Magnetoresistive memory in phase separated La0.5Ca0.5MnO3

JOAQUIN SACANELL; F. PARISI; P. LEVY; L. GHIVELDER

Trieste, Italia

Conferencia; Summer College and Conference on "Physics and Chemistry of Rare-Earth Manganites"; 2003

International Center for Theoretical Physics

Recently, it was shown in phase separated (PS) La5/8-yPryCa3/8MnO3 (y=0.3) and La1/2Ca1/2Mn1-zFezO3 (z = 0.05) the existence of a non-volatile magnetoresistive memory effect [1, 2]. Both systems studied are mostly ferromagnetic at low T, the effect was studied in the PS range, which is around 100 K for both compounds.The equilibrium ferromagnetic (FM) fraction f, can be changed by applying a low magnetic field (H < 1 T) and the change in f is persistent after the removal of H. The value of the applied H remains imprinted on the value of f.The effect is a consequence of the existence of an equilibrium PS state and a dynamical growth process of the coexisting phases. In this work we have studied the non-volatile memory effect on another PS compound, namely La0.5Ca0.5MnO3. Upon lowering temperature, this compound first undergoes a paramagnetic to FM phase transition at TC ~ 225 K, and then to a charge ordered antiferromagnetic phase at Tco ~ 150 K, so this system has an important COAFM fraction at low T. We have shown that the response of the system to the application of a magnetic field depends on the temperature range involved, due to the existence of three well-differentiated PS regimes [3]: ‘‘soft PS’’ for TC > T > 200 K , ‘‘intermediate’’ PS for 200 K > T > Tco , and ‘‘hard PS’’ for T < Tco.We have studied two ways of imprinting distinct values of H, hereafter referred as ZFC and FC procedures. ZFC procedure:The sample is cooled in zero field for resistivity (r) measurements and in a low H0 (~ 0.1 T) for magnetization (M). Then different values of Hap (>H0 for M measurements) are applied and turned off during cycles of ~1 hour. Either r or M are measured during the whole experiment.For T > 200 K and T < 70 K, the system has no memory of the applied Hap and recovers its H=0 state once the field is turned off. For 150 K < T < 200 K, the compound shows a reduction of the r and an increase of M during the application of distinct values of Hap. Once the field is turned off, the system does not recover its H=0 state and its subsequent M and r values depends on the previous Hap. The physical properties show no changes if, after the application of Hap1, Hap2 < Hap1 is applied. From these results, we infer an increase of f after each Hap-on / Hap-off procedure.For 70 K < T < 150 K, the effect is similar but it’s at least one order of magnitude smaller.FC procedure:The sample is cooled in the field cooling procedure (FC) in HFC ~ 1 T. Then H is lowered by the “application” of different values of dHap < 0 that are “turned on” and “turned off” during cycles of ~1 hour. The HFC serves to force a highly FM state [4, 5] and the application of dHap < 0, results in a reduction of f. So in this procedure, the memory is imprinted in an increase of non – FM regions. For T > 200 K and T < 70 K, the system has no memory and the process of apply and turn off the dHap is reversible.For 70 K < T < 150 K, the compound shows a huge reduction of M during the application of distinct values of dHap. Once the dHap is turned off, the system does not recover its HFC state and its M value depend on the previous dHap. The physical properties show no changes if, after the application of dHap1, dHap2 (with |dHap2|< |dHap1|) is applied. From these results, we infer a reduction of f after each dHap – on / dHap – off procedure. During the application of dHap, an evolution of M can be seen that indicates the increase of non – FM regions (after the rapid change due to the misalignment of FM domains).For 150 K < T < 200 K, the effect is similar but it’s at least one order of magnitude smaller.We have obtained that the memory effect is greater when we apply the field in the FC procedure. [1] P. Levy et al., Phys. Rev. B 65, 140401 (2002)[2] P. Levy et al., JMMM 258-259, 293 (2003)[3] R.S. Freitas et al., Phys. Rev. B 65, 140403 (2002)[4] F. Parisi et al., Phys. Rev. B 63, 144419 (2001)[5] J. Sacanell et al., Phys. B 320, 90 (2002)