Decoding the glass genome

https://doi.org/10.1016/j.cossms.2017.09.001Get rights and content

Highlights

  • Model-driven approach enables the improved and accelerated design of new glassy materials.

  • Combines physics-based and data-driven models, including machine learning techniques.

  • Approach for “decoding the glass genome” is demonstrated with the development of new chemically strengthened glass.

  • Additional fundamental research is necessary to continue advancing the field of glass science and technology.

Abstract

Glasses have played a critical role in the development of modern civilization and will continue to bring new solutions to global challenges from energy and the environment to healthcare and information/communication technology. To meet the accelerated pace of modern technology delivery, a more sophisticated approach to the design of advanced glass chemistries must be developed to enable faster, cheaper, and better research and development of new glass compositions for future applications. In the spirit of the U.S. Materials Genome Initiative, here we describe an approach for designing new glasses based on a mathematical optimization of composition-dependent glass property models. The models combine known physical insights regarding glass composition-property relationships together with data-driven approaches including machine learning techniques. Using such a combination of physical and empirical modeling approaches, we seek to decode the “glass genome,” enabling the improved and accelerated design of new glassy materials.

Introduction

Throughout human history, glass has been one of the most vital and influential of all materials, and the importance of glass is only growing [1]. While often taken for granted, glass windows enable visible light into buildings and vehicles while sheltering the occupants from harsh weather conditions [2]. Modern high-tech glasses, such as photochromic [3] and electrochromic [4] windows, can dynamically adapt to sunlight conditions to improve energy efficiency for applications in architecture and transportation. Moreover, the use of thinner, lightweight strengthened glass can improve fuel economy in vehicles without compromising safety [5].

In addition to glass windows, the development of glass lenses is another society-changing invention that has brought the world into focus for people suffering from vision problems. Glass lenses were also the essential invention for revolutionizing the field of astronomy and for enabling the discovery of microbiology [6]. Today, glass lenses are ubiquitous in computers and other personal electronic devices.

The invention of low-loss glass optical fibers was key in the development of the Internet, enabling exponentially growing levels of communication across the globe [7]. Glass has also played a revolutionary role in the display of information, from early televisions based on cathode ray tubes to modern flat panel displays [8]. As the resolution of these displays improves, the requirements on the high-tech glass substrates have become tighter and tighter [9]. New glasses will also be key for visualizing information through augmented and virtual reality devices.

Strengthened glasses with high chemical durability have been critical for healthcare applications such as pharmaceutical packaging [10]. Within the realm of healthcare, bioactive glasses have also been developed to address a wide range of medical problems, including bone repair, cancer therapy, soft tissue repair, and dental applications [11].

Owing to these many highly impactful applications of glass—addressing major global challenges in energy, communications, healthcare, information display, safety, and more—an argument has been made that we are now living in the Glass Age [12]. However, the current applications of glass represent only the beginning of what this versatile class of materials will offer.

Future applications necessitate the careful design of new glass compositions to meet stringent requirements for both product attributes and the properties required for large-scale manufacturing. To accelerate the design of new glassy materials, it is imperative that we make comprehensive use of available modeling tools [13]. The concept of the “materials genome” was first conceived by Prof. Zi-Kui Liu at the Pennsylvania State University [14] and then independently by Prof. Gerband Ceder and coworkers at MIT in the 2000s [15], [16]. The concept borrows from the idea of biological genomics, which involves the sequencing and analysis of the genomes of living organisms, with the overall goal of connecting the observable characteristics of the organisms with their underlying genetic chemistry. The concept of the materials genome is analogous to biological genomics, but applied in the field of materials science and engineering. In this case, the goal is to make quantitatively accurate predictions of material properties based on their underlying chemical makeup. For our current work, we use a combination of both physics-based and empirical models to help elucidate the origin of glass properties and pave the way toward composition optimization for emerging applications. By developing models for each property of interest, we can decode the “glass genome” to enable the design of new transformative materials to meet many of the grand challenges faced by the world today and in the future.

Section snippets

Glass design considerations

One of the defining characteristics of glass is its non-crystalline structure, lacking the long-range atomic ordering found in crystalline materials. Owing to this non-crystalline structure, a glass does not need to satisfy the same stoichiometric requirements as in crystal chemistry [17]. Coupled with the fact that nearly every element of the periodic table can conceivably be incorporated into a glass, there are an essentially infinite number of potential glass chemistries available. Since the

Model-driven glass design

Given a desired set of product- and manufacturing-related attributes that must be achieved, how can one obtain an optimized composition satisfying all of these requirements? Historically, researchers employed the so-called “cook and look” approach of glass design. Based on many years of knowledge and intuition and the available data in literature and lab notebooks, researchers would take an educated guess at an initial glass composition. This composition would be prepared in a crucible melt,

Model selection

Materials modeling broadly includes any model that predicts material properties or behavior based on the underlying chemistry and/or process conditions. Materials modeling may be performed at many different length scales (electronic, atomic, mesoscopic, or macroscopic) and across a wide range of time scales. Models may also incorporate different levels of physics, ranging from quantum mechanical models that capture detailed electronic band structures to purely empirical models that do not

Physics-based modeling

While purely empirical models do not assume that anything is known about the underlying physics governing material properties, often one can obtain models with broader applicability and better extrapolation ability by incorporating known physics, wherever possible. Glassy materials are well-known to pose significant challenges for physical modeling based on their three “non”s [25]. As non-crystalline materials, the atomic positions in a glass are not precisely known and must therefore be

Data-driven modeling

While physics-based models can provide accurate predictions of many glass properties, it is not always practical to rely solely on physical models when optimizing new glass compositions. Empirical modeling can help to supplement more theoretical approaches by making use of available experimental data. A wide variety of empirical models have been proposed for modeling the composition dependence of glass properties, including linear and polynomial regression [60] and various machine learning

Opportunities for future research

Despite the ubiquitous importance of glassy materials in our everyday lives and the many opportunities for developing new advanced glasses to address many of the important challenges faced by human society, the state of global glass research in most of the world today is regrettably weak. Fig. 5 plots the number of publications related to glass science and technology originating from the top ten most active countries in the field. The criteria and procedure for obtaining these results are

Conclusions

Glasses play a critical role in many areas of modern technology. The accelerated pace of modern technology development requires a new model-driven approach to materials design, making use of both high-quality experimental data and whatever physics are known regarding the composition dependence of material properties. By coupling together models for each material property of interest, the process of new glass design can be reduced to a mathematical optimization exercise. The success of this

Acknowledgements

I would like to acknowledge many valuable collaborations with A. Tandia, K.D. Vargheese, Y.Z. Mauro, D.C. Allan, A.J. Ellison, and M.S. Pambianchi at Corning Incorporated, as well as with M.M. Smedksjaer and Y.Z. Yue from Aalborg University. Some of the figures are provided courtesy Corning Incorporated. “Materials Genome®” is a registered trademark of Zi-Kui Liu.

References (71)

  • Y.D. Ha et al.

    Characteristics of dynamic brittle fracture captured with peridynamics

    Eng. Fract. Mech.

    (2011)
  • B.L. DeCost et al.

    A computer vision approach for automated analysis and classification of microstructural image data

    Comput. Mater. Sci.

    (2015)
  • A.K. Varshneya

    Fundamentals of Inorganic Glasses

    (2006)
  • M. Arbab et al.

    Glass in Architecture

    Int. J. Appl. Glass Sci.

    (2010)
  • N.F. Borrelli, N.T. Lönnroth, M. Prassas, P.M. Then, L.A. Zenteno, Multi-Pane Window with a Low Emissivity Layer and a...
  • T. Leonhard et al.

    Novel lightweight laminate concept with ultrathin chemically strengthened glass for automotive windshields

    SAE Int. J. Passenger Cars-Mech. Syst.

    (2015)
  • M. Bolt

    Glass: The eye of science

    Int. J. Appl. Glass Sci.

    (2017)
  • R.D. Maurer

    Glass fibers for optical communications

    Proc. IEEE

    (1973)
  • A. Ellison et al.

    Glass substrates for liquid crystal displays

    Int. J. Appl. Glass Sci.

    (2010)
  • Q. Zheng et al.

    Variability in the relaxation behavior of glass: impact of thermal history fluctuations and fragility

    J. Chem. Phys.

    (2017)
  • R.A. Schaut, W.P. Weeks Jr., Historical review of glasses used for parenteral packaging, PDA J. Pharm. Sci. Technol....
  • G. Kaur

    Bioactive Glasses: Potential Biomaterials for Future Therapy

    (2017)
  • D.L. Morse et al.

    Welcome to the glass age

    Int. J. Appl. Glass Sci.

    (2016)
  • J.C. Mauro et al.

    Accelerating the design of functional glasses through modeling

    Chem. Mater.

    (2016)
  • Z.-K. Liu

    Perspective on Materials Genome®

    Chin. Sci. Bull.

    (2014)
  • S. Curtarolo et al.

    Predicting crystal structures with data mining of quantum calculations

    Phys. Rev. Lett.

    (2003)
  • N. Nosengo

    The material code

    Nature

    (2016)
  • J.C. Mauro et al.

    A simplified eigenvector-following technique for locating transition points in an energy landscape

    J. Phys. Chem. A

    (2005)
  • S. Kirkpatrick et al.

    Optimization by simulated annealing

    Science

    (1983)
  • A.K. Varshneya

    Chemical strengthening of glass: lessons learned and yet to be learned

    Int. J. Appl. Glass Sci.

    (2010)
  • J.C. Mauro et al.

    Viscosity of glass-forming liquids

    Proc. Natl. Acad. Sci. USA

    (2009)
  • Q. Zheng et al.

    Viscosity of glass-forming systems

    J. Am. Ceram. Soc.

    (2017)
  • J.C. Mauro et al.

    Nonequilibrium viscosity of glass

    Phys. Rev. B

    (2009)
  • J.C. Mauro et al.

    Fictive temperature and the glassy state

    J. Am. Ceram. Soc.

    (2009)
  • P.K. Gupta et al.

    The laboratory glass transition

    J. Chem. Phys.

    (2007)
  • Cited by (102)

    • Materials genome strategy for metallic glasses

      2023, Journal of Materials Science and Technology
    View all citing articles on Scopus
    View full text