The first attempt to use more sophisticated techniques in order to take advantage of the extremely large overall driving force which appears in purely osmotic applications was made by Popper, Camirand, Nury, and Stanley (1966).
They used first generation reverse osmosis membranes, made of cellulose acetate and similar materials, in both a flat and tubular configuration.
Two different membrane housings were used: a tubular configuration to accommodate the tubular membranes and a plate and frame configuration for flat membranes.
The latter came from a suitable modification of a commercial filter press allowing circulation of the fruit juice and osmotic medium across the flat sheet membrane.
An average osmotic flux of about 2.5 kg/(m2 h) was reported by this team and they claimed the production of highly concentrated fruit juices.
Beaudry and Lampi (1990a,b) revisited the same process, which was named by Popper et al. (1966) dialysis .
They changed the name of the process from dialysis into direct osmosis and improved some of the engineering aspects.
They also utilized improved, new generation thin film composite (TFC) reverse osmosis membranes, which were properly modified.
Through additional developmental work, these membranes became thinner (overall membrane thickness 25-85 µm), while the membrane s top selective layer remained similar to that of tight reverse osmosis membranes (MWCO less than 100 Da), thus allowing no passage of other food ingredients besides water.
Furthermore, a tubular membrane module of specific design was constructed to provide housing for the sensitive ultrathin direct osmosis membranes.