Sciencemadness Discussion Board

Electrode material that won't oxidize?

Quince - 29-4-2005 at 00:58

I need to figure out what material I can use for electrodes, that will last at possibly close to 1600*C in air -- without burning up in the air or becoming covered by a nonconductive layer through reaction with the oxygen and/or nitrogen under these conditions. I intend it to be in half mm thin rectangular form.

Platinum seems to fit the conditions, but even a foil a tenth of the thickness I need costs $100 from what I found. What about tungsten? It can withstand much higher temperature, but it burns in air... So what to use? Is there some reasonably priced alloy or something?

neutrino - 29-4-2005 at 02:17

Why does it need to be this thick? If I were you, I'd go down to my local jewler, get a 1.5g platinum wedding band for $35, and hammer it out to the needed size. Remember that Pt is ductile, so this isn't so hard. If you don’t feel like hammering, you might want to consider a 1 gram Pt bar, they’re sold on Ebay for ~$35.

Else, I'd try plating a piece of something (tungsten wire is cheap) with Pt, that might work.

If you need real corrosion resistance (on the order of HF and aqua regia), get a piece of PtIr alloy. Unfortunately, this isn’t so cheap because both of these metals are precious metals.

Quince - 29-4-2005 at 03:14

I just need something to withstand that temperature in AIR, without getting an oxidized or nitride layer thick enough to prevent electrical flow into the discharge (nor, on the other hand, eroding away). I know platinum doesn't react with oxygen under normal temperature, but I couldn't find data for such high temperatures.

vulture - 29-4-2005 at 11:03

What kind of conductance are you aiming at? Because anorganic nitrides are conductors, albeit bad ones. Otherwise check tantalum.

uber luminal - 29-4-2005 at 11:07

Quote:
Originally posted by Quince
electrodes, that will last at possibly close to 1600*C in air I intend it to be in half mm thin rectangular form.


i am currious what you would be doing that only gets to 1600 C, runs in air, and uses electrodes? (and AC or DC?)

Copper would work, if it was able to get rid of heat quickly enough. (by either calculating volume of the Cu, the thermal conductivity and time as the variable, or fixed time with volume as the variable).

Ta, Pt, Ir, Pd, Moly, Tungsten, Zr (I think) to name a few. There are Moly/Ni steels that can withstand oxidation at those temps. Getting most of these metals into small flat sections might be a problem though. Ta and Pt are easy to roll, but Ir, Pd, Mo, W are all very brittle metals, Zr you might be able to get that way.

Where do you live? I think I have some tungsten sheet that might fit your dimensions.

12AX7 - 29-4-2005 at 11:26

FYI, tungsten only has a high brittle/ductile transition temperature. Someone I know has bent thin TIG rods (for heaters for a vacuum metallization project, of all things interesting) by heating to 300C or so and bending with leather gloves.

There are a lot of tungsten alloys and composites with varying properties. Conductive ceramics may also suffice.

(FWIW, I think he's planning a device with a small, continuous glow discharge.. kind of an odd project, and yeah, smacks of "not showing us the whole picture", but I digress.)

Tim

neutrino - 29-4-2005 at 13:06

Zr is very hard, I doubt you could roll that. Ir wouldn't work, it forms volatile oxides at those temperatures. I guess that the alloy I mentioned earlier might not work then.

Quince - 29-4-2005 at 15:31

The application is the same as the sapphire thread I also have in this forum. I need to make a microhollow plasma cathode as an electron source for a larger glow discharge. It's a thin insulator sandwitched between two thin electrodes, with a small hole going through the whole thing. Pendulum electrons between the two conductors start a stable discharge.

I suppose since the heat is concentrated near the hole, I might just need to have a resistant plating on the electrode near around the hole. In my experiments with copper and mica, both erode very quickly. Alumina would be better, but sapphire is even better than that. I'm guessing pure tungsten will burn up; as for platinum, it's super expenisve, but if I could use a small amound just for the hole area, and bond it to copper that helps carry heat away, maybe that would work.

[Edited on 29-4-2005 by Quince]

neutrino - 29-4-2005 at 15:55

Pt isn't really that expensive in small amounts, only about $25/gram.

Quince - 29-4-2005 at 16:04

Can I weld it to copper?

uber luminal - 29-4-2005 at 16:11

Tungsten takes a good amount of heat before it starts to react with air in a short amt of time (days is short amt of time). I have used the W sheets for projects related to what your doing.

Zr is hard, but I can get it into sheets if you want... or at least I can.

Tim, your buddy may have bent TIG rods, which may or may not be pure W, but if you took the rods to a microscope you would find fractures galore. its not ductile at all, until you make it plastic at temps much higher, but then again, you could say the same about most metals.

Quince - 29-4-2005 at 20:39

So I guess Pt is the best suggestion this far. But again, how do I bond it to the power leads? I can't solder obviously. Can Pt be welded to Cu?

12AX7 - 30-4-2005 at 04:14

Welding will probably result in a brittle intermetallic. Silver solder or brass braze it. Standard jewelry practice, BTW.

Tim

vulture - 30-4-2005 at 06:00

Zr and W are not suitable for this, they'll simply burn at the temperature mentioned.

The melting point of copper is 1083C, so I can't believe that people are mentioning it here...:o

uber luminal - 30-4-2005 at 09:23

yes, the MP of copper is below the constrained temp. However its thermal properties are great for getting rid of heat, provided you cool it with something... like cold water flowing past it.

I use a copper electrodes in my arc melters... which reach temps MUCH higher (6000 C on the cu side). you just have to calculate for volume, time and how quickly the heat can be transfered to the moving liquid/gas.

I dont think you can braze Pt to copper, as I doubt it would wet the pt. Silver solder might be used, but as soon as heat is applied to said joint, you have 3 metals, expanding and cooling at completely different rates. The same can be said about welding it, however weldinging it properly it could be done. But... Since Cu has a lower MP than Pt, you could have had LME in the weld joint, and when the heat is later applied the Cux/Ptx alloy (at the weld), the Cu may melt within the Pt, not only hosing the weld, but liquid Cu may LME the Pt, through the heating stress. So in effect, it might hose the entire pieces of Pt by embrittlment. But since Pt still doesnt react with air... eh who cares, its not like your supporting anything with it.

uber luminal - 30-4-2005 at 09:29

an inexpensive alternative... could the experiment be done under inert gas?

neutrino - 30-4-2005 at 10:20

Why not just weld a small piece of Pt wire to the foil and connect the end of the wire to something else in a cooler place? You could even just cut a small strip of the foil and use that as a wire.

Quince - 30-4-2005 at 15:13

neutrino, a wire is not a good heatsink.

12AX7, I can't silver braze as the silver will melt.

If I made the complete electrodes from platinum, that'd be $500, so I guess so much for that idea.

uber luminal, you mentioned earlier molybdenum and tantalum. However, I read that both metals oxidize at fairly low temperature. You mention some steels. Can you be more specific? Searching for molybdenum/nickel steels doesn't show any oxidation data for high temperature, so I don't see how to choose.

I see tungsten/copper and molybdenum/copper used as plasma electrodes, but I don't know if that's just to save on the more expensive material, and if it's oxidation resistant.

Alumel would seem to be perfect, excet that no one sells it in sheet or foil form...

It says alumel is not resistant to a reducing atmosphere. If there's significant amount of water vapor in the plasma, would freed H ions possibly create reducing conditions?

Finally, what if I make them out of tungsten and then have it platinum plated? Or is plating likely to erode?

[Edited on 30-4-2005 by Quince]

Quince - 30-4-2005 at 15:20

Here's a picture to explain my setup. On the left are five microhollow cathodes -- dielectric sandwitched between planar electrodes. Thicknesses are about half a mm, as are the microhollow diameters. It operates in air. In my experimentation using copper and mica for the electrodes and insulator, correspondingly, both eroded very quickly.

Notice that extreme temperatures are only reached around the holes, so if I could somehow bond a more expensive resistant material in only those areas, it should be fine. But it doesn't seem possilbe. I don't know if even heavy plating would not erode in that area.

[Edited on 30-4-2005 by Quince]

electrodes.png - 5kB

12AX7 - 30-4-2005 at 17:04

Your picture is corrupted.

Silver braze does not mean silver, although it could be used. It is an alloy usually of silver and copper, with variable amounts of zinc and/or cadmium, and is available in melting points ranging from 900 to 1500°F.

Tim

Quince - 30-4-2005 at 17:31

Firefox, Irfanview, Photoshop, and Paint open the picture just fine.

I melted copper and silver and mixed them for brazing the heater elements for my mantle, and I can still melt it with my lighter.

Hopefully someone can answer my chemistry question from the other post. If plasma makes molecules dissociate into their component atoms, even if I used a carbon dioxide/water vapor atmosphere (burned propane) instead of air, I'd still have lots of ozone, and the H ions can act as reducers and wrek alumel? Or will the presence of C make the O more likely to react with the C than the electrode/other O? Basically I'm asking if there'll be an improvement over air, as of course only a noble gas will be perfect (but I do not have the option of a closed system, and I can't afford helium tank running off into the outside).

And the platinum plating, would that erode quickly or is that my solution?

[Edited on 1-5-2005 by Quince]

neutrino - 30-4-2005 at 17:36

I didn’t mean to use the wire as a heatsink, rather as a simple electrical contact that could take the heat.

12AX7 - 30-4-2005 at 21:29

If you are intent on alumel, use air, nothing is less reducing than pure oxygen radicals.

The average condition is what matters, whether there is an excess of H & C vs. O. This is easily set in a burner. Since molecular interaction is random, your electrodes may eventually wear out....but you were expecting that.

BTW have you looked at conductive ceramics yet? At all? No? Get some graphite and/or silicon carbide and/or a composite with something else like tungsten (or a PGM) or WC. IIRC, SiC is at least a semiconductor.

Tim

vulture - 1-5-2005 at 05:45

You might want to look into inconel, it's a nickel, molybdenum, cobalt (IIRC) alloy which has extremely good wear, abrasion and oxidation resistance. It's being used in space applications for reentry modules.

Won't be cheap, but I doubt it'll be more expensive than platinum. It's also being used to make crucibles.

[Edited on 1-5-2005 by vulture]

unionised - 1-5-2005 at 06:34

If this page
http://hcrosscompany.com/metals/inconel.htm
is correct then it is a couple of hundred degrees short in the melting point.

vulture - 1-5-2005 at 10:32

Quote:

Typical applications include furnace components, chemical and food processing, nuclear engineering and sparking electrodes.


Interesting. If you could get the inconel to protect your copper...

Quince - 1-5-2005 at 23:42

These conductive ceramics 12AX7 mentioned, how do you bond them to metal? Can they be made at home?

uber luminal - 2-5-2005 at 05:37

can they be made at home? and can they be fused to metal (Cu esp).

I say yes. But even I would have difficulty in doing so, even if I used all the resources at my job. (which could be replicated for home use...)

How about a mechanical fastening?

Other note on semi-conductive ceramics or semi-conductors in general. In my experiences, it is difficult to establish plasma from a semi-conductor at any distance. Ussualy you have to strike the electrodes on eachother, or make the semi-conductor molten 1st. (not all semics are the same of course though)

melting Si is a perfect example. its your basic semic, it doesnt conduct much at room temp. It esp. wont establish plasma. However when arc melting it, a pull rod may be used(a copper pole that will accept plasma, that is positioned such that the plasma may be driven over the materials intended to except the plasma. like creating an imposed plasma. The small amounts of heat and radiation eventualy get the surface of the poor thermal conductive materials near its plastic phase and then a plasma can be established.)

unionised - 4-5-2005 at 00:16

The talk of desktop fusion using lithium tantalate set me thinking. Ta is stable at high temps and forms a non-stoichiometric oxide TaO (2 and a bit)which is a conductor. It might do the job.
Unfortunately, it's not cheap.

zoomer - 5-5-2005 at 18:35

Would gold plated onto copper work? I know that both melt below the target temp, but with creative heat management you might get away with it. You have to flash the copper with nickel first, or the gold will migrate into the copper with use, but I don't think that would interfere with your objective. Au over Cu would be a lot cheaper than Pt over titanium, and a lot easier. Both of those are difficult in plating apps.

Z

12AX7 - 5-5-2005 at 18:54

Now that I think of it, you can probably plate a layer of silver or copper (could even make it thick so it conducts heat better :) ) onto a semiconductor.

Tantalum oxide conductivity sounds nice, but nonetheless, Ta is listed as 480C max in air. :(

Tim

Quince - 5-5-2005 at 21:59

Tungsten oxidizes above 400*C, but it would provide a good substrate for plating as it won't melt. Now just to figure out how to platinum-plate it.

Quince - 6-6-2005 at 22:55

Well, I've got tungsten of a good thickness, 0.8 mm, but how the hell do I cut and drill this? Carbide bits don't last very long (and when trying to drill the sapphire I'm using as insulator, they barely scratched it). How do I handle this? The tungsten also breaks easily. What a crappy metal.

[Edited on 7-6-2005 by Quince]

12AX7 - 7-6-2005 at 00:30

Try heating it above the brittle-ductile transition temperature. One website lists it as 450-900°F so a torch will do well.

Tim

Quince - 7-6-2005 at 00:42

Great idea!

What about the sapphire? I got some diamond-coated drill bits for the holes, but I don't know how to cut it. Chemical etching would be too slow.

BTW, looking at this mutual characteristics graph for the 4X150A:
http://www.webace.com.au/~electron/tubes/FIG23.jpg
Now, I'm guessing for lower Va the curves would be similar. And it looks like it's only linear for Vg1 > 0...
What ways to linearize? Modulating Vg2 as in ultralinear wouldn't work because it takes the tube away from the current sink mode I intend. Maybe the diode tube predistortion as in the long lost linear gain stage thread at diyaudio can work here? Can Hawksford EC be applied to a transconductance output stage?

[Edited on 7-6-2005 by Quince]

12AX7 - 7-6-2005 at 05:31

Just do plain old SE...

If you still refuse to get that Al2O3 ultrasonic drilled, just etch the damn thing.

Tim

Quince - 7-6-2005 at 15:17

Well it is single ended, in that there's a single tube driving the output. But the load is in series with the tube, not parallel to it. This current drive is needed because of the nature of the plasma as a load (constant voltage or possibly negative resistance), and for safety (if the plasma turns to an arc, that's like a short circuit and things will fry, but current drive prevents that).

I didn't find any local jeweller with the ultrasonic thingamajig to cut the crystal.

12AX7 - 7-6-2005 at 21:20

Quote:
Originally posted by Quince
Well it is single ended, in that there's a single tube driving the output. But the load is in series with the tube, not parallel to it.


And this is different from a resistively loaded SE stage how?

If you must, add NFB, CC or CV, whichever works best.

Tim

Quince - 7-6-2005 at 21:46

In SE the load is in parallel with the output stage-ground connection. Here the load is between the B+ rail and output stage, so the output stage acts as pass device to ground.

In voltage output class A, the output device is usually loaded by a constant current source. Here that doesn't make sense, since it's current output, so it needs a constant voltage source. So the B+ must be very well regulated, given that it goes in one side of the load. It's gotta be very low output impedance, the way ground is in the voltage output case. One thing I'm not clear is if it might be better NOT to have a capacitor after the regulator. I remember for example one discussion at tubecad where Broskie noticed an improvement after removing the post regulator capacitance in a particular application. Hmm.

[Edited on 8-6-2005 by Quince]

Quince - 7-6-2005 at 22:23

Regarding platinum plating tungsten:

I read that after dissolving in aqua regia, repeated evaporation and addition of more HCL will result in most of the HNO3 evaporating, leaving chloroplatinic acid, which can be used for plating.

I came across this patent:
http://free.patentfetcher.com/Patent-Fetcher.php?submit=Fetc...
This patent is for plating tantalum, but I'm thinking it might work here. Instead of purely plating, a heating step is involved which diffuses platinum into the surface of the metal being plated, forming an abrasion resistant layer. What do you guys think? How would I modify the procedure for tungsten (or would it work at all)? What can I substitute for the lead acetate in his plating solution (I don't have any source of that material)? One thing I notice is that the heating step is done in a vacuum. Would air be a problem? The only thing I have that can reach that temperature is a propane torch...

[Edited on 8-6-2005 by Quince]

[Edited on 8-6-2005 by Quince]

zoomer - 8-6-2005 at 05:29

A local lapidary club could help with cutting the sapphire, probably sooner and much cheaper than most jewelers.

Z

Madandcrazy - 11-6-2005 at 06:44

The tungsten only stable in vacuum ;) but resistant against temperatures.


[Edited on 11-6-2005 by Madandcrazy]

Quince - 30-7-2006 at 18:49

Hey, I managed to machine the tungsten and sapphire. It's a lot of work, but if you use diamond bits under water it's doable.

not_important - 30-7-2006 at 20:21

Answers to what you seek could likely be found at this

http://ecsmeet2.peerx-press.org/jsp/mas/reportTechProg.jsp?M...

where I noticed a paper to be presented there, that is on the very material I would have suggested

The oxidation of electroformed iridium and iridium-rhodium alloys in stagnant air at 1300 and 1400ºC is investigated and described.
http://ecsmeet2.peerx-press.org/ms_files/ecsmeet2/2006/04/27...


WSi2 an MoSi2 could possibly work, they could be formed as surface layers on the base metal.

Any chance you could use N2 instead of air?

Quince - 30-7-2006 at 20:23

No, the plasma has to be open to the air without any barrier.

not_important - 30-7-2006 at 23:48

Makes it real tugh then There are various metal + Si/B/N compounds that are conductive enough, but won't take an oxidising environment at that temperature. There are oxides that are conductive enough, but not when they are cold. And the size restrictions make it difficult to cool the electrodes enough that you could get away with exotics just at the active region (it's fairly common to cool electrodes enought that more common materials con be used, with just a limit amount of exotics at the areas really stressed).

Tungsten base with a coating of MoSi2 and/or WSi2, or platinum or platinum/iridium or similar alloys are pretty much it. The tungsten will need the coating else it will oxidise away in a few hours.

There are things that might work, except that I suspect the size restrictions might prevent using the. Nickel oxide with a few atom-percent of lithium is a semiconductor; I've made it with a room temperature conductivity around that of Nichrome, just by firing a mix of nickel and lithium carbonates to cone 12 (which well below your 1600 C), A mix of aluminum and chromium oxides, with a few to 30 atom-percent of Cr, are also conductive. But making and fabbing those to your size needs might be difficult.

[Edited on 31-7-2006 by not_important]

Quince - 30-7-2006 at 23:55

My plan remains the platinum/plating+bake to diffuse method from the patent.

Though I'm using tungsten as the substrate instead of tantalum, so I'm not sure if it will work as well.

[Edited on 31-7-2006 by Quince]

not_important - 31-7-2006 at 08:39

How well it will work will depend on several things. The first is how good is your vacuum, as the cooking starts as platinum black which will give no protection until diffused into a uniform layer. If there's much O2 around it will start oxidising the tungsten long before you reach a temperature where the platinum starts to diffuse. You might be able to get away with an argon atmosphere.

Another issue regards the differences in tantalum and tungsten alloys with platinum, something I can't answer. If the thermal expansion of the alloy is greatly different than the base metal, adhesion problems may arise.

Ta2O5 tends to stay put, losing oxygen around 1500 C to give a lower oxide, while WO3 tends to sublimate which could lead to surface erosion. The Ta-Pt surface may end up as Ta2O5-Pt, protecting the underlaying metal, while W-Pt would turn into a network of Pt as the W eroded away until the Pt detachs.

You might be able to get around potential problems by repeateding the plate and diffuse cycle, starting with a very long diffusion time and shortening it in each following cycle. The idea is to get a graded concentration of Pt, hopefully avoiding too great a differences in thermal properties, and ending up with a nearly 100% Pt surface.

Nixie - 10-2-2007 at 19:41

I'm not too happy with the platinum plating, in that it's not very hard (though the baking process wasn't carried out optimally due to limitations in my equipment). I'm considering rhodium plating (and a bonus is that rhodium can take 200 degrees more heat), but since rhodium is over twice the price of platinum, I don't want to do it if it won't work well. I've come across the following:
Quote:
Except in extremely clean conditions and then only with very special proprietary electrolytes, all rhodium is deposited with cracks. Copper and Silver migrate right through the cracks and show up as brown (copper) or gray to black (silver). Put nickel under the rhodium as a diffusion barrier.

Now, my substrate being tungsten, will this still be a problem? Is a baking procedure as in the patent still useful or even possible here? I'm worried that thermal expansion will make the plating break off from the substrate if there's no diffusion. On the other hand, if continued diffusion takes place during operation, eventually more and more tungsten will migrate and the plating will lose effectiveness as an oxygen barrier.