GERHARD HUMMER was pondering a serious plumbing problem. He was
trying to unravel the inner workings of tiny proteins called
aquaporins, which are found in the walls of living cells. Each
aquaporin is threaded by a narrow pore that helps control the flow of
water into the cell. The pore is a complex thing, narrow in parts and
wide in others, lined with a variety of chemical groups that mostly
repel water. But it is basically a pipe.
And that realisation made Hummer, working at the US National Institutes of Health in Bethesda, Maryland, turn his attention to carbon nanotubes.
Consisting of curled-up sheets of carbon and just nanometres wide, they
are essentially smooth pipes of water-repelling graphite. Hummer hoped
that their simple structure might offer new insights into the way that
water travels through aquaporins. It proved a smart move. Nanotubes
have not only helped researchers like Hummer understand water flow in
proteins, but they are also enabling scientists to devise a host of
nanoscale plumbing parts – such as molecular pumps, gates and valves –
capable of moving and filtering everything from salty water and
hydrocarbon fuels to gases such as carbon dioxide. It seems that these
humble tubes could hold the key to cheap desalinated water, better fuel
cells and new strategies to tackle global warming.
study of fluid flow in nanotubes kicked off around a decade ago when
along with two colleagues he created a detailed computer simulation of
the way water moves inside a carbon nanotube just 0.8 nanometres wide.
When they dunked the tube into a tiny tank of virtual water, the
researchers found that a thin thread of water molecules rushed into the
interior of the tube. This was surprising, given the narrowness of the
nanotube’s pore and the water-repelling nature of its carbon surface.
Then when they tweaked the simulation, slightly increasing the
repulsion between the water molecules and the carbon atoms of the
nanotube, they were surprised to see that the tube emptied almost
instantaneously. When they decreased the strength of the repulsion, the
tube filled again. The ease with which they could fill or empty the
tube was unexpected, and their results – published in Nature in 2001 (vol 414, p 156) – implied that just small changes in charge or even tube geometry might be used to move water through real nanotubes.
and his colleagues then simulated an array of short carbon nanotubes,
again each one just 0.8 nanometres across, packed side by side in a
membrane. When pure water was added to one side of the membrane and
brine to the other, water immediately flowed down the nanotubes into
the brine, driven by the difference in salt concentration. What
surprised the researchers was the speed of the flow:
it seemed that the chain of water molecules passing through each
nanotube experiences virtually no friction, moving nearly ten thousand
times faster than theory predicts. What’s more, Hummer’s team found
that ions could not get through the pores in either direction. In
principle, the nanotubes were wide enough to let the ions through, but
it seems they could not make it when the water was confined by the tube.
The reason for this behaviour is actually straightforward (see diagram).
Charged ions, like those in brine, are surrounded by a network of water
molecules in a so-called “hydration shell”. But there is no space to
accommodate this network inside a nanotube. Instead, each water
molecule is hydrogen-bonded to just two others, one in front and one
behind, forming a continuous, organised chain. For an ion to enter a
nanotube, its hydration shell must be stripped away. Hummer’s results
suggest this costs too much energy, so the ions stay put.
This, in effect, is what occurs in a conventional desalination process called reverse osmosis
in which brine is filtered by a fine membrane so that pure water passes
through and the ions are left behind. However, this process requires
large amounts of energy to pump the water through the membrane, which
is one reason why desalinated water is expensive. The high flow-rates
measured by Hummer suggested that desalination would be more efficient
if it could harness the ion-blocking properties of nanotube membranes.