C. Patzig and T. Höche then present scanning and transmission electron microscopy, analytical tools, and the relevant analytical tools and other ancillary techniques available (Chapter 2.3).
From the results obtained with these methods, the next three chapters focus on the glass structure. Short‐range order is dealt with by J.F. Stebbins in a chapter where he defines such concepts as network formers and modifiers, bridging and nonbridging oxygens and their tetrahedral distribution around the network‐forming cations they coordinate, and he also discusses how they change with temperature and composition (Chapter 2.4). From the scale of atoms to that of macroscopic bodies, the increasing complexity found at longer distances is then examined by G.N. Greaves who also points out the basic differences found between oxide and metallic systems (Chapter 2.5). Because glasses have generally complex chemical compositions in both industrial and geological contexts, the relationships between the structures of simple and multicomponent systems are finally described by B.O. Mysen in a perspective aimed at understanding composition–property relationships (Chapter 2.6).
Returning to more theoretical aspects, the next chapter by P.K. Gupta stresses with the constraint theory the importance of simple topological considerations and of temperature‐dependent bond strengths to understand the glass structure and composition–property relationships. To conclude this section, two chapters review the rapid advances made in the growing field of atomistic simulations in two complementary ways. In the Monte‐Carlo and molecular‐dynamics simulations presented by A. Takada (Chapter 2.8), the interatomic potentials used to simulate the structure and properties of glass‐forming systems are determined a priori, with the advantage that up to a few thousand particles can be considered with current computing power. As reviewed by W. Kob and S. Ispas, atomic interactions are determined instead from first principles in the more fundamental and precise ab initio approaches, such as density functional theory; the price is that only relatively small systems can be currently investigated (Chapter 2.9). Ascertaining the energetics of a glass‐forming melt is fraught with considerable difficulties, however, because phase transformations, in which one is interested, are driven by small differences of typically a few kJ between the very large numbers yielded with significant uncertainties by ab initio methods.
Throughout this section the emphasis is generally put on oxide glasses. Metallic glasses are specifically dealt with in Chapter 7.10 and organic polymers in three other contributions (Chapters 8.8–8.10).
Reference
1 Plato, T. 55e–56d, transl. by D.J. Zeyl p. 1224–91 in Plato Complete Works. Indianapolis: Hackett Pub. Co., 1997.
2.1 Basic Concepts of Network Glass Structure
Alex C. Hannon
ISIS Facility, Rutherford Appleton Laboratory, Chilton, Didcot, Oxon, UK
1 Introduction
What is a glass from a structural standpoint? There are different answers dependent upon whether the emphasis is on structure, preparation methods, or thermodynamic properties. However, a simple structural definition is adopted here, according to which a material must be solid and have a noncrystalline structure to be called a glass.
Put simply, the description of a crystal involves a unit cell, containing a particular arrangement of atoms, which is then replicated periodically in three dimensions to build up the structure as illustrated by the 2‐D (two dimensional) representation of Figure 1. Crystal structures have translational symmetry because, if crystallite boundaries are neglected, the environment of a particular atom is the same as that of all equivalent atoms in all unit cells. Contrastingly, a glass lacks translational symmetry and a detailed understanding of its structure requires as much experimental information as possible. In addition to X‐ray (XRD) and neutron (ND) diffraction, inelastic neutron scattering and various other spectroscopies (X‐ray absorption, infrared, Brillouin, Mössbauer, and X‐ray photoelectron spectroscopies) may be used (Chapter 2.1), along with electron microscopy, physical property data (especially density), and theoretical modeling.
As is clear to any reader of this Encyclopedia, many types of materials can vitrify if they are solidified rapidly enough to avoid crystallization. Leaving aside metallic glasses (Chapter 7.10) or organic polymers (Chapters 8.7 and 8.8), however, the majority of useful glassy materials are formed by oxides or chalcogenides. This is the reason why this chapter is restricted to these materials and to the elementary concepts that are used to describe their structure.
These structures can be understood well in terms of the continuous random network (CRN) model, first propounded by Zachariasen [1], because the atomic bonds have some covalent (directional) character. The basic features of this model will thus be reviewed and related to fundamental structural information gathered for silica glass, the archetypal glass former, and the “mother” of all amorphous silicates. Because microcrystalline descriptions of glass structures in fact preceded Zachariasen's model, their basic limitations will also be summarized. The structural changes induced by the addition of so‐called network modifiers in oxide glasses will then be discussed at short‐ and medium‐length scales, along with the intermediate character of some oxides that may act as glass formers only when combined with some modifiers. Finally, the manner in which network glasses can depart from Zachariasen's model will be illustrated with chalcogenides.
List of Acronyms
2‐Dtwo dimensionalBObridging oxygenCRNcontinuous random networkFSDPfirst sharp diffraction peakIROintermediate‐range orderLROlong‐range orderMDmolecular dynamicsMRNmodified random networkMROmedium‐range orderNBOnon‐bridging oxygenNDneutron diffractionNMRnuclear magnetic resonanceRMCreverse Monte CarloSROshort‐range orderv‐SiO2 vitreous SiO2XRDX‐ray diffraction
Figure 1 Two‐dimensional representation of a crystal structure for a composition A2O3. Small dark spheres: A atoms; large light spheres: oxygens. Unit cell delineated by thick, dashed lines.
2 The Zachariasen–Warren Random Network Model
The structure of oxide and chalcogenide glasses is usually described by the Zachariasen–Warren random network model, thus termed because it was proposed by Zachariasen [1] in 1932 and subsequently supported by early XRD studies of glass structure made by Warren and coworkers (Chapter 10.11, [2]).
