基于统计数据的RRAM器件特性参数分析  

Analysis of RRAMI device characteristic parameters based on statistical data

作　　者：柯庆 代月花[1] Qing KE;Yuehua DAI(School of Integrated Circuits,Anhui University,Hefei 230601,China)

机构地区：[1]安徽大学集成电路学院,合肥230601

出　　处：《中国科学：信息科学》2024年第8期2021-2034,共14页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:62274002,61874001)资助项目。

摘　　要：本文统计了58款不同结构和不同材料的RRAM器件结构和电学特性参数,根据参数之间的联系建立了器件特性参数的计算模型,用统计学的百分位法计算了器件数据的差异性,获得了RRAM器件参数的定量指标.根据计算结果,我们发现RRAM器件的写时间、擦时间、写能量和数据维持时间均未达到预期指标, ECM器件低阻态时有量子力学效应,而VCM器件则没有量子力学效应. RRAM器件还需要进一步发展才能取代Flash技术.就发展趋势而言, ECM和VCM器件与NOR flash的差距最小,最可能取代NOR flash.This article provides statistics on the process and electrical parameters for 58 RRAM devices with different structures and materials.The calculation model of the device characteristics was built by these parameters,while the differences in device data were calculated by using statistical percentile method,and as a result,quantitative indicators of RRAM device parameters were obtained.Based on the calculation results,we found that the write time,erase time,write energy,and data retention time of RRAM devices did not reach the expected indicators.ECM devices exhibit the quantum mechanical effects in the low resistance state,while there is no quantum mechanical effect for VCM devices in the low resistance.RRAM devices still need further development to replace Flash technology.In terms of the RRAM device's development trend,ECM and VCM devices have the smallest gap with NOR fash and are most likely to replace NOR flash.

关 键 词：阻变存储器件 参数计算模型 百分位方法 统计指标 

分 类 号：TP333[自动化与计算机技术—计算机系统结构]