Nonequilibrium 1/f Noise in Low-doped Manganite Single Crystals G. Jung 1, B. Dolgin 1, D. Wu 1, 2, V. Markovich 1, Y. Yuzhelevski 1, M. Belogolovskii 3, Ya. M. Mukovskii 4 1 Department of Physics, Ben-Gurion University of the Negev, 84105 Beer-Sheva, Israel 2 Department of Materials Engineering, Monash University, Clayton, Australia 3 Donetsk Physical and Technical Institute, National Academy of Sciences of Ukraine, Donetsk, Ukraine 4 Moscow State Steel and Alloys Institute, Moscow, Russia For discussions and reprints please contact: [email protected] ac. il 1. Introduction ü In a complex phase diagram of colossal magnetoresistive (CMR) manganites FMM La 1 -x. Cax. Mn. O 3 (LCMO) the critical doping level separates ferromagnetic (FM) insulating ground state at x< x. C=0. 225 from FM metallic ground state above x. C. Moreover, for 0. 17TC. Intrinsic PS and metal-insulator transition at TC leads to percolation metallic-like conductivity in the FM state at T< TC. Low temperature resistivity is dominated by tunneling across intrinsic barriers associated with twins and grain boundaries, and/or with inclusions of insulating FM phase interrupting metallic percolating paths. . PMI percolating metal hopping FM/PM + M-I transition FMI + FMM TC= TM-I intrinsic tunneling Fig. 1. R(T), magnetic and transport properties of La 0. 78 Ca 0. 18 Mn. O 3 crystal vs. temperature. ü Noise data should reflect differences in transport mechanisms and magnetic states at different temperatures. 2. Transport and noise. ü Transport and noise measurements were performed with La 0. 78 Ca 0. 18 Mn. O 3 single crystals grown by a floating-zone method and cut into (8 x 3 x 1. 6) mm 3 bars with the longest dimension along <110> crystalline direction. . Fig. 2. Typical noise spectrum. Fig. 3. 1/f noise intensity vs. bias current. Fig. 4. 1/f noise intensity vs. voltage. ü The noise observed at all temperatures was of the 1/f-type, Fig. 2. ü At all temperatures, with exclusion of 77 K, the noise intensity scales as the square of the current indicating that the noise originates in current independent resistivity fluctuations converted into voltage fluctuations by dc current flow. ü Noise intensity plotted as a function of voltage collapses to a single line SV= l. V 2, where l=const. , meaning that temperature evolution of the noise seems to follow the R(T) curve, as shown in Fig. 5. . Indeed, the UPON # 1. Why at all temperatures, with exception of 77 K, resistance fluctuations constitute a fixed, temperature independent fraction of the total resistance, independently of the scattering mechanism and the magnetic state of the sample? Why different magnetic properties of the sample and different scattering mechanisms are not reflected in the noise? Fig. 5. R(T) and 1/f noise intensity vs. temperature. 3. Nonequilibrium 1/f noise at low temperatures. ü At 77 K the intensity of 1/f noise initially increases as I 2 but with further current increase starts to decrease, see Fig. 3. Transport mechanism at 77 K is dominated by inelastic tunneling. Nonlinear I-V curves are well described by the Glasman-Matveev (GM) model, Fig. 6. GM tunneling is a self organized process, number of involved states increases with increasing bias. Tunneling through higher number of intermediate states is characterized by higher conductivity, see Fig. 7. ü The number of involved states can be inferred from the analysis of power exponents of experimentally recorded I-V curves, see Fig. 8. Fig. 7. An increase of tunnel conductivity is marked by appearance of peaks in d 2 I/d. V 2. Fig. 6. GM model of indirect inelastic tunneling. Fig. 8. Principles and results of the analysis of the power exponent of nonlinear I-V curves. ü Noise data plotted together with d 2 I/d. V 2 as a function of bias enable one to tentatively associate noise maxima with changes in the number of the localized states in the tunnel barrier through which the tunneling occurs. ü The local noise maxima can be attributed to excess partition-like noise associated with additional degree of freedom of a tunneling carrier – a choice of two distinct paths of tunneling with different number of states. ü We have verified that the observed 1/f noise obeys Dutta-Horn-Dimon model at all temperatures, including 77 K. Decrease of the noise with increasing current can be therefore tentatively attributed to stress imposed on the trap states constituting elementary fluctuators for 1/f noise. The intensity of 1/f noise generated by an ensemble of asymmetric two-level elementary fluctuators will decrease with increasing asymmetry of the fluctuators. Initial quadratic increase of the noise with increasing bias reflects direct or resonant tunneling through N=1 impurities. The involved elementary two-level fluctuators are not located in the barrier and are not stressed by the bias. Fig. 9. d 2 I/d. V 2 and noise intensity as a function of bias voltage. Number N indicates number of states in the barrier inferred form the analysis of the I-V curves in Fig. 8. UPON # 2. What causes the nonequillibrium character of noise at low temperatures? Why the decrease of 1/f noise with increasing bias is observed only at 77 K when clear hallmarks of intrinsic tunneling can be seen already at T~ 120 K. Is then the proposed tentative mechanism of stressed fluctuators feasible and correct? Fig. 10. Energy structure of a symmetric (at low bias) and stressed (at high bias) elementary two-level fluctuator tentatively ascribed to impurities and defects forming inelastic tunneling states in intrinsic barriers dominating transport properties at low temperatures. up tup = tdn down up down tup >> tdn