Exploring Challenging Properties of Liquid Metallic Systems through Machine Learning: Liquid La and Li4Pb Systems.

In this machine learning (ML) study, we delved into the unique properties of liquid lanthanum and the Li4Pb alloy, revealing some unexpected features and also firmly establishing some of the debated characteristics. Leveraging interatomic potentials derived from ab initio calculations, our investigation achieved a level of precision comparable to first-principles methods while at the same time entering the hydrodynamic regime. We compared the structure factors and pair distribution functions to experimental data and unearthed distinctive collective excitations with intriguing features. Liquid lanthanum unveiled two transverse collective excitation branches, each closely tied to specific peaks in the velocity autocorrelation function spectrum. Furthermore, the analysis of the generalized specific heat ratio in the hydrodynamic regime investigated with the ML molecular dynamics simulations uncovered a peculiar behavior, impossible to discern with only ab initio simulations. Liquid Li4Pb, on the other hand, challenged existing claims by showcasing a rich array of branches in its longitudinal dispersion relation, including a high-frequency LiLi mode with a nonhydrodynamic optical character that maintains a finite value as q → 0. Additionally, we conducted an in-depth analysis of various transport coefficients, expanding our understanding of these liquid metallic systems. In summary, our ML approach yielded precise results, offering new and captivating insights into the structural and dynamic aspects of these materials.

Ab initio study of longitudinal and transverse dynamics, including fast sound, in molten UO2 and liquid Li-Pb alloys.

The disparity between the masses of the two components in a binary liquid system can lead to the appearance of a peculiar phenomenon named "fast sound," which was identified for the first time in Li4Pb several decades ago and later observed in other Li based alloys. However, the exact characteristics and nature of this phenomenon and the reasons behind its appearance have not totally been identified yet. In this work, we analyze the longitudinal and transverse current correlation functions of UO2, Li4Pb, and Li0.17Pb0.83, as obtained from ab initio molecular dynamics simulations. We find that fast sound appears to occur in the two former systems but not in the latter. Additionally, we discuss some of the properties of the liquid mixtures that may be related to the appearance (or absence) of the phenomenon, such as the composition, the polyhedral structure of the melt, and the type of bonding in the system.

An Efficient Deep Learning Scheme To Predict the Electronic Structure of Materials and Molecules: The Example of Graphene-Derived Allotropes.

Computations based on density functional theory (DFT) are transforming various aspects of materials research and discovery. However, the effort required to solve the central equation of DFT, namely the Kohn-Sham equation, which remains a major obstacle for studying large systems with hundreds of atoms in a practical amount of time with routine computational resources. Here, we propose a deep learning architecture that systematically learns the input-output behavior of the Kohn-Sham equation and predicts the electronic density of states, a primary output of DFT calculations, with unprecedented speed and chemical accuracy. The algorithm also adapts and progressively improves in predictive power and versatility as it is exposed to new diverse atomic configurations. We demonstrate this capability for a diverse set of carbon allotropes spanning a large configurational and phase space. The electronic density of states, along with the electronic charge density, may be used downstream to predict a variety of materials properties, bypassing the Kohn-Sham equation, leading to an ultrafast and high-fidelity DFT emulator.

First principles study of liquid uranium at temperatures up to 2050 K.

Uranium compounds are used as fissile materials in nuclear reactors. In present day reactors the most used nuclear fuel is uranium dioxide, but in generation-IV reactors other compounds are also being considered, such as uranium carbide and uranium mononitride. Upon possible accidents where the coolant would not circulate or be lost the core of the reactor would reach very high temperatures, and therefore it is essential to understand the behaviour of the nuclear fuel under such conditions for proper risk assessment. We consider here molten metallic uranium at several temperatures ranging from 1455 to 2050 K. Even though metallic uranium is not a candidate for nuclear fuel it could nevertheless be produced due to the thermochemical instability of uranium nitride at high temperatures. We use first principles techniques to analyse the behaviour of this system and obtain basic structural and dynamic properties, as well as some thermodynamic and transport properties, including atomic diffusion and viscosity.

Structure and dynamics of the liquid 3d transition metals near melting. An ab initio study.

The static and dynamic properties of several bulk liquid 3d transition metals at thermodynamic conditions near their respective melting points have been evaluated by using ab initio molecular dynamics simulations. The calculated static structure factors show an asymmetric second peak followed by a more or less marked shoulder which points to a sizeable amount of icosahedral local order. Special attention is devoted to the analysis of the longitudinal and transverse current spectral functions and the corresponding dispersion of collective excitations. For some metals, we have found the existence of two branches of transverse collective excitations in the second pseudo-Brillouin zone. Finally, results are also reported for several transport coefficients.

Orbital-free density functional theory simulation of collective dynamics coupling in liquid Sn.

The appearance of a second excitation mode in the longitudinal and transverse collective dynamics of a series of liquid metals has been observed recently, either by inelastic X-ray scattering (IXS) or by first-principles molecular dynamics (FPMD). The phenomenon's origin is still uncertain, although some theories have been used with relative success to reproduce the FPMD results as a means to find an explanation for it (e.g., mode-coupling (MC) theory in liquid zinc [B. G. del Rio and L. E. González, Phys. Rev. B 95, 224201 (2017)]). For liquid tin (l-Sn), the second excitation mode in the dynamic structure factor and longitudinal current spectrum was observed by IXS [S. Hosokawa et al., J. Phys.: Condens. Matter 25, 112101 (2013)]. By performing orbital-free density functional theory MD simulations of l-Sn, we confirm the existence of a second excitation mode in the longitudinal and transverse collective dynamics and provide a theoretical explanation based on MC theory. Moreover, we introduce a new binary term in MC theory to better capture the negative minima present in the memory functions of the collective dynamics. These results confirm that the origin of the second excitation mode exhibited by the longitudinal and transverse collective dynamics in some liquid metals involves an indirect coupling of the longitudinal and transverse modes.

Globally-Optimized Local Pseudopotentials for (Orbital-Free) Density Functional Theory Simulations of Liquids and Solids.

The accuracy of local pseudopotentials (LPSs) is one of two major determinants of the fidelity of orbital-free density functional theory (OFDFT) simulations. We present a global optimization strategy for LPSs that enables OFDFT to reproduce solid and liquid properties obtained from Kohn-Sham DFT. Our optimization strategy can fit arbitrary properties from both solid and liquid phases, so the resulting globally optimized local pseudopotentials (goLPSs) can be used in solid and/or liquid-phase simulations depending on the fitting process. We show three test cases proving that we can (1) improve solid properties compared to our previous bulk-derived local pseudopotential generation scheme; (2) refine predicted liquid and solid properties by adding force matching data; and (3) generate a from-scratch, accurate goLPS from the local channel of a non-local pseudopotential. The proposed scheme therefore serves as a full and improved LPS construction protocol.