Chapter 1: History of Germanate Glasses Free

Silicate glasses have been produced by natural processes since the formation of the earth.¹  The earliest humans produced tools and weapons by shaping naturally occurring silicate glasses. The production of synthetic silicate glasses with controlled compositions began several thousands of years ago. Examination of the literature reveals that silicate glasses have been the subject of far more study than any other family of glasses.^1,2 

Production of glasses based on other oxides is relatively new. The earliest report of a borate glass found by this author was published by Grenet³  in 1897, who reported that the thermal expansion coefficients of sodium borate glasses exhibit a minimum at about 20 mol% Na₂O. This minimum, which was confirmed in many later studies, gave rise to the term “borate anomaly” to indicate that such a minimum was not expected from the literature on similar silicate glasses.¹  The designation of maxima or minima in properties as a function of glass composition as “anomalies” has since been extended to numerous other property/composition trends.¹ 

The first report of a germanate glass dates from 1926, when Dennis and Laubengayer⁴  obtained a high quality vitreous GeO₂ sample by melting in an alundum crucible with a Pt sleeve liner in an induction furnace. Attempts to produce this glass prior to their work had only yielded samples with many impurities and other defects. The quality of the glass produced by Dennis and Laubengayer⁴  is confirmed by the agreement between their values of the density refractive index, and thermal expansion coefficient and those found in many subsequent studies of this glass (see Chapter 3).

In 1962, papers by Ivanov, Evstrop’iev and Dorokhova^5,6  and by Ivanov and Evstrop’iev⁷  reported the first observations of anomalies in the properties of alkali germanate glasses. Ivanov, Evstrop’iev and Dorokhova^5,6  reported the existence of maxima at about 15 and 10 mol%, respectively, in the densities of sodium and potassium germanate glasses. In a separate study, Ivanov and Evstrop’iev⁷  also reported a maximum in the density of the potassium germanate glasses as well as maxima in the refractive indices of the sodium and potassium germanate glasses. Maxima in the refractive indices were found at about 15 mol% Na₂O for the sodium germanate glasses and at about 12 mol% K₂O for the potassium germanate glasses.

The pioneering work of Ivanov and co-workers^5–7  was confirmed in 1964 by Murthy and Ip⁸  for both the densities and refractive indicies. Murthy and Ip extended the earlier work to include lithium, rubidium, and cesium germanate glasses. Maxima in the density were found at about 20 mol% Li₂O, 15 mol% Na₂O, 10 mol% K₂O, 15 mol% Rb₂O, and 22 mol% Cs₂O. The maxima in the refractive index were found at about 19 mol% Li₂O, 15 mol% Na₂O, and 12 mol% K₂O, Rb₂O, and Cs₂O. The positions of the maxima in both density and refractive reported by Murthy and Ip⁸  for the sodium and potassium germanate glasses were identical to those reported earlier by Ivanov and co-workers.^5–7 

In 1970, Nemilov^9,6  reported the existence of two anomalies in the transformation range viscosity of sodium and potassium germanate glasses. The temperature corresponding to a viscosity of 10¹² Pa s was found to drop very rapidly with the addition of very small amounts of Na₂O or K₂O. The initial drop was followed by an increase to a maximum at higher alkali oxide concentrations. The initial minimum was confirmed in 1972 by a detailed study of the viscosity of glasses containing between 0.022 and 2.8 mol% Na₂O.^10,6  The initial minima, or high GeO₂ anomalies, occur at 2 to 3 mol% Na₂O and K₂O. The maxima, or high alkali oxide anomalies, occur at 17 mol% Na₂O and 15 mol% K₂O.

Nemilov’s results were confirmed in 1974 by Shelby,¹¹  who extended the earlier work to include the rubidium and cesium germanate glasses. The existence of both anomalies was found for all four alkali germanate systems. The high GeO₂ anomaly occurs at 2 to 3 mol% R₂O for all four alkali oxides. The high R₂O anomaly occurs between 15 and 20 mol% Na₂O, at 18 mol% K₂O, and at 20 mol% Rb₂O. Since only 3 compositions were studied for the cesium germanate glasses, the location of the high R₂O anomaly could not be determined with any certainty.

The glass transformation temperature, T_g, would be expected to exhibit behavior identical to that of the transformation range viscosity. Hanada et al.¹²  reported the existence of a maximum in T_g at about 18 mol% Na₂O for sodium germanate glasses in 1973. Their work did not include any glasses with compositions containing less than 5 mol% Na₂O, i.e. did not include the region of the high GeO₂ anomaly. Shelby¹¹  published the results of a more detailed study of the thermal expansion behavior of sodium germanate glasses the following year. He found a minimum at about 2 mol% Na₂O, followed by a maximum at 18 mol% Na₂O. This work was extended to include potassium, rubidium, and cesium germanate glasses in 1975 and lithium germanate glasses in 1987.^13,14  In every case, a minimum was found at 2 to 3 mol% R₂O, with maxima at 17 mol% for the K₂O, Rb₂O, and Cs₂O germanate glasses and at 19 mol% Li₂O for the lithium germanate glasses.

In 1971, Shaw¹⁵  reported maxima at 16 mol% Na₂O in the bulk, shear, and Young’s moduli of sodium germanate glasses. The Knoop hardness of these glasses also exhibits a maximum at 16 mol% Na₂O. The existence of maxima in these moduli for sodium germanate glasses was confirmed in 1973 by Hanada et al.,¹²  although the locations of the maxima found in their work are uncertain due to the lack of data for glasses containing between 12.5 and 20 mol% Na₂O.

The existence of anomalies in the d.c. electrical conductivity of alkali germanate glasses is not as clear as those in the properties discussed above. The electrical conductivities of sodium and potassium germanate glasses were first reported by Ivanov et al.^5,6  in 1962. The isothermal electrical conductivity of the sodium germanate glasses increases by about 2 orders of magnitude upon addition of only 1 mol% Na₂O. The sharp initial increase in electrical conductivity is followed by a region where the electrical conductivity increases very slowly with increasing Na₂O content up to about 10 mol% Na₂O, followed by a region of much more rapid increase. In contrast, the electrical conductivity of the potassium germanate glasses initially decreases with K₂O concentration, passes through a minimum at 5 to 7 mol% K₂O, and then increases rapidly until it reaches a value near that of the corresponding sodium germanate glass at about 25 mol% K₂O.

The electrical conductivity of glasses is determined by the concentration of ions per unit volume and ionic diffusivity of the mobile ions present in the glass.¹  The concentration of these ions expressed as ions per unit volume is determined by the molar concentration of these ions and the density of the glass. The molar concentration is determined by the formula of the glass and increases monotonically as the alkali oxide concentration, expressed in mol%, increases, while the densities of germanate glasses pass through maxima, as discussed above. While these maxima in density do decrease the rate of increase of the ionic concentration, they do not reverse the direction of the trend for alkali germanate glasses.

The diffusivity of the mobile ions, however, can either increase or decrease with increasing concentration of those ions. In 1963, Evstrop’iev^16,6  reported that the ionic diffusivity of Na²³  determined by tracer measurements in sodium germanate glasses passes through a minimum between 5 and 10 mol% Na₂O, i.e. displays an anomaly. Analysis of the electrical conductivity data for the alkali oxide germanate systems using the Nernst–Einstein equation reveals that minima exist for all five systems, i.e. anomalies also exist in the ionic diffusivity of those glasses.¹  The depth of these minima increases from very shallow to very deep as the atomic number of the alkali ion increases.

Minima are also found in the permeability and diffusivity of helium in sodium, potassium, rubidium, and cesium germanate glasses.¹⁷  These minima decrease in depth and shift to lower alkali oxide concentrations as the atomic number of the alkali increases.

The thermal expansion coefficient is the only basic property of alkali germanate glasses which does not display clear evidence of an anomaly. While the thermal expansion coefficient either increases very slowly (K, Rb, and Cs) or passes through a small minimum between 5 and 10 mol% R₂O (Li and Na) before increasing more rapidly with increasing alkali oxide concentration, it is difficult to determine if the observed behavior can be attributed to a germanate anomaly.^11,13,14  Although this behavior suggests the existence of an anomaly in the thermal expansion coefficient, the lack of data from multiple sources precludes a more definitive conclusion.

Models for the structure of germanate glasses will be discussed in more detail in the following chapter. The materials presented here are intended to indicate the historic development of these models.

Vitreous germania is believed to be isostructural with vitreous silica in that both glasses consist of a network of tetrahedra connected at all four corners to form a continuous structure. This model for the structure of vitreous germania was first proposed by Warren¹⁸  in 1934 based on X-ray diffraction measurements. Warren found that the Ge–O distance was 0.165 nm, which is slightly longer than the Si–O distance of 0.160 nm. Leadbetter and Wright¹⁹  in 1972 improved the resolution of the diffraction pattern over that of Warren and found that the Ge–O distance was 0.174 nm, which agrees with that reported in 1969 by Lorch,²⁰  from neutron diffraction measurements. Leadbetter and Wright¹⁹  also determined that the intertetrahedral angle was about 133°. Micoulaut et al.,²¹  who reviewed the data on the structure of vitreous germania, suggested that this angle, which is smaller than that of vitreous silica, indicates that vitreous germania has a larger concentration of small (3-member) rings than found in vitreous silica.

The first model for the structure of alkali germanate glasses was proposed by Ivanov and Evstrop’iev⁷  in 1962 and by Evstrop’iev and Ivanov²²  in 1963. Murthy and Kirby²³  followed with an infrared study in 1964 which they interpreted as supporting the Ivanov/Evstrop’iev model. This model, which will be designated as the classic model throughout this book, proposes that vitreous germania is isostructural with vitreous silica, as discussed above. Addition of alkali oxides does not result in the formation of non-bridging oxygens (NBOs) as occurs for alkali silicate glasses but in the conversion of Ge–O tetrahedra to octahedra with 2 neighboring alkali ions for charge balance. It is assumed that these octahedra must be connected to 6 tetrahedra, i.e. two octahedra cannot share a bond, and that their formation will cease once the concentration of octahedra reaches a value which would require violation of that condition. Once the maximum permissible concentration of octahedra is reached, further additions of alkali oxides will result in the reversion of octahedra to tetrahedra with one or more non-bridging oxygens.

Henderson and Fleet²⁴  presented the first significant challenge to the classic model in 1991. They proposed that the initial addition of Na₂O to GeO₂ results in the formation of small, 3-member rings of germanium–oxygen tetrahedra. Germanium–oxygen tetrahedra with one non-bridging oxygen are formed as well. The maximum in density occurs when the network becomes saturated with 3-member rings, after which tetrahedra with one NBO continue to be formed. With further Na₂O additions, tetrahedra with two NBOs begin to form, increasing in concentration until the network breaks down and the limit of glass formation is reached.

Henderson and Wang²⁵  expanded the work of Henderson and Fleet²⁴  to include all five alkali oxides. They suggest that the shift in the position of the density maximum toward lower R₂O contents in the order Li > Na > K is due to a decrease in the concentration where the network becomes saturated with 3-member rings. Although this concentration would be expected to decrease even further for Rb₂O and Cs₂O, the greater masses of the Rb⁺ and Cs⁺ ions counter the decrease in network density due to depolymerization which occurs after saturation of the 3-member ring concentration, resulting in a shift of the density maxima toward higher R₂O concentrations. They also suggest that formation of 5-fold germanium–oxygen coordination units may occur at higher R₂O concentrations.

There have been many studies since that time which expand on this model. Most of these studies have focused on the role of 3-membered rings in the germanate anomaly and on the existence 5-fold or 6-fold germanium–oxygen units in the network. These studies will be discussed in detail in the following chapter.

Many other germanate glass systems exist. Alkaline earth germanate glasses are believed to have structures similar to those of alkali germanate glasses. Unfortunately, since the alkaline earth germanate glasses are phase separated²⁶  over the region from 1 or 2 to over 20 mol% RO, it is impossible to determine if this is true in that region. Other systems, e.g. alkali aluminogermanate glasses, demonstrate that addition of Al₂O₃ to the glass composition results in elimination of at least some of the germanate anomalies (see Chapter 11). Systems which do not contain modifier oxides, e.g. binary SiO₂–GeO₂ glasses, must have very different structures from those of the alkali germanate glasses. The models which have been proposed for these and other systems will be discussed in the following chapter.

Properties of some germanate glass systems have been reported in multiple papers by a variety of authors, while those of other systems have only been reported in one or two papers. The data for these cases where there are multiple sources frequently exhibit a significant degree of scatter, which makes the determination of the position and magnitude of the germanate anomalies difficult. In these cases, the author has used his judgement based on over 50 years of studying the properties of glasses to determine the “best values” for the property as a function of glass composition. These values are not determined from a mathematical fit to the data, but simply represent the opinion of the author of this book. The best values are listed in tables in the appropriate chapters and are used in many figures where the properties of more complex glasses or glasses containing less-common species are compared to those of more basic glasses. Best values were not determined for cases where only two sources are available.

While many germanate glassforming systems have been extensively studied by numerous researchers, others have been the subject of very limited work. In some cases, there has been only one report of glass formation and any properties of glasses in a specific system. In these cases, the results are designated as “unconfirmed” or described using similar terminology. This designation is not intended to imply that the data are incorrect, but simply to note that additional study is needed to confirm the reported values.

In a few cases, where there are a substantial number of studies of a given property for a given glass system, the results from one or more studies obviously differs significantly from the others. If this difference exceeds the normal limits of expected variability for that property, that work is designated as “questionable” or even as probably incorrect. The reader is free to form their own opinion regarding these studies.