Reverse Osmosis: How the Membrane Works and Where RO Actually Fits

Reverse osmosis (RO), also called inverse osmosis, is a membrane separation operation driven by a pressure difference that separates solvent from a solution. Apply pressure exceeding the osmotic pressure to the feed on one side of the membrane and the solvent flows against the natural direction: that is reverse osmosis. The low-pressure side yields the permeate, which is the solvent that passed through; the high-pressure side yields the concentrate. Treating seawater this way gives fresh water on the low-pressure side and brine on the high-pressure side.

What osmosis does when nobody pushes

Place equal volumes of a dilute solution, say fresh water, and a concentrated solution, say seawater or brine, on either side of a container separated by a semi-permeable membrane. Solvent from the dilute side passes through toward the concentrated side until the level difference creates a pressure equal to the osmotic pressure and the system reaches equilibrium.

Osmotic pressure depends on the concentrated solution's type, concentration and temperature, not on the membrane. Push harder than that and the flow reverses, which is the whole principle in one sentence.

The transport equations that size a plant

The solvent permeation rate, expressed as volume flux N, follows:

N = Kh (Δp − Δπ)

Here Kh is the hydraulic permeability coefficient, which increases slightly with temperature; Δp is the hydrostatic pressure difference across the membrane; and Δπ is the osmotic-pressure difference. For a dilute solution the osmotic pressure is π = iCRT, where i is the number of ions produced by dissociation of the solute molecule, C is molar concentration, R is the universal gas constant and T is absolute temperature.

Two practical consequences follow directly from that algebra. First, flux scales with net driving pressure, so raising feed pressure raises output but also raises energy consumption roughly in proportion. Second, as recovery increases the concentrate gets saltier, Δπ grows, and net driving pressure falls even though the pump has not changed. That is why the last elements in a pressure vessel always produce the least permeate.

Two theories, one dense film

RO membranes are not sieves. The active skin is treated as a dense, pore-free film, and two models are used to explain how anything crosses it.

The solution-diffusion model, proposed by Lonsdale and co-workers, assumes that solute and solvent dissolve in the homogeneous non-porous surface layer and diffuse through it, driven by chemical-potential differences arising from concentration or pressure. Differences in solubility and in diffusivity within the membrane phase govern transport. The sequence is adsorption and dissolution at the feed-side surface, independent molecular diffusion through the active layer along each species' chemical-potential gradient, then desorption at the permeate-side surface. Steps one and three are fast, so the rate is governed by diffusion. Selectivity comes from the solubility and diffusivity differences, which is why a membrane can separate gas or liquid mixtures without any pore being smaller than the molecules.

The preferred-adsorption and capillary-flow theory takes a different route. When a liquid dissolves different substances its surface tension changes differently: dissolved organics such as alcohols, acids, aldehydes and esters lower surface tension, while some inorganic salts slightly raise it. Because the solute distributes unevenly, more concentrated at the surface layer than in the bulk, surface adsorption occurs. When an aqueous solution contacts a polymer porous membrane that negatively adsorbs solute but preferentially adsorbs water, a film of pure water of a certain thickness forms at the membrane-solution interface. Under external pressure that water passes through the membrane's capillary pores, yielding pure water.

Where RO earns its capital cost

RO usually employs asymmetric or composite membranes, and the equipment is mainly hollow-fiber or spiral-wound separation units. The membranes reject inorganic ions, colloidal matter and macromolecular solutes, producing purified water, and they can also pre-concentrate macromolecular organic solutions.

Because the process is simple and energy-efficient it developed rapidly over the past two decades, and it is now used at scale for seawater and brackish-water desalination, boiler-water softening and wastewater treatment, and in combination with ion exchange for high-purity water. Its scope keeps expanding into concentration of dairy and fruit juices and into separation or concentration of biochemical and biological preparations.

In pretreatment and desalination duty, RO performs particularly well. It can cut the load on ion-exchange resin by over 90% and reduce regenerant consumption by more than 90%, which saves cost and reduces the chemical burden on the environment. It also removes particulates, organic matter and colloids, protecting the resin from fouling and extending its life.

The part that decides success

Almost every RO failure traced back through a plant ends in pretreatment. Oil, scale-forming ions, colloidal silica and biological growth all concentrate on the feed side, and the membrane cannot clean itself. That is why upstream units carry more weight than the membrane skid: remove the oil before it reaches the element, hold the saturation indices in check, and dose antiscalant with the same care applied to the membrane itself. Do that and an RO train runs for years between cleanings. Skip it and the same train fouls in months, with the membrane blamed for a problem it never caused.