SPICE Model of Memristive Devices with Threshold
SPICE Model of Memristive Devices with Threshold
复制标题
DOI:
--
复制
发表时间:
2012-04
期刊:
影响因子:
--
通讯作者:
Y. Pershin;M. Ventra
中科院分区:
文献类型:
--
作者:
Y. Pershin;M. Ventra
Abstract—Although memristive devices with threshold voltagesare the norm rather than the exception in experimentallyrealizable systems, their SPICE programming is not yet common.Here, we show how to implement such systems in the SPICEenvironment. Specifically, we present SPICE models of a popularvoltage-controlled memristive system specified by five differentparameters for PSPICE and NGSPICE circuit simulators. Weexpect this implementation to find widespread use in circuitsdesign and testing.Index Terms—Memristive devices, memristor, memristormodel, threshold dynamics. I. I NTRODUCTION I N the last few years, circuit elements with memory,namely, memristive [1], memcapacitive and meminductive[2] systems have attracted considerable attention from differentdisciplines due to their capability to store and manipulateinformation on the same physical platform [3]. However, whencombined into complex circuits, progress in this field signifi-cantly relies on the available tools at our disposal. One suchtool is the SPICE simulation environment, commonly used incircuit simulations and testing. While several SPICE modelsof memristive [4], [5], [6], [7], [8], [9], [10], memcapacitive[6] and meminductive [6] elements are already available, theytypically [4], [5], [6], [7], [8] rely on physical models withouta threshold (see, e.g., Refs. [11], [12]).Threshold-type switching is instead an extremely importantcommon feature of memristive devices (for examples, see Ref.[3]) and, due to physical constraints, likely to be common inmemcapacitive and meminductive elements as well. Indeed, itis the threshold-type switching which is responsible for non-volatile information storage, serves as a basis for logic oper-ations [13], [14], etc., and therefore, it can not be neglected.In the present paper we introduce a SPICE model fora memristive device with threshold voltage that has beenproposed by the present authors [15]. Using this type of mem-ristive devices, we have already demonstrated and analyzedseveral electronic circuits including a learning circuit [15],memristive neural networks [16], logic circuits [14], analogcircuits [17] and circuits transforming memristive responseinto memcapacitive and meminductive ones [18]. These pre-vious results thus demonstrate the range of applicability ofthe selected physical model. As a consequence, we expect itsSPICE implementation to find numerous applications as well.