Brazing question: Ceramic to Brass

Kent Brazing to ceramics is common in certain types of laser, although not brass. You may find someone who knows over on alt.lasers Martin.

Deutsch makes hermetic connectors and glass/metal seals. We used them extensively in E-beam equipment, bringing signals out of a hard vacuum to atmosphere. Probably a bit pricey.

snipped-for-privacy@deutsch.bellman.co.uk/

In article , Jim Stewart

One of the difficulties you would encounter is the difference in thermal expansion between the two materials - both during your operating temperature, and also at the actual braze temperature, which will be higher than several hundred deg F.

For this reason most ceramic- or glass-to-metal feedthroughs are engineered with a particular glass and a particular metal (typically Kovar, a good match) to prevent differential shrinkage, cracking, and leaks.

Depending on the design you *might* be able to do brass if you keep the length of the joint short. As others have said, often metal is vacuum deposited on the ceramic, and then a metal part is soldered or brazed to the metal.

I might wonder what your application is, because there are a

*large* variety of different types of glass to metal feedthroughs that are on the market, some pretty inexpensive. I would also say that at a few hundred deg F you might be able to get away with a silcone rubber o-ring seal for your feed-through.

They are available as multipin, high voltage, high current, coaxial, etc.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

When I worked in a laser factory, we used Indium seals.

Basically a special metal ring that goes between the ceramic and the tube. Pressure is applied and a good seal is achieved. Holds up to lots of G's and is working in space right now. I don't think we tested the seal past about

250 deg F. though.

dean s

jim rozen wrote: One of the difficulties you would encounter is the difference in thermal expansion between the two materials (clip) ^^^^^^^^^^^ One way to overcome a difference in coefficients of expansion is to resort to a "graded seal." (if my memory serves.) This consists of layers of materials such that the difference in expansion of any two adjacent layers is not too large. ^^^^^^^^^^^^^ (clip) I might wonder what your application is, because there are a *large* variety of different types of glass to metal feedthroughs that are on the market, some pretty inexpensive.(clip) ^^^^^^^^^^^ Every light bulb, vacuum tube and neon sign has glass to metal feedthroughs.

Tubes now have kovar feedthrough wires in the 'press' which is the portion of the tube where the wires are fed through a region where molten glass is pressed down over them.

In the olden days kovar alloy was not understood, so platinum wire was used instead. To keep the costs down, short (1/8 inch long) sections of Pt wire were butt-welded onto the copper wires, and embedded.

This feature can still be seen in very old vacuum tubes.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

That's because Indium melts a little above there. Certainly it's liquid by the 4-500F that the original poster mentioned.

All the tubes I've seen look coppery where the glass seals on the wire.

Tim

-- "That's for the courts to decide." - Homer Simpson Website @

Those are kovar. Copper will not form a leak-tight seal, given the processing temperatures for tube manufacture.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

I don't think this is true. When transistors cases were all metal except for the glass seal kovar was used. For glass diodes in the DO7 package either copperweld or dumet was used. I believe copper weld was developed by Sylvania for use in light bulbs. Copperweld was an inner core of iron with sheath of copper and the diodes were hermetic. Dumet was an alloy of iron, copper and some small amounts of something else. It had a copper color but was magnetic. It also made a hermetic seal. Some of the finer details escape me after 44 plus years.

Chuck Pilgrim

info on active brazing alloys that will wet ceramics but it's a pretty exotic process.

Unless you really need both high strength and high temperature it would be simpler to silver the ceramic and bond with silver bearing soft solder (standard practice with the ceramic component mounts in older Tektronix hardware).

There is a dreadful thermal expansion mismatch between brass and ceramic. In either case it would be essential to design the joint so that, when it cools down, the joint material in compression.

Jim

There may be other alloys that have a similar thermal contraction, but copper will not form a reliable press seal for vacuum tube manufacture.

Jim

================================================== please reply to: JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com ==================================================

When you say "gas-tight", are you talking below or above atmospheric pressure? Through hole or butt joint? Various high-vacuum technique books have procedures on how to make vacuum seals between metal and glass or metal and ceramic. Metal/ceramic seals involve either sputtering a metallic coating onto the surface or painting on a metallic glaze and firing it to give a surface that can be soldered to. There's also a mechanical joint, ala spark plug, that you could make up, assuming you're using the ceramic for electrical insulation. Look at how various high-power radio transmitting tubes are made, too. You might want to consider changing your rod out for something that has a lower expansion coefficient. Kovar would be ideal, it's made for the job but somewhat hard to find on a hardware store shelf. You can readily get tungsten rods as TIG electrodes in various sizes.

Stan