Melting and freezing are simple concepts that most people will have been given a grasp of in primary school. Melting is when particles is a solid gain enough energy to begin moving around each other (but not enough to break free entirely) and freezing is when the particles of a liquid lose energy and so fix in their place. Of course nothing in life can ever be that simple and so the exact atomic mechanism has been studied physicists and chemists for centuries. Despite this it is still very challenging to produce experimental observations of the atomic actions when going through these phase transitions.
The long standing theory, being made a long time ago, is called Classical Nucleation Theory (CNT). This says that one atom loses enough energy to freeze and then other particles will freeze onto it as the solid lattice forms. However there are many assumption made in this theory and so a more complex but superior explanation is Ostwald’s rule. This is the idea that in general there is a tendency for the least stable solid structure to form first and the more stable ones later. This has also been shown to be true in reverse when crystals melt. Theoretical models are also ineffective for modelling melting and freezing with some predictions of crystallisation rate being a hundred times off the experimental value.
This shows that more experimental data is required to back up any theory that can be created. First a set of bismuth (Bi) nanoparticles are grown on a base compound and observed by a high resolution electron microscope. The melting and freezing mechanisms can be brought into effect and images of the event can be taken. From the data and the analysis of it it was found that when the nanoparticles get near the phase change they enter a series of intermediate states which aren’t liquid or solid. The progression through the intermediates are essentially identical for melting and freezing and in fact on a small scale both can happen simultaneously with the net change of the local group representing the bulk phase change. Even though there are very clear differences between bismuth particles being excited by an electron beam and ice being warmed the mechanisms found still represent the microscopic activities that are likely to occur in all phase changes.