Solder is de-wetting in hot oil bath

A.

1. Your vendor is saying that the tin coating measures within specification on the coupon, but the coupon is not representative of the rest, so the rest is actually thicker? So the coupon doesn't de-wet, is that it?
2. The application of a hot oil reflow is a skill. I imagine that the vendor of the plating made this observation about the vendor of the hot oil?
3. Cyanide copper, cyanide cadmium, what kind of tin bath?
4. I would not eliminate the copper as the problem bath.
5. Have you done a x-section of the de-wetting, and bumps? Send us a photomicrograph and we will scan it, so we can all see the problem. Please provide the magnification of the shot. Send a photo of the de-wetted parts too!

A. Dear Gary,

De-wetting is a complicated picture; as you know perhaps already.

You can have different situations:

1. Dissolution of a wettable coating on a non-wettable substrate
   I think, the steel can be considered as a non-wettable coating. A prolonged exposure to the hot oil can cause a complete de-wetting.
   The question is if the product previously has been O.K. and is now showing this problem or if they have been bad always before.
2. Non-wettable areas on a wettable substrate
   The basic problems can be:
   - A too high organic content in the plating layers. (Here a carbon treatment can help)
   - A bad precleaning, remaining traces of oxides on the surface.
3. Formation of a thick intermetallic layer, causing de-wetting.
   In situation three, a shorter dipping time will improve the situation.

I understand that the hot oil dipping is for reflowing the finish. I have to say that you really need skills for that and that the use is questionable (But I think in your industry a spec is a spec and not so easy to change). However, if the parts previously have been O.K., there are several possibilities to find out what changed, such as:

• The oil used during stamping/forming/rolling the parts
• The composition of the electrolytes/cleaners, etc.
• The oil-temperature/dipping-time

I hope these remarks help you a bit. Please let us sometimes know if you solved the problem and how.

Kind regards,

A. If stripping the tin and replating usually solves the problem, I would say that, in the successful lots, you are not fusing tin and cadmium, which may lend credence to the overheating theory.

• Would a cross section of the part show distinct cadmium and fused tin layers?
• Would an analysis of the cadmium layer show a significant amount of tin (> than 1%)?
• Can you do this analysis for tin with Auger analysis in a Scanning Electron Microscope?

A. I suspect that the free tin has been completely consumed by formation of intermetallic with the cadmium. The intermetallic will not wet. If the larger coupons are used as a benchmark for the reflow, exposure time in the oil may be too long for the smaller parts.

A confirmation of this condition could be performed by Auger analysis as suggested by Tom. I think that a cross section analysis, if done properly, could provide an answer at lower cost. Another analytical alternative is direct microscopic examination of the reflow surface, before and after application of a chemical stripper that removes tin, but not the intermetallic. This latter technique shows a change is surface finish only if free tin has been removed. Either method would be less costly than Auger analysis.

A. That has been covered here in one form or another many times. Initially, the brightener is the bad guy. To get a bright finish, it plates a tiny amount of the organic material into the tin. While tiny, it can be enough to prevent wetting. Use a matte tin and your problem will go down significantly. Note that tin oxidizes over time and has the same effect. Dipping in a dilute stearic acid [on eBay or Amazon] /Xylene [on eBay or Amazon] solution significantly increases the solderability time.

A. Danny,

The answer to your question is yes, Zinc will readily migrate from the Brass substrate into the Tin coating thus leading to de-wetting of the deposit. Research conducted by ourselves using the Auger technique shows Zinc just nanometers from the surface within 24hrs, and increased rapidly when exposed to time / temperature eventually leading to a dark streaky deposit. This was on a 4um deposit straight onto brass. With regard to reworking a copper & tin deposit directly onto the existing coating will work sufficiently even with non active fluxes. We strongly advise all of our customers including Litton here in the UK that a barrier layer of Cu/Ni is a must. However please also bear in mind that a barrier layer will merely slow down the rate of diffusion and not totally eliminate it.

A. Danny,

You need a layer of either copper or nickel to stop the migration of the zinc from the brass. A good layer of nickel or copper will diminished and or eliminate de-wetting in your solder. For best results have the terminals stripped to the bare metal and replated.

Regards,

A. They're right about the migration. I would think that your plater should have made you aware of the issue before plating. A minimum of 50-100 microinches of Copper or Nickel is key in inhibiting migration and subsequent de-wetting problems.

A. Hi Kati,
Is it possible to change over to dull/matte as opposed to the existing semi-bright? What is the base material? Don't worry, be happy!

Hello,
Thank you for answer.
We didn't have possibility to replace our tin semi bright solution to matte, but I will do experiments with matte solution at supplier.
Our base material is Ag.
Thank you.

A. I am a retired chemist who worked for a company supplying chemicals to the PCB fabricators. Have seen too much de-wetting of solder in my time. All too often, de-wetting is caused by the Tin/Lead being plated on dirty or oxidized substrate metal. And sometimes this problem can be caused by simply leaving substrate in rinse water too long before Tin/Lead plating. Crucial that substrate be perfectly clean before Tin/Lead plating, or de-wetting is inevitable.