TY - JOUR
T1 - Kinetics and relaxation of electroresistance in transition metal oxides
T2 - Model for resistive switching
AU - Das, N.
AU - Tsui, S.
AU - Xue, Y. Y.
AU - Wang, Y. Q.
AU - Chu, C. W.
N1 - Copyright:
Copyright 2009 Elsevier B.V., All rights reserved.
PY - 2009/9/10
Y1 - 2009/9/10
N2 - The kinetics of electric-field-induced resistive switching across metal (Ag) -Pr0.7 Ca0.3 MnO3 interfaces has been investigated. The resistance hysteresis ΔR varies with the pulse amplitude V0 roughly as a step function with existence of a threshold voltage Vt for a fixed switching pulses width Tw. On the other hand, the ΔR varies with the pulse width (Tw) as a two-stage sequence at a fixed V0: an initial exponential rise with a time constant τS 2× 10-7 s and a slow linearly increasing tail. The slow linear part is dominant only in the quasi-dc switch (pulse width∼a few seconds) below Vt but negligibly small above it. The retentions of the ΔR corresponding to the two stages are also extremely different, indicating that different underlying processes are involved. The relaxation time (τR) is 108 s (∼year) or higher for the sub- μs switching, in strong contrast with the total disappearance of the ΔR after a few days for subthreshold (V0 Vt) quasi-dc switch. Different mechanisms, therefore, dominate the two different stages. More results obtained from time dependence study and impedance spectroscopy suggest that defect creation/annihilation, such as broken bonds under field, is likely the mechanism for the sub- μs switching and that a slow accumulative process (like diffusion) of defects may be responsible for the subthreshold quasi-dc switch. Many observations further suggest that the accumulative process is much more complicated than simple migration/diffusion of the pre-existing defects.
AB - The kinetics of electric-field-induced resistive switching across metal (Ag) -Pr0.7 Ca0.3 MnO3 interfaces has been investigated. The resistance hysteresis ΔR varies with the pulse amplitude V0 roughly as a step function with existence of a threshold voltage Vt for a fixed switching pulses width Tw. On the other hand, the ΔR varies with the pulse width (Tw) as a two-stage sequence at a fixed V0: an initial exponential rise with a time constant τS 2× 10-7 s and a slow linearly increasing tail. The slow linear part is dominant only in the quasi-dc switch (pulse width∼a few seconds) below Vt but negligibly small above it. The retentions of the ΔR corresponding to the two stages are also extremely different, indicating that different underlying processes are involved. The relaxation time (τR) is 108 s (∼year) or higher for the sub- μs switching, in strong contrast with the total disappearance of the ΔR after a few days for subthreshold (V0 Vt) quasi-dc switch. Different mechanisms, therefore, dominate the two different stages. More results obtained from time dependence study and impedance spectroscopy suggest that defect creation/annihilation, such as broken bonds under field, is likely the mechanism for the sub- μs switching and that a slow accumulative process (like diffusion) of defects may be responsible for the subthreshold quasi-dc switch. Many observations further suggest that the accumulative process is much more complicated than simple migration/diffusion of the pre-existing defects.
UR - http://www.scopus.com/inward/record.url?scp=70350605146&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=70350605146&partnerID=8YFLogxK
U2 - 10.1103/PhysRevB.80.115411
DO - 10.1103/PhysRevB.80.115411
M3 - Article
AN - SCOPUS:70350605146
VL - 80
JO - Physical Review B-Condensed Matter
JF - Physical Review B-Condensed Matter
SN - 1098-0121
IS - 11
M1 - 115411
ER -