4
\$\begingroup\$

I have been running small batches of production over a year without a problem. Today, I found a problem that is never happened before. I have changed the PCB to gold plated in my recent batch and I see mold build up at a certain location 20 out of 500 units. I am unable to understand the root cause.

The production environment is humid (it is in Asia) but for mold to build up more things are needed especially sugar or some other type of food which I believe hard to come by in such environment.

What would be the cause of such build up? Is gold plate more prone to organic exposure?

update

I look under the microscope and now I think it is corrosion. Here is the photo. img1 img2

What process mistake would cause a problem like this?

\$\endgroup\$
  • \$\begingroup\$ It should be entirely covered in solder mask and there shouldn't be exposed gold. Maybe something to do with flux cleaning? Should the units be sealed against the humidity? \$\endgroup\$ – pjc50 Nov 12 '14 at 9:40
  • \$\begingroup\$ @pjc50 it is always the same location. I am assuming some manufacturing issue. It is on the legs of a 144 pin ic. \$\endgroup\$ – TGG Nov 12 '14 at 9:45
  • 3
    \$\begingroup\$ Are you absolutely sure it is organic material? Have you looked at it through a microscope? This may also be tin whiskers, which are likely to form under certain very specific conditions. \$\endgroup\$ – user36129 Nov 12 '14 at 9:50
  • \$\begingroup\$ @user36129 it is mold. Very obvious. \$\endgroup\$ – TGG Nov 12 '14 at 9:52
  • 1
    \$\begingroup\$ It's not 'thin', but 'Tin' as in Sn whiskers. Yes, they grow over time, after assembly. They can be a really big hazard because they're made from conductive material and cause shorts either between the pins directly or when they are disturbed, fly off and land on other contacts. Look at wikipedia for more info on the how and why. \$\endgroup\$ – user36129 Nov 12 '14 at 15:35
3
\$\begingroup\$

Pictures are too bad to determine what is this. On second one - whitest "thing" looks like crystal (some salt?), that yellow may be tin sulfate (salt), but I don't know. Maybe you can figure this out with microscope or take sample and try some chemical analysis.

Use some rubber gloves and be careful. If this is made in Asia - they could use solder with lead. Lead salts are very dangerous.


If this is not fungus - this is corrosion caused by acid (from flux that was not cleaned?). Corrosion is salt (result of metal + acid reaction).

You can try to clean everything in 99%+ isopropanol, it will dissolve salts and acids.


If this is fungus

You can eradicate them by putting your boards in electric oven. Don't use gas oven - burning gas produce water vapor so gas ovens are not "dry".

Do this if you are sure that is not corrosion, because corrosion will speed up and "eat metals" on your boards in high temperature very fast.

If parts on the board are not vulnerable to heat "bake" them in dry air:

140°C for 3 hours or 160°C for 2 hours

This should sterilize boards.

If you have parts that don't like heat (plastic parts, ADC converters, voltage references) - you can try:

  1. 99%+ isopropanol bath for 15 minutes
  2. Dry it (not in oven!!!, be careful - isopropanol is flammable and ignites below 400°C)
  3. "Bake" in dry air - 100°C (or a bit above water boiling point) for few hours

I would not recommend to put these boards in your products.

\$\endgroup\$
  • \$\begingroup\$ I inspected a few hundred boards in the factory. They are al starting to show signs of this corrosion. (I think it is corrosion since the legs of ICs start to loose their silver color and show the copper underneath the legs metal). For units that show no sign, do you think I can conform coat them and still ship? Coating will stop corrosion for sure. (Btw this is a non leaded process). \$\endgroup\$ – TGG Nov 13 '14 at 4:21
  • \$\begingroup\$ Coat with what? Some soldermask paint? Whatever you do - dry these boards first. \$\endgroup\$ – Kamil Nov 13 '14 at 5:05
  • \$\begingroup\$ Yes. I fly out yesterday. The issue is how to salvage the existing boards. These boards are produced in late September and the issue only caught during the final test. Can I soak the existing boards (without sign of corrosion) in isopropyl and dry them in electric Owen and conform coat them? (Since they are produced about 50 days ago I cannot clean them and expect problem solved since there could be some corrosion already that is not visible in microscope). Will this avoid any corrosion in the future? \$\endgroup\$ – TGG Nov 13 '14 at 5:08
  • \$\begingroup\$ Kamil another question. I have been running production for a year without a problem. Same flux same everything. I moved to gold plating in this batch. Would gold plating require a different flux or additional process. Factory doesn't clean as they say this is a no clean process. \$\endgroup\$ – TGG Nov 13 '14 at 5:11
  • \$\begingroup\$ Don't dry them from alcohol in electric oven. Alcohol will evaporate and ignite/explode when vapor contact with heating element in oven! Dry them in normal air. \$\endgroup\$ – Kamil Nov 13 '14 at 5:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.