tungsten (?) filament tensile/fatigue strength

[mkstowegnv]

Hi, I would be very grateful to get your impressions and best guesses in response to the naive questions that follow.

I am creating a research apparatus which has two arms one above the other attached to a vertical axil so that when the axil turns CW and then CCW, a hot vertical filament that spans the 25 cm gap between the arms is waved back and forth (in humid room temperature air with little air movement). The filament is ~25 cm long, .05 mm diameter and current flowing through it keeps it at 450 to 490 C. Actually the arms wave and then stop (and the current goes off) before waving back, so that the filament is subject to vibration and bouncing, and to thermal cycling. It will not surprise any of you that pure platinum wire fails after a day of such abuse, when the object is to eventually to have many such devices running for years ideally (originally we thought we would be running at lower temperatures where the filaments would have stood up for longer).

It is relatively easy and not too expensive to get gold coated tungsten fine wire from Goodfellow which is what we will probably try next. (The gold coating to prevent corrosion and prevent any problems from a small amount of organic material that will be burning up on the surface of the filament.)

Can anyone suggest a better material?

I assume that it is best to get annealed tungsten wire (I haven't yet been able to get anyone at Goodfellow who can tell me whether or not their wire is annealed - how important is it that it is? and what would be the best way to anneal it if we have to do it ourselves?)

Depending on how prohibitive the minimums are, it might possible to get gold-coated Rhenium Tungsten alloy which has greater ductility and fatigue strength but less tensile strength at light bulb filament temperatures, but I don't know about 450-490 C and to be honest I am not sure which parameter is more important.

We have been keeping the filament loose so that it 'dances' a bit and is presumably subject to bending fatigue. Would it be better to keep it taught where it would be subject to less bending but more yanking from the vibration of the arms (mostly due to the stepper motor turning the axil which we have tried to minimize but can't eliminate completely)?

The diameter and length of the filament cannot be changed very much but it might be possible to put some kind of spring at either or both ends to keep it tense but give it some give.

Any thoughts?

Any suggestions are appreciated (even a pointer to a good book about lightbulb filament design).

Thanks so much for your help, Best wishes and happy holidays, Mark

 

[btrueblood]

If platinum is not working for you, tungsten is a step in the wrong direction, for two reasons. A. tungsten has poor high temp. oxidation resistance (that's why light bulbs are either evacuated and/or purged with inert gas). B. tungsten is a very brittle material with poor fatigue life. I doubt very much if a gold coating will give much corrosion resistance at that temp., and the gold's abrasion resistance for a cutting wire is pretty abysmal too. W/Re is a bit more ductile, but that's a pretty relative thing.

Why not use a Nichrome wire? About a 80/20 Ni/Cr (or higher Chrome content) should work about as well as anything else. So would any of the nickel-based superalloys (Inconel 600, 625). Even stainless steel would be an improvement. You will want to burn the carbon off after every cycle, to hopefully get rid of as much carbon as possible (it reacts with the chrome and will slowly deplete it at 900 F). You may be able to get it hotter (1300 F) and avoid this problem, people on this site will be able to tell you what alloys and temperatures to use to avoid carb/decarb issues (or if it's really possible).

Beyond that idea, you may want to try a silcon carbide fiber strand, with suitable oxide coating, or try a thin glass or other ceramic capillary tube surrounding a Nichrome filament.

Yes, applying some tension on the wire is probably better than letting it bounce around "loosely". Also, you should keep it "hot" until the filament is at rest again. The amount of tension you put on the wire is limited by the creep strength of the material; you could probably search for a few weeks in a local university library to find creep-vs.-temperature-vs.-time data for whatever alloy you pick, or just test different springs yourself (with or without the vibration loads) at your use temperature and find out how much creep will occur.

 

[unclesyd]

Based on personal experienced I would try annealed Tungsten wire.
The wire has to be annealed. There are some minimum bends, based on wire diameter, that have to adhered to. A spring will work in keeping the proper tension on the wire. We used and coil spring holding a spool of wire.
The wire will have a finite life.

 

[cocobolorose]

The tungsten wire will have a hard drawn, fibrous microstructure at first, and will be pretty flexible, otherwise it couldn't be coiled and re-coiled to make light bulb filaments. Recrystallization occurs at about 1200C, and then it will be very brittle at RT. The major benefit to using rheniated, thoriated or potassium-doped wire is for higher temperature applications than what you're interested in.
Tungsten begins to react with air at ~500C, forming a volatile surface oxide, so after a while at this temp you'll get a significant loss of material, so you're borderline here. However, gold coating to reduce oxidation might just do the trick. Good luck

 

[EdStainless]

You are at a rather modest temp. I would look at Nichrome or one of the similar Fe based high Cr alloys.
After all, these materials hold up as heating elements working in air for decades.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[mkstowegnv]

Thank you for your very helpful replies. We are going to try two types of filament to see which holds up the longest - the gold coated tungsten from Goodfellows, and '35N LT' superalloy from the kind folks at Fort Wayne Metals. Unclesyd's experience suggests making sure that the tungsten is annealed. The technical person at Goodfellows in the UK insists that tungsten cannot be annealed but that their's is "stress-relieved". Do you think this is effectively the same thing? Should we try to anneal the wire oneselves (how)?

Thanks again for your advice. Best wishes, Mark

 

[btrueblood]

Mark,

Your supplier is correct, you can't "anneal" tungsten in the classical sense, as heating the wire to recrystallization temperatures will result in a more brittle product. Also, you won't be able to do much heat treating of a gold-plated tungsten filament, since the gold will diffuse and/or evaporate long before the proper temperatures are reached.

Stress relieving creates some of the same effects as annealing (reduction of winding/drawing stresses); you can, however, end up with a more brittle material at RT, depending on what temperature (how close to the recrystallization temperature range, and how long the temp. is held) the stress relief is conducted at. Take a look at GTE/Sylvania and the GE websites.

http://www.sylvania.com/BusinessPro...ents/ElectronicsDisplayProducts/TungstenWire/

I still think you will be happier with a Nichrome wire, good luck!

 

[unclesyd]

I am still trying to get my data and so numbers on the exact temperatures we used to anneal Tungsten. My old company acts like they want to sell me the information that I developed for our process.

Here is a website that had some very good, almost everything you wanted to know, information.

http://aries.ucsd.edu/LIB/PROPS/ITER/AM01/AM01-0000.html

One wire supplier.

http://www.calfinewire.com/datasheets.asp?Metals=6&Alloys=metals/tds/tung99.htm