Local structure, thermodynamics, and melting of boron phosphide at high pressures by deep learning-driven ab initio simulations.

Boron phosphide (BP) is a (super)hard semiconductor constituted of light elements, which is promising for high demand applications at extreme conditions. The behavior of BP at high temperatures and pressures is of special interest but is also poorly understood because both experimental and conventional ab initio methods are restricted to studying refractory covalent materials. The use of machine learning interatomic potentials is a revolutionary trend that gives a unique opportunity for high-temperature study of materials with ab initio accuracy. We develop a deep machine learning potential (DP) for accurate atomistic simulations of the solid and liquid phases of BP as well as their transformations near the melting line. Our DP provides quantitative agreement with experimental and ab initio molecular dynamics data for structural and dynamic properties. DP-based simulations reveal that at ambient pressure, a tetrahedrally bonded cubic BP crystal melts into an open structure consisting of two interpenetrating sub-networks of boron and phosphorous with different structures. Structure transformations of BP melt under compressing are reflected by the evolution of low-pressure tetrahedral coordination to high-pressure octahedral coordination. The main contributions to structural changes at low pressures are made by the evolution of medium-range order in the B-subnetwork and, at high pressures, by the change of short-range order in the P-subnetwork. Such transformations exhibit an anomalous behavior of structural characteristics in the range of 12-15 GPa. DP-based simulations reveal that the Tm(P) curve develops a maximum at P ≈ 13 GPa, whereas experimental studies provide two separate branches of the melting curve, which demonstrate the opposite behavior. Analysis of the results obtained raises open issues in developing machine learning potentials for covalent materials and stimulates further experimental and theoretical studies of melting behavior in BP.

Freezing in two-length-scale systems: complexity, universality and prediction.

Two-length-scale pair potentials arise ubiquitously in condensed matter theory as effective interparticle interactions in molecular, metallic and soft matter systems. The existence of two different bond lengths generated by the shape of potential causes complicated behavior in even one-component systems: polymorphism in solid and liquid states, water-like anomalies, the formation of quasicrystals and high stability against crystallization. Here we address general properties of freezing in one-component two-length-scale systems and argue that solidification of a liquid during cooling is essentially determined by the radial distribution function (RDF) of the liquid. We show that different two-length-scale systems having similar RDFs freeze into the same solid phases. In some cases, the similarity between RDFs can be expressed by the proximity of two dimensionless effective parameters: the ratio between effective bond lengths,λ, and the fraction of short-bonded particlesφ. We validate this idea by studying the formation of different solid phases in different two-length-scale systems. The method proposed allows predicting effectively the formation of solid phases in both numerical simulations and self-assembling experiments in soft matter systems with tunable interactions.

Structure of the simple harmonic-repulsive system in liquid and glassy states studied by the triple correlation function.

An efficient description of the structures of liquids and, in particular, the structural changes that happen with liquids on supercooling remains to be a challenge. The systems composed of soft particles are especially interesting in this context because they often demonstrate non-trivial local orders that do not allow to introduce the concept of the nearest-neighbor shell. For this reason, the use of some methods, developed for the structure analysis of atomic liquids, is questionable for the soft-particle systems. Here we report about our investigations of the structure of the simple harmonic-repulsive liquid in 3D using the triple correlation function (TCF), i.e., the method that does not rely on the nearest neighbor concept. The liquid is considered at reduced pressure (P = 1.8) at which it exhibits remarkable stability against crystallization on cooling. It is demonstrated that the TCF allows addressing the development of the orientational correlations in the structures that do not allow drawing definite conclusions from the studies of the bond-orientational order parameters. Our results demonstrate that the orientational correlations, if measured by the heights of the peaks in the TCF, significantly increase on cooling. This rise in the orientational ordering is not captured properly by the Kirkwood's superposition approximation. Detailed considerations of the peaks' shapes in the TCF suggest the existence of a link between the orientational ordering and the slowdown of the system's dynamics. Our findings support the view that the development of the orientational correlations in liquids may play a significant role in the liquids' dynamics and that the considerations of the pair distribution function may not be sufficient to understand intuitively all the structural changes that happen with liquids on supercooling. In general, our results demonstrate that the considerations of the TCF are useful in the discussions of the liquid's structures beyond the pair density function and interpreting the results obtained with the bond-orientational order parameters.

Deep machine learning interatomic potential for liquid silica.

The use of machine learning to develop neural network potentials (NNP) representing the interatomic potential energy surface allows us to achieve an optimal balance between accuracy and efficiency in computer simulation of materials. A key point in developing such potentials is the preparation of a training dataset of ab initio trajectories. Here we apply a deep potential molecular dynamics (DeePMD) approach to develop NNP for silica, which is the representative glassformer widely used as a model system for simulating network-forming liquids and glasses. We show that the use of a relatively small training dataset of high-temperature ab initio configurations is enough to fabricate NNP, which describes well both structural and dynamical properties of liquid silica. In particular, we calculate the pair correlation functions, angular distribution function, velocity autocorrelation functions, vibrational density of states, and mean-square displacement and reveal a close agreement with ab initio data. We show that NNP allows us to expand significantly the time-space scales achievable in simulations and thus calculating dynamical and transport properties with more accuracy than that for ab initio methods. We find that developed NNP allows us to describe the structure of the glassy silica with satisfactory accuracy even though no low-temperature configurations were included in the training procedure. The results obtained open up prospects for simulating structural and dynamical properties of liquids and glasses via NNP.

Effect of copper concentration on the structure and properties of Al-Cu-Fe and Al-Cu-Ni melts.

We address a relationship between properties of liquid and solid states by comparing structural characteristics and viscosity in Al-Cu-Fe and Al-Cu-Ni melts. The former system forms an equilibrium quasicrystalline phase but the latter does not. We show that the concentration behavior of the viscosity, melting temperature and characteristics of the chemical short-range order correlate with each other. The main structural differences between the melts are related to the peculiarities of their electronic structure, which is the same for liquid and solid states near the melting temperature.

Nucleation instability in supercooled Cu-Zr-Al glass-forming liquids.

Few general models representing certain classes of real glass-forming systems play a special role in computer simulations of supercooled liquid and glasses. Recently, it was shown that one of the most widely used model glassformers-the Kob-Andersen binary mixture-crystalizes in quite lengthy molecular dynamics simulations, and moreover, it is in fact a very poor glassformer at large system sizes. Thus, our understanding of crystallization stability of model glassformers is far from complete due to the fact that relatively small system sizes and short time scales have been considered so far. Here we address this issue for two embedded atom models intensively used last years in numerical studies of Cu-Zr-(Al) bulk metallic glasses. Exploring the structural evolution of Cu64.5Zr35.5 and Cu46Zr46Al8 alloys at continuous cooling and isothermal annealing, we observe that both systems nucleate in sufficiently lengthy simulations, although critical nucleation time for the latter is an order of magnitude higher than that for the former. We show that Cu64.5Zr35.5 is actually unstable to crystallization for large system sizes (N > 20 000). Both systems crystallize with the formation of tetrahedrally close packed Laves phases of different types. We argue that nucleation instability of the simulated Cu64.5Zr35.5 alloy is due to the fact that its composition is very close to that for the stable Cu2Zr compound with a C15 Laves phase structure.

Polytetrahedral structure and glass-forming ability of simulated Ni-Zr alloys.

Binary Cu-Zr system is a representative bulk glassformer demonstrating high glass-forming ability (GFA). From the first glance, the Ni-Zr system is the most natural object to expect the same behavior because nickel and copper are neighbors in the periodic table and have similar physicochemical properties. However, it is known that the Ni-Zr system has worse GFA than the Cu-Zr one. To understand the underlying physics, we investigate the Ni α Zr1-α system in whole concentration range α ∈ [0, 1]. Doing molecular dynamic simulations with a reliable embedded atom model potential, we show that the simulated Ni-Zr system also has relatively low GFA, which is comparable to that for an additive binary Lennard-Jones mixture without any chemical interaction. Icosahedral local ordering in Ni-Zr alloys is known to be less pronounced than that in the Cu-Zr ones; we see that as well. However, the icosahedron is not the only structural motif responsible for GFA. We find that the local structure of glassy Ni α Zr1-α alloys at 0.3 < α < 0.65 can be described in terms of Z11-Z16 Kasper polyhedra with high density of topological defects including icosahedra as a part of this family. Concentration of topologically perfect Kasper polyhedra appears to be several times smaller than that in Cu-Zr. This is the reason for relatively poor GFA of the Ni-Zr system.

Cooling rate dependence of simulated Cu64.5Zr35.5 metallic glass structure.

Using molecular dynamics simulations with embedded atom model potential, we study structural evolution of Cu64.5Zr35.5 alloy during the cooling in a wide range of cooling rates γ ∈ (1.5 ⋅ 10(9), 10(13)) K/s. Investigating short- and medium-range orders, we show that the structure of Cu64.5Zr35.5 metallic glass essentially depends on cooling rate. In particular, a decrease of the cooling rate leads to an increase of abundances of both the icosahedral-like clusters and Frank-Kasper Z16 polyhedra. The amounts of these clusters in the glassy state drastically increase at the γmin = 1.5 ⋅ 10(9) K/s. Analysing the structure of the glass at γmin, we observe the formation of nano-sized crystalline grain of Cu2Zr intermetallic compound with the structure of Cu2Mg Laves phase. The structure of this compound is isomorphous with that for Cu5Zr intermetallic compound. Both crystal lattices consist of two types of clusters: Cu-centered 13-atom icosahedral-like cluster and Zr-centered 17-atom Frank-Kasper polyhedron Z16. That suggests the same structural motifs for the metallic glass and intermetallic compounds of Cu-Zr system and explains the drastic increase of the abundances of these clusters observed at γmin.

Hydrodynamic anomalies in supercritical fluid.

Using the molecular dynamics simulations we investigate properties of velocity autocorrelation function of Lennard-Jones fluid at long and intermediate time scales in wide ranges of temperature and density. We show that the amplitudes of both the leading and the subleading time asymptotic terms of velocity autocorrelation function, a1 and a2, show essentially non-monotonic temperature and density dependence. There are two lines on temperature-density plain corresponding to maxima of a1 (a2) along isochors and isotherms situated in the supercritical fluid (hydrodynamic anomalies). These lines give insight into the stages of the fluid evolution into gas.

Multistage structural evolution in simple monatomic supercritical fluids: superstable tetrahedral local order.

The local order units of dense simple liquid are typically three-dimensional (close packed) clusters: hcp, fcc, and icosahedrons. We show that the fluid demonstrates the superstable tetrahedral local order up to temperatures several orders of magnitude higher than the melting temperature and down to critical density. While the solid-like local order (hcp, fcc) disappears in the fluid at much lower temperatures and far above critical density. We conclude that the supercritical fluid shows the temperature (density)-driven two-stage "melting" of the three-dimensional local order. We also find that the structure relaxation times in the supercritical fluid are much larger than ones estimated for weakly interactive gas even far above the melting line.

Superfragile glassy dynamics of a one-component system with isotropic potential: competition of diffusion and frustration.

We investigate glassy dynamical properties of one-component three-dimensional system of particles interacting via pair repulsive potential by the molecular dynamic simulation in the wide region of densities. The glass state is superfragile and it has high glass-forming ability. The glass transition temperature T(g) has a pronounced minimum at densities where the frustration is maximal.