Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

SMTnet
»
Electronics Forum
»
Sn/Pb diffusion

Sn/Pb diffusion

Subscribe to Thread

I need to find out about tin/lead diffusion and how this cau... - Apr 29, 2002 by geoff_goring

See, I didn�t bite on you zinc diffusion trick that you trie... - Apr 29, 2002 by davef

Dave Many thanks for your very informative piece on SMT.c... - May 03, 2002 by geoff_goring

Interesting situation. Everything [ie, intermetallic comp... - May 03, 2002 by davef

As I thought..... This is probably why most journals and ... - May 15, 2002 by geoff_goring

Based on what you�ve said, comments are: * There is no reas... - May 17, 2002 by davef

Are we saying then that IMC can occur at the solder / Cu pad... - May 20, 2002 by geoff_goring

Q1: Are we saying then that IMC can occur at the solder / Cu... - May 20, 2002 by davef

geoff_goring

#19723

Sn/Pb diffusion | 29 April, 2002

I need to find out about tin/lead diffusion and how this causes solderability problems in device lead-outs.

This is not to be confused with an earlier thread of Zn/Pb diffusion - unfortunately this is was a typo.

reply »

davef

#19734

Sn/Pb diffusion | 29 April, 2002

See, I didn�t bite on you zinc diffusion trick that you tried to pull last week.

Understand, in this response, I am NOT trying to bust you. I�m trying to determine what you are looking to understand. [Some times I get so locked into talking a certain way, I do not connect like I should when people don�t take me down that familiar path.]

Diffusion. A material transport phenomena occurring in solids, and is caused by the continual physical motion of atoms from one position to another. This results in the flow of material from regions of high concentration to regions of low concentration.

So, when you talk diffusion and solderability, it gets confusing, because the diffusion usually occurs after the solder connection is made. Usually diffusion goes to strength of materials, reliability kind of talk. Whereas, solderability talks to the time when metal is liquid.

Taking a step backward, it�s perfectly reasonable to expect to see diffusion of metals on the leads on components [and solder pads on HASL boards] prior to assembly. So, is this diffusion that you're talking about affecting the solderability of components that you are receiving? [If so, please explain why you are not returning them to your supplier and getting replacements?]

Lead does not dissolve [or diffuse] in tin. [When you look at pictures of Sn/Pb section, the color of the tin is distinct from the lead.] We�re going to have to drill-down on this a little further also. What are you dissolving to tin or in lead?

Common metals that dissolve in tin are: * Antimony * Cadmium * Copper * Gold * Indium * Iron * Nickel * Platinum * Silver

Common metals that dissolve in lead are: * Bismuth * Gold * Indium * Platinum

Please describe: * What's going on? * What is the extent and breadth of this problem? * What are the materials involved? * Etc

reply »

geoff_goring

#19765

Sn/Pb diffusion | 3 May, 2002

Dave

Many thanks for your very informative piece on SMT.com

As we stand we have not yet 'seen' any problem. My issue is that we have ordered (LTB) a very large quantity of ASICS and our usage has taken a down-turn. We therefore need to keep the ones where the die has been packaged for AS LONG AS POSSIBLE. To this end we asked the manufacturer of the ASICS to predict a shelf life. Their response was that they will guarantee them for 1 year when, they say, Sn/Pb diffusion will start to have affected solderability. It was the QA manager of this supplier who originally quoted Zn/Pb diffusion and, foolishly believing he knew what he was talking about, I copied this info to the site. As you comment, it is Sn/Pb diffusion we need to consider. I presume there are, however, trace Zinc elements in the leads though??

Anyway, my main goal is to verify exactly what causes the solderability problem. Whether it is intermetallic reaction, poor wetting, crumbling surface or whatever?? Secondly, and as a follow-on to this, I need to verify for how long I can REALISTICALLY store these devices and how to 'age' them. I have a rough idea of how to this (IPC/EIA J-STD-002A). COupled to this I also have the problem of what to do with the bare dies that have been pre-ordered but not yet packaged?

Appreciate more advice.

Many thanks

Geoff A Goring IEng MIIE

reply »

davef

#19769

Sn/Pb diffusion | 3 May, 2002

Interesting situation.

Everything [ie, intermetallic compounds, solderability, etc] is speculative without knowing more about the materials of: * Leads. * Solderability protection.

... because suppliers use a fairly wide of materials. Your supplier's comments about zinc may be perfectly in-line, given their materials.

If you are truly going to "age" a component sample, then the aging method should produce either: * Same oxide conditions * Same intermetallic conditions ... OR * Combination of oxide/intermetallic conditions ...

... that the component sample would have after sitting in storage for X amount of time.

Unfortunately surface finish oxidation doesn't behave well in accelerated conditioning - typically you force/produce oxide thickness/species that are not present under normal storage conditions. Intermetallic growth is better behaved and there have been some good data published showing those relationships.

Consider the following conditioning methods: * Humidity tends to quickly degrade some [ie, tin, copper, etc] surface finishes, in terms oxide growth. Many use 85�C/85% RH as their humidity parameters for oxide growth - you get to pick the time. * Temperature highlights intermetallic growth problems. Investigators use either 125�C for 8 hours or 150�C for 16 hours as temperature parameters for intermetallic growth.

reply »

geoff_goring

#19931

Sn/Pb diffusion | 15 May, 2002

As I thought.....

This is probably why most journals and standards I have studied hold shy of even attempting to predict shelf life. They certainly hold off of any predictions when based on accelerated life testing too.

The component lead is a copper sub-straight with a coating (tinning) of the 85/15 Sn/Pb alloy ?(with possible Zn trace????). Given that most intermetallic growth or oxidisation occurs at room temperature and accelerates with increasing temp (using the equation 1 -e/kt ??) why can we not decelerate this diffusion or oxidisation by dramatically lowering the temp?

reply »

davef

#19969

Sn/Pb diffusion | 17 May, 2002

Based on what you�ve said, comments are: * There is no reason not to expect a trace of Zn in your 85/15 Sn/Pb solderability protection, but Sn/Cu IMC are more prevalent than Zn. * Intermetallic growth and oxidation are very different materials properties. For instance, IMC occur along the boundary between the copper pad and the solder and oxidation occurs on the surface of the solder. * No, �most intermetallic growth � occurs at room temperature� is incorrect. IMC are an artifact of a successful solder connection. You are correct that they keep growing at temperatures above 0�K, but they grow faster at soldering temperature than at room temperature. * Similarly metals oxidize faster at higher temperatures than lower temperatures. Although oxides are not requisite to soldering, practically they are an element of the process. * Sure at 0�K IMC stop growing and the rate of oxidation decreases, but that is not realistic.

It is not necessary to get in a froth about this. Our buyers very skilled at getting us components that are at least five years old. And we solder the blank of them as if we didn�t know the difference.

Consider: * Following the guidelines of J-020 for packaging these devices. * Storing the packaged devices in an adequate environment (20 to 30 C, 40 to 70% R.H.), maybe in an inert atmosphere. * Keeping the devices in a sealed package to eliminate air born contamination. * Setting-up a monitoring program to periodically assess the solderability of components.

See, running shelf life tests are not going to do dip, because these are usually comparative tests and you have nothing to compare the test against. Maybe spending the test money on N2 would be money better spent.

reply »

geoff_goring

#19980

Sn/Pb diffusion | 20 May, 2002

Are we saying then that IMC can occur at the solder / Cu pad join on the PCB when finished assembled? In our case we are being told that this process is starting before assembly. Is it fundamentally the same thing. i.e. will the Tin diffuse with the Cu over time anyway and in our case start at the join of the Sn/Pb coating and Cu sub-straight on the lead of the device?

Incidentally, if this is the case, what, in your opinion is the recognised process for making allowances for the IMC at the PCB Cu pad? Gold plating, conformal coating etc etc??

Only about 1 fifth of our dies have been packaged and do indeed fall into the J-020 category.

Are we also saying that it would be best to press on and get the other 4/5 of the stock (approx 150,000 devices, approx �0.5 million) packaged?

I appreciate your thoughts about not getting in a froth as this is life so to speak - but with half a million at stake I rather wanted to start to make as many informed and correct decisions as possible. 20 - 30 C 70% R.H is pretty normal - is it the inert atmosphere that will help most?

reply »

davef

#19988

Sn/Pb diffusion | 20 May, 2002

Q1: Are we saying then that IMC can occur at the solder / Cu pad join on the PCB when finished assembled? A1: Yes, but in a broader context I am saying that when someone solders correctly, they form an IMC.

Q2: Will the tin diffuse with the Cu over time anyway and in our case start at the join of the Sn/Pb coating and Cu sub-straight on the lead of the device? A2: Exactly, when your supplier fuses, dips, or what ever process to apply that solder to your component leads, they will form an IMC that will continue growing, regardless if you assemble the component or not. When the bare board is HASL, the copper pads on your bare board form an IMC in a separate process. In each of these cases, the IMC takes on a substantial growth spurt when the component is soldered to the board.

Q3: What, in your opinion, is the recognized process for making allowances for the IMC at the PCB Cu pad? A3: With SMT components, recognized process selects pads to have adequate strength, limits the time above solder liquidous temperature, and uses a fairly steep cool-down temperature ramp. With PTH components, there is no allowance, because the action of the wave washes the old IMC away from the solder connection [it becomes an element of your dross] and a new IMC is formed at the copper solder boundary.

Q4: Gold plating, etc etc?? A4: Beyond the issues of Q3, ENIG need to be plated properly. Issues are: * Nickel corrosion. * Thick gold. * Thin gold. * Black pad.

Search the fine SMTnet Archives for discussion on these topics.

Q4: Conformal coating etc etc?? A4: Unaware of conformal coating issues relating to IMC.

Q5: Are we also saying that it would be best to press on and get the other 4/5 of the stock (approx 150,000 devices, approx �0.5 million) packaged? A5: No, I would stay the course. I would NOT commit additional die to packaging, but obviously you need to be: * Reasonably sure that your supplier could package the nonpackaged devices to prevent deterioration. * Confident with your market prospects.

Q6: 20 - 30 C 70% R.H is pretty normal - is it the inert atmosphere that will help most? A6: No. The packaging according to J020 is your first line of defense. Maybe a thicker, rather than a thinner solder coating on the leads is either more important or at worst slightly less important. Finally, I�m suggesting that low temperature storage to limit IMC growth probably has limited benefits and that you probably would be better advised to the spend the money on low temperature store on other remedies, like nitrogen storage.

reply »