Homogeneous ice nucleation rate of deeply supercooled water
In plain words
Pure water droplets can stay liquid well below 0 C because ice needs a seed to start. How fast ice appears on its own at very low temperature, and which form of ice appears first, is still uncertain.
Precise statement
Determine the homogeneous ice nucleation rate $J(T)$ in pure H2O between about 230 and 240 K, where measurements by different techniques disagree by several orders of magnitude and simulations (seeding with TIP4P/Ice and similar models) depend on CNT-based extrapolation. Decide whether CNT with a temperature-dependent ice-water interfacial free energy $\gamma(T)$ reproduces $J(T)$, and confirm whether the critical nucleus is stacking-disordered ice.
What would settle it
Mutually consistent measurements of J(T) by independent techniques and simulations with an accurate water model that reproduce them without fitted interfacial free energies.
Status in the literature
Unverified note
Seeding simulations reproduce measured rates within several orders of magnitude and indicate stacking-disordered critical nuclei; precise agreement is lacking (2026).