arXiv subjects

Fausto Martelli

Publications and source records attributed to Fausto Martelli.

At least 19 recordsLinked to original sources

arXiv2026

Topological origin of peak splitting in the structure factor of liquid water

The splitting of the principal peak in the structure factor of liquid water is commonly interpreted as evidence of a competition between two distinct local environments. Here, we show that this peak splitting arises from medium-range topological features of the hydrogen-bond network. Using atomistic simulations, we systematically decompose the structure factor into contributions from hydrogen-bonded rings of different sizes. We find that 5-8-membered rings, which dominate the network topology of liquid water at low temperatures, can directly explain the experimentally observed bimodal scattering signal. Among these, 5-membered rings are particularly persistent, maintaining distinct structural signatures even above room temperature. Our findings establish a direct link between the network topology of liquid water and experimentally accessible diffraction features, clarifying the microscopic basis of water's behaviour and suggesting a broader conceptual framework for interpreting the anomalies in tetrahedral network liquids and glasses.

cond-mat.soft↗

arXiv2025

Predicting Thermodynamics of Liquid Water from Time Series Analysis

Thermodynamics, introduced over two centuries ago, remains foundational to our understanding of physical, chemical, biological, and engineering systems. Its principles are traditionally grounded in the statistical mechanics framework, which explains macroscopic behavior from microscopic states. In this work, we propose an alternative approach that interprets thermodynamic behavior through the lenses of time series analysis, an approach commonly used in other fields, including finance, climate, and signal processing. We perform classical molecular dynamics simulations of liquid water, the most complex, anomalous, and important substance known, over a wide range of its phase diagram. By examining the temporal evolution of the hydrogen bond network (HBN) topology, we demonstrate that the dynamics of microscopic topological motifs populating the HBN encode the system's macroscopic thermodynamic behavior. Furthermore, our approach enables the prediction of thermodynamic properties in regions beyond those directly sampled in our simulations. We achieve this result by leveraging artificial intelligence to uncover patterns in temporally resolved data that are often lost through conventional averaging. This work offers new insights into the fundamental behavior of water and network-forming materials more broadly, establishing a new paradigm for understanding material properties beyond the classical confines of statistical mechanics.

physics.chem-ph↗

arXiv2025

Devitrification and Melting Dynamics in Vapor Deposited Water Ice

The equilibration dynamics of ultrastable glasses subjected to heating protocols has attracted recent experimental and theoretical interest. With simulations of the mW water model, we investigate the devitrification and melting dynamics of both conventional quenched (QG) and vapor deposited (DG) amorphous ices under controlled heating ramps. By developing an algorithm to reconstruct hydrogen-bond networks, we show that bond ring statistics correlates with the structural stability of the glasses and allows tracking crystalline and liquid clusters during devitrification and melting. We find that QG melts in the bulk, whereas melting in DG preferentially begins near the free surface. During devitrification, the DG shows an excess of 5-membered rings near the free surface, which is consistent with its tendency to nucleate the crystal phase in this region. Additionally, the DG shows an Avrami exponent exceeding the standard 1+d behavior, while both glasses display the same sub-3d growth of liquid clusters across heating rates, indicating that the DG enhanced exponent stems from its higher kinetic stability.

cond-mat.soft↗

arXiv2025

Distillation of atomistic foundation models across architectures and chemical domains

Machine-learned interatomic potentials have transformed computational research in the physical sciences. Recent atomistic `foundation' models have changed the field yet again: trained on many different chemical elements and domains, these potentials are widely applicable, but comparably slow and resource-intensive to run. Here we show how distillation via synthetic data can be used to cheaply transfer knowledge from atomistic foundation models to a range of different architectures, unlocking much smaller, more efficient potentials. We demonstrate speed-ups of $> 10\times$ by distilling from one graph-network architecture into another, and $> 100\times$ by leveraging the atomic cluster expansion framework. We showcase applicability across chemical and materials domains: from liquid water to hydrogen under extreme conditions; from porous silica and a hybrid halide perovskite solar-cell material to modelling organic reactions. Our work shows how distillation can support the routine and computationally efficient use of current and future atomistic foundation models in real-world scientific research.

physics.comp-ph↗

arXiv2025

Force-Free Molecular Dynamics Through Autoregressive Equivariant Networks

Molecular dynamics (MD) simulations play a crucial role in scientific research. Yet their computational cost often limits the timescales and system sizes that can be explored. Most data-driven efforts have been focused on reducing the computational cost of accurate interatomic forces required for solving the equations of motion. Despite their success, however, these machine learning interatomic potentials (MLIPs) are still bound to small time-steps. In this work, we introduce TrajCast, a transferable and data-efficient framework based on autoregressive equivariant message passing networks that directly updates atomic positions and velocities lifting the constraints imposed by traditional numerical integration. We benchmark our framework across various systems, including a small molecule, crystalline material, and bulk liquid, demonstrating excellent agreement with reference MD simulations for structural, dynamical, and energetic properties. Depending on the system, TrajCast allows for forecast intervals up to $30\times$ larger than traditional MD time-steps, generating over 15 ns of trajectory data per day for a solid with more than 4,000 atoms. By enabling efficient large-scale simulations over extended timescales, TrajCast can accelerate materials discovery and explore physical phenomena beyond the reach of traditional simulations and experiments. An open-source implementation of TrajCast is accessible under https://github.com/IBM/trajcast.

physics.comp-ph↗

arXiv2024

Correlating ultrastability with fragility and surface mobility in vapor deposited tetrahedral glasses

Several experiments on molecular and metallic glasses have shown that the ability of vapor deposition to produce ultrastable glasses is correlated with their structural and thermodynamic properties. Here we investigate the vapor deposition of a class of tetrahedral materials (including silicon and water) via molecular dynamics simulations of the generalized Stillinger-Weber potential. By changing a single parameter that controls the local tetrahedrality, we show that the emergence of ultrastable behavior is correlated with an increase in the fragility of the model. At the same time, while the mobility of the surface compared to the bulk shows only slight changes at low temperature, with increasing the tetrahedrality, it displays a significant enhancement towards the glass transition temperature. Our results point towards a strong connection between bulk dynamics, surface dynamics and glass-ultrastability ability in this class of materials.

cond-mat.soft↗

arXiv2024

Link to Densify: Topological Transitions and Origin of Hysteresis During the (De)Compression of Amorphous Ices

In this Letter we study the phase transition between amorphous ices and the nature of the hysteresis cycle separating them. We discover that a topological transition takes place as the system transforms from low-density amorphous ice (LDA) at low pressures, to high-density amorphous ice (HDA) at high pressures. Specifically, we uncover that the hydrogen bond network (HBN) displays qualitatively different topologies in the LDA and HDA phases: the former characterised by disentangled loop motifs, while the latter displaying topologically complex and metastable Hopf-linked and knotted configurations. At the phase transition, the transient opening of the HBN topological motifs yields mechanical fragility on the macroscale. Our results provide a detailed microscopic description of the topological nature of the phase transition and the hysteresis cycle between amorphous ices. We argue that the topological transition discovered in this work may not only improve our understanding of amorphous ices but represent a generic mechanism for the densification of network-forming materials.

cond-mat.dis-nn↗

arXiv2023

High-dimensional order parameters and neural network classifiers applied to amorphous ices

Amorphous ice phases are key constituents of water's complex structural landscape. This study investigates the polyamorphic nature of water, focusing on the complexities within low-density amorphous ice (LDA), high-density amorphous ice (HDA), and the recently discovered medium-density amorphous ice (MDA). We use rotationally-invariant, high-dimensional order parameters to capture a wide spectrum of local symmetries for the characterisation of local oxygen environments. We train a neural network (NN) to classify these local environments, and investigate the distinctiveness of MDA within the structural landscape of amorphous ice. Our results highlight the difficulty in accurately differentiating MDA from LDA due to structural similarities. Beyond water, our methodology can be applied to investigate the structural properties and phases of disordered materials.

cond-mat.dis-nn↗

arXiv2023

Electrofreezing of Liquid Water at Ambient Conditions

Water is routinely exposed to external electric fields (EFs). Whether, e.g., at physiological conditions, in contact with biological systems, or at the interface of polar surfaces in countless technological and industrial settings, water responds to EFs on the order of a few V/{\AA} in a manner that is still under intense investigation. Dating back to the $19^{th}$ century, the possibility of solidifying water upon applying an EF instead of adjusting temperature and pressure -- a process known as electrofreezing -- is an alluring promise that has canalized major efforts since, with uncertain outcomes. In this work, we perform long \emph{ab initio} molecular dynamics simulations \textcolor{black}{of water at ambient conditions exposed at EFs of different intensities. While the response of single water molecules is almost instantaneous, the cooperativity of the hydrogen bonds induces slower reorganizations that can be captured by dividing the trajectories in disjoint time windows and by performing analysis on each of them separately. Upon adopting this approach, we find} that EFs of $0.10\leq$EFs$\leq0.15$~V/{\AA} induce electrofreezing \textcolor{black}{occurring after $\sim150$~ps. We observe a continuous transition to a disordered state characterized by frozen dynamical properties, damped oscillations, lower energy, and enhanced local structural properties. Therefore, we ascribe this state to} a new ferroelectric amorphous phase, which we term f-GW (ferroelectric glassy water). Our work represents the first evidence of electrofreezing of liquid water at ambient conditions and therefore impacts several fields, from \textcolor{black}{fundamental chemical physics to} biology \textcolor{black}{and} catalysis.

physics.chem-ph↗

arXiv2023

Structural signatures of ultrastability in a deposited glassformer

Glasses obtained from vapor deposition on a cold substrate have superior thermodynamic and kinetic stability with respect to ordinary glasses. Here we perform molecular dynamics simulations of vapor deposition of a model glass-former and investigate the origin of its high stability compared to that of ordinary glasses. We find that the vapor deposited glass is characterized by locally favoured structures (LFS) whose occurrence correlates with its stability, reaching a maximum at the optimal deposition temperature. The formation of LFS is enhanced near the free surface, hence supporting the idea that the stability of vapor deposited glasses is connected to the relaxation dynamics at the surface.

cond-mat.soft↗

arXiv2022

Wettability of graphite under 2D confinement

The thermodynamics of solid/liquid interfaces under nanoconfinement has tremendous implications for liquid transport properties. Here using molecular dynamics, we investigate graphite nanoslits and study how the water/graphite interfacial tension changes with the degree of confinement. We found that, for nanochannel heights between 0.7nm and 2.6nm, graphite becomes more hydrophobic than in bulk, and that the value of the surface tension oscillates before eventually converging towards a constant value for larger slits. The value of the surface tension is correlated with the slip length of the fluid and explained in terms of the effective and interfacial density, hydration pressure and friction coefficient. The study clearly indicates that there is a critical channel height of 0.9nm (achievable experimentally1) at which the surface tension reaches its highest value, but the water diffusion across the channel is at its minimum. The structural analysis shows that for this pore size a transition between a 2D and 3D hydrogen bond network is accompanied by an abrupt increase in conformational entropy. Our results show that the wettability of solid surfaces can change under nanoconfinement and the data can be used to interpret the experimental permeability data.

physics.flu-dyn↗

arXiv2021

The physics of Empty Liquids: from Patchy particles to Water

Empty liquids represent a wide class of materials whose constituents arrange in a random network through reversible bonds. Many key insights on the physical properties of empty liquids have originated almost independently from the study of colloidal patchy particles on one side, and a large body of theoretical and experimental research on water on the other side. Patchy particles represent a family of coarse-grained potentials that allows for a precise control of both the geometric and the energetic aspects of bonding, while water has arguably the most complex phase diagram of any pure substance, and a puzzling amorphous phase behavior. It was only recently that the exchange of ideas from both fields has made it possible to solve long-standing problems and shed new light on the behavior of empty liquids. Here we highlight the connections between patchy particles and water, focusing on the modelling principles that make an empty liquid behave like water, including the factors that control the appearance of thermodynamic and dynamic anomalies, the possibility of liquid-liquid phase transitions, and the crystallization of open crystalline structures.

cond-mat.soft↗

arXiv2021

Topology and complexity of the hydrogen bond network in classical models of water

Over the years, plenty of classical interaction potentials for water have been developed and tested against structural, dynamical and thermodynamic properties. On the other hands, it has been recently observed (F. Martelli et. al, \textit{ACS Nano}, \textbf{14}, 8616--8623, 2020) that the topology of the hydrogen bond network (HBN) is a very sensitive measure that should be considered when developing new interaction potentials. Here we report a thorough comparison of 11 popular non polarizable classical water models against their HBN, which is at the root of water properties. We probe the topology of the HBN using the ring statistics and we evaluate the quality of the network inspecting the percentage of broken and intact HBs. For each water model, we assess the tendency to develop hexagonal rings (that promote crystallization at low temperatures) and pentagonal rings (known to frustrate against crystallization at low temperatures). We then introduce the \emph{network complexity index}, a general descriptor to quantify how much the topology of a given network deviates from that of the ground state, namely of hexagonal or cubic ice. Remarkably, we find that the network complexity index allows us to relate, for the first time, the dynamical properties of different water models with their underlying topology of the HBN. Our study provides a benchmark against which the performances of new models should be tested against, and introduces a general way to quantify the complexity of a network which can be transferred to other materials and that links the topology of the HBN with dynamical properties. Finally, our study introduces a new perspective that can help in rationalizing the transformations among the different phases of water and of other materials.

cond-mat.soft↗

arXiv2021

Re-defining the concept of hydration water in water under soft confinement

Water shapes and defines the properties of biological systems. Therefore, understanding the nature of the mutual interaction between water and biological systems is of primary importance for a proper assessment of biological activity and the development of new drugs and vaccines. A handy way to characterize the interactions between biological systems and water is to analyze their impact on water density and dynamics in the proximity of the interfaces. It is well established that water bulk density and dynamical properties are recovered at distances in the order of $\sim1$~nm from the surface of biological systems. Such evidence led to the definition of \emph{hydration} water as the thin layer of water covering the surface of biological systems and affecting-defining their properties and functionality. Here, we review some of our latest contributions showing that phospholipid membranes affect the structural properties and the hydrogen bond network of water at greater distances than the commonly evoked $\sim1$~nm from the membrane surface. Our results imply that the concept of hydration water should be revised or extended, and pave the way to a deeper understanding of the mutual interactions between water and biological systems.

cond-mat.soft↗

arXiv2021

Network Topology in Water Nanoconfined between Phospholipid Membranes

Water provides the driving force for the assembly and stability of many cellular components. Despite its impact on biological functions, a nanoscale understanding of the relationship between its structure and dynamics under soft confinement has remained elusive. As expected, water in contact with biological membranes recovers its bulk density and dynamics at $\sim 1$ nm from phospholipid headgroups but surprisingly enhances its intermediate-range order (IRO) over a distance, at least, twice as large. Here, we explore how the IRO is related to the water's hydrogen bond network (HBN) and its coordination defects. We characterize the increased IRO by an alteration of the HBN up to more than eight coordination shells of hydration water. The HBN analysis emphasizes the existence of a bound-unbound water interface at $\sim 0.8$ nm from the membrane. The unbound water has a distribution of defects intermediate between bound and bulk water, but with density and dynamics similar to bulk, while bound water has reduced thermal energy and much more HBN defects than low-temperature water. This observation could be fundamental for developing nanoscale models of biological interactions and for understanding how alteration of the water structure and topology, for example, due to changes in extracellular ions concentration, could affect diseases and signaling. More generally, it gives us a different perspective to study nanoconfined water.

cond-mat.soft↗

arXiv2019

Unravelling the contribution of local structures to the anomalies of water: the synergistic action of several factors

We investigate the microscopic origin of water's anomalies by inspecting the hydrogen bond network (HBN) and the spatial organization of low-density-liquid (LDL) like and high-density-liquid (HDL) like environments. Specifically, we simulate --via classical molecular dynamics simulations-- the isobaric cooling of a sample composed by 512 water molecules from ambient to deeply undercooled conditions at three pressures, namely 1 bar, 400 bar and 1000 bar. \emph{In correspondence with the Widom line}, (i) the HDL-like dominating cluster undergoes fragmentation caused by the percolation of LDL-like aggregates following a spinodal-like kinetics; (ii) such fragmentation always occur s at a "critical" concentration of $\sim20-30\%$ in LDL; (iii) the HBN within LDL-like environments is characterized by an equal number of pentagonal and hexagonal rings that create a state of maximal frustration between a configuration that promotes crystallization (hexagonal ring) and a configuration that hinders it (pentagonal ring); (iv) the spatial organization of HDL-like environments shows a marked variation. Moreover, the inspection of the global symmetry shows that the intermediate-range order decreases in correspondence with the Wid om line, and such decrease becomes more pronounced upon increasing the pressure, hence supporting the hypothesis of a liquid-liquid critical point. Our results reveal and rationalize the complex microscopic origin of water's anomalies as the cooperative effect of several fac tors acting synergistically. Beyond implications for water, our findings may be extended to other materials displaying anomalous behaviours.

physics.chem-ph↗

arXiv2018

Water between membranes: Structure and Dynamics

An accurate description of the structure and dynamics of interfacial water is essential for phospholipid membranes, since it determines their function and their interaction with other molecules. Here we consider water confined in stacked membranes with hydration from poor to complete, as observed in a number of biological systems. Experiments show that the dynamics of water slows down dramatically when the hydration level is reduced. All-atom molecular dynamics simulations identify three (inner, hydration and outer) regions, within a distance of approximately 1 nm from the membrane, where water molecules exhibit different degrees of slowing down in the dynamics. The slow-down is a consequence of the robustness of the hydrogen bonds between water and lipids and the long lifetime of the hydrogen bonds between water molecules near the membrane. The interaction with the interface, therefore, induces a structural change in the water that can be emphasized by calculating its intermediate range order. Surprisingly, at distances as far as ~ 2.5 nm from the interface, although the bulk-like dynamics is recovered, the intermediate range order of water is still slightly higher than that in the bulk at the same thermodynamic conditions. Therefore, the water-membrane interface has a structural effect at ambient conditions that propagates further than the often-invoked 1 nm length scale. Membrane fluctuations smear out this effect macroscopically, but an analysis performed by considering local distances and instantaneous configurations is able to reveal it, possibly contributing to our understanding of the role of water at biomembrane interfaces.

cond-mat.soft↗

arXiv2018

Root-Growth of Boron Nitride Nanotubes: Experiments and \textit{Ab Initio} Simulations

We have synthesized boron nitride nanotubes (BNNTs) in an arc in presence of boron and nitrogen species only, without transition metals. We find that BNNTs are often attached to pure boron nanoparticles, suggesting that root-growth is a likely mechanism for their formation. To gain further insight into this process we have studied key mechanisms for root growth of BNNTs on the surface of a liquid boron droplet by ab initio molecular dynamics simulations. We find that nitrogen atoms reside predominantly on the droplet surface where they organize to form boron nitride islands below 2400 K. To minimize contact with the liquid particle underneath, the islands assume non-planar configurations that are likely precursors for the thermal nucleation of cap structures. Once formed, the caps are stable and can easily incorporate nitrogen and boron atoms at their base, resulting in further growth. Our simulations support the root-growth mechanism of BNNTs and provide comprehensive evidence of the active role played by liquid boron.

physics.chem-ph↗