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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

Electromigration Testing


Electromigration Testing | 27 June, 2004

I am a Quality Engineer which went from an Automotive QS9000 company manufacturing ceramic hybrid boards and now I am with a ISO PC mfg. company dealing with PCBA boards. I've been assigned a task in determining the contributing factor to high contamination levels (dendritic growth) found on our boards, 20 - 80 ug/in^2. The test,Ionic Contamination (Electromigration) Test , was performed on 5 PCBA boards by an outside lab using IPC-TM-650 2.6.14 Rev. A standards. My question is, what could be the root cause. Could it be the no-clean flux? If so, what type of flux should be used to reduce the contamination levels? As part of a containment action all boards will be washed. Please help. Thank you.

reply »


Electromigration Testing | 28 June, 2004

First, welcome to your new job.

Second, IPC-TM-650 2.6.14 is only loosely an �Ionic Contamination (Electromigration) Test�. * Electromigration Test, IPC TM 650 2.6.14, * Ion Chromatography for Ionic Cleanliness, IPC TM 650 2.3.28

Third, IPC-TM-650 is a series of test methods. [You can find them posted on ] It is not standard.

Further, your �20 - 80 ug/in^2� sounds like the result of a bulk solvent extract test, like Resistivity of Solvent Extract (ROSE) [Eurolanders call this Solvent Extract Conductivity (SEC)], rather than a Ion Chromatography (IC) Test, which usually talks about contamination level specific contaminants [ie, bromide, chloride, sulfates, etc]. If you know the type on contaminate, it would help focus your search.

Next, possible sources are: * Flux contamination from assembly process (typically a halide ion being present). * Inadequate cleaning after assembly especially under components. * Contamination on component body (handling for example). * Contamination on component leads * Contamination from inside the component that may be leaching out along the leads. * Board fabrication process

As you say, your NC flux could be the source of the problem. If it is, you are applying too much flux. Talk to your supplier for recommendations on proper control. This assumes that we�re talking about wave or hand soldering [SMT flux residues are pretty much well controlled] [In parallel with this, consider using the same IPC-TM-650 2.6.14 to test on inbound bare circuit boards.]

If your hand soldering people are using squirt bottles filled with wave soldering NC flux; shut-down the line, throw all the bottles in the dumpster, and plan your product recall. When using squirt bottles, the raw flux spreads across the board, under components, etc. It�s uncontrolled. The heat of the soldering iron makes it spread faster / further. * When using water washable fluxes, squirting flux �all over the board� is reasonable from a contamination control stand-point [not from an employee health / safety stand-point], because the board is cleaned after hand soldering. * In NC hand soldering, the only portion of this flux that is safe to ship is the small portion that comes to soldering temperature. The remainder is BIG trouble.

Finally, cleaning NC residues is non-trivial. Done improperly, cleaning could increase your contamination levels, if your flux is not meant to be cleaned or you don�t use proper processes and materials. Cleaning water soluble flux residues should be the consideration in your case.

reply »


Electromigration Testing | 30 June, 2004

Thanks for the information.


I will search for and review the Ion Chromatography for Ionic Cleanliness test method. This definitely is the test I want to perform. Do you know what the IPC specification is using this test method? The outside lab gave us a spec. of 10.7ug/in^2 using the Electromigration IPC TM 650 2.6.14 method. Do you know if this applies to PCBA manufactures that use no-clean flux or does it not matter? For the electromigration test, we ran two groups together (supplier A and supplier B) and supplier A ,which uses no-clean flux, had the high contamination levels (between 20 - 80 ug/in^2) and supplier B contamination levels were between 6-7 ug/n^2. Thanks again.

reply »


Electromigration Testing | 1 July, 2004

There is no specification for �Ion Chromatography for Ionic Cleanliness�. The performance of your product in your customer use environment determines the level that you select. The good part for you is that have at least one pole that you can put in the sand, where you know it�s unacceptable.

We're a bit peeved with your outside laboratory. Based on what you have told us, they have their heads up their patoots [and it's costing you money.] The "10.7ug/in^2 using the Electromigration IPC TM 650 2.6.14 method" is an upper limit based on high solids rosin fluxes. It is from MIL-STD-2000, which is obsolete. Additionally, it was developed before many people on this forum were born, which has zero correlation with modern circuit layout and fabrication technology and assembly components, processes, and materials.

Bulk solvent tests are not meant to determine the necessary cleanliness level for product reliability. They are intended as a process control measure after you have determined a proper level and a correlation between the proper cleanliness level and the bulk tester. Further, bulk testers only measure ionic contamination. They do not measure nonionic contamination.

Rather than go through a long dissertation here, you can find much of the information on residue detection methods in IPC-HDBK-001, Section 8. This handbook is the companion to J-STD-001.

Back to solvent extract testing, and this doesn�t go into the vagaries of measurement variations between different brands of bulk testers or operators [gauge R&R] and setup issues. EMPF [ ] published a good paper ("An In-Depth Look At Ionic Cleanliness Testing," RR0013, August 1993) on testers a decade ago. It should still be available.

When you say, �For the electromigration test, we ran two groups together (supplier A and supplier B) and supplier A, which uses no-clean flux, had the high contamination levels (between 20 - 80 ug/in^2) and supplier B contamination levels were between 6-7 ug/n^2.� * Does that mean that you use two assembly contractors and one of them is doing something bad to the dog? * What kind of flux does supplier B use?

reply »



Electromigration Testing | 18 August, 2004


Sorry for the late response. I was trying to gather more info.

The group with low ionic contamination levels came from one of our other facilities. We want to outsource more of our product due to increasing demand. Both groups use the same no clean flux and manufacturer, Aim. The only differences between the two groups that I know so far are:

Low Ionic Group B Ultrasonic Spray Flux Solder Paste (Sn62) Enclosed Area (clean room type)

Higher Ionic Levels - Group A Solder Paste (Sn63) Reg. Spray Fluxer Lowered the flux amount and improved ionic levels to 16 - 17 ug/in^2 although affected yields thus, can't reduce amount much more.

- both groups passed the SIR test. - I am also told that the test method Trace Labs performed only applies to PCB boards and not PCBA? If so, would that also mean that the spec. does not apply? The ionic contamination test was performed in a chamber at 85C for 500 Hrs. I am trying to figure out if this ionic contamination issue is actually an issue as our facility ,which makes the group B boards , made it out to be or there is no or low reliability risk to our customers. I appreciate yur help

reply »


Electromigration Testing | 19 August, 2004

This may sound dopey, but it may be the answer to your question. ROSE testers from different suppliers produce different results.

Look at MIL-STD 2000. Based on Table VII, you find the following upper control limits (micrograms NaCl / sq. in.): * MIL-P-28809 Beckman / Markson: 10 * Omegameter: 14 * Ionograph: 20 * Ion Chaser [Zero-Ion]: 32

What that means is that for different test methods, there are varying degrees of sensitivity (ionograph is more sensitive than omegameter).

reply »



Electromigration Testing | 31 August, 2004

Hi Davef, if by using Omegameter 600 SMD I am reading 1,56 microgram NaCL/sqcm, what is the real value I am measuring? considering the 14 sensitivity ?

This message was posted via the Electronics Forum @

reply »


Electromigration Testing | 31 August, 2004

You�re asking the wrong person to give remedial math lessons. Anyhow, here we go: * [1.56 microgram NaCl/cm^2]*[2.54 cm/in]*[2.54 cm/in] = 10.06 microgram NaCl/in^2 * [Omegameter] / [MIL-P-28809 Beckman / Markson] = [10.06 microgram NaCl/in^2] / [GS' normaized reading] = [14] /[10] * [GS' normaized reading] = [10] * [10.06 microgram NaCl/in^2] / [14] = [7.19 microgram NaCl/in^2]

reply »



Electromigration Testing | 4 September, 2004

Many Thanks Dave for your patience (GBY).

I was just trying to give a chance to our connector Supllier to read few micrograms less of NaCl found with the connectors. But even if using a "bonus formula" they are out of specs.

Following it could be interesting for readers:

a) Sample of Raw Connectors Brand XXXX.. as received. TIC test by Omegameter 600 SMD. Temp of media = 43�C Test time = 10 min value = 36.2 ugr/sqinch of NaCl equivalent (5.6/sqcm) b) sample same lot# as above raw connectors but washed off before TIC test value = 6.1 ugr/sqinc (0.94/sqcm)

(IPC-TM-650 2.3.28 Iono Cromatography Dionex4000i tels that most of bad boys are Br and Cl)

We are luky that those connectors are soldered on PCB by WS Flux so the contaminants get away during DI water wash off. But very hard any way to meet our customer Specs (1.00 ugr/cm2 NaCl ). But what about if those connectors have to be sldered by a NC process ? The above little history confirm that Ion contamination can come from everywhere !!

Regards GS

This message was posted via the Electronics Forum @

reply »


Electromigration Testing | 6 September, 2004

You're correct GS. Your's is a good cautionary tale for all NC flux users. Unfortunately, it's too easy to assume that boards and components are clean, even if our NC process is flawless.

reply »



Electromigration Testing | 15 September, 2004

Do anyone of you know if are there any Standard (IPC, JEDEC, etc) telling how much can be the acceptable level of TIC (ugr/cm2 NaCl equivalent) allowed on surace of Connectors, Components,before to be assembled on PCB ?

Many thanks Rgds GS

This message was posted via the Electronics Forum @

reply »

ICT Total SMT line Provider

Capillary Underfill Dispensing