Unconventional nanoscopic ferroelectricity in hafnia-zirconia thin-films

Hafnia-based thin-films exhibit ferroelectricity, which becomes more robust with film-thickness reduction. This behavior is unconventional and new, and seems to be less affected by factors that kill ferroelectricity at nanoscale such as depolarization fields. Such Si compatible nanoscopic ferroelectricity is very good news for both microelectronics and ferroelectric communites. However, the origin of this new kind of FE behavior remained elusive owing to the non-existence of single-phase, single-crystalline samples. Utilizing pulsed-laser deposition (PLD) we first optimized and synthesized the desired epitaxial single-phase samples. Through thorough structural characterization (XRD and STEM) we discovered a new polar phase with rhombohedral symmetry, stabilized by nanoparticle pressure and substrate strain, that is responsible for the ferroelectric behavior. I will discuss our most recent systematic studies on trying to replicate and formulate guidelines to obtain this novel phase on other substrates, especially Si[1,2]. These films show ferroelectricity all the way upto 2 nm film thicknesses, which lead us to create very homogenous multiferroic tunnel junctions[3]. The structure-property correlations of these tunnel junctions was investigated with some of the state-of-the-art microscopy techniques such as differential phase contrast STEM. These results will be presented.

REFERENCES
1.Wei, Nukala et al., Nature Mater., 17, 1095 (2018)
2.Nukala, Antoja et al., ACS Elec. Mater., 1, 2585 (2019)
3.Wei, Nukala et al.,Phys. Rev. Appl. 12., 031001 (2019)

Future agenda: top-down neuromorphic computation -  In the latter part of the talk, I will motivate my future plans from a neuromorphic computation point of view. I will try to convince that materials science has a lot to offer to this broad field, thus opening up opportunities to venture into unexplored territories. State-of-the-art in materials science includes understanding the physics of two or three terminal devices at a single device level, and further depending on the expertise of engineers to integrate these devices into a large scale neuromorphic architectures. I refer to such modular assembly of individual devices as a bottom-up approach. However, the human brain is a complex dynamical system, exhibiting recurring spatio-temporal electrical patterns emerging from the interaction between large number of non-linearelements (1011neurons and 1015synaptic connections), governed by scale-free neuronal avalanches. Such spatial and temporal complexity also occurs in several materials systems. To explore these systems in what can be referred to as a “top-down neuromorphic strategy”, the idea will be to first characterize the complexity arising from a large number of interacting elements in these systems. For this, I will identify some potential materials systems(e.g. networks of ferroelectric domains and domain boundaries), and propose not just structure-property experiments, but also charge avalanche counting experiments to characterize both global complexity and the local elements responsible for it. Some of these investigations are inspired from the neurobiology community. The hope is to be able to create or draw guidelines to creating “inorganic neuromorphic topologies” from self-assembled complex materials systems, and move towards understanding strategies to compute with them. From a more fundamental perspective, with the famous statement of Richard Feynman ‘what I cannot create, I cannot understand’ as an inspiration, we aim to create inorganic “analogues” of a part of mammalian brain and understand the operational principles of not just the inorganic systems under study, but also of their biological counterparts..