BGA and Gold Boards

Has anyone had experience with placing BGA on gold plated boards. I thought I read a while ago there was a problem with embrittlment causing cracked joints. If anyone has experience with this your feedback would be great. Mike

reply »

| Has anyone had experience with placing BGA on gold plated boards. I thought I read a while ago there was a problem with embrittlment causing cracked joints. If anyone has experience with this your feedback would be great. | Mike

Mike, We and industry has put a stop to this phenomonon. Yes, gold is available as an alternative to HASL and OSP but in a new form as "flash" plated. Actually, it is not electroplated in the old sense where large amounts were depositied on solder surface areas. Now gold, and other metals, are electrolessly deposited in small amounts (thicknesses) so as not to negatively impact soldering, solderability, or solder joint acceptance or long term reliability. The problem used to be, as with gold traditionally, that, as in your wave solder pot, gold would "contaminate" the solder, and resulting joints, in large enough amounts to do the damage. Now, gold, and other metals, are depositied in thickness ranging from 3 to 7 milllionths of an inch - thus rendering a flat, solderable surface without the deleterious affects it once had. The problem is the relative newness of the technology (hence your question and that of many others). Therefore, there are but a few, but growing by demand, number of suppliers that know how to manage the processes. Part of the problem with flash is the netative interaction between processes associated with these very good finishes with solder masks. Again, this is a function of effective process management, or not. Another issues involves applying approximately 100 millionths of nickel to the copper surface to prevent copper leaching into and contaminating the gold. There are two ways to apply the "new" metal finishes. One is to place solder mask over an all gold board. The other is to use the immersion processes to "selectively plate" only the termination areas found after stripping the LPI solder mask away from those areas. Key suppliers now consist of the "top shops" like Hadco, Praegetzer, Multek, Collins, and a host of smaller shops jumping in as demand increases. These surface finishes are an excellent alternative most of us have been seeking. Hope this helps, Earl Moon (Guys, I just happened to see this one and had to respond as it's a big part of my new job).

reply »

| Has anyone had experience with placing BGA on gold plated boards. I thought I read a while ago there was a problem with embrittlment causing cracked joints. If anyone has experience with this your feedback would be great. | Mike There has been a problem with nickel/gold plating both on PCB's as well as BGA package substrates. The problem only occurs with electroless nickel/gold plating. In electroless nickel plating, phosphorus is added to the plating baths. When you solder to the nickel/gold pad, the phosphorus migrates to the surface of the nickel, forming a very brittle phosphorus/tin intermetallic between the nickel and your tin/lead solder. If you use electrolytic plating you should not have a problem, since phosphorus is not necessary to plate for nickel with this plating method. Many of the early BGA's used electroless nickel/gold plating on the substrates solder pads. Those suppliers who are aware of the problem have switched to electrolytic plating. You should check with your BGA component supplier as well as your PCB supplier. If you have any questions you can call me at 512-933-5783.

reply »

| | Has anyone had experience with placing BGA on gold plated boards. I thought I read a while ago there was a problem with embrittlment causing cracked joints. If anyone has experience with this your feedback would be great. | | Mike | There has been a problem with nickel/gold plating both on PCB's as well as BGA package substrates. The problem only occurs with electroless nickel/gold plating. In electroless nickel plating, phosphorus is added to the plating baths. When you solder to the nickel/gold pad, the phosphorus migrates to the surface of the nickel, forming a very brittle phosphorus/tin intermetallic between the nickel and your tin/lead solder. If you use electrolytic plating you should not have a problem, since phosphorus is not necessary to plate for nickel with this plating method. Many of the early BGA's used electroless nickel/gold plating on the substrates solder pads. Those suppliers who are aware of the problem have switched to electrolytic plating. You should check with your BGA component supplier as well as your PCB supplier. If you have any questions you can call me at 512-933-5783.

Terry, I emailed you on this subject regarding problems I was assured were eliminated. I too would appreciate further information regarding this disclosure as I am committed to it. Also, how do current electroplating processes deposit a small amount of gold insufficient enough not to cause solder contamination and the embrittlement problem Mike questioned - without removing it before reflow. H-P, in the 1970's and 80's "flash" plated gold surfaces with as little as currently specified electroless processes - about 7 millionths". As far as early BGA's go, I do agree there were the problems you describe. Some of us did the usual with military type devices (gold plated LCCC's, etc) and that was remove in a solder dip before soldering them to boards to prevent the well known embrittlement problem. Thanks, Earl Moon

reply »

| Has anyone had experience with placing BGA on gold plated boards. I thought I read a while ago there was a problem with embrittlment causing cracked joints. If anyone has experience with this your feedback would be great. | Mike Mike, Your concerns, fuhgetaboutem. We went through a rigid process qualification for BGA using OSP, HASL, and immersion gold surface finishes. Several thermal cycling tests were performed. 1000 at -55 to 125C. 2000 at 0 to 85 C. 1000 from 0 to 85C under 85% humidity etc. Cross sectional analysis with SEM to check on crack formation and propagation. Stud pull tests. The whole nine yards. Bottom line, HASL gave the most variability. OSP and Gold did quite well. All three had 100% yield. There were uBGAs, CBGAs, PBGAs, and .3mm qfps all on the same substrate. (A printing nightmare but done successfully!!) Bottom line, qualify your board providers process. Don't buy gold from just anyone. Look for process controls, a decent customer base, and, most importantly, references. You'll have a good measure of success. Embrittlement only occurs when the gold is too thick. Make sure the vendor as a means of measuring plating thickness and shoot for 3-5 microns of plating thickness over the nickel. Regards, Justin Medernach Flextronics International

reply »

| | Has anyone had experience with placing BGA on gold plated boards. I thought I read a while ago there was a problem with embrittlment causing cracked joints. If anyone has experience with this your feedback would be great. | | Mike | Mike, | Your concerns, fuhgetaboutem. We went through a rigid process qualification for BGA using OSP, HASL, and immersion gold surface finishes. Several thermal cycling tests were performed. 1000 at -55 to 125C. 2000 at 0 to 85 C. 1000 from 0 to 85C under 85% humidity etc. Cross sectional analysis with SEM to check on crack formation and propagation. Stud pull tests. The whole nine yards. Bottom line, HASL gave the most variability. OSP and Gold did quite well. All three had 100% yield. There were uBGAs, CBGAs, PBGAs, and .3mm qfps all on the same substrate. (A printing nightmare but done successfully!!) Bottom line, qualify your board providers process. Don't buy gold from just anyone. Look for process controls, a decent customer base, and, most importantly, references. You'll have a good measure of success. Embrittlement only occurs when the gold is too thick. Make sure the vendor as a means of measuring plating thickness and shoot for 3-5 microns of plating thickness over the nickel. | Regards, | Justin Medernach | Flextronics International

Justin "baby", Did you read Terry's message regarding his work and studies at Motorola - on this subject? I too have done many studies and, on the surface, agree and have a vested interest in seeing gold become the coating of choice. One thing I did not know, as Terry points out, is the Phosphorus problem associated with the electroless nickel process. The gold is fine. It's the nickel - I want to know more about. Of course, it is possible to electroplate the entire panel with nickel and gold, in the specified quantities, amounts, or thicknesses and not see the Phosphorus problem. This again is as stated by Terry herein and via email. I am very disturbed by this disclosure by a leader in BGA technology. I'm sure many others also will be - concerning the electroless coating types that are becoming so much in demand. Go get em cowboy, no - Superboy, Earl Moon

reply »