The drive to increase circuit density with smaller PWB geometries and higher layer counts in multi-layer boards along with the increasing use of electronics in harsh environments for high reliability and safety critical applications (automotive, avionics, medical, military) have made short circuiting of PWBs due to growth of conductive filaments between biased conductors a major concern. In addition, the impending implementation of lead-free soldering processing, which may affect laminate stability and materials choices, can increase the potential for conductive filament formation (CFF) failures. To mitigate these catastrophic failures, it is necessary to understand the roles and synergistic effects of environmental conditions, material properties and manufacturing quality in accelerating or deterring CFF.

In this dissertation, four laminate types (including a halogen free) and three conductor spacings are tested at different voltages in accordance with IPC TM-650, allowing a ranking of these laminate types based on resistance to CFF. Demonstrated is the use of an innovative technique, the superconducting quantum interference device (SQUID), to verify and locate the internal short circuits due to CFF. The SQUID which detects magnetic fields generated by the current paths, displays images of the current density enabling identification of the shorted locations. With this technique, a new variant of filament formation in glass/epoxy composites: vertical filament formation (VFF) was identified. The conductive filaments found at the failure sites were observed during cross-sectioning techniques, verifying that the failures were due to CFF. A test standard to identify and quantify hollow glass fibers, potential paths for filament formation in laminated PWBs, was created. It was observed that board types, which show the longest time to failure due to CFF in PTH-PTH configuration, might not offer the best protection for PTH-plane geometry. Based on insulation resistance measurements, it was seen that the IPC-TM-650 test specification of monitoring every 24 hours could allow intermittent failures to go undetected. It was demonstrated that PTH-PTH dielectric breakdown voltage values followed the same trend as the time to failure observed for the PTH-PTH CFF failure data, suggesting that dielectric breakdown voltage can be an indicator of CFF susceptibility, saving considerable time and cost.