With classical scaling of CMOS transistors according to Dennard's scaling rules running out of steam, new possibilities to increase the functionality of an integrated circuit at a given footprint are becoming more and more desirable. Among these approaches the possibility to reconfigure the functionality of a transistor on the single devices level stand out, as by such an approach the same physical circuitry is enabled to perform different tasks in different configurations of the circuit. Reconfigurable transistors that allow the reconfiguration from a p-channel to an n-channel transistor and vice versa have emerged as an important example of such devices. The basic concepts required to built such devices have been proposed more then 20 years ago and the field has continuously developed ever since. In this article first the basic classification of reconfigurable field effect transistors is reviewed an described form a new angle. In the second part the important technology enablers to construct reconfigure field effect transistors are examined. Further the historical development, starting at the proposal of the main concepts up to the current status of device and circuit development are described. The most important additional features that have been introduced in the last years in order to even further increase the flexibility of the devices are discussed. Finally the application potential of reconfigurable transistors is described placing the spotlight on hardware security and neuromorphic applications.
|Published - Aug 2022