Magnetization reversal probed by spin-polarized tunneling

Moon-Ho Jo, N. D. Mathur, M. G. Blamire
2002 Applied Physics Letters  
We report magnetic reversal processes in the magnetic tunnel junction ͑MTJ͒ based on a half metallic manganite, La 0.7 Ca 0.3 MnO 3 by comprehensive spin-polarized tunneling ͑SPT͒ measurements. The large tunnel magnetoresistance up to 77% of (R ap ϪR p )/R ap in the present MTJ is highly sensitive to the local magnetization fluctuation in the ferromagnetic electrodes and thus enables us to establish an instantaneous one-to-one correlation between the magnetization reversal and the SPT with the
more » ... d the SPT with the two-dimensional SPT measurements. We have found the systematic angular variations of the SPT features in the resistance-field curves, and based on the observed angular dependences, we propose a crucial role of the edge-domain pinning and the resultant multi-domain breakup for the magnetization reversal. Magnetic tunnel junctions ͑MTJs͒, which exploit spinpolarized electron tunneling ͑SPT͒ between asymmetric spin subbands at the Fermi level (E F ) in metallic ferromagnets, have received considerable attention because of the underlying physics and its potential applications. 1,2 It is commonly assumed that there is a direct correlation between the bulk magnetization vector and the spin-quantized axis of conduction electrons in the SPT and, in general, the tunneling conductance has a certain functional form with a relative angle of the two magnetization vectors in the electrodes. 3, 4 The study of magnetization reversal in MTJ has mainly relied on the direct observation of magnetization processes using techniques based on electron transmission microscopy and magnetic force microscopy, in conjunction with magnetic measurements or micromagnetic simulation. 5 In principle, it is also possible to establish a one-to-one correlation between the magnetization reversal and the SPT features during SPT measurements, provided that the SPT response to the external field is both simple and sensitive enough to fully reflect the magnetization states in detail. 6, 7 In this letter we report magnetization reversal processes directly characterized with the highly sensitive SPT measurements based on half metallic La 0.7 Ca 0.3 MnO 3 (LCMO). 8 The reversal processes of such MTJs are dependent on the orientation of the external magnetic field, and a direct correlation between the magnetization reversal and the SPT was established from the systematic variation of such SPT features in two-dimensional angular measurements. The MTJs of this study were fabricated from the heteroepitaxial LCMO/NdGaO 3 ͑NGO͒/LCMO trilayers grown on NGO ͑001͒ substrates by in situ pulsed laser deposition. The details of the junction fabrication are described elsewhere. 8 The MTJ elements were patterned using optical lithography and Ar ion milling to produce micron-scale junctions of a 6ϫ6 m 2 size. The magnetic field dependent junction resistance (R j ϪH) curves were measured in fourterminal ac measurements at 77 K and the typical sweep rate of the applied magnetic fields was at 0.5-5 Oe/s. Figure 1͑a͒ sketches the geometry of the twodimensional angular measurement in this study, where each junction resistance (R j )-field ͑H͒ curve recorded at a certain angle was traced as a function of the in-plane rotation angle. A series of typical R j ϪH curves at various in-plane rotation angles is shown in Fig. 1͑b͒ . Note that the tunnel magnetoresistance ͑TMR͒ is defined as (R ap ϪR p )/R ap in this study and thus it ranges between 0% and 100%. Particularly, it should be pointed out that the sharp switching is persistent at all rotation angles, although the overall features systematically change. Thus, in the angular SPT measurements the relative alignment of magnetization between the top and the bottom ferromagnets at each angle can be investigated directly by the tunneling conductance. The angular dependence of the features of the R j ϪH curves was investigated based on three criteria as indicated in Fig. 1͑b͒ , namely, the angular variation of ͑1͒ the switching fields H c1 and H c2 , ͑2͒ the magnitude of the TMR, ⌬R/R ap * , and ͑3͒ the slope of the top plateau ͕R(H c1 )ϪR(H c2 )͖/(H c1 ϪH c2 ). Note that the TMR is normalized with R ap * which represents the average value of the R ap , i.e., R ap * ϭ͕R ap (H c1 )ϩR ap (H c2 )͖/2, as indicated in Fig. 1͑b͒ . We have measured several junctions of the various junction conductances and TMR, and they showed essentially the same angular dependence as in Fig. 1͑b͒ . Figure 2͑a͒ shows the angular variation of the switching fields and they were fitted to the second order sinusoidal oscillations of H c1 ϳK uni • sin 2 (Ϫ/4) and H c2 ϳK uni • sin 2 (Ϫ/4)ϩ(K bi /4)• sin 2 2 for the bottom and top electrode. It was verified that the anisotropy constant K uni in the twofold symmetry and K bi in the fourfold symmetry are related to the crystalline anisotropy, and the size and the shape of the top electrode, respectively. 9 Note that the slight deviation of sinusoidal oscillations from the experimental data is due to the order of anisotropy assumed for the fitting, thus it does not affect the major conclusions to be drawn. The variation of the TMR shown in Fig. 2͑b͒ exhibits a rather complex angular dependence. It seems to be inversely dependent on the variation in the H c2 , i.e., the TMR is maximal when the switching of the bottom electrode occurs at the lowest field. It is also weakly related to H c1 . Thus, it can be approximately fitted to Ϫ͕kH c1 ()ϩlH c2 ()͖, where k and a͒ Electronic mail: mhjo@fas.harvard.edu APPLIED PHYSICS LETTERS VOLUME 80, NUMBER 15
doi:10.1063/1.1469678 fatcat:r2kkoiwa45dzvi7ofmi2s4d4ry