The fractional coverage of images from all crystals vs time is shown in Fig. 5a for filled (e.g. Fig. 2c)
and in Fig. 5b for clean, unfilled (e.g. Fig. 2d) crystalmaps.
The fractional coverage of unfilled crystalmaps can be argued to approximately represent the solid fat content of the samples.
They produce the “s-shaped” curve typically found on SFC vs time plots with the final (equilibrium) coverage increasing with decreasing temperature.
These are similar in form to published SFC vs time plots obtained from DSC data [4], and the final values for the unfilled crystalmaps at 25 °C (the only directly comparable temperature) are also similar. However, the interaction of light with the crystals is not well known and so the precise edge (and therefore thresholding point that should be used when creating the clean crystalmaps) of the crystallites is not known exactly.
This is more of an issue with the speckly crystals found with palm oil (compared with tripalmitin) as there is a much larger interfacial area between the liquid and solid phases.
Consequently the thresholding value used to create Fig. 2d has a significant effect on the SFC values obtained. In this case the thresholding value used was chosen by eye to match the original images most closely and the same value was used for all runs.