3.1. Comparison of different observation methods
First, in order to validate our direct observation
method, some ice crystal sizes data were compared to
data resulting from two others methods, namely a
destructive method and an indirect method.
The destructive method was based on mixing the ice
cream sample in a suitable medium, dispersed at 10C,
and observed with an optical microscope by transmission
between two thin layers. Then, the ice crystals were
automatically analysed with a software based on the
principle ofsep aration ofadjacent crystals.
The indirect method adopted was also previously used
in our laboratory (Woinet et al., 1998; Faydi et al.,
2001). Frozen samples were first freeze-dried at temperatures
lower than the glass transition temperature to
avoid possible structure modifications during the sublimation
step. Then, the freeze-dried samples were
polished, sputter coated with a thin gold layer. The
samples were transferred under vacuum into the
scanning electron microscope, where they were viewed
(15 kV accelerating voltage) and photographed by using
different magnifications. In the case of samples with
overrun, we could differentiate clearly the air bubbles by
their spherical shape, and the pores generated after the
ice sublimation (Fig. 3).Fig. 4 compares the ice crystal mean size obtained by
the three microscopy methods—i.e. direct, destructive
and indirect methods—for ice cream samples frozen
under different process conditions. These three methods
converge globally to the same data.