An Appraisal of High Temperature Humidity Stress Tests for Assessing Plastic Encapsulated Semiconductor Components

Very high temperature, high humidity lifetests and saturated storage ('pressure pot') tests are increasingly being adopted as rapid replacements for the standard 85°C/85% RH lifetests used to assess the susceptibility of plastic encapsulated ICs to moisture. It has also been suggested that a reduction in test time could be achieved by the detection of debonding between the plastic and the die, since debonding may be required to initiate the processes which eventually lead to failure. To evaluate the alternative techniques, CMOS from 6 manufacturers has been subjected to a series of biased lifetests up to 131°C/90% RH and to unbiased saturated storage tests up to 131°C, and a capacitance technique was used to detect the onset of debonding between the plastic and the die. It is concluded that debonding is only a necessary precondition for functional failure during (i) saturated storage tests and (ii) lifetests on specimens having a passivation which is defect free and either contains no phosphorus or has a low phosphorus content, and that in both cases corrosion will be confined to the bond pads. Such corrosion was produced by lifetests on oxide-passivated components, but the majority of failures during these tests were associated with two other corrosion features, namely localised corrosion of passivated tracks and uniform corrosion of all cathodic tracks, neither of which required debonding. There is evidence that the localised corrosion is associated with passivation defects and the uniform corrosion with the presence of phosphorus.