Cleanliness Specifications -- New designs, component sensitivity and the elimination of rosin and solvent cleaning have forced new detection methods.

Historically, production floor cleanliness testing has been performed using equipment that utilizes the IPC test method 2.3.25, resistivity of solvent extract (ROSE). Developed during the era of rosin fluxing and solvent cleaning, such tools allowed for surface extraction of process residues and, by a standard conductivity probe, a general measure of the conductivity after exposure to the extraction solution for 10 to 15 minutes. Equipment allowed a wide-range of board sizes to be placed inside a large tank (extraction cell) of isopropyl alcohol (IPA) and deionized (DI) water. The 75 percent IPA and 25 percent DI water solution was then gently heated, sprayed in certain systems and circulated past the surface of the submerged assembly.

Detection Methods: Then Versus Now

Typically, 50,000 ml of solution would be sufficient for a board with a surface area of 100 sq. in. or less. The conductivity measurement would then be converted to an equivalence number, a sodium/chlorine (NaCl) standard then normalized by the surface area. Finally, the cleanliness value would be reported. At the time of the technique's inception, the industry had not experienced large amounts of electromigration shorting or failures due to electrical leakage.

New circuit designs, component sensitivity and elimination of rosin and solvent cleaning have greatly challenged the electronics industry. With today's flux technologies, one must understand the values reported by the tools mentioned previously. What is clean? What is dirty? A joint 1996 IPC and U.S. Navy study (RR0013) concluded that no common calculation factor could be established among the entire range of testing equipment available, nor could the cleanliness levels be correlated to electrical performance. The tools were labeled as process indicators for gross levels of change, but not denoted as a true measure of cleanliness.

To Clean or Not to Clean?

A few months ago, I performed a study that compares traditional rosin mildly activated (RMA) fluxes to new no-clean technology. My evaluation used IPC B24 boards-FR-4 boards with copper traces on one side with a hot air solder leveling (HASL) surface finish. A group of these coupons were cleaned in a saponified aqueous (DI water) inline cleaner and then wave soldered with one of the following fluxes: 1) a no-clean liquid flux (2.5 percent solid) that was not cleaned or 2) an RMA flux (25 percent solids) cleaned with a methanol aziotrope (Freon). The no-clean bare board assembly used both the cleaned and not cleaned conditions. RMA fluxed boards used only bare boards that were not cleaned.

The cleaned bare boards showed low ionic residue levels (chloride) and passed surface insulation resistance (SIR) tests by performing well with high resistance levels throughout the test. The no-clean fluxed and soldered boards showed electrical failures by the 96-hour mark and never recovered. These boards also showed multiple corrosion sites and dendrite growth, along with white residue in many areas of the board. The no-clean fluxed and soldered boards showed no corrosion sites, no metal migration, no white residue and good electrical performance.

The RMA fluxed and solvent cleaned boards showed good electrical performance with no signs of corrosion or metal migration. Numerous white areas existed on the board surface, signifying that moisture reacted with the rosin.

Standard ROSE testing showed acceptable levels for the RMA flux, but ion chromatography (IC) discovered very high chloride levels due to the activator in the RMA flux. Such levels demonstrate that to have good electrical performance, a large amount of rosin must be present to encapsulate the residue. ROSE testing showed acceptable levels for the no-clean assembly that failed and unacceptable levels for the no-clean and RMA assemblies that passed.

Conclusion

The previous experiment supports the idea that ROSE testers are process control tools-not a reliable measure of cleanliness. Therefore, the ROSE tester is not an effective tool to predict electrical performance in elevated humidity conditions. Current cleanliness levels should be based on a user's ability to collect data and compare that data to field performance and ESS testing conditions. Currently, the best way to understand cleanliness is to baseline the entire process with SIR and to periodically audit the process, to compare to the field performance, and use the ROSE testers to identify large changes in the process.

http://www.circuitsassembly.com

Copyright [copyright] 2001 Miller Freeman LLC