Desalination Process Choice: Why MSF Holds the Gulf and RO Takes the Rest

Two families, and one decision that follows from them

Every desalination route belongs to one of two families: you either boil the water out of the salt, or you push the water through a barrier that the salt cannot cross. Thermal plants evaporate and condense; membrane plants pressurise. More than twenty process variants exist on paper, but the commercial field narrows to multi-stage flash (MSF), low-temperature multi-effect distillation (LT-MED) and reverse osmosis (RO).

The choice between them is rarely about which process is more advanced. It is about the price of heat, the quality of the feed and the size of the unit. A plant sitting next to a power station has waste heat that costs nothing to use; a plant on a windy coast with cheap electricity has the opposite advantage.

Where each thermal process fits

MSF works by flashing: heated brine enters a series of chambers at progressively lower pressure and part of it evaporates instantly in each one, with the vapour condensed to freshwater. It still accounts for the largest output among global plants, and its attractions are maturity, operational safety and flexibility. It is normally built alongside a thermal power station and suits large and very large capacities, which is why Gulf states keep choosing it, with single units already producing 40,000-50,000 tons of freshwater per day. Its weaknesses are energy intensity and the cost of corrosion control at high temperature.

LT-MED saves energy, tolerates poorer feedwater and delivers high product quality. Multi-effect distillation reuses the latent heat of one effect in the next, and running at low top-brine temperature reduces scaling enough that pretreatment can stay simple. The trade-off is a larger plot area and a taller, more complex vessel arrangement than a membrane train.

Vapour compression sits between the two: a compressor raises the pressure and temperature of the vapour so it can drive the next effect. A 30 m3/d vertical-tube atmospheric unit, a 30 m3/d horizontal-tube negative-pressure unit running at 72 C, and a 30 m3/d OTE/VC variant were all built and tested in China during the 1980s, and the problems logged in that work shaped later designs more than the successes did. Vapour compression is popular for small and medium duties because it needs no steam supply.

Dew-point evaporation is a different idea again. A carrier gas is humidified with seawater and then dehumidified, and the heat of condensation is returned to the evaporation chamber rather than thrown away. Two experimental units with effective heat transfer areas of 9.6 m2 and 2.75 m2 were built and instrumented to establish the parameter correlations, but the process has not displaced the mainstream three.

Electrodialysis and the ion-exchange membrane

Electrodialysis does not separate water from salt; it moves salt out of the water. An ion-exchange membrane is a functional sheet 0.5-1.0 mm thick, classified as cationic or anionic by the sign of the charge it lets through. Stack cation and anion membranes alternately and you get alternating diluting and concentrating compartments, so freshwater and brine leave separately.

The economics favour brackish water over seawater, because the energy demand tracks the quantity of salt removed rather than the volume of water produced. Outside desalination the same stacks are used to concentrate and purify process streams in the chemical, pharmaceutical and food industries.

Freeze desalination and the triple point

Seawater has a triple point where vapour, liquid and solid coexist. Move away from it and the equilibrium breaks, with the phases reverting to one or two forms. Vacuum freeze desalination uses that behaviour to make seawater evaporate and freeze at the same time, with water itself acting as the refrigerant; the ice crystals are then separated and washed. Energy demand is lower than in distillation, corrosion and scaling are far less severe, pretreatment can be simple, and high-salinity feed is not a problem. The reason it remains marginal is mechanical: handling and washing an ice slurry at scale is difficult and the equipment is unfamiliar to most operators.

Reading the Chinese cost curve

China's programme ran through a sequence of institutions and targets, and the details explain why domestic costs fell as fast as they did. Research on electrodialysis began in 1958. In 1965 the chemistry department at Shandong College of Oceanology became the first in the country to work on CA asymmetric reverse osmosis membranes, and during the 1970s it studied hollow-fibre and spiral-wound elements and began to industrialise them. In 1984 the State Oceanic Administration (SOA) built the Hangzhou water treatment centre on the basis of its desalination research office, and in 1992 the State Science and Technology Commission established a national engineering centre for liquid separation membranes there.

Reverse osmosis stayed a foreign monopoly until the late 1990s. The demonstration projects changed that. Around 2000, 1,000-ton/day RO trains were commissioned on Changdao Island in Shandong and at Shengsi in Zhejiang, using work-exchange energy recovery devices that brought water production energy consumption down to 4 kWh/ton. In 2002, with support from a National Development and Reform Commission (NDRC) industrialisation project, a 10,000-ton/day plant followed at Rongcheng. Salt rejection across the domestic industry improved from 99.2% to over 99.7%, and by the end of 2010 more than 70 plants with a designed capacity of 600,000 cubic meters/day were in service.

The policy side moved with the technology. The twelfth five-year plan for the industry set five tasks, starting with stronger technological innovation and the R&D of core desalination technologies and key components, and continuing through equipment manufacturing, engineering demonstration, wider use of the product water, and standards.

Cost is the number that decides adoption: desalinated seawater now sits at 4-5 CNY/ton, which is inside the range most coastal industrial users will accept. For shipboard and remote duty the older units were far smaller, starting with 5 m3/d vapour compression and flash units built in the 1960s using diesel jacket-water waste heat.

The high-recovery claim, read carefully

One technology family markets itself on recovery rather than on energy. RSE technology is claimed to reach 95% recovery against 35% for conventional processes, to treat water with salinity as high as 16%, and to cost 16-27 US cents per 1,000 litres, with investment at a quarter of that for distillation or RO. Those are strong numbers and they deserve scrutiny: recovery and salinity tolerance trade against each other in every membrane process, so a unit that claims both needs its pretreatment and its brine disposal examined before the capital figure is believed.

How to specify, in the order that matters

Fix the product water use first, because it sets the allowable boron, chloride and TDS. Then fix the feed, including seasonal turbidity and algal blooms, since that decides whether pretreatment is a filter or a full dissolved air flotation train. Only then compare processes on total cost per cubic metre, with the heat and power price for that specific site.

Three general rules hold. Membrane methods dominate wherever power is cheap and feed quality can be controlled; thermal methods dominate wherever waste heat is free and feed quality cannot. The three main technologies will stay in industrial use for decades, but the share held by reverse osmosis will keep growing. And the concentrate is not automatically a waste: raising the freshwater yield concentrates the valuable elements in the brine, which is the starting point for chemical recovery rather than a disposal problem.