Materials chemomechanics at the atomic scale: modelling and experiments
CECAM-HQ-EPFL, Lausanne, Switzerland
In the context of solid-state/materials systems, the term chemomechanics refers to complex interrelated chemical and mechanical processes that originate at the atomic scale. In particular, we use the term to refer to chemical reactions that determine the mechanical behaviour of a material. Some of these reactions are desirable (e.g., leading to strengthening of metal matrices where carefully chosen impurities have been added) while others are not (e.g., as they may lead to the failure of a ceramic medical implant by stress-corrosion cracking). In virtually all cases, an improved understanding of these processes has a potential for huge technologic and economic impact.
However, the nanoscale mechanisms underlying many chemo-mechanical processes are still poorly understood. We believe this has in part a cultural origin: the training which chemists and solid state physicists have been receiving in the last few decades is traditionally resonant of technologies like catalysis and semiconductor electronics, where mechanical properties are not immediately prominent. At the same time, the atomistic point of view has never been central in the training of mechanical or materials engineers, i.e. the experts who specialise in optimising the mechanical performance of structural materials. While this has recently started changing, the process of establishing a proper interdisciplinary tradition for atomistic modelling of chemomechanical phenomena is still in its infancy, in spite of its enormous potential for impact and benefit to the modelling community’s profile. The present proposal, taking on from a workshop on brittle fracture (the first of its kind) which was held in CECAM in 2011, aims at being a step in this direction.
The problems which arise in chemomechanics are inherently multi-scale, which poses further, very distinctive technical difficulties. Describing chemo-mechanical phenomena requires algorithms which dynamically couple the macroscopic stress fields with the chemistry at the atomic scale, to reproduce experimental observations. This is difficult for three principal reasons. Firstly, the accuracy of QM-based techniques is needed in the chemically active regions, but cannot be afforded in the entire simulated system. Non-uniform precision schemes involving the simultaneous coupling of QM and classical force field techniques can help address this issue, but the information efficiency of these techniques is still generally poor (e.g., most of the QM based information calculated is used just once during the simulation). Secondly, chemomechanical reactions typically take place on time scales longer than those accessible via conventional molecular dynamics, which requires rare events techniques. The reaction pathways are furthermore complex, and well established methods like nudged elastic band might not be sufficient: advanced free energy sampling techniques (e.g., metadynamics) are thus necessary to make significant progress. Thirdly, the experiments are themselves also extremely challenging, because processes typically take place at buried defects or interfaces, hindering direct observation. As has become a rule in nanoscale science when complex chemistry is involved, fundamental theory is needed to guide and support experiment for producing new knowledge and significant progress. Feedback by key experimentalists is envisaged in the present proposed workshop.
Targets of the workshop
In this workshop, we envisage focusing on the following target subject areas:
1. Chemomechanics of ductile materials (e.g. hydrogen embrittlement, solution strengthening of metals, and fracture screened by plastic response)
2. Chemomechanics of brittle materials (e.g. stress corrosion cracking of glasses)
3. Tribochemistry (e.g. environmentally enhanced wear, formation of tribological layers and their influence on friction)
Many significant steps have been made both from the experimental and theoretical point of view, however, there is no established systematic approach to the study of chemomechanical processes in materials systems yet. The objective of the workshop is therefore to bring together experimentalists and modellers working at various levels of theory to discuss possible ways forward, and to help establish new collaborations. In doing so, we also aim at keeping a focus on scientific problems of “real” contemporary industrial interest, by inviting R&D delegates from key EU industries.
1. Chemomechanics of ductile materials
H Embrittlement takes place in a number of metals become brittle after exposure to hydrogen. Some of the modern high-strength steels, for example, have the potential for reducing the weight of cars significantly, but have found few applications because of their dramatic susceptibility to hydrogen embrittlement.
The small atomic weight and the high mobility of hydrogen make it difficult to detect for local experimental probes, such that modelling is of particular importance for a better understanding of hydrogen embrittlement. Hydrogen embrittlement has been simulated with a wide variety of models and approaches, covering all length scales relevant for materials. Density Functional Theory (DFT) calculations contribute to clarify the interaction of hydrogen with stress fields, interfaces, surfaces and point defects . Atomistic molecular dynamics simulations that are based on embedded atom method potentials simulate the hydrogen accumulation at crack tips or dislocations [2,3]. Monte Carlo simulations have been used to analyse the diffusion of hydrogen in different materials . Continuum finite element methods predict the hydrogen distribution on a macroscopic length scale .
While a large number of simulations have been made, examples of coherent scale-bridging modelling that take into account the bond chemistry on the atomic level and from this predict hydrogen embrittlement on the macroscale, are scarce. A closer interdisciplinary integration of modelling approaches across the length scales would yield a better, more quantitative understanding of the relation between material composition/microstructure and hydrogen embrittlement.
Solute Strengthening is a further prominent chemomechanical process relevant to the failure of ductile materials, where substitutional solute atoms are added to a metal matrix, leading to local bonding alterations and elastic mismatches which can cause dislocation pinning. Within a multiscale context, the interactions between solutes and dislocations have been modelled with DFT calculations using flexible boundary conditions to represent the long-range stress as an external pressure field .
Stress corrosion cracking is the name given to a broad class of complex and interrelated chemical and mechanical processes. These may e.g., lead to the failure of normally ductile metals and alloys when mechanically loaded and exposed to a chemically aggressive environment . Recent developments in chemically reactive interatomic potentials have allowed some chemical complexity to be included in MD simulations of fracture in ductile metals . The situation becomes more complex in brittle materials.
2. Chemomechanics of brittle materials
Stress corrosion cracking plays an arguably even more important role in the context of the failure of brittle materials, where cracks in ceramics and glasses exposed to a wet environment break under only very moderate (subcritical) loads. The overall lifetime of many ceramics and glasses is limited by slow growth of small defects due to these corrosion processes, meaning that a better understanding of stress corrosion cracking could have important technological applications. An important open problem is how to improve the lifetimes of biomedical implants, which are often made of ceramics such as silica and zirconia because of their high biocompatibility, despite their undesired brittleness. For example, Prozyr hip implants were recalled in 2002 after a large number of failures were attributed to rapid in vivo ageing of their thin-film zirconia coatings .
Key outstanding issues in this field include the nature of the sub-critical crack propagation mechanism: in particular, the exact chemical mechanisms by which Si-O bond frameworks break in the presence of water molecules are unknown, as is the role of deep penetration of water molecules within SiO2 amorphous structures . It e.g., is still unclear whether the crack tip is dry with a constant flux of H2O molecules to the tip as postulated by Wiederhorn  or whether it there is a water-filled cavity as proposed by Crichton et al. .
Nanoscale measurements can help to provide a link between atomistic simulations and the macroscale. In situ atomic force microscopy (AFM) studies of slow crack propagation have been carried out, showing that water penetrates up to 10 nm ahead of a moving crack tip . The applicability of ab initio techniques to these problems has been somewhat limited to date; whilst some studies have looked at the role played by stress on the rate of individual chemical processes using small model systems , it is clear that to make real progress a multiscale modelling approach is necessary. QM/MM studies of stress corrosion processes have shown the power of such techniques to meet the simultaneous requirements of accurately describing crack tip chemistry and using sufficiently large model systems to capture the long range elastic relaxation around the crack tip [15,16]. However, much more work will be needed to address key issues such as how the diffusivity of water in silica changes in the highly stressed regions near a crack tip.
Chemomechanical processes taking place upon relative sliding of two surfaces have a huge influence on the tribological properties of materials, such as the measured friction. The tribological contact between solid surfaces can induce structural and chemical modifications of the near-surface material layer that modify the mechanical properties of the material . Moreover, mechanically activated chemical reactions routinely occur at the material/environment and material/lubricant interfaces . These reactions can determine the friction properties of the system through e.g., chemical passivation of the surfaces, formation of adsorbed “tribofilms”, oxidative wear and even degradation of the lubricant.
Despite their importance, tribochemical processes are still poorly understood because they take place at a buried interface, where direct experimental observation is extremely challenging. However, recent advancements in both experimental , and multiscale atomistic simulation approaches [20,21] are paving the way towards a deeper understanding of tribochemistry at the atomic scale.
Recently, atomic-scale wear has been investigated both experimentally and by atomistic simulation. A first example is the wear of AFM silicon tips which causes their premature blunting. While stress-corrosion cracking in the presence of water is a possible wear mechanisms for oxidised silicon tips, high precision AFM experiments suggest that wear can proceed as an atom-by-atom process, where successive bond breaking reactions are promoted by the interaction with water molecules during tip sliding . However, the detailed reaction mechanisms are still not understood and atomistic simulations would be an ideal tool to complement the experimental work.
Diamond and diamond-like carbon
Molecular dynamics simulations were used to study the wear of diamond and diamond-like carbon surfaces [20,21]. It was observed that a sp3-to-sp2 rehybridisation takes place at carbon surfaces upon solid/solid contact. Crucially, the simulations reveal that the mechanical amorphisation of diamond surfaces is responsible for the anisotropic wear rate of diamond. Moreover, chemomechanical reactions between sp-hybridized carbon atoms and oxygen molecules were observed to cause the oxidative wear of the surface.
Tribologically induced chemical reactions can lead to chemical passivation of the surfaces, which can in turn drastically reduce friction and wear . Passivation through e.g., hydroxyl groups can lead to further physisorption of polar molecules and to the formation of soft tribological layers that can very significantly enhance lubrication. These and other reactions are believed to be responsible for the superlubricity of tetrahedral-amorphous carbon (ta-C) in contact with OH-containing molecules (such as glycerol) . However, neither experiments nor atomistic simulations have so far been able to reveal the fundamental mechanism underlying superlubricity.
A closer collaboration between experimentalists and modellers, as well as improved integration between different simulations scales and methods appear necessary for further improvement in all the subfields.
Ralf Drautz ( Ruhr University, Bochum ) - Organiser
Gianpietro Moras ( Fraunhofer IWM ) - Organiser
Alessandro De Vita ( King's College London and University of Trieste ) - Organiser
James Kermode ( University of Warwick ) - Organiser