Molecular Storage Elements for Proton Memory Devices

The one-bit memory-cell structure of choice for use in digital integrated circuit (IC) applications would be a single field-effect-transistor (FET) that includes a binary information storage element and acts as a memory-cell addressing switch. The storage element can be placed in two readily discernible physical states that modulate the transistor’s channel conductivity enabling data to be read electrically by sensing the current level of the transistor. The data sensing mode of memory transistors renders them robust against read-current distortions to the storage mechanism; an important advantage over their two-terminal counterparts. In the case of organic memory transistors, various combinations of storage elements, in the form of charged electrets or ferroelectric-like gate dielectrics, and organic semiconductor channels have been actively investigated for emerging large-area lowperformance logic circuit applications on flexible substrates. Regarding inorganic Si-based ICs, the polysilicon floating-gate (FG) storage element has been the dominant technology for the implementation of extremely durable nonvolatile memory transistors (currently referred to as FG flash transistors) for the past three decades. With the increasing challenge of scaling flash memory, various storage elements based on novel gate materials and physical storage principles have been proposed as modified flash cells. Significant development efforts have been devoted to storage elements exploiting: (a) the trapping of electrons and/or holes in nitride gate dielectrics or in isolated nanocrystals embedded in FET gate dielectrics (referred as nano-floating gate memory, NFGM); (b) the ferroelectric polarization of gate dielectrics (ferroelectric memory FET, FEMFET); and (c) the motion of protons within the gate dielectrics (proton memory). Apart from the benefits of using a modified flash memory architecture and of being radiation-tolerant, the proton memories also have the advantage, over FEMFETs and NFGMs, of being able to be programmed at much lower voltages, thus offering an attractive alternative for ultradense low-voltage, non-volatile data storage. While promising device results have been obtained, the conventional methods of forming protonmemories still face critical issues towards the development of commercial products. For example, all