| Memory devices constitute an important segment of semiconductor devices (about 20%) with Flash devices for nonvolatile memory, where the memory bit is stored on a floating gate embedded in the gate oxide of a MOSFET, forming a significant part of it. Flash scaling has kept pace with CMOS transistor scaling, lagging by a generation, with future scaling dependent on the ability to scale tunnel-oxide and inter-poly dielectric thicknesses while retaining long retention time, tasks at odds with each other.; As the memory device dimensions enter the quantum realm, novel device concepts are expected to play a dominant role in improving performance. Two concepts are expected to aid in Flash memory scaling, nano-floating gate memory and single electron memory. In the nano-floating gate memory, the memory bit is stored on nano-floating gates (crystals, nanodots) instead of a monolithic floating gate, thus aiding in retention of charge in the face of thinning tunnel oxides. Single electron memory, on the other hand, utilizes the classical phenomenon of Coulomb blockade which is a manifestation of the repulsion between like charges. These memory devices utilize an SET as a detector instead of a MOSFET. Nano-floating gate memory might serve as a bridge between Flash memory and the single electron memory.; However, single electron devices suffer from the problem of fluctuating background charges which change device thresholds randomly. One method to overcome these deleterious effects is based on the destructive readout of a few-electron charge bit from a floating gate, connected to a macroscopic charge reservoir by a tunnel barrier, and detected as oscillations in the SET. This dissertation reports the experimental realization of this device in Al/Al2O3 material system where the tunnel barrier is a small lateral gap between the floating gate and the charge reservoir. Further, utilizing the oxidation of aluminum in oxygen plasma, high threshold single electron memory devices are fabricated that have the potential of long retention time. Atomic layer deposited tunnel barriers are also investigated, for the ability to form barriers of any desired thickness. However, nucleation of the oxide needs to be improved for better insulating behavior. |
