Abstract
Metal oxide semiconductor (MOS) devices are limited by both parasitic capacitance and scaling difficulties. A memristor, a two-terminal nanoscale electronic device with non-volatile memory, has the promise to overcome these limitations. Resistance in a memristor represents data and it is able to retain its prior value even after power withdrawal, thus remaining non-volatile. The memristor concept was introduced in 1971, although it took until 2008 to create the first such device on titanium oxide. Memristors can be interfaced with existing CMOS technology since both share nearly similar fabrication properties. This article explores the programmable aspects of memristors which have resulted in developments of new sequential logic families, namely, four memristor-based logic operations – Material Implication (IMPLY), Memristor Ratioed Logic (MRL), Hybrid Memristor-CMOS (MeMOS) Logic, and Memristor-Aided Logic (MAGIC) – each of which presents its own advantages and disadvantages. By utilizing both memristors and CMOS, in particular, MRL can perform as a complete logic family while reducing the physical size. MeMOS logic similarly utilizes a memristor-CMOS hybrid for more efficient computations. In comparison to IMPLY logic circuits, MeMOS logic provides for faster speed within a more compact circuit. A memristive-aided logic family, MAGIC, provides an alternative for memristive-based logic where it uses a memristor for each input and an additional memristor for the output.