Beside membranes and heat sits a third family of methods — experimental, niche, or economic only for a particular feed. They still matter when you choose a process or watch where research is heading.
Methane hydrates and freezing
Hydrate crystallization builds a hydrogen-bonded water lattice around a guest gas. The crystals usually form at high pressure and raise the freezing point by at least 12°C. The water-to-gas ratio is about 6 to 17. Crystals are separated from brine, washed and melted; the clathrate gas is reused. Compared with freezing, the process runs warmer and uses less energy. A 1960s pilot failed because the crystals were too fine to separate. If it were made to work, the booklet put the cost at about $0.46–0.52 per cubic metre.
Taking fresh water from ice is thousands of years old among the peoples of Central Asia and Siberia. When brine is cooled to a critical point, ice crystals form and salt is pushed out, because the hydrogen bond between water molecules is stronger than the van der Waals bond with sodium chloride. Indirect freezing uses a refrigerant and compressor; direct freezing freezes the water itself below atmospheric pressure. The attractions are low energy, little corrosion and no chemicals. The drawbacks are capital cost and the mechanics of moving an ice–water slurry. Today the process is used more on industrial wastewater than on municipal drinking water.
Solar humidification
The first solar still is dated to 1551. The first public plant was built in 1872 by Charles Wilson at a nitrate mine in Chile: 4,700 m² of area, more than 2,300 litres a day, and about forty years of operation. The principle is the greenhouse effect. Vapor hits glass or plastic, condenses and runs into a gutter.
One cubic metre of still volume yields about 4 litres a day, so 4,000 m³/day would need around 100 hectares. Heat is free; cleaning plates, managing scale and pumping are not. The process is mainly economic for villages and farms without grid power.
Hybrid RO and ED, ion exchange and capacitors
RO recovery is capped by brine osmotic pressure, salt precipitation and supersaturation at the membrane. Feeding RO reject to electrodialysis can push recovery above 90% and shrink the final brine stream. In a parallel layout the RO membrane sits between two ED ion membranes; continuous ion removal lowers concentration polarization and can cut energy by about 15% versus RO alone.
Ion exchange swaps solution cations and anions for H+ and OH- until NaCl is removed. It is usually too expensive for public drinking-water desalination, but it works as RO pre- or post-treatment and for brackish water around 2,000 ppm. Seawater above 35,000 ppm needs several beds in series.
Flow-through capacitors resemble electrodialysis: ions migrate onto carbon electrodes while voltage is kept low enough to avoid faradaic reactions. When the electrodes fill, the current stops and the ions return to a concentrated stream. Energy to take TDS from 500 to 25 ppm is reported at about 0.7 kWh/m³.
Water from hydrogen plants
In a fuel cell, hydrogen enters at the anode and oxygen at the cathode. Electrons form an electric circuit; protons combine with oxygen to make pure water. A hydrogen power plant can therefore deliver steam, electricity, electrolyzed hydrogen fuel and freshwater together. The route is still a complement, not a substitute for large RO or MSF plants, but it shows desalination can sit inside a clean-energy system.
