Phase separation and controlled crystallization of non-oxide glasses

This paper reviews the experimental results associated with the phase separation and controlled crystallization of non-oxide glasses, mainly chalcogenide glasses and ZrF₄-based fluoride glasses. Experiments demonstrate that the phase separation is not always the precursor of the glass-ceramic process. It has been found that some non-oxide glasses could be converted into glass ceramics without phase separation. Different mechanisms of controlled crystallization of non-oxide glasses are discussed.     

Synthesis of CuO---Al2O3---SiO2 glasses from the sol-gel method and properties

The synthesis of CuO---Al₂O₃---SiO₂ glasses by the sol-gel method was investigated and the thermal expansion of bulk glasses was measured. Phase separation was observed in these glasses and apparently consisted of a copper-rich dispersed phase and a network former-rich matrix phase. The low thermal expansion coefficient of these glasses was governed by the matrix phase.     

The influence of the crystallization process on the internal friction of some oxide glasses

The influence of phase separation and crystallization on the internal friction of some oxide glasses is reviewed and discussed. In alkali-containing glasses, the internal friction peak caused by stress-induced diffusion of alkali ions decreases in magnitude and shifts to higher temperature slightly due to phase separation. And in alkali-free glasses phase separation only exerts a minor decrease upon the background of internal friction curves, whereas crystallization influences the internal friction of these glasses more strongly. Because of crystallization, in alkali-containing glasses alkali ions might diffuse in a residual glass phase and a crystal phase, respectively. This might cause corresponding internal friction peaks. And in alkali-free glasses, no evident internal friction peak is observed. However, the author found a high and wide internal friction peak at about 100°C in the crystallized MgO·Al₂O₃·SiO₂·TiO₂ and ZnO·Al₂O₃·SiO₂·ZrO₂ glasses. The peak occurring in the two glasses studied is probably connected with glass crystallization and crystallized crystals.     

Low expansion copper aluminosilicate glasses

The thermal expansion coefficients of Cu₂O---Al₂O₃---SiO₂ glasses have been measured. These glasses have very low expansion coefficients similar to that for SiO₂ glass, but their liquids temperatures are much lower. It was possible to reduce the liquids temperature by the addition of 2 mol% of Na₂O while maintaining low expansivity. In order to explain the low expansivity, the effects of cation size, valence, the Cu²⁺/Cu⁺ ratio, bond strength and phase separation were examined. Phase separation was observed in these glasses which probably consisted of a copper-rich dispersed phase and a network former-rich matrix phase. It was concluded that the overall expansion coefficients of the glasses were governed by the low expansion matrix phase.     

Further investigations of the effect of replacing lithium by sodium on lithium silicate scintillating glass efficiency

Ce³⁺ doped lithium (⁶Li) silicate glasses are thermal neutron detectors. Prior work showed that when sodium (Na) is substituted for Li the scintillation efficiency, under beta particle stimulation, increased and then decreased as the sodium (Na) content was increased. When all the ⁶Li was replaced by Na no scintillation was observed. Raman spectra, acquired using a visible excitation source, provided no evidence of anomalous behavior. SEM microscopy did show some phase separation, but there was no obvious correlation with the scintillation efficiency. We have reexamined these glass samples using deep UV Raman excitation which reduces fluorescence interference. The newly acquired spectra show evidence of phase separation in the glasses. Specifically we see a peak at 800 cm^? 1 Raman shift which can be assigned to a vitreous silica moiety that results from phase separation. There is a strong correlation between this peak's area, the scintillation efficiency, and the Na content. The observed trend suggests that phase separation enhances scintillation and addition of Na reduces the amount of phase separation. We also see evidence of at least two defect structures that can be tentatively assigned to a three-membered ring structure and an oxygen vacancy. The latter is fairly strongly correlated with enhanced scintillation efficiency.     

Raman spectroscopic study of SbxSe100−x phase-separated bulk glasses

The structure of Sb_xSe_100−x bulk glasses is investigated with the aid of Raman scattering over a wide composition range. The Raman spectra of the glasses exhibit unusual features when compared with other structurally similar binary glasses, owing to the phase separation of the present glasses in a certain composition range. The evolution of the Raman spectra and the depolarization ratio of the polarized and depolarized Raman intensities are consistent with the phase separation effect. The present findings are discussed alongside with calorimetric data from the literature that have been used up to now to extract structural information in an indirect way. The capability of Raman scattering as a tool for investigating phase separation in homogeneous media is demonstrated.     

Evaluation of phase separation in glasses with the use of atomic force microscopy

Glass forming melts frequently exhibit liquid–liquid immiscibility resulting in phase separation. The chemical and spatial variation of phase separated morphologies in glasses can range from a few angstroms to microns, often requiring very high magnification for detection. Historically, phase separated glasses have been characterized by transmission electron microscopy (TEM). This technique is very time consuming and costly, requiring specialized equipment and training. Atomic force microscopy (AFM) provides an inexpensive alternative to TEM and has proven to be a powerful tool in the characterization of phase separation in glasses. AFM provides rapid and accurate evaluation of the type, degree and scale of phase separation in glasses down to the nanometer level. Using a combination of topographical and phase imaging AFM we were able to quantitatively determine the microstructures of phase separated glasses with a resolution down to 50 nm. Additionally we were able to quantitatively confirm the time dependence of the chemical segregation and growth processes for phase separation in glass by spinodal decomposition. This paper will present sample preparation techniques and results for evaluation of phase separation in alkali borosilicate and sodium silicate glass systems.     

Bestimmung von Diffusionskoeffizienten aus der Röntgen-Kleinwinkelstreuung und Lichtstreuung von Gläsern

It is shown how to conclude from small angle X-ray scattering and light scattering of phase separated glasses what kind of phase separation process is occurring in the glass. In the cases of diffusion controlled particle growth and of spinodal decomposition, methods were presented permitting to find out from diffraction experiments the diffusion coefficients underlying these decomposition processes. The methods were demonstrated with three different glasses.     

Second-order optical non-linearity initiated in Li2O–Nb2O5–SiO2 and Li2O–ZnO–Nb2O5–SiO2 glasses by formation of polar and centrosymmetric nanostructures

Amorphous nanoheterogeneities of the size less than 100 Å have been formed in glasses of the Li₂O–Nb₂O₅–SiO₂ (LNS) and Li₂O–ZnO–Nb₂O₅–SiO₂ (LZNS) systems at the initial stage of phase separation and examined by transmission electron microscopy, small-angle X-ray and neutron scattering. Both LNS and LZNS nanoheterogeneous glasses exhibit second harmonic generation (SHG) even when they are characterized by fully amorphous X-ray diffraction (XRD) patterns. Chemical differentiation and ordering of glass structure during heat treatments at appropriate temperatures higher T_g lead to drastic increase of SHG efficiency of LNS glasses contrary to LZNS ones in the frame of amorphous state of samples. Following heat treatments of nanostructured glasses result in crystallization of ferroelectric LiNbO₃ and non-polar LiZnNbO₄ in the LNS and LZNS glasses, respectively. Taking into account similar polarizability of atoms in LNS and LZNS glasses, the origin of the principal difference in the second-order optical non-linearity of amorphous LNS and LZNS samples is proposed to connect predominantly with the internal structure of formed nanoheterogeneities and with their polarity. Most probably, amorphous nanoheterogeneities in glasses may be characterized with crystal-like structure of polar (LiNbO₃) phase initiating remarkable SHG efficiency or non-polar (LiZnNbO₄) phase, which do not initiate SHG activity. It gives an opportunity to vary SHG efficiency of glasses in a wide rage without remarkable change of their transparency by chemical differentiation process at the initial stage of phase separation when growth of nanoheterogeneities is ‘frozen’. At higher temperatures, LiNbO₃ crystals identified by XRD precipitate in LNS glasses initiating even more increase of SHG efficiency but visually observable transparency is impaired.     

The structure of borosilicate glasses studied by Raman scattering

Raman spectra of alkali and alkaline earth borosilicate glasses are reported. These spectra are used to discuss the molecular structure of the glasses. The influence of Al₂O₃ additions on the structure of borosilicate glass is also discussed. It is shown that the same type of groups are present in borosilicate glasses as in borate and silicate glasses. The presence of large borate groups such as tetraborate and metaborate groups is strongly suggested by the Raman spectra. It appears that boron ions are hardly taken up in the silicon-oxygen network. Our results suggest that the region of phase separation is larger than the region presently acknowledged.     

Formation and structure of titanate glasses

The formation of high titanium oxide (30–60 wt%) containing glasses was studied. Stable titanate glasses with high content of TiO₂ and BaO could be obtained even without other glass-forming oxides. It is demonstrated that the coloration of titanate glasses with high content of TiO₂ is due to oxygen loss during the melting. A systematic study of controlling melting and heat treatment conditions led to the successful decoloration of titanium oxide containing glasses. Infrared spectra and X-ray diffraction studies showed that Ti⁴⁺ in titanate glasses is in sixfold coordination. Phase separation was observed by electron microscopy when the titanate glasses were heat-treated at the temperature above T_g. The crystallization of titanate glasses is generally preceded by phase separation. The obtained crystalline phases are mainly different titanates.     

Amorphous nanostructuring in potassium niobium silicate glasses by SANS and SHG: a new mechanism for second-order optical non-linearity of glasses

Potassium niobium silicate (KNS) glasses xK₂OxNb₂O₅(1−2x)SiO₂ with x=0.167; 0.182; 0.200; 0.220 and 0.250 have been subjected to prolonged heat treatments in a wide temperature range above T_g. As a result, glasses exhibiting liquid-type phase separation phenomena have been isolated. Moreover for each glass composition, the temperature zones have been determined to produce transparent, opalescent or opaque materials which have been studied by X-ray diffraction (XRD), small-angle neutron scattering (SANS) and second harmonic generation (SHG) techniques. SANS data unambiguously point at nanostructuring of KNS glasses in the scale of 5-20 nm under appropriate heat treatments near T_g. In contrast to initial KNS glasses, nanostructured glasses exhibit SHG activity. For earliest stages of phase separation SHG-active glasses are characterized by fully amorphous XRD patterns. Further development of phase separation in glasses with increasing of their opalescence leads to diminishing SHG, and subsequently partial crystallization takes place giving opaque materials. Since relative maximum of SHG efficiency corresponds to non-crystalline nanostructured glasses, such new transparent second-order non-linear media may be of both scientific and practical interest. With regard to non-crystalline structure of nano-inhomogeneities, SHG mechanism in the glasses is supposed to be due to a combination of third-order non-linearity with a spatial modulation of linear polarizability.     

SAXS study of the micro-inhomogeneity of industrial soda lime silica glass

The microstructure of a variety of industrial multicomponent soda lime glasses for container and glazing and of a lead alkali silicate glass for pressed tableware articles was analyzed using small angle X-ray scattering (SAXS). No inhomogeneity with size >0.5 nm was found in the industrial samples, while ternary soda lime glass samples with a suitable composition when heat treated developed a phase separation which SAXS detected without difficulty. The results obtained disprove earlier reports which attributed problems of rheology and strength affecting container production (‘bad workability’) to microstructure formation.     

The structure of some chalcogenide glasses

A study using replica electron microscopy, scanning electron microscopy, electron diffraction, X-ray diffraction, sputter etching and differential thermal analysis of the structural properties of glasses having a range of compositions within the AsTeGeSi quaternary system has shown that phase separation generally occurs in the bulk material. The second phase is dispersed as inclusions in the non-crystalline matrix and is apparently non-crystalline away from the boundary of the glass forming region but is crystalline near to the high Te boundary. The crystalline second phase is probably Te. This quaternary system, AsTeGeSi is often used as source material for the fabrication of devices studied for their monostable electrical switching behaviour and the phase separation observed in these materials needs to be considered in the interpretation of the switching properties and for failure analysis of such devices.     

The structure of phosphate glass evidenced by small angle X-ray scattering

Numerous phosphate glass systems with compositions covering the entire glass forming range have been examined by small angle X-ray scattering. The experiments were carried out in a widely extended interval of s values between 0.055 and 31 nm⁻¹. A small angle scattering effect was detected for all samples investigated with the exception of the aluminium metaphosphate glass. The small angle scattering recorded indicates the presence of two differently-sized heterogeneity regions of electron density. The small-scale heterogeneities are present in magnesium and zinc phosphate glasses only. Their size is about 1 nm diameter in the magnesium and about 2 nm in the zinc phosphate glasses. The species do not result from liquid–liquid phase separation. Examinations to extract information on the nature of these heterogeneity regions are described in detail. There is no fundamental knowledge on the large-scale heterogeneities established. The small angle scattering suggests that their occurrence is related to water contamination. Anomalous small angle X-ray scattering experiments of a series of zinc polyphosphate glasses and a strontium metaphosphate glass sample are performed to examine the distribution of the Zn or Sr atoms, respectively.     

Non-isothermal crystallization and nanostructuring in potassium niobium silicate glasses

Potassium niobium silicate (KNS) glasses the composition of which is characterized by the K₂O/Nb₂O₅ molar ratio ranging from 0.85 to 1.2 and SiO₂ 50-54 mol% were examined in order to clarify the influence of chemical composition on formation of transparent nanostructured state of glasses. Differential thermal analysis, X-ray diffraction and scanning electron microscopy were used to study the non-isothermal crystallization of the KNS glasses as well as their morphological features. It was found that all glasses devitrify in three steps forming unidentified phases at the first two ones while at higher temperature (1000-1100 °C) the crystallization of K₃Nb₃O₆Si₂O₇ takes place. For prolonged heat treatment time (more than 5 h) at high temperature (1050-1100 °C) the transformation of this phase into the KNbSi₂O₇ ferroelectric one occurs in some extent. Nanostructuring occurs at the first stage of the devitrification process. It results from two partially overlapped processes: amorphous phase separation and subsequent crystallization. It was shown that only for the glass with the K₂O/Nb₂O₅ molar ratio equal to 0.85 and SiO₂ 50 mol% it is possible to separate the above processes by isothermal heat treatments at 680 °C obtaining fully transparent nanostructured samples. These samples contain nanocrystals 10 times smaller than the amorphous inhomogeneities of the phase separated matrix in which are dispersed.     

Raman and infrared spectra of As2Sx chalcogenide glasses with x ⩽ 3

Raman laser and far infrared spectra of As₂S_x glasses with x ? 3 are given and discussed. The purpose of the work is to bring a vibrational spectroscopic contribution to the study of these glasses and to the hypothesis, still under discussion in the literature, that they might be constituted either by a homogeneous vitreous phase or by a mixture of As₂S₃ and As₄S₄.Our results confirm a phase separation, with formation of ß-As₄S₄, below a certain value of x, which depends not only on the preparation method of the samples but also on other factors such as melting time. Laser irradiation of ß-As₄S₄ modifies its Raman spectrum. Such a phenomenon is attributable to two principal factors, either a partial polymerization or formation of a species richer in arsenic. A structural and formation model of the As₂S_x glasses is given, starting from a more generalized structural model of vitreous As₂S₃ which is an accord with the vibrational results and those by the diffraction method in the literature.     

Synthesis and characterization of M2O–MgO–WO3–P2O5 (M = K, Rb, Cs) glass system

Transparent glasses of the composition M₂O–MgO–WO₃–P₂O₅ (M = K, Rb, Cs), corresponding to the crystalline phases of M₂MgWO₂(PO₄)₂, have been prepared and studied by Raman and IR spectroscopy as well as DTA. Moreover, the thermally stimulated depolarization and dc conductivity have been measured. The glass transition temperature is 797, 795 and 773 K for the K-, Rb- and Cs-containing glass, respectively. Raman and IR studies have shown that these glasses have very similar structure. The main building blocks are pyrophosphate groups, WO₆ octahedra and magnesium–oxygen polyhedra. The dc conduction in these glasses is controlled by hopping of small polarons. The potassium containing glass was shown to be very stable whereas the rubidium and cesium glasses have significantly higher tendency for crystallization and phase separation. It seems, therefore, that the potassium containing glass is a suitable material for the preparation of samples containing non-linear and ferroelectric nanocrystals of the K₂MgWO₂(PO₄)₂ phosphate.     

Time-resolved fluorescence measurements on Eu3+- and Eu2+-doped glasses

Various fluoride, phosphate and borosilicate glasses with known properties and global structure have been doped with Eu³⁺ (4f⁶) at doping concentrations between 10¹⁸ and 10²¹ cm^?3. By applying reductive melting conditions Eu³⁺ could partially be transformed to Eu²⁺ (4f⁷). Four fluoroaluminate glasses with varying phosphate content between 0 and 20 mol% (FPx), a pure phosphate glass (P100) and two borosilicate glasses with low (DURAN^®-like) and high optical basicity (NBS1) have been used for these investigations. The time-resolved fluorescence in the visible range has been studied for both ions. Although static and dynamic fluorescence of Eu³⁺ and Eu²⁺ are dramatically different (f–f-transitions for Eu³⁺; d–f-transitions for Eu²⁺), glass structure changes have a similar influence on the dynamic fluorescence behavior of both ions. A strongly ionic surrounding due to fluorine ligands as in fluoroaluminate glass samples provides the longest fluorescence lifetime (about 7 ms for Eu³⁺; about 1.3 μs for Eu²⁺). Increasing phosphate content decreases the fluorescence lifetime due to more oxygen ligands. Interesting differences have been found for the two borosilicate glasses due to the difference in their optical basicity (Na₂O/B₂O₃ ratio). Measurements indicate a homogeneous distribution of europium ions in most FP samples. NBS1 measurements suggest that two different local europium sites are formed. For Duran-like samples only one specific europium site was found, although these samples show phase separation at high doping concentrations into a SiO₂-rich phase and borate- and europium-rich droplets. Fluorescence quenching due to energy transfer from Eu²⁺ to Eu³⁺ could be found for co-doped samples; Eu³⁺-doped samples show no fluorescence quenching.