Influence of sintering temperature on the magnetic

9 downloads 0 Views 4MB Size Report
Dec 1, 2017 - applied in power converters [1–3], power inductors [4], tun- able antennas ..... size, coercivity Hc decrease, while remanence Br increase. (8). Pv = fB2 m. p .... electroscience.com/sites/default/files/datasheets/40011.pdf. 11.
Influence of sintering temperature on the magnetic properties of LTCC ferrite tape for multilayer component applications Čedo Žlebič, Miodrag Milutinov, Ljiljana Živanov, Andrea Marić, Nelu Blaž & Goran Radosavljević Journal of Materials Science: Materials in Electronics ISSN 0957-4522 J Mater Sci: Mater Electron DOI 10.1007/s10854-017-8364-6

1 23

Your article is protected by copyright and all rights are held exclusively by Springer Science+Business Media, LLC, part of Springer Nature. This e-offprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to selfarchive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23

Author's personal copy Journal of Materials Science: Materials in Electronics https://doi.org/10.1007/s10854-017-8364-6

Influence of sintering temperature on the magnetic properties of LTCC ferrite tape for multilayer component applications Čedo Žlebič1   · Miodrag Milutinov1 · Ljiljana Živanov1 · Andrea Marić1 · Nelu Blaž1 · Goran Radosavljević2 Received: 10 July 2017 / Accepted: 1 December 2017 © Springer Science+Business Media, LLC, part of Springer Nature 2017

Abstract Influence of sintering temperature on the complex permeability, the core loss density and B-H hysteresis loop of ferrite lowtemperature co-fired ceramic (LTCC) multilayer toroidal cores is presented. Ferrite LTCC toroidal cores are fabricated from 22 layers of ESL 40011 tapes sintered at 885, 1000, 1100 and 1200 °C. Applying the sinusoidal excitation, the measurements are performed in the frequency range from 10 kHz to 5 MHz and for magnetic flux density from 2 to 100 mT. The measurements of sintered multilayer cores were done in order to investigate the difference in magnetic properties among those samples. This investigation enables to find optimal sintering temperature of multilayer ferrite cores which used in different multilayer magnetic component applications.

1 Introduction Characterization of magnetic properties of the low-temperature co-fired ceramic (LTCC) ferrite materials, such as the core loss density (CLD), permeability, and B-H hysteresis loop is very significant as these properties affect the operation of electronic devices. Ferrite LTCC materials have been applied in power converters [1–3], power inductors [4], tunable antennas [5], phase shifter [6], etc. Dependence of the magnetic characteristics of the frequency and alternating current (ac) magnetic flux density must be taken into account when designing electronic device since they affect the performance of the entire device. Ferrite LTCC materials (Ni–Zn ferrite) are commercially produced in tape format by tape-casting slurry on polymeric carrier substrates at varying thickness. The magnetic characteristics of Ni–Zn ferrite are mainly determined by their chemical compositions and microstructures, the latter of which are very sensitive to the sintering process and additives. Many researchers have been carried experiments on the magnetic behaviour of Ni–Zn ferrites. In [7] has been presented initial permeability and resistivity at 1 MHz for * Čedo Žlebič [email protected] 1



Faculty of Technical Sciences, University of Novi Sad, 21000 Novi Sad, Serbia



Institute of Sensor and Actuators Systems, Vienna University of Technology, 1040 Vienna, Austria

2

different sintering temperature in the range from 850 to 930 °C. The frequency dispersion of the permeability (from ­106 to 1­ 09 Hz) for three different sintering temperatures 900, 1000, 1100 °C has been given in [8]. In [9] has been shown as the sintering temperature increases (from 900 to 1050 °C), the permeability also increases (from 150 to 400) for ESL 40010. The property of the ferrite LTCC sample also depends on its preparation process. In our investigation, we used commercialized LTCC ferrite materials from Electro Science ESL 40011 [10]. Complex permeability dependence of ESL 40011 on ambient temperature (from 20 to 120 °C) is presented in [11]. Permeability and core loss density of ESL 40011 were measured on toroidal cores sintered at 885 °C [12] and the influence of superimposed dc bias on the magnetic property of ferrite laminates was evaluated. Affect of peak sintering temperature on ESL 40011 tape complex permeability and microstructure has been presented in [13]. The result shows significant increase in permeability value with increase of the peak sintering temperature (from 885 to 1200 °C). Similar research, but for the Mn–Zn ferrites, is presented in [14]. To our knowledge, no systematic study on the effect of sintering temperature on the magnetic behaviour of LTCC ferrite tape has been published. Because of that, we continue our research, published in [13], about the effect of sintering temperature on magnetic properties of toroidal cores fabricated from ESL 40011 ferrite tape such as: permeability, core loss density and B-H hysteresis loop.

13

Vol.:(0123456789)

Author's personal copy

Journal of Materials Science: Materials in Electronics

The aim of this paper is to examine the effect of different sintering temperatures of ESL 40011 material on the core loss density in the frequency range from 10 kHz to 5 MHz and for magnetic flux density from 2 mT up to 100 mT. This ranges of frequency and ac magnetic flux are needed for point-of-load (POL) converters (CLD measured at f = 1.5 MHz and f = 3 MHz for magnetic flux 1/100 mT) [2], buck converters (CLD measured at f = 1/5 MHz for magnetic flux 1/100 mT) [3], and power inductors (inductance measured in frequency range 100 Hz/1 MHz) [4]. Also, permeability as function of frequency at two levels of ac magnetic flux density and B-H hysteresis loop at frequency of 10 kHz for four different sintering temperatures is presented. Tested multilayer samples are sintered at temperatures of 885, 1000, 1100, and 1200 °C. These measurements enable to find the optimal sintering temperature for requested magnetic properties in determined frequency range and at needed level of ac magnetic flux density. The material fabrication process and material characterization of tested multilayer samples and measurement setup are described in Sect. 2 of this paper. Measurement results of the core loss density with the discussion are presented in Sect. 3. Results of the complex relative permeability measurements and B-H hysteresis loop results are presented in Sects. 4 and 5, respectively. Section 6 draws conclusions from the presented research.

2 LTCC core fabrication, material characterization and measurement setup The same samples that were used in [13] were also used in this paper for further measurements. Samples fabrication and material characterization are partly taken from [13] for a clearer presentation of the results.

Table 1  Geometric parameters, real and imaginary parts of permeability, grain sizes and chemical composition of samples obtained at different sintering temperature, summarized from [13]

13

2.1 Samples fabrication Four multilayer toroidal samples from ESL 40011 material were obtained after lamination of 22 ferrite tape layers from which by laser cutting were formed toroidal shapes. Uniaxial lamination process was applied for 4 min at a temperature of 75 °C with a pressure of 10.6 MPa. Samples were sintered at maximum temperatures Tmax of 885, 1000, 1100, and 1200 °C implemented sintering profiles as in [13]. Samples are labelled in the paper as S885, S1000, S1100, and S1200 where number represent sample sintering temperature. Applied temperature profiles differed only in the value of maximum temperature as well as in duration of its achieving. For all tested samples, heating rate from 0 to 450 °C was ~ 0.478 °C/min and heating rate from 450 °C to peak temperature was ~ 6.505 °C/min. The cooling rate was ~ 2.240 °C/min and duration of burnout phases and exposures to peak sintering temperatures was 2 and 3 h, respectively, for each peak sintering temperature.

2.2 Material characterization Geometric parameters of sintered toroidal samples are— inner diameter (ID)/outer diameter (OD) 4.18/6.42 and 1.2 mm in height. Real (µr′) and imaginary (µr″) part of permeability, the frequency at which µr′ and µr″ reach their maximum values (fmax_real and fmax_imag), the range of grain size and the ratio by weight percentage of each element of tested samples are listed in [13]. In [13] is shown that the real part of permeability increases significantly with the increase of maximum sintering temperature, from µr′ = 263.25 for the S885 to µr′ = 1022.09 for the S1200, which is result of larger grains sizes obtained at higher sintering temperatures. Grain sizes are smallest for the S885 and are in the range of 0.48–8.49 µm. The largest grains are obtained for S1200 (5.41–53.90 µm). All these values are shown in Table 1. Data in Table  1 are summarized from [13], where the same samples were analyzed as in this paper. For

ID/OD/h

S885 4.18/6.42/1.2 mm

S1000

S1100

S1200

µr′ (fmax_real) µr″ (fmax_imag) Grain size Element  Fe (%)  O (%)  Zn (%)  Ni (%)  Cu (%)

263.19 (2.77 MHz) 103.14 (11.74 MHz) (0.48–8.49) µm

383.41 (2.09 MHz) 148.27 (9.22 MHz) (1.25–10.50) µm

593.57 (1.58 MHz) 230.31 (5.98 MHz) (2.83–28.90) µm

1022.09 (0.82 MHz) 423.80 (3.04 MHz) (5.41–53.90) µm

48.36 23.75 17.73 7.11 3.05

47.58 26.16 16.74 6.84 2.68

48.69 24.08 16.90 7.37 2.96

47.91 25.10 16.50 6.77 3.72

Author's personal copy Journal of Materials Science: Materials in Electronics

determining the influence of sintering temperature on grain size of tested samples, Scanning Electron Microscope (SEM) is used [13]. Samples microstructure are presented in Fig. 1 and measured ranges of grain sizes are shown in Table 1. Tested samples were coated with 15 nm gold layer and subjected to Energy Dispersive Spectrometry (EDS). Influence of sintering temperature on variations in percentage ratio of individual chemical elements (Fe, O, Zn, Ni and Cu) is determined using EDS. EDS spectrum of ferrite samples are presented in Fig. 2 and percentage ratio of individual chemical elements are shown in Table 1. As it can be seen in Table 1 from the chemical composition of tested samples for different sintering temperatures, there are Ni–Zn ferrites with a small concentration difference of the similar elemental composition. Since the results of the EDS analysis depends on the place on the sample where it is performed, it can introduce a certain dispersion in the values of the elements concentration.

2.3 Measurement setup for magnetic properties determination Modified wattmeter method offers possibility of measurement all these values: CLD, permeability and hysteresis with same measurement setup. The core loss is measured with modified wattmeter method using digital oscilloscope Keysight Technology DSO90604 with low-noise and 50 Ω input impedance [15], used as a part of the measuring setup shown in Fig. 3a. The classic wattmeter method uses oscilloscope with high-input impedance. The modified wattmeter method is adapted to the oscilloscope with low-input impedance. Equivalent model of the modified two-winding wattmeter method is shown in Fig. 3b. Magnetization of the core under test is modelled with reactance Xm, while the core loss is modelled with resistance Rm connected in series. The oscilloscope with low-impedance in the classic two-winding method will cause a significant terminal current i2 in the secondary winding. In order to attenuate this current, the secondary winding is terminated with the resistances R3 and R2 connected in series, as shown in Fig. 3. As it is described in [16], the current in the excitation winding is measured using sensing resistor R1 while the induced

Fig. 1  Microstructures of tested samples. a Sample S885, b Sample S1000, c Samples S1100*, and d Sample S1200. [13]

13

Author's personal copy

Journal of Materials Science: Materials in Electronics

Fig. 2  EDS of tested samples for a Sample S885, b Sample S1000, c Sample S1100, and d Sample S1200

Fig. 3  a The measurement setup for the modified two-winding wattmeter method, based on the digital oscilloscope DSO90604A with R ­ osc = 50 Ω input impedance, b An equivalent model of the modified twowinding wattmeter method

voltage over the sensing winding is measured using a voltage divider made of resistors R2 and R3. The measuring voltages v1 and v2 over sensing resistors R1 and R2, are captured using passive probes with 50 Ω impedance. A disadvantage of this modified method is the presence of the current i2, which causes reduction of the magnetizing current im. The reduction of magnetizing current becomes more significant when the frequency is increased. The resistances R1 and R2 are 50 Ω to match the input impedance. The value of resistance R3 depends on the measuring frequency, dimensions and permeability of the core. It needs to be sufficiently high to reduce the current i2 as much as possible. The minimum of the current i2 is defined over the sensitivity of the oscilloscope and resistances R2 and Rosc = 50 Ω. As function generator of sinusoidal signals, an HP3314A is used [17]. The average core loss Pcore over the period T is determined by using (1), where the magnetizing current im is obtained from the measured voltages v1 and v2 by using (2), and the induced voltage um is determined by (3), where Re1 is the resistance equivalent to the sensing resistance R1 connected in parallel with the oscilloscope input resistance Rosc, and Re2

13

is the resistance equivalent to the sensing resistance R2 connected in parallel with the oscilloscope input resistance Rosc, as shown in Fig. 3b. T

Pcore

1 = u (t) ⋅ im (t) dt, T∫ m

(1)

0

im = i1 − i2 =

um =

v1 v − 2 , Re1 Re2

) v2 ( ⋅ R3 + Re2 . Re2

(2) (3)

The source of measurement error of proposed method is voltage measurement accuracy of the used digital oscilloscope. The voltage measurement accuracy of the DSO90604A is sum of gain accuracy (± 2% of full scale) and resolution (0.024% with averaging) [15]. According to the specification of the DSO90604A, the relative error of voltage measurement is ± 3%. The phase discrepancy is another source

Author's personal copy Journal of Materials Science: Materials in Electronics

of error that need to be added. High sampling rate of the DSO90604A introduce additional error of ± 2%. Therefore, the relative error of the core loss is ± 8%. The core loss density was calculated as

Pv =

Pcore P = core , Ae ⋅ le Ve

(4)

where Ae is the effective area of the cross section of the core, le is the effective length of the core, and Ve is core effective volume [18]. Real µr′ and imaginary µr″ part of complex relative permeability was calculated as

𝜇r� =

𝜇r��

=

Z− = m

{

} Im Zm 𝜔 ⋅ L0 ⋅ 𝜇0

,

(5)

,

(6)

{ } Re Zm 𝜔 ⋅ L0 ⋅ 𝜇0

max(um ) j