Gold Embrittlement in solder joints

Suggestion:
1. Lead-tin is normally matched by lead-tin. About 7 years ago, the computer manufacturers recommended that SIMMs shall have that kind of match to avoid gold plated or gold contacts at SIMMs. They have been working there for 7 years without any problem.
2. Perhaps, if you contact SIMMs manufacturers that may have some information about this since they practically faced this problem.
Try Computer Shopper listed SIMMs Manufacturers or over

Thanks for the tip. The NASA manual has a procedure for removing the gold by diluting it with solder and removing it. Very similar to pretinning. The question that has come up relates to the attachment of leads to the gold plated pads a on a die. Thei NASA procedues relate to spacecraft that will have to survive in a space environment for many years with assured survivability. I'm wondering how much of their requirements are 'overkill'? I believe that most bonding wires in IC's are spot welded so this isn't necessarily an issue? What if you are using a die that isn't in a case and you need to have good mechanical attachment to the bonding leads? This particular application will have to survive high G-loading after an extended shelf life so time is a factor. Rick Schmidt
ricksintuc@yahoo.com

Mining, Mechanical and Electronics engineer.

The removal of gold plating on components is covered in MIL spec, and is dealt with using a 'double dipping' method.
Darn frustrating too, you pay extra for the components, and then have to purchase and maintain equipment, and pay staff to remove it before the componet can be used.
That said, IPC 610 level three does not suggest that gold should not be used for high specification assemblies, and until an alternative finish with equal to or better solderability and shelf life properties than gold (sorry all those reps trying to sell tin, silver, and organic finishes), and as flat a surface, gold looks likely to stay with us.
Don't be scared of it though, gold joints are all around us, controlling lifts, computer systems, nuclear powerstations (now I'm scaring myself), and I know that boards I've made are actually floating around in satellites

Steve Bull
PCB and Quality Engineer

Suggestion: Visit
etc. for more info on somewhat related sites to finishes

I am a certified IPC-A-610C Class A Instructor. IPC/EIA J-STD - 001 C (the end item standard that 610C is inspecting to) states:

5.4.1 Gold Removal:
Gold shall be removed from solder terminals plated with 2.5 um [0.0984 mil] or greater gold thickness.

Before soldering to any gold plated surface, you must first remove the gold to conform to IPC-610. This is due to gold embrittlement of the solder connection. An acceptable method is to tin the surface and then remove the tinning with solder wick. Once is enough to leach out the gold.

Tin-lead solder normally used in electronics is not compatible with gold plating. Gold readily alloys with tin-lead, but when the gold concentration exceeds a few percent
the alloy becomes weak during thermal cycling. This phenomenon is called gold embrittlement. As long as the gold plating thickness is small (20-30 µin) this is not considered a problem. For thickness above 100 µin, indium based alloys rather than tin-lead solder are recommended
to ensure the integrity of solder joints.

-- Greg Robinson, CQE, Thoratec Corporation