the glass transition
/ Tg = tee-jee /
As a supercooled liquid is cooled, it gets thicker and thicker, like honey heading toward toffee heading toward rock. At some temperature the atoms are moving so sluggishly that, on the timescale of your experiment, they simply stop rearranging. From that point down the material is a rigid glass rather than a very stiff liquid. That crossover temperature is the glass transition, written Tg, and it is the moment the liquid structure gets frozen in for keeps.
The glass transition is not a true phase change like melting or boiling. Nothing snaps; no heat is suddenly absorbed. Instead, properties that depend on the liquid still flowing, such as the volume, change their slope smoothly but distinctly as you pass through Tg. Below Tg the glass contracts only by ordinary bond stretching; above it, extra room can open up as the network rearranges. By convention Tg is pegged to a viscosity of about 10^12 Pa s (roughly 10^13 poise), which is stiff enough that the glass behaves as a solid on human timescales. Because it depends on how long you wait, Tg is genuinely rate-dependent: cool ten times faster and you freeze the structure a little higher, so Tg creeps up by a few degrees.
Tg matters because it is the border between two worlds of processing. Above it, glass flows and can be blown, pressed, or drawn; below it, glass is elastic and brittle and must be cut or ground. Every named reference point on the viscosity curve, from the softening point down to the annealing and strain points, lives in the neighbourhood of Tg. For soda-lime window glass Tg is around 550 degrees C; for fused silica it is above 1100 degrees C, which is exactly why silica is so much harder to melt and shape.
Watch a glassblower: while the gather glows and is above Tg it stretches and swirls like taffy on the blowpipe; the instant it cools past Tg it stiffens and any further push cracks it, so the work must go back into the glory hole for reheating. Tg is literally the line between shape-it and shatter-it.
Above Tg glass is a workable liquid; below Tg it is a brittle solid. Crossing Tg switches the whole process.
Tg is not a fixed constant of the material; it shifts with cooling rate, so quoted values assume a standard rate. Treat Tg as a narrow band, not a razor-sharp temperature like a crystal's melting point.