Chemistry Reference
In-Depth Information
mixture of the same ingredients shows the same trend. The value of K ge also
decreases with the fat volume fraction for solid droplets [Figure 1(b)]; while this
effect is smaller than that seen for liquid droplets, it establishes the existence of
some positive interaction between solid fat droplets and aroma molecules. The
magnitude of the interaction with solids appears greater for smaller droplets,
suggesting it occurs at the surface rather in the bulk of the crystalline droplet.
The variation in K ge was modelled as a function of liquid oil volume fraction
using Equation (2) with independently measured bulk partition coefficients
(K go ¼ 6 10 6 , K gw ¼ 0.0182). Figure 1(a) shows that the equation fits the
data well over the whole concentration range, suggesting that the most impor-
tant factor is the volume partitioning of the aroma molecules, and that other
effects (e.g., Kelvin pressure effects, surface binding of the volatile, etc.) are not
important enough to be seen in the liquid oil emulsion data.
The particle-size effect seen in the solid fat data cannot be adequately
represented solely by a volume-based model. This suggests that interfacial
binding may be important. The presence of a binding surface can be incorpo-
rated into Equation (2) as an additional two-dimensional phase, 8,9
þ K iw A s
K gw
K ge ¼ f o
1
K go þ ð 1 f o Þ
ð 3 Þ
;
K gw
where A s is the interfacial area per unit volume of emulsion (expressed as 6 f o /
d 32 ), and K iw the surface binding coefficient defined as the ratio of the surface
excess concentration (volume of aroma compound per unit interfacial area) to
the aqueous concentration. As a first approximation, the EH was assumed to
interact just at the surface of the solid droplets and not with the bulk of the
crystal, i.e., we set f o ¼ 0. Equation (3) was then used to model the head-space
concentration of EH in equilibrium with the solid droplets. There is a reason-
ably good fit to the higher volume fraction samples, with the broken lines in
Figure 1(b) corresponding to the best fit value of K iw ¼ 4.5 10 6 m from a
statistical fit to the data with f o > 0.1. But the model overestimates the head-
space concentration at lower oil volume fractions.
As the effect of surface binding is much smaller than the effect of partitioning
into liquid oil, we are not able to discount the possibility of similar binding
occurring at the liquid oil-water interface. When Equation (3) is applied to the
liquid droplet data, a similar value of K iw (or even a value several orders of
magnitude greater) causes no appreciable change to the predicted curve. For
the remainder of the work we assume that surface binding occurs at all lipid-
water interfaces with this same binding coefficient regardless of whether the
droplets are solid or liquid. However, we only expect to see the importance of
surface binding in emulsions where there is no liquid oil present.
29.4 Nature of the Flavour-Binding Interactions
In an effort to better understand the nature of the interactions, a coarse
n-eicosane emulsion was prepared with a median droplet size of
10 mmto
B
Search WWH ::




Custom Search