Cooling rate effects in sodium silicate glasses: Bridging the gap between molecular dynamics simulations and experiments

Structural Causes of Brittleness Changes in Aluminosilicate Glasses with Different Cooling Rates

Numerous sources have already demonstrated that varying annealing rates can result in distinct toughness and brittleness in glass. To determine the underlying mechanisms driving this phenomenon, molecular dynamic (MD) simulations were employed to investigate the microstructure of aluminosilicate glasses under different cooling rates, and then uniaxial stretching was performed on them under controlled conditions. Results indicated that compared with short-range structure, cooling rate has a greater influence on the medium-range structure in glass, and it remarkably affects the volume of voids. Both factors play a crucial role in determining the brittleness of the glass. The former adjusts network connectivity to influence force transmission by manipulating the levels of bridging oxygen (BO) and non-bridging oxygen (NBO), and the latter accomplishes the objective of influencing brittleness by modifying the environmental conditions that affect the changes in BO and NBO content. The variation in the void environment results in differences in the strategies of the changes in BO and NBO content during glass stress. These findings stem from the excellent response of BO and NBO to the characteristic points of stress-strain curves during stretching. This paper holds importance in understanding the reasons behind the effect of cooling rates on glass brittleness and in enhancing our understanding of the ductile/brittle transition (DTB) in glass.

Atomistic mechanism of structural and volume relaxation below glass transition temperature in a soda-lime silicate glass revealed by Raman spectroscopy and its DFT calculations

To elucidate the atomistic origin of volume relaxation in soda-lime silicate glass annealed below the glass transition temperature (Tg), the experimental and calculated Raman spectra were compared. By decomposing the calculated Raman spectra into specific groups of atoms, the Raman peaks at 800, 950, 1050, 1100, and 1150 cm-1 were attributed to oxygen and silicon in Si-O-Si, non-bridging oxygen in the Q2 unit, bridging oxygen in low-angle Si-O-Si, non-bridging oxygen in the Q4 unit, and bridging oxygen in high-angle Si-O-Si, respectively. Based on these attributions, we found that by decreasing the fictive temperature by annealing below Tg - 70 K, a homogenization reaction Q2 + Q4 → 2Q3 and an increase in average Si-O-Si angle occurred simultaneously. By molecular dynamics simulation, we clarified how the experimentally demonstrated increase in average Si-O-Si angle contributes to volume shrinkage; increasing Si-O-Si angles can expand the space inside the rings, and Na can be inserted into the ring center.

Accurate and Transferable Machine Learning Potential for Molecular Dynamics Simulation of Sodium Silicate Glasses

An accurate and transferable machine learning (ML) potential for the simulation of binary sodium silicate glasses over a wide range of compositions (from 0 to 50% Na2O) was developed. The potential energy surface is approximated by the sum of atomic energy contributions mapped by a neural network algorithm from the local geometry comprising information on atomic distances and angles with neighboring atoms using the DeePMD code [Wang, H. Comput. Phys. Commun. 2018, 228, 178-184]. Our model was trained on a large data set of total energies and atomic forces computed at the density functional theory level on structures extracted from classical molecular dynamics (MD) simulations performed at several temperatures from 300 to 3000 K. This allows for the generation of a robust and transferable ML potential applicable over the full compositional range of glass formability at different temperatures that outperforms the empirical potentials available in the literature in reproducing structures and properties such as bond angle distribution, total distribution functions, and vibrational density of state. The generality of the approach enables the future training of a potential with other or more elements allowing for simulations of structures, properties, and behavior of ternary and multicomponent oxide glasses with nearly ab initio accuracy at a fraction of the computational cost.

Passivation of Mid-Gap Electronic States at Calcium Aluminosilicate Glass Surfaces upon Water Exposure: An Ab Initio Study

When a clean glass surface is exposed to humid air, a thin water layer forms on the hydrophilic surface. Using ab initio molecular dynamics, we simulate the changes in the electronic structure of a CaO-Al₂O₃-SiO₂ glass model upon vacuum fracture and subsequent exposure to H₂O. When the glass is fractured, dangling bonds form, which lower the band gap of the surface by ∼1.8 eV compared to the bulk value due to mid-gap surface states. When H₂O adsorbs onto the vacuum-fractured surface, the band gap increases to a value closer to that of the bulk band gap. Using two different hydroxylation methods, we find that the calculated band gap of the glass surface depends on the hydroxylation state. Surfaces with ∼4.5 OH/nm² have smaller band gaps due to unfilled surface states, and surfaces with ∼2.5 OH/nm² have larger band gaps with no apparent unfilled surface states. The resulting changes in the electronic structure, quantified by electron affinity and work function values, are hypothesized to play an important role in the electrostatic charge transfer based on the principles of surface state theory, which posit that the density of electronic surface states determines the amount of electronic charge transfer to or from material surfaces.

Influence of Magnesium on the Structure of Complex Multicomponent Silicates: Insights from Molecular Simulations and Neutron Scattering Experiments

A series of multicomponent glasses containing up to five oxides are studied using classical molecular dynamics simulations and neutron scattering experiments. The focus is on the role of magnesium in determining the structural properties of these glasses and the possible mixed effect during a sodium/magnesium substitution. Calculated structure functions (pair correlation function and structure factor) rather accurately reproduce their experimental counterpart, and we show that more fine structural features are qualitatively reproduced well, despite some discrepancies in the preferential spatial distribution between sodium and magnesium to aluminum and boron, as well as the nonbridging oxygen, distribution. The simulated systems offer a solid basis to support previous experimental findings on the composition-structure relationship, allowing for further analysis and property calculation. It is confirmed that the substitution of sodium by magnesium leads to the decrease of four-fold boron and a modification of the alkali coordinations with a significant change of the network structure. Specifically, magnesium coordination extracted from numerical simulations highlights a potential dissociation from penta- to tetra- and hexahedral units with increasing MgO contents along the glass series, which could not be resolved experimentally.

Insight into the structure-elastic property relationship of calcium silicate glasses: a multi-length scale approach

Based on a combination of molecular dynamics simulations, and Raman and Brillouin light scattering spectroscopies, we investigate the structure and elastic properties relationship in an archetypical calcium silicate glass system. From molecular dynamics and Raman spectroscopy, we show that the atomic structure at the short and intermediate length scales is made up of long polymerized silicate chains, which adjusts itself by closing the Si-O-Si angles and leaving more space to [CaO]_n edge shared polyhedra to strengthen the glass. Using Brillouin spectroscopy, we observe an increase of elastic constants of the glass with the calcium content, as the cohesion of the glass structure is enhanced through an increase of the binding between the cross-linked calcium-silicate frameworks. This result, albeit being simple in its nature, illustrates for the first time the implication of the calcium framework in the elastic behavior of the glass and will contribute substantially to the understanding of the composition-structure-property relationships in multi-component industrial glasses.

Structural origin of thermal shrinkage in soda-lime silicate glass below the glass transition temperature: A theoretical investigation by microsecond timescale molecular dynamics simulations

Microscopic dynamical features in the relaxation of glass structures are one of the most important unsolved problems in condensed matter physics. Although the structural relaxation processes in the vicinity of glass transition temperature are phenomenologically expressed by the Kohlrausch-Williams-Watts function and the relaxation time can be successfully interpreted by Adam-Gibbs theory and/or Narayanaswamy's model, the atomic rearrangement, which is the origin of the volume change, and its driving force have not been elucidated. Using the microsecond time-scale molecular dynamics simulations, this study provides insights to quantitatively determine the origin of the thermal shrinkage below T_g in a soda-lime silicate glass. We found that during annealing below T_g, Na ions penetrate into the six-membered silicate rings, which remedies the acute O-O-O angles of the energetically unstable rings. The ring structure change makes the space to possess the cation inside the rings, but the ring volume is eventually reduced, which results in thermal shrinkage of the soda-lime silica glass. In conclusion, the dynamical structural relaxation due to the cation displacement evokes the overall volume relaxation at low temperature in the glassy material.

Looking through glass: Knowledge discovery from materials science literature using natural language processing

Most of the knowledge in materials science literature is in the form of unstructured data such as text and images. Here, we present a framework employing natural language processing, which automates text and image comprehension and precision knowledge extraction from inorganic glasses' literature. The abstracts are automatically categorized using latent Dirichlet allocation (LDA) to classify and search semantically linked publications. Similarly, a comprehensive summary of images and plots is presented using the caption cluster plot (CCP), providing direct access to images buried in the papers. Finally, we combine the LDA and CCP with chemical elements to present an elemental map, a topical and image-wise distribution of elements occurring in the literature. Overall, the framework presented here can be a generic and powerful tool to extract and disseminate material-specific information on composition-structure-processing-property dataspaces, allowing insights into fundamental problems relevant to the materials science community and accelerated materials discovery.

Experimental method to quantify the ring size distribution in silicate glasses and simulation validation thereof

Silicate glasses have no long-range order and exhibit a short-range order that is often fairly similar to that of their crystalline counterparts. Hence, the out-of-equilibrium nature of glasses is largely encoded in their medium-range order. However, the ring size distribution-the key feature of silicate glasses' medium-range structure-remains invisible to conventional experiments and, hence, is largely unknown. Here, by combining neutron diffraction experiments and force-enhanced atomic refinement simulations for two archetypical silicate glasses, we show that rings of different sizes exhibit a distinct contribution to the first sharp diffraction peak in the structure factor. On the basis of these results, we demonstrate that the ring size distribution of silicate glasses can be determined solely from neutron diffraction patterns, by analyzing the shape of the first sharp diffraction peak. This method makes it possible to uncover the nature of silicate glasses' medium-range order.

Preparation of Plastics- and Foaming Agent-Free and Porous Bamboo Charcoal based Composites Using Sodium Silicate as Adhesives

Plastics and foaming agents are often used to prepare large-size and low-density bamboo charcoal (BC) based composites. In this study, a plastic-free and foaming agent-free BC based composite was prepared by substituting sodium silicate (SS) for plastics. The effect of both the BC particle sizes and the usage amount of SS on the mechanical and adsorptive properties of the BC/SS composites were investigated. The experimental results show that when the BC particle size is 270 μm and the mass ratio of BC to SS is equal to 10:5, the BC/SS composite has the optimal foaming effect and best comprehensive properties. In addition, the foaming pores of the composite are caused by water vapor, which has difficulty escaping the BC because of the blockage of SS during the hot pressing process. In the BC/SS composite (10:5), the static bending intensity and the compressive strength reach respectively 6.13 MPa and 5.5 MPa, and the average pore size and porosity are 557.85 nm and 52.03%, respectively. In addition, its formaldehyde adsorptionrate reaches 21.6%. In view of good mechanical properties, formaldehyde adsorption, and environmentally friendly performance, the BC/SS composite has a great potential as a core layer of interior building materials.

Molecular dynamics simulations of simplified sodium borosilicate glasses: the effect of composition on structure and dynamics

The fusion of valuable material properties has led to the acceptance of sodium borosilicate (NBS) glasses for nuclear waste immobilization. Although popular, the mechanisms associated with these properties are still only partially discovered and need further exploration. Bearing this in mind, the combination of experiments, molecular dynamics (MD) simulations and the Dell, Yuan and Bray model have been used to understand the role of composition variation for structural and physical aspects of vitrified borosilicate glasses. Experiments have been conducted to evaluate the macroscopic glass parameters of density (ρ), glass transition temperature (Tg) and thermal expansion coefficient (TEC). Experimentally observed trends for ρ, Tg and TEC with composition have been found in good agreement with the MD results. MD studies also provide a microscopic understanding of the glass structure and phenomena associated with the change in the glass composition. A detailed view of local structure and medium-range connectivity for the borosilicate glasses has been explored. Owing to a large B4 population, the results showed the abundant presence of BO4-BO4 connections, we hereby omit the generally accepted "B[4] avoidance rule" for glass. The relative propensity for connecting SiO4/BO3/BO4 structural motifs is in line with the predictions made by the Dell, Yuan and Bray model. Furthermore, the effects of composition on the mechanical integrity of NBS glasses, including the elastic nature, plastic distortion, yielding, breaking stress, and brittle fracture, have been explored by MD simulations. In addition, the glass dynamics have been evaluated by diffusion coefficient and the results suggest that Na+ is likely to be more mobile in the case of NBS1 as compared to NBS2 and NBS3 due to significant disruption in the glass network introduced by a larger amount of Na2O network modifier. Also, the diffusivity was reduced with increasing B2O3 due to the altered role of Na+ ions from network modifiers to charge compensators. The combined study of experiments, MD simulations and the Dell, Yuan and Bray model establish the correlation between the microscopic structure and macroscopic properties of NBS glasses with varied composition, which might be of great scientific use for future glasses in various applications including nuclear waste immobilization.

Dynamics of confined water and its interplay with alkali cations in sodium aluminosilicate hydrate gel: insights from reactive force field molecular dynamics

This paper presents the dynamics of confined water and its interplay with alkali cations in disordered sodium aluminosilicate hydrate (N-A-S-H) gel using reactive force field molecular dynamics. N-A-S-H gel is the primary binding phase in geopolymers formed via alkaline activation of fly ash. Despite attractive mechanical properties, geopolymers suffer from durability issues, particularly the alkali leaching problem which has motivated this study. Here, the dynamics of confined water and the mobility of alkali cations in N-A-S-H is evaluated by obtaining the evolution of mean squared displacements and Van Hove correlation function. To evaluate the influence of the composition of N-A-S-H on the water dynamics and diffusion of alkali cations, atomistic structures of N-A-S-H with Si/Al ratio ranging from 1 to 3 are constructed. It is observed that the diffusion of confined water and sodium is significantly influenced by the Si/Al ratio. The confined water molecules in N-A-S-H exhibit a multistage dynamic behavior where they can be classified as mobile and immobile water molecules. While the mobility of water molecules gets progressively restricted with an increase in Si/Al ratio, the diffusion coefficient of sodium also decreases as the Si/Al ratio increases. The diffusion coefficient of water molecules in the N-A-S-H structure exhibits a lower value than those of the calcium-silicate-hydrate (C-S-H) structure. This is mainly due to the random disordered structure of N-A-S-H as compared to the layered C-S-H structure. To further evaluate the influence of water content in N-A-S-H, atomistic structures of N-A-S-H with water contents ranging from 5-20% are constructed. Qⁿ distribution of the structures indicates significant depolymerization of N-A-S-H structure with increasing water content. Increased conversion of Si-O-Na network to Si-O-H and Na-OH components with an increase in water content helps explain the alkali-leaching issue in fly ash-based geopolymers observed macroscopically. Overall, the results in this study can be used as a starting point towards multiscale simulation-based design and development of durable geopolymers.

Revealing hidden medium-range order in amorphous materials using topological data analysis

Despite the numerous technological applications of amorphous materials, such as glasses, the understanding of their medium-range order (MRO) structure-and particularly the origin of the first sharp diffraction peak (FSDP) in the structure factor-remains elusive. Here, we use persistent homology, an emergent type of topological data analysis, to understand MRO structure in sodium silicate glasses. To enable this analysis, we introduce a self-consistent categorization of rings with rigorous geometrical definitions of the structural entities. Furthermore, we enable quantitative comparison of the persistence diagrams by computing the cumulative sum of all points weighted by their lifetime. On the basis of these analysis methods, we show that the approach can be used to deconvolute the contributions of various MRO features to the FSDP. More generally, the developed methodology can be applied to analyze and categorize molecular dynamics data and understand MRO structure in any class of amorphous solids.

Ionic self-diffusion and the glass transition anomaly in aluminosilicates

The glass transition temperature (Tg) is the temperature after which a supercooled liquid undergoes a dynamical arrest. Usually, glass network modifiers (e.g., Na2O) affect the behavior of Tg. However, in aluminosilicate glasses, the effect of different modifiers on Tg is still unclear and shows an anomalous behavior. Here, based on molecular dynamics simulations, we show that Tg decreases with increasing charge balancing cation field strength (FS) in the aluminosilicate glasses, which is an anomalous behavior as compared to other oxide glasses. The results show that the origins of this anomaly come from the dynamics of the supercooled liquid above Tg, which in turn is correlated to pair excess entropy. Our results deepen our understanding of the effect of different modifiers on the properties of the aluminosilicate glasses.

Precipitation of calcium-alumino-silicate-hydrate gels: The role of the internal stress

Concrete gains its strength from the precipitation of a calcium-alumino-silicate-hydrate (C-A-S-H) colloidal gel, which acts as its binding phase. However, despite concrete's ubiquity in the building environment, the atomic-scale mechanism of C-A-S-H precipitation is still unclear. Here, we use reactive molecular dynamics simulations to model the early-age precipitation of a C-A-S-H gel. We find that, upon gelation, silicate and aluminate precursors condensate and polymerize to form an aluminosilicate gel network. Notably, we demonstrate that the gelation reaction is driven by the existence of a mismatch of atomic-level internal stress between Si and Al polytopes, which are initially experiencing some local tension and compression, respectively. The polymerization of Si and Al polytopes enables the release of these competitive stresses.

Statistical topology of bond networks with applications to silica

Whereas knowledge of a crystalline material's unit cell is fundamental to understanding the material's properties and behavior, there are no obvious analogs to unit cells for disordered materials despite the frequent existence of considerable medium-range order. This article views a material's structure as a collection of local atomic environments that are sampled from some underlying probability distribution of such environments, with the advantage of offering a unified description of both ordered and disordered materials. Crystalline materials can then be regarded as special cases where the underlying probability distribution is highly concentrated around the traditional unit cell. The H_{1} barcode is proposed as a descriptor of local atomic environments suitable for disordered bond networks and is applied with three other descriptors to molecular dynamics simulations of silica glasses. Each descriptor reliably distinguishes the structure of glasses produced at different cooling rates, with the H_{1} barcode and coordination profile providing the best separation. The approach is generally applicable to any system that can be represented as a sparse graph.

Predicting the Young's Modulus of Silicate Glasses using High-Throughput Molecular Dynamics Simulations and Machine Learning

The application of machine learning to predict materials’ properties usually requires a large number of consistent data for training. However, experimental datasets of high quality are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young’s modulus of silicate glasses. We demonstrate that this combined approach offers good and reliable predictions over the entire compositional domain. By comparing the performances of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.

Evidence of a two-dimensional glass transition in graphene: Insights from molecular simulations

Liquids exhibit a sudden increase in viscosity when cooled fast enough, avoiding thermodynamically predicted route of crystallization. This phenomenon, known as glass transition, leads to the formation of non-periodic structures known as glasses. Extensive studies have been conducted on model materials to understand glass transition in two dimensions. However, despite the synthesis of disordered/amorphous single-atom thick structures of carbon, little attention has been given to glass transition in realistic two-dimensional materials such as graphene. Herein, using molecular dynamics simulation, we demonstrate the existence of glass transition in graphene leading to a realistic two-dimensional glassy structure, namely glassy graphene. We show that the resulting glassy structure exhibits excellent agreement with experimentally realized disordered graphene. Interestingly, this glassy graphene exhibits a wrinkled but stable structure, with reduced thermal vibration in comparison to its crystalline counterpart. We suggest that the topological disorder induced by glass transition governs the unique properties of this structure.

Modifier clustering and avoidance principle in borosilicate glasses: A molecular dynamics study

Oxide glasses are typically described as having a random, disordered skeleton of network-forming polyhedra that are depolymerized by network-modifying cations. However, the existence of local heterogeneity or clustering within the network-forming and network-modifying species remains unclear. Here, based on molecular dynamics simulations, we investigate the atomic structure of a series of borosilicate glasses. We show that the network-modifying cations exhibit some level of clustering that depends on composition-in agreement with Greaves' modified random network model. In addition, we demonstrate the existence of some mutual avoidance among network-forming atoms, which echoes the Loewenstein avoidance principle typically observed in aluminosilicate phases. Importantly, we demonstrate that the degree of heterogeneity in the spatial distribution of the network modifiers is controlled by the level of ordering in the interconnectivity of the network formers. Specifically, the mutual avoidance of network formers is found to decrease the propensity for modifier clustering.

New insights into the atomic structure of amorphous TiO2 using tight-binding molecular dynamics

Amorphous TiO₂ (a-TiO₂) could offer an attractive alternative to conventional crystalline TiO₂ phases for photocatalytic applications. However, the atomic structure of a-TiO₂ remains poorly understood with respect to that of its crystalline counterparts. Here, we conduct some classical molecular dynamics simulations of a-TiO₂ based on a selection of empirical potentials. We show that, on account of its ability to dynamically assign the charge of each atom based on its local environment, the second-moment tight-binding charge equilibration potential yields an unprecedented agreement with available experimental data. Based on these simulations, we investigate the degree of order and disorder in a-TiO₂. Overall, the results suggest that a-TiO₂ features a large flexibility in its local topology, which may explain the high sensitivity of its structure to the synthesis method being used.