No abstract

No abstract

For the practical use of resistive random access memory (ReRAM), many formation/rupture models of conductive paths are proposed. In this paper, we report both the probability of conductive path formation on grain surfaces and the marked drastic change in conductivity caused by a small number of atoms migrating on grain surfaces, determined by using experimental and calculation results complementarily. Experimental results of resistive switching operating modes suggest that resistance changes at grain boundaries, to which our calculation results can give an explanation. The energy for the conductive change from a low-resistance state to a high-resistance state is estimated to be about 0.05 eV per surface atom, which is much smaller than the formation and migration energies of vacancies (1.44–4.42 eV) and is comparable to the estimated temperature of the conductive path in the reset process.

Practical use of Resistive Random Access Memory (ReRAM) depends on thorough understanding of the resistive switching (RS) mechanism in polycrystalline metal oxide films. Based on experimental and theoretical results of NiO based ReRAM, we have proposed a grain surface tiling model, in which grain surfaces (i.e. grain boundaries) are composed by insulating and conductive micro surface structures. This paper reports the adequacy of our model to the NiO based ReRAM and universality of surface electronic properties in metal oxides of NiO, CoO and MgO. Experimental results of RS operating modes suggest that the resistance changes in the grain boundaries, supporting our model. First-principles calculation results suggest that our model can be adopted to other metal oxide materials and the RS from a low resistance to a high resistance can be caused at 1000 K, which agrees with previous experimental reports.

No abstract