How Does Desalination Work? Pushing Water Through a Membrane

Every method is either a phase change or a filtration, and the industry has moved decisively from the first to the second:

• •Thermal distillation, which evaporates water and condenses the vapour, leaving salt behind. Multi-stage flash distillation does this at successively lower pressures so that water boils repeatedly without additional heat, and multiple-effect distillation reuses the heat of condensation from one stage to drive the next
• •Reverse osmosis, which forces seawater at high pressure against a membrane that passes water molecules and rejects dissolved salts, and which now accounts for the large majority of new capacity
• •Electrodialysis, which uses an electric field to pull ions through selective membranes, efficient for brackish water with lower salt content and not for seawater
• •Freezing, which exploits the fact that ice crystals exclude salt, and which works and has never been made economic at scale
• •Solar stills, which are simple, use no purchased energy and produce quantities too small for anything but emergency and household use

Osmosis is the natural movement of water through a semipermeable membrane from a dilute solution toward a concentrated one, which is the process that moves water into plant roots and into cells. Reversing it means applying pressure to the salty side greater than the osmotic pressure pulling the other way, which for seawater is around twenty-seven atmospheres, and plants operate at fifty-five to eighty to achieve useful flow. The membranes are thin-film composites, with an extremely thin polyamide layer perhaps a tenth of a micrometre thick doing the separation, supported by porous backing, and wound in spirals into cylindrical elements that pack an enormous area into a pressure vessel. Roughly forty to fifty percent of the incoming seawater emerges as fresh product and the rest leaves as concentrated brine. The technology's economics were transformed by energy recovery devices, which transfer pressure from the outgoing brine stream directly to incoming feedwater, cutting the energy requirement by more than half and taking modern plants to around three kilowatt hours per cubic metre.

Desalination makes sense where the alternatives are worse, which is a narrower condition than its promoters suggest. The Gulf states built the first large capacity, having no rivers and cheap fuel, and the region still holds a substantial share of world capacity, historically thermal because the plants could be coupled to power stations and use waste heat. Israel now supplies a large majority of its domestic water from a handful of reverse osmosis plants and has become a technology exporter. Singapore uses it as one of four water sources alongside imports, catchment and recycled wastewater. Spain, Australia and California have built plants as drought insurance, several of which sit idle in wet years, which is expensive and is the point of insurance. Around twenty thousand plants operate worldwide producing something over a hundred million cubic metres a day, and the cost has fallen by roughly an order of magnitude since the 1970s while remaining higher than treating a river.

The waste stream is the most underdiscussed part. For every unit of fresh water produced, a plant discharges roughly one and a half units of brine at close to twice the salinity of seawater, warmer than the sea it enters and containing antiscalant chemicals, coagulants and traces of the metals the plant is made from. A 2019 study estimated global brine production at around fifty billion cubic metres a year, substantially higher than previous figures. Discharged carelessly into a shallow enclosed bay, it sinks, forms a dense layer and damages seagrass and benthic communities; discharged through a properly designed diffuser into an area with good mixing, the effect is measurable and local. Intake causes its own damage, drawing in fish, larvae and plankton, which subsurface intakes drawing through the seabed largely avoid. Neither problem is unsolvable and both cost money, which is the recurring theme of the whole technology.

The thermodynamic minimum for separating fresh water from seawater is around one kilowatt hour per cubic metre, and real plants achieve three to four, which means the room for improvement is real and bounded. That energy has to come from somewhere, and a plant running on gas converts a water problem into an emissions problem, which is why the proposals that make sense pair desalination with renewable generation and treat the plant as a flexible load that runs when electricity is abundant. The honest comparison is with the alternatives, and on cost per cubic metre desalination remains far above fixing leaks in distribution networks, which lose a quarter or more of supply in many cities, above water reuse, which treats wastewater to potable standard at lower energy cost than desalination, and above demand reduction through pricing and efficiency. Its case is strongest as a drought-proof supply of last resort in coastal cities with no other option, which is exactly where it has actually been built.

Desalination either boils water and condenses the vapour or forces it through a membrane under pressure, and reverse osmosis has displaced thermal methods almost entirely, operating at fifty-five to eighty atmospheres and recovering pressure from the outgoing brine to reach around three kilowatt hours per cubic metre against a thermodynamic floor of about one. Roughly half the intake becomes fresh water and the rest leaves as brine at twice seawater salinity, which damages enclosed waters if poorly discharged. It costs more than fixing leaks or reusing wastewater.