Hello,
I'm looking to determine the residual strain after the release of a MEMS structure.
I fabricated micro-bolometers above a sacrificial polyimide layer. On removal of the polyimide the bolometer membrane is deformed. Electrically the bolometer is still intact but from SEM images I can see it is under tensile stress while the metal supporting arms are under compressive stress.
I have attached an SEM image showing the bolometer after the sacrificial removal of the polyimide layer.
Obviously I'd like to tune this stress out of my structure. I could do this experimentally but what I think is best to do is simulate some strategies before committing to a laborious and costly fabrication effort.
I have access to COMSOL and have been looking at tutorials on how to determine the residual stress in a structure. However, knowing little about mechanical engineering, I'm still not entirely sure how best to tackle the problem.
Can anyone offer me some assistance?
Many thanks,
James
Any one of the technique, you can use:
X ray diffraction; Neutron diffraction/Ultrasonic techniques
Photo-stress techniques
It's the compressive stress in the arms giving the tensile stress in membrane. If you can measure the frequency of the membrane you can calculate the residual stress in the system from a Comsol model.
Your structures are curved, do they consist of two materials? If the structures can be approximated as free standing cantilevers and you know the layer thicknesses and the Young's modulus of the materials you can calculate the stresses from the curvature.
Measurement is the only effective way to have knowledge of residual stress : stop. From a theoretical point of view, XRD and neutron diffraction can be the best choice for crystalline materiala: these techniques are non destructive at all and give you a complete overview of residual stress in any direction. XRD can be performed in a lot of lab , it's quite cheap (in italy about 100 - 150 euro for a small component like your) but it's able to measure only a thin layer of 5 - 10 micron. Neutron diffraction is more effetive (down to 2 mm !!) , but it's possible to perform only in very few lab and itìs very expensive. If measurement can be destructive, also strain gauge technique are very effective and simple to perform
@Pieter Westerik - my membrane and legs are made of multiple materials.
@Francesco Chici - thanks for the info.
another technique, suitable for membrane, could be to trace a grid of point on surface, realize a small hole and detemine grid deformation. This result can be suitable for FEM modelling of residual stress.
in your physical model, you may be able to etch a fine square grid onto your surface. If you could make before and after measurements of your grid-you will have a measurement of the strain on the surface. this could give you a rough estimate of your residual stress. You could also measure the curvature of your device, to get an estimate of the bending moment and resulting strains. the SEM may allow you enough magnification to measure these items directly-at least roughly.
dear Metrovich , grid strain is not a "rough estimation" expecially if membrane remain flat
You just simulate the resonance frequency of the structure using any FEM solver like cosmos or coventorware. Thereafter you have to measure the resonant frequency of the structure. There is some option in coventorware where you can apply residual stress in the structure and simulate its resonant frequency. In this way you can calculate back the residual stress from the shift in resonance frequency.
- Badges
- Science topic
- Similar topics
- Engineering
- Mechanical Engineering