be mentioned that macroscopic surfaces (vessel walls, etc.) can catalyze
nucleation by the same mechanism as shown in Figure 14.5b.
Epitaxy. Equation (14.12) shows thatyhas to be small forjcatto
become really small. Fory¼90 degrees,fcatwill equal 0.5, which will not
greatly affect nucleation temperature, but fory¼15 degrees,fcat&0.0009,
and this would make an enormous difference, e.g., reducing the
undercooling needed for fast nucleation from 30 to 1 K. (Try to calculate
this for one of the systems given in Table 14.2.) It is, however, very unlikely
that such a thin spherical segment of crystalline material would form, since
it would only be one or two molecules thick. However, something like it can
occur if the crystal lattice of the material to crystallize is very similar to that
of one of the crystal faces of the impurity. In such a case, a monolayer of the
new phase can be deposited onto the impurity, and grow out. This is called
epitaxy. A well-known example is the nucleation of ice on silver iodide
crystals in cold air supersaturated with water vapor.
In foods something similar occurs during the crystallization of fats.
Natural fats have a wide compositional range, implying that several
different kinds of crystals have to be formed. This needs undercooling, but
as soon as the first crystals have formed, other crystals nucleate on the
existing ones. This must be due to epitaxy, since the various crystals are very
similar in lattice structure. It is indeed observed that the undercooling
needed is only by about 2 K.
A very small contact angle tends to be formed also if the solid is a
crystal, phaseais air, andbis the liquid phase of the crystal material. This
implies that the crystal acts as an effective catalytic impurity for the
formation of its own melt. This explains why very little overheating
generally suffices to induce the phase transition crystal!melt, as was
mentioned in Section 14.1.
Memory. Figure 14.5b, frame 5, shows that in a sharp crevice in a
particle or the vessel wall, a concave interface between crystalline material
(b) and solution or melt (a) can exist, as seen from phasea. This can occur if
the contact angleyis small. It implies that phasebwill spontaneously grow.
Moreover, if the system is heated to a temperature above Teq, some
crystalline material may remain in the crevice. If now the system is cooled
again, crystallization may occur without significant undercooling being
needed, since the crystalline phase is already present. This is called a
memory effect. The memory of the system can be ‘‘destroyed’’ by increasing
the temperature to several degrees aboveTeq. Often about 5K will suffice,
but there are examples of a more persistent memory.