Читать книгу Encyclopedia of Glass Science, Technology, History, and Culture - Группа авторов - Страница 247

For the future, the ideas about the extended structure of glass, introduced separately in this chapter, need eventually to be consolidated into a holistic description. This is implicit in the conceptual MRN, CRN, and DRPHS models [9, 13, 25] whose fixed geometries are infinitely perpetuated through rings of different sizes and modifier channels for insulating network glasses and through variations in icosahedra for densely packed metallic glasses. Nevertheless, for the last two decades of research, SRO, IRO, and LRO/MRO have been loosely defined in terms of static geometry between basic atomic building units, correlations between adjacent units, and more distant neighbors. This has created demarcations, often based on different experimental techniques, that avoids the central issue. The extended structure of glass should rather be seen as a continuous development from the microscopic atomic level through mesoscopic dimensions to the macroscopic scale relevant to applications.

The challenge in glass science is to associate, in a quantitative way, atomic structure with functionality [1, 3, 7]. In particular, heterogeneity is a common feature of functional glasses but is incompletely understood at the level of static structures. It has its antecedents in the rheology of the supercooled state and the dynamics of the glass transition. Empirical correlations have been reported, often controversially received, between the fragility m of supercooled liquids at the glass transition and solid‐state properties such as Poisson's ratio, between m and I_BP, and between m and f(0,T) [1]. Underlying all these relationships is the role played by DFs, which increase in amplitude at the dynamic crossover ~1.2T_g [26] and are frozen into the glassy state as α relaxation dynamics slow down. They appear to be central to nucleation processes, whether these are crystallization, phase separation, or polyamorphism. DFs have the dimensions of the acoustic wavelengths of the boson peak, which underscores, again, the cooperative nature of the many‐atom fast β dynamics in the condensation of glasses and their functionality.

Finally, the important technical drivers in meeting the challenge raised by the prospect of understanding the extended structure of glass as a predictor of functionality are the advances in experimental technology. First are those coming from light and particle beams in materials science – spallation neutron sources, coherent X‐ray sources, electron nanobeams, and atomic‐scale microscopy. These tools have already escalated in intensity and diversity in recent years, increasingly driven by the needs of the biosciences, where similar issues exist in handling complexity and aperiodicity on a grand scale. Second, and as important, are advances in high‐performance computing and the visualization of big data sets. Year on year larger and larger simulations are reported, sizes for MD, currently reaching several million atoms, already on a length scale commensurate with DFs. Ab initio techniques, which are the most reliable for predicting functionality, necessarily lag behind, but are currently heading toward the thousand atom mark – all of which concurs with the forward thinking of Alder and Wainwright, over 50 years ago, that “the behaviour of systems of many interacting particles cannot, in general, be dealt with theoretically in an exact way […] Since these difficulties are not conceptual but mathematical, high‐speed computers are well‐suited to deal with them” [29].