Abstract: In this paper, an overview is given of modeling investigations carried out in our research group for a better understanding of plasmas used for medical, environmental and nano applications. The focus is both on modeling the plasma chemistry and the plasma-surface interactions. The plasma chemistry provides the densities and fluxes of the important plasma species. This information can be used as input when modeling the plasma-surface interactions. The combination of plasma simulations and plasma – surface interaction simulations provides a more comprehensive understanding of the underlying processes for these applications.

Abstract: The exact structure and properties of the Si vertical bar SiO2 interface are very important in microelectronics and photovoltaic devices such as metal-oxide-semiconductor field-effect transistors (MOSFETs) and solar cells. Whereas Si vertical bar SiO2 structures are traditionally produced by thermal oxidation, hyperthermal oxidation shows a number of promising advantages. However, the Si vertical bar SiO2 interface induced in hyperthermal Si oxidation has not been properly investigated yet. Therefore, in this work, the interface morphology and interfacial stresses during hyperthermal oxidation at room temperature are studied using reactive molecular dynamics simulations based on the ReaxFF potential. Interface thickness and roughness, as well as the bond length and bond angle distributions in the interface are discussed and compared with other models developed for the interfaces induced by traditional thermal oxidation. The formation of a compressive stress is observed. This compressive stress, which at the interface amounts about 2 GPa, significantly slows down the inward silica growth. This value is close to the experimental value in the Si vertical bar SiO2 interface obtained in traditional thermal oxidation.

Abstract: The adsorption probability and reaction behavior of CHx plasma species on various nickel catalyst surfaces is investigated by means of reactive molecular dynamics (MD) simulations using the ReaxFF potential. Such catalysts are used in the reforming of hydrocarbons and in the growth of carbon nanotubes, and further insight in the underlying mechanisms of these processes is needed to increase their applicability. Single and consecutive impacts of CHx radicals (x={1,2,3}) were performed on four different Ni surfaces, at a temperature of 400 K. The adsorption probability is shown to be related to the number of free electrons, i.e. a higher number leads to more adsorptions, and the steric hindrance caused by the hydrogen atoms bonded to the impacting CHx species. Furthermore, some of the CH bonds break after adsorption, which generally leads to diffusion of the hydrogen atom over the surface. Additionally, these adsorbed H-atoms can be used in reactions to form new molecules, such as CH4 and C2Hx, although this is dependent on the precise morphology of the surface. New molecules are also formed by subtraction of H-atoms from adsorbed radicals, leading to occasional formation of H2 and C2Hx molecules.

Abstract: Monte Carlo (MC) methods have a long-standing history as partners of molecular dynamics (MD) to simulate the evolution of materials at the atomic scale. Among these techniques, the uniform-acceptance force-bias Monte Carlo (UFMC) method [ G. Dereli Mol. Simul. 8 351 (1992)] has recently attracted attention [ M. Timonova et al. Phys. Rev. B 81 144107 (2010)] thanks to its apparent capacity of being able to simulate physical processes in a reduced number of iterations compared to classical MD methods. The origin of this efficiency remains, however, unclear. In this work we derive a UFMC method starting from basic thermodynamic principles, which leads to an intuitive and unambiguous formalism. The approach includes a statistically relevant time step per Monte Carlo iteration, showing a significant speed-up compared to MD simulations. This time-stamped force-bias Monte Carlo (tfMC) formalism is tested on both simple one-dimensional and three-dimensional systems. Both test-cases give excellent results in agreement with analytical solutions and literature reports. The inclusion of a time scale, the simplicity of the method, and the enhancement of the time step compared to classical MD methods make this method very appealing for studying the dynamics of many-particle systems.

Abstract: The growth mechanism and chirality formation of a single-walled carbon nanotube (SWNT) on a surface-bound nickel nanocluster are investigated by hybrid reactive molecular dynamics/force-biased Monte Carlo simulations. The validity of the interatomic potential used, the so-called ReaxFF potential, for simulating catalytic SWNT growth is demonstrated. The SWNT growth process was found to be in agreement with previous studies and observed to proceed through a number of distinct steps, viz., the dissolution of carbon in the metallic particle, the surface segregation of carbon with the formation of aggregated carbon clusters on the surface, the formation of graphitic islands that grow into SWNT caps, and finally continued growth of the SWNT. Moreover, it is clearly illustrated in the present study that during the growth process, the carbon network is continuously restructured by a metal-mediated process, thereby healing many topological defects. It is also found that a cap can nucleate and disappear again, which was not observed in previous simulations. Encapsulation of the nanoparticle is observed to be prevented by the carbon network migrating as a whole over the cluster surface. Finally, for the first time, the chirality of the growing SWNT cap is observed to change from (11,0) over (9,3) to (7,7). It is demonstrated that this change in chirality is due to the metal-mediated restructuring process.

Abstract: Using self-consistent KohnSham density-functional theory molecular dynamics simulations, we demonstrate the theoretical possibility to synthesize NiC60, the incarfullerene Ni@C60 and the heterofullerene C59Ni in an ion implantation setup. The corresponding formation mechanisms of all three complexes are elucidated as a function of the ion implantation energy and impact location, suggesting possible routes for selectively synthesizing these complexes.

Abstract: In this paper, the effects of prolonged bias-enhanced nucleation (prolonged BEN) on the growth mechanisms of diamond are investigated by molecular dynamics (MD) and combined MD-Metropolis Monte Carlo (MD-MMC) simulations. First, cumulative impacts of CxHy+ and Hx+ on an a-C:H/nanodiamond composite were simulated; second, nonconsecutive impacts of the dominant ions were simulated in order to understand the observed phenomena in more detail. As stated in the existing literature, the growth of diamond structures during prolonged BEN is a process that takes place below the surface of the growing film. The investigation of the penetration behavior of CxHy+ and Hx+ species shows that the carbon-containing ions remain trapped within this amorphous phase where they dominate mechanisms like precipitation of sp3 carbon clusters. The H+ ions, however, penetrate into the crystalline phase at high bias voltages (>100 V), destroying the perfect diamond structure. The experimentally measured reduction of grain sizes at high bias voltage, reported in the literature, might thus be related to penetrating H+ ions. Furthermore, the CxHy+ ions are found to be the most efficient sputtering agents, preventing the build up of defective material.

Abstract: Metal-catalyzed growth mechanisms of carbon nanotubes (CNTs) were studied by hybrid molecular dynamics−Monte Carlo simulations using a recently developed ReaxFF reactive force field. Using this novel approach, including relaxation effects, a CNT with definable chirality is obtained, and a step-by-step atomistic description of the nucleation process is presented. Both root and tip growth mechanisms are observed. The importance of the relaxation of the network is highlighted by the observed healing of defects.

Abstract: In this paper, we present the results of combined molecular dynamics−Metropolis Monte Carlo (MD-MMC) simulations of hydrocarbon species at flat diamond (100)2 × 1 and (111)1 × 1 surfaces. The investigated species are considered to be the most important growth species for (ultra)nanocrystalline diamond ((U)NCD) growth. When applying the MMC algorithm to stuck species at monoradical sites, bonding changes are only seen for CH2. The sequence of the bond breaking and formation as put forward by the MMC simulations mimics the insertion of CH2 into a surface dimer as proposed in the standard growth model of diamond. For hydrocarbon species attached to two adjacent radical (biradical) sites, the MMC simulations give rise to significant changes in the bonding structure. For UNCD, the combinations of C3 and C3H2, and C3 and C4H2 (at diamond (100)2 × 1) and C and C2H2 (at diamond (111)1 × 1) are the most successful in nucleating new crystal layers. For NCD, the following combinations pursue the diamond structure the best: C2H2 and C3H2 (at diamond (100)2 × 1) and CH2 and C2H2 (at diamond (111)1 × 1). The different behaviors of the hydrocarbon species at the two diamond surfaces are related to the different sterical hindrances at the diamond surfaces.

Abstract: Molecular dynamics simulations using improved TersoffBrenner potential parameters were performed to investigate Cl+ etching of a {2×1} reconstructed Si(100) surface. Steady-state Si etching accompanying the Cl coverage of the surface is observed. Furthermore, a steady-state chlorinated reaction layer is formed. The thickness of this reaction layer is found to increase with increasing energy. The stoichiometry of SiClx species in the reaction layer is found to be SiCl:SiCl2:SiCl3 = 1.0:0.14:0.008 at 50 eV. These results are in excellent agreement with available experimental data. While elemental Si products are created by physical sputtering, most SiClx (0<x<4) etch products are produced by chemical-enhanced physical sputtering.

Abstract: In this paper, an overview is given of different modeling approaches used for describing gas discharge plasmas, as well as plasma-surface interactions. A fluid model is illustrated for describing the detailed plasma chemistry in capacitively coupled rf discharges. The strengths and limitations of Monte Carlo simulations and of a particle-in-cell-Monte Carlo collisions model are explained for a magnetron discharge, whereas the capabilities of a hybrid Monte Carlo-fluid approach are illustrated for a direct current glow discharge used for spectrochemical analysis of materials. Finally, some examples of molecular dynamics simulations, for the purpose of plasma-deposition, are given.

Abstract: In this paper, a new implementation of the Metropolis Monte Carlo (MMC) algorithm is presented. When combining the MMC model with a molecular dynamics (MD) code, crystal growth by plasma-enhanced chemical vapor deposition can be simulated. As the MD part simulates impacts of growth species onto the surface on a time scale of picoseconds, the MMC algorithm simulates the slower adatom surface processes. The implementation includes a criterion for the selection of atoms that are allowed to be displaced during the simulation, and a criterion of after how many MMC cycles the simulation is stopped. We performed combined MD-MMC simulations for hydrocarbon species that are important for the growth of ultrananocrystalline diamond (UNCD) films at partially hydrogenated diamond surfaces, since this implementation is part of a study of the growth mechanisms of (ultra)nanocrystalline diamond films. Exemplary for adatom arrangements during the growth of UNCD, the adatom surface behavior of C and C2H2 at diamond (111)1 × 1, C and C4H2 at diamond (111)1 × 1 and C3 at diamond (100)2 × 1 has been investigated. For all cases, the diamond crystal structure is pursued under the influence of MMC simulation. Additional longer time-scale MD simulations put forward very similar structures, verifying the MMC algorithm. Nevertheless, the MMC simulation time is typically one order of magnitude shorter than the MD simulation time.

Abstract: This paper outlines two different numerical simulation approaches, carried out by our group, used for describing hydrocarbon plasmas in their applications for either nanoparticle formation in the plasma or the growth of nanostructured thin films, such as nanocrystalline diamond (NCD). A plasma model based on the fluid approach is utilized to study the initial mechanisms giving rise to nanoparticle formation in an acetylene plasma. The growth of NCD is investigated by molecular dynamics simulations, describing the interaction of the hydrocarbon species with a substrate.

Abstract: The reaction behavior of species that may affect the growth of ultrananocrystal line and nanocrystalline diamond ((U)NCD) films is investigated by means of molecular dynamics simulations. Impacts of CHx (x = 0 – 4), C2Hx (x=0-6), C3Hx (x=0-2), C4Hx (x = 0 – 2), H, and H-2 on clean and hydrogenated diamond (100)2 x 1 and (111) 1 x 1 surfaces at two different substrate temperatures are simulated. We find that the different bonding structures of the two surfaces cause different temperature effects on the sticking efficiency. These results predict a temperature-dependent ratio of diamond (100) and (111) growth. Furthermore, predictions of which are the most important hydrocarbon species for (U)NCD growth are made.

Abstract: This paper presents the computational study of the metal-organic (MO) CVD of titanium dioxide (TiO2) films grown using titanium tetraisopropoxide (TTIP) as a precursor and nitrogen as a carrier gas. The TiO2 films are deposited under atmospheric pressure. The effects of the precursor concentration, the substrate temperature, and the hydrolysis reaction on the deposition process are investigated. It is found that hydrolysis of the TTIP decreases the onset temperature of the gas-phase thermal decomposition, and that the deposition rate increases with the precursor concentration and with the decrease of substrate temperature. Concerning the mechanism responsible for the film growth, the model shows that at the lowest precursor concentration, the direct adsorption of the precursor is dominant, while at higher precursor concentrations, the monomer deposition becomes more important.