Old germanium transistors, A health risk?

Am 23.10.20 um 23:40 schrieb Jeff Liebermann:

BeO is white, not pink, and as a sintered powder it would be white ceramic also. And on the box of the 20 BFQ139 transistors that I yesterday took a photo from, the warning was much larger in print than the number and type of transistors.

cheers, Gerhard

Vapour pressure not listed, of course.

A year isn't very long, and I doubt that you measured the bond line before disassembly, so although you're no doubt correct that there were no voids, that's different from being uniform.

Fancy modern TIMs can have problems with washout, where under thermal cycling fine particles get washed out of the stacks of filler flakes over time, dramatically increasing the thermal resistance.

Voids also form under thermal cycling, which is common in servers as cores get switched on and off. Thermal expansion pushes the TIM out when it's heated, but there's nothing to pull it back in when the temperature goes down again. High performance heat sinks have reservoirs of TIM, so it can flow in both directions without getting lost.

Cheers

Phil Hobbs

(Who used to work in the Packaging Research department at IBM Watson, and so got to help out with some of these sorts of things, and heard a lot more at talks and group meetings. Don't get me started on lead-free solder.)

Not necessarily, Cr-doped Al2O3 looks like that as well.

This is one thing. But I believe that the primary requirement for a general-purpose magnetron is "cheap". The second is most likely "cheap" as well. BeO technology is way too expensive.

Best regards, Piotr

I did an extensive search to try to find the vapour pressure of mineral oil and dimethicone.

Vapour pressure is a can of worms. It depends on the grade and application, and can vary from mineral oil used in high vacuum pumps to dimethicone used in women's makeup powder.

Just to show you I'm not lying, here are some of the searches I did:

I don't know what you mean by bond line. The mineral oil completely covers the cpu. You can see the shiny surface on both the cpu and the heat sink, and you can feel the slippery surface due to the mineral oil.

It is clear the mineral oil forms a very thin film that maintains contact between the cpu and the heat sink.

Please note I do not use the wimpy heat sinks supplied by Intel. I use a much better unit from Amazon, similar to this one:

Thermaltake CLP0556 CPU Cooler

This supplies considerable spring-loaded pressure to the heat sink. The pressure is uniform from one heat sink to the next, and the mineral oil is squeezed to a very thin film.

The distribution was complete and uniform.

I do not know what you mean by TIM. The heat sinks expand and contract under thermal cycling. Mineral oil is an incompressible liquid, and it maintains a very thin film that remains in contact with the cpu and heat sink despite the expansion and contraction.

You said earlier that dimethicone was the product of choice. It is impossible to get in the small quantities need for heatsinks. I use Walmart grade mineral oil, and it works fine.

Note also that regular thermal paste dries out and needs to be replaced. You see this when removing the heat sink on computers that have been in service for a number of years.

The application of new paste depends on the skill of the user. Mineral oil has no such problems.

I can monitor any evaporation of the mineral oil by checking the cpu temperature with the watch -n 2 sensors in Ubuntu. The results using mineral oil are very satisfactory.

Beryllium toxicity is much more slow acting than that unless you do something incredibly stupid.

Trouble is as a fine dust it could end up in your lungs which is very bad. I'd treat all old power transistors as potentially dangerous inside. It is a powerful skin sensitiser for contact dermatitis too.

It is safe enough as large scale crystalline materials like emerald and aquamarine but as a fine ceramic dust it is really quite nasty.

Sharpening beryllium copper tools in oil refineries was seriously life shortening until they realised that beryllium in the dust was so toxic.

Various single crystals of pure beryllium metal around the world from the early days of X-ray crystallography have killed people in the past. The wounds from a scratch off it take almost forever to heal.

I stand corrected. Not all pink ceramic contains beryllium.

Hopefully you're not referring to the ceramic covers over the transistor die, which are probably alumina and do not contain beryllium because they do not require superior heat transfer. The area that needs the best heat transfer is between the semiconductor die and the screw or flange mounting base. For example, the "ceramic substrate" in this drawing might be a suitable candidate for BeO:

At least one ceramic major supplier uses pink (probably from chromium doping) as either their trademark or possibly as a means of recognizing that the ceramic contains beryllium. From 1998: Beryllium oxide is normally white. I have never seen it in pink, although it may have been produced that way for some special application. Many electrical insulators made by Heany Ceramics are pink. In fact, this is their trademark color. I think most of these are Alsimag ceramics though.

made our ceramics stand out from the rest.

BeO Still A Force In RF Power Transistor Packaging

BeO Takes The Heat In RF Transistor Packages

On Sat, 24 Oct 2020 12:07:57 -0700, Jeff Liebermann wrote: (...) Yet another beryllium hazard.

"Potential Beryllium Contamination due to Wear of Beryllium/Copper Finger Stock Used for Radio-frequency Shielding"

I used quite a bit of finger stock while building several RF screen rooms and for RF shielding various radios. I cut myself in the leg on a sharp edge during one build. I don't recall if it was from the copper sheathing, beryllium-copper finger stock, or something else like a copper plated nail. The cut healed normally. If I'd known that I was risking my life, I probably would have asked for hazard pay or a salary increase.

Drivel: Beryllium-copper finger stock is quite springy. We installed some on a radio burn-in room around the door frame, similar to what is shown in the article. Before I could stop them, someone slammed the door shut to see if everything fit. It fit, but we couldn't get the door open again by pulling on the door handle. I eventually opened the door by climbing through ceiling fan duct and kicking the door open from the inside.

Alumina is also white, unless chromium-doped (and that turns it pink). BeO is tagged with an odd color (I've seen purple, and pink) as a warning, not because that is its natural color.

This is a microwave oven tube example