FREEZING GELS

The process of freezing of liquid in the pores of a gel is examined in detail. Gels typically begin to freeze at a small undercooling, which indicates that nucleation occurs by a heterogeneous, rather than a homogeneous, mechanism. If a crystal appears in excess liquid at the surface of a porous body, it will penetrate smaller and smaller pores as the temperature is decreased. However, if the pores are small, before the undercooling is great enough to permit penetration, there is a strong driving force for liquid to flow out of the pores to the crystal. It is shown that this creates a gradient in the pressure in the pore liquid, and produces stresses strictly analogous to those generated during evaporative drying. Therefore, cracking can occur during freezing; indeed, freezing stresses can make freeze-drying even more damaging than conventional drying. Considerable shrinkage can occur as liquid drains from the pores to the growing crystal, and this has important consequences for the measurement of pore size distributions by thermoporometry. The pore sizes may be underestimated by approximately 30% for a typical silica gel, if water is the liquid in the pores. Since the stresses are related to the entropy of fusion and crystal/liquid interfacial energy, organic liquids offer advantages over water for both freeze-drying and thermoporometry. The stress exerted on the network by a crystal is analyzed and a criterion for fracture of the network is presented. An analysis is presented of the kinetics of heterogeneous nucleation when the network itself acts as a nucleation catalyst; in this case, the crystals are expected to appear at nodes in the network. However nucleation by the network is not expected in most material systems.