INTerface reactivity, microstructure and stress Evolution during thin film GRowth: multi-scALe modelling and experimental validation – INTEGRAL

The elementary mechanisms of adsorption and diffusion, at the atomic level, as well as the corresponding energy barrier heights will be identified using both ab initio calculations of the DFT type and molecular dynamics. These events will then be incorporated into a kinetic Monte Carlo (kMC) code, so as to reproduce realistic growth conditions for magnetron sputtering (energy distribution and angular distribution of incident particles). The simulation code thus developed will be able to combine different length and time scales and predict the evolution of physical quantities, such as roughness, density and / or morphology of thin films, with the ultimate aim of treating process of generation and relaxation of intrinsic constraints in a single multi-physics simulation code. The early stages of growth are of prime importance for the subsequent development of the microstructure; thus, the case of metals with high mobility (Cu) and that of metals with low mobility (Mo) will be studied in parallel, in order to compare them with experimental observations showing different growth modes adopted, 3D vs 2D, respectively. The chemical reactivity at the interface between metal and crystalline silicon will be studied in order to better understand the mechanisms of formation of interface silicides. Finally, in the case of polycrystalline growth, the formation and evolution of grain boundaries will be approached, in relation to the diffusion of point defects. The simulation approach will bring together the rigid lattice kMC and off-lattice kMC models to address defect creation, interfacial chemical mixing and grain boundary formation during the growth of polycrystalline films. The modeling of the intrinsic stress will be approached using the kinetic activation-relaxation technique (k-ART), a self-learning off-grid kMC algorithm, in collaboration with the group of Professor Mousseau at the University of Montreal. Experimental data will be obtained from magnetron sputtering growth experiments using a unique palette of real-time and in situ growth studies, but also using Synchrotron facilities, allowing structural analysis in situ, by diffraction. or small-angle X-ray scattering. The strength of the project lies in the direct confrontation of the results of numerical modeling with the experimental data obtained on the structural, electrical and optical properties of thin films. The ultimate goal of the project is to assess the intrinsic stresses and to simulate their evolution during growth under energetic conditions, as such a study has not been successfully completed so far.

- Development of a 3D kinetic Monte Carlo code devoted to magnetron sputter-deposition of copper films (WP1); a manuscript under preparation
- In the framework of the 17th “Journée de la Matière Condensée - JMC17” the PI co-organized the mini-symposium “CPR24 Growth of thin films: multi-level modeling, experimental elaboration and characterization”, which brought together more than 30 participants in each session and 12 oral contributions.

- Obtaining of funding from the French government program “Investissements d’Avenir” (EUR INTREE, reference ANR-18-EURE-0010) to support a Master 2 internship.

The k-ART software of Mousseau’s team (Partner in the present project) has been recently used to study the trapping of vacancies in the film as a way to understand the stress. In the present work, stress effects will be incorporated into the kMC model in a twofold manner:
• by relating the number and type of defects obtained from kMC simulations to the macroscopic stress, using extended analytical model or kMC approach
• by modifying dynamically the energy barrier that control the diffusion of atoms into and out of the GB due to presence of stress using k-ART.
Boateng et al. have reported kMC simulations in which the misfit strain was taken into account to reproduce the heteroepitaxial growth of strained layers, but they did not address the intrinsic stress that develops during polycrystalline film growth. Polycrystalline thin films are more difficult to simulate because of the difficulty inherent to GB formation. GB formation has been already included into kMC simulations by the group of Prof. Huang at Northeastern Univ. but without considering the stress effects .Contacts have been established with the above-mentioned group, their insight has been offered to the group and their valuable advice will be available throughout the INTEGRAL project.

International Communications
1. C. Mastail, C. Furgeaud, F. Nita, A. Michel, G. Abadias, “A Multiscale Modelling Of Thin Film Growth: Application To Sputterdeposited Cu Films” Plasma thin film international union meeting – PLATHINIUM (Virtual), 13-17 September 2021
2. G. Abadias, C. Mastail, R. Mareus, F. Nita, C. Furgeaud, A. Michel, “Computational Modeling of 3D Thin Film Growth Morphology: Influence of Angular and Energy Distribution of Particle Flux”, Invited Talk, 47th International Conference on Metallurgical Coatings and Thin Films - ICMCTF(virtual), 26-30 Avril 2021 (invited talk))


National Communication 1. F. Nita, C. Furgeaud, C. Mastail, A. Michel, G. Abadias « Modélisation par Monte Carlo cinétique de la croissance de films minces de Cu dans des conditions énergétiques » 17èmes journées de la matière condensée –JMC17(virtual) 24-28 août 2020 (oral)

Organisation of the mini-symposium « CPR24 Croissance des films minces: modélisation multi-niveaux, élaboration et caractérisation expérimentales », in the framework of the 17th « Journée de la Matière Condensée - JMC17» which brought together more than 30 participants in each session and 12 oral contributions.

The “INTEGRAL” project aims to develop a robust, reliable and realistic multiscale computational modelling of metallic thin film growth when using physical vapor deposition, and in particular under energetic conditions such as sputtering deposition scales with the ultimate goal to study stress generation and relaxation processes. Elementary mechanisms at the atomistic level, adsorption and diffusion, and the corresponding energy barrier heights will be identified using both ab initio DFT and molecular dynamics calculations. These events will then be implemented in a kinetic Monte Carlo (kMC) code, so as to depict realistic growth conditions in magnetron sputtering (energy and angular distribution of incoming particles). The resulting simulation code will be capable of seamlessly bridging length and time scales, and predicting evolution of physical quantities, roughness, density and/or morphology, with the ultimate goal to address intrinsic stress generation and relaxation processes into a single, multi-physics simulation package.
The initial growth stages are of prime importance for subsequent microstructure development, therefore the cases regarding high mobility metals (as Cu) and low mobility metals (as Mo) will be studied in parallel, as experimental observations point to different adopted growth modes, 3D vs 2D, respectively. Chemical reactivity at the interface between metal and crystalline silicon will be investigated to gain insight in the mechanisms of silicide formation. In a final step, polycrystalline growth and grain boundary formation will be addressed, highlighting the relation between grain boundary evolution and diffusion of point defects.
The computational approach will encompass both on-lattice and off-lattice kMC models to resolve the interdependent issues of defect creation, chemical intermixing and grain-boundary formation during polycrystalline film growth. Stress modelling will be addressed using kinetic activation-relaxation technique (k-ART), an off-lattice self-learning kMC algorithm, in collaboration with the group of Prof. Mousseau at Montreal University. Data collection from growth experiments for in situ structural analysis will be conducted within the Pprime laboratory as well as externally, using Synchrotron facilities. The strength of the INTEGRAL project is the direct experimental validation on structural, electrical and optical properties of thin films, using a unique palette of in situ and real-time studies during growth that is available to the research group.
The ultimate purpose of the project is to simulate the stress build-up during growth under energetic conditions, as this has yet to be achieved. The fundamental knowledge obtained during the INTEGRAL implementation will provide a more clear view of the relationship between the evolution of the film microstructure (grain size and texture) and its properties (stress state, defect density, mechanical attributes). It will make accessible new paths for fundamental and technological concept design and implementation, and it will greatly reduce the time required from concept to industrial product.