Data-driven prediction of nonequilibrium chemistry in plasma enhanced atomic layer etching of silicon nitride

In the semiconductor industry, plasma enhanced atomic layer etching (PEALE) has attracted significant attention due to its potential for high quality etch in nanometer-scale and high-aspect ratio features, where atomic layer precision is critical. PEALE consists of two cyclic steps: 1) treatment of a surface with some precursor, and 2) ion bombardment-assisted removal of material. However, these processes can be challenging to study both experimentally and computationally; they involve timescales too short to be experimentally observable, but also involve complex dynamics that can be difficult to comprehensively model in simulations. Thus, fundamental mechanisms of PEALE are poorly understood. In this dissertation, a computational framework based on self-consistent-charge density functional tight binding (SCC-DFTB) and density functional theory (DFT) have been developed and applied to the case of PEALE of silicon nitride (SiN) with hydrofluorocarbons (HFCs). This framework includes an extension to the pbc-0-3 parameter set as well as tooling for automated execution and analysis of parallel TBMD simulations. We begin by investigating the possibility of thermal reactions between HFCs and SiN substrates, both under ambient conditions as well as under the influence of bombardment-induced local heating. Through this, we show that thermal reactions are unlikely to be the primary contributor to achieving PEALE of SiN with HFCs. Then, the structural and compositional evolution of SiN under ion bombardment without HFCs are assessed. Following this, we identify key mechanisms that govern bombardment-induced decomposition of physisorbed HFCs, with emphasis placed on analyzing the assisting role of the supporting SiN substrate. The surface reactions that can occur during bombardment, as well as after, have been described and used to predict the formation pathways of the quasi-equilibrium films that are generated during PEALE. Finally, the implications of these reactions for enhancing SiN etch through chemical and physical modification are discussed, especially the complexity of non-equilibrium etch products that play a key role. In this work, new atomic-level insight is gained into the mechanisms of SiN PEALE, with implications for process design in the semiconductor industry. Furthermore, this new computational approach is a general strategy that can be applied to many other plasma-driven processes, and for other non equilibrium chemistry as a whole.