Outgassing

[BenTC]

I was wondering about the vacuum requirements for experiments wrt outgassing from the surface and bake out requirement - and how each time you open up the vessel to modify the contents the surface can uptake atomspheric gases that might need to baked out again. To what degree might rebaking be required? (1)

I was reading the following http://en.wikipedia.org/wiki/Wikipedia: ... ovember_27 ...

Even small (10-cm diameter cylinder) vacuum chambers suffer from a variety of interesting phenomena - first, of course, is leakage and backflow through the pump system; so you switch off the roughing pump and turn on the bigger roughing pump or the turbo pump (which you can't even operate until you're at pretty low pressure!). Next, leakage through the real, non-ideal seals, flanges, and gaskets; and then when you get down to ultra high vacuum, weird things begin to happen. Your pumps have set up a steady state, you're accounting for leakage; but the pressure doesn't go down any more! Weird! Where did all these gas atoms come from! Every time you pump them out, something still registers. By now, your pressure is so low that hydrogen atoms which used to be embedded into the metal lattice of your structure start to outgas. So you turn off the turbo and turn on the ion pump; and you get most of the hydrogen out. But now some really weird things happen. Helium and Neon start percolating through the two inches of solid, hermetically sealed steel. Because these noble gases are sort of small and they don't really interact with the electrons of transition-metal atoms, they just "squeeze" through the pore spaces. And, because they aren't ions, you can't get rid of them with the ion-pump! So, you turn down the cryo and liquefy everything that is gas; and you can get rid of most of the neon... but the Helium won't go away! Well, your cryo is made of helium, so you can't get it much colder than that; and something about the Second Law of Thermodynamics is kicking in. No worries - turn on the optical laser system and start plucking any of the remaining atoms out of the vacuum chamber

...and wondering further about the noble gases whether there would be any advantage in putting a second skin outside the vacuum vessel holding hydrogen under pressure such that a pressure gradient exists across the first layer towards the centre of the device, so that hydrogen would "flow" into all the interstitial spaces so that when the device is open to atmosphere it wont be contaiminated with non-ionic gases.

...and further whether the gas under pressure between the two skins could be deuterium, such that only deutrium would leak into the chamber when it is operating.

Does description that make any sense?
Would it work physically? (2)
Would there be any value in it? (3)

I suppose there will be plenty of generated Helium to deal with anyway for any alphas that are neutralised within the main chamber. Even though that quote indicates a difference between the ionizability of hydrogen and helium, can the helium be ionized at all given a high enough electric field? (4)

[MSimon]

Turbo pumps.

[KitemanSA]

Turbo pumps.

That was stage 2ish of 5ish, keep up dude!

Oh, and by the way, steel is a lousy pressure chamber material. Its lattice (body centered cubic IIRC) has great big holes in it. You need a more densely packed crystal lattice (face centered cubic IIDC) like autenitic stainless or aluminum. There may be other issues, but wouldn't a few "skim coats" of vapor deposed Al make a lot of difference?

[MSimon]

Turbo pumps.

That was stage 2ish of 5ish, keep up dude! :wink:

Why 5 stages? I'd have a roots backing pump. A small turbo pump (pre compressor), and a large turbo sucking on the vacuum vessel. I've done the preliminary calcs.

It should be good enough to get you down to 1E-9 after bake out. Since the machine is running at about 1E-6 that gives you impurity gas (H2 or D2 and He mostly - everything else has higher compression ratios) of between .1% and 1%. Good enough? Considering H is fuel and He is waste - probably.

[chrismb]

I've done the preliminary calcs.

What fuel/power output did you use in your calculations?

[MSimon]

I've done the preliminary calcs.

What fuel/power output did you use in your calculations?

I'd have to look it up. It was for an experimental reactor so the maximum flow was on the order of 1 to 100 cc/sec STP.

The main thing I worked out was pressure ratios. A good turbopump optimized for H2 and He gives about a 3,000 to 1 compression ratio.

The backing pump gets you down to 1E-2 torr. Two turbo pumps in series (sized adjusted for actual volumes pumped) gets you down to 1E-9 torr.

And the size adjustment is: a 3,000 l/s pump could be followed by a 1 l/s pump. You would probably want to oversize it by a factor of 10. So say 10 l/s.

A production system will have to be custom probably. So I did not run those numbers. I wanted to use off the shelf parts.

[chrismb]

You do realise that 1cc/s at STP would require a 1E9 litre/s pump at 1E-9 torr, don't you?

[MSimon]

You do realise that 1cc/s at STP would require a 1E9 litre/s pump at 1E-9 torr, don't you?

Since we will be operating at about 1 E-6 torr that would be a 1 E6 l/s of pumping. About 300 3,000 l/s pumps. Say they take up a "box" 10" on a side. 100 sq inches. so 30,000 sq in total. About 14 ft on a side. Doable.

Of course it would be better to custom design a larger pump. But if you are in a hurry.....

[chrismb]

Since we will be operating at about 1 E-6 torr that would be a 1 E6 l/s of pumping. About 300 3,000 l/s pumps. Say they take up a "box" 10" on a side. 100 sq inches. so 30,000 sq in total. About 14 ft on a side. Doable.

OK. I can picture that. A few million quids worth of turbo pump lining the side of the chamber, sucking for all they're worth.

So if you need 1E6litres/s pumping capacity for some particular fusion power output, then you'd need twice that pumping capacity if you doubled your power output? Is that right, that the pumping requirement will be, to a first approximation, linear to fusion power? [Double the emitted alphas, double the helium out?]

[MSimon]

Since we will be operating at about 1 E-6 torr that would be a 1 E6 l/s of pumping. About 300 3,000 l/s pumps. Say they take up a "box" 10" on a side. 100 sq inches. so 30,000 sq in total. About 14 ft on a side. Doable.

OK. I can picture that. A few million quids worth of turbo pump lining the side of the chamber, sucking for all they're worth.

So if you need 1E6litres/s pumping capacity for some particular fusion power output, then you'd need twice that pumping capacity if you doubled your power output? Is that right, that the pumping requirement will be, to a first approximation, linear to fusion power? [Double the emitted alphas, double the helium out?]

Maybe/maybe not. I calculated flow rares at 10X the burn rate. (to maintain chamber pressure) Standard engineering rule of thumb to reduce a "problem" to relative insignificance. 5X might work too.

[chrismb]

Well, let's say it is 1:1 and you don't need any 'extra' pumping on top of just the reactor products.

If you're sucking up 3000l/s on a 1E-6 torr environment, then that'd be around 3E16 particles/sec. Is that right?

[MSimon]

Well, let's say it is 1:1 and you don't need any 'extra' pumping on top of just the reactor products.

If you're sucking up 3000l/s on a 1E-6 torr environment, then that'd be around 3E16 particles/sec. Is that right?

I'll take your word for it (often a dangerous proposition).

[chrismb]

Don't take my word for it! We need to know these things, don't we?

~1E22 particles per litre at STP. So 3000 litres is 3E25 particles at STP. So 3000litres/s at ~1E-9 atms (1E-6 torr) would be 3E25/1E9 = 3E16 particles/s.

Is that right?

[KitemanSA]

Might it be possible that a material like graphene is nearly transparent to ions at MeV but opaque to atoms at eV? If so, couldn't the fusion product be significantly constrained to a volume outside the main chamber and thus at higher pressure for easier pumping?

Not sure if this would work since there would still be a need to pump the central volume to very low pressure while NOT getting in the way of the fusion product/ direct conversion equipment. But it would be a matter of keeping it there against leakage, not fusion product.

May be simpler just to have massive pumps.

[mvanwink5]

I am not sure what dictates the vacuum requirements. If arc suppression is the need, then the chamber could be made larger. Also, for condensers at a power plant, there is an air removal section designed into the condenser box to facilitate removal of non-condensable gases. Because alphas can be directed, perhaps vacuum removal sections could be designed to create a zone where the pressure is higher, for easier removal. I guess my point is that there is more to vacuum design than first glance might suggest, and therefore costs could be greatly improved (or the reverse?).