Search

Yade is an extensible open-source framework for discrete numerical models, focused on the Discrete Element Method. The computation parts are written in c++ using a flexible object model and allowing independent implementation of new algorithms and interfaces. Python is used for rapid and concise scene construction, simulation control, postprocessing and debugging. Yade is located at yade-dem.org, which contains this documentation. Development is kindly hosted on launchpad and GitLab ; they are used for source code, bug tracking and source downloads and more. Building, regression tests and packages distribution are hosted on servers of the Grenoble Geomechanics group at Laboratoire 3SR, UMS Gricad and Gda\'nsk University of Technology. Yade supports high precision calculations and Python 3. The development branch is on GitLab.

This paper presents the design of the MuPIF distributed, multi-physics simulation platform and its performance in the context of the H2020 Composelector project. The description of MuPIF's model and data interfaces provides implementation and operational details that illustrate how MuPIF is a highly versatile and robust tool for various engineering applications. Its distributed design and the integration of models in a workflow allows MuPIF to simulate real industrial tasks. The design of a composite airplane frame for the Composelector project, focusing on innovative design of composite materials and structures, demonstrates MuPIF's capabilities and ability to be integrated into Business Decision Support Systems.

Scanning tunneling microscopy experiments show that the unstable growth morphology observed during molecular beam homoepitaxy on slightly vicinal Si(001) surfaces consists of straight step bunches. The instability occurs under step- flow growth conditions and vanishes both during low-temperature island growth and at high temperatures. An instability with the same characteristics is observed in a 2D Kinetic Monte Carlo model of growth with incorporated Si(001)- like diffusion anisotropy. This provides strong evidence that the diffusion anisotropy destabilizes growth on Si(001) and similar surfaces towards step bunching. This new instability mechanism is operational without any additional step edge barriers., Comment: 6 pages, 4 figures

The effects of adsorbates on nucleation and growth of two-dimensional islands is investigated by kinetic Monte Carlo simulations and rate equation theory. The variation of island morphology with adsorbate parameters is discussed and the temperature-dependence of island density in the case of immobile adsorbates is studied in detail. A set of rate equations for the description of nucleation in the presence of predeposited mobile and immobile adsorbates is developed., Comment: NATO ARW Colective surface diffusion coefficients under non-equilibrium conditions, October 2-6, 2000, Prague

The effect of impurities on epitaxial growth in the submonolayer regime is studied using kinetic Monte Carlo simulations of a two-species solid-on-solid growth model. Both species are mobile, and attractive interactions among adatoms and between adatoms and impurities are incorporated. Impurities can be codeposited with the growing material or predeposited prior to growth. The activated exchange of impurities and adatoms is identified as the key kinetic process in the formation of a growth morphology in which the impurities decorate the island edges. The dependence of the island density on flux and coverage is studied in detail. The impurities strongly increase the island density without appreciably changing its power-law dependence on flux, apart from a saturation of the flux dependence at high fluxes and low coverages. A simple analytic theory taking into account only the dependence of the adatom diffusion constant on impurity coverage is shown to provide semi-quantitative agreement with many features observed in the simulations., Comment: 11 pages, 11 Postscript figures, RevTex

We study the dynamics of island nucleation in the presence of adsorbates using kinetic Monte Carlo simulations of a two-species growth model. Adatoms (A-atoms) and impurities (B-atoms) are codeposited, diffuse and aggregate subject to attractive AA- and AB-interactions. Activated exchange of adatoms with impurities is identified as the key process to maintain decoration of island edges by impurities during growth. While the presence of impurities strongly increases the island density, a change in the scaling of island density with flux, predicted by a rate equation theory for attachment-limited growth [D. Kandel, Phys. Rev. Lett. 78, 499 (1997)], is not observed. We argue that, within the present model, even completely covered island edges do not provide efficient barriers to attachment., Comment: 7 pages, 2 postscript figures

Surface reconstructions can drastically modify growth kinetics during initial stages of epitaxial growth as well as during the process of surface equilibration after termination of growth. We investigate the effect of activation barriers hindering attachment of material to existing islands on the density and size distribution of islands in a model of homoepitaxial growth on Si(111)7x7 reconstructed surface. An unusual distribution of island sizes peaked around "magic" sizes and a steep dependence of the island density on the growth rate are observed. "Magic" islands (of a different shape as compared to those obtained during growth) are observed also during surface equilibration., Comment: 4 pages including 5 figures, REVTeX, submitted to Physical Review B

Computer simulations and scaling theory are used to investigate the damping of oscillations during epitaxial growth on high-symmetry surfaces. The crossover from smooth to rough growth takes place after the deposition of (D/F)^\delta monolayers, where D and F are the surface diffusion constant and the deposition rate, respectively, and the exponent \delta=2/3 on a two-dimensional surface. At the transition, layer-by-layer growth becomes desynchronized on distances larger than a layer coherence length proportional l^2, where l is a typical distance between two-dimensional islands in the submonolayer region of growth., Comment: to appear in The European Physical Journal B/Surfaces and Interfaces (5 pages, 4 figures)

Homoepitaxial growth is unstable towards the formation of pyramidal mounds when interlayer transport is reduced due to activation barriers to hopping at step edges. Simulations of a lattice model and a continuum equation show that a small amount of desorption dramatically speeds up the coarsening of the mound array, leading to coarsening exponents between 1/3 and 1/2. The underlying mechanism is the faster growth of larger mounds due to their lower evaporation rate., Comment: 4 pages, 4 PostScript figures

We propose deposition noise to be an important factor in unstable epitaxial growth of thin films. Our analysis yields a geometrical relation H=(RWL)^2 between the typical mound height W, mound size L, and the film thickness H. Simulations of realistic systems show that the parameter R is a characteristic of the growth conditions, and generally lies in the range 0.2-0.7. The constancy of R in late-stage coarsening yields a scaling relation between the coarsening exponent 1/z and the mound height exponent \beta which, in the case of saturated mound slope, gives \beta = 1/z = 1/4., Comment: 4 pages, RevTex Macros, 3 eps figures

The nucleation and growth of two--dimensional islands is studied with Monte Carlo simulations of a pair--bond solid--on--solid model of epitaxial growth. The conventional description of this problem in terms of a well--defined critical island size fails because no islands are absolutely stable against single atom detachment by thermal bond breaking. When two--bond scission is negligible, we find that the ratio of the dimer dissociation rate to the rate of adatom capture by dimers uniquely indexes both the island size distribution scaling function and the dependence of the island density on the flux and the substrate temperature. Effective pair-bond model parameters are found that yield excellent quantitative agreement with scaling functions measured for Fe/Fe(001)., Comment: 8 pages, Postscript files (the paper and Figs. 1-3), uuencoded, compressed and tarred. Surface Science Letters, in press.

We use analytic formulae obtained from a simple model of crystal growth by molecular--beam epitaxy to determine step--edge barriers to interlayer transport. The method is based on information about the surface morphology at the onset of nucleation on top of first--layer islands in the submonolayer coverage regime of homoepitaxial growth. The formulae are tested using kinetic Monte Carlo simulations of a solid--on--solid model and applied to estimate step--edge barriers from scanning--tunneling microscopy data on initial stages of Fe(001), Pt(111), and Ag(111) homoepitaxy., Comment: 4 pages, a Postscript file, uuencoded and compressed. Physical Review B, Rapid Communications, in press.

We report the first observation of re-entrant layer-by-layer etching based on {\it in situ\/} reflection high-energy electron-diffraction measurements. With AsBr$_3$ used to etch GaAs(001), sustained specular-beam intensity oscillations are seen at high substrate temperatures, a decaying intensity with no oscillations at intermediate temperatures, but oscillations reappearing at still lower temperatures. Simulations of an atomistic model for the etching kinetics reproduce the temperature ranges of these three regimes and support an interpretation of the origin of this phenomenon as the site-selectivity of the etching process combined with activation barriers to interlayer adatom migration., Comment: 11 pages, REVTeX 3.0. Physical Review Letters, in press.

We present results of numerical simulations of kinetic roughening for a growth model with surface diffusion (the Wolf-Villain model) in 3+1 and 4+1~dimensions using lattices of a linear size up to $L=64$ in 3+1~D and $L=32$ in 4+1~D. The effective exponents calculated both from the surface width and from the height--height correlation function are much larger than those expected based on results in lower dimensions, due to a growth instability which leads to the evolution of large mounded structures on the surface. An increase of the range for incorporation of a freshly deposited particle leads to a decrease of the roughness but does not suppress the instability., Comment: 8 pages, LaTeX 2.09, IC-DDV-93-004

Monte Carlo simulations of an atomistic solid-on-solid model are used to study the effect of lattice misfit on the distribution of two-dimensional islands sizes as a function of coverage $\Theta$ in the submonolayer aggregation regime of epitaxial growth. Misfit promotes the detachment of atoms from the perimeter of large pseudomorphic islands and thus favors their dissolution into smaller islands that relieve strain more efficiently. The number density of islands composed of $s$ atoms exhibits scaling in the form \mbox{$N_s(\Theta) \sim \Theta / \langle s \rangle^2 \, g(s/\langle s \rangle$)} where $\langle s \rangle$ is the average island size. Unlike the case of homoepitaxy, a rate equation theory based on this observation leads to qualitatively different behavior than observed in the simulations., Comment: 10 pages, LaTeX 2.09, IC-DDV-94-001

We investigate growth on vicinal surfaces by molecular beam epitaxy making use of a generalized Burton--Cabrera--Frank model. Our primary aim is to propose and implement a novel analytical program based on a perturbative solution of the non--linear equations describing the coupled adatom and dimer kinetics. These equations are considered as originating from a fully microscopic description that allows the step boundary conditions to be directly formulated in terms of the sticking coefficients at each step. As an example, we study the importance of diffusion barriers for adatoms hopping down descending steps (Schwoebel effect) during growth and post-growth equilibration of the surface., Comment: 16 pages, REVTeX 3.0, IC-DDV-94-003

A simple model of epitaxial growth proposed by Wolf and Villain is investigated using extensive computer simulations. We find an unexpectedly complex crossover behavior of the original model in both 1+1 and 2+1 dimensions. A crossover from the effective growth exponent $\beta_{\rm eff}\!\approx\!0.37$ to $\beta_{\rm eff}\!\approx\!0.33$ is observed in 1+1 dimensions, whereas additional crossovers, which we believe are to the scaling behavior of an Edwards--Wilkinson type, are observed in both 1+1 and 2+1 dimensions. Anomalous scaling due to power--law growth of the average step height is found in 1+1 D, and also at short time and length scales in 2+1~D. The roughness exponents $\zeta_{\rm eff}^{\rm c}$ obtained from the height--height correlation functions in 1+1~D ($\approx\!3/4$) and 2+1~D ($\approx\!2/3$) cannot be simultaneously explained by any of the continuum equations proposed so far to describe epitaxial growth., Comment: 11 pages, REVTeX 3.0, IC-DDV-93-006

We investigate the growth kinetics on vicinal GaAs(001) surfaces by making detailed comparisons between reflection high--energy electron--diffraction specular intensity measured near in--phase diffraction conditions and the surface step density obtained from simulations of a solid--on--solid model. Only by including a barrier to interlayer transport and a short--range incorporation process of freshly--deposited atoms can the simulations be brought into agreement with the measurements both during growth and during post--growth equilibration of the surface., Comment: 11 pages, REVTeX 3.0, IC-DDV-93-005