Discussion Forum

Hello Everyone,

I have a to design a beam resonator to find the residual stress in the material. The shape is like an H, due to the overhang of the etching. I made a 3D study on Comsol 5.1 even though the material is thin 35 nm (yes, nano), I made a swept for the thin part. I have a value of stress inside the material obtained from wafer curvatures that I inserted as plain stress in Initial values. I constrained the sides in the thickness and I let everything else free. I checked the conversion of the mesh, and was ok, and with the parametric sweep I found the right measures to have just the first eigenfrequency inside my desired range. The results is a H "squared" 20x20 um with a width of 5 um. Just to check that everything is correct, I removed the overhang and I have compute just for the single part so I have a beam 20x5x0.035 um obtaining a first eigenfrequency of 10.3 MHz and I tried to compare with the analitical solution from "https://www.researchgate.net/publication/231134551_Residual_stress_measurement_method_in_MEMS_microbeams_using_frequency_shiftdata" and from a PhD thesis on MEMS. Both, from different approaches give the same results that is about 3 times less than the eigenfrequency from Comsol. The error is actually sqrt(10) but I checked all the units many times on excel and I made in two different sheets. Am I doing something wrong on the FEM analysis? attached there are the pictures and the mph file.

Thanks, Emanuele