The Engineering of Desalination Plants

Chart 1: Energy consumption by desalination method. MSF = multi-stage flash, MED = multiple effect distillation, MVC = mechanical vapor compression, RO = reverse osmosis. Source: N43 and Hermes, compiled from IDA and IEA data.

01 The Scale of Thirst

Fresh water is 2.5% of the water on Earth. Of that 2.5%, most is locked in glaciers and ice caps; only about 0.5% of the planet's total water is accessible fresh water in lakes, rivers, and aquifers. The rest — 97.5% — is seawater, with an average salinity of about 35,000 parts per million dissolved salts. Human consumption, agriculture, and industry require fresh water, and the gap between demand and supply is widening. By 2025, an estimated 1.8 billion people will live in regions experiencing absolute water scarcity, defined as less than 500 cubic meters per person per year. Desalination — the engineered conversion of salt water to fresh water — is the only source that does not depend on the local hydrological cycle. It is, in principle, unlimited. The question is what it costs, in energy and money, to produce.

Global desalination capacity has grown from roughly 12 million cubic meters per day in 2010 to over 100 million cubic meters per day by 2025, serving roughly 300 million people across 177 countries. The growth is concentrated in the Middle East — Saudi Arabia, the UAE, Israel, and Kuwait — and in water-stressed coastal regions including Australia, Spain, California, and Chile. The largest single plant, the Taweelah desalination facility in Abu Dhabi, produces 909,000 cubic meters per day, enough to supply a city of three million. The expansion is driven by necessity, by improvements in membrane technology that have halved the energy cost of desalination over two decades, and by the co-location of desalination plants with power generation that allows waste heat to be used for thermal processes.

02 The Thermodynamic Minimum

Every desalination process must overcome a fundamental thermodynamic barrier: separating salt from water requires work, because mixing salt and water is spontaneous and entropy-increasing. The theoretical minimum energy to desalinate seawater at 35,000 ppm salinity and 50% recovery — producing one cubic meter of fresh water and one cubic meter of brine at 70,000 ppm — is approximately 1.06 kilowatt-hours. This is the floor below which no process can go, set by the second law of thermodynamics. Every real process exceeds this minimum by a margin determined by its engineering efficiency. The gap between the thermodynamic minimum and actual energy consumption is the engineering challenge of desalination.

That gap has narrowed dramatically for reverse osmosis. Early RO plants in the 1970s consumed 10–12 kWh/m³. Modern plants with energy recovery devices operate at 3–5 kWh/m³ — within a factor of three to five of the theoretical minimum. Thermal processes — multi-stage flash distillation, the workhorse of Middle Eastern desalination for half a century — consume 10–16 kWh/m³ when operating on primary energy, though they can use waste heat from adjacent power plants that would otherwise be rejected to the environment. The distinction between electrical and thermal energy matters: a thermal plant running on cogenerated waste heat may have a higher primary energy consumption but a lower marginal cost, because the energy input is already paid for by electricity generation.

03 Reverse Osmosis: The Membrane Revolution

Reverse osmosis is the dominant desalination technology for new construction, and the reason is simple: it uses less energy than any thermal alternative. The process works by forcing pressurized seawater through a semipermeable membrane that blocks salt ions but allows water molecules to pass. The natural process of osmosis moves water from a dilute solution to a concentrated one across such a membrane; reverse osmosis applies hydraulic pressure greater than the osmotic pressure of the feed water to drive the flow in the opposite direction. The osmotic pressure of standard seawater is about 27 bar; the practical operating pressure of an RO plant is 55–70 bar, enough to overcome osmotic pressure and the hydraulic resistance of the membrane and flow channels.

The membrane itself is the heart of the technology. Modern RO membranes are thin-film composite structures: a polysulfone support layer, an ultra-thin polyamide active layer roughly 100 nanometers thick, and a protective coating. The polyamide layer is where the separation occurs — its dense, crosslinked polymer matrix allows water molecules to diffuse through while rejecting dissolved ions. The membrane is wound into spiral-wound elements — flat sheets separated by feed spacers and permeate channels, rolled around a central permeate tube. A typical 8-inch-diameter element produces about 40 cubic meters per day of permeate. A large plant contains thousands of elements housed in pressure vessels, arranged in multiple passes to maximize recovery and water quality.

The key engineering advance that brought RO energy consumption down was the energy recovery device. In a single-stage RO system, the concentrated brine leaving the membrane — still at high pressure — contains roughly half of the input hydraulic energy. An energy recovery device transfers that pressure to incoming feed water, reducing the load on the high-pressure pump. Isobaric ERDs, such as the Pressure Exchanger developed by Energy Recovery Inc., achieve 95–97% transfer efficiency, meaning almost all of the brine's pressure is recovered. This single component reduced RO energy consumption by roughly 40% and made modern large-scale seawater desalination economically viable.

Chart 2: Global desalination capacity growth, 2000–2025. Membrane (RO) technology overtook thermal methods and now accounts for the majority of new capacity. Source: N43 and Hermes, compiled from IDA and GWI data.

04 Thermal Distillation: The Gulf Workhorse

Multi-stage flash distillation, the technology that built the Middle East's water infrastructure, works on a principle that is ancient in concept but sophisticated in execution: boil seawater, condense the vapor, collect the fresh water. The engineering innovation is the "flash" — instead of heating water to boiling at atmospheric pressure, MSF heats seawater to a moderate temperature and then steps it through a series of chambers at progressively lower pressures. At each stage, a fraction of the water "flashes" into vapor because the local pressure has dropped below the vapor pressure for that temperature. The vapor condenses on heat exchanger tubes carrying incoming feed water, transferring its latent heat and preheating the feed. A large MSF plant may have 20–24 stages and operates at top temperatures of 90–110°C.

Multiple effect distillation (MED) is a closely related thermal process that arranges evaporator effects in series, each at a lower temperature and pressure than the last. Vapor from the first effect condenses in the second effect's heat exchanger, providing the heat to evaporate more feed water there, and so on. MED operates at lower top temperatures (60–70°C) than MSF, which reduces scaling and corrosion but also limits the number of effects — typically 8–12 — because the final-stage temperature must remain above the cooling water temperature. Both MSF and MED can be coupled to power plants in a cogeneration configuration, where steam extracted from the power plant's turbine provides the thermal energy for desalination. This is the standard configuration in the Gulf states, where distillation plants have run alongside gas- and oil-fired power stations for decades, and where the waste heat would otherwise be dumped to the atmosphere via cooling towers.

05 Pretreatment: The Unseen Engineering

The reverse osmosis membrane is a sensitive component. It is fouled by suspended solids, scaled by sparingly soluble salts, degraded by oxidants, and clogged by biological growth. The engineering challenge is not just the membrane and the pressure — it is the pretreatment train that makes seawater clean enough to pass through a 100-nanometer polymer layer without destroying it. A typical pretreatment sequence begins with coarse screening to remove debris and marine organisms, followed by media filtration — gravity-fed sand and dual-media filters that remove particles down to about 10 micrometers. Many modern plants add ultrafiltration or microfiltration membranes as a second stage, producing feed water with a silt density index below 3, which is the threshold below which RO membranes can operate reliably.

Chemical pretreatment is equally important. Antiscalants are dosed to inhibit the precipitation of calcium carbonate, calcium sulfate, and barium sulfate on the membrane surface. Sodium bisulfite is added to neutralize chlorine, which is used for disinfection but which attacks the polyamide active layer of RO membranes within hours. The feed water pH may be adjusted to optimize membrane rejection and minimize scaling. The pretreatment system is typically 20–30% of a plant's capital cost and a major contributor to its operating cost, because the filters must be backwashed, the chemicals must be replenished, and the membrane elements — which have a lifespan of 5–7 years in seawater service — must be periodically cleaned and eventually replaced. The pretreatment train is the least glamorous and most critical subsystem in a desalination plant.

06 Brine and the Environmental Cost

For every cubic meter of fresh water a desalination plant produces, it discharges roughly 1.5 cubic meters of concentrated brine — the feed water minus the extracted fresh water, with its salt concentration roughly doubled. A 500,000 cubic-meter-per-day plant — a mid-size facility — discharges about 750,000 cubic meters of brine per day at approximately 65,000 ppm salinity. The discharge also carries pretreatment chemicals, cleaning agents, and heat — thermal plants discharge warm brine that is typically 5–10°C above ambient. The brine is denser than seawater and, if discharged without dispersion, sinks to the seafloor where it can create hypersaline, hypoxic conditions that kill benthic organisms.

Modern brine outfall systems are engineered to mitigate this. Multi-port diffusers distribute the brine across a wide area and use jet momentum to achieve rapid mixing with ambient seawater, diluting the plume to within 1–2 parts per thousand of ambient salinity within a defined mixing zone. Submarine outfalls extend hundreds of meters offshore to reach areas of stronger current that accelerate dilution. The environmental impact is localized but real, and it is the most consistent source of opposition to new desalination projects. In California, the Carlsbad desalination plant's permitting process required over a decade of environmental review and the implementation of intake modifications to minimize entrainment of marine organisms. The brine problem is not a reason to abandon desalination, but it is a design constraint that must be engineered into the project from the outset, not retrofitted after complaints.

07 The Economics and the Future

The cost of desalinated water has fallen substantially over the past two decades. In 2000, the cost of RO-desalinated seawater was roughly $1.50–2.00 per cubic meter. By 2025, the best modern plants produce water at $0.50–0.80 per cubic meter, with the reduction driven by improved membranes, energy recovery devices, economies of scale, and competitive procurement. This is still expensive compared to conventional water supply — surface water typically costs $0.10–0.30 per cubic meter, and recycled water costs $0.30–0.50 — but it is within the range that municipal water utilities in water-scarce regions can bear, particularly when the alternative is running out of water entirely.

The future of desalination engineering is likely to involve further integration with renewable energy. Solar-powered desalination — using photovoltaic electricity to drive RO or concentrated solar thermal to drive distillation — is being demonstrated at scale in the Middle East and Australia. The challenge is the variability of renewable supply: an RO plant designed for continuous operation must either have energy storage or accept reduced output during low-irradiance periods. Forward osmosis, membrane distillation, and other emerging technologies promise lower energy consumption but have not yet reached commercial maturity at scale. The practical trajectory is incremental: better membranes, more efficient pumps, smarter energy management, and larger plants that benefit from economies of scale. Desalination is not a technology awaiting a breakthrough. It is a mature engineering discipline that is steadily getting cheaper, and its deployment will track the global trajectory of water scarcity. The regions that need it most are the ones building it fastest.