What are Memristors?

What are Memristors and where are they used?

Professor Leon Chua developed the memristor theory in 1971. A scientist discovered the behavior of memristors when he was working in a lab at HP, trying to figure out crossbar switches. Memristors are also known as matrix switches, as they behave as a switch when connecting multiple inputs to several outputs. When Professor Chua looked at the assortment of resistors, capacitors, and inductors for the switching, he noticed a vital missing component, calling it the memory resistor or memristor. In 2006, Stanley Williams developed the practical model of a memristor.

One can consider the memristor as the fourth class of electrical component, after the familiar resistor, capacitor, and inductor. However, unlike the others, memristors exhibit their unique properties only at the nanoscale. Theoretically, memristors maintain a relationship between the time integrals of voltage across and current through two terminals of an element, as they are passive circuit elements.

The resistance of a memristor, or memristance, varies according to its function. It allows access to the history of the applied voltage via tiny read charges. The presence of hysteresis defines the material implementation of its memristive effects. This is its fundamental property, which looks more like a non-linear anomaly. Therefore, memristance is a simple charge dependent resistance, with a unit of Ohm. However, it has its own advantages.

The memristor technology offers lower heat generation as it utilizes less energy. Used in data centers, it offers greater resilience and reliability under interruptions of power. While not consuming any power when idle, memristors are compatible with CMOS interfaces. It is possible to store additional information as memristors allow higher densities to be achieved.

Physically, the memristor has two platinum electrodes across a resistive material, and its resistance depends on its polarity, magnitude, and length. The device retains its resistance even when the voltage is turned off, which makes it a non-volatile memory device. The resistive material can be titanium dioxide or silicon dioxide. As voltage is applied across the terminals, the oxygen atoms within the material disperse towards one of the electrodes. This activity stretches or contracts the material depending on the polarity of the applied voltage, thereby changing the resistance of the memristor.

Depending on their build, memristors are of two types—Iconic thin film and molecular memristors, and magnetic and spin based thermistors.

The material property of iconic thin film and molecular memristors rely more on different material properties of the thin film atomic lattices and application of charge makes these display hysteresis. Using these materials scientists make memristors of Titanium dioxide, ionic or polymers, resonant tunneling diodes, and manganite.

In contrast, magnetic and spin based memristors rely more on the property of the degree of electron spin. Therefore, these systems are aware of the polarization of electronic spin. There are two major types of such memristors—the spintronic memristors, and the spin torque transistor memristors.

With the practical demonstration of memristor manufacturing, their potential application has led to a rapid increase in research for using them in analog and digital circuits, such as programmable logic controllers, computers, and sensors. This has also led to development of theoretical models of memristors—Verilog-A, MATLAB, and Spice.