The Effects of Flux Residues on Reliability

Eric Bastow
Indium Corporation, Clinton, NY, USA

With the predominance of no-clean soldering processes and ever decreasing component standoff, the industry has had to consider the reliability of, what may be, partially activated or "gooey" flux residues under component bodies. Similarly, questions have also risen about the reliability of flux residues resulting from the reflow of no-clean solder pastes that are "entrapped" under RF shields or "cans", where escape of the volatile ingredients of the flux is greatly hindered.

In this paper, discussion will be made regarding an experiment designed to mimic the aforementioned conditions and how these conditions affected the SIR performance of the no-clean flux residues. A variety of no-clean solder paste flux residues will be discussed, including a halogen-containing, Pb-free solder paste flux; a halogen-free, Pb-free solder paste flux; a halogen-free, Pb-free solder paste flux with a residue optimized for pin probing; and a halogen free SnPb solder paste flux.

The results of this experiment do provide reasonably conclusive evidence that certain characteristics of the flux residue are altered solely as a result of being covered and/or entrapped at varying degrees. Other sensible conclusions can be made as well.

1) All the residues under the glass slides were "gooey" and all of the residues under the RF shields were "hard".
2) All of the fluxes cause discoloration on the underside of the RF shield.
3) Certain fluxes produce residues that have notable visual anomalies when covered with an RF shield.
4) Covered fluxes, be it with a glass slide or RF shield, produce lower SIR values than uncovered, albeit not necessarily failing readings (<1.00E+8 ohms per square).
5) Halogen-free flux chemistries do not necessarily provide a more electrically "reliable" residue than halide containing residues in these scenarios.

Variations in the SIR performance of a flux, in these scenarios, is likely possible from one flux formulation to another. For example, it may not be wise to anticipate that all halogen-free, pin-probable, Pb-free, no-clean fluxes will perform in the same manner. Some flux formulations may perform better than others under these conditions.

Long term reliability studies of non-hermetic semiconductor SSPA products containing no-clean flux in close proximity to the MMIC need not be performed. Answer is don't. The outgassing and residues left on the surfaces of the semiconductor die does degrade performance of Ku and Ka band products.

Thank you for you work described in the paper presented at IPC. A similar subject on No Clean Flux residue effect on active RF circuits needs to be studied and presented. I've had significant problems with trapped flux under parts in RF filters. It seems the flux has a high dielectric constant that results in added capacitance to the adjacent PCB layer, de-tuning the filter passband. Frequencies above a few MHz are prone to this effect, the higher frequencies, the greater the effects of trapped flux.

Is anyone aware of published studies or plans to start one for effects of trapped flux on active RF circuits?