Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

Dark Pad / Gold Embrittlement?


Dark Pad / Gold Embrittlement? | 26 July, 2000

I have read many of the past threads on both topics and I still am not convinced I know the answer to the soldering problem we are encountering. I am hopping someone can help. I have a dense SMT board we are building and have been building for some time. On a recent batch we started experiencing an extremely high ICT failure in the area of a ceramic BGA. Upon close examination we found the component was not bonding to the pads. We went into panic mode on reinsuring our profiles and paste and could not find any problems or improvement in the bonding of the components. In doing some further evaluations we found that we can pretty much pull any part off the board leaving the pads and fillets on the board. I first suspected Gold embrittlement. In researching Gold embrittlement on SMTnet I discovered �Dark Pad. A condition many of you are familiar with. With limited information in hand I sent batches of boards out for x-ray fluorescence testing and cleanliness testing. Here�s what I found Lab 1 ( big name independent testing lab) came back and said the gold was as thick as 45u in. (spec of course is 3-6) Lab 2 (smaller but very well equipped independent) said the gold was as thick as 16u in. Lab 3 (board fab�s lab) said it was 3-6 u in. Turns out big lab didn�t have the capabilities to test gold over nickel without special filters and recalled their gold results but did report contamination. Lab 2 stands behind their results, also found contamination. Lab 3 says contamination in cleaning caused the problem and a re-clean would solve the problem. The board fab re-cleaned the boards and returned them. We built a sample to only find we had the same problem. Now to the question. If gold over nickel is self limiting how do you get 16u in of gold, has anyone else seen this much gold before or is lab 2 tests flawed? Secondly if its not too much or too little gold and its not contamination what is it? Thanks for your time.

reply »

Steve Schrader


Re: Dark Pad / Gold Embrittlement? | 26 July, 2000


We've suffered through similar frustrations. One thing we found that might be of some help is uur supplier, unbeknownst to us, was outsourcing their immersion gold. Their supplier apparently lost control of their process for a period of a few months. As best as I can tell it was for the weeks of of 52/99 - 7/00. I've pasted below one of the email responses from our supplier:

"Once the boards in question were received back, it was quite obvious that the boards had a black pad issue. You are correct, it did take additional time to determine "why" it happened. The returned boards were sent to [our supplier] for SEM analysis. It was noted on the report that the boards did in fact have issues with the nickel plating (manifesting in "black pad"). I believe this to be isolated to the date codes from 5299 to 0700 as the supplier has not seen issues other then this time frame, and has review both bath make-up and controls."

They said something about Phosphorous being the cause, but I lost the details, maybe Dave F. can shed some light on this.

We couldn't live with the risks of black pad disease (we saw it from 2 vendors), so we switched to Immersion Tin and have not had any problems so far (in 3 months).

Our board vendor claimed that their immersion gold supplier was the largest in Southern CA and lots of board shops outsource their immersion gold to them. If that's the case, lots of people have had black pad problems from the date codes listed above.

Years ago, we had a vendor "accidentally" run our boards down their gold plate line instead of their immersion gold line and we ended up with 30+ u in. of gold. But they 'fessed up to it when we sent them back and they checked the thickness.

Steve Schrader

reply »


Re: Dark Pad / Gold Embrittlement? | 27 July, 2000

Hi Doug: Responding to your questions: � If gold over nickel is self limiting ... You didn�t say whether you gold plating was electroless or immersion. Electroless is not is self limiting. It can plate up to 5 thou. Expect immersion plated gold to self limit at 10 to 12 uin. � If its not too much or too little gold and its not contamination what is it? ... The clue for me is that you have a gold solder connection to the pad and a poor solder connection to all the components. That says: - Ni is clean and not badly oxidized - Au is greater than 3 uin thick and keeping the Ni from oxidizing - Solder flux is active and capable of removing oxides and residues on the pads - Solder alloys with the Au and Ni - Solder connections with various components are poor

... there�s gotta be story there!!!

It would be nice to have sumathat fancy-smancy analysis of the post-reflow solder material. But we don't get that stuff, do we? Anywho, I'd bag all "black pad" conversations and focus on your initial thought ... excess gold alloying with your solder is making unreliable solder connections.

Good luck

reply »

Software for SMT

ii-feed SMT Intelligent Feeder