In this paper, the authors focus on two well-identified switching mechanisms, namely, interfacial (or homogeneous) switching and filamentary switching. These switching mechanisms have been reported in various devices, but a broader analysis remains to be conducted. By comparing the performances of TiO2- and HfO2-based resistive switching devices in terms of variability, retention, controllability, and switching energy, the authors discuss how oxygen vacancies organization can determine a general set of properties that will define the range of applications that could be envision for each material/device technology.

No abstract

the latter and to achieve extremely high packing density. [9] The understanding of switching mechanisms at device level is a key factor for a technology to be viable at very large scale integration. Different types of resistive memory devices have been studied in the past years such as oxide-based memories (OxRAM) and conductive bridge random access memories (CBRAM). It should be noted that for both technologies it is very challenging to combine good cycling endurance, stable retention and high window margin (WM). Two distinct resistive states can be obtained based on a reversible filament formation (SET operation) and rupture (RESET operation) inside an insulating layer sandwiched between two electrodes; SET operation being the switching from a high resistive state (HRS) to a low resistive state (LRS) and RESET the reverse operation (HRS to LRS). In the case of CBRAM, metal ions coming from top electrode (TE) are introduced in the insulating layer. Working principle is based on a metallic filament formation and dissolution controlling performances. [10] These devices present high WM, relatively low endurance, and poor retention stability. [11,12] In the case of OxRAM, oxygen vacancies creation and annihilation inside the oxide dominates the switching mechanism. [13][14][15] This technology shows low WM combined to high endurance and stable retention. [3,16] While lots of effort have been done lately to improve switching speed and power consumption in RRAM, [17,18] several challenges need to be overcome, namely the high extrinsic (device to device) and intrinsic (cycle to cycle) variability in RRAM characteristics. [7] High WM could potentially help solving this variability by maximizing the ratio of HRS over LRS. Moreover, coupling high WM and high endurance (up to 10 8 cycles required for storage class applications [19] ) remains a critical issue. Combining CBRAM and OxRAM in one hybrid oxide-based CBRAM (hybrid-RRAM (HRRAM)) where filament can be composed of metal ions and oxygen vacancies could offer alternative performances such as high WM coupled with high endurance. Recent studies have identified materials issues in oxide and metal based RRAM. [20,21] However, material properties study is still lacking in HRRAM to guide stack choice (oxide vs electrodes) toward a given application. In a previous work, a trade-off between endurance, window margin, and retention was demonstrated. When comparing various HRRAM electrical performances and filament composition, Here, the impact of copper and oxygen vacancy balance in filament composition as a key factor for oxide-based conductive bridge random access memories (hybrid resistive random access memories (HRRAMs)) performances is investigated. To this aim, several RRAM technologies are studied using various resistive layers and top electrodes. Material analyses allow to highlight the hybrid aspect of HRRAM conductive filament. Density functional theory simulations are used to extract microscopic features and highlight differences from a material point of view. Integr...

In this work, we address Resistive RAM (RRAM) variability. To this aim, we investigate various RRAM technologies (Oxide RAM and Conductive Bridging RAM), integrated on kb 1T1R arrays. Impact of variability is evaluated and discussed for various RRAM features: window margin, switching speed, consumption, retention and endurance. Solutions are proposed in order to improve overall RRAM performances. Optimized programming schemes are discussed. Importance of controlling the memory operation and programming energy to improve RRAM reliability is put in evidence.

No abstract

No abstract

No abstract

No abstract