How to solder a gold wire inside a steel tube for 600¦C 3

[soptic]

I need to solder a 0.42 mm diameter gold wire inside and along a stainless steel tube OD/ID 1.5/0.45 mm and 20 mm long, to resist operating temperatures of 600°C. Any suggestion on how to do it?

 

[kenvlach]

By definition of the Am. Welding Society, temperatures above 450°C require brazing, not soldering. Two high temperature braze filler metals are *BNiCr (70Ni-16Cr-4B, solidus-liquidus range 1010-1066°C) and *BAgMn (85Ag-15Mn, 960-971°C). The first has better corrosion resistance, the second is lower melting. Each should be applied in a brazing furnace, with the SS having been prefluxed as well.

The special circumstances of small diameter and limited clearance warrant consideration of alternate methods. E.g., plate a very thin layer of nickel inside the SS tube (from a Wood’s Nickel solution, using an auxilliary anode wire), since it is much easier to bond to Ni than to SS. Also, plate a thin layer of Ag onto the Au wire. Then, insert the wire into the tube and heat above 962°C under vacuum or inert gas; the Ag will melt and braze the Au to the Ni.

*The prefix B indicates a Brazing filler metal.

 

[HIMES]

There are several precious metal braze alloys which offer good mechanical and chemical resistance at 600C, will readily wet clean bare SS (brazed in vacuum or hydrogen) and give you a good chance of filling the tight joint you have in your design. Some options are BAu-4 (82Au 18Ni - sol/liq 955C) and Palniro-7 -AMS4786 (70Au-22Ni-8Pd - liq 1037C). These alloys flow well, don't require the SS be plated, which will furhter reduce your gap, and have low risk of diluting/melting the gold wire. Be aware there is risk using hi Ni or Cu braze alloys near the Au melting temp, as both form a eutectic with gold and may reduce your wire thickness. I've seen excellent results from the BAu-4 brazed in vacuum. The only risk is the alloy may flow along the wire, away from the joint, if not stopped-off or something similar.

 

[metman]

I agree with HIMES and the BAu-4 in H2 or vacuum as a good choice which will wet the S.S. nicely. you will want to specify time at brazing temperature to a minimum like max of 2 minutes because of the potential dilution of your Gold wire as HIMES pointed out.

Jesus is THE life,
Leonard

 

[soptic]

Thank you ALL for your suggestions, the only problem is that my gold wire in reality is a fused silica optical fiber (diam. 0.36 mm) coated with 30 microns of gold to reach 0.42 mm in diameter. Is there any metal braze alloy which could be applied at 700 - 750 °C (max. temp. indicated by fiber manufacturer) to avoid damaging the gold coating? Thank you in advance for your help!

 

[HIMES]

The silver-copper and silver-copper-indium alloys are good choices below 800C. I think the silver-copper eutectic alloy BAg-8 is a good choice, but not sure how well it will wet bare SS, and you won't get near the strength of the other alloys mentioned before.

There is a major risk in your design due to the large CTE mismatch of fused silica and SS. The mismatch will open the gap between the wire and SS tube as T increases, which is helpful when you braze, but will cause hi stress at the braze joint everytime you go up to your 600C operational temperture. Will the part see many thermal cycles? Stress rupture and/or thermal fatique may be an issue if it does. One option to mitigate this risk is using a relatively thick, ductile metal interlayer between the silica and SS. This wili give you a compliant interlayer to accomodate the CTE mismatch induced stress. You could copper or (ductile) Ni plate the ID of a SS tube with a larger ID than your current design, or plate the wire, or both surfaces. BAg-8 would wet copper or Ni and gold plating well. A stress analysis based on the stress rupture of Cu or Ni at 600C should tell you if it will work. In any case, plating a small ID will be challenging - so maybe plating the wire, or both surfaces, is a better option. Anyone have other insight?

 

[tomwalz]

Try Lucas Milhaupt
5656 S. Pennsylvania Ave.
Cudahy, WI 53110
800 558-7006

or Prince & Izant
Cincinnati, OH 45624
800 634-0437

 

[HIMES]

One other option which may be simpler is to use an Fe-Co-Cr Stainless Invar tube instead of SS - if available. It will greatly reduce the CTE mismatch, but I think there is a phase change Invar goes through which dramtically raises CTE which may be below 600C. Also it may need to be Ni plated to increase wetting.

 

[soptic]

The precision tubes we are using are SS 1.4435 or 1.4301 availabe from medical needles manufacturers in many dimensions. I don't now if other materials are available, for example tubes in Invar or nickel as HIMES suggest. Further, the tubes or ferrules will have to support many thermal cycles between RT and 600°C. What about using a ceramic tube instead for CTE matching (alumina for example)?

 

[metman]

Hi purity Alumina will come a lot closer to the silica CTE than S.S. The following might be a solution:

1) Metalize the alumina tube using the Mo-Mn process (if you are unfamiliar with this, I can give you most of the details).

2) Green Nickel Oxide coat on Mo-Mn metalized surface fired in dry H2. This provides a pure Ni surface for brazing.

3) Pure Silver braze the Au plated silica to metallized tube at 1010C in dry H2.

This will give you matched CTE's with reasonably ductile interface.

Jesus is THE life,
Leonard

 

[kenvlach]

sensoptic,
Thanks for details of the application. (I was trying to visualize a corrosive situation where the SS needed grounding & provided mechanical protection)

Due to the thinness of the gold, brazing will create serious long-term alloying problems at 600°C, and most rapidly for the lower melting braze metals. I will mention that it is even necessary to plate a Ni barrier layer to prevent diffusion between Cu and Au in much lower temperature applications, e.g., circuit boards and jewelry.

With the new details and thermal mismatch issue, I suggest not doing any brazing at all; just gold plate the inside of the SS tube. Bare Au to bare Au will provide electrical contact, and w/o a rigid brazed connection, no CTE expansion problem. Note: to plate Au onto SS, first do a Wood’s Ni ‘strike’ onto the SS. You might be able to buy an already gold-plated pin (tube), they are used a lot in the aerospace industry.

 

[HIMES]

The use of an alumina tube would certainly remove the delta CTE risk, but makes brazing more difficult due the metallization metman described. An atlernative to metallization is using an active metal braze (ABA). ABA's, if you're not familiar, contain highly reactive metals like Ti or Al which reduce and react with the surfaces of non-metals and allow the braze alloy to wet and flow. One choice might be Silver-ABA which brazes at about 40C lower than pure silver. Such an alloy is available from Wesgo Metals, among others. Processing with ABA's is fairly typical and removes the need for laborious metallization.

 

[soptic]

Kenvlach,

If you suggest gold plating the inside of the tube and not doing brazing at all, how will the fiber be fixed to the tube?

 

[kenvlach]

Do you need to have a strong attachment or just make electrical contact?
To preserve your thin coating of gold long-term at 600 F (which I presume is essential to maintain your optic signal), it's advisable to not contact it with any other metal, which is why I suggested plating the tube ID with gold. For tighter contact, just put a little more gold on the ID.

What is the purpose of the SS tube? Could you use graphite (about the only conductor that can tolerate 600 C and not react with gold)?

 

[soptic]

Kenvlach,
We need a strong attachment, in fact, we don't need an electrical contact. The tube serves just as a ferrule to hold the fiber end in a 600°C environement without damage.

 

[metman]

sensoptic,

I did not read your requirements adequately. On rereading I see that max fiber temperature is 750C.

Is it possible to coat the the fiber a little thicker with Gold? If so, you could shrink fit the coated fiber inside an Invar or Kovar tube. If these materials are not available in tube form, you could make a tube by drilling and reaming round stock to size. If tube is custom made then you could stay with original fiber coated dimensions and size the tube hole for shrink fit using LN2 on the fiber and heat the tube.

Jesus is THE life,
Leonard

 

[HIMES]

Since you only have a 150C margin between processing and operation, brazing is tough as most alloys in that range (Cu-Ag-In) will dilute the Au. I like metman's idea of creating a mechanical joint by shrink fitting. If you also want a metallurgical bond, gold plate the ID of the tube then shrink fit the assembly. Gold diffusion bonds to itself at relatively low temperatures. Unless there is too much seperation of the wire and tube, the gold surfaces should bond at you operation temperature given enough time (probably less than 1 hour).

 

[metman]

In thread330-59538 Kalvin Kelly gave tjhe following response in which he states that thermal fits will relax when heated sufficiently. However HIMES has addressed that problem with his suggestion of Gold plate i.d. of tube and let diffusion take its course at operating temperature.

from thread330-59538:
Unfortunately, molten aluminum is the Universal Solvent for most metals. 309, and about anything else I can think of, will very quickly dissolve. It is hard even to find a ceramic that aluminum will not reduce.
Graphite is OK, used to line the pots where aluminum ore is electrically reduced to the metal.
Although it is academic here, thermal fits won't work at high temperature, they very quickly relax.
What other than aluminum will be present? That is, will there be any oxygen or oxidizing species present (CO2, H2O)?
If there is no partial pressure of oxygen at all then one might search a little further to find if some refractory metal might work.
James Kelly



Jesus is THE life,
Leonard

 

[kenvlach]

Ah, consensus toward not degrading the gold with brazing alloy.
Here is some thermodynamic data for braze elements alloying with the gold (& hence degrading the optical signal), and that further, the deduction that the more easily the braze, the more rapid the alloying with the gold:
The partial molar enthalpies of mixing for some elements in gold (at infinite dilution, i.e., into pure gold, in cal/gram atom):
Ag -4040
Al -30724*
Co +9590
Cu -2780
Fe +6090
Ga -22400
Mn -8500
Ni +5140
---Metallurgical Thermochemistry, 5th Edn., p.393-394 (1979).
A negative value indicates a large thermodynamic driving force for dissolution into gold.
These values also explain why Ni is used as a barrier layer between Au & Cu.
Indium is also a bad actor; Selected Values of the Thermodynamic Properties of Binary Alloys, p. 283 (1973), gives the partial molar enthalpy of mixing for gold (at infinite dilution, i.e., into pure indium, in cal/gram atom): -11200 cal/gram atom.

*agrees with James Kelly's comment in Thread330-59538 about liquid Al being a universal solvent [for metals].

 

[JimMetalsCeramics]

You may have to use a gold-based braze (try Wesgo Metals). Another option is to electroform the wire in place. Try contacting a goldsmithing house like Johnson Matthey, Heraeus or D. F. Goldsmith.