# Question Regarding Nucleation in Phase Change

1. When liquid turns into solid a nucleus is required. Is the same requirement present for solid-liquid (melting)?
2. Do condensation also require nucleation sites like in freezing? When I look at descriptions for phase equilibrium or condensation, the vapour phase just “goes” into the liquid phase. If I cool a closed container with no nucleation sites where liquid-vapour are in equilibrium, should I expect the vapour be supersaturated (kind of like supercooling a liquid)?
• For question 1, my understanding is that yes localised nucleation points will develop in real systems since theta phase will have the lowest free energy $G$ or $A$ ... they will then grow throughout the other phase until they are dominant... if left to equilibrate then the phase will become apparent at the macroscopic level. :) – AngusTheMan Oct 15 '14 at 19:13
• Also solids tend to melt from their surfaces inwards so even well below their melting temperature they will have a solid liquid phase ... As you heat it up the "amount" of this liquid phase increases and grows without limit... melting is actually a lot more complex than the liquid-vapour transition, especially if more than one species e.g not an elemental solid. Different types of interatomic bonds add extra degrees to consider. However this is all very qualitative talk .... look at "Lindemann hypothesis" for more info... – AngusTheMan Oct 15 '14 at 19:23