14] P. Kollár, D. Olekˇsáková, J. Füzer, J. Kovác, S. Roth, and K. Polanski, “The magnetic properties of powdered and compacted microcrystalline permalloy,” J. Magn. Magn. Mater., vol. 310, pp. 2609–2611, Mar. 2007.
