5th Users' Conference of IT4Innovations

Tuesday 9 Nov 2021, 09:00 → 16:00 Europe/Prague

Online (IT4Innovations)

5th Users' Conference of IT4Innovations will take place on November 9, 2021. 

After concerning the current situation with the ongoing pandemic of coronavirus disease and unpredictable future measures in the Czech Republic, we decided that the 5th Users' Conference of IT4Innovations will be fully virtual.

Timetable

9:00 - 10:00	Introduction to the latest advances at IT4Innovations towards the users of the IT4I systems
10:00 - 10:30	Users' forum
10:30 - 10:45	Break
10:45 - 12:15	Keynotes
12:15 - 13:00	Break
13:00 - 14:00	Users' talks
14:00 - 15:00	Poster session

 

Contribution types: 

Keynotes

3 keynotes will be presented at scheduled times during the conference. Each keynote is expected to take max. 30 mins (with discussion included).

Users' talks

Selected talks by our prominent users will be presented at scheduled times from 1 pm to 2 pm and will be divided into 3 separate sessions via ZOOM breakout rooms. Please, see the timetable of the conference to find the talks you wish to attend. 

Poster session

The poster session will take place in the breakout rooms via ZOOM. The poster session will be divided into two parts from 2 pm to 3 pm, for details see the timetable of the conference. Each presenter will have 30 minutes to present their poster in a separate breakout room.  

You can register for the conference now!

Attendees will discover more about our future upgrade plans, listen to talks given by our prominent users and can engage in discussions during the Users' Council meeting and poster session.

The link for the conference will be sent via e-mail on November 8, 2021 to registered participants only.

 

All presentations and educational materials of this course are provided under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license.

Poster Session

• Advances in the Evolutionary Design of Complex Cellular Automata_Bidlo.pdf
• A Reactive Force Field for Molybdenum Disulfide Crystallization_Ponomarev.pdf
• ATLAS Job Sumbission Systems for the IT4Innovation_Svatos.pdf
• Calculation of NMR spin-spin couplings for intrinsically disordered proteins_Gaffour.pdf
• Determining the elastic properties of single grains in the ultrafine-grained W-Cr composite_Kadzielawa.png
• Enhancement of Ion Acceleration by Laser-Pulse Shaping via Plasma Shutter_ Matys.pdf
• Enhancement of Ultrashort X-Rays Pulses_Maslarova.pdf
• Handling C++ Exceptions in MPI Applications_Jaros Jiri.pdf
• High-Order Magneto-Hydrodynamic Code with Spontaneous Magnetic Fields Generation_Nikl.pdf
• HPC Ising Model Solver_Application in Material Design_Dujovic.pdf
• Hydrogen effect on the strength of the Fe3Al alloy_Sestak.pdf
• Inversion of the satellite-observed tidal magnetic fields in terms of 3-D upper-mantle electrical conductivity_Sachl.pdf
• k-Dispatch's Performance Modules for Advanced Workflow Submission_Jaros Marta.pdf
• Peptide Elongation in Carbon Nanotubes_An Atomistic Approach_Nepomuceno.pdf
• Performance of k-Wave Toolbox on Multi-GPU Accelerated Clusters_Vaverka.pdf
• Preventing Deadlock i Non-uniform-timestep Distributed Pseudospectral Method_Kadlubiak.pdf
• Reactive Molecular Dynamics Study on the Tribological Properties of Selected Vanadium Oxides_Dasic.pdf
• Segregation of Oxygen Impurities in Transition Metal Nitride Superlattices_Cerny.pdf
• Theoretical study of sublimation, local disorder and strain energies of crystalline caged hydrocarbons_Tous.pdf
• Towards Solvent Interactions in Modelling Infrared Spectra of Rhenium(I) Complex_Melcak.pdf

Presentations Heyda-IT4I-5thUserConference-2021.pdf

09:00 → 10:00 Welcome and IT4Innovations news - Branislav Jansík & Vít Vondrák, IT4Innovations

10:00 → 10:30 Users' Forum: New election of executive board members

10:30 → 10:45 Break

10:45 → 11:15 Keynote I: R.Vácha: Enhanced translocation of amphiphilic peptides across membranes by transmembrane proteins

11:15 → 11:45 Keynote II: S. Böhm: HyperQueue: Simplifying Usage of PBS/SLURM Clusters

11:45 → 12:15 Keynote III: M. Kourniotis: Disruption of giant molecular clouds by the stellar feedback: hydrodynamic simulations

12:15 → 13:00 Break

13:00 → 14:00 Users' Talk II: Chairman: J. Jaroš

13:00 PeptideCS: Close to Complete Mapping of the Conformational Space of All Dipeptides by the Quantum Chemical Method

To understand how the conformational space of small peptide fragments determine the formation of protein three-dimensional structures is one of the important goals of modern biochemistry and structural biology. Achieving this goal will allow significant progress, for example, in the design of specific enzyme-based catalysts which would greatly simplify chemical synthesis in industry. Full control of the conformational behavior of protein fragments may also represent a way how to significantly deepen our understanding of protein folding and protein-ligand interactions.

To reliably elucidate the key determinants of conformational spaces of peptide fragments, we needed an extensive dataset. We used data from Peptide Conformational Samples dataset (PeptideCS), developed in our group, consisting of over 400 milions of dipetide off-equilibrium structures and 100 milions equilibrium structures, all computed at IT4I. The dataset contains structure, QM energy, QM energy in solvent (water), and gradients. To eliminate the redundancy in the equilibrium structures Because we found (equilibrium) structures of local minima to be redundant, we used machine learning clustering algorithm to select a set of unique representative minimas.

Enormous set of data represented by the PeptideCS database enabled us to investigate the impact of various physico-chemical factors determining the properties of the conformational space and of the minima (equilibrium structures) found. The factors include charge of the chain charge of the particular residue, its sterical requirements, and backbone-side chain interaction. These factors determine the shape of the conformational energy window, energy and histograms. In addition, we analysed and compared the energy surfaces of the dipeptides, defined by the 400 million off-equilibrium structures and investigated the secondary structure preferences, including energy barriers in the forbidden region of the Ramachandran diagram.

The extensive study should considerably improve the prediction of three-dimensional protein structure from the first principles (ab initio) which may complement recent successful efforts of the AI algorithms, such as AlphaFold2.

Speaker: Tadeas Kalvoda (Institute of Organic chemistry and Biochemistry of the CAS)

13:20 Tuning Nanoscale Intrinsic Friction via Quantum Mechanical Descriptors

Study of friction and energy dissipation always relied on direct observations. Actual theories provide limited prediction on the frictional and dissipative properties if only the material chemistry and geometry are known. We here develop a framework to study intrinsic friction and energy dissipation based on the only knowledge of the normal modes of the system at the equilibrium. To this aim, we conducted quantum-mechanic based investigation on prototypical TMDs. We combined the structural and dynamic information from group theoretical analysis and phonon band structure calculations with the characterisation of the electronic features using non-standard methods such us orbital polarization [1] and the recently formulated bond covalency [2] and cophonicity [3] analyses. We outline phonon-scattering selection rules [4], and propose a phonon-mode based method, named Normal-Modes Transition Approximation [5], to identify possible sliding paths from only the analysis of the phonon modes of the stable geometry and to tune the corresponding sliding energy barriers. We show how to characterize the frequency content of observed physical quantities [6-7] and individuate the dissipative processes active during experimental measurements. As a case study, we consider the relative sliding motion of atomic layers in transition metal dichalcogenides (TMDs) dry lubricant, and discuss how to extract information on the potential energy landscape from Atomic Force Microscopy signals.
The presented framework switches the investigation paradigm on friction and energy dissipation from dynamic to static studies, paving new avenues to explore for the design of novel anisotropic tribological and thermal materials [8-10]. Finally, thanks to the general formulation of our approaches, the present outcomes can be promptly used to finely tune physical properties for the design of new materials with diverse applications beyond tribology.

The present research has been published in impacted scientific journals and received support from the GACR and Horizon-2020 funding bodies.

References
[1] A. Cammarata et al. “Octahedral Engineering of Orbital Polarizations in Charge Transfer Oxides”, Physical Review B 87, 155135 (2013)
[2] A. Cammarata et al. “Covalent Dependence of Octahedral Rotations in Orthorhombic Perovskite Oxides, The Journal of Physical Chemistry 141, 114704 (2014)
[3] A. Cammarata et al. “Tailoring Nanoscale Friction in MX2 Transition Metal Dichalcogenides”, Inorganic Chemistry 54, 5739 (2015)
[4] A. Cammarata “Phonon–phonon Scattering Selection Rules And Control: An Application To Nanofriction And Thermal Transport” RSC Advances 9, 37491 (2019)
[5] A. Cammarata et al. “Overcoming Nanoscale Friction Barriers In Transition Metal Dichalcogenides”, Physical Review B 96, 085406 (2017)
[6] A. Cammarata et al. “Atomic-scale Design Of Friction And Energy Dissipation”, Physical Review B 99, 094309 (2019)
[7] A. Cammarata et al. “Control Of Energy Dissipation In Sliding Low-Dimensional Materials”, Physical Review B 102, 085409 (2020)
[8] A. Cammarata et al. “Fine control of lattice thermal conductivity in low-dimensional materials”, Physical Review B 103, 035406 (2021).
[9] A. Cammarata et al. “Effect of Noninteracting Intercalants on Layer Exfoliation in Transition-Metal Dichalcogenides”, Physical Review Applied 15, 064041 (2021)
[10] B. Perotti, A. Cammarata et al. “Phototribology: Control of Friction by Light”, ACS Appl. Mater. Interfaces 13, 43746 (2021)

Speaker: Dr Antonio Cammarata (Czech Technical University in Prague)

Cammarata_IT4I2021.pdf

13:40 Performance of VASP.6 on IT4I GPU accelerated supercomputers

VASP software (Vienna ab-initio simulation package) is one of the widely used Density Functional Theory implementations to predict and design materials properties. As such, it is present almost on all supercomputers and it is also a numerical code that heavily uses computational resources. On new architectures, however, the default compilations of VASP are often sub optimal, resulting in an inefficient use of computational resources. The parallel efficiency of the code also depends on correct choice of some code parameters. In this presentation I will address the problem of tuning the performance of the latest version of VASP code by optimal choice of compilers, compilers options, numerical libraries and provide some guide for correct choice of the program parameters.

Speaker: Sergiu Arapan (IT4Innovations VŠB - TU Ostrava)

13:00 → 14:00 Users' Talk III: Chairman: D. Legut

13:00 A first-principles study of pressure-induced changes in magnetism of Ni-Mn-Sn alloys

We have performed a quantum-mechanical study of a series of stoichiometric Ni2MnSn structures focusing on pressure-induced changes in their magnetic properties. Our study concentrated on the role of point defects, in particular Mn-Ni, Mn-Sn and Ni-Sn swaps. For most defect types we also compared states with both ferromagnetic (FM) and anti-ferromagnetic (AFM) coupling between (i) the swapped atoms and (ii) those on the original sublattice. Our calculations show that the swapped Mn atoms can lead to magnetic moments nearly twice smaller than those in the defect-free Ni2MnSn. Further, the defect-containing states exhibit pressure-induced changes up to three times larger (but also smaller) than those in the defect-free Ni2MnSn. Importantly, we find both qualitative and quantitative differences in the pressure-induced changes of magnetic moments of individual atoms even for the same global magnetic state. Lastly, despite of the fact that the FM-coupled and AFM-coupled states have often very similar formation energies (the differences only amount to a few meV per atom), their structural and magnetic properties can be very different. For details see M. Friák et al., Materials 14 (2021) 523, doi:10.3390/ma14030523.

Acknowledgments:
The authors acknowledge the Czech Science Foundation for the financial support received under the Project No. 20-16130S entitled “Multifunctional properties of powdered Ni-Mn-Sn Intermetallics”. Computational resources were made available by the Ministry of Education, Youth and Sports of the Czech Republic under the Project of the IT4Innovations National Supercomputer Center (project “e-Infrastructure CZ-LM2018140”) within the program Projects of Large Research, Development and Innovations Infrastructures.

Speaker: Martin Friák (Institute of Physics of Materials, Czech Academy of Sciences, Brno, Czech Republic)

13:20 FeSe superconductor with excess Fe: insight from first principles calculations

Iron-based unconventional superconductors with quasi-two-dimensional crystal structure have attracted intense interest after the critical temperature of FeSe was enhanced by more than one order of magnitude when thicknesses were reduced to the monolayer limit and placed on top of an insulating oxide substrate. In heterostructures comprising interfaces of FeSe with topological insulators, additional interesting physical phenomena are predicted to arise e.g. in the form of topological superconductivity [1].
Importantly, the tetragonal FeSe phase relevant for superconductivity is stabilized by excess Fe, leading to non-stoichiometric Fe(1+δ)Se compounds. However, the number of first-principles computational studies considering excess Fe is limited. We have studied Fe(1+δ)Se employing the coherent potential approximation and the tight-binding linear muffin-tin orbital method, which are well suited for disordered systems and can treat systems with even a very small off-stoichiometry without the need for a large supercell. It also allows us to explicitly address the impact of chalcogen vacancies.
Furthermore, we have studied how the FeSe bandstructure is modified by a prototypical interface to Si(001). We have also examined the effect of chalcogen height (Fe-Se planes distance) for both FeSe and Fe(1+δ)Se, since this is a parameter likely to be varied at an interface. This parameter has been determined with only a limited accuracy so far, and it appears to affect the band structure significantly here. Calculated band structures are compared to experimental ARPES data [2].

[1] LIU, X., et al. 2015. Electronic structure and superconductivity of {FeSe}-related superconductors. J. Phys.: Condens. Matter 27, 183201
[2] FIKÁČEK J. , et al, 2020. Step-edge assisted large scale FeSe monolayer growth on epitaxial Bi Se thin films. New J. Phys. 22, 073050

Speaker: Karel Carva

13:40 Antiferromagnetic magnons and local anisotropy: Dynamical mean-field study

We present a dynamical mean-field study of antiferromagnetic magnons in a one-, two-, and three-orbital Hubbard model of square and bcc cubic lattice at intermediate coupling strength. We investigate the effect of anisotropy introduced by an external magnetic field or single-ion anisotropy. For the latter we tune continuously between the easy-axis and easy-plane models. We also analyze a model with spin-orbit coupling in cubic site-symmetry setting. The ordered states as well as the magnetic excitations are sensitive to even a small breaking of SU(2) symmetry of the model and follow the expectations of spin-wave theory as well as general symmetry considerations.

Speaker: Dr dominique Geffroy (Masaryk University, Brno.)

13:00 → 14:00 Users` Talk I: Chairman: B. Jansík

13:00 Weakly Hydrated Anions Bind to Polymers but Not Monomers in Aqueous Solutions

Weakly hydrated anions help to solubilize hydrophobic macromolecules in aqueous solutions, but small molecules comprised of the same chemical constituents precipitate out when exposed to these ions. Herein, this apparent contradiction is resolved by systematically investigating the interactions of NaSCN with polyethylene oxide oligomers and polymers of varying molecular weight. A combination of spectroscopic and computational results reveals that SCN - accumulates near the surface of polymers, but is excluded from monomers. This occurs because SCN - preferentially binds to the center of macromolecular chains, where the local water hydrogen bonding network is disrupted. These findings suggest a link between ion-specific effects and theories addressing how hydrophobic hydration is modulated by the size and shape of a hydrophobic entity.

Speaker: Jan Heyda (University of Chemistry and Technology, Prague)

13:20 Molecular dynamics simulations of electron-transfer through tryptophan chains and a rhenium chromophore in modified azurins

Dynamical factors controlling hopping speed and efficiency in two modified azurin constructs that include a rhenium(I) sensitizer, Re(His)(CO)3(dmp)+, one or two tryptophans (W1, W2) and CuI center were studied by femtosecond time-resolved spectroscopy. Experimental kinetics investigations showed that the two closely spaced (3-4 Å) intervening tryptophans dramatically accelerated long-range electron transfer (ET) from CuI to the photoexcited sensitizer. In order to interpret electron-transfer processes connected with evolution of electronic states (3CT, 3CS1 and 3CS2) the complex systems were investigated by QM/MM molecular dynamics (MD) simulations in solution. The quantum part (QM) was Re(His)(CO)3(dmp) and part of protein chain up to W2, with the rest of the system treated by MM. The QM calculations employed DFT techniques with the PBE0 functiona and D3 dispersion correction.
It was found that TDDFT QM/MM/MD trajectories of low-lying triplet excited states of ReI(His)(CO)3(dmp)+–W1(–W2) exhibited crossings between sensitizer-localized (Re) and charge-transfer [ReI(His)(CO)3(dmp•–)/(W1•+ or W2•+)] (CS1 or CS2) states. Avoided crossing between the lowest Re and CS states on TDDFT trajectories demonstrated that dynamical fluctuations of solvated Re-tryptophan-azurins create situations where oxidation of the tryptophan by an electronically excited Re complex is energetically feasible, leading to a localized CS state. Calculating electronic coupling Hab between their diabatic approximates in the crossing region enabled us to assess whether the studied ET approaches the adiabatic or nonadiabatic limit.

Speaker: Dr Stanislav Záliš (J. Heyrovský Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, Dolejškova 3, CZ-182 23 Prague, Czech Republic)

13:40 Peptide deformylase as a probe of long-range allostery between ribosome surface and interior

Ribosomes are complex biomachines responsible for protein synthesis in all organisms. The ribosome consists of three strands of RNA and dozens of ribosomal proteins. It is organized into two subunits, where the small subunit reads the mRNA and the large subunit catalyzes peptide bond formation. The nascent protein translocates from the catalytic center to the ribosome surface through a 10 nm-long exit tunnel.
Protein synthesis is a highly regulated process. We have hypothesized a long-distance allostery within the ribosome, which could bring information about the tunnel content to the ribosome surface. The anticipated allosteric path engages ribosomal protein uL22, one of two ribosomal proteins that span the large subunit from the surface deep into the exit tunnel. On the ribosome surface, uL22 binds the peptide deformylase (PDF), an enzyme which initiates protein maturation in bacteria by removing the formyl group from the leading methionine of the nascent protein.
We have used all-atom explicit-water molecular dynamics simulations in equilibrium and non-equilibrium regimes, and describe the effect of PDF binding on the structure and dynamics of uL22. PDF binds through its C-terminal alpha-helix. The binding induces a conformational change of uL22 loops, one located on the surface and another in the exit tunnel. Although the signal-to-noise ratio in the equilibrium simulations is modest, the non-equilibrium simulations suggest that the PDF binding also affects the flexibility of the uL22 loop exposed to the tunnel. This could hint that PDF binding might affect the rate of nascent chain translocation through the exit tunnel.

Speaker: Michal H. Kolar (University of Chemistry and Technology)

14:00 → 14:30 Poster session: I

14:00 Enhancement of ultrashort directional X-ray pulses

Short directional X-rays have been extensively studied due to their high potential in many
applications such as high-quality phase-contrast imaging of biological samples and spectroscopy at
femtosecond timescale. Interaction of very intense laser pulses with plasma medium enables the
production of high energy photons in X- and gamma-ray range by several mechanisms, such as
betatron radiation, Compton scattering and bremmstahlung. A basic principle is that a relativistic
electron radiates light at a very short wavelength, which is a consequence of a relativistic Doppler
shift. In the frame of this project, the betatron radiation from laser wakefield acceleration (LWFA) of
electrons will be studied. In LWFA, electrons are injected into a plasma wave (wakefield), generated
and dragged by a few-tens-of-fs, ultra-intense laser pulse (driver) in optically transparent plasmas.
The betatron radiation is naturally generated due to the transverse oscillatory motion of electrons during the acceleration. In order to introduce LWFA X-ray sources fully for practical purposes, the production of more photons is required. To address this problem, plasma density modifications were examined. The process was studied for standard parameters feasible with current sub-100 TW laser systems by means of numerical particle-in-cell simulations. We show that the intensity of radiation increases when the plasma density increases.

Speaker: Dominika Maslarova

14:00 Exploring the complete conformational space of all 400 dipeptides by quantum chemical methods: Can we predict the protein structure "ab initio"?

To understand how the conformational space of small peptide fragments determine the formation of protein three-dimensional structures is one of the important goals of modern biochemistry and structural biology. Achieving this goal will allow significant progress, for example, in the design of specific enzyme-based catalysts which would greatly simplify chemical synthesis in industry. Full control of the conformational behavior of protein fragments may also represent a way how to significantly deepen our understanding of protein folding and protein-ligand interactions.

To reliably elucidate the key determinants of conformational spaces of peptide fragments, we needed an extensive dataset. We used data from Peptide Conformational Samples dataset (PeptideCS), developed in our group, consisting of over 400 milions of dipetide off-equilibrium structures and 100 milions equilibrium structures, all computed at IT4I. The dataset contains structure, QM energy, QM energy in solvent (water), and gradients. To eliminate the redundancy in the equilibrium structures Because we found (equilibrium) structures of local minima to be redundant, we used machine learning clustering algorithm to select a set of unique representative minimas.

Enormous set of data represented by the PeptideCS database enabled us to investigate the impact of various physico-chemical factors determining the properties of the conformational space and of the minima (equilibrium structures) found. The factors include charge of the chain charge of the particular residue, its sterical requirements, and backbone-side chain interaction. These factors determine the shape of the conformational energy window, energy and histograms. In addition, we analysed and compared the energy surfaces of the dipeptides, defined by the 400 million off-equilibrium structures and investigated the secondary structure preferences, including energy barriers in the forbidden region of the Ramachandran diagram.

The extensive study should considerably improve the prediction of three-dimensional protein structure from the first principles (ab initio) which may complement recent successful efforts of the AI algorithms, such as AlphaFold2.

Speaker: Tadeas Kalvoda (Institute of Organic chemistry and Biochemistry of the CAS)

14:00 Handling C++ Exceptions in MPI Applications

Handling error states in C++ applications is managed by exceptions. In distributed applications, it is necessary to inform the other processes, that something wrong happened, and the application should either recover from the faulty state, or report the error and terminate gracefully. Unfortunately, the MPI standard does not provide any support for distributed error handling. This poster presents a new approach for exceptions handling in MPI applications. The goals are to (1) report any faulty state to the user in a nicely formatted way by just a single rank, (2) ensure the application will never deadlock, (3) propose a simple interface and ensure interoperability with other C/C++ libraries. The code was tested with several injected errors into multiple ranks such as non existing input file, disk quota exceeded, wrong rank in the MPI call, and standard system exceptions. In all situations the code has worked properly.

Speaker: Jiri Jaros (Brno University of Technology)

14:00 High-order magneto-hydrodynamic code with spontaneous magnetic fields generation

The magneto-hydrodynamic description is adopted in many disciplines, where the dynamics of magnetized fluids is investigated. It plays a vital role in plasma physics, where high-temperature plasma becomes highly electrically conductive. Its modelling is then essential for the applications like inertial and magnetic confinement fusion, laboratory astrophysics and many others. For this purpose, we recently developed the resistive magneto-hydrodynamic extension of the multi-dimensional simulation code PETE2 (Plasma Euler and Transport Equations version 2) [1,2]. The Lagrangian nature of the code means that the computational mesh follows the flow of the matter unlike the traditional codes. Its numerical description is based on the high-order curvilinear finite elements, which provide high precision, computing efficiency, flexibility and robustness. The latest addition is the model of spontaneous magnetic fields, which are generated during the laser--target interaction or at the fronts of cosmic jets and elsewhere. The construction of the code is reviewed and examples of physically relevant simulations are given.

[1] J. Nikl, M. Kuchařík, and S. Weber. High-order curvilinear finite element magneto-hydrodynamics I: A conservative Lagrangian scheme. Journal of Computational Physics, 2021. Submitted.

[2] J. Nikl, M. Kuchařík, M. Holec, and S. Weber. Curvilinear high-order Lagrangian hydrodynamic code for the laser-target interaction. In S. Coda, J. Berndt, G. Lapenta, M. Mantsinen, C. Michaut, and S. Weber, editors, Europhysics Conference Abstracts – 45th EPS Conference on Plasma Physics, volume 42A, page P1.2019. European Physical Society, 2018.

Speaker: Jan Nikl (ELI Beamlines Centre, Institute of Physics, Czech Academy of Sciences, Czech Republic and Institute of Plasma Physics, Czech Academy of Sciences, Czech Republic and Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University, Czech Republic)

14:00 Hydrogen effect on the strength of the Fe$_{3}$Al alloy

We performed a quantum-mechanical molecular-dynamics (MD) study of Fe$_{3}$Al with and without hydrogen atoms under conditions of uniaxial deformation up to the point of fracture. Addressing a long-lasting problem of hydrogen-induced brittleness of iron-aluminides under ambient conditions, we performed our density-functional-theory (DFT) MD simulations for T = 300 K (room temperature). Our MD calculations include a series of H concentrations ranging from 0.23 to 4 at. % of H and show a clear preference of H atoms for tetrahedral-like interstitial positions within the D0$_{3}$ lattice of Fe$_{3}$Al. In order to shed more light on these findings, we performed a series of static lattice-simulations with the H atoms located in different interstitial sites. The H atoms in two different types of octahedral sites (coordinated by either one Al and five Fe atoms or two Al and four Fe atoms) represent energy maxima. Our structural relaxation of the H atoms in the octahedral sites lead to minimization of the energy when the H atom moved away from this interstitial site into a tetrahedral-like position with four nearest neighbors representing an energy minimum. Our ab initio MD simulations of uniaxial deformation along the ⟨001⟩ crystallographic direction up to the point of fracture reveal that the hydrogen atoms are located at the newly-formed surfaces of fracture planes even for the lowest computed H concentrations. The maximum strain associated with the fracture is then lower than that of H-free Fe$_{3}$Al. We thus show that the hydrogen-related fracture initiation in Fe$_{3}$Al in the case of an elastic type of deformation as an intrinsic property which is active even if all other plasticity mechanism are absent. The newly created fracture surfaces are partly non-planar (not atomically flat) due to thermal motion and, in particular, the H atoms creating locally different environments.

Acknowledgement
The authors acknowledge the Czech Science Foundation for the financial support received under the project No. 20-08130S. Computational resources were provided by the Ministry of Education, Youth, and Sports of the Czech Republic under projects e-INFRA CZ (ID:90140) at the IT4Innovations National Supercomputing Center and e-Infrastruktura CZ (e-INFRA LM2018140) at the MetaCentrum as well as the CERIT-Scientific Cloud (project No. LM2015085), all granted within the program Projects of Large Research, Development, and Innovations Infrastructures. M.F. and M.Š. acknowledge the support provided by the Czech Academy of Sciences (project No. UFM-A-RVO:68081723).

Speaker: Dr Petr Šesták (Central European Institute of Technology (CEITEC), Brno University of Technology, Technická 2, CZ-616 69 Brno, Czech Republic)

14:00 HyperQueue: Simplifying Usage of PBS/SLURM Clusters

In recent years, HPC workloads and communities have undergone substantial paradigm shifts. There is an increasing amount of users that want to leverage HPC clusters to execute many simple and embarrassingly parallel tasks as easily as possible. However, due to the limitations of traditional HPC job managers, these users must often resort to manual aggregation of tasks into a smaller number of jobs to reduce job manager overhead. This approach is both labour-intensive and inefficient, as it lacks dynamic load balancing required to fully utilize computational nodes with tens or hundreds of cores. We introduce HyperQueue - a task scheduling runtime that can execute a large amount of tasks on top of an HPC job manager by automatically aggregating tasks into jobs and dynamically load balancing them across all allocated nodes and CPU cores. HyperQueue is an open-source tool that is designed for ease of use and deployment.

Speaker: Jakub Beránek (IT4Innovations)

14:00 Inversion of the satellite-observed tidal magnetic fields in terms of three-dimensional upper-mantle electrical conductivity

The interaction of the oceanic tidal flows with the Earth's main magnetic field provides a powerful natural source of electromagnetic energy suitable for sub-oceanic upper-mantle electrical conductivity sounding. We have developed a new frequency-domain, spherical harmonic-finite element approach to the inverse problem of global electromagnetic (EM) induction. It is set up for an effective inversion of satellite-observed tidally-induced magnetic field in terms of three-dimensional structure of the electrical conductivity in the sub-oceanic upper mantle. The numerical code is parallelized using the OpenMP standard and it uses either the Math Kernel Library (MKL) or AMD Optimizing CPU Libraries (AOCL) to compute the Fourier transform effectively.

In order to demonstrate that the new approach can successfully reconstruct the 3-D upper mantle conductivity, we performed synthetic tests using a 3-D conductivity model WINTERC-E (Fullea et al., 2021) as a testbed. The WINTERC-E model is independent of any EM data and thus it represents an ideal target for synthetic tests of the 3-D EM inversion. In the next step, we proceed to the inversion of satellite-derived (Swarm) models of tidal magnetic signatures. We explore different datasets and different regularization settings in the inversion.

Speaker: Dr Libor Šachl (Department of Geophysics, Faculty of Mathematics and Physics, Charles University, V Holešovičkách 2, 180 00, Praha 8, Czech Republic)

14:00 k-Dispatch's Performance Modules for Advanced Workflow Submission

Complex scientific workflows describing challenging
real-world problems are composed of a lot of computational tasks
which require high performance computing or cloud facilities to
be computed in a sensible time. Most of such tasks are also
written as distributed parallel programs being able to run across
multiple compute nodes. The amount of requested resources
per task influences both the overall execution time (makespan)
and the computational cost. Optimal resource assignment to
particular task is thus crucial. Since the exact execution time
cannot be measured for every possible combination of task, input
data, and assigned resources, its prediction may be challenging. This poster introduces the idea of performance modules implemented within k-Dispatch that employ space-searching methods together with fitting and machine learning methods. The optimal amount of resources is found using those methods and evaluated using a cluster simulator that estimates the workflow makespan quickly. Using the performance modules, the execution plans for the input workflows are found and can be submitted to the real cluster. k-Dispatch then monitors those remote calculations within a provided service.

Speaker: Marta Jaros (Faculty of Information Technology, Brno University of Technology)

14:00 PeptideCS: An extensive dataset mapping the potential energy surface of mono- and dipeptides

Recent successes of deep-learning based approaches to the protein folding problem^1,2 have greatly relied on the existence of large protein structural³ and genomic⁴ databases. However, since these deep-learning based approaches cannot predict structure exactly, and the sufficiently precise DFT-based methods are computationally too expensive, the final structure refinement step still relies on empirically-constructed force fields. To substitute these force fields for more-accurate machine-learned DFT-based force fields, combining the speed of the empirical force fields with the accuracy of modern DFT methods, suitable training data are needed. Datasets of DFT-based molecular properties of small molecules were first available for equilibrium geometries (QM7, QM-9 and related datasets),⁵ and later extended to off-equilibrium geometries (e.g. ANI-1).⁶, which are a typical occurrence in proteins, but the molecules in these datasets are usually quite small and thus not representative of protein structures. Therefore, we believed that accurate energetic description of protein folding might benefit from a more specialized dataset. Therefore, we created Peptide Conformational Samples dataset (PeptideCS), where we sampled conformations with different dihedral angles in the main chain and side chains of -NH-Me and -CO-NHAc capped amino acids and dipeptides. Similar to others,⁷ we used DFT-B-based GFN2-xTB method⁸ to optimize appropriately constrained geometries, and then calculated energies, energy gradients and atomic charges at the BP86^9,10/DZVP-DFT¹¹ level with COSMO¹² solvation model (water) and, additionaly, energies in water, 1-octanol, N,N-dimethylformamide, and n-hexane solvents with the COSMO-RS¹³ solvation model. Our dataset consists of over 400 million non-equilibrium structures, uniformly sampled in all relevant dihedral angles. We also ran optimization at the GFN2-xTB level to obtain minima on the potential energy surface, resulting in additional 100 million structures. The resulting dataset should thus extensively cover all possible arrangements in these simple peptide building blocks and can be used in development and validation of protein force fields.

References (1) Science 2021, 373, 871–876. (2) Nature 2021, 596, 583–589. (3) Nucleic Acids Res. 2019, 47, D520–D528. (4) Nucleic Acids Res. 2005, 33, 154–159. (5) Sci. Data 2014, 1, 1–7. (6) Sci. Data 2017, 4, 1–8. (7) J. Chem. Inf. Model. 2020, 60, 6135–6146. (8) J. Chem. Theory Comput. 2019, 15, 1652–1671. (9) Phys. Rev. A 1988, 38, 3098–3100. (10) Phys. Rev. B 1986, 33, 8822–8824. (11) J. Chem. Theory Comput. 2017, 13, 3575–3585. (12) J. Phys. Chem. 1995, 99, 2224–2235. (13) J. Phys. Chem. A 1998, 102, 5074–5085.

Speaker: Erik Andris (IOCB Prague)

14:00 Performance of k-Wave Toolbox on Multi-GPU Accelerated Clusters

Realistic ultrasound simulations are becoming integral part of many novel medical procedures such as photoacoustic screening and non-invasive treatment planning. The common denominator of all these applications is the need for cheap and relatively large-scale ultrasound simulations with sufficient accuracy. Typical medical applications require full-wave simulations which take frequency-dependent absorption and non-linearity into account.

Our poster investigates performance of k-Wave acoustic simulation toolbox across evolving IT4Innovations supercomputer infrastructure. Our primary focus is on new clusters equipped with multi-GPU nodes – Barbora and Karolina. The behavior of k-Wave code should be useful in performance predictions for broader class of local Fourier basis pseudo-spectral codes. Our results from these new clusters are compared to other accelerated machines such as Salomon and Piz Daint and dense many-GPU node DGX-2. Finally, these results are put into the context of medical applications utilizing k-Wave toolbox.

Speakers: Filip Vaverka (Faculty of Information Technology, Brno University of Technology), Jiri Jaros (Brno University of Technology)

14:00 Preventing Deadlock in Non-uniform-timestep Distributed Pseudospectral Method

Non-uniform-timestep Distributed Pseudospectral Method allows different timestep sizes in different sub-domains of simulation. This requires special care when marching overlap regions forward in time. Overlap regions have to be evaluated at timesteps to match the neighbor's temporal discretization. This process involves the extrapolation of such values using a modified integration scheme. To minimize introduced error, extrapolation should be executed as late as possible. Under certain circumstances, this leads to a deadlock. In this poster, I will present necessary and sufficient condition that allows the deadlock to occur, detailed schematics of such case and simple solution that resolves the issue.

Speaker: Kristián Kadlubiak (Brno University of Technology)

14:00 Towards Solvent Interactions in Modelling Infrared Spectra of Rhenium(I) Complex

Ultrafast intersystem crossing (ISC) in transition metal complexes represents a challenging process for both experimental and theoretical investigation. Two computational approaches have been employed to study the relaxation processes during the ISC of complex [Re(bpy)(CO)$_\rm3$Cl] (bpy = 2,2’-bipyridine) in explicit dimethylsulfoxide solvent. Several relaxation processes has been identified along with their time scales and physical nature. Using state-of-the-art non-adiabatic molecular dynamics we have described the relaxation of molecular spin-orbit wave packet, and successive vibrational relaxation in non-adiabatic regime. We have connected the non-adiabatic simulations to the experiment through the fluorescence decay. An unconventional method for infrared spectra at sub-ps time scale was adapted. In accord with the experiment, we have observed the fast shift of frequencies and revealed the correlation with proximal solvent relaxation.

Next time scale, accounts for the vibrational relaxations. We restricted ourselves only to the monitoring of carbonyl modes, which provide connection to the experiment. Simulation results indicate that about 2 ps are needed for carbonyl modes to approach the new equilibrium frequencies. Longest time scale corresponds to the solvent relaxation. We found that 5 ps are enough to approach the new equilibrium distribution. Mechanistically, the redistribution of solvent follows the charge redistribution, i.e., negatively charged oxygen atoms are approaching closer to the carbonyl ligands and are pushed further from the bipyridine ligand.

Speaker: Martin Melčák

14:30 → 15:00 Poster session: II

14:30 Advances in the Evolutionary Design of Complex Cellular Automata

Evolutionary design has become a successful concept in the areas where a solution to a problem needs an exploration of extensive search spaces (i.e. the analytical approach is intractable or not known). It has been shown during recent years that innovative or still unknown solutions can be obtained automatically using proper evolutionary setups. Cellular automata represent a massively parallel computational concept allowing us to simulate various complex systems or event to perform computations (cellular automaton represents a universal computing model). The cellular automaton consists of a regular structure of many simple computing elements - cells, each of which may, at a given moment, acquire a state from a given finite set of states. The cell states are updated synchronously in discrete time steps according to transition rules specified for a given task to be solved. This process is called a development or computation of the cellular automaton. However, the design of suitable transition rules represents a difficult problem especially due to the fact that common programming paradigms are hardly applicable for the cellular automata. Moreover, the number of transition rules (and the number of transition functions determining the global behaviour of the cellular automaton) grows exponentially depending on the number of states and the simulation of the target behaviour typically requires a significant amount of time (it is needed to calculate the states of all cells for a sequence of several time steps). Therefore, the design of a cellular automaton is often treated as a search problem in the space of potential transition functions by means of suitable (meta)heuristic algorithms, e.g. by means of evolutionary techniques. This paper summarises some recent results regarding the research of representation techniques for the evolutionary design of complex two-dimensional cellular automata. Two representations will be mentioned: a conventional table-based method and an advanced approach utilising conditionally matching rules. In both cases the evolution strategy will be applied to design cellular automata for solving selected benchmark and real-world problems. In particular, the pattern development problem and the problem of filtering gray-scale images corrupted by a given type of noise are considered as case studies. It will be shown that using a proper settings of the evolutionary algorithm, interesting results can be obtained representing solutions of the given problems that have not been known before. Some observations from the analysis of resulting cellular automata will be presented which indicate, for example, that in some cases the behavior of the resulting automata is totally different depending on the representation applied. Specifically, the table representation exhibits a rather chaotic development of the cellular automaton during which a target pattern emerges at a single specific moment. On the other hand, the conditionally matching rules showed an ability to achieve behaviors that progressively construct the target pattern from a given initial state which, in addition, may represent a final state of the cellular automaton. Moreover, the latter method also exhibits significantly higher success rate which represents one of its advantages and proves an importance of systematic research in this area.

Speaker: Michal Bidlo (Brno University of Technology, FIT)

14:30 ATLAS job submission systems for the IT4Innovation

The distributed computing of the ATLAS experiment at LHC is allowed to opportunistically use computing resources of the Czech national HPC center IT4Innovations. Two submission systems based on the ARC-CE are used. Those will be the main focus of this contribution including description of their structures, functionalities and requirements, comparison of their advantages and disadvantages, and demonstration of significance of opportunistic resources contribution.

Speaker: Dr Michal Svatos (FZU, AV CR)

ATL-SOFT-SLIDE-2021-621.pdf

14:30 Calculation of NMR spin-spin couplings for intrinsically disordered proteins: a prospective tool to facilitate experimental NMR studies

Intrinsically disordered proteins (IDPs) are identified by polypeptide chains that do not have a stable single well-defined structure. They are responsible for the development of neurogenetic diseases such as Alzheimer and Parkinson. Structural characterization of IDPs has to be facilitated by both NMR experiment and computational techniques. The functionality of standard techniques such as X-ray crystallography is limited due to the high flexibility of IDPs. Thus the application of quantum mechanics (QM) calculations combined with molecular dynamics (MD) simulations is highly recommended.
In this contribution, we focus on the calculation of spin-spin couplings. The prediction of J-couplings typically builds on empirically parameterized Karplus equations. Alternatively, quantum mechanics (QM) can be applied if the empirical parametrization is prevented by the lack of training experimental data. We design a computational protocol that combines the molecular dynamics (MD) calculations with density functional (DFT) calculations along with fragmentation techniques.

The poster contribution builds on the application of the adjustable density matrix assembler (ADMA) for the protein fragmentation. We will discuss the effect of the DFT method and basis set as well as the effect of the size of surroundings on the computed spin-spin couplings. In addition, the impact of statistical averaging and ensemble size will be demonstrated for an example of Tau protein fragment.

Speakers: Amina Gaffour (faculty of pharmacy charles university- hradec kralove), Mr Bera Krishnendu (Central European Institute of Technology and Masaryk University, Brno)

14:30 Crystallization of amorphous metals through molecular dynamics simulations

Amorphous materials, such as amorphous alloys and metallic glasses (MGs) [1-4] are of current interest due to their peculiar internal structures and unusual properties [5-8] like high strength, excellent corrosion and wear resistance, localized deformation by shear banding and much higher hardness than crystalline alloys of comparable elastic modulus. Because glassy alloys do not exist in thermodynamic equilibrium, they undergo crystallization with the supply of thermal energy. However, despite decades of experimental and theoretical efforts, many questions regarding the details of these processes remain still open.

Due to the small space scales involved, the experimental investigation of the mechanisms underlying the phase formations caused by strain/stress as well as during nanoindentation on amorphous materials is difficult. Fortunately, the fast increase in available computation power is today allowing more and more accurate and larger size molecular dynamics (MD) simulations of almost any kind of system.

In this work, the crystallization of refractory metals (W, Nb, Ta, V, Mo) has been examined by means of molecular dynamics simulations. All these metals are more stable in their usual bcc structure, so under indentation, a growth of bcc crystals around the indenter (and occasionally in the whole sample) is observed. The velocity at which this process occurs depends on various factors like the type of metal, initial conditions on the amorphous sample, etc. The dependence on the initial density, melting point, and cohesive energy, has been investigated, as well as the structure of the resulting grain boundaries and grain size average. A simple model correlating crystal-forming ability (CFA) with thermodynamic properties will be presented.

References:
1. Klement, W., Willens, R. H. & Duwez, P. Non-crystalline structure in solidified gold-silicon alloys. Nature 187, 869–870 (1960).
2. Greer, A. L. Metallic glasses, in Physical Metallurgy, 5th edn, (eds. Laughlin, D. E. & Hono, K.), Elsevier,305–385 (2014).
3. Johnson, W. Bulk glass-forming metallic alloys: science and technology. MRS Bull 24, 42–56 (1999).
4. Inoue, A. Stabilization of metallic supercooled liquid and bulk amorphous alloys. Acta Mater. 48, 279–306 (2000).
5. Gaskell, P. H. A new structural model for transition metal-metalloid glasses. Nature 276, 484–485 (1978).
6. Miracle, D. B. A structural model for metallic glasses. Nat. Mater. 3, 697–702 (2004).
7. Sheng, H. W., Luo, W. K., Alamgir, F. M., Bai, J. M. & Ma, E. Atomic packing and short-to-medium-range order in metallic glasses. Nature 439, 419–425 (2006).
8. Ma, E. Tuning order in disorder. Nat. Mater. 14, 547–552 (2015).

Speaker: Prashant Dwivedi (Czech Technical University in Prague)

14:30 Determining the elastic properties of single grains in the ultrafine-grained W-Cr composite

While the elastic properties of a composite microstructure are relatively straightforward in obtaining (e.g., using the resonance ultrasound spectroscopy), the features of a single grain are not directly measured. Here, we show our approach to determine these properties by employing large-scale ab-initio modeling techniques. Furthermore, we quantitatively reproduce the observed values of the whole microstructure, basing our model-grain compositions on the X-ray diffraction patterns, thus substantiating the utility of this procedure in the design of novel alloys.

We acknowledged the financial support by the Czech Science Foundation through grant No. 20-18392S as well as the Czech MŠMT via project e-INFRA CZ (ID:90140).

Speaker: Andrzej Piotr Kądzielawa (IT4Innovations, VŠB-TUO; Jagiellonian University, Institute of Theoretical Physics, Poland)

14:30 Enhancement of Ion Acceleration by Laser-Pulse Shaping via Plasma Shutter

A plasma shutter [1,2] is usually a thin solid foil that is placed in front of the main target in the laser-target interaction. The laser pulse with its prepulses then needs to burn through it which can improve the laser pulse contrast and its intensity profile. This includes the generation of a steep rising front [3] and intensity increase [4] . The new pulse shape can improve the consequent ion acceleration from the additional target [5] and reduce the development of short-wavelength instabilities [6]. In this work, we study the effects of the plasma shutter on ion acceleration from an additional target with the help of 2D and 3D PIC simulations using code EPOCH [7] . We also consider a setup using two plasma shutters, where the first of them is expanded by prepulses simulated by hydrodynamic simulations. We demonstrate a substantial increase in maximal energy of accelerated heavy ions when a thin Si3N4 plasma shutter is added into the interaction of a PW-class laser pulse with the silver main target [5]. Our simulations are also visualized in our custom made web-based application with a virtual reality mode [8].

Supported by HIFI (CZ.02.1.01/0.0/0.0/15_003/0000449), Panosc (No. 823852) and IT4I (e-INFRA CZ, ID:90140)

References:
[1] S. A. Reed et al., Applied Physics Letters 94, 201117 (2009).
[2] S. Palaniyappan et al., Nature Physics 8, 763-769 (2012).
[3] V. A. Vshivkov et al., Physics of Plasmas. 5, 2727 (1998).
[4] M. Jirka et al., Physical Review Research 3, 033175 (2021).
[5] M. Matys et al., Proc. SPIE 11779, 117790Q (2021).
[6] M. Matys et al., High Energy Density Physics 36,100844 (2020).
[7] T. D. Arber et al., PPCF 57, 113001 (2015).
[8] M. Danielova et al., EuroVis 2019 - Posters, 20191145 (2019).

Speaker: Martin Matys (ELI Beamlines / Czech technical university in Prague)

14:30 HPC solver of Ising model: application in Material Design

Modeling thermal properties of new materials using computational methods based on Quantum Mechanics usually provides us with a well-defined ground state of the system, but not its transitional temperatures. For obtaining these properties, we propose a generic and high-performance method for mapping an ab-initio system onto the Heisenberg model using our original code (github.com/Mellechowicz/JorG).
The most crucial part of the algorithm is identifying a set of metastable states of the system, which is achieved through simulated annealing of a ferromagnetic 3D Ising model. This being also the most resource-intensive part of the algorithm, parallelizing it would significantly accelerate the entire process.
To accomplish this, we intend to implement parallel simulated annealing utilizing both distributed (Message Passing Interface) and shared (OpenMP/OpenACC) memory models. In this way, we will provide researchers with a fast and automatic scheme for assessing transitional temperatures that is well-adjusted for running in parallel on supercomputers, accelerating the design of new materials.

Speaker: Aleksandra Dujović

14:30 Peptide Elongation in Carbon Nanotubes: An Atomistic Approach

Proteins are the most ubiquitous biomolecules found in nature. Proteins are responsible for catalyzing reactions, transducing signals, structural properties, and much more. During their lives, proteins are found in various environments. It is not rare that proteins are found in enclosed spaces, such as pores, channels, and tunnels. For instance, in the first seconds of their existence, proteins are confined to the exit tunnel of the ribosome, the cell’s protein assembler. When in such spaces, proteins behave differently than when free in solvent. One of the simplest environments to mimic a confined space and study those differences is the carbon nanotube (CNT).
In this work, an automated pipeline was developed to emulate the growth of a peptide in confined environments through all atom molecular dynamic (MD) simulations and computational alchemy. By using the pipeline, we can observe the interactions and forces that are relevant in that environment after the addition of every amino acid.
Starting with a model enclosed space, a CNT with a 1.2 nm diameter, we performed several elongations with different polypeptide sequences (polyalanine, polyserine, polyglycine). The CNT was treated as an infinite structure in the elongation axis to avoid finite system interactions. 12 elongations (replicates) were performed per sequence, with 10 ns of free simulation per added amino acid (70 ns per elongation). Different starting points were randomly selected to ensure a better sampling of the conformational space. To evaluate the trajectories obtained, the end to end distance, radius of gyration, Root-Mean-Square Deviation (RMSD), intramolecular interactions, intermolecular interactions and the rate of progress of the peptide through the CNT were measured and compared. Smaller side chain amino acids, such as glycine, were shown to diffuse faster than slightly more complex side chains, like serine. It was also seen that the end to end distance was much larger in peptides which contained less intramolecular interactions. Furthermore, simpler amino acids moved through the CNT faster than more complex amino acids. This shows that more intramolecular interactions lead to shorter end to end distances, slower elongation rates and, therefore, to a slower progression of the peptide through the CNT. By having this data, and expanding to other amino acids, sequences, and CNT sizes, it would be possible to shed light on the elongation process of proteins in confined spaces, such as the ribosome, with atomistic resolution.

Speaker: Felipe Castro Nepomuceno

14:30 Reactive Force Field for Molybdenum Disulfide Crystallization

Molybdenum disulfide, MoS$_2$, is a layered material from transition metal dichalcogenide (TMD) family. Its range of applications includes tribological coatings, materials for electronics, and catalysis.
TMD thin films are often prepared via deposition processes, that originally yield an amorphous material[1]⁠. Tribological applications mostly rely on the ability of MoS$_2$ to crystallize in the course of exploitation. Other applications require MoS$_2$ to be crystallized in a certain way: in mild conditions for flexible stretchable photodetectors[2]⁠ or with specific defects for catalysis applications[3]⁠.
Accurate simulations at a large scale provide insights into collective events and structural changes in the course of transformation and isolate the effects of varying conditions. Those findings can be very useful for guiding experimental search of the treatment conditions.
Reactive Force Field (ReaxFF)[4]⁠ is an empirical potential, aiming to bridge the gap between first principle methods and simple empirical potentials. The former, such as DFT, are very accurate, but very computationally expensive: simulations are limited to thousands of atoms and hundreds of picoseconds. The latter, like Tersoff[5]⁠, are suitable for handling millions of atoms for nano- and even microseconds. One can achieve good results for bulk properties, but these potentials aren’t very accurate on the atomistic level. ReaxFF is supposed to be able to handle tens and hundreds of thousands of atoms for nanoseconds, while being at DFT level of accuracy. However, ReaxFF is a complicated potential, that comprises 39 general parameters, 32 parameters per atom, 24 parameters per bond, 7 parameters per angle and torsion. These parameters have to be carefully fitted to experimental and DFT-computed data to achieve the desired quality of the description.
Several ReaxFF parameterizations exist for Mo-S element system. They were used to study bending of MoS$_2$ layers[6]⁠, crystallization of a single layer of MoS$_2$[7,8]⁠, and formation MoS$_2$ from MoO$_3$ and sulfur[9]⁠. None of those, however, yielded a layered MoS$_2$ in our crystallization simulations outside the single-layer setup, producing a not-discovered experimentally and non-stable within DFT rock-salt type MoS.
In our search for V-O parameter set we found a parameterization that was producing desired layered structure in melt-quench and oxidation simulations. We used it as a basis for future development, using state-of-the-art parameters for S atom and applying Monte-Carlo parameter fitting vs. DFT-computed data. Convex-Hull diagram computed for Mo$_x$S$_y$ system within our new ReaxFF matches DFT results. In simulations, that imitated tribological conditions, we observed crystallization of layered MoS2. Our parameter set is a basis for future development and expanding the set of elements to study practically relevant compositions.

[1] Polcar T, et al. Rev Adv Mater Sci 2007;15:118
[2] Wuenschell JK, et al. J Appl Phys 2020;127.
[3] Hu J, et al. Nat Catal 2021;4:242+.
[4] Van Duin ACT, et al. J Phys Chem A 2001;105:9396.
[5] J. Tersoff. Phys Rev B 1988;37.
[6] Ostadhossein A, et al. J Phys Chem Lett 2017;8:631.
[7] Chen R, et al. J Vac Sci Technol A 2020;38:022201.
[8] Chen R, et al. J Phys Chem C 2020;124;50:27571.
[9] Hong S, et al. Nano Lett 2017;17:4866.

Speaker: Ilia Ponomarev (Czech Technical University in Prague)

14:30 Reactive Molecular Dynamics Study on the Tribological Properties of Selected Vanadium Oxides

Achieving lubrication at high temperatures and pressures, as well as in oxidative environment, is a challenging problem of modern tribology, relevant for a wide range of applications such as turbomachinery, machining tools and aerospace industry [1, 2]. A promising solution for the conditions of high temperatures/pressures and presence of oxygen is to use a hard and oxidation-resistant coating (Cr-N, Ti-N, Cr-Al-N, Ti-Al-N) containing an additional element, i.e., lubricious agent, that can diffuse to the surface of the coating and form an oxide which reduces friction [3]. Out of several metals used as lubricious agents, vanadium became a standard choice since its oxide melts at considerably low temperatures, hence providing liquid lubrication [4].

Vanadium reacts with oxygen and may form oxides with different stoichiometries. The motivation of this study is to investigate how does the oxidation state of vanadium impact the tribological properties of vanadium oxides. The crucial point is related to the effectiveness of vanadium oxides as lubricants: how low is the coefficient of friction and does it depend on the stoichiometry.

We have performed a reactive molecular dynamics study on the tribological properties of five selected vanadium oxides ($V_xO_y$) under the conditions of elevated temperatures $\{600, 800, 1000\}$ K and pressures $\{1, 2, 3, 4\}$ GPa, in 5 independent runs for each stoichiometry, temperature and pressure, hence totaling in 375 distinguishable simulations. We included the stoichiometries ($V_2O_3, V_3O_5, V_8O_{15}, V_9O_{17}, VO_2$) observed in the experimental studies of vanadium oxide-based coatings [5, 6]. The simulation setup consists of two rigid $V_2O_5$ layers and amorphous vanadium oxide $V_xO_y$ confined between them [7]. The two $V_2O_5$ layers are used to impose the normal load and a constant sliding velocity. Our simulations were implemented using the reactive force field (reaxFF) [8] in the reax/c package [9] of the LAMMPS code [10] and they were run at IT4I’s Barbora supercomputer.

By applying a linear fit on the dependence of the sliding force $F_x$ on the normal load $F_z$:
\begin{equation}
F_x = COF \cdot F_z + F_x^{0}
\end{equation}
we extracted the coefficient of friction $COF$ and the sliding force at zero load $F_x^{0}$. For all stoichiometries the coefficient of friction decreases with the increase of temperature and takes the values $COF < 0.2$, which is consistent with the experimental findings [11]. In the offsets we obtained a clear separation and ordering, depending on the stoichiometry of vanadium oxides. We have explained those results with a structural analysis, i.e., we computed the average coordination number estimating the bonds between vanadium in amorphous $V_xO_y$ and oxygen in rigid $V_2O_5$ layers. As the oxidation state of vanadium in $V_xO_y$ increases (in the order $\{V_2O_3 < V_3O_5 < V_8O_{15} < V_9O_{17} < VO_2\}$), there are less bonds between vanadium atoms of amorphous $V_xO_y$ and oxygen atoms of rigid $V_2O_5$ layers.

The key finding of our study is that each of the considered vanadium oxides provides liquid lubrication under the imposed conditions. This finding represents a valuable information relevant for the design of coatings containing vanadium as the lubricious agent.

Speaker: Dr Miljan Dašić (Faculty of Electrical Engineering, Czech Technical University in Prague)

DasicMiljan_OSTRAVA2021.pdf

14:30 Segregation of oxygen impurities in transition metal nitride superlattices

Using a first-principles approach we investigated an effect of oxygen as substitutional impurity on cohesion of three transition metal nitride multilayers. Namely, we studied AlN/VN, AlN/TiN and VN/TiN systems in a rock-salt structure (B1) with (0 0 1) interfaces. Preferred oxygen positions were determined with the help of calculations of free energy that also included vibrational entropy terms. Subsequent calculations of cleavage energy for all possible cleavage planes enabled us to identify the weakest link and to assess the impact of the oxygen impurity on a cohesion of each of the studied multilayers. Supercells of different size were used to estimate possible effect of oxygen concentration. The results indicate that oxygen prefers to replace nitrogen atoms in interfacial planes and that these impurities do not reduce the multilayer cohesion. Moreover, in the case of AlN/TiN system, their presence was found to increase the cleavage energy of the interface.

Speaker: Miroslav Černý (CEITEC, Brno University of Technology)

14:30 Theoretical study of sublimation, local disorder and strain energies of crystalline caged hydrocarbons

Caged hydrocarbons exhibit interesting properties, rendering these materials suitable candidates for a variety of uses. Their carbon skeleton usually experiences a large strain unlike other, chained hydrocarbons – this prompts discussion about their use as energy storage media, precursors for pharmaceuticals or explosives. Despite this, relevant thermodynamic data for these compounds is relatively scarce or subject to significant inconsistencies. The primary aim of this work is to evaluate sublimation properties such as sublimation enthalpy and sublimation pressure using ab initio and density functional theory calculations. Calculations are performed both for the solid and gaseous phase and their results processed using statistical-mechanical tools. Furthermore, an ab initio fragment-based additive scheme is utilized to accurately assess the cohesion energy of the crystals. This scheme relies on the additivity of pairwise or higher-order interactions to calculate bulk properties using higher levels of theory, which would be otherwise unaffordable in bigger systems. Additionally, disorder in the solid phase of caged hydrocarbons is also studied. Carboxylated derivatives of caged hydrocarbons exhibit proton transfer in their respective solid phases, with nonsymmetrical potential surfaces thanks to the crystal lattice. This happens seemingly even at room temperatures, if the solid phase is stable. The corresponding energy barriers are studied using the dimer method to sample the potential energy landscape. Some caged hydrocarbons may also exhibit a form of rotational disorder, as their molecules possess symmetrical shape and zero dipole moment, enabling to rotate in the crystal lattice without significant energy penalty. Both of these phenomena can potentially affect measurements of atomic positions, resulting in inaccurate or incomplete crystallographic data. The aforementioned strain energy of various caged hydrocarbons is also evaluated as the reaction enthalpy of the corresponding homodesmotic decomposition reaction and by comparison with a molecular equivalent constructed from non-strained group increments.

Speaker: Petr Touš