Ice surfaces have a thin layer of water below its melting temperature of 0℃. Such premelting phenomenon is important for skating and snowflake growth. Similarly, liquid often crystallizes into a thin layer of crystal on a flat substrate before reaching its freezing temperature, i.e. prefreezing. The thickness of the surface layer usually increases and diverges as approaching the phase transition (such as melting and freezing) temperature. Besides premelting and prefreezing, whether similar surface phenomenon exists as a precursor of a phase transition has rarely been explored.
Han’s team at HKUST proposes that a polymorphic crystalline layer may form on a crystal surface before the crystal-crystal phase transition and names it pre-solid-solid transition. If carbon atoms near the surface of a diamond could rearrange into a graphite lattice before reaching the diamond-graphite transition temperature, then it would be a so-called pre-solid-solid transition. The mechanism is essentially the same as premelting or prefreezing: the newly formed surface layer lowers the crystal’s surface energy. Han’s team pointed out that it is possible when two polymorphic crystals can form a coherent interface, i.e. the two lattices with appropriate lattice spacing and orientations match perfectly at the interface. Thus, the surface of the denser polymorphic crystal can form a layer of less dense crystal because the newly formed crystal-crystal interface is coherent and costs almost no energy.
Han’s team further confirmed the pre-solid-solid transition in experiment and computer simulation. They found that the surface of a thin-film colloidal crystal with a triangular lattice can form a square lattice as their interface is coherent. The thickness of the surface layer grows with temperature in a power law similar to premelting. These are further confirmed by their simulation about atoms with different interactions.
Premelting has been firstly conjectured by Michael Faraday, the father of electricity, in 1842, but not experimentally confirmed unambiguously until 1980s. The second type of phenomenon, prefreezing, has been proposed and observed in 1950s-70s. The pre-solid-solid transition proposed and observed by Han’s team is the third type of surface wetting precursor of phase transition.
Although it is a phenomenon under thermal equilibrium, they find that surface crystalline layer can also exist in nonequilibrium processes after an abrupt temperature change, such as melting, freezing and polycrystal annealing. The surface layer facilitates the processes, thus may have utility in material fabrication and processing. The trivial solid-solid transition occurring on the free surface after crossing the solid-solid transition point is ruled out from the polycrystal annealing, phase rule or other evidence. In addition, they found the novel double surface layers of liquid and square lattice in the overlapped temperature regime of premelting and pre-solid-solid transition.
Colloids, such as milk, paint, blood, are usually liquid suspensions of small micron-sized particles with Brownian motions. The particles thermal-motion trajectories can be tracked under optical microscopy even inside the 3D bulk of crystals or liquids, which can hardly be obtained in atomic systems. Thus, colloids have been used as a powerful model system to measure the microscopic processes in phase transitions. “Our work shows that colloids are also effective in discovering new type of phenomenon even for the well-known phase behaviors under thermal equilibrium.” Han said, “Its mechanism is simple and thus could be proposed decades ago, but seems to be a blind spot in material science. One future direction is to search for this phenomenon in atomic or molecular crystals. Unlike premelting generally exists in most crystals, pre-solid-solid transition can only exist in polymorphic crystals with low-energy coherent interface. We suggested several candidates of atomic and molecular crystals with coherent interfaces.”
Polymorphic crystals often have different properties, e.g., graphite is soft, black and electric conductive, while diamond is hard, transparent and non-conductive. A crystal with a polymorphic crystalline layer would have tunable surface properties and such layer can regenerate after being worn or corroded off. Thus, it should be a very useful material in applications.
The work is funded by the Hong Kong Research Grants Council and the Guangdong Basic and Applied Research Foundation, and has been published recently in Nature Physics.