Lanthanoides in Glass and Glass Ceramics

1 Introduction

Glass is one of the oldest materials used by mankind. The first glass was produced in ancient Egypt about 4,000 years B.C. Nowadays, it exists in plenty of types and it can be either a mass product with millions of tons of annual production, but also a specialty product with only a few kg of annual production. Currently, about 100 million tons of glass is produced worldwide. There are different types of glass [1]:

1. Container glass (44 % total market amount by weight);

2. flat glass (29 % total market amount by weight);

3. glass fibers (12 % total market amount by weight); and

4. specialty and technical glass and glass ceramics (15 % total market amount by weight).

Glass fibers and in particular specialty and technical glass can contain a much larger number of elements than container glass and flat glass. When the composition of the glass is developed, one of the unique material properties of glass is used. Glass as an amorphous component can dissolve almost all elements in considerable amounts. Therefore, glass chemists have developed glass compositions, containing almost all elements of the periodic system of elements. Modern glass compositions can consist of more than 15 intentionally added cations, each playing a unique role in the final glass or glass ceramics.

Nevertheless, being in a metastable state, glass has always a tendency to crystallize. While this is an unwanted effect for typical container and flat glass or glass fibers, the crystallization of glass is willingly triggered in the case of glass ceramics in order to get new materials with new and specific properties.

Due to the very broad and large application of specialty glass, lanthanoides can be added to the glass for many different reasons, the most relevant being:

1. Coloring agents, amounts <2–3 wt% [2]

2. Mechanical properties, amount <10–15 wt% [3]

3. Tuning optical properties, amount <60 wt% [4]

4. Crystallization agent, amount <3 wt% [5]

5. High Abbe number glass, amount <50 wt%

6. Laser glass, amount <5 wt%

Regarding all aspects, a precise knowledge of the chemical composition is mandatory in order to establish lanthanoide recycling strategies.

2 Literature survey of rare earth chemical analysis in glass

3 Analytical methods for the determination of main components of glass (except lanthanoides)

Details of X-ray fluorescence (XRF) and ICP are presented in “Rare Earth Elements” edited by Alfred Golloch. Comparing methods, it is to be mentioned that it is easier to prepare a multielement standard solution for ICP than to produce or acquire a solid standard for XRF or ICP analysis. On the other hand, direct measurements of solids need less time that a procedure with decomposition of the solid. The reproducibility of XRF is better than of ICP. The detection limits are lower for ICP, especially ICP-MS.

There are some specific features of glass analysis:

In the case of XRF, glass can either be measured directly or, because of its relative low melting point, converted to a fused bead by remelting without fluxing agent. The benefit is a higher sensitivity that means lower detection limits compared to sample preparation by diluting with a fluxing agent.

Some glass contains B₂O₃ or Li₂O or both of them. If a complete analysis of the glass sample (a sum of 100 %) is required, these components have to be determined. For physical reasons, it is impossible to measure Li₂O by XRF. B₂O₃ can be determined in glass samples prepared without fluxing agent. Because of an information depth of the boron radiation below 1 μ m, it is necessary to have planar polished surfaces [39]. Otherwise, the B₂O₃ determination is performed as the Li₂O determination, e.g., by ICP-OES in solution after decomposition (see next chapter).

For the analysis with ICP, glass can be decomposed by hydrofluoric acid digestion which leads to an acid solution free of Si (in most cases, the main component of the glass) and B. Therefore, a higher sample weight can be used which lowers the detection limits.

SiO₂ is determined gravimetrically after melting the sample with sodium carbonate, dissolving the melt in HCl and precipitation of the SiO₂. The precipitated SiO₂ is filtered, ignited, and then weighed [40, 42].

B₂O₃ is determined by ICP-OES after melting the sample with sodium carbonate, dissolving the melt in HCl and filtering from precipitated material [41, 42].

Fluorine in glass can be determined by XRF or by wet chemistry [42, 43]. One digestion method is the pyrohydrolysis in a tube furnace at 1,050°C with superheated water vapor in presence of a catalyst. The fugitive fluorine components are distilled into water. The second method is sintering the glass powder with a mixture of sodium carbonate and zinc oxide at 1,000°C. After cooling, the melt is extracted. In both cases, the fluoride ion can be measured electrometrically (fluoride-sensitive electrode), by ion chromatography or photometrically (alizarin–ethylenediaminetetraacetic acid complex).

4 Preparation of sample solutions for glass analysis by ICP-OES

5 ICP-OES analysis of rare earth elements

ICP-OES measurements can be performed according to DIN 51086-2 [45]. Specific wavelengths for some rare earth elements are given (Table 1). DIN 51086-2 is being extended for the missing rare earth elements (Table 2).

6 Analysis of special optical glass

Two samples of optical glass were analyzed in ISC by hydrofluoric acid–perchloric acid digestion described in Chapter 4.1.3. SiO₂ and B₂O₃ were determined according to DIN [40, 41]. The results are given in Table 3.

It can be seen that the sum of determined components for both samples is near to 100 %. Due to the fact that the measurement precision for ICP-OES is higher compared to ICP-MS, definitely no important component was overlooked.

Li₂O cannot by determined by XRF. The measurement of B₂O₃ is difficult or impossible in case of borate-fused beads. The XRF quantification to 100 % can hardly be performed without information from other methods, e.g., ICP-OES.

7 Analysis of glass by topochemical analysis

A precise analysis of element contents in glass and glass ceramics is obtained by methods described above. Sometimes, it might be useful to determine the element contents locally too, i.e., on a scale of micrometers or nanometers. This is not necessary for glass of a high-grade purity and homogeneity, but e.g., for glass ceramics. Glass ceramics are formed by tempering glass that contains nucleating agents, such as titanium or zirconium oxide. Crystals are formed during a defined heat treatment because of these nucleating agents. At the end of the production process, a microstructure, consisting of crystals and glass is formed (Figure 1). During development of glass ceramics, local analyses of the elemental concentrations are mandatory to understand what happens during heat treatment. In case of recycling of glass ceramics, the total concentration of the single elements rather plays the predominant role. To measure them exactly, it is necessary to chemically dissolve first the glassy as well as the crystalline phases and then analyze them by ICP. The knowledge of the chemical composition of the various glassy and especially crystalline phases may be useful here to develop appropriate chemical dissolution and precipitation routines. For these two reasons, i.e., for the development of glass ceramics on one hand and for the development of chemical dissolution and separation processes on the other, a closer look at a sample is needed by using the so-called topochemical analysis EDX. Topochemical analysis EDX is usually performed in two ways: First, an EDX device is mounted on a scanning electron microscope (SEM) or a transmission electron microscope (TEM). Second, it can be combined with a WDX device, which is either a stand-alone device or it is mounted on a SEM. The functionality of these two methods is described in other chapters of this book. This chapter describes the procedure of the analysis and should be considered in more detail.

Figure 1

Transmission electron micrograph of a glass ceramic. Image Fraunhofer ISC.

In both methods, a focused electron beam impinges on the sample surface and penetrates into the sample. By inelastic interaction processes between the primary electrons (the electrons of the electron beam) and the electrons of the atoms of the sample, heat energy arises (amongst others) that leads to a local heating of the sample. The degree of heating depends on the kinetic energy of the primary electrons, the current intensity (beam current), i.e., how many primary electrons per unit time are incident on the sample surface, as well as the cross section of the electron beam, i.e., the size of the incident on the sample. In extreme cases, temperatures of 1,000°C may arise during the analysis by EDS in TEM. Figure 2 shows the resulting temperature during the analysis by EDS in TEM, depending on the current strength and thermal conductivity. It may happen that samples melt or vaporize locally when no appropriate care in the analysis is taken.

Figure 2

Possible local sample temperatures during the analysis in TEM by EDS, depending on the beam current and the thermal conductivity k [51].

During analyses in a SEM by EDS or WDS, the resulting temperatures are lower, but can still be significant (Figure 3).

Figure 3

Local sample temperature depending on the diameter of the electron beam and the electron beam power [47].

The local stoichiometry changes by melting or even vaporizing, but this is trivial. It is also soon obvious that increased atomic mobility – that means increased diffusivity – arises due to the increased temperature. But another effect happens, namely the formation of a space charge inside the sample. It arises because here a part of the primary electrons is quasi landfilled. Spray and Rae described this effect in 1995 [47]. The following (Figure 4) is obtained from this article.

Figure 4

Formation of a space charge zone during the EDS or WDS analysis [47].

Due to the increased diffusivity by temperature rise and the driving force created by the space charge region, the very mobile ions such as Na⁺, but also Li⁺, diffuse deeper into the sample and the stoichiometry in the analysis area changes.

Lanthanoides are not directly affected by this effect due to their usually relatively low diffusion rate. But as the concentration of other elements changes in the environment of lanthanoides, a false concentration of lanthanoides is measured. This effect should be also taken into consideration.

There is another important parameter to be considered during local analysis: the pressure. In the electron microscope is a very low pressure of about 10–5 mbar.

If the vapor pressure of some elements is very high, then a change in the stoichiometry is carried out by evaporation also.

To prevent these two effects, a consistent cooling of the samples is recommended during the analysis in order to avoid an increase in temperature and therefore an increased diffusivity.

A final aspect concerns the different hydrolytic resistance of glass. This has to be considered in the preparation that should be water-free; it should be carried out with petroleum for example. If water is used and the glass is very sensitive to an aqueous attack, then the alkali and alkaline earth metal ions are especially leached. With the standard soda-lime glass for windows, changes of 10 % in the sodium concentration may arise.

In summary, the following recommendations for a good local chemical analysis with EDS and WDS can be made:

1. Use cooling stages (liquid nitrogen) to avoid high sample temperature.

2. Avoid water during grinding and polishing.

3. If possible, avoid high beam current and focusing of the beam.

4. Observe the time dependent peak development during the measurement.

5. Don’t forget that the surface of glass can be strongly different from the bulk.