Quantitative insights into the mechanisms of nucleation during crystallization

Nucleation during first order phase transitions, such as a lake freezing in winter, formation of snowflakes, and bio-mineralization, occur ubiquitously in nature. Traditionally nucleation is understood in terms of the classical nucleation theory (CNT) in which the rate of nucleation is given by J = Bexp(−G/kT), where B is a pre-exponential factor and G is the barrier to form a nucleus of critical size which depends on the interfacial free energy γ and the chemical potential difference between the solid and liquid. Many factors that control J are determined by a material's electronic structure. However, experimental evidence suggests that depending on growth conditions (i.e. extrinsic factors such as temperature, pressure, chemistry of solution, and concentration of solute) nucleation can involve kinetically favored pathways that deviate from homogeneous nucleation in CNT. For instance, crystallization can involve the formation of crystallites (i) with lower γ and lattice symmetries that are different from the final crystallite (e.g. two-step nucleation), or (ii) by the coalescence of smaller crystallites by oriented attachment (e.g. crystal growth by particle attachment). Thus, in practice the predicted nucleation rates are usually off by orders of magnitude. Our ability to predict crystallization therefore hinges on how various kinetic pathways are affected by growth factors. To this end, I will present new insights on two aspects of nucleation under highly driven conditions; (a) polymorph selection and (b) the competition between diffusion and dislocation activity during crystal growth by particle attachment. I will illustrate how these evidences challenge the traditional interpretations of how crystals nucleate and growth.