Thermodynamische Modelering van Chemische Reacties bij Materiaal Groei

Across the evolution of different technology nodes, the continuous reduction of the dimensions of the building blocks of CMOS has been pushing the physical limits of the active components in transistor and memory devices into numerous ends. Consequently, a straightforward linear scaling of the dimensions does not apply anymore to all the aspects of the active components. An educated selection of new materials with improved performances and high-quality interfaces is hence needed to sustain the development of innovative approaches. The introduction of these new materials leads to a source of development challenges in terms of growth and etching. In these uncharted waters, issues related to the identification and/or the design of chemical precursors compatible with growth reactors, while accounting for the silicon integration limits, is getting more and more difficult. For other material classes, it is also fundamental to identify and to understand the reactivity underlying their deposition process, their substrate dependency and how this can be used to engineer the deposition selectively. Similarly, the development of strategies to etch selectively multi-component materials, as typically used in complex magnetic metal layers or amorphous semiconductors, requires the identification of process window(s) and reactant(s) favoring the formation of volatile compounds such that the occurrence of an unwanted material (re)deposition is prevented. A large part of these challenges arises from our inability to forecast the occurrence of dominant chemical reactions in multi-component systems, where thousands to millions of reactions can potentially occur simultaneously, while the thermodynamics of these reactions is unknown for most of the components. This Ph.D. topic aims at solving this challenging issue by using the predictive power of first-principles simulations to build a thermodynamic reference database connecting the thermodynamic properties of new materials, their deposition/etching molecules and their associated by-products. This database will be then combined with numerical minimization methods to identify the most likely reactions to occur during material growth and etching. The insights gained will provide precious information in terms of precursor selection and process control such as concentration, temperature and pressure dependency. The outcome of the simulations will be used systematically to drive the development of new material and related processes and will be benchmarked against experimental measurements.