Design Of New Glasses Research Articles

Machine learning-driven development of niobium-containing optical glasses

High refractive index glasses are essential for old and new optical systems, such as microscopes, telescopes and novel augmented reality lenses and micro projectors. However, a fair portion of these glasses use toxic components, such as PbO, BaO, As2O3, and TeO2, which lead to high refractive indexes and facilitate the melting operation, but are harmful for human beings and the environment. On the other hand, it is known that niobium significantly increases the refractive index and is a non-toxic element. The objective of this paper was to develop new optical glass compositions containing Nb2O5 with a relatively high refractive index (nd > 1.65), intermediate Abbe number (35 < Vd < 55) and fair glass transition temperature, Tg. To this end, we used a machine learning algorithm titled GLAS, which was recently developed at DEMA-UFSCar to produce new optical glasses composition. After running the algorithm 13 times, two of the most promising compositions were chosen and tested for their glass forming ability and other properties. The best composition was analyzed in respect to the refractive index, glass transition temperature and chemical durability. A comparison between the laboratory results and predictions of the artificial neural network indicates that the GLAS algorithm provides adequate formulations and can be immediately used for accelerating the design of new glasses, substantially reducing the laboratory testing effort. Also, the results indicate that niobium glasses might offer some advantages over its main competitor (La2O3).

Designing optical glasses by machine learning coupled with a genetic algorithm

Engineering new glass compositions have experienced a sturdy tendency to move forward from (educated) trial-and-error to data- and simulation-driven strategies. In this work, we developed a computer program that combines data-driven predictive models (in this case, neural networks) with a genetic algorithm to design glass compositions with desired combinations of properties. First, we induced predictive models for the glass transition temperature (Tg) using a dataset of 45,302 compositions with 39 different chemical elements, and for the refractive index (nd) using a dataset of 41,225 compositions with 38 different chemical elements. Then, we searched for relevant glass compositions using a genetic algorithm informed by a design trend of glasses having high nd (1.7 or more) and low Tg (500 °C or less). Two candidate compositions suggested by the combined algorithms were selected and produced in the laboratory. These compositions are significantly different from those in the datasets used to induce the predictive models, showing that the used method is indeed capable of exploration. Both glasses met the constraints of the work, which supports the proposed framework. Therefore, this new tool can be immediately used for accelerating the design of new glasses. These results are a stepping stone in the pathway of machine learning-guided design of novel glasses.

Data-driven design of glasses with desirable optical properties using statistical regression

In this study, we used a data-driven approach to build models for assisting the design of new glasses with high refractive index and low dispersion. Our models, which are based on multiple linear regression and kernel ridge regression, achieved high accuracy in predicting optical properties of glasses based on their composition alone. Using the predictions of these models as a guide, we fabricated new glasses in our laboratory. In agreement with model predictions, these glasses had promising optical properties. This work therefore demonstrates a successful example of data-driven materials design and can be used as a template for designing glasses or other materials with other desirable properties.

Determining the liquidus viscosity of glass‐forming liquids through differential scanning calorimetry

AbstractViscosity at the liquidus temperature (TL), ηL, is a critical parameter for the design of new glasses, particularly for industrial glass production where crystallization must be suppressed. However, a direct viscometric determination of ηL for a glass‐forming system is difficult due to crystallization. Here we propose an alternative approach for determining ηL through differential scanning calorimetry (DSC). Specifically, DSC is used to measure both the viscosity curve and liquidus temperature of a glass‐forming system and then derive its ηL value. The ηL values determined using DSC are found to be in excellent agreement with those measured through viscometry. The DSC approach is applicable to various glass‐forming systems covering a wide range of fragilities and ηL values spanning over five orders of magnitude. Other advantages of this approach are its accuracy and small sample requirements.

Machine learning for glass science and engineering: A review

The design of new glasses is often plagued by poorly efficient Edisonian “trial-and-error” discovery approaches. As an alternative route, the Materials Genome Initiative has largely popularized new approaches relying on artificial intelligence and machine learning for accelerating the discovery and optimization of novel, advanced materials. Here, we review some recent progress in adopting machine learning to accelerate the design of new glasses with tailored properties.

Machine learning for glass science and engineering: A review

The design of new glasses is often plagued by poorly efficient Edisonian “trial-and-error” discovery approaches. As an alternative route, the Materials Genome Initiative has largely popularized new approaches relying on artificial intelligence and machine learning for accelerating the discovery and optimization of novel, advanced materials. Here, we review some recent progress in adopting machine learning to accelerate the design of new glasses with tailored properties.

Tailoring structure and properties of silica glass aided by computer simulation

By using a combination of experiments and molecular dynamics simulations, our studies show that the elastic response of silica glass to initial compression gradually changes from abnormal to normal with increasing quench pressure, helium content or alkali modifier added in the glass matrix. We uncovered the structural origin of the elastic anomaly in silica glass as localized structural transitions between motifs of different stiffness that are similar to those found in its crystalline counterparts. Pressure-quenching, helium-stuffing, or alkali-modifying plays a different role in changing the structure of silica glass, but all of the resulting structures reduce the propensity for such local structural transitions to take place, thus the degree of elastic anomaly. Our studies demonstrate that by processing in ways that gradually eliminates the elastic anomaly, the degree of silica glass to undergo irreversible densification can be eventually eradicated. This provides a solid foundation for the bottom-up design of new glasses with tunable structure and properties.

Microscopic Origins of Compositional Trends in Aluminosilicate Glass Properties

Revealing and understanding the microscopic origins of the macroscopic properties of aluminosilicate glasses is important for the design of new glasses with optimized properties. In this work, we study the composition‐structure‐property relationships in 20 MgO/CaO sodium aluminosilicate glasses upon Al2O3‐for‐SiO2 and MgO‐for‐CaO substitutions. We find that some properties (density, molar volume, Young's modulus, and shear modulus) are linear through the investigated range of Al2O3 compositions, while others (refractive index, coefficient of thermal expansion, Vickers hardness, isokom temperatures, and liquid fragility index) exhibit a change in the slope around the composition with [Al2O3] = [Na2O], which is especially pronounced for the glasses containing MgO. We discuss these phenomena based on structural information obtained by NMR spectroscopy and topological considerations.

Designing Glasses with Tunable Structure and Properties by Computer Simulation

AbstractA normal solid becomes stiffer when squeezed and softer when heated. In contrast, silica glass behaves the opposite way: its elastic moduli decrease upon compression and increase upon heating. Silica glass is also known to densify under compression and radiations. These have been long-standing mysteries in materials science. Using molecular dynamics simulation, we uncovered the structural origins of the anomalous thermo-mechanical behaviors and mechanisms of permanent densification in silica glass. Accordingly, these anomalies can be attributed to localized structural transitions, analogous to those that occur in the crystalline counterparts. The irreversible densification in silica glass is achieved through structural transition involving bond breaking and re-formation under a combination of high pressure and temperature. We further revealed that the anomalous thermo-mechanical behaviors are inherently connected to the ability of the glass to undergo permanent densification. Our computer simulations demonstrate that by processing in ways that gradually eliminates anomalous thermo-mechanical behaviors, degree of the glass to undergo densification can be eventually eradicated. This provides the conceptual foundation for the bottom-up design of new glasses with tunable structure and properties.

The effects of radiation on the density of an aluminoborosilicate glass

Glasses used for nuclear waste immobilization are subjected to high levels of radiation, and this may affect their physicochemical properties. Alpha radiation is responsible for an important fraction of the radiation energy dissipated in these glasses. It has been reported previously that some borosilicate glasses increase their density during irradiation while the density of other glasses decreases. Although the density increase of silica after irradiation has been understood, thanks mainly to molecular dynamics calculations and diffraction experiments, the processes involved in more complex glasses could be more varied. In this work we irradiated an aluminum-borosilicate glass which is a candidate for the aforementioned purposes and which increases density during alpha irradiation from the 10B (n,α) 7Li reaction. We studied the effects of alpha irradiation on its microstructure, using several experimental techniques, and subsequently correlated the results. Small angle X-ray scattering (SAXS) measurements revealed the presence of inhomogeneities of about 10 Å in the untreated samples. After annealing these samples, TEM images displayed a contrast structure and helium pycnometry revealed density changes, both typical of glass phase separation. After irradiation, the glass density increased and the SAXS intensity decreased, indicating a compositional homogenization process in the samples subject to a higher dose of irradiation. Atomic displacements were calculated by means of the TRIM [1] computer code. The number of displacements produced by each 10B(n,α) 7Li reaction was estimated at 580 and involved distances of up to 15 Å. An increase in the density of the irradiated samples can be explained in terms of the atomic displacements produced by the nuclear reaction cascades of the reaction 10B (n,α) 7Li, in the scenario of pre-existing phase separation in the samples. In the case of the aluminum-borosilicate glasses studied here, which exhibit a fine phase separation, the density of the Si-rich phase increases with the incorporation of Na and B atoms. The B-rich phase also increases its density with the flow of Si atoms from the matrix. Vacancies created by irradiation in the glass structure, are responsible for a density decrease. The final effect is due to the sum of all contributions described, which in this case results in a net density increase of the irradiated samples. An understanding of this phenomenon can lead to the design of new glasses which overcome radiation with a minimum of density change.

Crystallization kinetics in Fe−B−Si metallic glasses

Crystallization kinetics of Fe-B-Si metallic glasses, as characterized by the effective activation energy Ec and the Avrami exponent n have been studied using Differential Scanning Calorimetry. The data show that in the hypoeutectic (Fe-rich) region, the glasses crystallize into α-(Fe,Si) and Fe3B in that order. The growth of crystallites is diffusion controlled, and Ec falls off rapidly away from the eutectic trough. In the hypereutectic region, both crystalline phases are obtained simultaneously. Thermal stability is greatly enhanced by addition of Si(∼3 at. %) and the growth process is no longer purely diffusion controlled. The behavior of Ec and n versus composition is discussed. Effects on Ec and n of additions of small amounts of Ni and Mo are also reported. Such additions have direct relevance to the design of new glasses for applications at high frequencies.