J. Am. Ceram. Soc., 84  2905–908 (2001) journal Effect of the Magnetostrictive Layer on Magnetoelectric Properties in Lead Zirconate Titanate/Terfenol-D Laminate Composites Jungho Ryu,*,† Shashank Priya, Alfredo Vazquez Carazo, and Kenji Uchino** ´ International Center for Actuators and Transducers, Materials Research Laboratory, The Pennsylvania State University, University Park, Pennsylvania 16802 Hyoun-Ee Kim* School of Materials Science and Engineering, Seoul National University, Seoul, 151-742 Korea Magnetoelectric laminate composites of piezoelectric/magne- were used as a piezoelectric material and magnetostrictive mate- tostrictive materials were prepared by stacking and bonding rial, respectively. The composites were manufactured by sand- together a PZT disk and two layers of Terfenol-D disks with wiching and bonding a PZT disk between two layers of Terfenol-D different directions of magnetostriction. These composites disks. These composites demonstrated a superior magnetoelectric were studied to investigate (i) dependence on the magnetos- voltage coefficient (dE/dH 4.68 V/cm Oe) in comparison with triction direction of the Terfenol-D disk and (ii) dependence on previously reported magnetoelectric composites.2–5 In previous the direction of the applied ac magnetic field. Three different work, we pointed out that the most important factors for obtaining types of assemblies were prepared by using two types of disks: a high magnetoelectric property were the high piezoelectric volt- one with magnetostriction along the radial direction, the other age coefficient (gij) and thickness ratio of the PZT layer over the with magnetostriction along the thickness direction. The max- Terfenol-D layer. imum magnetoelectric voltage coefficient (dE/dH) of 5.90 In addition to their high dE/dH, the laminated magnetoelectric V/cm Oe was obtained for a design where the composite was composites have a very simple structure and a relatively simple made by two Terfenol-D layers with a radial magnetostriction fabrication method, i.e., bonding each disk. The laminated mag- direction. netoelectric composites can be easily applied to practical applica- tions, such as magnetic field sensing devices, leak detectors for microwave ovens, and current measurement of high-power electric I. Introduction transmission systems. In this study, we investigate the directional dependence of last three decades, many particulate1–3 and in-situ- I N THE 4,5 grown magnetoelectric composite materials have been devel- oped by using piezoelectric materials and magnetostrictive ferrite magnetostriction of the Terfenol-D disk and of ac magnetic field on the magnetoelectric response of the PZT/Terfenol-D laminate composites. This helps in optimizing the performance of laminate materials to overcome the problems of single-phase magnetoelec- magnetoelectric composites. tric materials, i.e., low magnetoelectric response and requirement of low temperature.6,7 These composite materials used the product property of piezoelectric and magnetostrictive effects of each phase.7,8 This can be explained as follows: the magnetostrictive II. Experimental Procedure phase is strained when a magnetic field is applied to the composite, PZT (APC– 840, American Piezoceramics, Inc., PA) is used as and this strain induces stress on the piezoelectric phase, generating a piezoelectric material for our composites. This PZT material has an electric field in the piezoelectric phase. Thus, the product a high piezoelectric voltage coefficient (g33 26.5 mV m/N). property in such composites is the magnetoelectric effect, i.e., an Regarding the magnetostrictive material, TbDyFe2 (Terfenol-D) applied magnetic field induces an electric field. These particulate giant magnetostrictive alloy (Etrema Products, Inc., Ames, IA) is or in situ magnetoelectric composites made of piezoelectric mate- one of the most ideal magnetostrictive materials for these com- rials and magnetostrictive ferrite materials show higher magneto- posites because of its high magnetostriction and coupling con- electric properties compared with single-phase magnetoelectric stants.10 To investigate the dependence of the magnetostriction materials, such as Cr2O3.1–5 However, these composites still have direction, we used two types of Terfenol-D disks: one with some problems in reproducibility and reliability, such as control of principal magnetostriction along the radial direction, the other with the connectivity and chemical reaction between two phases, low magnetostriction along the thickness direction. electric resistivity, and an insufficient magnetoelectric voltage PZT/Terfenol-D magnetoelectric composite samples were pre- coefficient for practical applications.3 pared by stacking and bonding together the PZT and Terfenol-D Recently, laminated magnetoelectric composites made by using disks with silver epoxy. Detailed fabrication procedures have been piezoelectric and magnetostrictive materials have been studied by described in a previous paper.9 The dE/dH was determined by our group.9 Lead zirconate titanate (PZT) and Terfenol-D disk measuring the electric charge generated across the sample when an ac magnetic field and dc bias were applied. The magnetoelectric property was measured in terms of the variation of the coefficient dE/dH as a function of dc magnetic bias field. An electromagnet S.-I. Hirano—contributing editor (Model GMW 5403 Magnet, Power and Buckley, Inc., New Zealand) was used for generating the bias field up to 0.45 T (4.5 kOe). The ac magnetic field of 1 kHz was applied using Helmholtz coils superimposed on the dc bias field, both parallel to the sample Manuscript No. 188019. Received January 16, 2001; approved July 9, 2001. thickness axis. To investigate the dependence of the applied ac *Member, American Ceramic Society. **Fellow, American Ceramic Society. magnetic field direction, samarium– cobalt permanent magnets † Visiting Researcher from Seoul National University, Seoul, Korea. were used instead of the electric magnet for applying the dc 2905 2906 Journal of the American Ceramic Society—Ryu et al. Vol. 84, No. 12 magnetic bias. In this case, the applied dc magnetic bias to the composite was 0.075 T (750 Oe). A signal generator (Model DS340, Stanford Research Systems, Sunnyvale, CA) was used to drive the Helmotz coils and generate the ac magnetic field. The electric charge generated from the piezoelectric layer was measured through a charge amplifier (Model 5010B Dual Mode Amplifier, Kistler Instrument Co., Amherst, NY). This amplifier was designed for converting a charge signal from the piezoelectric transducer into a proportional output voltage. The output voltage from the amplifier was mea- sured with an oscilloscope (Model 54645A, Hewlett–Packard Co., Palo Alto, CA). The measured voltage represented the electric charge from the piezoelectric PZT layer under a short-circuit condition. The output voltage was obtained from the charge and the capacitance of the PZT layer of the composite using V Q/C (1 kHz). The output voltage divided by the thickness of the PZT layer and the ac magnetic field gave the dE/dH of the samples. III. Results and Discussion Fig. 2. The parameter dE/dH as function of applied dc magnetic bias with different assembly. (1) Magnetostriction Direction Dependence To explore the direction dependence of magnetostriction, three types of laminate composites were prepared using two types of Figure 2 illustrates the variation of the dE/dH as a function of Terfenol-D disks. These are as follows: the dc magnetic bias for three different composites. The frequency (1) Composite with two Terfenol-D disks, which have mag- of the ac magnetic field was 1 kHz for the measurement. The netostriction along the thickness direction (denoted as Comp. T-T). dE/dH values of all composites increased with increasing dc bias (2) Composite with one Terfenol-D disk with thickness mag- and saturated at 4 kOe. The Comp. R-R showed the most netostriction direction and the other Terfenol-D disk with radial superior magnetoelectric properties, and the maximum dE/dH was magnetostriction direction (denoted as Comp. T-R). 5.90 V/cm Oe under a magnetic bias 4.2 kOe dc. In the (3) Composite with two disks, which have magnetostriction Terfenol-D disks, the magnitude of strain in the principal magne- along the radial direction (denoted as Comp. R-R). tostriction direction was higher than in the other directions.10 Figure 1 shows the schematic illustrations of each composite According to the results reported in our previous work,9 the structure (Fig. 1(a)) and a picture of the sample (Fig. 1(b)). The piezoelectric voltage coefficient (gij), strain in the Terfenol-D dielectric polarization direction of the PZT disk and the applied along the radial direction, and the thickness ratio between the PZT magnetic field direction were in the thickness direction for all the and Terfenol-D disks are three important factors for producing a composites. All composites were fabricated with the same thick- superior magnetoelectric response in this type of magnetoelectric ness (total 2.5 mm; two disks of 1 mm thick Terfenol-D and one laminated composite. The highest dE/dH in Comp. R-R originates disk of 0.5 mm thick PZT) and diameter (12.7 mm), which were from high magnetostriction along the radial direction. optimized in a previous study.9 (2) Magnetic Field Direction Dependence In applications, such as magnetic field sensing devices, the dependence of the magnetoelectric response on the magnetic field direction is an important factor. To examine the magnetic field direction dependence, we measured the dE/dH by changing the applied magnetic field direction. The configuration applied for determining the direction dependence of the magnetic field is schematically represented in Fig. 3. The dc magnetic bias was Fig. 1. (a) Schematic illustrations for three different PZT/Terfenol-D Fig. 3. Magnetic field directions to monitor dE/dH dependence with composites and (b) their photograph. (Ruler is in inches.) applied ac magnetic field direction. December 2001 Effect of Magnetostrictive Layer on Magnetoelectricity in Pb(Zr,Ti)O3 / Terfenol-D Composites 2907 applied in the thickness direction during all the experiments and fixed at 750 Oe. By rotating the sample in the Helmholtz coils, the direction of the applied ac magnetic field could be changed. The angle of the applied ac magnetic field (1 kHz) indicated the difference in angle between the ac magnetic field and dc magnetic bias (thickness direction of the sample). The dependence of the dE/dH of the composites on the applied ac magnetic field direction is shown in Fig. 4. The dependence of the ac magnetic field direction exhibited similar behavior for all the composites, with a moderate increase to 25o– 45o. Above that, the dE/dH decreased with increasing ac magnetic field direction. As discussed in the previous section, magnetoelectricity is the combined effect of magnetostriction and piezoelectricity. The magnitude of the dE/dH is dependent on the strain produced by the magnetostrictive material. Thus, the estimation of strain gives an idea regarding the magnitude of the dE/dH. Because strain is a second-rank tensor, we need to consider six components: three extensional strains and three shear strains. Let the coordinate axes be defined as shown in Fig. 5(a). The strain components can be transformed as follows on rotation of the coordinate axes by an angle, : x ij a ik a jl x kl (1) i, j where alm is the direction cosine between the l and m directions, xij the strain component after rotation, and xij the initial strain component. In the case of rotation along the 2-axis, these strains are given by x1 cos2 x1 2 sin cos x5 sin2 x3 (2) x2 x2 x3 sin2 x1 2 sin cos x5 cos2 x3 For these laminate composites, the effective strain to generate electric charge from the PZT layer is not along the 3-axis but is areal strain along the 1- and 2-axes. Areal strain ( A/A) is given by A x1 x2 (3) A where A is the area of the material. In the above equation, the Fig. 5. Definition of coordination axes for (a) theoretical calculation and strain components are written in the reduced notation. After (b) areal strain as function of applied magnetic field direction from Eqs. performing the rotation operation with respect to the 2-axis, areal (4), (5), and (6). strain is given by A x1 x2 Hence, if the strain components are known, areal strain can be A determined as a function of orientation. For simplicity, choosing a cos2 x1 2 sin cos x5 sin2 x3 x2 (4) polycrystalline ferromagnetic material with three axes as the direction of magnetic polarization (magnetic bias) and assuming that there is no externally applied stress, strain is given by11 x1 0 0 d 31 x2 0 0 d 31 H1 x3 0 0 d 33 H2 (5) x4 0 d 15 0 H3 x5 d 15 0 0 x6 0 0 0 where dij are the magnetostrictive constants, sometimes referred to as piezomagnetic constants, and Hi the components of the applied field in the ith direction. The components along the 1- and 3-axes of the magnetic field can be written in terms of the applied magnetic field (Happ): 3 3 H app H1 H3 H1 H app sin (6) Fig. 4. The parameter dE/dH as function of applied ac magnetic field direction. H3 H app cos 2908 Journal of the American Ceramic Society—Ryu et al. Vol. 84, No. 12 The magnitude of the constants d31 5.3 10 8, d33 showed a moderate increase up to 25°– 45° tilting. Above this 11 10 8, and d15 28 10 8 m/A are reported for angle, the dE/dH decreased significantly. This behavior seems to Terfenol-D in Ref. 12. Substituting the values of the strain originate from the contribution of relatively large shear mode components from Eq. (5) into Eq. (4), and taking into account strains. Eq. (6), the relation for areal strain is obtained in terms of the magnetostrictive coefficients, orientation angle, and applied References magnetic field. This expression is plotted as a function of the 1 orientation angle in Fig. 5(b). The maximum areal strain occurs J. van den Boomgaard and R. A. J. Born, “A Sintered Magnetoelectric Composite at the orientation angle, 51°. The magnetoelectric coeffi- Material BaTiO3–Ni(Co,Mn)Fe3O4,” J. Mater. Sci., 13, 1538–48 (1978). 2 J. van den Boomgaard, A. M. J. G. Van Run, and J. Van Suchtelen, “Magneto- cient should show a maximum around this angle. This behavior electricity in Piezoelectric–Magnetostrictive Composites,” Ferroelectrics, 10, 295–98 originates from the contribution of relatively large shear mode (1976). 3 strains, i.e., d15 d33 or d31 . J. Ryu, A. Vazquez Carazo, K. Uchino, and H.-E. Kim, “Piezoelectric and This situation is analogous to the piezoelectric strains ob- Magnetoelectric Properties of Lead Zirconate Titanate/Ni–Ferrite Particulate Com- posites,” J. Electroceram., in press. served in several single-crystal perovskites. Kuwata et al.13,14 4 J. Van Suchtelen, “Product Properties: A New Application of Composite Mate- reported the maximum apparent piezoelectric coefficients in rials,” Philips Res. Rep., 27, 28 –37 (1972). 5 Pb(Zn1/3Nb2/3)O3–PbTiO3, where 57° is the optimized angle J. van den Boomgaard, D. R. Terrell, R. A. J. Born, and H. F. J. I. Giller, “An In between spontaneous polarization and the electric field. Re- Situ Grown Eutectic Magnetoelectric Composite Material: Part I, Composition and Unidirectional Solidification,” J. Mater. Sci., 9, 1705–709 (1974). garding PbZrO3–PbTiO3, Du et al.15 reported the optimum 6 T. H. O’Dell, “A New Class of Materials,” Electron. Power, 11, 266 (1965). angle is 50°– 60°, depending on composition. In both cases, 7 K. Uchino, “Piezoelectric Composites”; pp. 523–32 in Comprehensive Composite large shear strains determine the optimum cant angle between Materials, Vol. 5, Ch. 5.24. Edited by L. Carlsson, R. L. Crane, and K. Uchino. the electric field and crystal orientation. Elsevier, New York, 2000. 8 R. E. Newnham, “Composite Electroceramics,” Ferroelectrics, 68 [1/4] 1–32 (1986). 9 J. Ryu, A. Vazquez Carazo, K. Uchino, and H.-E. Kim, “Magnetoelectric IV. Conclusions Properties in Piezoelectric and Magnetostrictive Laminar Composites,” Jpn. J. Appl. Phys., 40  4849 –51 (2001). 10 A. B. Flatau, M. J. Dapino, and F. T. Calkins, “Magnetostrictive Composites”; Magnetoelectric laminate composites of a piezoelectric PZT pp. 563–74 in Comprehensive Composite Materials, Vol. 5, Ch. 5.26. Edited by L. disk and two magnetostrictive Terfenol-D disks have been fabri- Carlsson, R. L. Crane, and K. Uchino. Elsevier, New York, 2000. cated by stacking and bonding to obtain a superior magnetoelectric 11 G. Engdahl, Handbook of Giant Magnetostrictive Materials; pp. 127–35. Aca- response. The highest magnetoelectric voltage coefficient (dE/dH) demic Press, San Diego, CA, 2000. 12 ibid, p. 175. at 1 kHz of the composite was 5.90 V/cm Oe under a dc magnetic 13 J. Kuwata, K. Uchino, and S. Nomura, “Phase Transition in the Pb(Zn1/3Nb2/3) bias of 4.2 T. This maximum dE/dH of the composite, which is the O3–PbTiO3 System,” Ferroelectrics, 37, 579 – 82 (1981). highest ever reported, was obtained when both sides of the 14 J. Kuwata, K. Uchino, and S. Nomura, “Dielectric and Piezoelectric Properties of Terfenol-D had magnetostriction along the radial direction. The 0.91Pb(Zn1/3Nb2/3)O3– 0.09PbTiO3 Single Crystals,” Jpn. J. Appl. Phys., 21, 1298 – 302 (1982). dependence of the applied ac magnetic field direction was also 15 X. H. Du, U. Belegundu, and K. Uchino, “Crystal Orientation Dependence of determined by changing the sample orientation in the Helmholtz Piezoelectric Properties in Lead Zirconate Titanate: Theoretical Expectation for Thin coils. The dE/dH changed by the applied ac magnetic field and Films,” Jpn. J. Appl. Phys., 36, 5580 – 87 (1997).