Thermocouple K high temperature measurement.

[Onno]

Dear reader

We are measuring temperatures between 750 and 1200 C using a thermouple type K (wire = 0.2 mm). The thermocouple touches the bottom of a small sample cup made of Molybdenum (h= 11 mm and d= 7 mm).
We are discussing the possibility that at these high temperatures the metals from the thermocouple and Mo cup will diffuse into each other, changing the thermocouple alloy, giving a false temperature reading (easily -100 C). We experienced mechanical connecting, bonding of the couple and Mo cup several times. We did not see any sign of melting. The bond could be broken easily by hand.
The thermocouple K consists of a Chromel (90% Ni, 10% Cr) and a Alumel (95 % Ni) electrode.
The cup is radiation isolated using 2 or 3 radiation shields around and on top/bottom of the cup.

Ni, Cr and Mo are all components in stainless steel and I think therfore that diffusion in our set up may take place.

Thanks for reading.

 

[Onno]

I forgot to mention that it concerns a vacuum system: p < 1e-6 mbar.

Thanks for reading

 

[dvd]

I have had good luck referring this type of question to Omega, all of the technical rep's I have talked with were knowledgable.

The following link will also provide some info.

http://www.omega.com/temperature/Z/pdf/z021-032.pdf

 

[btrueblood]

Yes, at the high temperature listed the type K thermocouple can degrade pretty quickly (due to diffusion, and due to annealing); the way to tell is to change the thermocouple (using fresh, new thermocouple wire) and see if the temperature changes...obviously you will be working on a new cool-down and reheat cycle of the test article, so you need to use your judgement...or compare to an optical pyrometer reading. You can get an idea of the degradation by doing a calibration at lower temperatures before and after your testing, but this doesn't tell you what the error is at higher temperatures. You also are probably seeing some conduction loss down the leads, tending to cool the thermocouple junction; try routing the thermocouple on a tortuous path through the heat shields to help cut down on radiant loss path as well.

 

[btrueblood]

Oh, and another link to check out is

http://www.nanmac.com

 

[Transient1]

What about a high temperature IR thermometer:

http://www.pyrometer.com/PyroLaser.html

 

[Onno]

Dear reader

We tested a new thermocouple type K connection 3 times. The measured temperature just goes down at each high temperature heating step. 150 C difference between measurement #1 and #3.

About the pyrometer. We installed a Hamamatsu infrared photo diode above the cup with 760 nm pass filter. There is a linear correlation with the measured temperature.

Thanks

 

[Transient1]

Onno,

If I understand your post correctly, you have determined that both the thermocouple and pyrometer do not measure the same temperature after each cycle. Also, that there is a linearly decreasing trend to the measurements.

 

[btrueblood]

If something forces you to continue using a thermocouple, consider using a larger diameter wire, and/or a sheathed probe. The idea is to slow the effect of heat-related changes to the wire cross section (by reducing thermal gradients, and mostly by limiting oxidation attack), and hopefully reduce the run-to-run error to something more acceptable. Irregardless, consider investing in a big spool of TC wire.

The IR or photo-pyrometer approach works well, if you can calibrate the emissivity of the target, or manufacture a hohlraum cavity on the part to function as a known emissivity target.

Another approach that might work for you is to use a Platinum/Platinum-Rhodium thermocouple, or a platinum wire RTD; again, sheathing these probes will help.

 

[Onno]

Dear all

We are considering a tungsten rhenium thermocouple:type C. We found out that it is used in 2 other set ups inside our laboratory to measure temperatures in vacuum furnaces between 800 and 2000 C. It appears to be a stable type couple for use in high temperature vacuum furnaces. I am investigating small diameter wire (0.1 mm) and small diameter (1 mm) Aluminumoxide protective sheating (purity > 99%). We are tsudying conduction losses through the protective sheathing and wire. We are hopefull in coming to a usefull solution.Yesterday I removed the used type K thermocouple from the furnace where it has been subjected to temperatures up to 1200 C and inspected the inner side of the alumunimumoxide protective sheath we now use. There appears to be some metal deposited on the inner side of the Al2O3 tube. On the flat end of the tube that comes in contact with high temperature also metal deposition. The other flat end (ambient temperature) does not have metal deposition. There is no difference between the positive and negative thermocouple leg. Both show metal deposition inside the Al2O3 tube.
Maybe there is some metal evaporation from the type K metal that changes its emf properties neagtively to what we expereinece: a lower measured temperature each time the type K couple is subjected to temperatures up to 1200 C.

We are still investgating but the photodiode we now use seems to output the same voltage at each use and power setting of the vacuum furnace. The thermocouple is used for calibration purposes only. The diode will do contactless temperature sensing once it is calibrated.

That is the way to go we think a type C thermocouple. Type K does not seem to be able to handle the combination of high vacuum (<1e-5 mbar) and high temperature (Tmax = 1250 C)


Interesting to read is :

http://www.osti.gov/bridge/servlets/purl/7216061-sQRyHT/7216061.PDF

A Thermocouple Evaluation Model and Evaluation of
Chromel-Alumel Thermocouples for High-Temperature
Gas-Cooled Reactor Applications
bY
Bev. W. Washbum

Thanks to all.

 

[btrueblood]

We had a lot of trouble keeping the tungsten/rhenium wire stable over temperature cycles, the tungsten oxidizes with each exposure to air, and the oxide vaporizes when reheated in vacuum. The platinum-based thermocouples are a bit more reistant to the oxide cycle damage. But, YMMV. Good luck, you are looking at the right things, I think.

 

[Onno]

Dear btrueblood

That is interesting waht you have to say about the oxidizing of tungsten with exposure to air. The furnace we are working with is under high vacuum: p<1e-6 mbar. And we try to keep it clean from dust and fingerprints. It is good practice to bring up the furnace chamber back to atmospheric pressure using clean and dry nitrogen gas. I do not know what nitrogen may do when it adsorbs on a clean tungsten surface. The tip of the thermocouple seems most vulnerable to oxidizing because it has been subjected to the highest temperature which has produced a clean and high vacuum outgassed tungsten surface. I suppose the surface can be chemically very reactive. Well, I ordered the necessary parts to install a type C (tungsten/rhenium) thermocouple. We will just try it out.


Cheers

 

[hacksaw]

forget type K at 1200 C, they are just not stable

type R,S etc

50-100 C reading errors for type K at those temps is common. there are somewhat more stable thermocouples (type N) but not at 1250C

some benefit may be obtained by increasing wire size, but that in it self creates other problems

with vapor condensation going on, you may have to chemically clean the probes regularly

 

[btrueblood]

Tungsten forms an oxide film at ambient temp. and pressure in air, and the film thickness increases with time (although the rate drops exponentially). Every time the tungsten gets hot (over 1000 F) the oxides vaporize, leaving fresh, sparkly-clean tungsten...which gets re-oxidized on exposure to air.

We also had issues with the initial calibration curve for W/W-Re thermocouples, we could never get a calibration to match the published curves...never did figure out if that was an issue with the wire, or how we were forming the beads, or what exactly.

 

[Onno]

Dear reader

As a last thing to add to this thread I will show the measured results for the degradation of type K thermocouple as used in a vacuum furnace (P < 1e-6 mbar). It is quite destructive.
The protective sheath, Al2O3, at the tip of thermocouple that faces the thermocouple threads was coated after some temperature cycles with metal.We have not researched this matter further. We wil contine our heating experiments using a type C thermocouple.

PS: The current indicated in the graph is the electrical current delivered to the laser we use for heating.