Author Topic: EEVblog #186 - Gold embrittlement  (Read 4537 times)

0 Members and 1 Guest are viewing this topic.

Online coppiceTopic starter

  • Super Contributor
  • ***
  • Posts: 9527
  • Country: gb
EEVblog #186 - Gold embrittlement
« on: August 01, 2014, 10:45:35 am »
There is some good material in blog #186, but I found the recommendation of gold coated PCBs rather odd. Gold and solder are a horrible combination, because of gold embrittlement. Tiny amounts of gold can make normally malleable solder - both traditional and lead free solder - become brittle, and crack under thermal stress. This seems counter intuitive, as both the gold and solder are soft materials..... until you mix them.

If your PCBs are not subject to aggressive temperature cycling you will probably never have experienced PCB failures due to gold embrittlement. If you have worked heavily with power, and especially RF power, electronics you will know this as one of the major pitfalls for reliability. RFis worse because component makers love coating everything with gold, and experienced assemblers of RF power electronics go to considerable lengths to strip off all that gold before assembling their equipment.

Solder dissolves gold pretty well, so you can remove the gold from components, and leave a surface coated in solder, by dipping the parts in a solder bath. People typically use several solder baths, and dip the parts in sequence. When the first bath gets too polluted with gold you send its contents to a metal refiner, to extract the gold, and add a new bath of clean solder at the other end of the line.

A lot of cheap connectors have gold all over the contacts. You might notice the good ones only have gold on the contact end of the pins, and the end you solder to is supplied coated in something solder like (e.g. tin plated). I have seen a number of engineers complain about nasty connectors which aren't gold all over. Its actually the other way around.
 

Offline LabSpokane

  • Super Contributor
  • ***
  • Posts: 1899
  • Country: us
Re: EEVblog #186 - Gold embrittlement
« Reply #1 on: August 02, 2014, 06:02:24 am »
In my experience, the gold plating is usually requested as primarily to prevent oxidation of the PCB or contact between opening the sealed bag and assembly. In places such as China or Mexico, ambient air is rather corrosive, so the gold helps there.
 

Online coppiceTopic starter

  • Super Contributor
  • ***
  • Posts: 9527
  • Country: gb
Re: EEVblog #186 - Gold embrittlement
« Reply #2 on: August 02, 2014, 07:29:32 am »
I find people usually specify gold because its not that expensive, and sounds like its high quality. The electronics industry has a bad track record of trends being pushed by people with a poor understanding of metallurgy.
 

Offline Psi

  • Super Contributor
  • ***
  • Posts: 10234
  • Country: nz
Re: EEVblog #186 - Gold embrittlement
« Reply #3 on: August 02, 2014, 08:16:53 am »
but it looks so pretty...
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 29472
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: EEVblog #186 - Gold embrittlement
« Reply #4 on: August 02, 2014, 08:56:47 am »
I find people usually specify gold because its not that expensive, and sounds like its high quality. The electronics industry has a bad track record of trends being pushed by people with a poor understanding of metallurgy.
First I had heard of it...
Quick search on the topic:
http://www.semlab.com/goldembrittlementofsolderjoints.pdf
No doubt there is much more out there.
From the above pdf:

Case History #1
A failure analysis effort in September 2003 [5] raised some interesting questions
regarding the relationship of gold embrittlement to voiding is SMT solder joints. The
client reported that SMT components were literally falling off of the PCBA.
Examination of the fracture surfaces in the SEM (Fig. 3) revealed that the joints were
heavily voided and that the voids contained some entrapped soldering flux. Closer
examination revealed that there were thin Au-Sn intermetallic platelets widely distributed
over the fracture surface and in the surfaces of the voids. Most of the Au-Sn IMC in
these solder joints appeared to be in a very thin (~ 0.1-0.2 microns) plate morphology,
which could be expected to exaggerate the embrittlement effect for a given amount of
gold compared with coarser plates or needle morphologies. It was hypothesized that the
numerous thin Au-Sn IMC plates trapped solder paste volatiles during reflow resulting in
the voided condition of the solder joints. It was unclear what specific conditions lead to
the observed morphology of the IMC in these joints. Calculation of the gold
concentration in these joints based on the 33 microinches measured on unassembled
PWBs gave 1.65 wt%Au, which is significantly lower than the 3 wt% threshold for gold
embrittlement. This suggests that there may be embrittlement and other process issues
for lower concentrations of gold if the IMC segregates to certain areas of the solder joint.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 13382
  • Country: gb
Re: EEVblog #186 - Gold embrittlement
« Reply #5 on: August 02, 2014, 10:09:57 am »
VERY interesting. I am one of the many who believed that Gold plated PCB's are a sign that the manufacturer takes quality seriously. I was not aware of the issues with soldering to Gold. As was stated, I suspect that the kit I have seen with Gold plated tracks likely did not have gold on the pads before soldering ?

Quality transistors used to have gold plated leads but then 'way back in the day', transistors were often mounted in PCB sockets that also contained gold receptacles, so it made sense.

Thanks for this information. I have learnt something new today  :)

Aurora
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf