sealing air evacuated glass blocks

[GasSavers_maximilian]

Air evacuated panels can obtain R 70 per inch, but they have to take the 14.7 psi of the atmosphere, so the internal support systems tend to rule out windows. If I just cared about visible light and solar gain, I was thinking that structural glass blocks are heavy enough to take the pressure, but I'd need a way to seal them permanently after the air is sucked out. The hole could be on the edges, so it wouldn't need to be pretty (or transparent). Caulk would allow air in slowly over time, so it needs to be something more substantial.

So far I have this: A metal disk with a rubber gasket on it surrounded by some sort of adhesive that won't allow air to osmose through it. The gasket forms a temporary seal, giving the adhesive time to cure once it's removed from the vacuum generator. Any ideas for such an adhesive? My knowledge of long term air impermeability is rather limited (metal, glass, ?). Thanks.

[Jay2TheRescue]

[GasSavers_maximilian]

Thanks for the suggestion. I emailed them to make sure it'd stop air over decades of exposure, since many plastics won't.

[GasSavers_maximilian]

Managing to find some data on permeability to gasses for various materials. Unfortunately, the tables tend to concentrate on the really excellent ones, so I can't find much (yet) on plastics or epoxies. It does give me info on glass, though, so I can predict the vacuum's life span in the block. Apparently enamel is quite good (used to coat steel in some tanks), so maybe I can just cover the seal with that, assuming it bonds to glass well. I can also make the hole extremely tiny, which will help whatever I go with.

[Nrggeek]

I think the main mode of heat transfer through a glass block would be the conduction through the glass (sides of the block), and the mortar around it, not the air. Without calculating it (heat transfer class was 10 years ago - I'm rusty), I doubt you will get much improvement over the ~R 1.9 by creating a (partial) vacuum.

[theholycow]

What about glass? It should be possible to work up enough heat to melt some glass to seal it.

[GasSavers_maximilian]

But how do I maintain a vacuum while I do it? I rejected that before, but now that I think, though, maybe it's not so impossible. All I need is an electric heater to melt some glass right above the hole. Throw a switch and wait until it has dripped on, then let it cool.

I'd need to bake out gasses absorbed into the glass blocks as well, but 150 C is all it'd take for that apparently (in a vacuum). My other approach is to go with glass panels with a support grid, siliconed shut and attached to a vacuum pump and switch, topping it off as needed. It's unclear if it'd look better or worse than glass block. Depend on the grid spacing, I guess.

[GasSavers_maximilian]

Originally Posted by Nrggeek

I think the main mode of heat transfer through a glass block would be the conduction through the glass (sides of the block), and the mortar around it, not the air. Without calculating it (heat transfer class was 10 years ago - I'm rusty), I doubt you will get much improvement over the ~R 1.9 by creating a (partial) vacuum.

Given the small size of the blocks, you're probably correct. That's another reason to explore the panel approach, since there'd be so much less edge involved. I should run the conduction numbers (need dimensions, though).

[GasSavers_maximilian]

Apparently a lot of glass block is partially evacuated (just by virtue of the super heated air during the manufacturing process). The installed end result is an R value of 1.96, almost certainly because of all the boundary leakage. I'm sure increasing the level of vacuum wouldn't effect it too much. Great call, Nrggeek!

A system hooked to a pump is a lot less demanding from a perfection standpoint, although making sure to resist all the forces involved will be interesting. I'm concerned about loads at the edges rather than on the face of the panels. Sounds like small scale prototype time.

[GasSavers_RoadWarrior]

IIRC the major mode of heat transfer in sealed glass spaces is actually convection currents, not the actual conductivity of the air. Kill the convection currents and you gain lots of R value.