Design and discovery of new piezoelectric materials using density functional theory

Abstract

Piezoelectric materials find applications in microelectromechanical systems (MEMS), such as surface acoustic wave (SAW) resonators, radio frequency (RF) filters, resonators, and energy harvesters. Using density functional theory calculations, the present study illustrates the influence of alloying and co-alloying with different nitrides on piezoelectric and mechanical properties of an existing piezoelectric material such as aluminum nitride (AlN). Besides improving the performance of existing piezoelectric material, a high-throughput screening method is used to discover new piezoelectric materials. AlN has several beneficial properties such as high temperature stability, low dielectric permittivity, high hardness, large stiffness constant, high sound velocity, and complementary metal-oxide-semiconductor (CMOS) compatibility. This makes it widely accepted material in RF and resonant devices. However, it remains a challenge to enhance the piezoelectric modulus of AlN. The first part of this thesis establishes that the piezoelectric modulus of AlN could be improved by alloying with rocksalt transition metal nitrides such as scandium nitride (ScN), yttrium nitride (YN), and chromium nitride (CrN). As the content of the rocksalt end member in the alloy increases, the accompanying structural frustration enables a greater piezoelectric response. This structural frustration is also accompanied by thermodynamic driving forces for phase separation which, with increased alloy concentration, lead to the destruction of the piezoelectric response upon transition to the (centrosymmetric, cubic) rocksalt structure. Thus, it becomes necessary to identify suitable alloying elements that may yield highest piezoelectric response with minimal alloying additions. The study reveals that the alloying with CrN would lead to the lowest transition composition (occurring at approximately 25\% CrN concentration) between the wurtzite and rocksalt structures, thereby allowing piezoelectric enhancements at alloying levels that are easier to stabilize during the synthesis. The present study indicates that, for 25\% CrN alloying, the piezoelectric modulus is about 4 times larger than that of pure AlN. Thus, it is proposed to use Cr$_x$Al$_{1-x}$N as a suitable replacement for AlN. One of the adverse effects of transition metal nitride alloying for improvement of the piezoelectric response is accompanied by the softening of AlN lattice. This would make the material less desirable for resonant applications. Our subsequent research establishes that co-alloying with YN and BN would enable the most superior combination of piezoelectric and mechanical properties of AlN. The rest of the thesis describes our efforts for identifying new piezoelectric materials. This requires high-throughput screening of inorganic materials for their piezoelectric properties. Our quest to search for new piezoelectric compounds is motivated by the notion that soft materials have the opportunity to exhibit large piezoelectric response (piezoelectric modulus \textbf{d}), and the knowledge that the van der Waals (vdW) layered materials are typically soft in the layer stacking direction. From a pool of 869 vdW structures, we have discovered 50 new compounds such as In$_2$Te$_5$, GeTe, and CuVO$_3$ having piezoelectric response larger than AlN. Remarkably, 70\% of our discovered piezoelectric candidates do not contain any toxic elements like Pb. Our analysis reveals that the large components of $\mathbf{d}$ always couple with the deformations (shearing or axial) of van der Waals ``gaps'' between the layers. This research establishes a wide scope of synthesizing vdW materials for applications demanding high piezoelectric modulus.