No Strain, No Piezoelectric Gain—Pushing Piezoelectric Boundaries via Strain Engineering in Composition-Tuned Wurtzite Scandium-Doped Aluminium Nitride in Neuromorphic Devices

Piezoelectric materials play a crucial role in the development of neuromorphic devices, particularly for emulating biological synapses and enabling adaptive bio-interfacing. In this study, we investigate the structural and piezoelectric properties of scandium-doped aluminum nitride (Al_1-<i>x</i>Sc<i>_x</i>N) using density functional theory (DFT) calculations, with a focus on strain engineering for optimizing piezoelectric performance.<br/>Our results reveal a phase transition from the wurtzite to the rock-salt structure at approximately 60% Sc composition. Interestingly, we find that, actually, between 50% to 60% Sc, the wurtzite structure transitions to a hexagonal layered structure, which does not display any piezoelectric response. Below 50% Sc, the wurtzite structure exhibits desirable piezoelectric properties, making it an ideal candidate for neuromorphic applications. This finding highlights the importance of carefully controlling the Sc doping level to maintain the desired wurtzite phase and piezoelectric behavior. Subsequently, we explore the effects of various Sc doping levels below 50% Sc on the structural optimization and piezoelectric coefficients using the special quasirandom structure (SQS) approach. This is followed by the investigation of biaxial strain engineering which reveals that the tensile strain required to achieve maximum piezoelectric coefficients decreases with increasing Sc content. This observation suggests that Sc doping can be leveraged to enhance the strain-induced piezoelectric response, potentially leading to improved performance in neuromorphic devices that rely on piezoelectric actuation or sensing.<br/>By combining compositional tuning and strain engineering, this study provides valuable insights into the design and optimization of high-performance piezoelectric materials based on Al_1-<i>x</i>Sc_<i>x</i>N for neuromorphic computing applications. The ability to tailor the piezoelectric properties through controlled doping and strain engineering opens up new avenues for developing adaptive bio-interfaces and bioinspired information processing systems.