Water vapor penetration rate into enclosures with known air leak rates

The rate at which water vapor penetrates a faulty seal in a "hermetically sealed" device has been examined both experimentally and theoretically in this study. The experimental data were generated by studying the rate of moisture ingress into two types of packages; a TO-5 can and a ¼ × ¼ in ceramic flatpack. The TO-5 packages were fitted with capillary tubes of varying diameters with known air leak rates and represented the gross leak range (down to 10
^-5
atm . cm
³
/s) while the ¼ × ¼ in ceramic flatpacks were used to study the fine leak range (10
^-6
to 10
^-8
atm . cm
³
/s). Both type packages were fitted with miniature moisture sensors (surface conductivity type) and calibrated at known relative humidities prior to sealing. The finished test packages were then exposed to constant high-humidity conditions and monitored for moisture content as a function of time. A theoretical analysis was performed by using conventional leak rate equations to convert air leak rates, at standard test conditions, to water vapor leak rates at partial pressure differentials seen typically at room ambient conditions. These data compared favorably with the experimental results in the range tested and suggests that present hermeticity specifications are too lenient for long-term reliability requirements.