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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Noclean soldering to gold

Russ

#8325

Noclean soldering to gold | 3 December, 1999

Hello all, I am in the process of converting our facility to noclean processes. So far I am experiencing troubles with soldering to gold finishes. The finish is an H.P. type1. The solder joints appear dull and grainy. They seem to be very strong and reliable but look somewhat poor at higher magnifications. I have tried numerous reasonable profiles at reflow and nothing seems to make any significant change. Does anyone out there have any suggestions?

Russ

reply »

#8326

Re: Noclean soldering to gold | 4 December, 1999

Hello Russ,

I believe that the dullness in color of the solder joint is related to the gold finish on the pads. I have had done alot of thru hole soldering of gold pins as well as SMT on gold finish and have seen the condition that you are talking about. Question: do you get the same result with an OA flux type? The no clean flux has a much lower acidity level but I have seen this condition using no clean, OA and RMA flux types. I belive it is caused by a small amount of the gold being absorbed in to the solder. I believe that this is common and I do not believe that it is a problem provided that the intergral strength of the solder joint is not degraded. Hope this helps Chris

reply »

Chris May

#8327

Re: Noclean soldering to gold | 6 December, 1999

Hi Russ,

I have a similar situation. I have been told that this can be caused by the nickel, over which the gold is plated, can result in suspect joints, embrittlement etc. Some joints can even be weak and broken without much force on some of my boards. I have double checked every aspect of the processes involve from paste life, temperatures, humidity, profile as per data sheets. The only thing it could be is the paste/flux to board finish relationship.

In the old days with HASL boards and active fluxes there didn't appear to be a mix and match scenario to go through.

I am using a Koki no clean paste at the moment over the immersion gold.

I am getting an aqueous cleaning system for the new year, so I can trial some more active fluxed paste.

I have also spoken to one of my bare board suppliers who is installing an immersion silver line/tank. This does not use nickel.(I believe)

I will be trialling this on a batch of boards.

Anyway, if your solder joints are only cosmetically affected, I wouldn't worry too much. If they survive pull testing your OK. If you really want shiny joints, try 2% silver paste.

Enough of my ramblings.It's been a long weekend. I'm sure you will hear from people more versed in the physics and chemistry of your problem than I am and I will read with interest what they have to say. It could also help me.

Good luck,

Regards,

Chris.

reply »

#8328

Re: Noclean soldering to gold | 6 December, 1999

Hi Russ, assuming that only your flux changed and the PCB-finish stayed the same it seems to me that the solderability of that finish is somehow poor. About Ni/Au it can be said that the solderconnection is made between Ni and Sn. The Au-finish is just an protection layer for the Ni and dissolves into the Sn just prior to the actual soldering between Ni and Sn what results in a more dull look of the joint. It�s also recommended to give the whole process more time over liquidous so that the Au can disperse uniformly into the joint and is not forming areas with higher concentration which result in weaker solder connections. ( Au should be 5-10� )

It is essential that the solderability of the Ni is good before applying the Au-coating otherwise poor wetting can occur with nocleans. Sometimes using higher activated flux improves things.

Results: Ni/Au leads to a more dull appearance of the joint which is a cosmetical issue as long as wetting and strenght are OK. Nocleans require good Ni solderability. We use Ni/Au with noclean with almost no problems. In two cases we encountered bad wetting of Ni.

My 2 cents Wolfgang

reply »

#8329

Re: Noclean soldering to gold | 6 December, 1999

Russ: Three things:

1 Wolfgang is 100% correctamundo about the slow dissolution rate of gold 2 Shear stength of the gold 63/37 solder connection decreases rapidly, as the gold content of the solder connection increases above 2%. 3 There are many reasons for dull, grainy solder connections. Check the archives for greater deatil.

My2�

Dave F

reply »

Russ

#8330

Re: Noclean soldering to gold | 6 December, 1999

Christopher, in response to your request concerning if I have the same issue using OA fluxes the answer would be no. This finish works well with all of our current Water Soluble fluxes. As a note, I have done some further testing and so far it seems that by shortening the "soak" and TAL while using a 2% silver allow that the joints now look good! I agree with everyone in reagrds that "looks don't matter" but unfortunately we are a C.M. and our customers like shiny. I have also found that nitrogen eliminates this problem alltogether but we don't have the gas either. I hope this also helps you with some of your issues.

Russ

reply »

#8331

Re: Noclean soldering to gold | 7 December, 1999

Hi Russ, there�s something I don�t understand about the measures you took. I guess that TAL means time above liquidous and shortening of that time can result in bad dissolution of the protective Au-layer. The necessary time for the Au to mix with your solder is reduced which leads to Au enriched zones above the IMC formed by Sn and Ni and stays there as a Au "saturated" area with weak physical strength. This may cause early failures due to "embrittlement" or whatever it is called. The shiny look you get speaks for this I think.

About the silver I read somewhere that the silver sits right there where the gold is supposed to be sitting after dissolution and may interfere with that process. I can�t judge if that�s true but it sticks somewhere in my mind and pops up now and than when discussing this problem. Maybe some soldergurus have somthing enlightning to say about that.

I am just wondering and may be awfully wrong Answers appriciated Wolfgang

reply »


RDR

#8332

Re: Noclean soldering to gold | 7 December, 1999

Wolfgang, I appreciate your continued feedback! The main reason that I shortened the Time Above Liquidous (TAL) was it appeared that I was burning off all of the flux from the paste. By shortening the TAL to 45 seconds and decreasing peak temp to 210 deg. C from 60 sec. and 220 deg. C I was left with a little liquid around the leads after reflow. With the land design of this particular product we can only apply a very small amount of paste (9.5mil wide by 72 mil length, 6 mil thick stencil)which I believe may be the real root of my problem. My preliminary non-scientific dental pick testing of the joint strength shows that the leads are strong. As far as the siver "sitting where the gold should be" I also have heard that and that is basically the reason that I tried it. It still sounds like I need to perform more reliability testing with this issue however. I will keep you posted if I find any earth shattering news. Thanks again for your support, I am new to the SMTNET thing and have been finding it very valuable due to people like yourself.

Russ

reply »

Dennis O'Donnell

#8333

Re: Noclean soldering to gold | 13 December, 1999

I can't figure out why everyone is plating everything with gold. According to the J Standard, gold plating should be removed before soldering to the surface. The gold, when left on the surface to be soldered will contaminate the solder joint and cause the joint to be brittle.

Remove the gold or tin the gold surface before soldering.

Dennis

reply »

#8334

Re: Noclean soldering to gold | 13 December, 1999

Hi Dennis The gold plating on a circuit board is present to protect the nickle plating from corrosion. The layer is so thin that as it goes into solution in the solder it constitutes only a small percentage of the total joint. It is commonly accepted that gold embrittlement will not occur with less than 3% gold. The gold that must be removed from component leads is much thicker and can result in too high a percentage in a joint. For reliability this plating must be removed, not just tinned. It is suggested to use a dedicated pot (or two)for this purpose and to monitor the gold level. John Thorup

reply »

Travson

#8335

Re: Noclean soldering to gold | 13 December, 1999

Russ,

Regarding your no clean soldering to gold. We have been doing that for about three years now. We have been using 62/36/2% silver solder paste. I have not done any thorough testing as to joint strength but the finish is shinier. It looks like we get better wetting also.

Travson

reply »

Fluid Dispense Pump Integration

 Reflow System