The accepted mechanism for the formation of a deuterium-tritium (D-T) ice layer is that mass evaporates (sublimes) from the warmer regions of the shell and deposits in the cooler regions. Recent observations of the early-stage formation of single-crystal ice layers in OMEGA targets show that the rate and direction of crystal growth are influenced by liquid wicking to the crystal growth surface. This behavior is attributed to the ice-liquid interface possessing a lower surface energy than the ice-vapor interface, and the amount of liquid transported by this process is determined by the size, position, and growth rate of the initial seed crystal. Appreciating this behavior allowed us to define an improved cooling ramp that balances the rate at which heat was removed from the target with the supply of liquid to the crystal growth surface. The time and temperature parameters used to form a seed crystal and then grow the crystal into a complete ice layer are presented. One benefit of this process may be fewer defects in the ice layer. The target was cooled to 0.6 K below the temperature where it was formed before strain-induced crystallographic features developed. An estimate of the extent of fractionation of D2, D-T, and T2 isotopes during the freezing cycle was based on the thickness uniformity of the ice layer and how the crystal grew. The region where the ice layer initially formed was 4% thinner than the region where its formation was complete. The alignment of this perturbation to the ice layer with the growth axis of the crystal suggests, to a first-order approximation, that the area of the crystal that first formed possessed a higher fraction (~4%) of tritium atoms.