Ferroelectric Tunnel Junctions at the Ultimate Scaling Limit

Ferroelectric Tunnel Junctions (FTJs) sandwich a ferroelectric (FE) tunnel barrier between two electrodes and exhibit analog resistance states depending on the degree to which the polarization of the FE domains are aligned.This project examines the scientific challenges related to scaling down FTJs based on HfZrO4 (HZO), by two objectives: Realize HZO thickness below 3 nm with retained FE and low defect levels.Using laterally nanoscale FTJs to probe and understand the switching dynamics and interactions of FE domains.

To reach the 1st obj. we will use ns laser annealing combined with advanced material design, supported by a comprehensive simulation model established early in the project. In-depth insight into defect formation/mobility will be gained by x-ray spectroscopy at e.g. MAX IV, in FTJs under electric stress as well as in-situ during HZO deposition.

For the 2nd obj. we will investigate nanoscale FTJs (< 50 nm), probe electrically and model the switching properties of individual domains and their stochasticity, and through unique 3D stacked devices study the interaction between neighbouring domains to gain insight into the mechanisms that determine FE domain size in HZO.FTJs can become the key building block in analog compute-in-memory electronics, a revolutionary tech that may bring enormous energy efficiency gains (10000x) to high-performance computing, such as AI.

Solving the challenges addressed here will be vital for the viability of FTJs in such technology.