Hacker News Reader

Comment by knolax | Hacker News Reader

So what replaced this stuff in modern applications? Better couplers? Special epoxy?

A number of things. Metal seals, welded and vacuum brazed joints, a variety of good epoxies. Even good O-rings. And helium leak chasers, so you can actually find and correct leaks.

To get down below a certain pressure, you have to heat everything and "bake out" residual volatile materials. This precludes a wax from even being tested at the lowest pressures.

If your system is cryogenic you can absolutely still use wax and wax-like products. I sealed an aluminum can for something like 30 LHe cool downs with a mixture of glycerine and soap.

Future cryogenicists: glycerine+soap is amazing because you can wash it away with hot water before you finally take the part to get re-welded. Most epoxies/sealants can't be cleaned so easily, fouling any future welds.

Just adding one more thing, I had very good lock with Swagelok Ultra-Torr fittings, which are an o-ring seal that works on either metal or glass tubing. This is how I managed the metal to glass seal when needed. I didn't do much glasswork, though. My dad is a retired research chemist, and he did all of his own glasswork unless it was something really tricky.

If needed I could silver braze the Ultra-Torr into a larger object. It's just a bit of a puzzle to figure out what sizes of glass tubing mate with the various metal fitting sizes, because glass tubing sizing can be variable.

Back when I worked in a vacuum lab we used copper O-rings between tubes, and the edges of the tubes had a knife edge to cut into the copper.

It was pretty expensive - you’re only supposed to use each O-ring once. We had a big box of them that we’d occasionally send off to recycle.

We also figured out how to make direct glass-to-metal joints at some point. Widely used in mass-manufactured stuff, but I think you can also buy just a glass-to-metal transition or feedthrough by itself if you're doing custom glasswork.

https://www.youtube.com/watch?v=OWw32BLodjY

There are a few people who make most of the copper to borosilicate glass joints in the world. It's really hard to do, and they're quite good at it. Don't try to make your own as a cost-saving measure!

A major challenge is matching thermal expansion coefficients. Copper expands quite a lot, and borosilicate barely moves at all, so they make a joint to a glass that moves less than copper, but more than boro, then they add a little bit of another glass, and so on until they've gradually transitioned from the copper to the borosilicate. Those well-made joints can be cooled to liquid helium temperature (4 K) without issue.

Aren't there also a bunch of alloys like Kovar that have expansion pretty similar to borosilicate's?

Yes, but you need to be able to bond glass to the alloy (note the color shift in that housekeeper's seal), and nickel alloys like kovar tend to have a really high magnetic permeability, limiting their application in experiments that need highly controlled magnetic fields.

Indium wire is sometimes used to make gaskets, especially for cryogenic systems. You have to make a new gasket every time you open the seal, though. It's much more ductile than copper and handles vaguely like stiff clay. You can sort of sculpt it together when joining the seam.

Cu o-rings that flow as you tighten the seals.

Reply on news.ycombinator.com