Three-dimensional kinetic Monte Carlo simulations of cubic transition metal nitride thin film growth

Abstract

A three-dimensional kinetic Monte Carlo (KMC) model has been developed and used to simulate the microstructure and growth morphology of cubic transition metal nitride (TMN) thin films deposited by reactive magnetron sputtering. Results are presented for the case of stoichiometric TiN, chosen as a representative TMN prototype. The model is based on a NaCl-type rigid lattice and includes deposition and diffusion events for both N and Ti species. It is capable of reproducing voids and overhangs, as well as surface faceting. Simulations were carried out assuming a uniform flux of incoming particles approaching the surface at normal incidence. The ballistic deposition model is parametrized with an interaction parameter r 0 that mimics the capture distance at which incoming particles may stick on the surface, equivalently to a surface trapping mechanism. Two diffusion models are implemented, based on the different ways to compute the site-dependent activation energy for hopping atoms. The influence of temperature (300-500 K), deposition flux (0.1-100 monolayers/s), and interaction parameter r 0 (1.5-6.0Å) on the obtained growth morphology are presented. Microstructures ranging from highly porous, [001]-oriented straight columns with smooth top surface to rough columns emerging with different crystallographic facets are reproduced, depending on kinetic restrictions, deposited energy (seemingly captured by r 0 ), and shadowing effect. The development of facets is a direct consequence of the diffusion model which includes an intrinsic (minimum energy-based) diffusion anisotropy, although no crystallographic diffusion anisotropy was explicitly taken into account at this stage. The time-dependent morphological evolution is analyzed quantitatively to extract the growth exponent β and roughness exponent α, as indicators of kinetic roughening behavior. For dense TiN films, values of α ≈ 0.7 and β = 0.24 are obtained in good agreement with existing experimental data. At this stage a single lattice is considered but the KMC model will be extended further to address more complex mechanisms, such as anisotropic surface diffusion and grain boundary migration at the origin of the competitive columnar growth observed in polycrystalline TiN-based films.

Effects of surface vibrations on interlayer mass transport: Ab initio molecular dynamics investigation of Ti adatom descent pathways and rates from TiN/TiN(001) islands

“…Molecular dynamics simulations inherently resolve the problems mentioned above by integrating Newton's equations of motion for each atom at any temperature of interest. CMD/AIMD reaction rates for adspecies intra-and interlayer migration [22,59,60,68,111] or desorption [63,112], as well as point defects in bulk [110,113], can be employed in kinetic Monte Carlo simulations [114] to efficiently probe the effects of precursors fluxes and ion-to-metal ratios on film growth modes.…”

Predicting structure zone diagrams for thin film synthesis by generative machine learning

“…The goal is to achieve a reliable prediction of complex, realistic microstructures based on given properties like composition and relevant process parameters. Microstructures can be predicted by simulations, e.g., kinetic Monte Carlo [29][30][31][32][33] or molecular dynamic simulation 34,35 , which depend on selection of model architectures, the selection of initial values and are computationally expensive. The interpretation of the overlap between simulation and experimental results remains to be performed by human assessment.…”

Изучение Процессов Структурообразования Керамических Покрытий Кинетическим Методом Монте-Карло

“…Однако с учетом большого количества элементов, параметров, а также в связи со сложностью изучения свойств наноструктурных материалов процесс разработки покрытий является трудоемкой задачей. Методы компьютерного моделирования могут существенно сократить не только время поиска, но и расширить понимание физических процессов роста покрытий [4][5][6].…”

“…Вычисление энергии диффузии каждого перехода происходило путем суммирования начальной энергии E 0 , определяемой температурой подложки и потенциалом смещения, и разницы между энергией взаимодействия атома с соседними в новом положении и текущем положении: E = E 0 − E. Такой подход позволяет реализовать анизотропность диффузии атомов. Для вычисления плазменного потока Ti и Cr использовались данные, полученные ранее численным моделированием в [5]. Определение потока атомов азота выполнялось в приближении постоянного равенства соотношения концентраций c Ti + c Cr = c N .…”