JOURNAL OF APPLIED PHYSICS VOLUME 86, NUMBER 5 1 SEPTEMBER 1999 Magnetic properties of chromium doped rare earth manganites Ln0.5Ca0.5Mn1 xCrxO3 Ln Pr, Nd, Sm and 0.05 x 0.10... by soft x-ray magnetic circular dichroism O. Toulemonde,a) F. Studer, A. Barnabe´, and B. Raveau Laboratoire CRISMAT, UMR 6508 associe´e au CNRS, ISMRA, boulevard Mare´chal Juin, 14050 Caen Cedex, France J. B. Goedkoopb) ESRF, BP 220, 38043 Grenoble Cedex, France Received 23 November 1998; accepted for publication 24 May 1999 Soft-x-ray magnetic circular dichroism SXMCD at Mn, Cr L2,3 and Ln M4,5 edges of Ln0.5Ca0.5Mn1 xCrxO3 (Ln Pr, Nd, Sm and 0.05 x 0.10 bulk polycrystalline samples have been performed at T 20 K below the ferromagnetic Curie temperature. We show the existence of magnetic sublattice on each of the probed cations. Chromium cations order at low temperature antiparallel to the manganese subnetwork and rare earth cations likely exhibit a sperimagnetic ordering. These results are compared with magnetization measurements and a tentative correlation with magnetoresistance properties is discussed. This work also demonstrates that SXMCD can probe element and site specific magnetic properties of multicomponent systems. © 1999 American Institute of Physics. S0021-8979 99 02917-5 I. INTRODUCTION polaronic type model,5 based on the possible existence of dynamic Jahn­Teller distortion of the MnO6 octahedra, was The soft x-ray magnetic circular dichroism SXMCD proposed to explain such CMR properties. has now become a basic tool to investigate the magnetic Up to now, neutron studies of these manganites showing properties of solids.1 The SXMCD observed in near-edge lattice evolution and variation of isotropic Debye­Waller core absorption processes is related to the magnetic moment factors at Tc ,6,7 extended x-ray absorption fine structure of the photoexcited atom when the core electron is promoted studies showing changes in the Debye­Waller factor at Tc ,8 into final states that are responsible for ferro- or ferrimag- and XAS studies at the O K9 and Mn L2,3 edges10 showing netic properties of the system. SXMCD benefits also from changes in the local distortion of the MnO6 octahedra sup- the double selectivity of the x-ray absorption spectroscopy port strongly this small polaron model which includes the XAS : it is sensitive to the chemical species of the absorb- Jahn­Teller distortion. ing atom and to the symmetry of the unoccupied electronic Small angle neutron scattering11 and calorimetric states probed by the photoelectron taking into account the measurements12 were also interpreted as evidence of the for- electric dipole selection rules. mation of magnetic polarons. At high temperature, an acti- In this work, we used SXMCD to investigate the mag- vated conduction with a Curie­Weiss behavior of the mag- netic properties of chromium doped rare earth manganites netization would be observed whereas, just above Tc , the Ln0.5Ca0.5Mn1 xCrxO3 (Ln Pr, Nd, Sm 2 by determining conductivity would follow Mott's variable range hopping the element specific magnetic ordering. model and spin-cluster formation would be responsible for Extensive studies of the manganese oxide perovskites magnetization behavior.12 Below Tc , Park et al. have re- Ln1 xAxMnO3 (Ln Re; A Ca, Sr, Ba or Pb were carried cently observed a spin dependence of the Fermi level by high out especially in the past 5 years after the discovery of giant resolution spin resolved photoemission on La0.7Sr0.3MnO3 or even colossal magnetoresistance CMR in these thin films.13 These results are in agreement with spin polar- compounds.3 Doping the LnMnO3 antiferromagnetic insula- ized band structure calculations,14 which strongly suggest a tor by alkaline­earth cation leads to mixed Mn3 /Mn4 va- semimetallic behavior for the lanthanum manganites lency, ferromagnetism, and metallic conductivity, which below Tc . could be due to hole doping in a 2p oxygen band. This was Here, we present a direct investigation of the local mag- explained within the double-exchange mechanism.4 How- netic moment carried by manganese, chromium, and rare ever, the double-exchange model, considering primarily the earth cations probe at L2,3 edges for transition elements and spin-dependent hopping mechanism, turns out to be insuffi- M4/5 edges for rare earth elements. A correlation between cient to explain the insulator behavior above Tc . Thus, a magnetization measurements and SXMCD signals is pro- posed in connection with magnetoresistance properties. a Author to whom correspondence should be addressed; electronic mail: oliver.toulemonde@ismra.fr II. EXPERIMENT b Present address: Faculty of Mathematics, Computer Science, Physics and Astronomy, University of Amsterdam, Valckenierstraat 65, NL 1018 XE, The chromium doped rare earth manganites Amsterdam. Ln0.5Ca0.5Mn1 xCrxO3 (Ln Pr, Nd, Sm were prepared in 0021-8979/99/86(5)/2616/6/$15.00 2616 © 1999 American Institute of Physics J. Appl. Phys., Vol. 86, No. 5, 1 September 1999 Toulemonde et al. 2617 the form of sintered pellets following a classical method of solid state chemistry. Thorough mixtures of oxides CaCO3, Cr2O3, Mn2O3, Nd2O3, Sm2O3, or Pr6O11 were first heated in air at 950 °C for 12 h. The samples were then pressed into pellets and sintered first at 1200 °C and then at 1500 °C for 12 h in air. X-ray powder diffraction measurements showed single phase patterns. Magnetization curves M(T) were established with a vi- brating sample magnetometer. Samples were first zero field cooled before applying 0.6 T at 5 K. Measurements were carried out upon warning. The x-ray magnetic circular di- chroism and x-ray absorption studies of these phases were performed systematically on the samples previously studied for their transport and magnetic properties. The structural characterization of the samples was realized by x-ray difrac- tion, electron diffraction, and electron microscopy.15 X-ray absorption spectra at Mn, Cr L2,3 edge and Ln M4,5 edges were recorded using circular polarized light at the Dragon beamline ID12B of the ESRF Grenoble, France . To re- move residual asymmetries, XAS and XMCD spectra were obtained by measuring the total yield signal at each photon energy at the opposite direction of the applied magnetic field of 0.55 T; spectra were then recorded for both light helicities FIG. 1. SXMCD signal of Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd, Sm and with polarization between 80% and 90% 5% measured Sm0.5Ca0.5Mn0.90Cr0.10O3 at Mn L2/3 edges at T 20 K. Shoulder A is likely due to Mn4 (3d3) Ref. 19 . between 700 and 900 eV.16 All the spectra were recorded after a zero field cooled process at T 20 K below the ferro- magnetic Curie temperature of the samples. The samples nism and favors an antiferromagnetic ordering of the manga- were scraped in situ before each of the measurements. The nese subnetwork which reduces the SXMCD signal. base pressure in the spectrometer chamber was close to One also observe an increase of the magnetic moment 10 10 mbar at the beginning of the experiment. carried by the manganese cation for Sm The SXMCD signals were corrected for partial circular 0.5Ca0.5Mn1 xCrxO3 (x 0.05 and 0.10 . Introduction of chromium then causes polarization and are normalized to L3 edge after background spin realignment of part of Mn3 to create a ferromagnetic removal following a procedure described by Chen et al.17 manganese sublattice. B. SXMCD at Cr L III. RESULTS AND DISCUSSION 2,3 edges SXMCD at Cr L A. SXMCD at Mn L 2,3 edges of the same chromium doped 2,3 edges rare earth manganites are shown in Fig. 2. The first qualita- SXMCD at Mn L2,3 edges of chromium doped rare earth tive but important information that can be drawn from manganites Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd, Sm and XMCD signals is the relative orientation of both Cr and Mn Sm0.5Ca0.5Mn0.90Cr0.10O3 is shown in Fig. 1. The strong magnetic subnetworks. The magnetic moment carried by the negative peak at about 645.5 eV already observed by manganese cations, which is set parallel to the propagation Pellegrin18 exhibits a shoulder A, which corresponds to the vector of the soft x rays, was taken as reference. The Cr one observed on the XAS spectra. In agreement with a pre- SXMCD spectra clearly show that Cr atoms possess a net vious assignment,19 shoulder A is likely due to Mn4 (3d3). moment antiparallel to the Mn moments as indicated by the Earlier XAS measurements at the Mn K edge of the same positive SXMCD signal in the Cr L3 range. Moreover, what- samples showed that a formal charge of manganese is only ever the chromium ratio, SXMCD spectra show the same slightly larger than Mn3.5 and considered equivalent to fine structure as that calculated for Cr3 in octahedral crystal whatever the rare earth cation for a doping ratio is.20 Thus, field symmetry with 10 Dq 2 eV.21 These Cr L2,3 edge the variation of manganese formal charge cannot induce the atomic calculations confirm the Cr3 formal charge which observed changes in the SXMCD spectra at Mn L2,3 edges was already observed at Cr K edges.20 for the various rare earths studied in the In the Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd, Sm series, Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd, Sm series. Therefore, the second important point is that SXMCD intensities at the reduction of the SXMCD signals when going from Cr L2,3 edges vary in the same way as the ones observed at praseodyium to samarium manganite can be correlated to the the Mn L2,3 edge, resulting in agreement with the simulta- decrease of the mean A site cation radius ra , which in turn neous presence of chromium and manganese atoms on the can induce a reduction of the Mn­O­Mn angle already ob- same perovskite B site. Thus, both magnetic sublattices in- served by x-ray and neutron diffraction. This reduction of the teract with each other to give an average magnetic moment mean Mn­O­Mn angle reduces the double exchange mecha- carried by the perovskite B site. 2618 J. Appl. Phys., Vol. 86, No. 5, 1 September 1999 Toulemonde et al. FIG. 2. SXMCD signal of Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd, Sm and Sm0.5Ca0.5Mn0.90Cr0.10O3 at Cr L2/3 edges at T 20 K. Finally, for Sm0.5Ca0.5Mn1 xCrxO3 (x 0.05 and 0.10 , one also observed an increase of the magnetic moment car- ried by the chromium cation with the doping ration so that an increase of the ferromagnetic ordering on the chromium sub- lattice, as already observed on manganese sublattice, can be deduced from XMCD measurements. C. SXMCD at rare earth M4,5 edges The SXMCD signal at rare earth M4,5 edges FIG. 3. a SXMCD signal at rare earth M4,5 edges of of Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd and Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd and Sm0.5Ca0.5Mn0.90Cr0.10O3 at T Sm0.5Ca0.5Mn0.90Cr0.10O3 are plotted in Fig. 3 a , whereas the 20 K. Rare earth M5 edge energy was substracted: 929 eV for praseody- SXMCD integrated signals are shown in Fig. 3 b . In com- mium, 980 eV for neodymium, and 1083 eV for samarium. b Integrated parison, the respective rare earth edge energies were sub- SXMCD signal at rare earth M4,5 edges of Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd and Sm stracted: 929 eV for praseodymium, 980 eV for neodymium, 0.5Ca0.5Mn0.90Cr0.10O3 at T 20 K. and 1083 eV for samarium at the M5 edge. Taking as refer- ence atomic calculations of the dichroic signals at M4,5 edges of trivalent rare earth ions due to Goedkoop,22 Pr and Nd 4 L dE magnetic sublattices appear aligned parallel to the Mn mo- L3d 3 L2 z 3 dE ments, whereas samarium magnetic sublattice is aligned in L3 L2 the reverse way antiparallel to the Mn moments. in agreement with previous calculations18,19 where n3d is the 3d electron occupation number of the considered cation, L D. Sum rules application 3 L2 denote the integration ranges, and ( ) is the absorption cross sections of the sample when the magnetic By applying a sum rule,23 one should be able to estimate field is applied parallel antiparallel to the propagation di- the operators of the orbital magnetic moment Lz and the rection of the x rays with fixed circular polarization. Because operators of the spin magnetic moment Sz for each probed of the weakness of the SXMCD signal, applying the sum rule element and edge. to the Cr L2,3 edges provides a more rough estimation of the In the case of manganese and chromium oxides, the orbital magnetic moment L3d z , which turns out to be of the spin-orbit splitting of the core hole is not large enough to same order of magnitude as the one found in the case of the prevent mixing of the J contributions to the L3 and L2 manganese L2,3 edge. These orbital magnetic moment values edges24 such that an error in the determination of Sz as correpond to a quenched orbital contribution as usual in the large as 200% has been shown by the multiplet case of transition metal. calculations.25 So, one can only give Table I an estimation In the case of rare earth M4,5 edges, the estimation of the of the orbital magnetic moment: orbital magnetic moments and spin magnetic moments are J. Appl. Phys., Vol. 86, No. 5, 1 September 1999 Toulemonde et al. 2619 TABLE I. Orbital and spin magnetic moment carried by manganese 3d shell and rare earth 4 f shell. Manganese Chromium Rare earth Rare earth Rare earth L3d 3d 4 f 4 f 4 f z Lz Lz Sz Mz Compounds B B B B B Pr0.5Ca0.5Mn0.95Cr0.05O3 3.0 10 2 4.9 10 2 0.9 10 1 0.3 10 1 3 10 2 Nd0.5Ca0.5Mn0.95Cr0.05O3 5.6 10 2 2.0 10 2 1.5 10 1 1.1 10 1 7 10 2 not not not Sm0.5Ca0.5Mn0.95Cr0.05O3 0.3 10 2 1.0 10 2 measured measured measured Sm0.5Ca0.5Mn0.90Cr0.10O3 6.6 10 2 1.0 10 2 0.4 10 1 0.1 10 1 2 10 2 also given in Table I. The estimation of the spin magnetic depending on the applied magnetic field. Neutron diffraction moment is indirectly given by the second sum rule studies have estimated a weak ferromagnetic moment to 0.4(5) L 2 dE B on Pr ions and parallel to the Mn moment27 for Z M4 M5 Pr0.7Ca0.3MnO3 but, antiparallel to the Mn moment in the 2 Se 2 M dE 3 dE , case of Nd ions-with the same order of magnitude-for 5 M4 Nd where and are always the absorption cross sections 0.72Ba0.28MnO3.26 Thus the rare earth magnetic sublattice can be seen as a distribution of magnetic moments on a cone of the sample when the magnetic field is applied parallel and with the main axis aligned in the direction of manganese antiparallel, respectively, to the propagation direction of the magnetic moments. The axis of this conic distribution of rare x rays with fixed circular polarization. M4 and M5 denote the earth magnetic moments would then be parallel or antiparal- integration ranges and Se Sz 3 Tz , where Tz is the lel to the ferromagnetic manganese sublattice, depending on expectation value of the magnetic dipole operator which de- the sign of the magnetic exchange interaction, i.e., depending scribes correlations between the spin and position of each on the rare earth nature. The statistical distribution of the rare electron and Sz is the expected spin magnetic moment. earth cation and the mismatch on the perovskite A site, in- Hence, whereas the first sum rule provides a direct estima- ducing frustration in the magnetic interaction, would be at tion of L4f 4 f z , the estimation of Sz from the second sum the origin of the weak coupling and change of orientation of rule requires the knowledge of Tz . Tz is difficult to mea- rare earth magnetic moments. sure but theoretical values of the ratio Tz / Sz for the rare Recent SXMCD measurements at the Pr M earths have been published by Carra et al.23 T 4,5 edge on z / Sz is es- Pr timated at 0.58 for Pr3 (N 0.7Sr0.3MnO3 thin film29 showing an increase of the 4 f 2, L 5, S 1 , 0.14 for Nd3 SXMCD signal with the increase of the applied magnetic (N4f 3, L 6, S 3/2 , and 0.23 for Sm3 (N4f 5, L 5, field, could be interpreted also as the result of the reduction S 5/2 . Hence, as m 4 f 4 f orb Lz B and mspin 2 Sz , of the cone angle provided that the magnetic domain satura- one can estimate the total magnetic moment equal to morb tion occurs in low magnetic field. Increase of the applied mspin Table I . Lz and Sz are antiparallel and nearly magnetic field will tend to align the Pr moment parallel to compensate each other explaining the weak total magnetic the direction of the propagation vector of the soft x rays, thus moment. The weak magnitude of the total magnetic moment reducing the cone angle around the applied magnetic field is in agreement with the one calculated from neutron diffrac- direction. tion experiments for Nd3 and Pr3 (M3 Ln 0.45 B at 4 K on Ln0.7A0.3MnO3.26,27 Taking into account that the spin- orbit splitting of the core hole is certainly not large enough to prevent a mixing of the J contributions to the M5 and M4 edges, atomic calculations of Ln3 would be necessary to obtain the moment by scaling as it were done for Er3 .28 The sign of the operators of the orbital magnetic moment Lz and of the spin magnetic moment Sz for the probed rare earth is linked to the qualitative information provided by SXMCD concerning the orientation of the rare earth mag- netic moment parallel or antiparallel to the ferromagnetic manganese sublattice. IV. MODEL OF MAGNETIC ORDERING IN DOPED MANGANITES A. Rare earth magnetic moments To propose a magnetic model for the rare earth in the manganites, one must take into account that previous SXMCD work on Nd 19 0.72Ba0.28MnO3 has shown the exis- FIG. 4. Thermal variation under 0.6 T of the magnetization of the samples tence of a flipping of the neodymium magnetic sublattice Ln0.5Ca0.5Mn0.95Cr0.05O3 (Ln Pr, Nd, Sm and Sm0.5Ca0.5Mn0.90Cr0.10O3. 2620 J. Appl. Phys., Vol. 86, No. 5, 1 September 1999 Toulemonde et al. TABLE II. SXMCD asymmetries at 645.5 eV for Mn L2,3 edge at 580.7 eV for Cr L2,3 edge and at M4 edge for the various rare earths studied and samples calibrated. Weighted Mn SXMCD Cr SXMCD asymmetry on Ln SXMCD asymmetry asymmetry B site asymmetry Compounds % % pervoskite % Pr0.5Ca0.5Mn0.95Cr0.05O3 5.9 6.2 5.2 9 0.7 Nd0.5Ca0.5Mn0.95Cr0.05O3 4.7 4.3 4.2 5 0.7 Sm0.5Ca0.5Mn0.95Cr0.05O3 1.4 1.9 1.2 3 / Sm0.5Ca0.5Mn0.90Cr0.10O3 4.1 5.7 3.1 2 0.5 Nevertheless, this work confirms that, except for lantha- result is also usually observed in the CMR rare earth man- num, rare earth cations carry a magnetic moment due to 4 f ganites upon increase of the applied magnetic field. This is electrons. However, the existence of magnetic moments on likely due to the distribution of magnetic moments induced the rare earth cation seems to be uncorrelated with magne- by a variable spread in energy of magnetic polarons.11 toresistance properties which take place at higher tempera- ture than the estimated Curie temperature of the magnetic V. CONCLUSION ordering of the rare earth sublattice. This result was already Existence and orientation of local magnetic moments of pointed out in the SXMCD study of Nd0.72Ba0.28MnO3, specific atoms have been obtained from soft-x-ray magnetic where the observed flipping of the neodymium magnetic mo- dichroim measurements SXMCD in doped CMR mangan- ment with increasing applied magnetic field did not affect the ites. Qualitative information about the magnetic moment magnetoresistance properties. magnitudes of the probed cations has also been extracted However, magnetization measurement recorded at 0.6 T from the spectra. Soft XMCD of the doping transition metals for various samples studied by SXMCD is shown in Fig. 4. appears to be a unique technique to give informations about Considering the SXMCD results, one can now better under- the magnetic behavior of elements in weak amounts inside stand the variations of the saturation magnetization at 5 K. complex materials. However, the detailed interpretation of The total magnetization is the addition of manganese, chro- the data will require thoroughful multiplet calculations, es- mium, and rare earth moments. Table II presents the pecially for the estimation of the spin contribution to the SXMCD asymmetries ( )/( ) at 645.5 eV dichroic signal. for Mn L2,3 edge, at 580.7 eV for Cr L2,3 edge and at the rare The correlation between the total magnetization and the earth M4 edge. Contrary to the magnetic multilayers soft XMCD measurements allows us to better understand the system,30 it is difficult to determine an average magnetic magnetic ordering inside the transition metal doped CMR moment on the pervoskite B site by comparison with a ref- rare earth manganites. erence manganite because of the intrinsic properties of the The chromium magnetic sublattice appears to be antipar- manganese­oxygen bonds.31 In manganeses oxides, the allel to the manganese magnetic sublattice taken as a refer- magnetization depends on the Mn­O distance and on the ence and a reversal of Sm magnetic moments with respect to O­Mn­O angle which is linked to the mean A site cation Pr and Nd moments has been observed. The small magni- radius ra , which induces a reduction of the O­Mn­O tudes of the total moment of Pr and Nd observed by neutron angle and on the filling of the Mn(3d) ­ O(2p) molecular diffraction are in agreement with the small amplitude of the orbitals, i.e., on the Mn4 /Mn3 ratio. Nevertheless, consid- total magnetic contribution found in SXMCD spectra and let ering an average magnetic moment on the B site estimated us think of a conic distribution of rare earth magnetic mo- by the sum of the weighted asymmetries found at the Cr and ments around the direction of the applied magnetic field (H Mn L2,3 edge, these asymmetries show a qualitative agree- 0.6 T). Such a result may account for the easy reversal of ment with the variations of the total magnetization which the Nd magnetic moments upon the increase of the applied decreases from Pr to Sm. Moreover, one should take into magnetic field observed in the study of the Nd account the rare earth magnetic contribution to the total mo- 0.72Ba0.28MnO3 CMR manganite. ment which can be parallel or antiparallel to the manganese Further experiments will be necessary to determine the magnetic moment. Then, by adding to the average magnetic ordering temperature of the chromium and rare earths mag- moments on the perovskite B site, a positive magnetic mo- netic sublattices to correlate magnetic moment carried by ment for Nd0.5Ca0.5Mn0.95Cr0.05O3 and a negative magnetic chromium with magnetoresistance properties. moment for Sm0.5Ca0.5Mn0.90Cr0.10O3 in agreement with the relative orientation of the observed magnetic moment carried 1 by the perovskite A site, one can find a second reason for the P. Rudolf, F. Sette, L. H. Tjeng, G. Meigs, and C. T. Chen, J. Magn. Magn. Mater. 109, 109 1992 ; C. Giorgetti et al., Phys. Rev. B 48, 17, variation of the total magnetization at low temperature be- 12732 1993 ; J. Vogel, A. Fontaine, V. Cros, F. Petroff, J. P. Kappler, G. tween these two samples. Krill, A. Rogalev, and J. Goulon, Phys. Rev. B 55, 3663 1997 ; W. F. Finally, the transition width from the paramagnetic to Pong, Y. K. Chang, M. H. Su, P. K. Tseng, H. J. Lin, G. H. Ho, K. L. the ferromagnetic state increases drastically with the reduc- Tsang, and C. T. Chen, ibid. 55, 11409 1997 ; A. Kimura, S. Suga, T. Shishidou, S. Imada, T. Muro, S. Y. Park, T. Miyahara, T. Kaneka, and T. tion of the mean A site cation radius ra Fig. 4 . Such a Kanomata, ibid. 56, 6021 1997 ; G. Schutz, P. Fischer, E. Goering, K. J. Appl. Phys., Vol. 86, No. 5, 1 September 1999 Toulemonde et al. 2621 Attekofer, D. Ahlers, and W. Rossl, Synchrotron Radiat. News 10, 13­26 13 J. H. Park, E. Vescovo, H. J. Kim, C. Kwon, R. Ramesh, and T. Venkate- 1997 . san, Nature London 392, 794 1998 . 2 B. Raveau, A. Maignan, and C. Martin, J. Solid State Chem. 130, 162 14 S. Satpathy, Z. S. Popovic, and F. R. Vukajlovic, Phys. Rev. Lett. 76, 960 1997 ; A. Barnabe, A. Maignan, M. Hervieu, F. Damay, C. Martin, and 1996 . B. Raveau, Appl. Phys. Lett. 71, 3907 1998 . 15 A. Barnabe´, M. Hervieu, C. Martin, A. Maignan, and B. Raveau, J. Mater. 3 R. M. Kuster, J. Singleton, D. A. Keon, R. M. Greedy, and W. Hayes, Chem. 8, 1405 1998 . Physica B 155, 362 1989 ; R. Von Hemmolt, J. Wecker, B. Holzapfel, L. 16 M. Drescher, G. Schnell, U. Kleineberg, H. J. Stock, N. Muller, U. Hei- Schultz, and K. Samwer, Phys. Rev. Lett. 71, 2331 1993 ; H. L. Ju, C. nzmann, and N. B. Brookes, Rev. Sci. Instrum. 68, 1939 1996 . Kwon, Q. Li, R. L. Greene, and T. Venkaten, Appl. Phys. Lett. 65, 2108 17 C. T. Chen, Y. U. Idzerda, H-J. Lin, N. V. Smith, G. Meigs, E. Chaban, G. 1994 ; A. Maignan, Ch. Simon, V. Caignaert, and B. Raveau, Solid State H. Ho, E. Pellegrin, and F. Sette, Phys. Rev. Lett. 75, 152 1995 . Commun. 96, 623 1995 ; R. Mahesh, R. Mahendiran, A. K. Ray- 18 E. Pellegrin et al., J. Phys. IV 7, C2-405 1997 . chaudhury, and C. N. R. Rao, J. Solid State Chem. 114, 297 1995 ; S. Jin, 19 O. Toulemonde et al., J. Magn. Magn. Mater. 190, 307 1998 . H. M. O'Bryan, T. H. Tiefel, M. McCormack, and W. W. Rhodes, Appl. 20 O. Toulemonde, F. Studer, A. Barnabe´, A. Maignan, C. Martin, and B. Phys. Lett. 66, 382 1995 . 4 Raveau, Eur. Phys. J. B 4, 159 1998 . C. Zener, Phys. Rev. 82, 403 1951 ; P. W. Anderson and H. Hasegawa, 21 G. Van Der Laan and B. T. Thole, Phys. Rev. B 43, 13401 1991 . ibid. 100, 975 1955 ; P. G. De Gennes, ibid. 118, 141 1960 . 22 5 A. J. Millis, P. B. Litlewood, and B. I. Shraiman, Phys. Rev. Lett. 74, J. B. Goedkoop, thesis, University of Nijmegen 1989 . 23 5144 1995 ; A. J. Millis, B. I. Shraiman, and R. Mueller, ibid. 77, 175 B. T. Thole et al., Phys. Rev. Lett. 68, 1943 1992 ; P. Carra et al., ibid. 1996 . 70, 694 1993 . 24 6 V. Caignaert, E. Suard, A. Maignan, Ch. Simon, and B. Raveau, J. Magn. Y. Teramura, A. Tanaka, and T. Jo, J. Phys. Soc. Jpn. 65, 1053 1996 . 25 Magn. Mater. 153, L260 1996 ; J. L. Garcia-Munoz, M. Suaidi, J. Font- E. Pellegrin private communication . 26 cuberta, and J. Rodriguez-Carvajal, Phys. Rev. B 55, 34 1997 . D. E. Cox, P. G. Radaelli, M. Marezio, and S. W. Cheong, Phys. Rev. B 7 P. Dai, J. Zhang, H. A. Mook, S. H. Liou, P. A. Dowben, and E. W. 57, 3305 1998 . 27 Plummer, Phys. Rev. B 54, R3694 1996 . F. Fauth, E. Suard, C. Martin, and F. Millange, Physica B 241, 427 8 C. Meneghini, R. Cimino, S. Pascarelli, S. Mobilio, C. Raghu, and D. D. 1998 . Sarma, Phys. Rev. B 56, 3520 1997 ; C. H. Booth, F. Bridges, G. H. 28 L. M. Garcia, S. Pizzini, J. P. Rueff, J. Vogel, R. M. Galera, A. Fontaine, Kwei, J. M. Lawrence, A. L. Cornelius, and J. J. Neumeier, Phys. Rev. J. P. Kappler, G. Krill, and J. B. Goedkoop, J. Appl. Phys. 79, 6497 Lett. 80, 853 1998 . 1996 . 9 O. Toulemonde, F. Millange, F. Studer, B. Raveau, J.-H. Park, and C-T. 29 O. Toulemonde, F. Studer, J. Wolfman, B. Mercey, Ch. Simon, and J. Chen, J. Phys.: Condens. Matter 11, 109 1999 . Goedkoop unpublished . 10 F. Studer, O. Toulemonde, V. Caignaert, P. Srivastava, J. B. Goedkoop, 30 Y. U. Idzerda, L. H. Tjeng, H. J. Lin, C. J. Guttierrez, G. Meigs, and C. T. and N. Brookes, J. Phys. IV 7, C2-529 1997 . Chen, Phys. Rev. B 48, 4144 1993 ; S. Andrieu, E. Foy, H. Fischer, M. 11 J. M. De Teresa, M. R. Ibarra, P. A. Algarabel, C. Ritter, C. Marquina, J. Alnot, F. Chevrier, G. Krill, and M. Piecuch, ibid. 58, 8210 1998 ; M. A. Blasco, J. Garcia, A. del Moral, and Z. Arnold, Nature London 386, 256 Tomaz, W. J. Antel, Jr., W. L. O'Brien, and G. R. Harp, ibid. 55, 3716 1997 . 1997 . 12 P. S. Anil Kumar, P. A. Joy, and S. K. Date, J. Phys.: Condens. Matter 10, 31 J. L. Garcia-Munoz, J. Fontcuberta, B. Martinez, A. Seffar, S. Pignol, and L269 1998 . X. Obradors, Phys. Rev. B 55, R668 1997 .