Thursday, December 10, 2009

When is Pyrex not pyrex?

When sold to American housewives, of course.

I recently had to look up the composition of Pyrex glass, and in the process I came across this interesting article in Consumer Affairs.

As everyone who has worked in a lab knows, Pyrex is a borosilicate glass. This is important because this glass composition has a much lower thermal expansion coefficient than standard soda-lime glasses, making it more resistant to rapid changes in temperature.

Evidently, Corning got out of the consumer products business a decade ago, and then sold licenced the US rights to the Pyrex brand to another company. They now use the pyrex brand on a non-borosilicate glass product. These new products are tempered ordinary soda-lime glass, which is more heat resistant than ordinary window glass, but can still explode, according to consumer reports.

Tempered glass is made by cooling the outer layers of glass faster than the glass can conductively equilibrate. As the core of the glass then cools, it contracts, drawing the outer layers of the glass into compression, while putting the core under tensile stress.

Figure 1. Tempering glass

Just like rocks, compression decreases the ability of fractures to propagate, so the glass is stronger.

The tempered ovenware manufacturer has been very lawyerly about catastrophic dish failure, which means that it is up to us bloggers to speculate about causes.

In order for tempered glass to work, the glass must not undergo annealing at its operating temperature. The compressed and tensile layers are not in equilibrium, so if the rate of internal diffusion becomes fast enough to allow the strain to anneal out, the glass will at best lose its temper, and at worst fail catastrophically.

Diffusion rates obviously temperature dependent, but they also depend on composition. We also know that on historical timescales, hydrogen diffuses into window glass while alkalis leach out.* So a testable hypothesis is this:

Some combination of H-rich foods and baking temperatures accelerates the diffusion of H into the bakeware. This lowers the annealing point below baking temperatures, causing differential annealing and catastrophic failure.

This hypothesis should be easily testable** by even the most cash-strapped geology departments, so if any of you folks are setting up summer undergraduate research projects, here’s your chance to combine rheology and home economics.

* My dad must have read this study when I was a kid, because he depth-profiled Na and H in the shards when I hit a baseball through the garage window...

** If you go for it, please BE SAFE! Exploding hot glass = huge safety issue.


John Spevacek said...

H-rich foods? So you mean baked sauerkraut? Anyone doing this is going to need a serious fume hood.

Kim said...

I don't think that breaking glass in an unconfined space would make a good student project. (Says the person who nominally advised the student who wanted to test whether Upheaval Dome could be a impact crater by shooting layers of - I forget what, now - with his grandfather's 30.06. Sometimes, one just has to get out of the way of student enthusiasm. Literally.)

I'm trying to think of how the diffusion of H through the glass would be measured. My first reaction was "if I had an SEM with EDS, we could cut sections through the glass and look at diffusion profiles." But H is too light for EDS. (Plus I don't have an SEM.)

EliRabett said...

Not quite sure, but if you mean acidic that should be amusing, just pour some vinegar into the dish and bake. The fact that the water will evaporate and concentrate the acid, will make it even more fun.

Chuck said...

Kim, one would hopefully design the experiment to confine any explosions to the oven. Which we blow stuff up in all the time.

And eli, starting with a high temperature acid, like sulphuric, lets you run all the way up to the dry annealing temperature. Once you get a temperature-time curve for that, you try citric, if the 170C sulfuric time is practical.

Kim, you can measure Na out with a SEM, H is generally done with depth profiling SIMS. For large distances, cross-sectional FTIR might work, but good luck cutting tempered glass to the right sample geometry.

Chris Phoenix said...

I've read of "pyrex" dishes exploding untouched at room temp. So it might not explode in the oven, then explode later even after it's cooled.

I'd think that a relatively safe setup would be to enclose the dish in metal insect screening, run the experiment, then smash it and throw away the pieces still contained. But the experimenter is responsible if this turns out to be insufficiently safe.