The extended structures of glasses generally define their functionality. In silicate and aluminosilicate glasses, for example, tetrahedral bonding through corner‐sharing oxygens ensures that the gap between the occupied oxygen p states and unoccupied antibonding sp3 levels is retained everywhere (this is the so‐called HOMO–LUMO gap between the highest occupied and lowest unoccupied molecular orbitals). The existence of relatively few mid‐gap defect states in turn guarantees visible and UV transparency for windows, optical components, laser glasses, etc. In metallic glasses, the cohesive potential between the delocalized electron gas and the ion cores is maintained isotropically throughout the bulk. These features underwrite electrical conductivity and, without the dislocations of crystalline metals, extremely high levels of mechanical hardness and toughness. Likewise, the spin–spin exchange interaction is retained in aperiodic metallic structures, supporting the ferromagnetism exploited in low‐loss magnetic metallic glass‐transformer cores. Additionally, because most glass formers – either metallic or not – derive from reasonably strong liquids, they have a wide supercooled liquid range and exhibit superplasticity at the softening point (108 Pa·s) but virtual rigidity at the glass transition (1012 Pa·s). Through Newtonian viscous flow, their monolithic structures enable the easy fabrication of components with essentially any shape, from the industrial dimensions of windscreens down to those of MEMS and nanotechnology.
By contrast, in crystalline materials even the largest industrial single crystals are extremely small by comparison to glass sheet. Large-area crystalline films comprise micron‐sized powders such that the facets of polycrystals generate a microstructure of interlacing grain boundaries where functionality generally resides. For example, the ductility of metals, the hardness of ceramics, the strength of steels, and the mesoscale magnetism of metal films mainly derive from the properties of grain boundaries that are often structurally disordered and anisotropic. By comparison glasses are structurally isotropic, and the cracks that affect their mechanical strength are usually restricted to the surface where the atomic structure terminates. The extended structure of glass links the SRO of atoms, molecules, and metallic clusters and geometrically underpins its intrinsic isotropic properties – optical, electronic, dielectric, electrolytic, magnetic, mechanical, etc. This can be visualized through computer‐simulated structures (Figure 1), where the open network of a silicate glass is contrasted with the dense close‐packed arrangement of a metallic glass. The respective atomic volumes are 7 and 11 Å3 and are typical of these very different families of glass, but where each share an extended defect‐free structure. Reviews of network glasses and metallic glasses can be found in [1, 3], respectively.
Figure 1 Visualizing the extended atomic structure characteristic of glass. (a) Molecular dynamics simulation of the network structure of NaKSi2O5 glass [1, 2]; Na and K atoms are dispersed within depolymerized silicate network, forming percolating alkali channels. (b) Reverse Monte Carlo simulation of the close‐packed structure of the metal‐metalloid glass Ni80P20 [3, 4]; P atoms also cluster, forming percolating channels through the Ni dense random packed structure. The scale bars indicate the start of long range order (LRO).
Source: (b) Reproduced from [4] © 2006 Nature Publications.
In this chapter we consider how several experimental techniques are required to access the extended structure of glass (Section 2), from diffraction and inelastic spectroscopies that reveal relationships between SRO, IRO, and LRO structure and dynamics, to microscopy that probes the average projected structure in real space. We then turn to the different types of structural order that characterize network and metallic glasses (Section 3): starting with the SRO, progressing through the configurations of adjacent local structural units that define IRO, and extending through MRO to LRO, the topology of larger agglomerations – clusters, rings, channels, and chains. Beyond these dimensions are those of density fluctuations (DFs) (Section 4), frozen into glasses from the liquid state, which reflect the degree of non‐ergodicity frozen in at the glass transition. In particular, DFs are the agents at supercooled temperatures that promote phase separation, either in density or in composition. Models of extended glass structure (Section 5) are next described and include conceptual models, devised before the advent of computational methods but still useful heuristically, and large computerized models that have been developed since. Using this approach, we show how structural heterogeneity in glasses (Section 6) can be modeled in terms of minority-component channels percolating through the majority network or metallic structure. Here, as elsewhere, the extended structure of glass is linked with its applications. Finally, we outline perspectives for future work (Section 7).
Acronyms
AFMatomic force microscopyBObridging oxygensBPboson peakCNcoordination numberCRNcontinuous random networkDAS NMRdynamic-angle spinning nuclear magnetic resonanceDFdensity fluctuationsDFTdensity functional theoryDRPHSdense random packing of hard spheresEXAFSextended X‐ray absorption fine structureHR TEMhigh‐resolution transmission electron microscopyINSinelastic neutron scatteringIXSinelastic X‐ray scatteringIROintermediate-range orderLROlong-range orderMAS NMRmagic angle spinning nuclear magnetic resonanceMDmolecular dynamicsMROmedium-range orderMRNmodified random networkNBEDnanobeam electron diffractionNBOnonbridging oxygensPSGphase‐separated glassesRMCreverse Monte CarloSROshort-range orderVDOSvibrational density of states
2 Extended Structure of Glass: The Need for a Multiplicity of Techniques
In addressing the extended structure of glasses, a wide portfolio of techniques has developed [1, 3, 5]. For many years the principal experimental method has been X‐ray and neutron scattering [6], initially concentrating on the radial distribution function (RDF) from which the radially averaged local structure T(r) can be determined, as illustrated in Figure 2 for silica glass and the metallic glass Ca60Mg25Cu15. The maxima identify interatomic correlations, first between nearest neighbors (SRO) defining the polyhedral or icosahedral building units and then between adjacent units (MRO or IRO), as spelt out in the cartoons. The SRO and IRO in glasses are often similar to their crystalline cousins. On the other hand, topology influences LRO – ring statistics for network glasses [1] and icosahedral packing for metallic glasses ([3], Chapter