Periodic Reporting for period 1 - TRANQUIL (TRANsport with QUantum nuclei in Ionic Liquids)
Okres sprawozdawczy: 2021-06-01 do 2023-05-31
Transport phenomena are fundamental and irreversible processes that occur due to the random motion of particles in condensed and gas phases of matter out of equilibrium, according to the second law of thermodynamics. The exchange of quantities such as energy, charge, and mass between different domains of a system is characterised and quantified, in the linear response regime, by transport coefficients, such as thermal and electrical conductivities, diffusivity, etc., which represent the proportionality factor between an external “thermodynamic force” (like a gradient of temperature, electrical potential or concentration) and the induced flux of heat, charge, or particles of a given type.
Transport phenomena are ubiquitous in material science and technology, influencing the efficiency of devices, fuel cells, heat exchangers: for instance, in the current quest for solid-state electrolytes for next generation batteries, a tradeoff between the flow of ionic charge - needed for fast charging and large powers - and a safe Joule-heat dissipation must be found to avoid overheating or explosions. The ability to understand, and then tune, materials’ transport properties is thus of paramount importance for next future technological development with vast societal impact.
From a more speculative standpoint, transport properties play a crucial role also in planetary science, where they govern the behaviour of materials at the extreme conditions typical of the interior of celestial bodies, allowing to construct evolutionary models of planets able to explain their current characteristics, like their temperature profile, luminosity or electromagnetic fields.
Despite their significance in different branches of physical, chemical, and materials sciences, as well as in their applications, our understanding of transport coefficients, including their dependence on microscopic chemical composition, pressure, and temperature, remains incomplete. Moreover, the experimental measurement of transport properties is often challenging for novel materials in the energy and industrial sectors, hazardous substances, or systems in extreme geophysical conditions.
Therefore, the development of accurate theories and numerical simulations, that account for the quantum nature of particles and interactions among them, is not only a subject of scientific speculation but also serves a practical purpose in predicting and tailoring properties that would otherwise be inaccessible. This action aimed to achieve fundamental advancements in this research domain, with a specific emphasis on ionic liquids and systems, like superionic materials, characterised by the presence of at least a diffusive species.
Transport phenomena are ubiquitous in material science and technology, influencing the efficiency of devices, fuel cells, heat exchangers: for instance, in the current quest for solid-state electrolytes for next generation batteries, a tradeoff between the flow of ionic charge - needed for fast charging and large powers - and a safe Joule-heat dissipation must be found to avoid overheating or explosions. The ability to understand, and then tune, materials’ transport properties is thus of paramount importance for next future technological development with vast societal impact.
From a more speculative standpoint, transport properties play a crucial role also in planetary science, where they govern the behaviour of materials at the extreme conditions typical of the interior of celestial bodies, allowing to construct evolutionary models of planets able to explain their current characteristics, like their temperature profile, luminosity or electromagnetic fields.
Despite their significance in different branches of physical, chemical, and materials sciences, as well as in their applications, our understanding of transport coefficients, including their dependence on microscopic chemical composition, pressure, and temperature, remains incomplete. Moreover, the experimental measurement of transport properties is often challenging for novel materials in the energy and industrial sectors, hazardous substances, or systems in extreme geophysical conditions.
Therefore, the development of accurate theories and numerical simulations, that account for the quantum nature of particles and interactions among them, is not only a subject of scientific speculation but also serves a practical purpose in predicting and tailoring properties that would otherwise be inaccessible. This action aimed to achieve fundamental advancements in this research domain, with a specific emphasis on ionic liquids and systems, like superionic materials, characterised by the presence of at least a diffusive species.
The theoretical playground for the study of transport phenomena via numerical simulations is offered by the so-called Green-Kubo theory (GKT) of linear response. According to GKT, the transport coefficient of interest can be obtained as the asymptotic-time integral of a given correlation function (ACF), computed at thermodynamic equilibrium, of the fluxes of interest. The ACF can then be sampled along an equilibrium molecular dynamics (EMD) simulations.
The action produced modern, simulation-oriented reviews of GKT, with emphasis on recent theoretical developments on invariance principles, new results on the statistical properties of estimators in EMD, and a fully quantum-mechanical formulation of charge transport in ionic liquids. Another main result of the action was the development of machine learning (ML) methods to try and circumvent accuracy and computational cost issues, by providing relatively cheap potentials targeting the accuracy of ab-initio calculations. Different ML potentials were developed to simulate typical solid-state electrolytes that have emerged as non-toxic and cheap candidates for next generation rechargeable batteries. For Li3ClO, the performance of ML potentials has been also validated against different methods, highlighting the dominant role of vacancies and anharmonic effects in operating temperature regimes. The role of different DFT functionals in predicting different properties was also investigated, by leveraging the ability of constructing ML models for hybrid functionals, whereas a direct investigation via ab-initio calculations would have been prohibitively expensive. A thorough investigation of finite-size effects (FSE) in EMD of ion diffusion, heat transport, and thermal motion in superionic materials was carried out, which emphasised the need for large-scale simulations
Another objective which clearly emerged along the deployment of the action was the urgence for common ontologies, standards, languages and practices in data-driven techniques for functional materials’ engineering and discovery. Together with international colleagues and experts in the field, the action’s experienced researcher (ER) proposed and submitted a related COST action which has been recently approved.
The results of the action have been disseminated in peer-reviewed articles and reviews in high-impact journals, and at national and international conferences and workshops, some of which were organised by the ER. Data, codes and script have been made available according to the FAIR strategy on open access databases. The action’s activity was continuously spread via social media (Twitter and Researchgate). The ER was also engaged in outreach activities aimed at the general public and students, from elementary to high school.
The action produced modern, simulation-oriented reviews of GKT, with emphasis on recent theoretical developments on invariance principles, new results on the statistical properties of estimators in EMD, and a fully quantum-mechanical formulation of charge transport in ionic liquids. Another main result of the action was the development of machine learning (ML) methods to try and circumvent accuracy and computational cost issues, by providing relatively cheap potentials targeting the accuracy of ab-initio calculations. Different ML potentials were developed to simulate typical solid-state electrolytes that have emerged as non-toxic and cheap candidates for next generation rechargeable batteries. For Li3ClO, the performance of ML potentials has been also validated against different methods, highlighting the dominant role of vacancies and anharmonic effects in operating temperature regimes. The role of different DFT functionals in predicting different properties was also investigated, by leveraging the ability of constructing ML models for hybrid functionals, whereas a direct investigation via ab-initio calculations would have been prohibitively expensive. A thorough investigation of finite-size effects (FSE) in EMD of ion diffusion, heat transport, and thermal motion in superionic materials was carried out, which emphasised the need for large-scale simulations
Another objective which clearly emerged along the deployment of the action was the urgence for common ontologies, standards, languages and practices in data-driven techniques for functional materials’ engineering and discovery. Together with international colleagues and experts in the field, the action’s experienced researcher (ER) proposed and submitted a related COST action which has been recently approved.
The results of the action have been disseminated in peer-reviewed articles and reviews in high-impact journals, and at national and international conferences and workshops, some of which were organised by the ER. Data, codes and script have been made available according to the FAIR strategy on open access databases. The action’s activity was continuously spread via social media (Twitter and Researchgate). The ER was also engaged in outreach activities aimed at the general public and students, from elementary to high school.
The action contributed to significant progress beyond the state of the art in several key areas: the inclusion of electronic quantum excitations into Machine Learning Interatomic Potentials (MLIPs) offered new possibilities for more precise modelling and large-scale simulation of materials at planetary conditions. Significant advancements in electrochemistry were achieved, with an analytical theory of electrostatic screening of an ionic liquid in a temperature gradient, and a study of transport properties of solvated electrons.
Concerning applications, the action allowed the first thorough assessment of the thermal conductivity in realistic solid-state electrolytes, highlighting the role played by quantum effects of the solid matrix, vacances, ionic diffusion, DFT functionals and finite-size effects. In addition to these scientific advancements, the project has addressed, via new quantitative metrics, the robustness of local predictions made by machine learning models.
The results of the action hold great potential for socio-economic impact and wider societal implications, contributing to the development of more efficient energy materials with tailored properties. The action’s theoretical results and methodological developments, largely disseminated through international conferences and workshops, ensure wide accessibility and utilisation of data and codes within the scientific community, thanks to the open-access strategy pursued in the action.
The ER’s network and management capabilities were also reinforced during the action, which resulted in an accepted proposal for a COST action on data-driven applications towards the engineering of functional materials.
The project's outreach activity fostered scientific curiosity and promoted the importance of STEM subjects among the younger generations, and conveyed the significance of transport phenomena and of sustainable materials.
Concerning applications, the action allowed the first thorough assessment of the thermal conductivity in realistic solid-state electrolytes, highlighting the role played by quantum effects of the solid matrix, vacances, ionic diffusion, DFT functionals and finite-size effects. In addition to these scientific advancements, the project has addressed, via new quantitative metrics, the robustness of local predictions made by machine learning models.
The results of the action hold great potential for socio-economic impact and wider societal implications, contributing to the development of more efficient energy materials with tailored properties. The action’s theoretical results and methodological developments, largely disseminated through international conferences and workshops, ensure wide accessibility and utilisation of data and codes within the scientific community, thanks to the open-access strategy pursued in the action.
The ER’s network and management capabilities were also reinforced during the action, which resulted in an accepted proposal for a COST action on data-driven applications towards the engineering of functional materials.
The project's outreach activity fostered scientific curiosity and promoted the importance of STEM subjects among the younger generations, and conveyed the significance of transport phenomena and of sustainable materials.