Two new papers by Corning Incorporated glass scientist John Mauro and his colleagues reveal the science behind satisfying football-watching. (Credit: Corning Museum of Glass)
According to the calendar, it is fall, known to many of us as football season. That means it is also flat screen season. During halftime, while football wonks, such as those above (credit: pennstatenews), are deconstructing the first half and updating you on the other games around the nation, take a moment to consider the science behind the LCD glass substrates of your flat screen TV. Two recent papers in the ACerS International Journal of Applied Glass Science by a Corning Inc. team show us that good science done well is essential to commercial success.
Designing new oxide glasses for the next generation of widescreens, smartphones, and tablets calls for a “delicate simultaneous tuning of a wide range of different properties, such as viscosity, coefficient of thermal expansion, elastic moduli, chemical durability, etc.,“ according to one of the papers. Corning glass scientist and paper coauthor John Mauro says in an email that the most important properties to consider are liquidus temperature and melt viscosity, that is, the balance between thermodynamics and kinetics.
Liquidus temperature matters for commercial glass producers. Lower-melting compositions require less energy, but obviously this cannot come at the price of compromising on glass quality. Molten glass exiting the furnace must be at a temperature above the liquidus temperature to prevent crystallization. Mauro says in an email, “Crystallization, of course, means that the liquid has failed to become a homogeneous glass, and it also leads to significant problems in any glass manufacturing process.” However, he continues, “What is more important than the liquidus temperature is really the liquidus viscosity, i.e., the viscosity of the glass-forming liquid at its liquidus temperature.”
Why? The more fluid the glass is, the more mobile the atoms within are, and the more easily the glass can crystallize. High-viscosity melts present a large kinetic barrier to crystallization.
The first paper, “Liquidus temperature of SrO–Al2O3–SiO2 glass-forming compositions,” (DOI: 10.111/ijag.12017) reports results from a two-year study at Corning on the system, which is an important ternary for commercial manufacture of glass for LCD substrates—SrO is a constituent in all LCD glass, for example. Surprisingly, there was no phase diagram in the ACerS-NIST Phase Equilibria Diagrams database for this ternary. The group focused its attention on commercially relevant areas of the phase diagram and determined liquidus temperatures and primary devitrification phases for 24 compositions with SiO2 contents of 65–80 mol% and alumina contents of 10–15 mol%.
They found that liquidus temperature decreases as the SrO/Al2O3 ratio increases. Similar trends occur in the MgO–Al2O3–SiO2 and CaO–Al2O3–SiO2 systems. In an era of simulation, the authors determined this the old-fashioned way. “All in all, this took about two years of fairly tedious work, since many hundreds of samples had to be made of various compositions at various temperatures, and each one had to have X-ray diffraction analysis to determine the identity of any crystalline phase,” Mauro says.
With the phase diagram providing information about thermodynamic possibilities, the next question is whether the kinetics are favorable. For glassmakers, the most important parameter is viscosity at the liquidus temperature. The Corning researchers were particularly interested in the correlation of viscosity with temperature and composition. Accurate prediction of viscosity is important for processing but challenging because it ranges over 12 orders of magnitude between melting and forming temperature regimes.
The complexity of multicomponent oxide glasses prevents first principles modeling, so the Corning group used topological constraint theory (pdf) to establish a “phenomenological model” for viscosity based on just two parameters—glass transition temperature and liquid fragility. According to the paper, which is set to appear in the next issue of the Intl. J. App. Glass Science, “The model is based on temperature-dependent constraint theory, where the composition is treated in terms of a network of bond constraints.”
Turning to Corning’s extensive database, the group tested the model against 7,141 actual viscosity measurements for 760 silicate glass compositions with three to eleven oxide components.
A plot of actual isokom (constant viscosity) temperature against predicted isokom temperature revealed a root mean square error of only 6.55 K in the isokom temperature, effectively validating the topological model. (For a good online review of viscosity as it relates to glass, see Missouri S&T professor R. K. Brow’s Cer. Eng. 103 notes (pdf).)
“Putting these two articles together, we have a complete picture of both the thermodynamic (i.e., liquidus temperature) and kinetic (i.e., viscosity curve) contributions to liquidus viscosity,” Mauro concludes in his email.
I can hardly wait for Saturday. Go Illini!
The second paper is accessible through “Early View” on the Wiley website (free to all ACerS members). See “Topological model for the viscosity of multicomponent glass-forming liquids,” (DOI: 10:1111/ijag.12009)