Thermal conductivity (ceramics)

Jan 02, 2013 71 Replies

There's been a design for a little transistor/diode temp. sensor floating around in my head.



In one version, there'll be a direct ground connection.



I wanted to have an electrically floating version too. I want to try a circuit on flex (kapton). Using the kapton as the insulator looks like a good idea. But the thermal conductivity 'stinks'. (~0.15 W/m-K) So driving home tonight I was thinking about using an Al2O3 washer instead. "What's the thermal cond. I said to myself, 100, 50 W/m-K?" Googleing Al2O3, thermal conductivity Coorstek, I found this,



formatting link
(I might have to print it out and pin it up.)



So for this application I need a 'decent' electrical insulator. Which looks to rule out all the carbide's. But what about the direct sintered SiC?



And then what about Aluminum and Silicon Nitrides? Has any one used those?



I'm looking for a little thin insulating washer, that will go from



77K to 400K. (Maybe 500?). A hole for a 2-56 screw.

George H.


AlN is great stuff, and not usually very expensive. Thermal conductivity is close to that of common aluminum alloys.

I've seen references to it decomposing on contact with water, into Al and ammonia or some such, so I left a slab in water for a week. Nothing happened. I think it's OK as regards outgassing in most apps.

I've seen stock TO220 insulators in AlN.

Hard anodized aluminum is a good insulator and a good thermal conductor, especially as a 1 mil thick layer is usually enough.

Another trick is to use a filler (microsphere, glass bead, fine sand, even monofilament) in epoxy as a spacer between an alum or copper heat spreader and a baseplate. Epoxy is OK as long as you keep it thin.

John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation

Neat, I'm already using TO220 transistors as temp sensors. (Prepacked hole who could resist?)

Interesting. Does anyone sell anodized Al washers?

No, I need a sensor that students can move around. I'm thinking I need some sort of spring (Belville washer?) in the stack-up.

George H.

ed text -

If it is very thin (and I think you can get .0005" thickness tapes -- certainly .001" because I use that, myself) then it may not be so bad. What would be the area? (I can't work out the W/K without that.)

Have you considered the usual suspects from Cotronics? Their

810 potting compound is about 0.67 W/m-K, I think. It's volume resistivity is on the order of 10^11 (ohm-cm.)

Jon

Probably. I've seen anodized aluminum TO3 and TO220 insulators.

John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation

Yeah, I ran the numbers with thin tape. I worry about it breaking with repeated use. Alumina has 100 times the conductivity, so the same area, but ten times thicker, (A 10 mil washer) still has 10 times the conductivity.

I'd like something moveable.

(Ge varnish is a classic low temp, thermal 'tie down'.)

George H.

Film? Adhesive on polyimide tapes is normally 1.5+ mils thick (per coated side, of course).

AlO2 99.8% washers 0.25 O.D. 0.08" I.D. 0.03" thick $3 apiece Polyimide " 0.25 O.D. 0.09" I.D. 0.005" thick 70 cents ea.

The AlO2 would have to be about 6x better conductivity to make up the increased thickness. Plus, it doesn't conform, so it won't make as good contact. (Un?) fortunately, it's about 60x better, so, ignoring contact issues, it will be 10x better overall, at

4.3x the cost. If the surface is not very flat, and the applied pressure fairly even, the ceramic will probably break.

Best regards, Spehro Pefhany

"it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com

ea.

"The Journey is the reward"

formatting link

eff.com

Cracking the ceramic is an issue. For the tape, as far as I can tell the glue on the tape is worse than the tape itself.

Any 'pinions on anodized Al?

George H.

formatting link

I only barely recall seeing 1/2 mil somewhere (didn't want it at the time.) I'm using the 1 mil variety. You are right, as the silicone adhesive is way too thick and can be a problem. I just imagine it possible to get the film without the adhesive. Someone must have it.

Pretty much agree. But without knowing exactly what George is about (I don't have a clear picture to be honest), I can't say if Cotronics stuff _could_ be useful or not. I've used it, though, and like it. And it is AlO2.

Jon

Yes, the silicon glue is as thick or thicker than the film itself. As Spehro immediately pointed out. As I also said, I suspect that the film by itself should be available from someone. Just not "as available" as the silicone glued tapes are. Could mechanically thin the adhesive, I suppose. I might try that here.

None from me except that I'd be worried about bumps and air acting as a thermal insulator.

Jon

Glass bead is what they use in chips for electrical isolation with heat transfer. Though I've used it in production, I never opened a package to see what the bead material looks like.

I like it, but don't know where to get hard anodized alumin(i)um washers in small quantities. I'm pretty sure my favorite OEM washer supplier could make them in moderate quantity, but I don't think that helps you.

DIY is possible (these guys just the other side of Rochester have all the stuff)

formatting link
but I bet your H&S folks would make things complicated.

Best regards, Spehro Pefhany

"it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com

A beryllium oxide washer is probably what you're looking for, and here's one source among many: http://www.aavid.com/product-group/interface/insulators/beryllium Google Beryllium Oxide washer for lots more

Just make sure to keep them from making contact with a cut, or any other internal contact.

RS Components, for one...

formatting link

...that's one thou/mil and is sold as thermal insulating film!

Cheers

Syd

There are tapes made for this. They are thermal paste coated.

One applies the tape (a pre-cut piece), and then removes the film separator, and applies the device.

The thermal coat melds into all the surface deformities. One ends up with a serviceable, replaceable 'gap tight' thermal mating.

I will look at the branding on the roll at work and post it.

Well I guess that depends on what's a 'moderate' quantity. Hundreds certainly and maybe 1,000's. But that's 'small potatoes' for washers.

But we have a metal 'stamping' supplier that could do the aluminum and then have them anodized by someone else...

Yuck... No, (but thanks) let someone else do the messy bits.

George H.

"The Journey is the reward"

formatting link

eff.com

Forgot to mention... The tape to which I refer is Aluminum.

That should change your image of what I offered.

Thanks John, I don't think I need the best, nor even the penultimate thermal conductor. (The MSDS is a bit off putting for student use.)

George H.

Hey, Seastrom will make Al washers with an anodized finish.

formatting link

George H.

"The Journey is the reward"

formatting link

eff.com

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required