Predicting MEMS Performance under Package Warp with i-ROM MODELBUILDER

The package interface of the i-ROM MODELBUILDER enables you to simulate precisely the influence of package warp on the performance of MEMS sensors and actuators. Temperature dependent bias-offset drift and sensitivity changes can be predicted before manufacturing. It enables you to optimize the performance of inertial sensor products without costly and time consuming experiments on prototypes.

The figure shows a highly enlarged deformation profile of a capacitive accelerometer. Out-of-plane warp components change the overlapping area of comb cells and the electrode gap of bottom plate capacitors.

More significant are usually in-plane displacements which move anchors and fixed comb cells towards or away from the chip center. It contributes strongly to capacitance changes in the electrical domain and stiffness changes in the mechanical domain. Both phenomena are the root cause of thermal drift effects and a non-repeatable behavior in harsh environments.

The following example comes with our User Manual and explains the modeling procedure and the theoretical background.

In the left figure, both anchors move in horizontal direction as shown by the orange arrows.  The springs attached to the mass bodies marked by the red stars observe tensile stress and the springs attached to the anchors marked by the blue stars observe compressive stress. Tensile stresses cause stiffening of bending springs and compressive stresses softening or even buckling effects.

Non-linear mechanical simulation results are illustrate in the right figures. Without warp, all strings have the same stiffness and the deformations (bending lines) are identical.

At large package stress, the upper and lower springs become very stiff and bend only a little. Just the inner springs at the anchors contribute to mechanical displacements at acceleration loads.

Package warp simulations can also be performed on more complicate MEMS products as gyroscopes, micro-mirrors or microswitches.

The observed deformation profiles or the variation of internal strain energy density at test load situations indicate what shape elements of the existing design might cause problems during thermal cycling tests.

Design changes can be applied and evaluated in shortest time thanks to implement non-linear reduced order modeling procedures.

If you need help with your designs please contact us.