A possible explanation could be that the shell is swelling and that the liquid fraction in the shell increases. To reach the maximum observed BrTAG levels of 5–7% the shells need to absorb about 80–140% (calculated on the fat content) additional liquid fat of the filling fat composition which would lead to a swelling of between 35% and 60% of the shell (note that the peek values only have been obtained in limited domains of the shell).
Unfortunately, we did not characterise the relative volumes of shell and filling to confirm this hypothesis, still, swelling of the shells was observed.
Further, swelling has been observed previously of CB matrixes put in contact with liquid fat.