P.H. Gaskell

Cationic environment in silicate glasses studied by neutron diffraction with isotopic substitution

The method of neutron diffraction coupled with isotopic substitution is presented and recent investigations on the environment around cations in silicate (Ti in K2O·TiO2·2SiO2, Ca and Ni in 2CaO·NiO·3SiO2) and aluminosilicate (Li in Li2O·Al2O3·2SiO2) glasses are reviewed. The examination of the cation-centered pairs obtained from the first difference function presents striking similarities for all investigated cations. These functions indicate a well-defined short- and medium-range environment around cations. The local site generally presents a lower coordination number than that found in the crystals of similar composition. The environment around Ti in vitreous K2O·TiO2·2SiO2 corresponds to a square-based pyramid and direct TiO5–TiO5 linkages were observed experimentally in the second difference function, contrary to crystals. A detailed description of the cation site distortion for Li and Ca may be given by this method. The distribution of cations at medium range, which can be extracted by the double difference method, reveals the presence of cation-rich regions in silicate glasses. The cation–cation distances often indicate a two-dimensional character in the cationic organization. On the contrary, Li-aluminosilicate glass shows a more homogeneous cation distribution, in relation with the charge-compensating role of Li in this glass. This non-homogeneous distribution of cations may be related to the nano-inhomogeneities proposed in the models of supercooled liquids.

Structure and properties of glasses — how far do we need to go?

Abstract A thorough understanding of the properties of glasses in sufficient detail to offer predictions requires adequate knowledge of the atomic structure. This is common ground among those researching in this field. What is less clear is just what constitutes adequate knowledge of the structure of glasses — in order to adequately understand their properties? How far do we need to go? Is there any alternative to fully characterised and verified computer-generated atomic models with accurate interatomic potentials? For every major type of glass? We address these questions and review the different classes of physical and chemical properties in relation to the structural data available from experiment and simulation. The conclusion is that much can be done, as has already been demonstrated, to understand those properties that depend on the local or the long-range structure of the glass — optical properties and ionic diffusion in mesoscopic structures are examples. Investigation of medium-range structure is more hazardous. The difficulty of obtaining good structural data is matched by the limited usefulness of those data in predicting and controlling properties — unless the questions are very carefully chosen. When questions are properly posed, then results can give very valuable insights. Moreover, solutions to this part of the structural puzzle could lead to spectacular changes in fundamental concepts of the nature of glasses, amorphous solids and even liquids.

Cationic ordering in oxide glasses: the example of transition elements

Abstract Structural data have been obtained on the cation surroundings in multi-component silicate and borosilicate glasses using chemically selective spectroscopic and scattering methods, such as extended X-ray absorption and neutron scattering with isotope substitution (NSIS). Transition elements such as Ni or Ti may occur in unusual 5-coordinated sites which coexist with other coordination numbers, depending on glass composition. Distribution of cationic sites in the glassy structure is responsible for unusual spectroscopic properties, as shown by Fe2+ Mössbauer spectroscopy. The environment of cations such as Zn, Zr or Mo, has been determined by EXAFS and discussed using the bond valence theory, which predicts the way to charge compensate the oxygen neighbours and which indicates the linkage of cationic sites with the silicate framework. Cation-cation correlations are given by NSIS up to ~8 Å , indicating an extensive Medium Range Ordering (MRO) with corner- and edge-linked cationic polyhedra, for Ti and Ni-bearing glasses, respectively. This heterogeneous cationic distribution in glasses is consistent with the presence of two-dimensional domains in which cation mixing may occur, as shown in a Ca-Ni metasilicate glass. Three-dimensional domains have also been found by Ni-K edge EXAFS in the case of low alkali borate glasses, with a local structure which mimics some aspects of crystalline NiO. The presence of ordered cationic domains, clearly illustrated by Reverse Monte Carlo simulations helps to rationalize the physical properties of multi-component silicate glasses.

Local and medium range structures in amorphous alloys

Abstract Experimental evidence suggests that the local environment of certain atoms in amorphous alloys is well-defined. Thus the structures of transition metal-metalloid glasses appear to be based on a 9-atom polyhedron centred on the metalloid. There are some indications that this polyhedron (nearly always distorted) is best represented by a trigonal prism. There are no clear signs that directional bonds are important in defining the environment of, say, B in NiB alloys and isolated “molecular” trigonal prisms are not particularly more stable in a central force field than other candidate structures (e.g. Archimedean antiprisms). Both polyhedra are also relatively unstable unless they are fully capped. The existence of trigonal prismatic units in a-metals may thus be related to efficient, space-filling packing of atoms — in which trigonal prisms are a consequence of a “medium-range” organisational scheme. Several medium-range structures consisting of packed trigonal prisms, tetrahedra and octahedra are examined. These include structures based on crystals and non-periodic clusters with five-fold symmetry similar to the “curved space” structures proposed as models for amorphous semiconductors and oxide glasses. It is concluded that while details of the particular medium-range structure present in the glass cannot be established, as yet, a medium-range structure probably exists and represents a constrained subset of possible configurations. The implications for glass formation in transition metal-metalloid alloys are discussed and it is suggested that the composition-dependence of medium-range structural constraints are intimately involved.

A neutron diffraction, isotopic substitution study of the structure of Li2O·2SiO2 glass

The structure of Li 2 O.2SiO 2 glass has been studied in detail by neutron diffraction with isotopic substitution of Li. The first Si-O and O-O distances are 0.163 ± 0.001 and 0.266 ± 0.001 nm, respectively. Li is found in a relatively well-defined tetrahedral site, with 3.2 ± 0.2 oxygen atoms at 0.197 ± 0.001 nm and a further 0.8 ± 0.5 in a broad tail extending to about 0.22. The Li-Li first neighbour distance in a silicate glass has been measured directly, for the first time, and the value of 0.310 ± 0.002 nm yields strong evidence for a non-random distribution of the Li ions in the glass. The short- and medium-range structure exhibit considerable similarity to those of crystalline (c-) lithium disilicate.

Structure, glass formation and properties☆

Abstract Recent experimental results, chiefly from neutron scattering with isotopic substitution have shown that several silicate and phosphate glasses, containing cations of Groups I, II or transition metals, have a degree of local and medium-range ordering that is strongly reminiscent of compositionally-equivalent crystalline phases. The consequences of these results are considered and that it is shown that models based on the structure of the corresponding crystal allow a reasonable rationalisation of the structure of the glass, providing constraints associated with the liquid state are included. Borrowing structural principles associated with the crystalline and liquid phases places amorphous materials where they belong: centrally in the spectrum of liquids, glasses and crystals.

Anomalous wide angle X-ray scattering study of strontium silicate and aluminosilicate glasses

Anomalous wide angle X-ray scattering (AWAXS) is used to investigate the short and medium range order in Sr silicate and Sr aluminosilicate glasses. Strong SrSr correlations at 4 A are found in the silicate glass, whereas they are much weaker in the presence of aluminium. The first SrO distance and coordination number and further SrSr distances in the silicate glass are consistent with edge sharing octahedra. The SrAl correlations cannot be exactly determined, but the role of charge compensator imposed on the Sr by tetrahedral Al3+ rapidly reduces the order found in the silicate glass.

The environment of CA ions in silicate glasses

Abstract Although it is now clear that the local environment around strongly covalently bonded atoms like silicon or boron is almost as ordered in glasses as in crystals, there is much less information about the surroundings of “network-modifying” ions. The general view, stemming from the classic Zachariasen-Warren model, is that the latter are randomly distributed in suitable interstices within the silicon-oxygen framework of silicate glasses, and that such interstices are variable in shape and size. The local environment can be established directly in favourable cases by neutron scattering with isotopic substitution. Calcium has suitable isotopes and, as it is a typical alkaline earth “modifier”, the structure around Ca gives a valuable insight into the local structure in oxide glasses generally. Glasses containing almost 50% CaO have been examined by neutron scattering with substitution of Ca. The resulting correlation functions- comprising weighted sums of the Ca-centered partial functions only, - gives a direct representation of the first neighbour shell of oxygens. This is found to be well-defined and similar to that for crystalline CaSiO 3 . Information on the second neighbour shell is also obtainable and the CaCa first neighbour distance, inferred indirectly, does not correspond to random substitution but, again, is similar to that observed in the crystal.

Relationships between the medium-range structure of glasses and crystals

Abstract The known structure of a crystalline phase is almost always useful in investigating the unknown structure of the compositionally equivalent glass. For the local environment around elements like Si, B and P, the correspondence between site geometry and symmetry can be impressively close. Beyond near neighbours, any relationship becomes less obvious − at least in real-space data. Progress in understanding the medium-range structures of glasses has been painfully slow as a result. One essential clue is given by reciprocal-space features at low Q (scattering vector) in X-ray or neutron scattering data, which are clearly related to the medium-range structure. Interpretation of these features as ‘quasi- Bragg’ scattering allows direct comparison between the structures of the glass and equivalent crystalline phases. Applications of this method will be illustrated in borates and silicates, together with some chalcogenide glasses. Correspondence between low-Q features for these glasses and compositionally-equivalent crystals is qualitatively good. In some cases there is semi-quantitative agreement too. Thus the essential flavour of the medium-range structure of several typical glasses appears to be interpretable, rather easily.

The titanium environment in a potassium silicate glass measured by neutron scattering with isotopic substitution

Abstract The structure of a titanium potassium silicate glass has been studied by neutron scattering with isotopic substitution. The first difference function gives the Ti-centered partial correlation functions and indicates a square-based pyramid for the Ti site (with four oxygens at 1.96 A and 1 oxygen at 1.68 A). The second difference function is dominated by the TiTi signal and suggests some clustering of these polyhedra with a first TiTi distance of 3.5 A.