A membrane that can remove salts from water more efficiently

ALAN TURING was no slouch. He laid the mathematical groundwork for modern computing. He led the successful effort to crack Germany’s Enigma code during the second world war. And he also, though it is less well known, made an important contribution to chemistry with a paper winningly entitled “The chemical basis of morphogenesis”. In it he described how the diffusion of two chemicals that react with each other can, in certain circumstances, produce complex patterns of blobs and striations. These patterns, now called “Turing structures”, bear an uncanny resemblance to many that are found in nature: a zebra’s stripes, for example, or a ladybird’s spots.

The extent to which such processes are involved in the embryonic development of animals is debated. But, on a more practical note, Zhang Lin of Zhejiang University in Hangzhou, China, and his colleagues now hope to turn Turing’s chemical insights to the task of improving desalination, a process that provides drinking water for around 300m people. To do so, they have made a membrane laced with microscopic Turing patterns that can remove salts from water up to four times faster than commercial alternatives. Their research is published this week in Science.

During desalination, seawater is often pumped first through a porous “nanofiltration” membrane made of a substance called polyamide. This removes bulky ions, such as magnesium and sulphate, as well as bacteria and other large particles. After that, it passes through a second membrane which has even tinier pores. This step, called reverse osmosis, removes ions smaller than magnesium and sulphate, particularly the sodium and chloride ions that make up common salt and that give seawater its characteristic taste.

The membranes employed for reverse osmosis are rough, and so have a large surface area through which water can pass. Nanofiltration membranes, by contrast, are smooth. That, Dr Zhang reasoned, meant that they might be improved. To introduce the necessary roughness he needed some way to modify the chemical reaction by which the membranes are made. This process, known as interfacial polymerisation, involves two chemicals. One, piperazine, is soluble in water. The other, trimesoyl chloride (TMC), can be dissolved only in an organic solvent such as decane, an oily hydrocarbon.

When piperazine and TMC meet, they react to form polyamide. But if the one is dissolved in water and the other in oil, which famously do not mix, then the reaction can happen only at the surface where the oil is floating on the water. The result is a polyamide sheet. In practice, in industrial conditions, this reaction is usually performed on a porous support that is first soaked in piperazine before one side is exposed to TMC. The polyamide sheet then forms on that side of the support. Dr Zhang’s insight was to see that this arrangement might be modified to be the type of two-chemical system that Turing described in his paper—and that if it could be, the resulting patterns would act as surface-area-increasing bumps.

For a system to form Turing patterns, two chemicals must diffuse at different rates through the medium in which the reaction is taking place. The rates cannot, however, be too different. The ideal discrepancy is about a factor of ten. To achieve this, Dr Zhang added polyvinyl alcohol to the piperazine solution, to make it more viscous and slow piperazine’s diffusion.

The upshot was the creation of polyamide sheets full of either tiny, hollow bubbles or interconnecting tubes, depending on the concentration of polyvinyl alcohol used. These are just the sorts of surface-area-increasing features that Dr Zhang had hoped for. And they did the job. The best were able to handle a fourfold increase in flow rate with no loss of performance.

Flushed with success, Dr Zhang is now turning his attention to the reverse-osmosis membranes. Though these are already rough, he thinks he can make them rougher. They, too, are made by interfacial polymerisation, so he may well be able to do so. And if both sorts of membrane can be improved, the process of desalination, which is likely to become more important as demand for water increases, will be made cheaper and more effective.