The idea is this:
To build, atom by atom, a tube of diameter of a chosen atom.
From the famous experiments with scanning tunnelling microscope, you can see that scientists have a limited ability to place atoms as they please on a flat surface. This is accomplished by a combination of ingenuous materials like piezodrives, lots of time and pacience, and very controlled circunstances (high ultra vaccum, and so on).
Aside from writing and drawing stuff, that isn’t much more than a display of tech marvel. We have yet to build complex, useful nanostructures like a nanoassembler, for example.
I’ll be streching this STM idea, and making some guesses along the way – since I don’t have first hand experience with such a microscope. There is only so much I can check because most papers are blocked behind a pay firewall.
So we can place atoms where we please over a substrate, cool. What if go arranging them side by side, leaving just a space in the middle, of precisely the size of the atom. This is the layer A.
As an example, let’s pick gold atoms on a copper substrate:
On another layer B, atoms are just placed side by side, filling the entire surface. These two layers are then sandwiched, facing each other, and you apply pressure to keep them together.
Note that this would be time consuming, since it took forever just to write the NIST logo. Advances in autonomous atom assemblers (AAA) are useful here.
The size ‘L’ of this should be as large as it’s possible to make it, so that it’s easier to manipulate. I suspect that as you try to make it bigger you will enconter more and more defects on your flat surface. ‘d’ ends up being just a little bit smaller than the atomic diameter.
To know what ‘D’ to use requires a study in particular, we will get back to it later.
The substrate is a a very carefully grown monocrystal with the least defects possible, to avoid gas difusion.
By doing that, luckily, you will end up with a very expensive artificial atomic tube. Choose a different type of atom and subtrate, you have changed the diameter of the tube. It really didn’t became clear to me if placing atoms in multiple layers is possible, but that should increase the possible diameters you could make – up to molecules size.
I can think of two possible applications for this:
First, the idea is that atoms that fit this size will find this tube and tunnnel through it. So you’d have the means of placing atoms in a precise location, by changing the position of this tube with a piezodrive. You’ve have taken atoms into the gas state in a chamber above, this filter in between, and just below it the surface you are building. This is based on the idea of atomic layer deposition.
With two, independent, blocks like this on top of each other, you would have the means of closing and opening the tube at will. Maybe this will bring us closer to building things in the nano-scale.
- Will atoms pass on this tube, and how would they behave in there?
- What is the thickness ‘D’ of this tube that provides this desired effect, of passing only certain atoms?
- Does more or less atoms pass if you increase D?
- Do they also leak somewhere else?
These are all question that could be answered by an experiment (or possibly by someone more educated than myself in these matters).
The second application is atomic separation. Assume you stack several of this units side by side forming a sieve. The ultimate sieve, by the way. If you place smaller and smaller ones in sequence, you can sort out the individual atoms of a gas or liquid.
It would be very nice to get 99.999999% pure samples from this, or to separate gold ions suspended in a solution, for example.
[IMAGE: STM constant current image of Ag, at 160K image credits]