Mass spectrometer

This is the refining and construction part of the fab. It takes elements and directs them atom at a time to specified locations.

This page includes time to build per mol of product, which in turn shows the frequency of the atom stream at the deposition end.

Scaling up might involve a lot of deposition ends.

We decided the average mol of raw material is 54g so as a rough guide, 1kg is 20mols.

There's no getting away from the fact that we need a lot of these. Each beam can handle 1E10 atoms/s at most, with 1GHz discrimination, whether it is the 3D printer unit or the refinery. So long as the size of these is small enough, production will scale to usable periods.

Quote: "Since the ions of the different isotopes have the same electric charge but different masses, the heavier isotopes are deflected less by the magnetic field, causing the beam of particles to separate out into several beams by mass, striking the plate at different locations." The Wikipedia article on Oak Ridge's Calutron, which was doing much the same task, but we can't make a magnetic field. What we're going to do is cheat a bit.

Let's start with the refinery process

If we batch the ionized plasma, we can have a pulsed group of ions.

We can string that out with an accelerating voltage, so that at the end of the acceleration phase all the light ions are at the front of the beam and all the sluggish heavyweights are at the back. Okay so far? A long chain arriving at the deflection plate gap, in time order of isotopes.

What we now do is to deflect them. The trouble is that a constant voltage across the plates will exactly undo the sorting we've achieved. The light ones, because they get there first, are going faster so they deflect less. The heavy ones are going slow so they're being pushed sideways for longer and they end up taking exactly the same path as the light ones did. The beam all ends up in one place.

But. Here's the cheaty bit. We vary the voltage on the deflector plates with a sawtooth waveform. We hit the light ones at the front with a big kick, the heavy ones at the back with a light kick, and we force the light ones to bend a lot and the heavy ones to bend less.

If we shape the waveform correctly, the isotope separation by mass number will be linear. Okay, so it's not exactly a sawtooth but it's quite like. The sawtooth repeats just as the next strung-out batch of ions arrives from the next bang event at the plasma gun.

Now, the construction part

We have the same control over a single atom in that we can direct it to an exact spot on the item we're fabricating. We might have X and Y plates for this stage. The atom has a speed, we apply a deflection and it goes where it's been sent.

What we have at the plasma gun end is various heaps of refined isotopes, so we can send a wave all of one isotope, then a wave of another, and a third and a fourth and so on, building up layers in specified places as needed. The trick here is that we're not inertia-separated, the stream of a given isotope from a single heap will all arrive at the deflection plates at an identical velocity but still in a time-separated procession.

It's no longer a bang, it's a single-isotope stream. The plasma generator is no longer taking raw feedstock, but it will have instead to direct the vaporizing and ionizing energy to a nominated heap each time it generates the next part of the stream. The timing of the arrival will have to be calculated for each batch because the timing of the stream is now dependent on the mass number. The generator will need a process, currently unspecified, to produce the stream over a time period instead of an single moment.

Jowan and I took a good half hour to come up with this, we hope you're impressed. John (talk) 22:43, 22 January 2020 (UTC)