Melting.: Loss of crystallinity of a solid reactant through melting, or eutectic formation, or dissolution in a product often results in an enhanced rate of decomposition as a consequence of the relaxation of the bonding forces responsible for lattice stabilization. The appearance of a liquid phase during the decomposition of solid reactants (e.g. metal nitrates or oxychlorides and organic substances) is often accompanied by an increased rate of product evolution. The presentation and discussion of observations for thermal decomposition reactions of solids should normally include due consideration of the possible occurrence of melting and the role of any liquid phase formed.
There is no generally acceptable comprehensive theory of melting. A feature of the fusion process, which is usually regarded as important in theoretical treatments of the subject, is the inability of a solid to superheat, and only a very small number of exceptions to this generalization are known [2]. This almost universal onset of liquefaction immediately upon reaching the melting point is in sharp contrast with the reverse process since supercooling of the vast majority of liquids can be demonstrated under appropriate conditions.
In a review of the subject, Ubbelohde [3] points out that there is only a relatively small amount of data available concerning the properties of solids and also of the (product) liquids in the immediate vicinity of the melting point. In an early theory of melting, Lindemann [4] considered that when the amplitude of the vibrational displacements of the atoms of a particular solid increased with temperature to the point of attainment of a particular fraction (possibly 10%) of the lattice spacing, their mutual influences resulted in a loss of stability. The Lennard-Jones—Devonshire [5] theory considers the energy requirement for interchange of lattice constituents between occupation of site and interstitial positions. Subsequent developments of both these models, and, indeed, the numerous contributions in the field, are discussed in Ubbelohde’s book [3].
Christian [6] treats melting as a nucleation and growth process, and discusses the possibility that the surface may be its own nucleating agent and lattice defects or impurities retained in such regions may similarly facilitate the formation of the melt. The melting process is, therefore, always effectively a two-phase phenomenon and any theoretical explanation must be based on consideration of the interactions between phases which differ in the degree of ordering.
Kuhlmann-Wilsdorf [7] provided a new theoretical approach in which melting was ascribed to the unrestricted proliferation of dislocations at the temperature for which the free energy of formation of glide dislocation cores becomes negative. Several physicists have shown interest in this model which has not so far been accorded similar attention in the chemical literature.
The theory of melting continues to be the subject of recent publications, including consideration of vacancy concentrations near the melting point [8,9], lattice vibrations and expansions [8, 10–12]. Meanwhile, the phenomenon also continues to be the subject of experimental investigations; Coker et al. [13], from studies of the fusion of tetra-n-amyl ammonium thiocyanate, identify the greatest structural change as that which occurred in the phase transition which preceeded melting. Solid and molten states of the salt are believed to possess similar structures, the significant difference being identified as the ability of the hydrocarbon chain to kink and unkink after fusion. The electrical conductivity of the solid increased as the melting point was approached. Allnatt and Sime [14] similarly found an anomalously large increase in the electrical conductivity of sodium chloride in the vicinity (i.e. within ∼4 K) of the melting point. Clark et al. [15] considered premelting transitions in the context of a general theory of disorder in plastic crystals.
As with other properties of solids, the increased relative significance of surface energy in very small (i.e. micrometre-sized) crystals influenced the melting points [2,16,17] and diffusion at this temperature. Quantitative studies of rates of melting of solids are impracticable since superheating is effectively forbidden and the rate of the endothermic phase change is determined by the rate of heat supply and the thermal conductivity of the solid.
Recrystallization.: The recrystallization of a solid may result in the production of a higher temperature lattice modification, which permits increased freedom of motion of one or more lattice constituents, e.g. a non-spherical component may thereby be allowed to rotate. Such reorganizations are properly regarded as premelting phenomena and have been discussed by Ubbelohde [3]. The mechanisms of phase transitions have been reviewed by Nagel and O’Keeffe [21] (see also Hannay [22]).
Recrystallization may also result in the elimination of regions of local lattice distortion, e.g. dislocations, grain boundaries etc., without change in chemical composition or, indeed, lattice configuration: such behaviour has been described by Christian [6]. Where two phases are present, certain particles of the discontinuous phase may grow at the expense of other (smaller or less perfect) crystallites so reducing the total interfacial energy [23,24]. Reactions involving the precipitation of a new phase from supersaturated solution are of great commercial importance, e.g. carbide formation during the manufacture of ferrous alloys [25], and the crystallization of glasses [1246].