Silicate glasses have been used in practical applications for thousands of years, with glass-based relics dating back to the Stone Age. Modern glasses are often treated with chemical additives, a modification process known as "tempering" that can increase the hardness of the material while maintaining its optical transparency. Characterizing this process is paramount toward understanding the final results and micro-XRF is a straightforward and convenient method for performing this analysis.
One popular modification takes place via exposure of the glass to heat and potassium salt. This results in a potassium (K+) - sodium (Na+) ion exchange, with K+ flowing into the glass and Na+ flowing into the solution via diffusion.1 The introduction of potassium into silicon (Si) rich glass results in a change in its chemical structure, resulting in the formation of a compression-tension stress gradient that is highest at the material surface.1 This makes the glass more resistant to fracturing and to scratching, with the internal stress gradient discouraging microcrack propogation.1 This toughening process can result in stark increases in mechanical hardness, with many commercially tempered glasses reaching 9H on the Mohs scale: a value similar to tungsten carbide or silicon carbide.
As the toughening process occurs via a diffusive process, it has several degrees of freedom. Factors that have been shown to influence the tempered glass final properties include the processing temperature, processing time, and the composition of the salt.1,2
Conventionally, a potassium nitrate salt is used to treat the glass, though research has shown the addition of calcium may result in a mechanically superior product. However, it has also been shown if the calcium concentration is too high it can have a negative impact.2 For this reason, the concentrations of salts must be carefully controlled to ensure the glass produced is of the optimal scratch/fracture resistance and hardness.
An understanding of the material’s compositional profile allows researchers to gain a better understanding of the diffusion process used and this, in turn, can help to optimize the properties of the glass produced.
The rest of this article covers a study into the characterization of three different tempered glasses, sourced from different manufacturers, used to understand the diffusion behavior in different processes. Specimens used in the study were commercially-sourced mobile phone screen protectors, each with a self-reported Mohs hardness of 9H.
Micro-X-ray fluresence spectroscopy (micro-XRF) was used to analyze the specimens' global and local elemental compositions and to carry out a quantitative comparison that highlighted differences in the product formulations and the processes used to make them.
Micro-XRF was carried out using a Sigray AttoMap laboratory micro-XRF spectrometer. The AttoMap uses unique developments in optics and X-ray production to deliver a spatial resolution at the ~10 µm level and elemental sensitivity that reaches below parts-per-million (sub-ppm).
Specimens used for investigation were commercially-available smartphone screen protectors, which were analyzed along their machined edge in order to characterize their elemental profiles. XRF was chosen because it helps mitigate the impact of contamination and surface-specific features on the overall results.
In addition, the use of charge-neutral X-ray photons prevented charging issues from occurring (which are common when working with bulk insulators like glass), meaning extensive specimen preparation was not necessary.
The X-ray source was set to use a copper (Cu) target operating at 50 kV. Specimens 1-DT and 2-GG were analyzed over a 50 µm wide region with an entire cross-section, whereas 3-HG was analyzed over an 800 µm wide region. Each collected data set had background subtraction and peak-fitting applied to reduce noise and facilitate accurate quantitative analysis.
Figure 1. Optical micrographs of the tempered glass cross-sections: 1-DT, 2-GG, and 3-HG.
Results and Discussion
Quantitative analysis of each specimen's exposed surface layer allowed the weight percentages of the main components in each specimen to be compared (Table 1).
Table 1. Weight percentage of elements measured within the bulk of each glass using the fundamental parameters approach. 1-DT and 2-GG are measured to have similar compositions, while 3-HG exhibits a different composition.
Specimens 1-DT and 2-GG were found to have similar concentrations of Si, Ca and K, with trace levels of Ti, whereas 3-HG had a noticeably different composition that also included Sn, Pt and Al. This suggested that 1-DT and 2-GG were designed using similar principles, whereas 3-HG may have been produced using a different approach.
These measurements gave bulk information on the specimens, with the following mapping measurement providing localized information.
The AttoMap’s localized elemental mapping was used to determine the elemental composition of the specimens throughout their cross sections. ‘Heat maps’ that show the distribution of Si, Ca and K throughout specimen 1-DT are shown in Figure 2, where high brightness corresponds to high concentration. The same is shown for 2-GG in Figure 3.
Figure 2. Micro-XRF mapping of the 1-DT cross-section. A blended tri-color image is shown, as well as individual contributions from Si, K, and Ca.
Figure 3. Micro-XRF mapping of the 2-GG cross-section.
In general, each cross-section profile shows a similar elemental distribution throughout the cross section, with each specimen showing a multi-layer structure of a silicon-rich layer, a gap, and a Si, K and Ca containing layer. The furthest right section of each map was identified as the tempered layer and, for this reason, was chosen as the point of interest for subsequent investigation.
Further analysis of the data for 1-DT and 2-GG showed there were significant differences between the specimens when examined for K/Ca concentration gradients. As the tempering processes used relies on the diffusion of K and Ca into the Si-rich glass it may be expected that the highest K and Ca concentrations should occur on the specimens' exposed surfaces.
For 1-DT (Figure 2) the Ca and K signals were at similarly high levels on the RHS and LHS of the toughened layer, with reduced levels between these sides. For 2-GG (Figure 3) there was a quasi-periodic concentration gradient with alternating high, low, high, low signals across the specimen from left to right. This suggests that the toughened layer may have been produced using multiple independently tempered layers; each tempered from one side only, e.g. by the deposition of Ca/K followed by heat treatment to encourage diffusion. Fusion of these individual layers could then have been achieved via thermal treatment/fusing before bonding to untreated glass.
Figure 4. Micro-XRF mapping of the 3-HG sample.
Upon examination of the 3-HG specimen some unique features were noted (Figure 4). There was some localization of Ca and K at each of the specimen's surfaces and at its midpoint which, using the same logic discussed previously, could be assumed to be initial points of contact between K/Ca (e.g. as Ca-doped KNO3) and the bulk Si glass.xxxxxxxxxxxxxxxxxxxxxxx
If this assumption is correct then the results suggest that the Ca did not uniformly diffuse through the material, resulting in Ca “hot spots” and indicating chemical modification may not have been complete. The K signal was also found to be at its highest on either edge of the specimen and in the specimen's center, suggesting that 3-HG may have been produced using a similar fabrication method as 2-GG, i.e. by fusion of two independently tempered layers.
The 2D heat maps were converted to 1D numerical plots, which show the global concentration profiles of Ca and K more effectively. The whole width of each measurement was averaged across the cross-section to convert each 2D map into a 1D average concentration profile that ran from left to right. Normalization of the concentration values was carried out to provide maximum values for each dataset, yielding a quasi-quantitative, numerical comparison between the three different specimens.
It was found that the average concentration gradients of K and Ca follow similar trends for each specimen (Figure 5), which is consistent with the diffusive method used to produce them.
However, restricting the plot to the primary diffusing species (K) and to the diffusion zone more differences can be seen (Figure 6). This viewpoint shows the two phases of K-concentration in 2-GG and 3-HG, and the single phase in 1-DT. This result confirms numerically the differences observed in the 2D maps, providing a quantitative insight into the impact of different fabrication methods on the concentration gradient.
Figure 5. Left graph plots the normalized concentration of Ca and right graph plots the normalized concentration of K for all three samples. Concentration gradients of Ca and K follow similar trends for each specimen and are consistent with expected behaviour of a diffusing species.
Figure 6. Zoom-in of the plot of potassium (K) within the difussion zone to more clearly show the spatially-resolved variation in K concentration as a function of position. 1-DT has a single peak in concentration, while 2-GG and 3-HG have two periods.
The similarity in the data for 2-GG and 3-HG indicates they may have been fabricated using similar methods, whereas the different profile of 1-DT indicates a different method may have been used.
This research showed how the Sigray AttoMap microXRF can be used to quantitatively characterize differences in the composition of different tempered glass specimens of the same reported hardness.
The AttoMap’s micron-scale resolution and sub-ppm detection limit allowed the concentrations of different trace additives to be measured and the gradients of diffused species to be determined, including the detection of regions of incomplete chemical modification.
The research found that two of the glass specimens had similar elemental compositions, with the other having a significantly different composition. The averaging of each columnar result to produce a 1D global concentration profile allowed differences and similarities in each specimen’s Ca and K profile to be observed. This data suggested that two of the glass products could have been fabricated using a similar approach, with the third product being produced using a different fabrication method.
All three glass specimens were stated by their respective manufacturer to have the same hardness, showing that different methods of fabrication, and the diffusion profiles they produce, can still result in similar mechanical properties.
The insight that the Sigray Attomap micro-XRF spectrometer provides allows researchers to rapidly and accurately determine elemental composition, facilitating a better understanding of the diffusion processes and chemical treatments used to mechanically strengthen commercial glasses.
References and Further Reading
- R. Gy “Ion exchange for glass strengthening,” Materials Science and Engineering: B, vol. 149, no. 2, pp. 159–165, Mar. 2008.
- V. M. Sglavo, A. Talimian, and N. Ocsko, “Influence of salt bath calcium contamination on soda lime silicate glass chemical strengthening,” Journal of Non-Crystalline Solids, vol. 458, pp. 121–128, Feb. 2017.
This information has been sourced, reviewed and adapted from materials provided by Sigray, Inc.
For more information on this source, please visit Sigray, Inc.