Closer look at lead-free joints
07 March 2006
The quest for knowledge about lead-free solder joints takes many forms. In this paper M.H. Biglari and F. Kox discuss their experimental process and results for assessing voiding in joints.
Manufacturers continue to have concerns about the long-term effects of using lead-free solders, and research to continues to determine what these long-term effects might be. In the work described below, PCBs with Ni/Au and immersion Sn finishes have been soldered with SAC and SnPb alloys. Reflow and wave soldering have been performed on three type of industrial. The components soldered to the board were: ceramic resistor, ceramic capacitor, coil, diode, jumper, plastic spacer, hybrid module, flexible PCB, QFP, LED, electrolytic capacitor, capacitor, oscillator, cable or connector. The component finishes encountered were: pure Sn, SnPb, SnPb40 and AgPd. Environmental and bump tests have been conducted and analyzed by means of metallurgical research. Visual and light microscopy using instruments from Olympus, as well as in-circuit tests have been preformed.
Experimental
Different types of PCB used in industrial applications (e.g. lighting, medical and laboratories equipments) were soldered by Evic Electronics BV (Neways). Reflow soldering was carried out using Vitronics-Soltec reflow equipment (Type Quantis pro). Wave soldering was performing by Vitronics-Soltec wave equipment (Type Delta 6622-C).
The desired components for light microscopic research were cut out from the selected PCB with a combitool (Ferm type FCT-300) using a cutting disc. Then the component was embedded with cold embedding material (Technovit 4074 Solid: Liquid 2:1). The sample was cured for about twenty minutes. The upper layer was ground with a grinding disc (grid 240) until the desired cross section surface was reached. The sample was then ground in an up-and-down motion. Turned 90 degrees to the rotation direction every time, the sample was subjected to increasingly finer grinding (grids: 240, 320, 400 and 600). The sample was washed off with water and alcohol and put in an ultrasonic bath, filled with alcohol, for 2 minutes. It was washed again with alcohol and this was dried with pure nitrogen. The sample was polished using a turning diamante disc. First, the biggest grain size (9 æm) was applied to the disc at the slowest rotation speed (150 rpm) and wetted with alcohol. The sample was turned in the opposite direction of the rotation, with about twenty to thirty rotations completed. Before polishing the sample in the same manner on another disc with a finer grain, the sample was washed as described above after the grinding. In total, three polishing steps (9, 3 and 1æm) were used. After polishing, the sample was washed once more and put on OP-chem disc with Struer 0.05æm aluminum oxide emulsion.
Light microscopy and stereo microscopy were performed using an Olympus stereomicroscope with a zoom range of 3.25-52x and an Olympus BX60M light microscope, with objectives covering 12.5-1000x.
Three types of PCB were soldered using both lead-free and conventional SnPb solder alloys. For lead-free process the solder used was SAC (SnAg4.0Cu0.5) alloy. For each type, twenty five boards were soldered.
Type 1 was a RF4 board with a chemical Sn finish. Fitted components were: ceramic resistor, capacitor, coil, diode, jumper or connector. The component finish was pure Sn, SnPb, SnPb40, SnPb38Bi2 or AgPd. Twenty five boards were preheated by IR to approximately 110§C and wave soldered with SnAg4.0Cu0.5 using a spray flux. The solder bath had wave nozzle held under nitrogen at a bath temperature of 265 §C and a contact time of 3.7 s. When defects were noticed, repair was done by hand soldering. For traceability during the tests, the boards were numbered with an extra proto assembly number.
Type 2 was a RF4 board with a Ni/Au finish. Fitted components were: ceramic resistor, capacitor, coil, diode, jumper, plastic spacer, co-ax cable or connector. The component finish was SnPb, SnPb38Bi2, SnPb40, or AgPd.
Type 3 was a RF4 board with a chemical Sn finish. Fitted components were: ceramic resistor, capacitor, coil, diode, jumper, plastic spacer, hybrid module, flexible PCB, cable or connector. The component finish was SnPb, SnPb40, SnPb38Bi2 or AgPd.
Two boards of each type, one soldered with lead-free solder alloy and one soldered with SnPb solder alloy were chosen to be undergo metallographic research as 0-hrs references. The other boards were subjected to environmental and mechanical trials.
Results and discussion
The visual inspection showed that the fluidal behavior of the lead-free solder was far from optimal. Solder spatters and superfluous solder were observed. The metallographic analyses revealed that voiding is more common and larger in SAC soldered boards compared to SnPb soldered boards. All boards were subjected to a visual inspection. Poor wetting at different solder-component interfaces was observed. Within the solder Ag3Sn and Cu6Sn5 phases were observed within an Sn matrix.
In general, the lead-free process produced many more voids than the lead process, which were also larger. This can be contributed to the interaction between flux and lower density of lead-free solder. The following phases have been observed in the solder AgSn3, Cu6Sn5. Two line scans were made, one at the underside from the ceramic part crossing the solder into the copper pad and one scan at the side from the ceramic into the solder. The scans reveal the AgPd substrate and Sn mixing very much at the thin underside. The holes are caused by the diffusion of Sn into the AgPd substrate. The gaps the migrated Sn atoms leave in the structure, coalescence to bigger visible holes. This effect is known as the Kirkendall9 effect.
Conclusion
The lead-free soldering process of various components on PCBs with different finishes for industrial applications have been investigated and showed several notable observations. In general, voiding is more common and larger in the lead-free process in comparison to the lead process. The tin-lead process showed relatively few and generally small voids. By lead-free soldering process, there were more and larger voids in PTH components, caused by flux and the lower density of lead-free solder. The most voids are present at the interface between solder and PCB finish and/or at the interface between solder and component. It is essential therefore that manufactures carefully asses the impact of lead-free solders on their components. To do this properly, requires the use of both stereo (e.g. Olympus SZ2 and SZX ranges) and compound microscopes (e.g. Olympus BX2 Material and MX2 ranges) capable of high resolutions at various magnifications.
Contact Details and Archive...
Related Articles...
Most Viewed Articles...