ARTICLE IN PRESS

Journal of Magnetism and Magnetic Materials 321 (2009) 3808–3812

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials

0304-88

doi:10.1

� Corr

E-m

journal homepage: www.elsevier.com/locate/jmmm

High permeability and low power loss of Ti and Zn substitution lithiumferrite in high frequency range

Vivek Verma a, S.P. Gairola a, Vibhav Pandey a, J.S. Tawale a, Hua Su b, R.K. Kotanala a,�

a National Physical Laboratory, Dr. K.S. Krishnan Road, New Delhi 110012, Indiab State Key Laboratory of Electronic Thin Films and Integrated Devices, University of Electronic Science and Technology of China, Chengdu 610054, China

a r t i c l e i n f o

Article history:

Received 10 April 2009Available online 6 August 2009

PACS:

75.50.�y

61.10.Ht

75.50.Gg

07.55.�w

76.90.+d

Keywords:

Magnetic

Structural

Ferrite

Complex permeability

Relaxation

53/$ - see front matter & 2009 Published by

016/j.jmmm.2009.07.044

esponding author. Tel.: +911145608599.

ail address: rkkotnala@nplindia.org (R.K. Kota

a b s t r a c t

The effect of Zn and Ti on the magnetic, power loss and structural properties of Li0.5ZnxTixMn0.05

Fe2.45�2xO4 ferrites (x ¼ 0.0 to 0.30 in step of 0.05)+0.5 wt% Bi2O3, prepared by standard ceramic

technique, has been investigated. Complex permeability (m* ¼ m0�jm00) has been analyzed at room

temperature in frequency range from 1 to 103 MHz. It was found an enhancement in permeability with

Ti and Zn concentration in Li0.5ZnxTixMn0.05Fe2.45�2xO4 and exhibits the maximum value 106 for

x ¼ 0.20 sample. Complex permeability of these ferrites exhibits stable frequency response up to 7 MHz

beyond which the real part decreases sharply and imaginary part increases to have a peak at the

relaxation frequency. Power loss measurements have been carried out in induction condition

(B ¼ 10 mT) in frequency range of 50 kHz to 3 MHz. Power loss has been found to be quite low with

the substitution of Ti and Zn in lithium ferrite.

& 2009 Published by Elsevier B.V.

1. Introduction

Pure and substituted lithium ferrites are low cost materials andhave important magnetic and electrical properties for technolo-gical applications. Ferrites assume special significance in the fieldof electronics and telecommunication industry because of theirnovel electrical properties which make them useful in radio-frequency circuits, high quality filters, rod antennas, transformercores, read/write heads for high digital tapes and other devices[1–3]. Mn–Zn ferrites and Ni–Zn ferrites are well known materialsfor power electronic equipments. Development of a high quality,low cost and low loss for high frequency ferrite materials forpower applications is an ever challenging aspect for investigation.Substituted lithium ferrites may be useful material for suchapplications because of their high saturation magnetization andlow power loss (Pcv) [4,5]. Magnetic and electrical properties offerrites were found to be sensitive to their composition andprocessing techniques. The main types of losses encountered inferrites are eddy current loss, hysteresis loss and residual loss.Consequently the requirements of a power ferrite are highresistivity to keep the eddy current losses low, high permeability

Elsevier B.V.

nala).

to reduce hysteresis losses and a high resonance frequency toreduce the residual losses, which consist mainly of resonance–relaxation losses [6,7]. In this work we have studied the effect ofZn and Ti on the microstructure, permeability and power loss oflithium ferrites.

2. Experimental

Polycrystalline samples of substituted lithium ferrite with thestoichiometric formula Li0.5ZnxTixMn0.05Fe2.45�2xO4 (x ¼ 0.0 to0.30 in step of 0.05) were prepared by conventional ceramictechnique. A small amount (0.5 wt%) of Bi2O3 was also added toreduce the sintering temperature, which prevents the volatiliza-tion of Lithia and enhances the densification [8]. The startingmaterials were AR grade Li2CO3, Fe2O3, ZnO, MnO2, TiO2 andBi2O3. An appropriate amount of these materials were taken andground/mixed thoroughly. The resultant mixtures were dried andcalcined at 750 1C for 10 h. The powder was ground and pressed inthe form of pellets and toroids using a small amount of PVA as abinder with an applied pressure of 10 ton. The final sintering wascarried out at 1050 1C for 5 h. Heating and cooling rate wascontrolled at 5 1C/min. The structural characterization of sampleswas carried out by the X-ray diffraction (XRD Rigaku Miniflex II,step size ¼ 0.02) technique using CuKa radiation (l ¼ 1.5406 A).

ARTICLE IN PRESS

V. Verma et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 3808–3812 3809

Power loss and permeability response in frequency range of1 MHz–1 GHz were measured by B–H analyzer (IWATSU, SY 8232)and Agilent 4284/Agilent 4285 respectively. Toroid shape samples(i.d. ¼ 5.5 mm, o.d. ¼ 11 mm, t ¼ 3 mm) were used to calculate thepower loss and permeability. Magnetic measurements wereperformed at room temperature using vibrating sample magnet-ometer (Lake Shore 7304) for all samples. Resistivity of sampleswas calculated by two probe method using KEITHLEY 4200 SCS.

3. Results and discussion

The X-ray diffraction patterns of the samples Li0.5ZnxTixMn0.05-

Fe2.45�2xO4 with x ¼ 0.0, 0.05, 0.10, 0.15, 0.20, 0.25 and 0.30) areshown in Fig. 1, which represents the typical spinel structure. The(h k l) values corresponding to the diffraction peaks are marked infigure. The lattice parameters ‘a’ for the sintered samplescalculated from the (311) diffraction peak, are shown in Table 1.Lattice constant increases with the Ti and Zn concentration up tox ¼ 0.20, because Zn2+ (0.82 A) is replaced by the Fe3+ (0.67 A)from tetrahedral side and Ti4+ (0.60 A) to Fe3+ (0.67 A) fromoctahedral site [9,10] but it begins to decrease with furthersubstitution which may be due to Ti4+ start to replace Fe3+ fromtetrahedral site also [11].

Complex permeability (m* ¼ m0�jm00) response curves for Znand Ti substituted Li0.5ZnxTixMn0.05Fe2.45�2xO4 ferrites for fre-quency range 106–109 Hz have been shown in Fig. 2(a) and (b). It isobserved that the value of real part of permeability (�28) ofLi0.5Mn0.05Fe2.45O4 sample increases with the concentration of Tiand Zn in Li0.5ZnxTixMn0.05Fe2.45�2xO4 and attains the maximumvalue of �106 for x ¼ 0.20 sample. Complex permeability of these

Fig. 1. X-ray diffraction patterns of Li0.5ZnxTixMn0.05Fe2.45�2xO4.

Table 1Microstructure and magnetic properties of Li0.5ZnxTixMn0.05Fe2.45�2xO4 samples.

Composition Li0.5ZnxTixMn0.05Fe2.45�2xO Lattice constant (A) Permeability (m)

x ¼ 0.00 8.321 27.9

x ¼ 0.05 8.342 46.5

x ¼ 0.10 8.347 64.4

x ¼ 0.15 8.358 59.8

x ¼ 0.20 8.357 105.5

x ¼ 0.25 8.348 51.4

x ¼ 0.30 8.340 45.5

ferrites exhibits stable frequency response up to 7 MHz beyondwhich the real part decreases sharply and imaginary partincreases to have a peak at the relaxation frequency. The changein permeability for different compositions is defined on thedependence of permeability, saturation magnetization andaverage grain size is expressed by following relation (1) [12]:

mipm0M2s Dm=½K1 þ ð3=2Þlss�b1=3d ð1Þ

at frequency 1 MHz Saturation magnetization (emu/g) Grain size (mm)

58.5 3.9

56.2 6.1

55.9 6.4

51.9 7.0

52.6 7.8

54.5 6.2

40.9 5.1

Fig. 2. Complex permeability curves of (a) real and (b) imaginary part of

Li0.5ZnxTixMn0.05Fe2.45�2xO4.

ARTICLE IN PRESS

Fig. 3. SEM micrographs of Li0.5ZnxTixMn0.05Fe2.45�2xO4 samples.

V. Verma et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 3808–38123810

where mi is the initial permeability, Ms and Dm are the satura-tion magnetization and average grain size respectively, K1 isthe magneto-crystalline anisotropy, ls is the saturation magneto-striction constant, s is the inner stress, b is the volume concen-tration of impurity and d is the thickness of domain wall.

The above relation shows that initial permeability stronglydepends on saturation magnetization (Ms), size of grain (D) andmagneto-crystalline anisotropy (K1). In our samples it is observedthat the saturation magnetization decreases slowly with theTi and Zn concentration in Li0.5ZnxTixMn0.05Fe2.45�2xO4, butx ¼ 0.30 sample shows a sudden decrease. This behaviour canbe explained on the basis of Neel’s two sub-lattice model [13].Grain growth is very pronounced in Ti and Zn substituted sampleswhich can be observed from SEM micrographs Fig. 3. The grain

size 3.9mm (sample x ¼ 0.0) increases to 7.8mm (sample x ¼ 0.20)with Ti, Zn concentration. Grain growth is uniform for samplex ¼ 0.20 sample and grains start to melt with further substitutedsamples x ¼ 0.25 and 0.30, which may be due to the concentrationof Zn that promotes grain growth and lowers the sinteringtemperature [14]. The value of K1and ls decreases and grain sizeincreases as content of Zn, Ti increases in ferrites [15].Permeability decreases with further substitution of Ti and Zn inlithium ferrites for the samples x ¼ 0.25 and 0.30, which may bedue to the decrease in magnetization and microstructure also. Thefrequency response of permeability shows a typical relaxationcharacter. This behaviour may be defined by the phenomenon ofreversible displacement of domain walls and also due to therotation of magnetization dipole inside the domain. The

ARTICLE IN PRESS

Fig. 5. Variation of Pcv and frequency at 10 mT of Li0.5ZnxTixMn0.05Fe2.45�2xO4

samples.

V. Verma et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 3808–3812 3811

permeability response at higher frequency is mainly determinedby domain wall displacement resonance and spin rotationresonance. It is also known that spin rotation resonancefrequency is higher than domain wall displacement resonancefrequency. In 1 MHz–1 GHz frequency range the probability ofdomain wall displacement resonance is less while spin rotationresonance probability is high.

The frequency dispersion of permeability of lithium ferritesamples is governed by Snoek’s limit om0 ¼ constant [16], wherem0 stands for the real part of permeability and o the resonancefrequency for domain wall motion above which m0 decreases.

It is well known that power loss of ferrite materials can bedivided into hysteresis loss (Ph), eddy current loss (Pe) andanomalous or residual loss (Pr). Hysteresis loss is dominant at dcor low frequencies. To reduce the hysteresis loss we have tominimize the anisotropy and magnetostriction as to maximize thepermeability of material. Uniform grain growth and low porosityare also desired properties to reduce hysteresis loss. Substitutionof Zn and Ti reduces the hysteresis loss by improving the graingrowth and microstructure of ferrites. Hysteresis loss can beevaluated from the area inside of the dc hysteresis looprepresented by the relation [17]:

Ph ¼ f

ZHdB

where f is the frequency, H the magnetic field strength and B themagnetic induction.

The eddy current losses become an important factor as theferrite is used at higher frequencies. Eddy current losses becometoo high at higher frequencies reducing the performance con-siderably. The eddy current losses can be reduced by increasingthe resistivity of material. A small amount of Mn avoids theformation of Fe2+ ions which improves the resistivity of ferrites[18]. Eddy current loss also can be reduced by increasing theresistivity of the polycrystalline ferrite by increasing grainboundary resistivity. It is observed that dc resistivity varies from106 to 107 orders with small concentration of Zn and Ti inLi0.5ZnxTixMn0.05Fe2.45�2xO4. Eddy current loss can be determinedby the relation:

Pe ¼ CB2f 2d2=r

Fig. 4. Variation of Pcv and flux density at 50 kHz of Li0.5ZnxTixMn0.05Fe2.45�2xO4

samples.

where C is the proportionality constant, B the flux density, f thefrequency, f the resistivity and d the thickness of the material.Therefore the total loss of the ferrites can be expressed as:

Pcv ¼ Ph þ Pe þ Pr

The variation of power loss (Pcv) with exiting flux density (Bm)at 50 kHz of Li0.5ZnxTixMn0.05Fe2.45�2xO4 samples have beenshown in Fig. 4. The variation of power loss increases with fluxdensity for all samples. However we observed decrease in powerloss with the substitution of Zn and Ti in lithium ferrite andexhibits a minimum loss by x ¼ 0.25 sample of Li0.5ZnxTixMn0.05

Fe2.45�2xO4. Frequency dependent power loss (Pcv) for the samplesat exciting condition of Bm ¼ 10 mT is shown in Fig. 5. We candepict from Fig. 5 that the power loss reduces with theconcentration of Zn and Ti up to x ¼ 0.25 in theLi0.5ZnxTixMn0.05Fe2.45�2xO4. The variation of power loss is smallup to frequency 1 MHz but it changes abruptly beyond this in theexiting flux density of 10 mT for all samples. This is evident fromFig. 5 that the hysteresis loss is predominant loss mechanism atlower frequencies approximately below 500 kHz and above thesefrequencies the eddy current loss increases gradually withincreasing frequency and becomes predominant in power loss.

4. Conclusions

The substituted lithium ferrites Li0.5ZnxTixMn0.05Fe2.45�2xO4

with the different amounts of Ti4+ and Zn2+ ions of x ¼ 0.0, 0.05,0.10, 0.15, 0.20, 0.25 and 0.30+0.50 wt% Bi2O3 were prepared byconventional ceramic method. There is significant enhancementin permeability with Ti and Zn doping lithium ferrites and shows amaximum value of 106 for x ¼ 0.20 sample, which is a veryimportant feature for power applications of ferrites. Grain growthmechanism is also one of the important features in Zn- and Ti-doped lithium ferrites. The grain size 3.9mm for x ¼ 0.0 sampleincreases to 7.8mm for x ¼ 0.20 sample with Ti, Zn concentrationin Li0.5ZnxTixMn0.05Fe2.45�2xO4. We also obtained reduction inpower losses in substituted lithium ferrites. These materialsexhibited low power losses up to 1 MHz, making them moresuitable for power applications.

ARTICLE IN PRESS

V. Verma et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 3808–38123812

Acknowledgement

The authors are grateful to Dr. Vikram Kumar, Director‘‘National Physical Laboratory’’ New Delhi for providing constantencouragement and motivation to carry out this work.

References

[1] B. Gillot, Eur. Phys. J. Appl. Phys. 91 (2002) 10.[2] C. Snelling, Soft Ferrites, ILITFE Books Ltd., London, 1969.[3] M. Sugimoto, J. Am. Ceram. Soc. 82 (1999) 269.[4] Jing Jiang, Yan-Min Yang, Liang-Chao Li, Phys. B 399 (2007) 105–108.[5] Yen-Pei Fua, Chin-Shang Hsu, Solid State Commun. 134 (2005) 201–206.[6] Anjali Verma, M.I. Alam, Ratnamala Chatterjee, T.C. Goel, R.G. Mendiratta, J.

Magn. Magn. Mater. 300 (2006) 500–505.[7] H. Shokrollahi, J. Magn. Magn. Mater. 320 (2008) 463–474.

[8] Zhong Yu, Ke Sun, Lezhong Li, Yunfei Liu, Zhongwen Lan, Huaiwu Zhang, J.Magn. Magn. Mater. 320 (2008) 919–923.

[9] Mamata Maisnam, Sumitra Phanjoubam, H.N.K. Sarma, Chandra Prakash, L.Radhapiyari Devi, O.P. Thakur, Mater. Lett. 58 (2004) 2412–2414.

[10] Zhenxing Yue, Ji Zhou, Xiaohui Wang, Zhilun Gui, Longtu Li, J. Eur. Ceram. Soc.23 (2003) 189–193.

[11] Sumitra Phanjoubam, Deepika Kothari, J.S. Baijal, Pran Kishan, Solid StateCommun. 68 (6) (1988) 549–554.

[12] J. Smit, H.P.J. Wijn, Ferrites, Philips Technical Library, Eindhoven, Netherlands,1959, p. 244.

[13] N.S. Satya Murthy, M.G. Natera, Phys. Rev. 181 (1969).[14] J. Simit, H.P.J. Wijn, Ferrites, Philips Technical Library, Eindhoven, Netherlands,

1959, p. 221.[15] J.K. Galt, W.A. Yager, J.P. Remeika, F.R. Merritt, Phys. Rev. 81 (1951) 470.[16] Gary G. Bush, J. Appl. Phys. 63 (8) (1998) 3765.[17] C.R. Hendricks, V.W.R. Amarakoon, D. Sullivan, Ceram. Bull. 70 (1991)

817–823.[18] Xiwei Qi, Ji Zhou, Zhenxing, Zhilun Gui, Longtu Li, Mater. Sci. Eng. B 99 (2003)

278–281.