If I put your encapsulated caffeine in a few finished product prototypes, how best can I measure or determine how much caffeine survived?
Regarding a dissolution curve for caffeine, if the coating is purely lipid and coated caffeine is particulate, simply filtering the product and assaying the filtrate for protein (by the kjeldahl procedure) would be a simple indirect method for checking the amount of caffeine no longer encapsulated. If the product is in a product containing protein, that system would not work.
Because caffeine is a purine, it has a UV absorption spectra that also can be used for quantification. It absorbs at 276 nm.
One could simply scan a filtered water extract and, providing other purines are not present in large amounts or do not absorb at this wavelength, this should be the simplest, most direct assay procedure for determining a dissolution curve.