-
Crystal structure identification with 3D convolutional neural networks with application to high-pressure phase transitions in SiO$_2$
Authors:
Linus C. Erhard,
Daniel Utt,
Arne J. Klomp,
Karsten Albe
Abstract:
Efficient, reliable and easy-to-use structure recognition of atomic environments is essential for the analysis of atomic scale computer simulations. In this work, we train two neuronal network (NN) architectures, namely PointNet and dynamic graph convolutional NN (DG-CNN) using different hyperparameters and training regimes to assess their performance in structure identification tasks of atomistic…
▽ More
Efficient, reliable and easy-to-use structure recognition of atomic environments is essential for the analysis of atomic scale computer simulations. In this work, we train two neuronal network (NN) architectures, namely PointNet and dynamic graph convolutional NN (DG-CNN) using different hyperparameters and training regimes to assess their performance in structure identification tasks of atomistic structure data. We show benchmarks on simple crystal structures, where we can compare against established methods. The approach is subsequently extended to structurally more complex SiO$_2$ phases. By making use of this structure recognition tool, we are able to achieve a deeper understanding of the crystallization process in amorphous SiO$_2$ under shock compression. Lastly, we show how the NN based structure identification workflows can be integrated into OVITO using its python interface.
△ Less
Submitted 8 May, 2024;
originally announced May 2024.
-
Mechanical Characterization of Superelastic NiTi Nanofoams by Molecular Dynamics Simulations
Authors:
Arne J. Klomp,
Karsten Albe
Abstract:
Nanoporous metals or nanofoams are a promising material class that is considered for sensing, actuation, and catalysis. To date, they mostly based on simple noble metals such as nanoporous gold, which exhibit peculiar stress-strain response different from the bulk material. At the same time bulk alloys such as NiTi feature a reversible martensitic phase transition giving rise to interesting shape…
▽ More
Nanoporous metals or nanofoams are a promising material class that is considered for sensing, actuation, and catalysis. To date, they mostly based on simple noble metals such as nanoporous gold, which exhibit peculiar stress-strain response different from the bulk material. At the same time bulk alloys such as NiTi feature a reversible martensitic phase transition giving rise to interesting shape memory and superelastic effects. Combining the rich mechanics of NiTi with the geometrical features of a nanofoam is expected to improve the mechanical performance of this material. In this atomistic study we explore the behavior of a NiTi nanofoam at varying temperature and its reaction to (cyclic) compression. Using molecular dynamics simulations we track the microscopic processes enabling reversible deformation as well as the mechanical failure mechanisms of the NiTi nanofoam.
△ Less
Submitted 28 July, 2022;
originally announced July 2022.
-
The nature of deformation-induced dislocations in SrTiO3: Insights from atomistic simulations
Authors:
Arne J. Klomp,
Lukas Porz,
Karsten Albe
Abstract:
The nature of mechanically induced dislocations in SrTiO3 at low temperatures has been a disputed matter for a long time. Here we provide a systematic overview of the existing knowledge on dislocations in stoichiometric SrTiO3 complemented by computational models of several of the proposed dislocation types. Based on atomistic simulations we reveal their structure and mobility and put our focus on…
▽ More
The nature of mechanically induced dislocations in SrTiO3 at low temperatures has been a disputed matter for a long time. Here we provide a systematic overview of the existing knowledge on dislocations in stoichiometric SrTiO3 complemented by computational models of several of the proposed dislocation types. Based on atomistic simulations we reveal their structure and mobility and put our focus on the splitting into partial dislocations, because this mechanism is held responsible for the good dislocation mobility in strontium titanate. Our results reveal that dislocations with a {1-10} glide plane (types A, B, and C) show easy glide behavior due to their glide dissociated configuration. The motion of dislocations on {001} glide planes (type E), however, is prohibited due to high Peierls barriers, as is the case for the {1-1-2} glide plane (type D). For dislocations with (partial) edge character the dissociation into partials is shown to be very sensitive to the charge state of the dislocation core. Since, in turn, dislocation mobility is strongly related to the dissociation of dislocations, the macroscopic ductility may in some cases have a direct relation to the charge at the dislocation core.
△ Less
Submitted 27 July, 2022;
originally announced July 2022.
-
Dynamic stability of nano-scale ferroelectric domains by molecular dynamics modeling
Authors:
Arne J. Klomp,
Ruben Khachaturyan,
Theophilus Wallis,
Anna Grünebohm,
Karsten Albe
Abstract:
Ultra-dense domain walls are increasingly important for many devices but their microscopic properties are so far not fully understood. In this study we combine atomistic and coarse-grained molecular dynamic simulations to study the domain wall stability in the prototypical ferroelectric BaTiO3. We transfer the discussion of the field-driven nucleation and motion of domain walls to thermally induce…
▽ More
Ultra-dense domain walls are increasingly important for many devices but their microscopic properties are so far not fully understood. In this study we combine atomistic and coarse-grained molecular dynamic simulations to study the domain wall stability in the prototypical ferroelectric BaTiO3. We transfer the discussion of the field-driven nucleation and motion of domain walls to thermally induced modifications of the wall without an external driving force. Our simulations show that domain wall dynamics and stability depend crucially on microscopic thermal fluctuations. Enhanced fluctuations at domain walls may result in the formation of critical nuclei for the permanent shift of the domain wall. If two domain walls are close - put in other words, when domains are small - thermal fluctuations can be sufficient to bring domain walls into contact and lead to the annihilation of small domains. This is even true well below the Curie temperature and when domain walls are initially as far apart as 6 unit cells. Such small domains are, thus, not stable and limit the maximum achievable domain wall density in nanoelectronic devices.
△ Less
Submitted 24 June, 2022;
originally announced June 2022.
