Séminaire de Physique Théorique
Van der Waals ferroelectric heterostructures for in-memory computing and emergent electronicsJean-François Dayen (Université de Strasbourg)
Thursday 05 December 2024 14:15 - Amphi F21 - Batiment F
Résumé :
2D ferroelectric materials are attracting fast-growing interest for the implementation of complex more-then-Moore and beyond-Moore architectures that are challenging to design with standard
thin film technology.1 Here, I will present recent developments on the coupling of a 2D vdW electron gas with various ferroelectric gate controls. We will discuss how these systems allow for rethinking circuit topology and memory-logic interaction, opening up new research directions in the area of frugal computational enhancement and neuromorphic computing for AI. I will first detail how, by making use of the switchable polarization state of two split ferroelectric gates,
the electrical potential landscape within a semiconductor channel can be permanently and reconfigurable modified.2 While using the non-volatile ferroelectric states encoded in each gate, the ferroelectric logic circuits can function as six alternative logic gates, while CMOS circuits are
limited to a single function. Such Re-FeFET circuits demonstrate high compactness, with an up to 80% reduction in transistor count compared to standard CMOS design. Moreover, the device can operate as a photodiode and generate photovoltaic energy. Finally, I will present how light-structure interactions in vdW systems allow for the implementation of non-volatile electrical and optical
control of ferroelectric polarization in ferroelectric/semiconductor heterostructures.3 The wavelength-dependent study unveils ferroelectric polarization control and decouples the mechanisms driven by photogenerated carriers for each material and at the interfaces. Following, long-term potentiation/depression and spike rate-dependent plasticity are shown using electrical AND optical controls, enabling optically stimulated and optically assisted synaptic devices.
References
[1] Jin, T. et al. ACS Nano 2022, 16, 9, 13595–13611. [2] A. Ram. et al. ACS Nano 2023, 17, 21,
21865–21877. [3] M. Soliman, et al., ACS Appl. Mater. Interfaces 2023, 15, 12, 15732.2 (2018).
Liens :