Cadotte's FilmTec membrane grew its salt-blocking skin at the meeting point of two liquids

Reverse osmosis became practical with a single-material membrane: the integrally skinned asymmetric cellulose acetate membrane made by Loeb and Sourirajan at UCLA in 1959 and patented in 1960. One polymer had to do both jobs — reject salt and pass water — so gains in rejection came at the cost of flux, and gains in flux at the cost of rejection. John Cadotte, working at North Star Research & Development Institute, had been making ultrathin polymer film laminates, patented in 1970, and filed his own reverse osmosis membrane patents in 1975 and 1977.

In 1972 Cadotte developed the first interfacially polymerised polyamide thin-film composite membrane. The recipe appears in the application filed on 22 February 1979 and issued on 7 July 1981 as US Patent 4,277,344, assigned to FilmTec Corporation: coat a porous support with an aqueous solution of a polyamine such as phenylenediamine, then bring the coated support into contact with a polyacyl halide such as trimesoyl chloride. Polymerization starts at the interface, in situ on the support, and the dried product is a composite — a crosslinked aromatic polyamide laminated to that support.

The reactants are held in immiscible liquids, so they meet only in a plane. The film forms there and then stops growing, because it separates the aqueous diamine from the organic acid chloride; the selective layer stays ultrathin while the support underneath takes the pressure. Using an acyl halide of at least trifunctional character crosslinks the polyamide; the patent shows this gives lower elongation and lower solubility than linear polymers — a dense skin that survives operation and cleaning. The patent also records good chlorine resistance, improvable with a chlorinating agent.

That architecture won. Today the great majority of reverse osmosis and nanofiltration membranes are polyamide thin-film composites, and FilmTec, the company that grew out of North Star, is now part of DuPont. The 1981 patent sits at the end of a documented line — North Star's 1970 ultrathin film laminate, Cadotte's 1975 and 1977 reverse osmosis membranes — and its claims describe the selective layer as a material made where two liquids touch rather than one poured, cast or extruded as a sheet.

Grow the selective skin where two liquids meet

A single-material membrane has to trade water flux against salt rejection; a composite hands each job to a different material.

Interfacial polymerization self-limits: the film it forms separates the reactants, so the selective layer stays ultrathin without growing thick and slow.

A tri- or higher-functionality acyl halide crosslinks the polyamide, which the patent shows gives lower elongation and lower solubility than linear analogues.

The porous support carries the mechanical load, so the chemistry can be delicate and expensive while the substrate stays cheap and strong.

Chlorine resistance decides whether a membrane survives real cleaning regimes; the patent covers improving it with a chlorinating agent.

Don't ask one material to do two jobs. Put the selective chemistry in a layer too thin to resist flow, let a thicker one take the pressure — and build the thin layer in place.

Polyamide thin-film composites made by interfacial polymerization became the reference architecture for reverse osmosis and nanofiltration, and FilmTec Corporation is now part of DuPont. The same reaction — an aqueous diamine against an organic acid chloride, polymerizing only at the interface — is still the standard way to form the selective layer, and has since been extended to nanofiltration, to solvent-resistant membranes and to layers built around nanoparticles. Reviews of desalination membrane materials trace the commercial line back to this composite.