Membranes for an ancient mariner

Membrane engineers are developing ways to filter drinkable water out of polluted, salty and contaminated supplies. Jon Evans reports

While swathes of Southern Europe were parched by drought, the summer of 2007 brought the worst flooding in 60 years to the UK. Homes and businesses in Northern Ireland, Yorkshire and the Midlands disappeared under water. When a water treatment plant in Gloucestershire fell victim to the deluge, 350,000 people were left without access to clean water. The painful irony of this situation was captured in headlines quoting Coleridge’s ancient mariner - ’Water, water everywhere, nor any drop to drink’. 

While the water company was forced to distribute and ration bottled water, people pumped filthy floodwater out of their houses and offices. But with small-scale water treatment devices, this filthy floodwater could have been transformed into clean drinking water. The ability to filter seawater into his own drinkable supply is something that Coleridge’s ancient mariner would certainly have been grateful for.  

Such neat, compact water-cleaning devices don’t yet exist - the treatment of dirty water currently involves a range of processes, including filtration, activated sludge and flocculation, that can only take place in a large water treatment plant. But such devices aren’t that far-fetched. 

One US company is already marketing small plastic bags that can transform dirty water into a sugary drink, which they distributed during the flooding caused by Hurricane Katrina in the US in 2005. The central component of these bags is a semi-permeable membrane that allows the passage of water molecules but blocks all unwanted particles and ions. 

Similar membranes are used in some large-scale water treatment plants, but the next generation of membranes promise to be a great deal more efficient. Such advanced membranes could eventually lead to the development of small-scale water treatment devices and even help avert a looming water shortage crisis. 

Thirsty world 

Water consumption is currently doubling roughly every 20 years, due to both population growth and industrial expansion. According to Population Action International, a US advocacy group, this could result in the number of people living in water-scarce or water-stressed conditions increasing from 745 million in 2005 to over 3 billion by 2025. 

At the moment, those experiencing water scarcity tend to live in regions that have historically suffered from a lack of water, such as the Middle East and North Africa. To deal with this scarcity many Middle Eastern countries have turned to desalination, whereby drinkable (also known as potable) water is extracted from seawater. Kuwait, for example, now obtains practically all of its fresh water from the sea. 

But more and more countries, many of which don’t currently suffer from water scarcity, are beginning to build desalination plants, as a way of dealing with increasing demand. In July 2007, while the floods were still raging, the UK water company Thames Water received the go-ahead for construction of the UK’s first desalination plant in east London.  

There are two main types of desalination plant. The first uses a distillation process, in which seawater is heated to evaporate off pure water, which is then condensed and collected. The second uses reverse osmosis (RO), which involves passing seawater through a semi-permeable membrane that allows the passage of water molecules but blocks salt molecules. To overcome the natural tendency of water to travel from low-salt solutions to high-salt solutions (via a mechanism known as osmosis), the water is forced through the membrane at high pressures (55-80 bar).  

Distillation-based desalination plants used to dominate, accounting for 90 per cent of desalination capacity in the 1970s, but have since lost market share to RO plants, following the development of more efficient membranes in the 1980s and 1990s. Most modern RO plants utilise thin film composite (TFC) membranes, which comprise a thin layer of polyamide (only around 1 m m thick) coated onto a microporous polymer, usually polysulfone (which is then supported on a thick support layer of polyester).  

Nevertheless, both types of desalination plant require a great deal of energy (to produce heat for distillation and pressure for RO). As a result, building and operating a desalination plant is still a costly business, with desalination by reverse-osmosis currently two to three times more expensive than standard water treatment. 

This means that desalination plants are currently only available to rich countries, which explains why they have been mostly taken up by the oil-rich but water-scarce Gulf States. It also helps to explain why RO is mainly used for desalination, rather than for normal water treatment. But this could all change if more efficient membranes became available. 

Filtering molecules 

Thomas Mayer and his colleagues at Sandia National Laboratories in Albuquerque, US are developing such membranes. ’We want to reduce the amount of energy required to purify water by desalination,’ says Mayer. ’So in the case of membrane technology, this means reducing the pressure needed to transport water and reject salt in a reverse osmosis process.’  

To achieve this, Mayer and his team studied the way that water and ions are transported across cell membranes in nature, which tends to be accomplished by membrane-spanning protein structures. ’These structures are essentially small pores that do their job very selectively and effectively,’ he explains. This seems to be due both to the small size of the pores, which are only 1-2nm in diameter, and the fact that they have a natural electrostatic charge. Their small size means that water molecules have to line up in a row in order to pass through, providing an orderly and efficient transport mechanism, while the electrostatic charge actively repels specific ions. ’We’re now trying to understand enough about how they work to build a synthetic version that has the same function,’ says Mayer. 

This has led Mayer and his team to investigate a self-assembled silica oxide material containing pores only a few nanometres in diameter. They are now looking at ways to modify the diameters of these pores and also add functional groups, such as charged molecules, to the pore walls, in order to mimic natural pores even more closely.  

This work is still at an early stage, with a working membrane based on this material still a couple of years away. Nevertheless, the potential efficiency gains offered by this kind of nature-inspired membrane are impressive. ’If we assume we can make a pore with more-or-less equivalent transport rates [to natural versions], multiply that by the number of pores we can get into our artificial structure, we end up with a water transport efficiency of about 25 times what can be achieved with current reverse osmosis membranes,’ Mayer explains. 

A similar approach is being taken by researchers from Lawrence Livermore National Laboratory, California, US, led by Olgica Bakajin, but they are creating their nanoscale pores using carbon nanotubes (CNTs). Their work is also slightly ahead of Mayer’s research, because they have actually produced a working membrane. To do this, Bakajin and her team first used chemical vapour deposition to grow vertical arrays of double-walled CNTs. They then covered this array in silicon nitride, which formed an impermeable barrier between the CNTs. Finally, they used ion milling and etching to expose the ends of the CNTs, creating pores through the membrane. 

The CNT pores are a similar size to Mayer’s pores (around 1.6nm in diameter) and have similarly impressive water transport efficiencies. ’The gas and water flows that we measured are 100 to 10,000 times faster than what classical models predict,’ says Bakajin. The fact that the CNTs have atomically smooth walls may also help speed the water molecules through the membrane. 

Bakajin and her team are now in the process of adapting these membranes for practical applications, including desalination. ’We are testing ion removal with our membranes and comparing them to commercial membranes,’ says Bakajin. ’At this point, I can safely say that we are getting very promising results.’ 

Much closer to the market are the nanocomposite membranes being developed by Eric Hoek and his colleagues at the University of California, Los Angeles (UCLA). These are based on conventional membrane polymers, like the polyamides currently used in TFC membranes, into which Hoek and his team incorporate engineered nanoparticles, such as zeolite nanoparticles. These nanoparticles are both highly hydrophilic (water-loving) and highly porous, acting as additional pores through the polyamide. 

’These nanoparticles act just like a tunnel for the water to pass through and they’re so hydrophilic that they will take water out of air,’ explains Hoek. ’So when we integrate this material into the polymer membrane and expose it to water, the nanoparticles will literally suck water in like a sponge.’ An added benefit of these nanoparticles is that they make the whole membrane hydrophilic, which helps to prevent particles from sticking to it. This kind of fouling with organic particles and minerals can quickly degrade the performance of RO membranes. 

Although Hoek’s membranes don’t offer the same efficiency gains as Mayer’s or Bakajin’s membranes, allowing water to pass through at twice the rate of existing RO membranes, they are at a much later stage of development and should prove more straightforward to manufacture. ’We make the membranes in the laboratory in a way that is consistent with the way membranes are commercially manufactured,’ says Hoek. As such, UCLA has already licensed the technology to a company called NanoH₂O, which was specially set up to commercialise Hoek’s membranes. According to Hoek, NanoH₂O will test these nanocomposite membranes in the field over the next few years. 

But it’s not just RO systems that could benefit from advanced membranes; so could a form of distillation known as direct contact membrane distillation (DCMD). This involves passing heated sea water along the surface of a hydrophobic (water-hating) polymer membrane. As the water evaporates, it passes through the pores in this membrane and enters a stream of cold potable water passing along the other side of the membrane. The main advantage of this form of distillation is that the vapour pressure produced by the hot seawater is much greater than that produced by the cold distillate stream, and this difference in vapour pressure helps to push water molecules through the membrane. 

Kamalesh Sirkar, a professor of chemical engineering at New Jersey Institute of Technology, Newark, US, has recently improved the efficiency of this process by coating the polymer membrane with an even more hydrophobic material, such as fluorosilicone. This helps to prevent water droplets from coating the membrane, which can degrade its ability to transport water molecules. 

’We have been able to achieve very significant fluxes for very different concentrations, including very high concentrations of salt,’ says Sirkar. Using his membranes, Sirkar has been able to desalinate brine with a salt concentration above 5.5 per cent, whereas current RO technologies can only deal with concentrations below this figure. This is because the extremely high pressures required for higher salt concentrations tend to damage the RO membranes. Sirkar and his team are now conducting pilot plant studies and are in discussions with a number of companies about commercialising this DCMD technology. 

Go with the flow  

An even more radical option, and one that is creating quite a bit of interest among water treatment researchers, is to focus less on the membrane and more on how to take advantage of osmosis rather than battle against it. This is the idea behind forward osmosis, which offers the possibility of a highly energy-efficient means for extracting drinkable water from dirty or salty feed waters.  

As its name suggests, forward osmosis is essentially the opposite of RO. A semi-permeable membrane separates salty or dirty water from water containing an even higher concentration of some salt-like molecule. This concentration difference means that water molecules in the salty or dirty water will naturally travel across the membrane via osmosis, without the need for any kind of external pressure. 

The obvious problem with this method is that it doesn’t immediately produce pure water, but rather water containing a salt-like molecule. There are two basic approaches for dealing with this problem: either you use the water with the molecules still in it or you try to remove them.  

The first approach has been adopted by a US company called Hydration Technologies, which has developed bags that can transform dirty water into a sugary drink. These bags consist of two pouches separated by a hydrophobic membrane; in one pouch is a sugar-based syrup, while dirty water is added to the other pouch. The highly concentrated syrup draws water molecules through the membrane, diluting the syrup to produce a sugary drink. 

The second, more challenging approach is being attempted by Apaclara, a UK start-up company based in Bristol. ’Our scheme is to have a material which is magnetically separable,’ explains Barnaby Warne, technical director of Apaclara. ’So you will have a nanoscale magnetic particle, and attached to this will be an osmotic agent, which at a high concentration will draw water across the membrane.’ The nanoparticles can then be removed from the water by simply applying a magnetic field. With funding from the US Office of Naval Research, Apaclara is currently developing a prototype small-scale forward-osmosis unit for use by military personnel out in the field.  

Commercial developers and governments will be watching the progress of membrane engineers carefully. Over 70 per cent of our planet is covered by water, and the ability to convert this cheaply into a drinkable supply for a burgeoning population is an exciting prospect. Small devices could prevent the difficult and incongruous task of providing and rationing stocks of clean water to people deluged by flood. Should the UK weather hold any future unpleasant surprises for Gloucestershire residents, a small company just down the road in Bristol could provide some invaluable help.  

Jon Evans is a freelance science writer based in Bosham, UK