Nonvolatile Reconfigurable Transistor via Ferroelectrically Induced Current Modulation

Abstract

New logic in memory (LiM) architectures, able to unify memory and logic functionalities into a single component, are highly promising for executing self-learning algorithms such as artificial neural networks (ANNs), with lower energy requirements. The multigated reconfigurable field effect transistor (RFET) is a novel type of logic device that can be fully reconfigured at run-time, promising to be a very versatile platform for logic applications. If equipped with a memory element, then it would represent the ideal building block for LiM-enabling hardware with embedded self-learning capabilities. To reach this goal, here we investigate the integration of a ferroelectric Hf0.5Zr0.5O2 (HZO) layer onto dual top gated RFETs. We demonstrate that HZO polarization charges can be successfully employed to tune the height of the two Schottky barriers, influencing the injection behavior and allowing the selection of the majority carriers, thus defining the transistor mode and switching it between a fully p-type transport to a prevalently n-type one. Moreover, we show that the modulation strength is strongly dependent on the height of the pulse used to polarize the ferroelectric domains, allowing for the selection of different current levels. All the different achievable states show a good retention over time, owing to the stability of the HZO polarization. The limitations of the produced devices are discussed, alongside possible mitigation strategies. The presented results demonstrate that ferroelectric HZO can modulate the carrier injection across Schottky barriers in RFET devices. This approach paves the way for the future realization of a fully optimized nonvolatile RFET, an ideal building block for novel LiM hardware, enabling low-power circuits for ANN execution.