A new device has been recently realized, though it was postulated years ago as one of the fundamental passive circuit elements. This new device alters its conductivity based on its electronic history and is referred to as a memory resistor or "memristor". In this implementation, instead of encoding "0" and "1" as the amount of charges stored in a traditional silicon-based cell, the low and high conductivity condition is interpreted as the OFF and ON states. The growing interest in memristors has resulted in the development of several concepts for the use of these devices, such as utilizing memristive systems for reconfigurable logic circuits, or new computer memory concepts. One of the very appealing aspects of this device is that both information retention and processing could be combined in a single device, a potentially massive advantage for achieving a reduced decoder size. Its not just academia that is interested, the majority of efforts in memristors are being advanced by industry and can be read about here: CNN Report
Just as living creatures alter their decision making process based on past experiences, memristors alter their response to a probe voltage based on their previous voltage exposure. This voltage path dependence can be exploited in a collection of circuit elements to achieve adaptive network systems though one of the most promising applications for memristors is the emulation of synaptic behavior.
The task of building an electronic version of a biological brain is a daunting task. Researchers are beginning to appreciate the complexity inherent in the "soft" aspects of adaptive memory and accept their lack of understanding at this point in time, but in addition, the "hard'' aspects of circuit density and power use for a biological brain are beyond the realm of current silicon technology. For example, how can the biological synapses, which occupy 10 Giga synapses per cubic cm in the cortex, be mimicked with an electronic version. In addition, the biological synapses consume minuscule power; have complex, non-linear dynamics; and, in some cases, can maintain their memory for decades. These characteristics, when combined with additional problems in our current understanding in the formalisms for adaptive learning, have until recently translated into making an electronic facsimile of a biological brain an unreachable goal. The advent of memristive devices has offered an alternative path to build an electronic brain architecture that can adaptively interact with the world.
As early as the late 1960's and early 1970's, a number of researchers presented the conductance switching properties of thin organic films and since then, organic bistable switching devices have been presented that were based on small molecules, polymers, as well composite constructs. These system were predominately developed to serve as a memory component in a von Neumann computer architecture. The mechanism by which the transitions occur are under much debate and is speculated to be vastly disparate for differing materials and/or device constructs. The mechanism for the switching in organic systems has been attributed to a range of phenomena, including filamentary conduction, charge transfer, ionic conduction, space charge and traps, as well as conformational change. The attached figure presents the hysteretic current-voltage response of poly(9-(9H-Carbazol-9-yl)nonyl methacrylate) [inset presents the corresponding monomer], where the directionality of the voltage scans indicated by red arrows. The Foulger group continues to pursue a number of different research thrusts into memristors that are presented by a review of our publications.