Almost three quarters of Earth's surface is covered with water, but most of it is too salty to drink. And the 2.5 percent that is freshwater is locked up either in soil, remote snowpacks and glaciers or in deep aquifers. That leaves less than 1 percent of all freshwater for humans and animals to drink and for farmers to use to raise crops—and that remnant is shrinking as rising global temperatures trigger more droughts. The upshot: it's becoming increasingly difficult to slake the world's thirst as the population grows and water supplies dwindle. Analysts at the investment bank Goldman Sachs estimate that worldwide water use doubles every 20 years.

So the search for new water sources is on. One proved candidate is desalinization—technologies that extract the salt from brine drawn from the oceans or saline aquifers to create potable water. But the historically high price of desalinization has largely kept it at bay, a situation that's changing as technology improves and growing demand squeezes freshwater supplies .

"The two main desalinization techniques are distillation and reverse osmosis, or RO," says Menachem Elimelech, an environmental engineer at Yale University. "Distillation, in which the raw water is evaporated and then condensed as freshwater, is energy-intensive, so it's mainly used in the Middle East where oil is abundant." Thermal salt-removing processes require high temperatures so they tend to be expensive (more than $1 per cubic meter of freshwater), but the use of rejected "waste" heat from other industrial or power plant operations for co-generation can cut energy expenditure.

More commonly, however, desalinization plants rely on RO, which is based on high-tech polymer membranes that are permeable to water, but reject the passage of dissolved salts, Elimelich says. When a saline solution sits on one side of a semipermeable membrane and a less salty solution is on the other, he explains, water diffuses through the membrane from the less concentrated to the more concentrated side. Scientists call this phenomenon osmosis, which tends to equalize the salinity of the two solutions.

In the 1950s and '60s researchers realized that they could reverse the process by applying pressure to the more concentrated solution, causing water molecules there to traverse the membrane, leaving behind a condensed brine. To counter the osmotic pressure that arises between the solutions and force water back through the membrane, desalinization plants must utilize high pressures of 7,000 to 8,300 kilopascals (71 to 86.5 kilogram-force per square centimeter or 1,000 to 1,200 pounds per square inch), he notes.

Common RO membranes are thin-film composites that combine a mechanically robust support layer made of microporous polysulfone with a micron-thick polyamide "filter" layer through which water molecules can pass but nothing else. The latter substance is "a second cousin to DuPont's Kevlar—the super-strong aramid polymer fiber used in lightweight body armor," says Bill Mickols, senior research scientist at Dow Water Solutions (DWS) of Edina, Minn., the biggest supplier of such products. RO membranes have matured during the past two decades, he says, with marked improvements in water permeability, salt-rejection capability, operating life (now as long as three to five years) and cost.

These advances, in combination with energy-recovery devices that take pressure from the concentrated brine stream and transfer most of it to the incoming water flow, have made desalinization more affordable. Current RO facilities desalinize seawater for 68 to 90 cents per cubic meter. The average delivery price of municipal water in the U.S. is around 60 cents a cubic meter, according to the American Water Works Association.

Other improvements currently in the works include measures to maintain process flow. RO plants must filter seawater and inject chemicals to eliminate particles that could cause clogging, and the membranes are washed regularly to lessen scale formation and biofilm fouling, says Benny Freeman, a chemical engineer at the University of Texas at Austin. "Chlorine is added to sterilize the water," he says, "but operators usually need to dechlorinate it afterwards to protect the membrane from chemical degradation." Freeman and James McGrath, a polymer scientist at Virginia Polytechnic Institute and State University in Blacksburg have modified chlorine-resistant polysulfone to serve as a desalinization membrane.

Researchers elsewhere are meanwhile attempting to work around the RO's reliance on high pressure. Elimelech and entrepreneur Robert McGinnis have advanced a process called forward osmosis (FO) that could reduce the energy needed to purify water by 90 percent. FO takes advantage of the osmotic pressure difference between a concentrated "draw" solution and a raw water stream to drive water through a semipermeable membrane.

"The right draw solution means you don't have to do all the work with pressure," McGinnis notes, who recently formed a start-up company called Oasys to commercialize the technology. The main challenge, he adds, is to select a nontoxic draw solute that may be simply and economically removed.

Oasys plans to use an ammonia and carbon dioxide mixture as a draw solute. When the solution is heated, the dissolved ammonium carbonate and related salts decompose into their precursor gases, enabling easy removal. Oasys' process, McGinnis says, can run efficiently on small quantities of electrical power combined with "waste" heat (less than 120 degrees Fahrenheit, or 50 degrees Celsius) from industrial operations. When fully scaled up, FO desalinization is expected to cost only 37 to 44 cents a cubic meter.

More speculative desalinization research aims to create "superflux" membranes that allow water to pass through more easily, says Mark Shannon, director of the Center of Advanced Materials for the Purification of Water with Systems (CAMPWS) at the University of Illinois at Urbana–Champaign. Investigators have shown, for example, that carbon nanotubes can convey water at unexpectedly high flow rates. These so-called water wires may be able to pass a greater volume of water at a given pressure than existing membranes while still blocking out hydrated salts.

Biomimetic membrane technologies are also under development, Shannon says. These materials try to imitate the ability of the minute pores in biological cell membranes to selectively allow water to flow through while preventing the passage of salt ions. The Danish firm Aquaporin, for instance, is embedding natural aquaporins (water channels) extracted from green plants in membranes that it hopes to market this year. Others, including CAMPWS scientists, are working on artificial active nanopore structures.

Recently, the Madrid Institute for Advanced Studies agreed to collaborate with Valladolid, Spain–based engineering company, PROINGESA, to design a capacitive de-ionization process that applies an electrical potential to raw water to attract dissolved salt ions toward oppositely charged electrodes, where they are adsorbed and later removed.

Some 13,000 desalinization plants capable of producing 52.3 million cubic meters (13.8 billion gallons) of potable water a day are currently in operation, according to the International Desalination Association (IDA). But that is only a half a percent of global daily water use, a figure that would grow faster if process costs could be further reduced. Nevertheless, construction of desalinization facilities rose at an annual clip of 17 percent since 1990, the IDA reports.

The trend worries many local environmental groups, such as California's Surfrider Foundation or Australia's Nature Conservation Council of NSW, which are concerned about protecting nearby ecosystems by safely disposing the concentrated brine left from the process as well as increased fossil-fuel use and the resulting greenhouse gas emissions.

In any case, a new market analysis by Lux Research forecasts that the global desalinated water supply will grow at a compound annual growth rate of 9.5 percent during the next decade as Australia, Israel, Singapore, California and others build desalinization plants for seawater and inland brackish water as well as for water recycling. This means that output will reach 54 billion cubic meters a year (54 trillion liters/year) by 2020, or triple what it had been in 2008. CAMPWS's Shannon agrees, predicting, "We're going to see exponential growth in desalinization over the next few decades."