relations predominantly depend on food composition, and the basic factors
governing them are given in Section 8.1. Figure 8.3c shows, for example,
that the presence of small molecules or ions has considerable effect on the
isotherm, in agreement with Eq. (8.2); curd contains several solutes
(altogether about 4 mmol per g protein), and cheese contains added salt as
well (about 5 mmol per g protein).
For true adsorption, several equations for adsorption isotherms have
been derived, based on various theories. Such equations often are applied to
water sorption isotherms of foods as well. However, one cannot speak of
adsorption in the case of most foods, as mentioned above, because there is
no (or a very limited) phase surface onto which water can adsorb. Moreover,
most foods contain numerous components; even if phase surfaces were
present they must be very inhomogeneous. In the author’s opinion, it
therefore makes little sense to use such equations. Only for relatively simple
and homogeneous systems, like pure starch granules, can some theories be
more or less applicable, but not for real foods. Mathematical fitting of
experimental data may be useful for practical purposes, and since the
equations generally have three or four adjustable parameters, a reasonable
fit can often be obtained. But one cannot attribute physical significance to
the parameters derived in this way, such as a ‘‘monolayer water content.’’
Another approach is to proceed from the chemical composition. This
may work for some fairly dry foods, although it implicitly assumes an
absorption mechanism, which is that certain chemical groups ‘‘bind’’ certain
amounts of water, and by determining the concentration of these groups, the
water sorption can be calculated. It concerns especially ionized groups (a
few water molecules per group) and dipoles, such as a peptide bond (< 1
water molecule per group). This method works reasonably well for proteins
aroundaw¼0.5.
Hysteresis. A sorption isotherm is in principle determined by
placing a small sample of the food of known water content in air of a given
humidity and temperature and then determining the weight of the sample
after various times. After the weight does not change any more, which often
takes several days, the equilibrium water content is considered to be
reached. By doing this experiment at a range of air humiditiesðawÞ,an
isotherm is obtained. One can do this with samples that are successively
brought to a higher or to a lower water content, and the curves so obtained
are usually not identical. There is hysteresis between the ‘‘desorption’’ and
the ‘‘adsorption’’ isotherm, as illustrated in Figure 8.4a. This means that
thermodynamic equilibrium is not obtained, at least at low water content. The
consequence then is that the water activity is undefined, since awis by
definition an equilibrium property. This would mean that the scale of the x-