Projects Awarded under PRACE Project Access – Call 22

On this page you will find the projects that were awarded under Call 22 for Proposals for PRACE Project Access in April 2021.

Biochemistry, Bioinformatics & Life Sciences

Project Title: Sequence determinants of biological polyelectrolyte interactions

Project Leader: Benjamin Schuler, University of Zurich, Switzerland

Resource Awarded

  • 16 300 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Research Field: Biochemistry, Bioinformatics and Life sciences

Collaborators

  • Milos Ivanovic, University of Zurich, Switzerland
  • Robert Best, National Institutes of Health, United States of America
  • Thomas Dannenhofer-Lafage, National Institutes of Health, United States of America
  • Louise Pinet, University of Zurich, Switzerland
  • Rohan Eapan, University of Zurich, Switzerland
  • Daniel Nettels, University of Zurich, Switzerland

Abstract

Interactions between intrinsically disordered proteins are now recognized to play key roles in biology. A very important class of such interactions, particularly in the cell nucleus, are between highly charged proteins, which are involved in chromatin structure, transcriptional regulation and membraneless organelles amongst other functions. Interactions between highly charged proteins can yield disordered complexes of extremely high affinity, yet we currently do not have good predictive models for either the structure of these complexes or their affinity. Such models would allow the effects of sequence variations to be interrogated computationally and improve our understanding of complex structure, dynamics and function. To develop such a model, we will start with state-of-the-art atomistic simulations of highly charged disordered proteins and their complexes. Based on these data, we will develop an accurate coarse-grained molecular simulation model for interactions between charged proteins, including effects of salt. The model will be further tuned (and validated) against extensive experimental data. HPC will be indispensable, both for atomistic simulations and the large number of sequences and pairwise interactions that need to be computed for optimizing the coarse-grained model. Such a model will have broad applicability to investigating the role of disordered interactions, which are abundant in biology and widely involved in health and disease.

Project Title: ModFil – Modulating the structure and dynamics of bacterial filaments

Project Leader: Massimiliano Bonomi, Institut Pasteur – CNRS, France

Resource Awarded

  • 107 580 828 core hours on Piz Daint hosted by CSCS, Switzerland

Research Field: Biochemistry, Bioinformatics and Life sciences

Collaborators

  • Yasaman Karami, Institut Pasteur, France

Abstract

Many aspects of bacteria’s life involve proteins that are assembled into long and dynamic filaments. Characterizing the molecular mechanisms that ensure structural integrity and confer the flexibility needed to perform their functions is crucial. This knowledge can open new ways to modulate biological functions, for example by designing small molecules that interact with relevantly populated conformations or mutations that alter dynamic properties. In our previous PRACE allocation, we characterized by extensive atomistic simulations the dynamics of two fibers that play a crucial role in bacteria: the type-4 pilus (T4P) and the type-2 secretion system pseudopilus (T2SS-Ps). Our simulations revealed some of the molecular mechanisms that control the dynamic behavior of these filaments. In this proposal, we build on our previous results to design and test in silico and experimentally several small molecules and mutations that can interfere with the function of T4P and T2SS-Ps. Given the size of these systems and the time scales under investigation, the PRACE Tier-0 computing performance is required to execute our proposal. Overall, this PRACE allocation will enable performing a total aggregated simulation time of 0.7 milliseconds.

Project Title: Protein dynamics and toxicity in light chain amyloidosis

Project Leader: Carlo Camilloni, Università degli Studi di Milano, Italy

Resource Awarded

  • 68 000 000 core hours on Piz Daint hosted by CSCS, Switzerland

Research Field: Biochemistry, Bioinformatics and Life sciences

Collaborators

  • Cristina Paissoni, University of Milano, Italy

Abstract

The goal of the present project is to shed light on the physico-chemical principle of light chain amyloidosis. Light chain amyloidosis is a protein misfolding disease caused by deposition of monoclonal immunoglobulin light chains as fibrillar aggregates in the hearth, kidney and/or other target organs of a patient. The extreme variability among light chain sequences, caused by genetic rearrangement and somatic hypermutation, together with the strong structural similarity, makes extremely challenging to understand the determinants of light chain aggregation.

The aim of the present proposal is to systematically determine high-resolution structural ensembles, by means of Metadynamics Metainference simulations with small angle X-ray scattering data, for 8 light chains, 4 aggregation prone and 4 not. The use of a relatively large number of proteins is mandatory to try to account for the large sequence variability of these proteins. From the systematic comparison of the structural dynamics of these proteins we expect to determine the physico-chemical principles for light chain amyloidosis and design a set of experiments to test our findings. In this first year of project we have been determining the conformational ensembles of 4 light chains, 2 aggregation prone and 2 not, partial results will be presented in the following. Given the sequence variability of light-chains 4 more proteins will be needed as justified in the following to clarify to which extent are related to aggregation and to which extent are instead related to genetic rearrangements.

Project Title: Unveiling the free energy landscape of WT and mutated K-Ras dimerization with Raf effectors on a model membrane using naive and adaptive MD simulations

Project Leader: Zoe Cournia, Biomedical Research Foundation, Academy of Athens, Greece

Resource Awarded

  • 52 000 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Biochemistry, Bioinformatics and Life sciences

Collaborators

  • Camilo Velez-Vega, Novartis Institutes for Biomedical Research, United States of America
  • John Manchester, Novartis Institutes for Biomedical Research, United States of America
  • Ioannis Andreadelis, Biomedical Research Foundation of Academy of Athens, Greece

Abstract

Approximately one third of all human cancers harbor an oncogenic mutation in at least one RAS gene, rendering Ras proteins the most commonly mutated oncoproteins. Among them, K-Ras is the most frequently mutated isoform in human cancers (83% of RAS mutations), while in some pancreatic adenocarcinomas it is mutated in nearly 90% of tumors. It has been proposed that when K-Ras dimerizes on the membrane, Raf kinase uses K-Ras as a scaffold to anchor and form dimers on its own, which are essential for Raf activation leading to cell proliferation. Unfortunately, K-Ras remains so far undruggable due to the very tight binding of its substrate. New approaches such as targeting the K-Ras mutant dimerization, which is critical for signal transduction and uncontrolled cell proliferation in tumorigenesis could open new avenues in targeting this oncogene. The aim of this project is to fully describe the free energy landscape and kinetics of the dimerization of WT and one of the most common mutants of KRAS, G12D, in the presence of Raf kinase on a model membrane by using naïve and adaptive MD simulations Through adequate sampling of the conformational space of this 1M atom-system with Tier 0 resources we will be able to identify key conformations and kinetics of the K-Ras dimer that hold biologically critical information, which can be further exploited in the discovery of novel therapeutics.

Project Title: Active learning and free energy perturbation for drug discovery

Project Leader: Jonathan Hirst, University of Nottingham, United Kingdom

Resource Awarded

  • 30 000 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Biochemistry, Bioinformatics and Life sciences

Collaborators

  • Arnaldo Filho, University of Nottingham, United Kingdom
  • Ellen Guest, University of Nottingham, United Kingdom
  • David Rogers, University of Nottingham, United Kingdom

Abstract

Our project will harness the power of both machine learning and of physics-based molecular simulation to accelerate the discovery of compounds with predicted therapeutic values as leads in drug discovery efforts. The need to be able to discover new medicines quickly has clear scientific, societal and technological benefit, which, with the emergence of new diseases, is more apparent than ever. Molecular modelling has offered much assistance and promise over several decades. Very recent advances in machine learning and in molecular simulation are presenting some timely opportunities.

Free energy perturbation calculations, whilst still comparatively expensive, provide quantitative predictions of protein-ligand binding free energies. In our project, we will exploit the throughput of top-end high performance computing coupled with effective “active learning” strategies for exploring chemical space to shorten the time required to find potent compounds that are straightforward to make in a chemical laboratory and which have a profile of physicochemical properties appropriate for a drug. Our findings will allow us to assess the level of predictive accuracy needed to shorten the discovery cycle. We will validate our computational results, using a well-characterised anti-cancer target of current interest, through collaboration with synthetic chemists and assays conducted by GlaxoSmithKline.

Project Title: The Mechanistic Connection Between DNA Modifications and Epigenetics

Project Leader: Modesto Orozco, IRB Barcelona, Spain

Resource Awarded

  • 2 500 000 core hours on JUWELS Booster hosted by GCS at FZJ, Germany

Research Field: Biochemistry, Bioinformatics and Life sciences

Collaborators

  • Miłosz Wieczór, IRB Barcelona, Spain

Abstract

The project aims to identify and quantify major mechanistic connections between a range of DNA modifications and DNA readout, either through altered nucleosome accessibility or changes in affinity for transcription factors. As recent findings revealed periodic mutation patterns in cancer genomic data, the coupling between DNA damage, nucleosome positioning and DNA accessibility for repair has now been considered a potential factor in cancer progression. To disentangle these correlations at a molecular level, we need to study the thermodynamics of positioning effects exerted by individual modifications in the nucleosomal context. At the same time, oxidative DNA lesions are being more widely recognized as regulators of gene expression, besides their known detrimental role in mutagenesis. Besides their well-studied indirect effects on the stability of non-canonical DNA structures, they can also act directly, disrupting the binding sites for regulatory factors at sites inherently prone to oxidation. In this project, we propose a series of massive alchemical free energy calculations aimed at reliable quantification of these effects, identification of major contributors to physiologically observable outcomes, and experimental validation of our computational predictions. We also plan to establish standardized protocols for the alchemical treatment of DNA modifications for the DNA damage & repair field.

Project Title: Development of machine learning potential for Ta-Ti-V-W high-entropy alloys

Project Leader: Jan Wróbel, Warsaw University of Technology, Poland

Resource Awarded

  • 6 100 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Chemical Sciences and Materials

Collaborators

  • Mihai-Cosmin Marinica, CEA (Commissariat à l’énergie atomique et aux énergies alternatives), France
  • Duc Nguyen-Manh, Culham Centre for Fusion Energy, United Kingdom
  • Damian Sobieraj, Warsaw University of Technology, Poland

Abstract

Within this project, the theoretical investigation of microstructure evolution of Ta-Ti-V-W high-entropy alloys (HEAs) will be performed by using a combination of density functional theory (DFT), machine learning (ML) methods and molecular dynamics (MD) simulations. Thousands of DFT calculations performed using PRACE-ICEI computing services will be used to investigate the most important properties of HEAs, such as elastic properties, migration energies of vacancies in those alloys and the relative stability between bcc structures and liquid configurations as well as to create the input database for ML methods. The crucial and most challenging theoretical task will be the development of an accurate interatomic potential for MD simulations for Ta-Ti-V-W alloys using ML methods. MD simulations will enable the investigation of time evolution of the microstructure and its elastic properties as well as the diffusion of atoms in HEAs as a function of alloy composition and temperature. They will allow also the investigation of formation of solid phases in the melted high-entropy alloys. The results of MD simulations will enable not only a better understanding of processes existing during an additive manufacturing and annealing of these materials but also will help in a design of alloy compositions.

Project Title: Water Splitting at Ruthenium Oxide/Water Interfaces: forefront DFT-MD and Machine-Learning-Metadynamics Investigations

Project Leader: Sandra Luber, University of Zurich, Switzerland

Resource Awarded

  • 60 000 000 core hours on Piz Daint hosted by CSCS, Switzerland

Research Field: Chemical Sciences and Materials

Collaborators

  • Fabrizio Creazzo, University of Zurich, Switzerland
  • Rangsiman Ketkaew, University of Zurich, Switzerland

Abstract

In the proposed project, state-of-the-art ab initio Molecular Dynamics and advanced metadynamics techniques coupled with innovative machine learning techniques will provide an unprecedent and interdisciplinary understanding of ruthenium oxide aqueous interfaces in catalyzing the water splitting process in the context of solar light driven technology.

We will provide a fully assess of the ruthenium dioxide RuO2 as water splitting catalyst to help into the modelling of highly efficient photoelectrochemical (PEC) systems based on ruthenium oxides, looking for the first time at both water oxidation and hydrogen evolution reaction (HER) at the (110), (100), (001) RuO2 catalytic surfaces. We will step-by-step reveal the energetics, kinetics and thermodynamics behind the OER and HER mechanisms on ruthenium oxide (aqueous) catalysts carefully and rationally via forefront computer simulations and machine learning techniques.

For the first time, in this project, a study of the water splitting will be presented not only by looking at the catalysts, but also by addressing finite temperature effects and the role of an explicit water environment in the catalytic process. This project will be carried out in close cooperation with experimentalists.

Project Title: Understanding the Role of Aluminium in Cements by Ab-initio Modelling

Project Leader: Enrico Bodo, Department of Chemistry, University of Rome “La Sapienza”, Italy

Resource Awarded

  • 3 200 000 core hours on JUWELS Booster hosted by GCS at FZJ, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Mohammed Salha, University of Rome “La Sapienza”, Italy
  • Gregory Chass, Queen Mary University of London, London, UK, United Kingdom

Abstract

Portland cement is used globally due to its short setting time and its ability to yield strength quickly, however, it lacks lasting toughness compared to its ancient Roman counterpart. Its main binding phase consists of a calcium-silicon-hydrate gel (C-S-H) which has a porous and amorphous structure. The source of Roman cements extraordinary toughness lies in the higher aluminium content which is generally substituted into C-S-H forming C-A-S-H.

Al is thought to introduce angular flexibility counteracting strain due to pore expansion that would otherwise induce fracture in a cement gel. The aim of this project is to use a Tier0 facility for molecular dynamics simulations to see how such toughening mechanisms function in real time and to ascertain Al’s role in stabilising pore networks. Understanding the structure and the dynamics of Al inside the amorphous gel can represent the gateway to an improved long-term reliability of building materials.
Given the complex, amorphous and heterogenous nature of cement gels, HPC access is pivotal in pursuing the in-silico modelling necessary to understand the role of Al, whilst keeping the necessary level of accuracy for reliable predictions.

Chemical Sciences & Materials

Project Title: Ab initio molecular dynamics for nanoscale osmotic energy conversion

Project Leader: Gabriele Tocci University of Zurich, Switzerland

Multi-year Proposal: 3rd or 3 years

Resource Awarded

  • 82 000 000 core hours on Piz Daint hosted by CSCS, Switzerland

Multi-year Proposal: 3rd or 3 years

Research Field: Chemical Sciences and Materials

Collaborators

  • Marcella Iannuzzi, University of Zurich, Switzerland

Abstract

A vast amount of energy, so-called blue energy, may be harnessed from the mixing of salty and fresh water at river estuaries. It is estimated that about 2 Terawatt can be extracted in principle at river outlets, the equivalent of approximately 1000 nuclear power plants. Yet, blue energy remains an unexplored source, due to the limited efficiency of conventional membranes. Recent experiments have reported on exceedingly high power generated from ionic transport across two-dimensional nanopores and nanotube membranes. Although the chemical nature and electronic properties of these materials has been suggested to be highly relevant for nanoscale osmotic energy conversion, its role for blue energy applications is not known. In this project, we will investigate the role of the electronic structure of materials on fluid and ionic transport at the nanoscale using first principles quantum mechanical simulations. In particular, we will couple accelerated ab initio molecular dynamics simulations performed on HPC facilities (Piz Daint) to hydrodynamic theories in order to compute transport properties from these simulations. We will focus on different regimes of liquid and ionic transport and investigate. In particular we will look into the linear tranport regime, where the generated electrical current is linear with the potential drop between two reservoirs, as well as at the nonlinear regime in connection with electronic transport in microelectronics. We will further explore a large variety of interfaces investigated in recent experiments or that have been predicted by computational studies. Extensive use of HPC facilities is required due to the substantial cost of ab initio simulations of liquid/solid interfaces. By the end of the three years we will have established the key principles for the development of nanomembranes for osmotic power generation. Further areas that will benefit from the proposed research are water desalination, transport in biomembranes and DNA sequencing through nanopores.

Project Title: Biodegrading Plastics

Project Leader: Maria Ramos, University of Porto, Portugal

Multi-year Proposal: 2nd of 3 years

Resource Awarded

  • 24 000 000 core hours on MareNostrum 4 hosted by BSC, Spain

Multi-year Proposal: 2nd of 3 years

Research Field: Chemical Sciences and Materials

Collaborators

  • Ana Oliveira, University of Porto, Portugal
  • Rui Neves, University of Porto, Portugal
  • Joao Coimbra, University of Porto, Portugal
  • Pedro Fernandes, University of Porto, Portugal
  • Alexandre Magalhaes, University of Porto, Portugal
  • Krzysztof Biernacki, University of Porto, Portugal
  • Oscar Passos, University of Porto, Portugal

Abstract

We want to be able to biodegrade polyethylene terephthalate (PET) plastic because it is one of the most abundantly produced plastics, widely used in packaging and textiles (where it is known as polyester), which is accumulating in the environment at a staggering rate, and for which no green, environment and economically sustainable recycling strategy exists. Over 60 million tons PET plastic are produced every year, from which ~60% is used in non-recyclable textiles and ~30% in plastic bottles. Over 500 billion plastic bottles are produced every year, and more than half of them are never recycled. The bacterium Ideonella sakaiensis was found to have the fantastic ability of feeding on plastic. Very recently, the x-ray structures of the two bacterial enzymes responsible for this feat, PETase and MHTase, have been made available in the protein databank. These bacterial enzymes represent a very promising tool to solve the issue of PET plastic pollution because they exhibit a strong ability to biodegrade PET plastic at room temperature. This is the greenest way of biodegrading PET plastic used in medicinal purposes such as medical sutures, in which a small amount of the enzymes is necessary. However, PET plastic biodegradation on a larger scale e.g. for all plastic debris as referred above, needs upscaling engineering, where large amounts of enzyme have to be immobilized in solid surfaces to generate a PET-plastic degrading reactor. In that sense, detailed knowledge on the stability of the enzyme when adsorbed on specific solid surfaces, as done in industrial settings, is needed. Furthermore, the chemistry of these enzymes have to be understood, so that new enzyme mutants more resistant to immobilization and with faster PET plastic degradation rate are can be rationally designed, in order to accelerate the process and therefore lower the cost of the production. For this purpose, computer molecular dynamics simulations of enzyme adsorption on solid surfaces, such as graphene, are needed to evaluate enzyme unfolding and to design immobilization-resistant mutants. Additionally, quantum mechanical/classical mechanical computer simulations can be used, to gain an atomic-level picture of the enzyme’s chemical mechanism, and based on it to design enzyme mutants that are highly active even when immobilized. As enzymes are fully biodegradable themselves, and operate at room temperature and pressure, the development of efficient enzymes through computer simulations will greatly facilitate and accelerate the development of new PETase and MHTase enzymes that can be used in industrial degradation of PET plastic waste, in a green, energy-saving and ecologic way.

Project Title: Role of electron holes in promoting the oxygen evolution reaction at the hematite/water interface: a first-principles study based of DFT and TD-DFT

Project Leader: Simone Piccinin, National Research Council (CNR), Italy

Resource Awarded

  • 102 000 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Chemical Sciences and Materials

Collaborators

  • Matteo Farnesi Camellone, National Research Council, Italy
  • Giulia Righi, National Research Council, Italy
  • Xuejun Gong, SISSA, Italy
  • Travis Jones, Fritz-Haber Institute of the Max Plank Society, Germany

Abstract

Photoelectrochemical water splitting is promising strategy to convert and store intermittent energy sources like solar energy. Hematite (Fe2O3) is one of the most investigated and best performing materials for the oxygen evolution reaction (OER), the bottleneck of the whole process, but its performance is still way below the theoretical limit, for reasons that are still not well understood. Recent experiments based on time-resolved operando optical spectroscopy have offered unprecedented insights into the mechanism of OER on hematite, correlating spectroscopic fingerprints to reaction rates, hinting at a multi-electron mechanism. The missing link to achieve a full understanding of this photocatalytic system is to correlate the structure of the hematite/water interface under illumination to the spectroscopic features detected experimentally, and to understand the origin of the multi-electron mechanism. To achieve these goals, we will investigate the changes in the optical absorption spectrum of the hematite/water interface upon formation of holes, using DFT-based molecular dynamics and time-dependent DFT. We will then model the reaction mechanism of OER using NEB calculations at varying hole coverages, to unravel the origin of the change in reaction order detected in experiments. The project will lead to key insights both for reaction mechanism and for structure/function correlations.

Project Title: Computational investigations of electrochemical carbon dioxide reduction towards high value products

Project Leader: Karen Chan, Technical University of Denmark, Denmark

Multi-year Proposal: 1st of 3 years

Resource Awarded

  • 37 000 000 core hours on JUWELS hosted by GCS at FZJ, Germany

Multi-year Proposal: 1st of 3 years

Research Field: Chemical Sciences and Materials

Collaborators

  • Nitish Govindarajan, Technical University of Denmark, Denmark
  • Georg Kastlunger, Technical University of Denmark, Denmark
  • Sudarshan Vijay, Technical University of Denmark, Denmark

Abstract

The devastating effects of fossil fuel consumption on the climate and their limited supply have created an urgent need for efficient conversion of sustainable carbon sources, such as carbon dioxide (CO2) and lignocellulosic biomass to high-value fuels and chemicals. Since electrochemical conversion can occur under ambient conditions, it is an attractive route to store intermittent energy from renewable sources, and in a decentralized way as needed.

The objective of this proposal is to realize this vision through fundamental mechanistic and large-scale catalyst screening studies of electrochemical carbon dioxide (CO2) conversion. Ab-initio electronic structure (DFT) and molecular dynamics simulations (AIMD) can provide detailed insights on reaction mechanisms and predict catalytic activity, but are computationally demanding and necessitate HPC resources. Our envisioned large-scale simulations will address the following three interconnected themes: (1) fundamental investigations of the electrochemical environment, (2) detailed mechanistic studies, and (3) large-scale screening studies to predict active electrocatalysts.

This project will provide an in-depth fundamental and mechanistic understanding of the reaction processes critical to a sustainable carbon cycle, and to guide the rational design of highly active electrocatalysts.

Project Title: Polymer physics

Project Leader: Carsten Svaneborg, University of Southern Denmark, Denmark

Multi-year Proposal: 1st of 3 years

Resource Awarded

  • 43 000 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France

Multi-year Proposal: 1st of 3 years

Research Field: Chemical Sciences and Materials

Collaborators

  • Ali Karimi, Continental, Germany
  • Ralf Everaers, Ecole Normale Supérieure de Lyon, France

Abstract

Polymer materials are used in a vast number of technological and industrial applications and are also of fundamental scientific interest. We propose to generate reference simulation data for polymer liquids and networks. With these data, we will improve theories of polymer physics. For example, theories predicting the emergence of elastic and viscoelastic properties from the molecular structure and its dynamics. We will also use these data to predict results of advanced experiments, that are used to investigate polymer materials. These synthetic data will be analyzed as experimental data and used to improve experimental analysis methods. Tires are an important application of polymer materials. Abrasion from tire threads is the cause of about 30% of the to micro-plastic emissions to the environment. Tires consume 20-30% of the fuel due to rolling resistance. We will also simulate realistic models of tire materials. With these data we aim to provide a much improved understanding of the molecular scale physics of abrasion, and rolling resistance. In summary, the proposed project will both have a strong scientific, societal and technological impacts.

Note while I apply for time at the skylake partition below, the benchmarks shows nearly identical scaling behaviour for the rome partition. Hence this project equally well run on that partition, or split between them. Whatever is most convenient and effective for the Juliot-Curie admin.

Project Title: ClaTherm: Computational Discovery of Clathrate Thermoelectrics

Project Leader: David Scanlon, University College London, United Kingdom

Resource Awarded

  • 46 400 000 core hours on MareNostrum 4 hosted by BSC, Spain

Research Field: Chemical Sciences and Materials

Collaborators

  • David Scanlon, University College London, United Kingdom
  • Daniel Davies, University College London, United Kingdom
  • Maud Einhorn, University College London, United Kingdom

Abstract

The Paris Climate Agreement 2015 states that the energy efficiency of existing processes must be significantly increased to mitigate the most catastrophic impacts of climate change. It is estimated that 52% of global energy is lost to the environment as waste heat. Thermoelectric devices present an opportunity to harvest wasted thermal energy and yield electricity. Clathrates are a class of compounds which show promise as thermoelectric materials, exhibiting low thermal conductivities while maintaining favourable charge transport properties. The chemical space of clathrates is largely unexplored, with much of the experimental work related to binary clathrate materials and more recently some ternary analogues containing group III elements. This family of compounds have large unit cells, which makes them intractable to computational investigation from first principles without the use of large scale HPC facilities. Our proposed calculations will identify high-performance thermoelectric clathrates using ab initio techniques, focussing on 124 new clathrates containing group III and V elements, which have been hitherto unreported experimentally. We will investigate stability, electronic properties and thermal transport properties to thoroughly assess their thermoelectric potential.

The design of new inorganic clathrates for this application area is a grand challenge due to their compositional and structural diversity. There exists an enormous search space for new candidate materials: By using simple electron counting rules based on the Zintl concept, we estimate that over 1500 different valence-allowed compositions are possible for Type I clathrates alone. So, while there likely exist some highly promising candidate materials in search space, it is intractable to both high-throughput experiments and, until now, atomistic computations.

The aim of this project is to identify new clathrates with excellent thermoelectric performance via a high-throughput workflow of atomistic simulations and data-driven techniques. Specifically, we have used our materials informatics approach (SMACT) to screen through the search space using low-cost heuristic tools. We will carry out ab initio simulations on the subsequently refined search space in order to accurately determine (i) stability and (ii) charge transport and thermal properties crucial to thermoelectric performance.

Project Title: Simulation of the boundary of the TCV tokamak

Project Leader: Paolo Ricci, EPFL, Switzerland

Resource Awarded

  • 46 000 000 core hours on SuperMUC hosted by GCS at LRZ, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Christian Theiler, EPFL, Switzerland
  • Gilles Fourestey, EPFL, Switzerland
  • Maurizio Giacomin, EPFL, Switzerland
  • Davide Galassi, EPFL, Switzerland
  • Nicola Varini, EPFL, Switzerland
  • Emmanuel Lanti, EPFL, Switzerland
  • Nicolas Richart, EPFL, Switzerland

Abstract

We propose to perform first-of-a-kind simulations of the plasma dynamics in the boundary region of diverted discharges performed in the TCV tokamak. Based on a two-fluid model for the plasma and a kinetic model for the neutral particles, these simulations are enabled by key improvements of the GBS code. By using a state-of-the-art validation methodology and leveraging an unprecedented set of experimental measurements allowed by recent diagnostics upgrades, the simulations will be validated against TCV experimental results. Through dedicated parameter scans and the comparison with TCV, the simulations we propose will assess our understanding of the complex processes occurring in the tokamak boundary, such as the mechanisms that determine the heat flux on the vessel walls, the operation in detachment regimes, and the L-H mode transition. Ultimately, we will lay solid foundations to the development of predictive capabilities for key quantities, such as the exhaust heat flux width, the conditions necessary for entering in the detachment regime, and the L-H power threshold, that are of fundamental importance for ITER operation and future fusion devices design.

Project Title: WateR lUBrIcated chaNnel (RUBIN)

Project Leader: Alessio Roccon, TU Wien, Austria

Resource Awarded

  • 101 900 000 core hours on Hawk hosted by GCS at HLRS, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Alfredo Soldati, TU Wien, Austria
  • Francesco Zonta, TU Wien, Austria

Abstract

Turbulence friction causes a significant reduction of the flow-rate and a consequent increase of the pumping cost in a wide range of applications involving the transportation of high viscous fluids. Among the possible drag reduction (DR) techniques developed in recent years, water-lubricated pipelining has emerged as one of the most promising. This technique takes advantage of the natural tendency of water to migrate towards the wall and thus to lubricate the flow. In this project, we want to investigate numerically the performance of this technique performing large-scale simulations of a turbulent channel in which two near-wall layers of water lubricate the core-flow of oil. The simulations will adopt an innovative approach based on direct numerical simulations of turbulence coupled with a phase-field method to describe the complex dynamics of the system, which is governed by the interplay of phenomena occurring on a wide range of spatial and temporal scales. The accurate description of these phenomena requires the adoption of high-resolution grids and thus large high-performance computing infrastructures are needed.

Project Title: Diffusion Monte Carlo calculation of the Fermi liquid parameters of the three-dimensional homogeneous electron gas

Project Leader: Sam Azadi, King’s College London, United Kingdom

Resource Awarded

  • 55 000 000 core hours on Hawk hosted by GCS at HLRS, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • William Matthew Foulkes, Imperial College London, United Kingdom
  • Tom Yates, Imperial College London, United Kingdom
  • Christopher Bradley, Imperial College London, United Kingdom

Abstract

Electronic systems are at the centre of many of the most important and topical areas of applied science, including materials physics, chemistry, nano-electronics, and quantum computation. The electron liquid paradigm is the foundation of most of our current understanding of the physical and chemical properties of these systems.

The purpose of this project is to deploy stochastic methods to solve the quantum mechanical Schrödinger equation that governs the fundamental properties of the three-dimensional quantum electron liquid. Stochastic approaches have fundamental advantages when applied to quantum mechanical systems with many degrees of freedom because they scale much better with system size than comparably accurate deterministic approaches. Another advantage is that the well-established mathematical theories of probability and stochastic processes provide bounds on the errors in the properties calculated and help us devise methods to reduce those errors. In our simulations, the quantum mechanical electrons move along stochastic trajectories, and expectation values are formulated as ensemble averages over the space of these trajectories. In practice, we exploit the similarity between the Schrödinger equation in imaginary time, which is a linear, parabolic partial differential equation, and the diffusion equation.

Project Title: The effect of roughness on superstructures in sheared Rayleigh-Bénard

Project Leader: Detlef Lohse, University of Twente, Netherlands

Resource Awarded

  • 71 000 000 core hours on MareNostrum 4 hosted by BSC, Spain

Research Field: Chemical Sciences and Materials

Collaborators

  • Roberto Verzicco, University of Tor Vergata, Italy
  • Richard Stevens, University of Twente, Netherlands
  • Chong Ng, University of Twente, Netherlands
  • Kai Chong, University of Twente, Netherlands
  • Christopher Howland, University of Twente, Netherlands
  • Martin Assen, University of Twente, Netherlands
  • Qi Wang, University of Twente, Netherlands
  • Robert Hartman, University of Twente, Netherlands
  • Guru Sreevanshu Yerragolam, University of Twente, Netherlands
  • Giulia Piumini, University of Twente, Netherlands
  • Naoki Hori, University of Twente, Netherlands
  • Rui Yang, University of Twente, Netherlands

Abstract

Sheared Rayleigh-Bénard convection is the canonical model problem to study the interplay between buoyancy and shear in turbulent thermal convection. Studying this system is relevant to get a better fundamental insight into natural instances of convections, such as atmospheric flows and ocean currents, which are often dominated by large-scale flow structures known as turbulent superstructures. Besides, sheared thermal convection is relevant for the optimization of many industrial applications.

Detailed studies on the characterization of turbulent superstructures, and their effect on the heat transfer, exist for classical Rayleigh-Bénard. However, while previous studies on sheared Rayleigh-Bénard revealed the development of beautiful and intriguing flow structures, the properties of these flow structures are not carefully studied. Here we plan to study the effect of these superstructures, as characterized by their energy and coherence spectra, on the heat transfer and wall friction coefficient. In particular, we want to see what changes in the flow structures are responsible for the surprising minimum heat transfer observed for intermediate shear rates. Furthermore, we want to study the effect of surface roughness on the heat transfer, flow structures, and wall friction coefficient in sheared Rayleigh-Bénard convection.

Project Title: Electronic Structure and response functions in semimetals Cd3As2 and TiSe2

Project Leader: Mark van Schilfgaarde, King’s College London, United Kingdom

Resource Awarded

  • 9 000 000 core hours on JUWELS Booster hosted by GCS at FZJ, Germany
  • 800 000 core hours on JUWELS hosted by GCS at FZJ, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Swagata Acharya, Raboud University, Netherlands
  • Mikhail Katsnelson, Raboud University, Netherlands
  • Mark van Schilfgaarde, King’s College London, United Kingdom
  • Jack Gahan, King’s College London, United Kingdom
  • Dimitar Pashov, King’s College London, United Kingdom

Abstract

In this project we will use the Quasiparticle Self-Consistent GW (QSGW) approximation within the Questaal package (questaal.org) to investigate the complex effects impurities and disorder, induce in the properties of two classes of semimetals: Dirac/Weyl semimemtals (for which Cd3As2 is the representative material) and TiSe2, a member of transition metal dichalcogenide family, MX2, which have weak bonding between planes and bear similarities with other well known 2D materials.

Cd3As2 can be profoundly changed by even modest disorder: in the Dirac semimetals, the symmetry can be disrupted by point defects, especially ones that break time-reversal symmetry. Symmetry breaking is also of central importance in TiSe2 : it forms a charge-density wave at low temperature. If the CDW can be suppressed, superconductivity can appear.

Studies of these materials are every challenging: very high fidelity is needed with small bandgaps and large dielectric response. QSGW is a gold standard for fidelity in electronic structure, but until recently it was too heavy a method and precluded studies of large systems requiring large unit cells to model disorder. However, a recent implementation with multitier parallelization and GPU cards makes these calculations possible on high performance machines with advanced GPU cards.

We will look at the role of disorder on symmetry-breaking in part through changes in the the band struture in both these materials. Further, we will develop a novel way of combining the lattice disorder into an effective self-energy, and use our QSGW++ implementation to study supercondutivity in TiSe2 with an effective potential that includes both nulclear disorder and electron scattering.

Project Title: QSGW study of complex magnetism in rare-earth garnets

Project Leader: Joe Barker, University of Leeds, United Kingdom

Resource Awarded

  • 41 600 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Chemical Sciences and Materials

Collaborators

  • Jerome Jackson, Science and Technology Facilities Council, United Kingdom
  • Dimitar Pashov, King’s College London, United Kingdom

Abstract

Predictive, highly accurate ab initio electronic structure calculations will be carried out on the series of rare-earth garnets leading to a detailed description of the fundamental electronic and magnetic properties. The nature of f-electrons in these materials will be revealed and the complex interaction between f-electron moments on rare-earth sites with d-electron moments will be made clear. By employing quasiparticle self-consistent GW, a gold standard for high accuracy electronic structure calculations, we will be able to provide an accurate and rich description of these materials, including information that cannot be obtained experimentally. These calculations are truly ab initio (parameter free) but require HPC resources to tackle complex materials with such large unit cells. We will answer long standing questions about the nature of magnetism in these materials by linking details, such as the localisation and energy range of f-electron bands, their crystal field splittings and exchange interactions, of the electronic structure to the resulting magnetic behaviour. By constructing Heisenberg models for these materials, we will create a robust description of finite temperature magnetism, including the (temperature dependent) magnon spectrum. This investigation represents a significant step forward in our basic understanding of this class of materials and will provides a microscopic description that will be of significant value to the many researchers in spintronics, magnonics and magneto-optics.

Project Title: PropHeTAM

Project Leader: Evelyne Martin, ICube, France

Resource Awarded

  • 27 600 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Research Field: Chemical Sciences and Materials

Collaborators

  • Mauro Boero, IPCMS, France
  • Guido Ori, IPCMS, France
  • Carlo Massobrio, IPCMS, France

Abstract

A deep understanding of the thermal behavior of materials is the key to limit generation of waste heat, reduce energy consumption, increase efficiency. In crystals, reduction of thermal conductivity at the nanoscale benefits thermoelectricity but damage electronic devices. Up to now, size effects in amorphous materials have only been confirmed in amorphous silicon. Numerous amorphous materials are used as key components of non-volatile memories, photovoltaic cells, thermoelectric devices, batteries. We have recently developed an atomic-scale approach to assess thermal conductivities using considerably shorter trajectories than within more traditional methods. This breakthrough enabled the use of first-principles molecular dynamics (FPMD) to calculate the thermal conductivity of three amorphous systems (GeTe4, Ge2Sb2Te5, SiO2). In each case, the results were in quantitative agreement with the measurements. They also reveal a reduction of thermal conductivity at small size related to propagative modes on distances well above short or intermediate structural range order. The present project aims at determining whether the existence of propagative modes in amorphous materials is a universal feature and to which extent. We further aim at quantifying sizes down to which a significative (>50%) reduction of thermal conductivity occurs. The required FPMD trajectories call for high performance computing, as justified below.

Project Title: The function, mechanism and inhibition of Two-Pores intracellular Channels

Project Leader: Matteo Ceccarelli, University of Cagliari, Italy

Resource Awarded

  • 40 200 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Chemical Sciences and Materials

Collaborators

  • Stefan Milenkovic, University of Cagliari, Italy
  • Igor Bodrenko, CNR, Italy
  • Velia Minicozzi, University La Sapienza, Italy

Abstract

Intracellular channels are key elements of cellular functions and diseases. New experimental techniques to probe their functions and the availability of new high resolution structures are transforming them in attractive targets for searching pharmacological inhibitors.

Two pore channels (TPCs) are cation-selective channels expressed in cellular organelles of plants and mammals and regulate trafficking and the endo-lysosome system in cells. Thus they are of primary importance for viral infections. Indeed viruses make use of the cell machinery to enter, infect the cell itself and exit to spread out, by using the endo-exocytosis pathway. Genetic ablation or TPC blockers have been previously shown to reduce viral infections in EBOLA, Mers-CoV and very recently SarS-CoV-2 infections. Though the development of a vaccine seems the obvious priority for stopping today’s COVID-19 pandemy, the development of inhibitors still represents another promising and independent option to fight virus infections. Further, TPC is also key for Parkinson, mast cell-related diseases, and anti-angiogenic therapeutic strategies, representing a very interesting and unexplored target.

The mechanism under which TPCs operate remains unknown and debated in literature. Our previous experience suggests that the cornerstone of TPCs functions is the control of the narrow constriction region of the channel, the hydrophobic gate. Moreover, the mechanism of Ca2+ transport by TPCs in mammals remains unrevealed. Here we want to exploit computational power of HPC and investigate the function, mechanism and eventually inhibition of human TPC2 via molecular simulations and docking. The results obtained will shed light on possible fighting strategies against the COVID19 and also to have methodological significance to better understand ion transport in TPCs.

Project Title: Multiphase Rayleigh–Bénard convection

Project Leader: Luca Brandt, Royal Institute of Technology (KTH), Sweden, Sweden

Resource Awarded

  • 51 400 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Chemical Sciences and Materials

Collaborators

  • Andreas Demou, Royal Institute of Technology (KTH), Sweden
  • Marco Crialesi Esposito, Royal Institute of Technology (KTH), Sweden
  • Nicolo Scapin, Royal Institute of Technology (KTH), Sweden
  • Dimokratis Grigoriadis, University of Cyprus (UCY), Cyprus
  • Roberto Verzicco, Università  di Roma “Tor Vergata”, Italy

Abstract

The thermally driven flow inside a fluid layer heated from below and cooled from above, known as Rayleigh-Bénard convection, is a widely studied physical problem due to its similarities with a range of real-life applications and physical phenomena. Even in its simplest form, the complexity of the flow increases rapidly with the Rayleigh number, and the resolution requirements become overwhelming. Moreover, when additional complexities are included in the configuration, such as two fluid layers and variable thermophysical properties, the numerical solution becomes extremely challenging. The present proposal aims to simulate, for the first time, the two-layer Rayleigh-Bénard convection with variable properties, in a three-dimensional configuration. To carry out such demanding simulations, the very efficient, massively parallel numerical tool CaNS with state-of-the-art capabilities will utilize the full potential of the accelerated system MARCONI100. The results will focus on the characterization of the variable property effects on the main flow features, in addition to assessing the validity of reduced two-dimensional simulations. Overall, this study will provide a more relevant frame of reference to geophysical and other engineering applications, where the multiphase nature of the flow and the property variations cannot be ignored.

Project Title: A Molecular Dynamics Study of Ionic Thermoresponsive Fluids for Desalination by Forward Osmosis

Project Leader: Robinson Cortes Huerto, Max Planck Institute for Polymer Research, Germany

Resource Awarded

  • 32 700 000 core hours on JUWELS hosted by GCS at FZJ, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Pietro Ballone, University College Dublin, Ireland
  • Nancy Carolina Forero Martinez, Max Planck Institute for Polymer Research, Germany
  • Luis A. Baptista, Max Planck Institute for Polymer Research, Germany
  • Ravi C. Dutta, Max Planck Institute for Polymer Research, Germany

Abstract

Among the most promising materials being considered for this process are organic salts of low melting temperature, generally known as room temperature ionic liquids (RTILs). RTILs represent a vast family of compounds, spanning a wide range of properties, and our aim is to contribute to the choice of suitable compounds optimising the energetic, economic and environmental aspects of the problem. The first requirement, however, is that the draw solution has to demix with increasing temperature, or, in other terms, has a lower critical solution temperature (LCST) at T<100 C. A further requirement is that demixing is as complete as possible, i.e., ends up into nearly pure salt and water.

The basic task we pose for ourselves is the computation of the thermodynamic properties for a series of RTIL compounds known or expected to have an LCST when mixed with water. Examples are several RTIL compounds based on the phosphonium cation, combined with a variety of organic or inorganic anions. In these systems, demixing takes place at a critical temperature through a continuous transition. The LCST of various phosphonium RTILs is in the range of 50-60 C, achievable for instance by solar heating, and fully suitable for desalination even in remote locations.

The computational project aims at: (1) characterising the structural, thermodynamic and kinetic properties of the mixed (low temperature) and demixed (higher temperature) phases; (2) simulating the demixing transition. Since this transformation is continuous, fluctuations are expected to be very important, hence large sizes (~25 nm, corresponding to ~ one million atoms) and correspondingly long simulation times (up to 1 microsecond) are required.

Project Title: DAM-DeLoMaMaL – Defects in austenitic materials: from DFT/DLM to ML potentials

Project Leader: Christophe Domain, EDF R&D, France

Resource Awarded

  • 23 700 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Research Field: Chemical Sciences and Materials

Collaborators

  • Julien Vidal, EDF R&D, France
  • James Kermode, University of Warwick, United Kingdom
  • Charlotte Becquart, Lille University, France
  • Lakshmi Shenoy, University of Warwick, United Kingdom
  • Adithya Nair, Lille University, France
  • Antoine Michel, EDF R&D, France
  • Gilles Adjanor, EDF R&D, France
  • Miroslav Fokt, EDF R&D, France

Abstract

Under extreme conditions (e.g. irradiation), the mechanical properties of structural materials evolve and the natural ageing of materials can be altered and/or accelerated significantly. The microstructure evolution is governed by the formation of defects and the evolution of defect clusters and their interaction with the solute atoms in the material. Microstructure modelling predictions performed at the mesoscale are strongly dependent on these atomistic phenomena. A multi-scale modeling scheme is developed to simulate the ageing of the materials in order to better predict their behavior under irradiation in nuclear power plants.

The objective of this project is to contribute to the multi-scale modeling of ageing of austenitic stainless steels used in nuclear power plants, subject to irradiation assisted stress corrosion cracking (IASCC).

The objective of this project is to better characterize the ageing mechanisms using state of the art atomic-scale calculations (Density Functional Theory – DFT) that require very large simulations on HPC machines.

The project will focus on many configurations of FeCrNi concentrated alloys, by studying different defects and defect clusters with static and dynamic calculations. The paramagnetic properties of FeCrNi alloys will be treated within the Disorder Local Moment (DLM) model.

The main outcomes of the project will be: i) an open access database of DFT calculations, ii) the first machine learning potential for the FeCrNi alloy which will be used to estimate irradiation primary damage (for estimation of end-of-life irradiation microstructure), model diffusion of irradiation defects (for estimation of swelling), stability of dislocation loops and dislocation fragments (for estimation of creep), interaction between dislocation and irradiation defects (for estimation of hardening and of IASCC), iii) contribution for Quantum Computing of paramagnetism.

The VASP code, one of the state of the art DFT codes for accurate calculations in metals with good scalability on HPC machines, will be used. These new calculations will allow an extended characterization of defect formation properties and plasticity consequences to improve the quality of the microstructure prediction of FeNiCr alloys. This work will contribute to increasing the precision and robustness of computational models used in the multi-scale modelling of structural materials with the challenge of bridging the gap between different length and time scales.

Project Title: BIG Turbulence

Project Leader: Alexandros Alexakis, Ecole Normale Supérieure, France

Resource Awarded

  • 40 000 000 core hours on Joliot-Curie (KNL) hosted by GENCI at CEA, France

Research Field: Chemical Sciences and Materials

Collaborators

  • Pablo D. Mininni, University of Buenos Aires, Argentina
  • Raffaele Marino, École Centrale de Lyon, France

Abstract

Geophysical and planetary flows display a magnificent pallet of coexisting behaviours that vary from the appearance of large scale organized structures such as jets and hurricanes; to small scale turbulence. In recent years it has been shown that this diversity of flow behaviour is in part linked to the presence of a bidirectional turbulent cascade where part of the energy is transferred to small scales similar to a classical three dimensional turbulence and part of the energy is transferred to the large scales like two dimensional turbulence. This hybrid behaviour between three and two dimensional turbulence is due to the strong anisotropy imposed in the planetary flows by rotation, stratification and the intrinsic domain anisotropy. Indications of such bidirectional cascades have been observed in field measurements. Numerical simulations of geophysical flows were able to demonstrate the presence of forward and inverse cascades but limitations in resolution prevent them from reaching desired parameter regimes and establishing clear inertial ranges. In this project we propose to perform unprecedented numerical simulations of rotating and stratified turbulence in an anisotropic domain, indisputably establishing the presence of bidirectional cascades in geophysical flows and providing a link to compare with observations in planetary flows.

Project Title: Solvation and Charge Transfer Processes at Semiconductor/Liquid Water Interfaces

Project Leader: Philipp Schienbein, University College London, United Kingdom

Resource Awarded

  • 40 000 000 core hours on Hawk hosted by GCS at HLRS, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Philipp Schienbein, University College London, United Kingdom
  • Jochen Blumberger, University College London, United Kingdom

Abstract

Interfaces between metal oxides and liquid water play a crucial role in numerous systems from heterogeneous catalysis to biological processes. Especially charge transfer processes across the interface are of great importance because the oxide can catalyse reactions in the adjacent liquid phase. A promising application is photocatalytic water splitting. Here, photons are absorbed by an oxide semiconductor and their energy is used to split water into hydrogen and oxygen. So far, relatively little is known about the exact microscopic processes, e.g. how water molecules interact with the semiconductor or how charges are localized. These properties are accessible by ab initio molecular dynamics simulations which are capable to realistically model the structural dynamics of an oxide/water interface. Since the complex electronic structure is still demanding for modern HPC facilities, we utilize recent Machine Learning techniques to accelerate the simulations. The goal of this project is to gain a deep understanding of the structure, dynamic and reactivity of liquid water at an oxide material using the example of BiVO4. The results are generalized to make predictions for other oxide/water interfaces. Moreover, the simulations complement existing experimental results and are thus an integral contribution to understand the latter on a microscopic level.

Project Title: CoSuFA – Computational Design of Surfactants for Flow Assurance

Project Leader: Stephan Mohr, Nextmol (Bytelab Solutions SL), Spain

Resource Awarded

  • 2 000 000 core hours on JUWELS Booster hosted by GCS at FZJ, Germany

Research Field: Chemical Sciences and Materials

Collaborators

  • Rémi Pétuya, Nextmol (Bytelab Solutions SL), Spain
  • Pablo Navarro, Nextmol (Bytelab Solutions SL), Spain
  • Ioannis N. Tsimpanogiannis, Centre for Research & Technology Hellas, Greece
  • Monica de Mier, Nextmol (Bytelab Solutions SL), Spain

Abstract

Surfactants are molecules used to alter interfacial properties. They play an important role in many industrial processes and products such as oil extraction, cosmetics, detergents, fabric softeners, lubricants, and many more, acting for instance as anti-agglomeration, dispersing, cleaning, wetting, emulsifying, foaming, or lubricating agents. Designing new and better surfactants is crucial to increase the efficiency of industrial processes and products and to enhance environmental friendliness (“green chemistry”). In spite of this importance, surfactants are still designed mainly using an experimental-based trial-and-error approach in the lab. However, this method has become too slow, costly and inefficient. Thanks to the ongoing digital transformation, computer simulations have become an attractive alternative. They allow to design new surfactants on the computer, which is not only faster, cheaper and more efficient, but also gives additional insights. With the CPU-hours asked for in this proposal we would like to contribute to this digitalization of surfactant development. The ultimate goal is to determine the characteristics and performance of a surfactant in silico, i.e. without the need to synthesize and test it in a laboratory. We will focus on one specific use case, namely hydrate anti-agglomeration additives, which is of paramount importance for the Oil & Gas industry.

Earth System Sciences

Project Title: The Quasi-Biennial Oscillation in a changing climate (QUBICC)

Project Leader: Marco Giorgetta, Max Planck Institute for Meteorology, Germany

Multi-year Proposal: 2nd of 2 years

Resource Awarded

  • 4 000 000 core hours on JUWELS Booster hosted by GCS at FZJ, Germany

Multi-year Proposal: 2nd of 2 years

Research Field: Earth System Sciences

Collaborators

  • Marco Giorgetta, Max Planck Institute for Meteorology, Germany
  • Luis Kornblueh, Max Planck Institute for Meteorology, Germany
  • Ulrich Achatz, Goethe University Frankfurt, Germany
  • Manfred Ern, Forschungszentrum Jülich, Germany

Abstract

The tropical quasi-biennial oscillation (QBO) is one of the most prominent dynamical phenomena in the stratosphere. The theory stipulates that wave-meanflow interaction between vertically propagating waves and zonal jets creates the downward propagating easterly and westerly jets of the QBO. Existing simulations of the QBO in general circulation models (GCMs) rely on the parametrized convective heating as a source for resolved tropical waves and gravity wave parametrizations for sub grid scale gravity wave drag. Recent studies showed that the uncertainty originating from the parametrizations and their tuning effectively hinders the understanding of the full QBO cycle in the current climate and consequently obstructs the assessment of climate change effects on the QBO. We therefore propose a first direct simulation of the QBO in a deep convection resolving GCM that by construction is independent of parametrizations for convection and gravity waves. By comparison of analyses and the direct QBO simulations for current and future climate conditions we expect to understand the key factors that can change the QBO, and thus to overcome the impasse from the parametrized GCMs.

Project Title: Volcanic ash hazard and forecast

Project Leader: Arnau Folch, Barcelona Supercomputing Center (BSC), Spain

Multi-year Proposal: 2nd of 2 years

Resource Awarded

  • 5 300 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France

Multi-year Proposal: 2nd of 2 years

Research Field: Earth System Sciences

Collaborators

  • Leonardo Mingari, Barcelona Supercomputing Center (BSC), Spain
  • Sara Barsotti, Icelandic Meteorological Office (IMO), Iceland
  • Manuel Luzón, Icelandic Meteorological Office (IMO), Iceland
  • Laura Sandri, Istituto Nazionale di Geofisica e Vulcanologia (INGV), Italy
  • Antonio Costa, Istituto Nazionale di Geofisica e Vulcanologia (INGV), Italy
  • Beatriz Montesinos, Istituto Nazionale di Geofisica e Vulcanologia (INGV), Italy
  • Jacopo Selva, Istituto Nazionale di Geofisica e Vulcanologia (INGV), Italy
  • Matteo Cerminara, Istituto Nazionale Geofisica e Vulcanologia (INGV), Italy
  • Federico Brogi, Istituto Nazionale Geofisica e Vulcanologia (INGV), Italy
  • Tomaso Esposti-Ongaro, Istituto Nazionale Geofisica e Vulcanologia (INGV), Italy

Abstract

Many parts of Europe are posed to volcanic hazards that can impact on regions close to the volcano (e.g. tephra fallout at Naples with 3 million people at risk from Vesuvius and Campi Flegrei) or even at continental scale (e.g. impacts from ash clouds on civil aviation like the massive shutdown during the 2010 Eyjafallajökull eruption in Iceland). Forecasting what will occur in the next hours when a volcano is erupting or quantifying potential impacts from a future eruption are relevant issues to aviation stakeholders and to civil protection agencies and governmental bodies. HPC plays a major role on making forecasts compatible with the time-space constraints of aircraft operations (emergency management scenarios and related urgent computing) and to perform physically-based modelling approaches, thereby reducing uncertainties on impacts and related economic loss estimations. This multi-year project aims at using HPC to increase the resolution of current operational model configurations by one order of magnitude and at overcoming the current limits of high-resolution physics-based Probabilistic Volcanic Hazard Assessments (PVHA). Outcomes will be shared with aviation and civil protection authorities in Italy and Iceland. High quality videos will be produced to disseminate results among the general public and potential users.

Project Title: (An)elastic Global Full-Waveform Inversion

Project Leader: Daniel Peter, King Abdullah University of Science and Technology (KAUST), Saudi Arabia

Resource Awarded

  • 48 400 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Earth System Sciences

Collaborators

  • Ebru Bozdag, Colorado School of Mines, United States of America
  • Emanuele Casarotti, INGV Rome, Italy

Abstract

High-resolution seismic images are essential to understand the structure and thermochemical composition of the mantle to interpret its dynamics, which directly control surface processes such as earthquakes and volcanoes. Seismic tomography is at a stage where further refinements require the use of full physics of wave propagation. Adjoint tomography efficiently takes advantage of 3D wave simulations leading to pure data-driven seismic models by avoiding commonly used approximations and corrections in classical tomography. After the publication of the first-generation global adjoint models, which are elastic and transversely isotropic in the upper mantle, constructed based on only traveltimes, our goal is to focus on a new global anelastic mantle model by the simultaneous inversion of anelastic and elastic parameters based on adjoint tomography including amplitudes of waveforms. To that end, we aim to complete a new azimuthally anisotropic global adjoint model with the requested allocation. As anelasticity causes physical dispersion, the new elastic/anelastic model also allows for locating earthquakes and other seismic sources more accurately. This will result in a much improved Earth model with drastically sharper mantle images attempting to answer long- standing questions on the origin of plumes, hotspots and the water content of the upper mantle.

Project Title: ImMEDIAT

Project Leader: Camille Lique, Ifremer, France

Resource Awarded

  • 35 000 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Research Field: Earth System Sciences

Collaborators

  • Claude Talandier, CNRS, France
  • Tina Odaka, Ifremer, France
  • Yann Meurdesoif, CEA, France

Abstract

In the global ocean, mesoscale eddies are ubiquitous. They account for most of the turbulent kinetic energy and are key to the long-term equilibrium of the large-scale circulation, ocean ventilation of tracers, upper-ocean biology and pollutant dispersion. In the Arctic Basin, however, observations taken under sea ice have pointed out that the energy at mesoscale in the Arctic interior is relatively low when compared to characteristic midlatitude open ocean dynamics. At first order, the mesoscale activity could be fundamentally different between ice-covered and ice-free regions.

Owing to the challenge of modelling at such small deformation radius (less than 10 km in the Arctic), there is currently no existing eddy resolving models of the Arctic Basin, preventing us from fully apprehending the specificity of the mesoscale features in the Arctic, and their importance for the Arctic dynamics. This is particularly important as interactions between mesoscale eddies and sea ice could potentially represent an important, yet currently ignored, mechanism, via which the ocean might contribute to the on-going and future sea ice retreat.

In this context, the overarching goal of ImMEDIAT is to improve the quantification and our fundamental understanding of the mesoscale dynamics in the presence of sea ice. To that aim, we will make use of the newly developed ocean-sea ice configuration SEDNA (Sea ice – EDdy resolving ocean paN-Arctic configuration), based on the NEMO modelling platform and with a horizontal resolution of less than a kilometre in the Arctic Basin. We will run and analyze simulations covering periods with very different seasonal cycle of the sea ice conditions. These pioneer simulations will allow to characterise all processes important for the interplay between the sea ice and the mesoscale eddy in the Arctic, to test theory related to these processes and quantify their effects on the large scales. We will be able to determine for the first time if the mesoscale dynamics is fundamentally different in the ice-cover regions and in open ocean and explore ways of parameterizing them for climate-scale ocean models that are unlikely to resolve mesoscale features in the polar regions in the near future.

Project Title: Kilometer-resolution climate modeling on GPUs (kmCLIM2)

Project Leader: Christoph Schär, ETH Zurich, Switzerland

Resource Awarded

  • 68 000 000 core hours on Piz Daint hosted by CSCS, Switzerland

Research Field: Earth System Sciences

Collaborators

  • Roman Brogli, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Jacopo Canton, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Marie-Estelle Demory, Institute for Atmospheric and Climate Science, ETH Zurichh, Switzerland
  • David Leutwyler, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Shuping Li, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Christian Steger, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Jesùs Vergara-Temprado, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Matthieu Leclair, Institute for Atmospheric and Climate Science, ETH Zurich- Switzerland
  • Ruoyi Cui, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Christoph Heim, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Shuchang Liu, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Daniel Regenass, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Ruolan Xiang, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland
  • Christian Zeman, Institute for Atmospheric and Climate Science, ETH Zurich, Switzerland

Abstract

There is agreement in the scientific community regarding the important role of man-made greenhouse gases in our climate system. However, while observations and climate models qualitatively agree for the past, there are large uncertainties for the future. One of the key modelling challenges is the representation of convective precipitation (e.g. rain showers) and convective clouds (e.g. cumulus). Due to the lack of adequate computational resolution, conventional climate models are unable to explicitly represent convective motions and use semi-empirical cloud parameterization schemes. Major efforts are underway towards refining the horizontal grid spacing of global and regional climate models to about 1 km. This development opens exciting prospects. In particular, it enables the explicit representation of deep convective clouds, and allows for a more adequate representation of precipitation systems and extreme events. In the current project, we exploit a continental-scale climate modeling capability at a horizontal resolution of about 2 km. This resolution is 10 to 50 times higher than in conventional models. The simulations are conducted with a version of the COSMO model, which is able to run efficiently on emerging hardware architectures using Graphics Processing Units. The respective climate modeling capability has been developed in a previous PRACE project, and has been used to conduct the first decade-long European climate simulation at km-scale. Research will address three key issues of climate change: (1) Over Europe, we will investigate potential changes in the occurrence of thunderstorms and address severe weather events (e.g. flash floods, wind storms, lightning, hail). The research will support the generation of physically-informed projections of the future occurrence of these extremes, and will provide guidance for impact assessment and climate change adaptation strategies. (2) Over the tropical Atlantic, we will address future changes in cloud cover to reduce the uncertainties in global climate projections relating to cloud-radiative feedbacks. The research will assess to what extend cloud cover may decrease or increase in response to climate change (thereby amplify or moderate global warming, respectively). (3) Over Asia, we will address the simulation of Monsoon circulation, which is an essential source of precipitation for more than 1.5 billion people. Results indicate that km-scale models have an excellent potential for these three purposes.

Engineering

Project Title: NanoGAS – Gas-induced drying of nanopores: From biology to chromatography

Project Leader: Alberto Giacomello, Sapienza University of Rome, Italy

Multi-year Proposal: 1st of 2 years

Resource Awarded

  • 48 000 000 core hours on Marconi100 hosted by CINECA, Italy

Multi-year Proposal: 1st of 2 years

Research Field: Engineering

Collaborators

  • Gaia Camisasca, Roma Tre University, Italy
  • Antonio Tinti, Sapienza University of Rome, Italy
  • Alberto Gubbiotti, Sapienza University of Rome, Italy
  • Emanuele Telari, Sapienza University of Rome, Italy
  • Gonçalo Paulo, Sapienza University of Rome, Italy

Abstract

The kinetics drying in nanopores — the formation of a vapor phase from the liquid one in nano-scale confinement — can be dramatically modified by the presence of hydrophobic solutes such as poorly soluble gases: even a single hydrophobic gas atom can decrease and, in some cases, abate the drying free-energy barrier, causing the instantaneous drying of the nanopore.

Gas-induced drying of nanopores plays a fundamental role in different scenarios. In biology, some kinds of ion channels have hydrophobic gates (HG), where the flux of ions or other molecules through the channels is blocked by the formation of a vapor phase; these channels are often targets of volatile anesthetics which facilitate drying of the HG. In technology, the performance of many devices is deteriorated by drying of nanopores: gases can alter the functioning of heterogeneous lyophobic systems, while drying decreases the performance of reversed-phase liquid chromatography (RPLC) columns.

NanoGAS aims at elucidating the gas-induced drying process taking place in nano confinement, revealing how the type of volatile drugs, the geometrical/constructive parameters of nanopores, and operating conditions influence this process. The ensuing fundamental understanding will allow the design of new anesthetic drugs and procedures as well as the rational design of efficient devices whose functioning relies on the wetting/drying properties of nanopores. A large and multidisciplinary community is therefore expected to benefit from the results obtained by the results of NanoGAS.

NanoGAS will leverage specialized rare-events molecular dynamics simulations to reveal the mechanisms and assess the kinetics of gas-induced drying in nanopores. During two years, NanoGAS will focus on drying induced by different gases (Year 1) and on the effects of the pore characteristics – surface chemistry, pore size and connectivity (Year 2).

Summarizing, NanoGAS aims at understanding in quantitative terms the physics of the gas-induced drying of nanopores, which has important biological and technological implications. To achieve these ambitious fundamental and technological goals, PRACE allocation is required to guarantee the necessary Tier-0 computational resources.

Project Title: Hydrogen to Leverage Ultra-Lean Combustion in Internal Combustion Engines

Project Leader: Frédéric Ravet, Renault, France

Resource Awarded

  • 5 300 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France
  • 9 700 000 core hours on Joliot-Curie (SKL) hosted by GENCI at CEA, France

Research Field: Engineering

Collaborators

  • Quentin Malé, CERFACS, France
  • Olivier Vermorel, CERFACS, France
  • Thierry Poinsot, IMFT, France

Abstract

While the global demand for transport energy is large and is increasing, improving Internal Combustion Engine (ICE) is of major importance because transport almost entirely relies on ICE burning petroleum-derived fuel. Lean combustion is sought in many modern designs for its low emissions and high energy efficiency. However, lower burning velocities and harder ignition prevent classical spark ignition engines to be operated in very lean regimes: spark ignition of lean mixture causes erratic combustion, misfires and partial burns.

Experiments show that pre-chamber ignition systems can pave the way towards lean combustion in ICE. These systems produce multiple high energy ignition and turbulence sources which result in a reliable ignition and a fast combustion of the main charge. If combined with hydrogen, extreme air dilution levels can be achieved.

However, aerothermochemical processes need to be investigated and well understood to design the pre-chamber properly. Only kinetically detailed Large Eddy Simulations (LES) can be used to study the complex unsteady phenomena at play in real engines. However, these simulations are very costly, which motivates the present proposal.

Project Title: Direct numerical simulation of flow in rough pipes

Project Leader: Paolo Orlandi, Sapienza University of Rome, Italy

Resource Awarded

  • 50 000 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Engineering

Collaborators

  • Sergio Pirozzoli, Sapienza University of Rome, Italy
  • Roberto Verzicco, University of Twente, Netherlands
  • Massimiliano Fatica, NVIDIA Corporation, United States of America
  • Josh Romero, NVIDIA Corporation, United States of America

Abstract

This research project aims at studying turbulent flow in a rough pipe through direct numerical simulation (DNS), all the way from the smooth to the transitional and fully rough regimes. Computer resources granted by PRACE will be exploited to explore for the first time the range of fully rough flow over small roughness elements (R/k > 40), in a full pipe configuration, as in most key experiments. Numerical simulations will also include the solution of the transport equation for a passive scalar to model heat and mass transfer processes. Studying rough pipe flow at high Reynolds number allows the observation of physical phenomena of interaction between the near-wall and the outer layers, and detailed information about vortex dynamics near the roughness elements. We also expect to shed additional light into the validity of Tonwsend’s outer-layer similarity assumption, with important reflection on the development of wall models for rough surfaces. Improved understanding of the flow physics of rough flows at high Reynolds number may yield improvement in the prediction of drag and heat transfer in a huge number of technological applications, whose error bars are still quite large, with incurred large benefit in terms of energy saving.

Fundamental Constituents of Matter

Project Title: Breaking the Strong Interaction: Towards Quantitative Understanding of the Quark-Gluon Plasma

Project Leader: Chris Allton, Swansea University, United Kingdom

Multi-year Proposal: 3rd or 3 years

Resource Awarded

  • 35 000 000 core hours on Hawk hosted by GCS at HLRS, Germany

Multi-year Proposal: 3rd or 3 years

Research Field: Fundamental Constituents of Matter

Collaborators

  • Gert Aarts, Swansea University, United Kingdom
  • Simon Hands, Swansea University, United Kingdom
  • Benjamin Jäger, University of Southern Denmark, Odense, Denmark
  • Michael Peardon, Trinity College, Dublin, Ireland
  • Jon-Ivar Skullerud, National University of Ireland, Maynooth, Ireland

Abstract

There are four fundamental forces that describe all known interactions in the universe: gravity; electromagnetism; the weak interaction (which powers the sun and describes most radioactivity); and, finally the strong interaction – which is the topic of this research. The strong interaction causes quarks to be bound together in triplets into protons and neutrons, which in turn form the nucleus of atoms, and therefore make up more than 99% of all the known matter in the universe. If there were no strong interaction, these quarks would fly apart and there’d be no nuclei, and therefore no atoms, molecules, DNA, humans, planets, etc. Although the strong interaction is normally an incredibly strongly binding force (the force between quarks inside protons is the weight of three elephants!), in extreme conditions it undergoes a substantial change in character. Instead of holding quarks together, it becomes considerably weaker, and quarks can fly apart and become “free”. This new phase of matter is called the “quark-gluon” plasma. This occurs at extreme temperatures: hotter than 10 billion Celsius. These conditions obviously do not normally occur – even the core of the sun is one thousand times cooler! However, this temperature does occur naturally just after the Big Bang when the universe was a much hotter, smaller and denser place than it is today. As well as in these situations in nature, physicists can re-create a mini-version of the quark-gluon plasma by colliding large nuclei (like gold) together in a particle accelerator at virtually the speed of light. This experiment is being performed at the Large Hadron Collider in CERN. Because each nucleus is incredibly small (100 billion of them side- by-side would span a distance of 1mm) the region of quark-gluon plasma created is correspondingly small. The plasma “fireball” also expands and cools incredibly rapidly, so it quickly returns to the normal state of matter where quarks are tightly bound. For these reasons, it is incredibly difficult to get any information about the plasma phase of matter. To understand the processes occurring inside the fireball, physicists need to know its properties such as viscosity, pressure and energy density. It is also important to know at which temperature the quarks inside protons and other particles become unbound and free. With this information, it is possible to calculate how fast the fireball expands and cools, and what mixture of particles will fly out of the fireball and be observed by detectors in the experiment. This research project will use supercomputers to simulate the strong interaction in the quark-gluon phase. We will find the temperature that quarks become unbound, and calculate some of the fundamental physical properties of the plasma such as its conductivity, symmetry properties of baryons and response of hadronic excitations to the chemical potential. These quantities can then be used as inputs into the theoretical models which will enable us to understand the quark-gluon plasma, i.e. the strong interaction past its breaking point.

Project Title: NeatQCD – Nucleon structure at the precision frontier using twisted mass lattice QCD

Project Leader: Giannis Koutsou, The Cyprus Institute, Cyprus

Multi-year Proposal: 2nd of 2 years

Resource Awarded

  • 66 000 000 core hours on Marconi100 hosted by CINECA, Italy

Multi-year Proposal: 2nd of 2 years

Research Field: Fundamental Constituents of Matter

Collaborators

  • Constantia Alexandrou, University of Cyprus, Cyprus
  • Jacob Finkenrath, The Cyprus Institute, Cyprus
  • Simone Bacchio, The Cyprus Institute, Cyprus
  • Kyriakos Hadjiyiannakou, University of Cyprus, Cyprus
  • Ferenc Pittler, The Cyprus Institute, Cyprus
  • Davide Nole, The Cyprus Institute, Cyprus
  • Florian Manigrasso, University of Cyprus, Cyprus
  • Antonino Todaro, University of Cyprus, Cyprus
  • Shuhei Yamamoto, The Cyprus Institute, Cyprus
  • Karl Jansen, DESY, Germany

Abstract

The proton is particularly important in understanding the fundamental properties of matter since, being a stable particle, it can be studied experimentally. Notably, low-energy, high-intensity experiments being conducted at MAMI in Mainz, as well as Jefferson Laboratory and FermiLab in the US are providing precision results on proton structure that may reveal new physics through small discrepancies in the Standard Model (SM). For this precision frontier of particle physics, a major challenge is to determine accurately the contributions due to the strong interaction component of the Standard Model, governed by the theory of Quantum Chromodynamics (QCD). The only formalism that allows us to compute the properties of the proton and its partner the neutron (collectively called nucleons) starting directly from the QCD Lagrangian is the lattice QCD formulation. Lattice QCD has seen tremendous progress in the past years, with large-scale simulations now able to access quantities in the precision frontier of nucleon structure. The goal of this two-year project is to compute, to high precision, nucleon structure observables in this regime using state-of-the-art lattice QCD ensembles, simulated using two degenerate light, strange, and charm twisted mass fermions with masses tuned to their physical values. In particular, we target observables that can probe physics beyond the Standard Model (BSM), including the nucleon tensor and scalar charges, the proton charge radius, and the neutron electric dipole moment. Tier-0 supercomputers, such as the Marconi successor in CINECA targeted in this proposal, are mandatory for such high precision, large scale calculations. In particular, our analysis program, which has been improved continuously during the past five years, relies on large scalable parallel systems of GPUs, that can only be delivered via large allocations such as PRACE.

Project Title: Nuclear Physics from the Standard Model

Project Leader: Assumpta Parreño, University of Barcelona, Spain

Resource Awarded

  • 121 000 000 core hours on Marconi100 hosted by CINECA, Italy

Research Field: Fundamental Constituents of Matter

Collaborators

  • Phiala Elisabeth Shanahan, Massachusetts Institute of Technology, United States of America
  • William Detmold, Massachusetts Institute of Technology, United States of America
  • Marc Illa, Universitat de Barcelona, Spain
  • Michael Louis Wagman, Fermi National Accelerator Laboratory, United States of America
  • Zohreh Davoudi, University of Maryland, United States of America

Abstract

Atomic nuclei make up more than 99% of the visible matter in the universe. Understanding the physics of these nuclei is thus central to understanding the world around us. Our knowledge of nuclei is derived from a century of innovative experiments that revealed nuclei, formed of protons and neutrons, at the cores of atoms. Over the last decades, the underlying structure of protons and neutrons in terms of more fundamental constituents, quarks and gluons, has also been unveiled. Our theoretical understanding of the emergence of nuclei from these quarks and gluons—to the best of our understanding, the fundamental degrees of freedom of nature—lags this impressive phenomenology.

Particularly important are the questions: How does the complex structure of a nucleus arise from the dynamics of the quarks and gluons that are the fundamental objects in the Standard Model which, along with gravity, is our current description of Nature? How do nuclei interact with each other, and with other Standard Model particles? How can new physics beyond the Standard Model be discovered using nuclear isotopes as targets? The Lattice Quantum Chromodynamics calculations proposed in this project will address parts of these broad questions and further elucidate the nuclear realm. In particular, we will calculate the ground-state spectrum and matrix elements of nuclei made by up to 4 baryons (A = 4). This will allow extraction of nuclear and hypernuclear forces relevant for determinations of the nuclear equation of state that governs neutron star structure and merger dynamics, and will constrain theoretical understanding of dark matter scattering from nuclei in terrestrial detectors, representing a significant step forward in showing how nuclei emerge from the intricacies of Standard Model dynamics.

Project Title: RadLepHdec – Electromagnetic corrections to leptonic D- and B-meson decay rates in LQCD

Project Leader: Christoph Lehner, University of Regensburg, Germany

Resource Awarded

  • 46 000 000 core hours on SuperMUC hosted by GCS at LRZ, Germany

Research Field: Fundamental Constituents of Matter

Collaborators

  • Davide Giusti, University of Regensburg, Germany
  • Sebastian Spiegel, University of Regensburg, Germany
  • Andreas Hackl, University of Regensburg, Germany
  • Daniel Knüttel, University of Regensburg, Germany

Abstract

Precision Flavor Physics is a particularly powerful tool for exploring the limits of the Standard Model of Particle Physics and in searching for inconsistencies which would signal the presence of New Physics. An important component of this endeavor is the determination of the elements of the Cabibbo-Kobayashi-Maskawa (CKM) matrix from a wide range of weak processes. The precision in extracting CKM matrix elements is generally limited by our ability to quantify hadronic effects and the main goal of large-scale simulations using the lattice formulation of QCD is the ab-initio evaluation of the non-perturbative QCD effects in physical processes. In the last few years calculations of hadronic observables from first principles, taking systematic errors under very good control, have been carried out through large-scale QCD simulations on the lattice, leading to a precision approaching O(1%) for a number of quantities. In order to make further progress lattice simulations must include isospin-breaking effects and reach the physical pion mass point. The aim of the present project is to compute, from first principles and in a broad kinematical range, the form factors contributing to the amplitudes for the radiative leptonic weak decays of D(D_s) and B(B_s) mesons. This will allow accurate predictions to be made at O(α_em) for the corresponding leptonic decay rates, leading to significantly improved precision in the determination of the CKM matrix elements. Precise predictions for the emission of a hard photon are very interesting especially for the decays of heavy mesons for which currently only model-dependent predictions are available in literature.

Project Title: From cosmos to colliders – multi-scale particle physics with GAMBIT

Project Leader: Anders Kvellestad, University of Oslo, Norway

Resource Awarded

  • 40 000 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Research Field: Fundamental Constituents of Matter

Collaborators

  • Pat Scott, University of Queensland, Australia
  • Tomas Gonzalo, RWTH Aachen University, Germany
  • Are Raklev, University of Oslo, Norway
  • Torsten Bringmann, University of Oslo, Norway
  • Marcin Chrząszcz, Polish Academy of Sciences, Poland
  • Farvah Mahmoudi, Université Lyon 1, France
  • Peter Athron, Monash University, Australia
  • Csaba Balazs, Monash University, Australia
  • Martin White, University of Adelaide, Australia
  • Felix Kahlhoefer, RWTH Aachen, Germany
  • Christoph Weniger, The University of Amsterdam, Netherlands
  • Andy Buckley, University of Glasgow, United Kingdom
  • Julia Harz, Technical University Munich, Germany

Abstract

The Global and Modular Beyond-the-Standard Model Inference Tool (GAMBIT) project aims at producing the most complete and rigorous assessments possible of new theories in particle physics. This is achieved through large-scale statistical analyses that combine the latest experimental results from dark matter searches, high-energy collider experiments such as the LHC, flavour physics, cosmology and neutrino physics. Using the massively parallelized GAMBIT software and cutting-edge optimization methods, these experimental results are compared against detailed theoretical predictions based on high-accuracy quantum field theory calculations and simulations.

The GAMBIT codebase has been developed over a period of eight years by a growing team of experimentalists, theorists, statisticians and computer scientists, working in very close collaboration. Today the GAMBIT team counts 70 members. It includes experts from nearly all the major particle and astroparticle experiments, and developers from many of the field’s major software packages for theory calculations, simulations and optimization.

The recent addition of the cosmology module CosmoBit to the GAMBIT codebase significantly extends our ability to perform truly multi-scale and interdisciplinary physics analyses. We can now test new theories against observations all the way from cosmology and gravitational waves, via astroparticle physics and dark matter searches, down to the fine details of the particle interactions at the LHC and in neutrino experiments. With PRACE Tier-0 computing power we will be able to fully utilize the numerous complementarities between data from the largest and the smallest scales of physics, to perform world-leading investigations into a diverse set of promising extensions of the Standard Model of particle physics.

Project Title: Scaling properties of conserved charge fluctuations in QCD

Project Leader: Olaf Kaczmarek, University of Bielefeld, Germany

Resource Awarded

  • 120 768 000 core hours on Piz Daint hosted by CSCS, Switzerland

Research Field: Fundamental Constituents of Matter

Collaborators

  • Frithjof Karsch, University of Bielefeld, Germany
  • Mugdha Sarkar, University of Bielefeld, Germany
  • Anirban Lahiri, University of Bielefeld, Germany
  • Marius Neumann, Univeristy of Bielefeld, Germany
  • Heng-Tong Ding, Central China Normal University, China
  • Sheng-Tai Li, Central China Normal University, China

Abstract

The goal of this project is to understand better the chiral phase transition in the limit of two massless quark flavors through the calculations of various fluctuation observables. Earlier calculations by us close to the chiral limit shows that mass as well as volume dependence of various quantities can well be described by finite-size scaling functions of O(4) universality class, even at finite lattice spacings. These observations already give evidence in favor of a second order phase transition in chiral limit for (2+1)-flavor QCD. Conserved charge (or flavor) fluctuations, which are derivatives of partition function w.r.t. to the corresponding chemical potential, are expected to carry the signals of any possible criticality with their own characteristic exponents. Using O(4) scaling functions it has been argued that second and fourth order cumulants of baryon number will be finite at the transition temperature in the chiral limit. Although cumulants starting from sixth order and higher order cumulants will be diverging at the chiral phase transition, in particular the mass scaling of the minimum of the sixth order cumulant in the chirally restored partonic phase is expected to be universal and that will establish the nature of the chiral scaling on a firm basis. Moreover these studies will allow us to understand the pattern of fluctuations which have been observed for physical pion mass and will give a better handle to compare with results of fluctuations obtained from experiments like STAR at RHIC and ALICE at LHC.

In this project we will address the above-mentioned topics by performing numerical simulations with Highly Improved Staggered Quarks (HISQ) action with quark masses lower than physical values, corresponding to pion masses about 110 MeV and 80 MeV. A systematic analysis on mass dependence will be performed using existing data for other masses.

Project Title: Long large-eddy-simulations of magnetized binary neutron star mergers: from the turbulent regime to the creation of large-scale magnetic fields

Project Leader: Carlos Palenzuela, Universitat de les Illes Balears, Spain

Resource Awarded

  • 43 500 000 core hours on MareNostrum 4 hosted by BSC, Spain
  • HLST Support

Research Field: Fundamental Constituents of Matter

Collaborators

  • Daniele Viganò, Institute of Space Sciences (ICE-CSIC), Spain
  • Borja Miñano, Universitat de les Illes Balears, Spain
  • Wolfgang Kastaun, Max Planck Institute for Gravitational Physics Hanover, Germany
  • Ricardo Ciolfi, INAF, Osservatorio Astronomico di Padova, Italy
  • Ricard Aguilera-Miret, Universitat de les Illes Balears, Spain
  • Jay Vijay Kalinani, University of Padova, Italy

Abstract

Binary neutron star (BNS) systems are unique astrophysical laboratories to study gravity, plasma physics and dense matter under very extreme conditions. The simultaneous observation of gravitational and electromagnetic counterparts during the event GW170817 + GRB170817A + AT2017gfo, consistent with the merger of a BNS system, started an era of multi-messenger astronomy that will enhance our understanding on the parameters of the system and the physical processes at play, allowing us to test our theories and validate our astrophysical models. The project proposed here is based on the study of the full dynamics of magnetized BNS mergers through extremely accurate numerical simulations, focusing on the physical mechanisms that are relevant for the formation of detectable electromagnetic signals, in particular short Gamma-Ray Bursts and Kilonovae. Our simulations allow to improve the effective resolution via employment of a sub-grid-scale modeling and will shed light on the different processes and instabilities increasing the strength of the magnetic field during the merger, as well as its conversion from small to large scales through winding and magneto-rotational-instability mechanisms. The proposed work extends a previous successful project, focused on the turbulent stage only, to achieve a complete global picture.

Project Title: PULSAR-PIC

Project Leader: François Courvoisier, CNRS, FEMTO-ST Insitute, France

Resource Awarded

  • 35 000 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Research Field: Fundamental Constituents of Matter

Collaborators

  • Kazem Ardaneh, CNRS/ FEMTO-ST Institute, France
  • Pierre-Jean Charpin, CNRS/FEMTO-ST, France
  • Remo Giust, University of Franche-Comte/ FEMTO-ST, France

Abstract

The current project is part of the ERC-funded project PULSAR. It aims at developing novel ultrafast laser materials processing technologies. Its context is mass fabrication (solar panels, consumer electronics, microelectronics) where new technologies are needed to process materials that are always new, on extremely large surfaces, with sub-micron resolution. Ultrafast laser processing is ideally positioned and is also much greener than conventional lithography which uses numerous harmful toxic chemicals.

However, even with high pulse energy, laser processing remains limited to point by point removal of ultra-thin nanometric layers from the material surface. This is because the uncontrolled laser-generated free-electron plasma shields out the incoming light and prevents reaching extreme internal temperatures at very precise nanometric scale.

PULSAR aims at breaking this barrier by developing a radically different concept of laser material ablation regime based on controlling the generation of nanoplasmas which create intense micro-explosions inside the bulk of materials. We use Particle-In-Cell (PIC) to simulate in 3D the interaction between spatially shaped femtosecond laser pulses and nanometric plasmas generated in the bulk of transparent materials. High performance computing is therefore necessary to run the high-resolution codes. Preliminary results with relatively low resolution indicate that we have identified basic phenomena for controlling the plasmas. However, reaching our objectives require much more computational power.

Project Title: The Effective quark-antiquark potential in Quark Gluon Plasma

Project Leader: Rasmus Larsen, University of Stavanger, Norway

Resource Awarded

  • 6 200 000 core hours on JUWELS Booster hosted by GCS at FZJ, Germany

Research Field: Fundamental Constituents of Matter

Collaborators

  • Alexander Rothkopf, University of Stavanger, Norway
  • Peter Petreczky, Brookhaven National Laboratory, United States of America
  • Swagato Mukherjee, Brookhaven National Laboratory, United States of America

Abstract

The study of nuclear matter under extreme conditions of temperature and pressure, similar to those close to the Big Bang is a central focus of collider experiments at the Large Hadron Collider (LHC) at CERN and the Relativistic Heavy Ion Collider (RHIC) at Brookhaven National Laboratory (BNL). One of the key efforts in these experiments is the measurements of the yields of quarkonia (bound states of heavy quarks and anti-quarks), which is believed to be sensitive to the properties of the created medium. This project sets out find the effective potential between the heavy quark and an anti-quark, which greatly enhances our ability to predict the yields of quarkonia in heavy ion experiments at LHC and RHIC. To reach this goal we need to solve a quantum field theory, called quantum chromodynamics (QCD), which amounts to evaluating integrals with more than 100M dimensions using Monte-Carlo methods. With the availability of multi-GPU computing enabled by super computers, we are able to significantly extend these computations, which in turn allow us to extract the effective potential from first principles in unprecedented accuracy. This study will provide new insight into the binding mechanism and real-time dynamics of heavy quarkonium.

Project Title: Strong gravity beyond general relativity

Project Leader: Pau Figueras, Queen Mary University of London, United Kingdom

Resource Awarded

  • 35 000 000 core hours on SuperMUC hosted by GCS at LRZ, Germany

Research Field: Fundamental Constituents of Matter

Collaborators

  • Tiago França, Queen Mary University of London, United Kingdom
  • Chenxia Gu, Queen Mary University of London, United Kingdom
  • Lorenzo Rossi, Queen Mary University of London, United Kingdom
  • Hans Bantilan, Queen Mary University of London, United Kingdom
  • Tomas Andrade, Institut de Ciencies del Cosmos, Universitat de Barcelona, Spain
  • Llibert Aresté Saló, Queen Mary University of London, United Kingdom
  • Luis Lehner, Perimeter Institute, Canada
  • Ramiro Cayuso, Perimeter Institute, Canada
  • Guillaume Dideron, Perimeter Institute, Canada

Abstract

The detections of gravitational waves produced in mergers of compact objects have revolutionised the field of gravitational physics, giving rise to the era of gravitational wave astronomy. With the recent upgrades of the detectors, gravitational wave detections are made on an almost weekly basis. This offers a new opportunity to test Einstein’s theory of general relativity in the strong field regime. Black holes are fundamental objects in any theory of gravity and they precisely capture the strong field regime of the theory. In this project we will focus on studying these objects and their dynamics.

One of the main challenges in carrying out these tests is that almost nothing is known about the strong field regime of alternative theories of gravity. Arguably, the highly dynamical, strong field regime of gravity is where potential deviations from general relativity are more likely to show up, thus revealing new aspects of the fundamental nature of gravity. This is the regime that we are going to focus on in this project and the key tool to access it is numerical relativity. Recent progress in the field has led to well-posed formulations of various alternative theories of gravity of physical interest, and hence they can be potentially simulated in a computer and their predictions extracted.

Another aspect of the strong field regime of gravity, as described by general relativity, is that in certain situations such as higher dimensions or different spacetime asymptotics, black holes can be unstable. There is strong evidence that the endpoints of such instabilities are naked singularities. The latter are important because not only they can reveal fundamental aspects of gravity, but they can also provide a glimpse of quantum gravity. Numerical relativity is again a perfectly suited tool to study them.

The aim of this proposal is to make progress in our fundamental understanding of gravity by:

1) Simulating black hole binaries in alternative theories of gravity and extracting the corresponding waveforms. This will tell us about what are the types of deviations from general relativity we should be looking for and thus potentially detect them.

2) Study the details of singularity formation in black hole instabilities in general relativity to reveal the nature of gravity at its fundamental level. This can allow us to constraint the role of quantum gravity.

Each of these projects will significantly advance our knowledge of gravity at its fundamental level.

Mathematics and Computer Sciences

Project Title: SWOP: the Submesoscale-permitting World Ocean Project

Project Leader: Clément Bricaud, Mercator Ocean International, France

Resource Awarded

  • 24 000 000 core hours on MareNostrum 4 hosted by BSC, Spain

Research Field: Mathematics and Computer Sciences

Collaborators

  • Doroteaciro Iovino, CMCC, Italy
  • Vineet Sony, MERCATOR OCEAN INTERNATIONAL, France
  • Miguel Castrillo, BSC, Spain
  • Julien Le Sommer, Institut des Géosciences de l’Environnement, France
  • Laurent Brodeau, Ocean Next, France

Abstract

Observations are crucial in the understanding of the oceans. They are relevant for planning human-scale activities and necessary for the climate change monitoring and management of marine resources. They are available on diffusion platforms, like the Copernicus Marine Environment Monitoring Service (CMEMS), one of the six pillar services of the EU (European Union) Copernicus programme. With the increase of observing platforms and the improvement of their accuracy, the resolution of observations has been improved. However, these observations data are concentrated at the surface and unevenly sampled. That’s why numerical models are necessary to integrate these data and provide a complete 4-dimensional view of the Ocean. The current numerical models are not adapted to integrate information coming from this new generation of sensors. Therefore, Ocean General Circulation Models (OGCM) must evolve to be able to resolve the global ocean flows at kilometric scale. With the increase of computational domain sizes, OGCMs must also evolve to massively parallel architectures and strong I/O demand. These questions are addressed in the H2020 projects IMMERSE and ESIWACE2 and some changes will be performed in NEMO OGCM. Now, it is conceivable to develop a new global configuration at 1/36° resolution for a future ocean forecasting system which will be operated in the framework of the Copernicus Marine Environment Monitoring Service.
In order to assess and stabilize such a configuration, it is important to perform a multi-year simulation. Here, we propose to realize a forced hindcast simulation of this global configuration at 1/36°horizontal resolution (called ORCA36) based on the last official release of the state-of-the-art modelling framework NEMO version 4 (https://www.nemo-ocean.eu). A 3-year spin up will be performed, then a 6-year simulation will be performed. The ocean-ice system will be forced with an atmospheric dataset generated by the ECMWF IFS forecasting system. The spin up will cover the 2013-2015 period and the simulation the 2016-2021 period. During the simulation, high frequency model outputs will be generated.

Universe Sciences

Project Title: SuperStars – Self-consistent Supernova Driven Star Formation

Project Leader: Paolo Padoan, University of Barcelona, Spain

Multi-year Proposal: 3rd or 3 years

Resource Awarded

  • 3 700 000 core hours on Joliot-Curie (KNL) hosted by GENCI at CEA, France
  • 19 400 000 core hours on Joliot-Curie Rome hosted by GENCI at CEA, France

Multi-year Proposal: 3rd or 3 years

Research Field: Universe Sciences

Collaborators

  • Troels Haugbølle, University of Copenhagen, Denmark
  • Åke Nordlund, University of Copenhagen, Denmark
  • Mika Juvela, University of Helsinki, Finland
  • Liubin Pan, Sun Yat-sen University, China
  • Veli-Matti Pelkonen, Universitat de Barclona, Spain
  • Lu Zujia, Universitat de Barclona, Spain

Abstract

Star formation is a fundamental and still largely unsolved problem of astrophysics and cosmology. Its complexity stems from the non-linear coupling of a broad range of scales, the interaction of turbulence, magnetic fields and gravity, and from the onset of different feedback mechanisms from massive stars, such as stellar winds, ionizing radiation and supernovae. This complexity defies an analytical approach. This project tackles the multi-scale nature of star formation with state-of-the-art adaptive-mesh-refinement methods, addressing three key questions: 1) What causes the disruption of molecular clouds, thus setting the local efficiency of star formation? 2) How can we explain the dichotomy between the global (Galactic) and local (molecular clouds) star-formation rates? 3) What is the expected variance of the star-formation rate at different scales? The computational model is ground-breaking, as it provides a self-consistent description of star formation and supernova-feedback for the first time. This is achieved by resolving the formation of individual massive stars, so the location and position of the supernovae is determined self-consistently by the star-formation process. To properly resolve the turbulent cascade driven by supernova explosions, the formation of individual massive stars, and the evolution of supernova remnants, the dynamic ranges of space and time scales are 0.01 pc to 250 pc and 0.01 yr to 70 Myr, respectively. This represents a challenging high-performance computing problem even with state-of-the-art codes and supercomputing systems. With our own version of the Ramses adaptive-mesh-refinement code on Skylake nodes, we can achieve our goal with approximately 49.5 Million core hours, which we break into three early allocations of 16.5 Million core hours. Datasets from this computational model will provide a numerical laboratory for star-formation studies, thanks to the very large sample of star-forming regions formed and evolved ab-initio (with realistic, self-generated initial and boundary conditions) in the simulation. We will generate synthetic catalogs of hundreds of molecular clouds and stellar clusters and thousands of massive stars and supernova remnants.

Project Title: Turbulent energy transfer across the Earth’s magnetospheric boundary layers

Project Leader: Takuma Nakamura, Austrian Academy of Sciences, Austria

Resource Awarded

  • 30 000 000 core hours on MareNostrum 4 hosted by BSC, Spain

Research Field: Universe Sciences

Collaborators

  • Kevin Alexander Blasl, Austrian Academy of Sciences, Austria
  • Rumi Nakamura, Austrian Academy of Sciences, Austria
  • Julia Stawarz, Imperial College London, United Kingdom
  • William David Nystrom, Los Alamos National Laboratory, United States of America

Abstract

In this project, a series of large-scale fully kinetic simulations will be performed to understand energy transfer physics in space plasmas. Space between planets, stars, and galaxies is filled with plasma, a collection of high-energy, charged particles with its density small enough to neglect particle collisions. In such a collisionless system, the boundary layer between regions with different plasma properties plays a central role in transferring energy and controlling the dynamics of the system. In the Earth’s magnetosphere, a representative collisionless plasma system, the energy input from the solar wind is transferred through different physical processes at various boundary layers, which eventually control the global dynamics of the magnetosphere related to many space weather phenomena like auroral substorms and geomagnetic storms. On the other hand, plasma turbulence has been commonly observed in many locations in space, and understanding how the energy cascades between different spatiotemporal scales in the turbulence is key for understanding the energy transfer in collisionless plasmas. Indeed, the recently launched high-resolution Magnetospheric Multiscale (MMS) mission, the first mission to resolve electron-scales in-situ, very frequently observed turbulence at each boundary layer in the magnetosphere. This project will systematically investigate the realistic energy transfer processes across the turbulent boundary layers in the solar wind-magnetosphere system based on large-scale fully kinetic simulations through the comparisons with latest observations by MMS. Since the sizes of the magnetospheric boundary layers are basically in magneto-hydrodynamic (MHD) scales (>104km), to quantitatively understand the energy transfer processes in the magnetosphere, it is required to handle the MHD-scales resolving the smallest electron-scales (10-100 km) where the energy is expected to be eventually dissipated. The scientific focus of this project is to quantitatively investigate these MHD-scale energy transfer processes resolving the electron-scales using large-scale fully kinetic simulations and their comparisons with the MMS observations. This project is timely because the combination of our high-performance fully kinetic simulation code, our new techniques for the direct, quantitative comparisons with MMS, and the requested Tier-0 resources allow us to perform such large-scale simulations and their quantitative comparisons with the real observations.