Surface insulation resistance methodology for today's manufacturing technology

Abstract

Corrosive contaminants left on a circuit from assembly and manufacturing processes present reliability problems. Contemporary surface insulation resistance (SIR) measurement procedures consist of daily resistance measurements across a comb pattern on samples that age in environmental chambers. Yet these tests lack information on the corrosiveness of the contaminants and often exhibit inconsistency as quantitative measures. A dc continuous measurement method is used here to study the fundamental science behind these measurements for ionic contaminants on a printed circuit board. For ionic contaminants, such as those left from low-solids-fluxes (LSF), the SIR values exhibited continuous and irreversible changes during the test. The dc voltage causes mobile ions to migrate toward the electrodes and thus deplete these ions from the insulating surface. Therefore, only the initial measurement on a virgin sample gives the true quantitative measure of these contaminants. Subsequent rise in SIR value does not indicate an improvement in reliability but rather the presence of mobile ions. Once the board is depleted of ions, reversing the applied voltage polarity cannot restore the initially low SIR value. Instead, it only gives a slow drop followed by a slow rise in the SIR value. These changes are responsible for many measurement anomalies commonly observed under the dc-biasing schemes in existing SIR measurement specifications and standards where the SIR values are recorded only once daily, An alternate SIR measurement using a large ac voltage to monitor product reliability needs be explored. An ac voltage or a small dc voltage causes little net ion migration and therefore gives more consistent results. However a small ac or dc voltage does not produce the desired aging effects of the large dc voltage used in present SIR measurement standards. Preliminary results from the large ac voltage measurements here also gave much slower changes in the measured SIR values. Yet the large ac voltage used may still cause voltage-acceleration of some real-life degradation and failure mechanisms.