Quantitative Hermeticity Assessment of Packages with Micro to Nano-liter Cavities

Abstract

Hermeticity is a measure of the "leak-proof ness" of packages with internal cavities and is critical for ensuring proper operation of the devices/circuits enclosed in them. The most widely used hermeticity detection technique in the industry is the helium fine leak test. The exiting conduction based governing equation is examined to investigate the volume dependant limits of the test when applied to metal sealed MEMS packages. The results clearly indicate that the test has limited applicability for small internal volumes (1 nanoliter - 1 microliter). The limited applicability of the guidelines specified in Method 1014.11 of the MIL-STD-883F document for hermeticity characterization is also characterized.

To cope with these limitations, a regression analysis based procedure is developed and implemented to extract the true leak rate from the apparent leak data. While the apparent leak rate obtained directly from the He mass spectrometer changes with the test parameters, the true leak rate remains constant and this can be used as a metric to evaluate a package seal.

The hermeticity of polymer sealed MEMS packages is also studied. Unlike metal sealed packages, gas transport in polymer sealed packages occurs via diffusion. A gas diffusion based model is proposed to study the hermetic behavior of these packages. An effective numerical scheme is developed to implement this model and simulate the change in cavity pressure as gas flows into or out of the cavity through the polymeric seal. An optical interferometry based leak test is developed to experimentally measure this change in cavity pressure. The experimental data is used to verify the validity of the proposed numerical scheme and the assumption of adiabatic boundary conditions made in the numerical model. An inverse method is presented to determine the two diffusion properties, diffusivity and solubility, of the polymeric seal by using the experimental data iteratively with the numerical data. The proposed method offers unique advantages over the routinely practiced/existing gas diffusion property measurement techniques.