Water in Dairy Products - Advanced Dairy Chemistry - page 470

misleading in the case of systems, such as aqueous solutions, where one
component is a network former (Angell, 2002). Although the controversy
continues, several arguments, including the annealing behaviour of hyper-
quenched water, are presented in support of a water T g 165 K (Velikov et al.,
2001; Angell, 2002). From the study of dielectric properties of water ''con-
fined'' in vermiculite clay or bread, a state allowing the study of supercooled
water in a temperature range where crystallization normally precludes obser-
vation, it was suggested that what is observed by DSC at 136 K is a local
-like process (Cerveny et al., 2004). In mixtures of water and solutes break-
ing up the H-bond network of supercooled water, the T g should be deter-
mined by this -like relaxation process. On the contrary, extrapolation of the
T g values obtained by DSC for mixtures of water with n-propylene glycols,
which expand the network instead of breaking it, gives a T g 162 K (Cerveny
et al., 2004; Jansson et al., 2005a).
The heat capacity increment at T g for water may be even more influen-
tial, when the Couchman and Karasz expression is used to predict T g for a
mixture, or to determine the (DCp) Tg of the dry solute (Borde et al., 2002),
because very different values were reported. For pure water (conventional
T g ¼ 136 K), reported values range between 0.09 (hyperquenched liquid)
(Hallbrucker et al., 1989) and 1.94 J/g//K (vapour-deposited liquid) (Sugisaki
et al., 1968). Whereas the latter value may be falsely high due to the presence
of a sharp relaxation overshoot (Angell and Tucker, 1980), the first one is
only 10% of the value extrapolated to pure water with various salt-water
mixtures, which are in the range 1.06-1.39 J/g/K (Angell and Tucker, 1980).
Extrapolation to pure water of the values obtained with galactose solutions
gave 0.93 J/g/K (Blond and Simatos, 1991).
The question of glass transition in proteins deserves some clarification.
The observation of glass transition on DSC thermograms is most often
difficult with proteins. The Cp jump, which is usually considered as the
signature of glass transition, is more clearly observed, in a given range of
water content, with fibrous proteins (gelatine: Marshall and Petrie, 1980;
gluten: Kalichevsky et al., 1992; Noel et al., 1995; Morales and Kokini,
1997) or polypeptides (poly- L -asparagine: Green et al., 1994) or after dena-
turation with globular proteins (legumin: Sochava and Smirnova, 1993;
bovine serum albumin: Farahnaky et al., 2005). Due to the low value of the
Cp jump, compression of the powder submitted to DSC is most useful (casein:
Kalichevsky et al., 1993). Identification of this feature with glass transition is
confirmed by thermomechanical spectroscopy, showing the characteristic
drop in E 0 and maximum in E 00 in a similar temperature range (Kalichevsky
et al., 1992-1993, De Graaf et al., 1993). The variation of the T g or T thus
determined as a function of water content exhibits the expected plasticization
behaviour, with values ranging from 80-1008C to 0-408C for a water content