We have three fundamental circuit elements today; the resistor, the capacitor and the inductor. In 1971 it was suggested that there should be a fourth element by Professor Leon Chua of the University of California Berkeley. It would be called the memory resistor, or memristor. The memristor would have properties that could not be mimicked by any combination of the other three elements.
Basically, a memristor would change its level of resistance when current is applied, and this could possibly be used for storing data. If the resistance was high it could equal 1, if it’s low it would be the opposite, in this case 0. And the resistance would be the same if you cut power and then re-apply power. It’s a variable resistor that can be controlled through current and it remembers the resistance.
This is actually very similar to phase-change memory, which should hit the market later this year. The difference is that the resistance is changed through heat, not current, with phase-change memory. This requires a lot more energy, than a future memory built with memristors would.
While memristance has been observed in instances, the theory has remained a theory until now. A group led by, HP Senior Fellow and director of the Information and Quantum Systems Lab, Stanley Williams not only formulated a model of the memristor, but also built nanoscale devices that could be tested and prove that the memristor is in fact real.
“To find something new and yet so fundamental in the very mature field of electrical engineering is a big surprise,” said R. Stanley Williams, an HP Senior Fellow and director of the Information and Quantum Systems Lab (IQSL).
The resistor is build from two layers of titanium oxide, TiO2. The lower layer is complete, while the upper has an oxygen deficiency of less than 1%. The missing oxygen atoms leaves small holes in the layer, which alters the resistance of the layer. The layers are sandwiched between two wires. When current is applied to the upper wire, oxygen atoms will jump from the lower layer to the upper layer and change the resistance of the layer. The resistance can be reversed with subsequent currents.
The article was published in Nature, April 30 (abstract).
