Boston University

Optical Characterization and Nanophotonics Laboratory

Home

About

People

Papers

Research

Facilities

Alumni

Raman measurements of Stress and strain in Silicon

Members: Anna K. Swan

 

Alumni: Bennett B. Goldberg

 

Introduction

MEMS technologies are evolving at a rapid rate with applications in a wide variety of sensors and actuators. There is currently great interest in extending these technologies to fabricate micro-machines for power generation, propulsion, cooling, and fluid pumping. The resulting structures are often subjected to high mechanical stresses, which if too high, will cause failure. Hence knowing the local stress is an important design consideration. Here we are using Raman scattering to locally measure the stress in a bent silicon cantilever and comparing with calculations.

Test structure

We are using Raman spectroscopy to measure the stress and strain in a silicon flexure to compare the result with simple analytical expressions as well as numerical models. This image shows a SEM image of the root of the flexure structure we have been using. At the end of the long flexure (dimensions: ~3mm long, 150 um across and 480um deep), a silicon ball is weged beween two flexures. The resulting deflection (69um) causes different amount of stress/strain along the leg of flexure.

Measurements

The strong Raman signal at ~520 -1 from silicon can be used to measure stress or strain applied to the silicon structure. We are using a Renishaw Ramascope 1000B for our measurements. A shift of the Raman peak to higher energies indicate compressive stress. Likewise, a shift to lower energies indicate tension. The spectral resolution is 0.02 cm-1 which corresponds to 10 MPa stress or strain. The spatial resolution is 1um. Hence, the stress is measured locally in a 1um area. The flexure will have the strongest strain and stress field at the root of the flexure.

Models

Two approaches were used. In the first, the flexure is considered to be an Euler-Bernoulli cantilevered beam of uniform rectangular cross-section of length L, width b, thickness h, and tip deflection delta. The analytical Euler-Bernoulli (E-B) model is expected to be valid for small deformations as is the case here. However, the E-B beam model is expected to be invalid in a region approximately 0 < x < h due to the finite compliance of the supporting silicon structure. In order to take these effects into consideration, numerical analyses were performed using the method of finite elements (FEM). The image shows the calculated stresses taking support compliance into account from the FEM.

Results

All the models match measurements to 25 - 35 MPa at locations 600 - 1000 um from the support, but significantly overestimate the stresses at distances <100 um. Including the support compliance in the numerical model leads to a better match with measurements, as is shown in this figure, comparing the measurements and calculations at 100 um distance from the root.

Publications

V. T. Srikar, A. K. Swan, M. S. Ünlü, B. B. Goldberg, and S. M. Spearing, "Micro-Raman measurement of bending stresses in micromachined silicon flexures," IEEE Journal of Microelectromechanical systems, Vol. 12, No. 6, December 2003, pp. 779-787

V. T. Srikar, A. K. Swan, B. B. Goldberg, M. S. Ünlü, and S. M. Spearing, "Microscale measurement of stresses in a silicon flexure using Raman spectroscopy," Materials Research Society Symposium - Proceedings, Vol. 741, 2-4 December 2002, pp. 219-224


Collaborators

V. T. Srikar and S. M. Spearing , Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.


Support

Micro machines at Sandia national laboratory
MEMS Clearinghouse


© 2007 Trustees of Boston University. All rights reserved.  |  Last modified April 16, 2007 at 12:00 AM EDT