A01 - Solid Solution effects on the Slip System Selection and Dislocation Activity
- Stefanie Sandlöbes
- Dierk Raabe
The sub-project A01 will focus on investigating the dislocation- and stacking fault-solute interactions and their effects on the mechanical properties. More specific, the solute-dependent nucleation mechanisms and mobility of dislocations on different slip systems and their effects on the ductility and work hardening will be studied considering also aspects related the change in dislocation core structure and cross slip, solute clustering, ordering phenomena and possible effects associated with impurity elements such as O.
For this purpose we will systematically test and characterize solid solution bulk and thin-film (B02) Mg-Al-Ca samples with varying Al and Ca concentrations. Mechanical and micro-mechanical testing will be utilized to identify promising alloy compositions which will then be in-depth characterized to identify (i) active dislocation systems, (ii) stacking fault energies and stacking fault configurations, (iii) the local chemistry and structure at these defects using scanning electron microscopical methods as electron backscatter diffraction (EBSD) and electron channelling contrast imaging (ECCI), conventional transmission electron microscopy (this sub-project), high resolution transmission electron microscopy (A03) and atom probe tomography (B02 and this sub-project).
The motivation of project A01 is to provide a more fundamental understanding of how the presence of dilute solutes influences both, the dislocation core structure and the associated dislocation kinetics to enable activation of non-basal slip modes that do not occur in pure Mg. Specifically, the combined effects of Al and Ca on dislocation core energies, configuration and mobility in Mg will be investigated using mechanical testing, electron microscopy and atom probe tomography. A special focus will be placed on the nucleation mechanisms and mobility of dislocations on different slip systems and their effects on the ductility and work hardening.
The key scientific questions in this project are:
- Do Al and Ca atoms form clusters?
- Where are these atoms located, more specific, how are Al and Ca located at defects such as stacking faults and dislocations and what is there is their local concentration?
- What is the nucleation mechanism of the observed <c+a> dislocations in the presence of Ca and Al?
- What is the nature of the interaction among dislocations on different slip systems?
The ultimate aim of this sub-project is to provide general ductilization descriptors for solid solution Mg alloys in close cooperation with theoretical work in sub-projects (A02, A06, C05).
- Mechanical testing and micromechanical testing
- Electron microscopy (SEM, EBSD, ECCI, conventional TEM)
A02 - Atomistic Simulations of Dislocation Processes
- Erik Bitzek
- Julien Guénolé
Some of the most fundamental characteristics and properties of dislocations in the dominant precipitate phases of the Mg-Al-Ca system remain currently largely unexplored. In contrast, dislocations in the hcp matrix phase have been increasingly well studied, particularly by atomistic simulations. Nonetheless, the effect of segregating atoms to dislocation cores on their properties and on dislocation-defect interactions has not yet been studied in detail or discussed in the framework of defect phase diagrams. In this project, we plan to gain, in close collaboration with the experimental projects (A03-05, B01, C02), mechanistic insights into the fundamental dislocation processes in complex phases and the microstructures that contain them. We also plan to provide quantitative information on dislocation properties to higher-scale models (C01) and to study the interaction of dislocations with other defects, in particular interphase boundaries (IPBs), in order to obtain a better understanding of the role of local chemical composition at defects on the mechanical properties.
- Determine the natural slip systems for dislocations in Mg17Al12 and (Mg,Al)2Ca phases, their core structure and properties. This requires the development of novel defect analysis methods for complex crystal structures which will be realized together with project A04 and applied in modified form to HREM data with project A03.
- Analyze the effect of local chemical composition on the structure and properties of dislocations in the matrix and precipitate phases and on IPBs as well as on their mutual interactions.
- Perform large-scale atomistic simulations of mechanical tests in complex precipitate containing microstructures, including in conjunction with project A04 the simulation of fracture and crack propagation in Mg17Al12, Mg2Ca, and along the IPB.
- Massively-parallel molecular dynamics (MD), also in combination with Monte Carlo (MC) steps. Additional nudge-elastic-band (NEB) calculations will be performed to reveal unknown plasticity mechanisms and to determine activation energies, e.g. for dislocation glide and dislocation cross-slip.
- In the first part of the project we will use existing recent reliable MEAM potentials for the Mg-Al and Mg-Ca systems. In the second part, we will test and apply the Mg-Al-Ca moment tensor potentials developed in A06.
A03 - Atomic Scale Characterisation of Dislocations, Planar and Point Defects
- Joachim Mayer
In the present project, the most advanced high-resolution imaging and analytical techniques based on aberration corrected transmission electron microscopy will be employed to elucidate the structure-property relationships. Sample preparation and conventional TEM analysis will be performed at the Central Facility for Electron Microscopy of RWTH Aachen University and advanced TEM will be applied at the Ernst Ruska-Centre at Forschungszentrum Jülich.
The project aims at providing the following information about the investigated partners:
- Quantitative description of positional parameters and three dimensional nature of defect core structures
- Column-resolved chemical description and bonding information of the core structures
- Quantitative matching of experimental and simulated images for structural refinement and improvement of the precision of structural parameters.
- Aberration-corrected high-resolution TEM (HRTEM) and scanning TEM (HRSTEM), as well as energy loss spectroscopy (EELS)- and energy dispersive X-ray (EDX)-spectroscopy will be the main techniques, supported by convergent beam electron diffraction (CBED), tomography and novel techniques like the application of pixelated detectors in STEM.
A04 - Unsupervised Crystal Analysis of High Resolution Image Data
- Benjamin Berkels
Since the understanding of defects is at the core of the proposed CRC, an important necessity towards this goal are proper mathematical tools to detect, categorize and quantify defects in atomic resolution data, both experimental data (found in A03+B01) and data obtained by computational simulation (found in A02+A06+C05)). Here, defects are deviations from a perfect crystal lattice. Thus, to characterize defects, the crystal lattice needs to be characterized at least in the area directly surrounding the defect. The development of tools for the latter, i.e. the analysis of crystalline structures (including the quantification of displacements induced by defects) from atomic resolution data is the goal of this project. Considering the sheer amount of data that will be available in the proposed CRC, a manual analysis forbids itself. Therefore, the methods to be developed need to be able to work in an unsupervised, or at most in a semi-supervised manner.
- Unsupervised segmentation of atomic resolution images of crystals into individual grains or crystal types without prior knowledge of the unit cell
- Robust characterization of the crystal unit cells as part of the segmentation
- Quantification of the deviation of an individual crystals from the perfect lattice
- Method for defect characterization of atomic resolution image data that simultaneously segments image data into crystals and quantifies deviations
- A variational approach based on the Mumford-Shah model and local Fourier transforms as a high dimensional feature extractor will be the basis for the segmentation.
- The deviation of individual crystals from the perfect lattice is modeled as non-rigid distortion field ψ and determined using variational techniques while curl ψ encodes information about dislocations and grain boundaries.
- 2D synchrosqueezed transforms serve as an alternative to extract information on defects, rotations and the gradient of the lattice deformation from crystal images.
A05 - Dislocations in Complex intermetallic Crystals
- Sandra Korte-Kerzel
In this project, the dislocation-mediated plastic deformation mechanisms of several complex crystals in the Mg-Al-Ca system, namely the MgCa2, AlCa2 and (Mg,Al)Ca2 Laves phases as well the Mg17Al12 phase, will be investigated by nanomechanical testing and electorn microscopy.
Plastic deformation in complex crystals is poorly understood. In most cases this is due to the intrinsic difficulty of moving dislocations through a complex lattice and the resulting brittleness of the crystals. However, a detailed understanding of how complex crystals deform plastically is essential for at least two reasons: to identify phases that posess desried properties (e.g. appreciable ductility) and to purposefully manipulate and engineer the complex components found in many high preformance alloys.
- Characterization of the dislocation structures (slip systems), associated critical stresses and underlying mechanisms of motion in considered Mg-Al-Ca intermetallic phases
- Development of an efficient experimental analysis to obtain this data reliably and efficiently for a large number of experimental variables, such as temperature, crystal structure orientation, stoichiometry
- Derivation of general concepts describing dislocation plasticity in complex crystals in close collaboration with other TPs
- Nanomechanical testing (nanoindentation and microcompression) at room and elevated temperature
- Slip trace analysis from correlative indentation and EBSD
- Slip plane/system analysis by conventional TEM (BF, DF, g*b analysis/LACBED)
A06 - Ab Initio Accuracy at Large Scales by Machine Learning
- Liam Huber
- Jörg Neugebauer
A realistic description of the interaction between solutes and extended defects requires often system sizes that go well beyond what is presently achievable by first principles density functional theory calculations. In contrast, classical potentials that are capable of simulating millions of atoms – and thus complex defects – are not as accurate or transferable, especially for multi-component alloys. In this project we will address this gap by using machine learning techniques to develop optimized interatomic potentials. Using these computationally highly efficient potentials we will study the interaction between solvents and defects as function of solvent concentration. Based on these studies concepts to efficiently construct defect phase diagrams will be developed.
- Development of high-accuracy machine-learned interatomic potentials for Mg-Al-Ca
- Compute and analyse structure and energetics of various low-symmetry defects at T=0K and use them to construct phase diagrams for 1D and 2D defects.
- Provide support for data storage and analysis using our in-house framework pyiron
- Density functional theory
- Molecular dynamics and statics using classical atomic potentials
- Machine learning of interatomic potentials, active learning
- Tools to handle, store and process big data from experiment and simulations