ReRAM stands for Resistive Random Access Memory. It is a kind of non volatile Random Access Memory or RAM which, as suggested by the name, stores memory as it changes the resistance across a dielectric material. This dielectric material is usually called a memristor.
A ReRAM is essentially based on a three layer structure, comprising of a top electrode and a bottom electrode, along with a switching medium that is sandwiched in between the two electrodes. When a certain voltage is applied to the two electrodes on the top and the bottom respectively, a filament is formed in the silicon based switching medium which creates a resistance. This is a conduction path that is formed between two insulating layers due to the application of a sufficiently high voltage. If the filament that has been created is reset or broken, it produces a high resistance. On the other hand, if t is being set, then it is a low resistance system. This variability in resistance can be achieved by varying the level of voltage applied. A ReRAM essentially creates defects within a thin oxide layer with the movement of the oxygen ions and the oxide vacancies acting as analogs to the movement of electrons and holes in a semiconductor.
Multiple materials can be used as the switching material between the two electrodes when it comes to the structure of a ReRAM. These include chalcogenides, perovskites, metal oxides of binary transition metals, solid state electrolytes, organic charge transfer complexes or donor acceptor systems, as well as two dimensional insulating materials such as boron nitride.
ReRAM is often preferred over other systems such as PRAM because it operates at a much faster timescale when it comes to switching which can take as little as 10 nanoseconds. It also has a simpler cell structure which is relatively smaller in size along with consuming a lesser amount of power. In fact, it is much better suited in low power systems as compared to flash or racetrack memory since it needs little voltage to operate. It is also frequently used to design caches because of its small access latency as well as high density.
The purpose for which ReRAM was initially introduced was to replace flash memory and NAND, although its adoption has not been as rapid and widespread as of yet. Companies are still exploring as to what materials can be used with ReRAM and what works best, after which increased use of this technology is expected. The goal is to be able to master the physics and science behind the switching process so that the reliability and performance levels of this technology can be increased to acceptable levels for widespread use. Currently, it is expected to find a majority of its application in the embedded memory industry.