|
 |
|
|
 |
 |
|
 |
|
 |
|
 |
|
 |
|
|
LargeSys
|
|
|
|
|
|
|
||||||||||||||||||
|
LargeSys |
||||||||||||||||||||||
|
|
|
|||||||||||||||||||||
|
|
LARGE SCALE ATOMISTIC MODELING FOR MATERIALS ENGINEERING AND APPLIED CHEMISTRY (LargeSys)
 Giorgio Pastore Luciano Colombo Alessandro De Vita Paola Gallo Mauro Rovere Enrico Smargiassi Laura Zoppi Furio ErcolessiMain research lines:
The main aim of this activity is the development of concepts, simulation schemes and computational tools suitable for large systems and their applications to relevant Material Science and Applied Chemistry problems. Atomistic modeling of physical systems at atomic scale is a well established field of research in condensed matter and materials physics. In recent years, progress in the development of quantum mechanical simulation methods, proposals of improved algorithms and the huge enhancement of the available computing power, have allowed to describe complex condensed systems from an ab-initio perspective. In spite of these recent advances, a fully atomistic description of large scale processes is not yet feasible using purely first-principle techniques. Modeling real materials is characterized by a big range of time and length scales, spanning many orders of magnitudes, from the atomistic to the laboratory scale. These systems require accurate and effective techniques to transfer the ab-initio description of the atomic components into the meso/macroscopic properties of complex assemblies. There is an evident need for novel ways of tuning and validating empirical potentials as well as effective techniques to incorporate accurate quantum information into large scale simulations. Such need is widespread over research areas including condensed matter physics, materials engineering and applied chemistry. The main focus of the present activity will be on themes in this area. In particular, effort will be devoted to develop computational methods to study the catastrophic failure of brittle materials, the creation, interaction and annihilation of materials microstructures under various external conditions, the conditions of formation and stability of colloidal systems, the solvation and transport properties of polar solutions, the bottom-up fabrication of supramolecular assemblies. One explicit unifying approach and a methodological goal of the present activity is the exploitation of synergies between different but related computational techniques for studying systems reaching up to the micrometric size scale. The proposed CRS can provide an ideal environment for this, by favoring close interaction between the different research lines of the activity and an effective interchange of expertise with the personnel of related activities on ab-initio simulations, high pressure systems, low dimensional, biological systems and software optimization. 1 ATOMISTIC SIMULATION OF MECHANICAL PROPERTIES OF MATERIALS The main focus of the present line will be on the role of microstructure evolution as the crucial element leading to the macroscopic mechanical response. By microstructure evolution we mean the competing creation, interaction and annihilation of dislocations, grain boundaries, phase boundaries, precipitates, microcracks, microvoids, and point defects under the various external conditions of temperature, pressure, and stress. As we commented above, the attempt at understanding such an enormously complex interplay of phenomena requires the convergence of several theoretical and computational methods capable of spanning disparate length and time scales. This concept amounts to say that a unified multiscale approach is indeed needed for any further fundamental improvement of our basic understanding of mechanical properties of materials. A multiscale approach should be able to manage a given problem on all the relevant time/length scales either by hyerarchically combine or by integrate continuum (finite elements) and atomistic (molecular dynamics) methods[1,2]. Within the present line, we will especially focus on the atomistic modeling through large-scale molecular dynamics simulations, based either on model-potential or tight-binding force models. However, our ultimate goal will consist in rooting the engineering view of a given material (i.e. an elasto-plastic continuum) into a physical picture of the same material as an assembly of interacting atoms. As far as applications are concerned, our focus will be on the microstructrure evolution which naturally identifies the mesoscopic scale (i.e., approximately the micron length range) as the central ground where the atomistic methods can provide valuable information. The classes of materials of interest will range from metals and metallic alloys to semiconductors and - in somecase - thermodynamically stable amorphous and nano-/micro-structured phases. As far as methodological issues are concerned, we will devote specific efforts, in collaboration with the other lines, in developing efficient numerical codes for really large-scale simulations. Furthermore, we will work to derive a reliable coupled classical/quantum atomistic simulation scheme. The key idea relies on the set-up of a full atomistic description of the relevant phenomena where the core region (i.e. the volume interested by quantum features affecting interatomic bonds) is described at the semi-empirical tight-binding molecular dynamics level[3], while the proper embedding volume (i.e. the region surrounding the core volume) is treated by means of model-potential MD simulations. According to the above scheme, we should be able to describe fully atomistically any selected event spanning several lenght scales, from 1 nm (core region) to 100 nm (embedding region). In other words, no need to match the elastic continuum regime will be further required. 2 MODELING MOLECULAR ASSEMBLY AND LARGE LOW-DIMENSIONAL STRUCTURES We plan to investigate 2D molecular self-assembly by using empirical potentials developed on the basis of first principles calculations and validated by direct comparison of the predicted structures with STM images from controlled experiments [4,5]. We also plan to study mechanical properties of brittle solids[6,7] and other chemical processes relevant for the characterization and engineering of low dimensional systems such as interfaces and surfaces[8]. In particular, we plan to keep developing and applying a novel hybrid (classical and quantum) molecular dynamics simulation technique we recently proposed [9], based on a flexible embedding scheme in which the parameters of an empirical interatomic potential are fitted on the fly , where and when needed, to precise electronic-structure level information (tight binding, and first-principle) relevant to localized system's sub-portions. The scheme will be used at first to study the mechanical response of materials, investigating phenomena involving low-dimensional chemically active regions. These include e.g. crack propagation [6] and dislocation mobility [7] in Si. Later, we plan to investigate processes where the mechanical behavior results from a higher chemical complexity (e.g., tribo-chemical phenomena), of direct interest for the engineering and characterization of low dimensional structures such as e.g., nano-structured surfaces[8]. We note that the embedding concept is a general and effective way to distribute the computational effort, and is thus of obiquitous potential use in all fields of atomistic modeling of materials, e.g., biological molecules with specific chemically active sites. The development of collaborations and cross-disciplinary applications with a number other activities of the Center is therefore expected. 3 SIMULATION MODELS FOR FLUID PHASES OF MATERIALS Atomistic simulation studies for bulk and confined fluid phases are crucial for the understanding of complex problems in applied soft matter physics and are becoming current tools in the chemical engineering community[10,11]. The present line will focus on three main topics: a) modeling the interactions and investigation of the structural and dynamical properties of close-to-critical dilute aqueous solutions; b) computer simulation studies of the phase stability of mixtures of simple and complex fluids; c) computer simulation studies of confined fluids. In recent years, there has been an increasing interest of the chemical industry in close-to-critical aqueous solutions as extraction media or as special reaction media[10]. Moreover, from the physical point of view, these systems are very interesting. Aqueous solutions play also a prominent role for biochemical and applied chemistry research[11]. Since full ab-initio studies for large systems as those required by the study of diluted solutions are presently not feasible, a key ingredient is the accurate modeling of the interactions. Polarizable models have been developed in recent years, but the field is still far from being in a mature state. We plan to devote a serious effort toward the development of new dissociable-polarizable potentials for water (and other polarizable solvents) which may embody the ab-initio level of description in the interaction parameters of polarizable ions. Such topics is obviously connected with theme b) in the case of the determination of solubility curves. More in general, simulation methods for predicting phase diagrams in multicomponent systems have been witnessing a great progress in the last decade. Applications to complex problems are becoming frequent in the current literature even though further work is required to improve the robustness and efficiency of the simulation algorithms. We plan to develop efficient simulation tools for chemical engineering applications. Interest in computer simulations of confined fluids, like thin films of water or other hydrogen-bond liquids confined in zeolites, biological hydrogels or vescicles, is rapidly growing because of the close connection with many relevant areas of biology, chemical engineering and geology[12]. The specific differences in the behaviour of confined fluids are, among others, due to different interactions with the substrate, the size distribution and geometry of the confining region, the size and polydispersity of the particles composing the liquid. In order to make contact with experiments and to make the computer simulation really predictive in this field, we intend to explore the influence of the pore connectivity on the phase stability of the confined fluid, use realistic interaction models, geometries and compositions and we want to take in consideration the dynamics of the substrate, when it can be relevant in determining the phenomena[13], like in the case water at contact with biological matter. BIBLIOGRAPHY [1] Advances in Materials Theory and Modelling: Bridging Over Multiple-Length and Time Scales L. Colombo et al. Eds., Mat. Res. Soc. Symp. Proc., vol. 677 (2001) [2] Theory and simulation of fracture R. Selinger and D. Farkas Eds. Mat. Res. Society Bullettin (May 2000) [3] Tight Binding Molecular Dynamics Special issue ed. by L.Colombo, Comp. Mat. Sci., vol. 12 (1998) [4] M.Bohringer et al., Phys.Rev.Lett. 83, 324 (1999). [5] J.Weckesser et al., Phys. Rev. Lett. 87, 096101 (2001). Symposium. Mater. Res. Soc., (Warrendale, PA, USA) 473 (1998). [5] R.Perez and P.Gumbsch, Phys. Rev. Lett.84, 5347,(2000). [7] G.Csanyi et al., Phys. Rev. Lett. 80, 3984 (1998). [8] S.Schintke et al., Phys. Rev. Lett., in press [9] A. De Vita and R. Car, Tight-Binding Approach to Computational Materials Science [10] C.A. Eckert et al., Nature 383, 313 (1996) [11] A.A.Chialvo, P.T.Cummings, Adv. Chem. Phys. 109, 115 (1999) [12] T.M. Truskett and P.G. Debenedetti J. Chem. Phys. 114, 2401 (2001) [13] P. Gallo et al. Phys. Rev. Lett. 85, 4317 (2000) |
|
|
|
||||||||||||||||||||||||||
|
|
|
INFM crs DEMOCRITOS
SISSA - Via Beirut 2-4 34014 Grignano, Trieste, ITALY Tel. +39 0403787443 Fax. +39 0403787528 |
|
|
|
webmaster: webadmin@democritos.it
|
![[X]](/immagini/busta.gif){kind=small}
