Choosing a Desalination Process: RO, MSF and MED Compared

The first question in any desalination project is which process to use. In practice only three technologies are deployed at scale: reverse osmosis (RO), multi-stage flash (MSF) and multi-effect distillation (MED). Everything else — electrodialysis, membrane distillation, freezing — is either niche or still at demonstration stage. Pick the wrong process and the electricity and maintenance bills will punish you for the next fifteen years.

1. How the three processes differ

RO is pressure driven. A high-pressure pump pushes seawater to 5.5–6.8 MPa, above its osmotic pressure, forcing water molecules through a semi-permeable membrane while salt, colloids and bacteria are rejected. There is no phase change, which is why energy consumption is low.

MSF is heat driven. Seawater is heated to 90–120 °C and passed through a series of chambers at progressively lower pressure; part of the water flashes to vapour and is condensed as distillate. It runs on heat rather than pressure and is usually tied to power-station or petrochemical waste heat.

MED reuses its own steam. The vapour produced in one effect becomes the heating medium for the next, so thermal efficiency is higher than MSF and operating temperature drops below 70 °C — which cuts scaling risk substantially.

2. The parameters that decide the argument

3. Match the process to the project

Vessels, islands, fishing villages and small plants (0.5–200 m³/d)

There is essentially one answer: reverse osmosis. Compact, containerisable, starts producing within minutes of power-up, and day-to-day work is limited to changing cartridges and periodic membrane cleaning. Marine units with energy recovery can get down to 2.5–3 kWh/m³.

Coastal power stations and large petrochemical sites (10,000+ m³/d with low-grade steam)

MED or an MED + RO hybrid is usually preferred. The steam is already waste heat, so the marginal cost of making water is very low, and distillate purity is high enough for boiler make-up. New MSF builds are rare today; most MSF capacity is legacy plant in the Middle East.

Municipal supply and industrial parks (1,000–50,000 m³/d, no waste heat)

RO remains the default. Scale brings specific energy down to around 3 kWh/m³ when combined with energy recovery and variable-frequency control. Several plants now deliver water at roughly USD 0.6/m³, competitive with long-distance water transfer.

High-salinity brine and zero liquid discharge (ZLD)

Typically RO for concentration followed by evaporation and crystallisation. No single process closes the loop on its own.

4. Three selection traps we see repeatedly

Comparing equipment price instead of water cost. Membrane systems are cheap to buy but membranes need replacing every 3–5 years and the high-pressure pump is roughly half of operating cost. Thermal systems are expensive up front but the heat is free. Do the maths over a ten-year life cycle.

Ignoring seawater temperature. RO output falls noticeably as feed temperature drops. With 5 °C winter seawater, output can be 30–40% below the design figure, so membrane area must be checked against the coldest condition — otherwise you run short every winter.

Cutting pretreatment to save money. Around 80% of RO failures trace back to pretreatment. Multimedia filtration, 5-micron cartridge filtration and antiscalant dosing are not optional; whatever you save here you will spend on cleaning and replacement membranes.

5. Our recommendation

Shenzhen Duruofei Technology Co., Ltd. has built marine and island RO systems for years, and the pattern is consistent: for small and medium capacity, choose RO and put the money into pretreatment design, energy recovery and automatic control. Those three decide both reliability and water cost. If your site has waste heat available, we will model an MED hybrid with you. There is no universally correct process — only the one that fits the site.