Strong uniaxial in-plane magnetic anisotropy of (001)- and (011)-orientedLa0.67Sr0.33MnO3thin films onNdGaO3substrates

PHYSICAL REVIEW B 79, 214425 共2009兲 Strong uniaxial in-plane magnetic anisotropy of (001)- and (011)-oriented La0.67Sr0.33MnO3 thin films on NdGaO3 substrates H. Boschker,1,* M. Mathews,1 E. P. Houwman,1,† H. Nishikawa,1,2 A. Vailionis,3 G. Koster,1 G. Rijnders,1 and D. H. A. Blank1 1MESA⫹ Institute for Nanotechnology, University of Twente, 7500 AE Enschede, The Netherlands 2B.O.S.T., Kinki University, 930 Nishi-Mitani, Kinokawa 649-6493, Japan 3Geballe Laboratory for Advanced Materials, Stanford University, Stanford, California 94305, USA 共Received 17 December 2008; revised manuscript received 23 April 2009; published 17 June 2009兲 Epitaxial La0.67Sr0.33MnO3 共LSMO兲 ferromagnetic thin films were coherently grown on NdGaO3 共NGO兲 substrates with different crystal orientations of the surface plane. On the 共110兲o- and 共001兲o-oriented substrates, the film grows in the 共001兲pc orientation, and on the 共100兲o-, 共010兲o-, and 共112兲o-oriented substrates the film is 共011兲pc oriented 共we will use subindices o and pc for the orthorhombic and pseudocubic crystal structures, respectively兲. The lattice parameters and pseudocube angles of the deformed LSMO pseudocube have been determined from x-ray diffraction measurements. The in-plane magnetic easy and hard directions of these films have been determined from the dependence of the remnant magnetization on the angle of the in-plane applied field. For all substrate orientations there is a strong in-plane uniaxial magnetic anisotropy, determined by the crystal directions of the substrate surface. The easy and hard magnetic-anisotropy directions are explained consistently by the 共bulk兲 inverse magnetostriction model, except for the film on NGO 共112兲o. DOI: 10.1103/PhysRevB.79.214425 PACS number共s兲: 75.30.Gw, 75.47.Lx, 75.70.Ak I. INTRODUCTION The La1−xAxMnO3 manganites, where A is a divalent alkali earth element 共A = Ba, Sr, Ca兲, have been intensively studied, instigated by their observed colossal magnetoresistance 共CMR兲 effects near the ferromagnetic-paramagnetic transition temperature, TC. La0.67Sr0.33MnO3 共LSMO兲 is a ferromagnetic half-metal, which is expected to show near 100% polarization of the conduction-electron spin and therefore often used as electrode material in magnetic tunnel junctions 共MTJs兲.1–3 In MTJs the device current is sensitive to the relative orientation of the magnetization vector in the two electrodes. Any deviation from full 共anti兲parallelism will result in a reduced tunnel magnetoresistance ratio 共TMR兲. The preferential direction of the magnetization in a ferromagnet is determined by the magnetic anisotropy, which includes shape anisotropy, surface anisotropy, and magnetocrystalline anisotropy. For LSMO grown on 共001兲c 共we will use subindices c, o, and pc for the cubic, orthorhombic, and pseudocubic crystal structures, respectively兲 SrTiO3 共STO兲 different mechanisms compete, such as a uniaxial contribution from the surface steps, the substrate strain-induced biaxial magnetocrystalline anisotropy, and the shape anisotropy.4–6 As was recently pointed out, this competition between the different mechanisms results in nanoscale magnetic-domain formation in these films.7 In order to obtain electrodes with a well-defined magnetization direction and abrupt switching behavior, LSMO films with strong uniaxial anisotropy are required. Apart from the magnetic properties, the surface termination of the electrodes at the interfaces with the barrier of an MTJ is also considered to be of major importance for the MTJ characteristics. In contrast to the 共001兲pc surface which has a 共MnO2兲d− or 共La0.7Sr0.3O兲d+ terminating layer, the 共011兲pc surface has a 共LaSrMnO兲4+ or a purely 共O2兲4− termi1098-0121/2009/79共21兲/214425共6兲 nation. For an MTJ with 共011兲pc interfaces between the electrodes and the barrier material, the top and bottom interfaces are expected to be symmetrical and without interfacial charge transfer.8,9 The magnetic properties of LSMO films are known to be very sensitive to the strain imposed by the lattice mismatch between the film and the substrate.10–13 Magnetic properties of LSMO thin films are mainly studied in systems with 共001兲pc-oriented LSMO, which is grown on SrTiO3 共001兲c, NdGaO3 共NGO兲 共110兲o, and LaAlO3 共LAO兲 共001兲pc. Generally compressive strain 共LAO兲 enhances out-of-plane magnetization, whereas tensile strain 共STO兲 increases the in-plane magnetization component. The NGO 共110兲o substrate imposes compressive strain, which competes with the effect of demagnetization, and both out-of-plane10 and in-plane12 magnetizations have been observed. Here, we report on the determination of the in-plane magnetic anisotropy in thin 共⬍50 nm兲 LSMO films on NGO with different surface crystal planes 关NGO 共110兲o, 共100兲o, 共010兲o, 共001兲o, and NGO 共112兲o兴. LSMO is 共011兲pc oriented on NGO 共100兲o, NGO 共010兲o, and NGO 共112兲o, and on the other substrates the orientation is 共001兲pc. The different surface-plane orientations of the substrate impose a specific strain on the film which changes the LSMO crystal structure and therefore the magnetocrystalline anisotropy. In all cases the in-plane magnetization shows a strong uniaxial anisotropy. The LSMO pseudocube lattice parameters are determined by x-ray diffraction 共XRD兲 and the easy and hard axis directions are determined by the substrate in-plane crystal directions. In all cases, except for the film on NGO 共112兲o, the easy axis directions follow from the inverse magnetostriction effect. II. CRYSTAL STRUCTURE The orthorhombic crystal structure of NGO 共Ref. 14兲 has lattice parameters a = 5.43 Å, b = 5.50 Å, and c = 7.71 Å. 214425-1 ©2009 The American Physical Society PHYSICAL REVIEW B 79, 214425 共2009兲 BOSCHKER et al. TABLE I. In-plane and out-of-plane lattice directions for LSMO on NGO with different surface-plane orientations. The lattice mismatch is calculated for the in-plane directions. Figures 共A兲–共E兲 show the surface plane of the substrate indicated in gray for the NGO 共110兲, NGO 共001兲o, NGO 共010兲o, NGO 共100兲o, and NGO 共112兲o substrates, respectively. Because the NGO lattice parameters are all different, there are various in-plane strain states possible for the LSMO films, depending on the substrate surface plane orientation. The in-plane lattice mismatch between LSMO and NGO is defined as m关abc兴 = 共aNGO − aLSMO兲 / aLSMO, where aLSMO and aNGO are the lattice constants of the LSMO pseudocube and the NGO substrate, respectively, in the direction 关abc兴o of the substrate. The calculated in-plane lattice mismatch for the different orientations of the NGO substrate is given in Table I. The NGO 共110兲o substrate orientation results in a 共001兲pc oriented LSMO film due to the “cube-on-cube” stacking, as shown schematically in figure A in Table I. The in-plane sides of the LSMO pseudocube are aligned along the 关1̄10兴o and the 关001兴o lattice directions of the NGO substrate. The only other NGO orientation that results in LSMO 共001兲pc growth is NGO 共001兲o. In that case the LSMO cube is rotated in plane over 45° with respect to the NGO 关100兴o direction and the pseudocube is in plane aligned along the 关110兴o and 关11̄0兴o, as is shown in figure B in Table I. The NGO 共010兲o 共figure C兲, 共100兲o 共figure D兲, and 共112兲o 共figure E兲 substrates result in growth in the 共011兲pc direction of the LSMO, with different values for the lattice mismatch along the two in-plane directions for each substrate orientation 共see Table I兲. III. EXPERIMENTAL RESULTS The LSMO thin films 共⬍50 nm兲 were grown on the NGO substrates of different orientations 关共110兲o, 共001兲o, 共100兲o, 共010兲o, and 共112兲o兴 by pulsed laser deposition. The surface treatments necessary for a single terminated surface are described elsewhere.15 During the deposition material was ablated from a stochiometric target with a laser fluence of 3 J / cm2. The oxygen background pressure was 0.35 mbar and the substrate temperature was 750 ° C. The target to substrate distance was fixed at 4 cm. After LSMO deposition, the films were cooled to room temperature at a rate of 10 ° C / min in a 1 bar pure oxygen atmosphere. Atomic force microscopy measurements showed smooth surfaces with unit-cell high steps. The step heights were determined to be ⬃3.9 and ⬃2.7 Å for LSMO 共001兲pc and LSMO 共011兲pc, respectively. XRD measurements were used to determine the directions of the crystal axes of the NGO substrate and the structure of the LSMO film. The length of the pseudocubic lattice vector in the out-of-plane direction was obtained from XRD ␪-2␪ measurements. In comparison with the bulk LSMO value an out-of-plane elongation of the unit cell is found for all surface plane orientations, indicative of the compressive strain in the films. In order to determine the length of the in-plane lattice vectors reciprocal space mapping of an asymmetric reflection has to be performed. Figure 1 shows the reciprocal space maps of a 50 nm film of LSMO grown on NGO 共110兲o at the 共260兲o 关共024兲pc兴, 共620兲o 关共02̄4兲pc兴, 共444兲o 共共204兲pc兲, and 共444̄兲o 关共2̄04兲pc兴 reflections. All the maps show that the in-plane component of the film peak is equal to that of the substrate indicating coherent growth. For the complete determination of the unit cell not only the lengths but also the angles between the vectors have to be measured. The angle 214425-2 PHYSICAL REVIEW B 79, 214425 共2009兲 STRONG UNIAXIAL IN-PLANE MAGNETIC ANISOTROPY… FIG. 1. 共Color online兲 Reciprocal space maps of LSMO grown on NGO 共110兲o at 共a兲 the NGO 共260兲o, 共b兲 the NGO 共620兲o, 共c兲 the NGO 共444兲o, and 共d兲 the NGO 共444̄兲o reflections. Here we used Q = 4␲ sin ␪ / ␭, where ␪ is the Bragg angle and ␭ = 1.5406 Å. Q has an out-of-plane component in the 共110兲o direction and the in-plane components are in the 关11̄0兴o direction 关cases 共a兲 and 共b兲兴 and 关001兴o direction 关cases 共c兲 and 共d兲兴. Indices of the film reflections are shown in bold. between the two in-plane lattice directions follows from the substrate surface structure. This angle is 90° for LSMO grown on NGO 共110兲o, NGO 共100兲o, and NGO 共010兲o but not for LSMO grown on NGO 共001兲o and NGO 共112兲o. The angle between an in-plane and the out-of-plane lattice vector can be obtained from the difference in the lattice spacing between an asymmetric reflection with a positive and one with the same negative in-plane contribution.16 In Fig. 1 关for LSMO on NGO 共110兲o兴 the peaks corresponding to the 共204兲pc and 共2̄04兲pc reflections have the same lattice spacing, whereas the peaks corresponding to the 共024兲pc and 共02̄4兲pc reflections have unequal lattice spacing. From this difference an angle ␣ of 89.6° is derived. For LSMO grown on the other four surface-plane orientations of NGO the angles between the in-plane and out-of-plane directions are all 90°. The lengths and angles of the pseudocubic unit cell of LSMO grown on the different surface-plane orientations of NGO are summarized in Table II. The measurements show that all films are fully coherently strained to the substrate surface. Now the crystal structure of the different films is known, we turn to the magnetic properties. Usually torque measurements are used to determine the easy axis directions and the anisotropy strength. However, such measurements did not provide conclusive information on the anisotropy in the thin films considered here because the torque signal was dominated by the substrate signal.17 Instead we used vibrating sample magnetometer 共VSM兲 magnetization measurements to determine the easy axis directions.18 Room-temperature hysteresis loops were measured as a function of the in-plane field angle, ␾H. The loops show the typical features of a uniaxial anisotropy: a square loop in the easy direction and an approximately linear M-H dependence in the hard direction. The remanence versus field angle 共M r − ␾H兲 curves of all samples show the expected behavior of in-plane uniaxial anisotropy, described by M r共␾H兲 = M r共␾easy兲兩cos共␾H − ␾easy兲兩, where ␾easy is the direction of 共the in-plane component of兲 the magnetic easy axis. Figure 2 shows hysteresis loops along both easy and hard directions and typical M r − ␾H curves of LSMO films on both NGO 共100兲o and NGO 共001兲o. The magnetic easy and hard axes are found to be along in-plane NGO-crystal directions. For LSMO films on NGO 共110兲o and NGO 共010兲o the easy and hard directions are aligned with the in-plane NGO-crystal directions as well. For the films on NGO 共112兲o it is found that the easy axis rotates away from the 关1̄1̄1兴o direction with increasing film thickness. This is presented in detail elsewhere.19 A summary of the easy and hard axis directions is given in Table III. We do not observe any influence of the surface steps on the magnetic anisotropy directions, as was observed earlier for LSMO on STO 共001兲c.4 We repeated the experiments at low temperature 共150 K兲 and found no changes in the easy and hard directions. Temperature-dependent saturation magnetization measurements follow closely the Brillouin-functional dependence for a Weiss ferromagnet with TC ⱖ 350 K and a low-temperature saturation magnetization which is approximately equal to 3.5␮B / Mn. IV. DISCUSSION From the XRD measurements it is concluded that LSMO thin films grow coherently on the five different NGO substrate surfaces, discussed here. Due to the lattice mismatch the LSMO pseudocube is deformed, creating different strain values in the different in-plane crystal directions. For all films a strong in-plane uniaxial anisotropy is found, which dominates any other anisotropies, e.g., the biaxial magnetocrystalline anisotropy and surface-step-induced anistropy. It is evident that this anisotropy is due to the in-plane anisotropic strain in the films induced by the substrate. In general the magnetoelastic energy due to the inverse magnetorestrictive effect can be described with the formula20 Ei = 2 2 − 23 ␭100␴i共␣21␥i1 + ␣22␥i2 兲 − 3␭111␴i共␣1␣2␥i1␥i2兲, assuming that the magnetization is predominantly in the plane of the film. Here ␣ j and ␥ij are the direction cosines of the magnetization and the stress ␴i with the pseudocubic lattice vectors. ␭100 and ␭111 are the magnetostriction in the 关100兴pc and 关111兴pc directions, respectively. For uniaxial anisotropy the in-plane anisotropy easy direction is determined by the lowest energy state. To account for the case of compressive strain ␧1 in one in-plane direction, tensile strain ␧2 in the orthogonal in-plane direction, and resulting out-of-plane strain ␧3, the energy for a given in-plane magnetization direction is written as being due to the superposition of the two strain states, Eeasy,hard = E1共␴1兲 + E2共␴2兲. 关The energy due to the out-of-plane strain equals zero, E3共␴3兲 = 0, for in-plane magnetization.兴 In this ␴i = Y␧ជi is taken into account straightforway the sign of ជ wardly. The calculated anisotropy directions are given in Table III using 共the room-temperature values兲 ␭ ⬇ 1.3 214425-3 PHYSICAL REVIEW B 79, 214425 共2009兲 BOSCHKER et al. TABLE II. Pseudocubic lattice parameters of LSMO, grown on NGO with different surface-plane orientations, as determined from XRD measurements. The in-plane lattice parameters are equal to those of the corresponding substrate lattice parameters. The error in the length is 0.005 Å and the error in the angle is 0.1°. Length 共Å兲 LSMO pseudocube lattice parameters LSMO pseudocube angles LSMO 共001兲pc on NGO 共110兲o 3.85 a = cNGO / 2 共in plane兲 2 2 + bNGO 共in plane兲 b = 1 / 2冑aNGO c 共out of plane兲 3.86 3.91 3.86 2 2 + bNGO 共in plane兲 b = 1 / 2冑aNGO c 共out of plane兲 3.86 3.91 a = cNGO / 2 共in plane兲 b c 兩共b − c兲兩 = bNGO 共in plane兲 兩共b + c兲兩 共out of plane兲 LSMO 共011兲pc on NGO 共100兲o 3.85 3.89 3.89 5.50 5.51 a = cNGO / 2 共in plane兲 b c 兩共b − c兲兩 = aNGO 共in plane兲 兩共b + c兲兩 共out of plane兲 LSMO 共011兲pc on NGO 共010兲o 3.85 3.88 3.88 5.43 5.55 aNGO 2 2 + bNGO 2 89.6 ␣ ␤ ␥ 90 90 90 90 89.3 ␣ ␤ ␥ ⬔关a , 共b − c兲兴 ⬔关a , 共b + c兲兴 ⬔关共b − c兲 , 共b + c兲兴 89.9 90 90 90 90 90 ␣ ␤ ␥ ⬔关a , 共b − c兲兴 ⬔关a , 共b + c兲兴 ⬔关共b − c兲 , 共b + c兲兴 88.7 90 90 90 90 90 3.86 3.88 3.88 ␣ ␤ ␥ 89.4 89.6 89.6 5.46 5.52 ⬔关a , 共b − c兲兴 ⬔关a , 共b + c兲兴 ⬔关共b − c兲 , 共b + c兲兴 89.5 90 90 LSMO 共011兲pc on NGO 共112兲o 2 2 + bNGO 共in plane兲 a = 1 / 2冑aNGO b c 冑冉 冊 冉 冊 ␣ ␤ ␥ LSMO 共001兲pc on NGO 共001兲o 2 2 a = 1 / 2冑aNGO + bNGO 共in plane兲 兩共b − c兲兩 = Angle 共deg兲 2 + 共cNGO兲2 共in plane兲 兩共b + c兲兩 共out of plane兲 ⫻ 10−5 共Refs. 21 and 22兲 and Y = 1.3⫻ 1011 N / m2,23 obtained from magnetostriction and elasticity measurements on LSMO. It is assumed that ␭100 = ␭111 = ␭. For all cases we find complete correspondence with the experimentally determined easy axis directions, except for the case of NGO 共112兲o. The in-plane easy axis is in the in-plane direction with 共largest兲 tensile strain or in the case that both strain directions are compressive, in the direction with smallest compressive strain. Earlier reports ascribe the occurrence of in plane 共out-of-plane magnetization兲 关for example, in the case of LSMO on STO 共001兲 and LAO 共001兲, respectively兴 as being due to the in-plane tensile 共compressive strain兲. The resulting deformation of the oxygen octahedra surrounding the Mn, being compressed 共extended兲 in the out-of-plane direction, should then be the cause for the anisotropy. In the cases presented here the in-plane strain is unequal in both orthogonal directions giving rise to a further in-plane deformation of the octahedra, causing the uniaxial in-plane anisotropy. The easy axis direction in the case of LSMO/NGO 共112兲o is not described correctly. We ascribe this to the strong distortions of the pseudocube, changing the angles between all 214425-4 PHYSICAL REVIEW B 79, 214425 共2009兲 STRONG UNIAXIAL IN-PLANE MAGNETIC ANISOTROPY… netic domains and angle dependent coercivity measurements consistent with magnetization reversal by domain-wall motion 共both not shown here兲. Also the inverse magnetostriction model predicts out-of-plane anisotropy, which competes with demagnetization. Both aspects, rotation of the magnetization vector out of plane and the presence of multiple domains, and the associated magnetization reversal by domain-wall motion prohibit the use of the integration method20 here. Even though the model cannot be used to calculate the uniaxial-anistropy constant, it is successful in describing the easy axis directions. This is because the in-plane easy axis directions depend only on the sign of the energy difference between the in-plane magnetization in two orthogonal directions while the anistropy constant cannot be determined without taking the energy of the out-of-plane component of the magnetization into account. FIG. 2. Top panel: 共a兲 hysteresis loops of a 14-nm-thick LSMO/ NGO 共100兲o film along in-plane easy and hard directions. 共b兲 Remanence vs in-plane field angle for this film at room temperature. Arrows denote easy and hard directions. Bottom panel: 共c兲 hysteresis loops of a 25-nm-thick LSMO film grown on NGO 共001兲o along in-plane easy and hard directions. 共d兲 Remanence vs in-plane field angle at room temperature. Arrows denote easy and hard directions. pseudocube axes, whereas in the other cases most angles are orthogonal. This further reduction in the symmetry of the unit cell is clearly important for the anisotropy and is not taken into account in the simple model above. We find a significant discrepancy between the values for the in-plane anisotropy constants calculated from the model above and the values obtained from the magnetization loops with the integration method. We ascribe this difference to the fact that the latter method can only be used if the magnetization loop can be completely described mathematically.20 For our samples this is not the case as an out-of-plane contribution of the magnetization is present as well 共especially at low-field values兲 resulting in domain formation in the out-ofplane direction. This is concluded from magnetic-force microscopy 共MFM兲 measurements showing out-of-plane mag- V. CONCLUSIONS We have grown successfully epitaxial 共001兲pc- and 共011兲pc-oriented LSMO thin films on different NGO substrates. The LSMO growth orientation and crystal structure depends on the crystal orientation of the substrate surface plane. The films have strong in-plane uniaxial anisotropy with the easy axis directions related to the crystal directions of the substrate. The easy axis directions are explained by magnetostriction induced by the anisotropic in-plane strain. 共011兲pc-oriented LSMO films grown on NGO may have advantages over common 共001兲pc films when they are used as electrodes in MTJs since they combine strong uniaxial anisotropy with a surface termination that allows interfaces without interfacial charge transfer. ACKNOWLEDGMENTS We acknowledge financial support from NanoNed, the nanotechnology network in the Netherlands, the Dutch Technology Foundation 共STW兲, the Netherlands Organization for Scientific Research 共NWO兲, and the EU program NanOxide. TABLE III. Magnetic easy and hard axes for LSMO grown on NGO. For LSMO films on NGO 共112兲o the anisotropy is uniaxial, but the easy and hard axes are not aligned with the NGO-crystal directions 共Ref. 19兲. The model easy and hard directions have been derived from the inverse magnetostrictive effect. LSMO 共001兲pc on NGO 共110兲o LSMO 共001兲pc on NGO 共001兲o LSMO 共011兲pc on NGO 共100兲o LSMO 共011兲pc on NGO 共010兲o LSMO 共011兲pc on NGO 共112兲o In-plane direction Lattice mismatch 共%兲 Expt. Model 关11̄0兴o 关001兴o 关100兴o 关010兴o 关010兴o 关001兴o 关100兴o 关001兴o −0.47 −0.70 −1.11 +0.16 +0.16 −0.70 −1.11 −0.70 Easy Hard Hard Easy Easy Hard Hard Easy Easy Hard Hard Easy Easy Hard Hard Easy 关1̄10兴o −0.47 Approximately hard Easy 关1̄1̄1兴o −0.59 Approximately easy Hard 214425-5 PHYSICAL REVIEW B 79, 214425 共2009兲 BOSCHKER et al. 12 Joonghoe *j.a.boschker@utwente.nl †e.p.houwman@utwente.nl 1 J.-H. Park, E. Vescovo, H. J. Kim, C. Kwon, R. Ramesh, and T. Venkatesan, Nature 共London兲 392, 794 共1998兲. 2 M. Bowen, M. Bibes, A. Barthélémy, J. P. Contour, A. Anane, Y. Lemaitre, and A. Fert, Appl. Phys. Lett. 82, 233 共2003兲. 3 J. M. De Teresa, A. Barthélémy, A. Fert, J. P. Contour, R. Lyonnet, F. Montaigne, P. Seneor, and A. Vaurès, Phys. Rev. Lett. 82, 4288 共1999兲. 4 M. Mathews, F. M. Postma, J. C. Lodder, R. Jansen, G. Rijnders, and D. H. A. Blank, Appl. Phys. Lett. 87, 242507 共2005兲. 5 K. Steenbeck and R. Hiergeist, Appl. Phys. Lett. 75, 1778 共1999兲. 6 Z. H. Wang, G. Cristiani, and H. U. Habermeier, Appl. Phys. Lett. 82, 3731 共2003兲. 7 E. P. Houwman, G. Maris, G. M. De Luca, N. Niermann, G. Rijnders, D. H. A. Blank, and S. Speller, Phys. Rev. B 77, 184412 共2008兲. 8 Y. Mukunoki, N. Nakagawa, T. Susaki, and H. Y. Hwang, Appl. Phys. Lett. 86, 171908 共2005兲. 9 I. C. Infante, F. Sanchez, J. Fontcuberta, S. Fusil, K. Bouzehouane, G. Herranz, A. Barthelemy, S. Estrade, J. Arbiol, F. Peiro, R. J. O. Mossanek, M. Abbate, and M. Wojcik, J. Appl. Phys. 101, 093902 共2007兲. 10 P. Lecoeur, P. L. Trouilloud, G. Xiao, A. Gupta, G. Q. Gong, and X. W. Li, J. Appl. Phys. 82, 3934 共1997兲. 11 C. Kwon, M. C. Robson, K.-C. Kim, J. Y. Gu, S. E. Lofland, S. M. Bhagat, Z. Trajanovic, M. Rajeswari, T. Venkatesan, A. R. Kratz, and R. Ramesh, J. Magn. Magn. Mater. 172, 229 共1997兲. Dho, Y. N. Kwo, Y. S. Hwang, J. C. Kim, and N. H. Hur, Appl. Phys. Lett. 82, 1434 共2003兲. 13 F. Tsui, M. C. Smoak, T. K. Nath, and C. B. Eom, Appl. Phys. Lett. 76, 2421 共2000兲. 14 L. A. Vasylechko, W. Morgenroth, U. Bismayer, A. Matkovskii, and D. Savytskii, J. Alloys Compd. 297, 46 共2000兲. 15 M. Mathews, Ph.D. thesis, University of Twente, 2007. 16 A. Vailionis, W. Siemons, and G. Koster, Appl. Phys. Lett. 93, 051909 共2008兲. 17 K. Steenbeck 共private communication兲. 18 Torque measurements have to be performed with the sample fully magnetized in the direction of the applied field, which requires a large applied field. This field induces significant magnetization in the paramagnetic substrate which dominates the measurements. For angle dependent VSM measurements, however, the main information about the anisotropy is obtained from the remanence in the magnetization loop where the substrate magnetization is zero. 19 H. Nishikawa, E. P. Houwman, H. Boschker, M. Mathews, G. Rijnders, and D. H. A. Blank, Appl. Phys. Lett. 94, 042502 共2009兲. 20 S. Chikazumi, Physics of Magnetism 共Wiley, New York, 1964兲. 21 N. Zhang, X.-M. Yin, M. Wang, T. Schneider, and G. Srinivasan, Chin. Phys. Lett. 23, 463 共2006兲. 22 R. V. Demin, L. I. Koroleva, and A. M. Balbashov, J. Magn. Magn. Mater. 177–181, 871 共1998兲. 23 V. Rajendran, S. Muthu Kumuran, V. Sivasubramanian, T. Jayakumar, and B. Raj, Phys. Status Solidi A 195, 350 共2003兲. 214425-6