1. Introduction
In the ice cream manufacturing process, the freezing
step is one ofthe most important steps which determines
the final texture of this frozen food. The final quality of
the product and particularly the smooth texture or the
cooling sensation perceived by the consumers when
eating ice creams is largely conditioned, beyond its
formulation, by its structure, namely by the distribution,
the size and the morphology ofthe air bubbles and the
ice crystals. Nevertheless, during hardening, storage and
commercial distribution, some important physical
changes occur in the ice cream structure. Concerning
the air phase, two main mechanisms ofthe air phase
redistribution have been observed: coalescence and
drainage (Chang & Hartel, 2002a). Concerning the ice
phase, the ice crystal texture is strongly affected by the
recrystallization phenomena, which depend principally
on formulation factors, on freezing process and on
storage temperature conditions (Donhowe & Hartel,
1996). A good quality control ofthe ice texture and
particularly ofthe ice crystals growth relies on a large
number ofsmall crystals. The literature data show that
the air bubbles diameter range is located between 30 and
150 mm with a mean diameter around 40 mm (Chang &
Hartel, 2002b), and, the ice crystals mean size is between
20 and 75 mm with a mean value around 40 mm (Berger,
Bullimore, White, & Wright, 1972; Hagiwara & Hartel,
1996; Hartel, 1996; Russel, Cheney, & Wantling, 1999).
Concerning the lipid components, the fat is present in
the mix as a fine emulsion which partially coalesces
during the freezing. The published data indicate a range
ofthese globule diameter to be 0.04–4.0 mm (Berger
et al., 1972).
To approach the quality criteria optimization, various
methods types have been used to characterize the texture
and the structure of frozen foods. These methods
include, for example, sensorial analysis, textural analysis,
differential scanning calorimetry, nuclear magnetic
resonance spectroscopy or imagery and various microscopic
methods. Among these numerous methods, it
seems that the microscopic methods are the most
ARTICLE IN PRESS
*Corresponding author. Tel.: +33-4-72-43-1843; fax:+33-4-72-43-
1682.
E-mail address:
[email protected] (J. Andrieu).
0023-6438/03/$30.00 r 2003 Swiss Society ofFood Science and Technology. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/S0023-6438(03)00094-X
suitable and adequate and, consequently, are the ones
most used and implemented in industrial control
laboratories.
Three types ofmicro scopic techniques are currently
used: destructive methods, indirect methods and direct
methods. The first methods type is based on the optical
observation ofa sample mixed with a suitable medium
for dispersing the observed phase (ice crystals or air
bubbles) and dissolving the other phases (Donhowe,
Hartel, & Bradley, 1990; Sztehlo, 1994). Its great interest
relies in the following facts: (1) the experimental time is
short and (2) the information on the particles size
distribution is easy to collect. Nevertheless, due to the
sample dissolution, no data about the relative position
and about the volume fractions of the different phases
can be obtained. Besides, as the sample is totally
destroyed or melted, the images obtained cannot
represent all the frozen system texture morphology.
The second type are ofindirect methods that use mainly
electronic microscopy techniques. They provide clear
images ofthe structure and they destroy only partially
the sample by cryo-substitution, cryo-fixation, freezeetching
or freeze-drying (Berger & White, 1971; Goff,
Verespej, & Smith, 1999; Woinet, Andrieu, Laurent, &
Min, 1998). Their main advantage relies on high
magnification that allows one to observe fine structural
details, but the electron microscopy is expensive and the
sample preparation time is quite long.
The last method used to characterize the structure of
frozen foods is more original and quite unknown in
industrial laboratories. It is the photonic microscopy
method with episcopic coaxial lighting that preserves at
best the sample original texture ofthe frozen sample. It
is based on direct light reflection on the surface of an ice
cream sample. A direct observation in situ at cold
temperature allows maintaining the sample in its native
state and minimizes the potential artefacts of all the
other methods. Compared with the electron microscopy,
the main drawback ofour method is its low magnification,
which does not allow, for instance, the observation
ofthe fat particles. So, to investigate the influence ofthe
freezing conditions on the ice cream structure, the direct
optical microscopic method by reflexion was developed
and set up in our laboratory (Faydi, Andrieu, &
Laurent, 2001) with model ice creams without overrun.
In the same project, some experimental data obtained
have been checked and validated by indirect and
destructive methods. The present paper concerns the
extension ofthis method to commercial ice creams with
standard formulations.