Random first-order transition theory

Random first-order transition theory (RFOT) is a theoretical framework for understanding the glass transition as a true thermodynamic phase transition, rather than a purely kinetic phenomenon. Developed from the work of Kirkpatrick, Thirumalai, and Wolynes in the 1980s, RFOT extends spin glass concepts to structural glasses, proposing that the glass transition is a random first-order transition—a transition to a state of broken ergodicity with a discontinuous order parameter.

In RFOT, the configurational entropy of a supercooled liquid decreases with cooling not because the system loses states gradually, but because it becomes dynamically trapped in a finite number of metastable states. At the Kauzmann temperature T_K, the configurational entropy would vanish in equilibrium, signaling an ideal glass transition. In practice, the system falls out of equilibrium at a higher temperature, but the underlying thermodynamic singularity at T_K shapes the dynamics and provides the driving force for the observed glass transition.

The theory predicts the existence of mosaic length scales—domains beyond which the system can rearrange independently—and relates the activation barriers for relaxation to the configurational entropy. RFOT has been successful in predicting the scaling of relaxation times and the onset of dynamical heterogeneity, though direct experimental tests remain challenging.

RFOT is the most mathematically sophisticated theory of the glass transition, but its sophistication is also its vulnerability. The theory is built on mean-field approximations that may not survive in finite dimensions, and its predictions rely on extrapolations that are as bold as the ones that created the Kauzmann paradox itself. A theory that resolves a paradox by making equally strong assumptions has not resolved it—it has relocated it.