Anirban Ghosh

Ph.D. thesis

Modern computing based on Von Neumann architecture and storage devices are based on detecting a change in the state of a material. Hence, bistable materials e.g. ferromagnets and ferroelectrics become a natural choice to achieve this objective in real devices. Since these materials possess two stable states which can be switched from one to another and are also non-volatile in nature they can be used both as a memory element and for non-volatile RAM. However, with the demand for realizing brain inspired computing driven by applications beyond what is achievable by conventional computers e.g. pattern recognition etc. novel material properties need to be explored. Human brain unlike computing devices are analog in nature and dynamically processes data. Hence analog computers in principle can overcome the limitations of digital computers. In order to realize brain like computation in real devices one would need to create multiple switchable non-volatile states. Switching in ferroelectric thin films takes place via heterogeneous nucleation (at the ferroelectric - electrode interface) and subsequent growth of domains. In this thesis it is shown how we can create multiple states in a capacitor structure that comprises ferroelectric PZT by controlling the switching through the manipulating the statistics of the nucleation energy. We achieved this by controlling the local electric field at the ferroelectric-electrode interface by coupling the switchable polarization of PZT and the non-switchable polarization of ZnO.