There are many sort of sieves, from the ones I uses daily to get rid of the water after the "spaghetti" have cooked to the ones used to separate trouts in fish farms.
The concept behind a sieve is quite clear. You have a surface that is impermeable and you have holes in that surface that will let a substance move from one side of the sieve to the other. By changing the size of the holes you can selectively filter a substance. In my case the sieve holes are too big to let spaghetti go through so they remain in the sieve whilst the water flows through it.Chemists have made much more sophisticated sieves, with holes that can let only certain kind of molecules (like proteins) go through the filter. The characteristics of a filter are related to its holes (their size and in some situations to their shape)and to the thickness of the surface. The thicker the surface the less permeable it is (the more resistance it offers to what it is supposed to go through the holes. Thatis because as "things" get across the holes there is a drag on the inner surface of the hole, as if the hole is actually a pipe. Now, if you are dealing with spaghetti you will never think of the holes in the sieve as pipes (or tubes) but if you imagine having a sieve that is supposed to let a certain type of molecule pass and not another than you are down to nanometers radius holes whilst the surface thickness may be in the order of a tenth of a mm, that is some 10,000 times the size of the hole. Just for comparison, if you have a sieve to let tennis ball go through you will need holes of about 7cm (6.7 to be precise but that might be too tight...) the length of the hole would be something like 700 meters! That is quite a pipe! And you can imagine that the aspect of permeability gets pretty relevant.
All this preamble to introduce a news from ETH Zurich where researchers have been working together with colleagues at Empa and LG electronics to create a sieve that is just two atoms thick. It is made by two layers of graphene that are overlaid one another.
Now, the resulting membrane is really thin! Two atoms thick means less than half a nanometer (to be exact it can be 2 Angstrom thick but giving some breath for bumps and non planar disposition of atoms you can average the thickness to 0.5nm) that is 50,000 times thinner than a sieve based on a sheet of paper! The permeability is as good as it can get, and therefore the sieve is as good as it can get.
The drill the holes of the exact dimensions researchers have used the same technology used for etching wafer for chips. A stream of helium or gallium ions are beamed (focused ion beam milling) on the graphene surface to create the holes and they can be made to incredible precision, as an example to let a certain kind of molecule go through and block the others.
The researchers had to use two layers of graphene because the structure of a single layer can present some missing atoms and it may be impossibile to drill holes of the exact shape. By overlaying another layer of graphene it gets statistically very unlikely that two defects in the two layers are exactly overlapping.
The researchers think that an industrial production of these filters can make water desalinisation possibile at an affordable cost. Today desalinisation requires a considerable amount of energy (to remove the chlorine atoms and other undesirable molecules. By drilling holes that only H2O molecules can pass you can create pure water. Because of the extremely high permeability the energy required is reduced to a minimum and the velocity of desalinisation increases considerably.
The key word here is: “industrial production”. So far the sieve has a surface area of one hundredth of a square mm. To have a filter than can be used for desalinisation you need to have hundreds of square cm filter surface, that is some million times the size of what they have achieved today. Still, it is an important step in the right direction.
With the investment being poured on graphene in Europe (one of the two flagship projects, 1B€ investment, is on graphene) and in the US and Japan, South Korea we might reasonably expect to have this kind of filter available by the end of this decade.