Gaz flow in the reaction chamber (source B. Caussat, ENSIACET)

The success of ALD processes mostly rely on the control of the chemical interactions between precursor molecules and surfaces. A precise atomic scale understanding of the elementary mechanisms involved during ALD growth thus constitutes a crucial point for the implementation and optimization of the overall deposition process; it should be noted that the ALD must lead to a high quality of the elaborated materials (conformal layers, purity, stoichiometry). Two decades of research in the field clearly shows that this understanding is supported by advanced means of in situ and ex situ characterization, in conjunction with the use of multi-scale modeling means. Mutiple questions can be addressed, choice of precursors, identification and quantification of growth mechanisms during the early stages of deposition and during steady state groth, modification / preparation of surfaces before deposition, to cite just a few. As a result, literature often shows a complex reality, leading to discrepancies with regard of ALD fundamental concepts: complexity of reaction mechanisms, incomplete dissociation, gas phase reactions, layer contamination, steric hindrance of grafted precursors, more complex intra-layer reactions, etc. In order to reduce these differences, to anticipate the choice of precursors, to evaluate the quality of the layers produced, their structuring at the atomic scale, but also to provide the elements of understanding and parameters for feature scale models, modeling and characterization at the nanoscale must be used in synergy. This is the aim of this GDR axis: to gather expertise at the national level in in situ and ex situ characterization (IR, XPS, LEIS, XRD …), modeling techniques at the atomic scale such as DFT techniques (for depicting elementary chemical reactions), kinetic Monte Carlo (mesoscopic scale and process simulation) and feature scale modeling. These tools must be able to answer the fundamental questions related to nanostructure / properties relationship and to push forward the use of ALD techniques.


Top view of HfO2 deposited by ALD onto Si : a kinetic Monte Carlo simulation (source A. Estève, LAAS-CNRS)
DFT cluster model of HfCl4 interacting with SiO2 and showing chlorine contamination (source A. Estève, LAAS-CNRS)