1. Northwestern University H. D. Espinosa ME 495 Testing of MEMS Materials Using Thermal Actuation, AFM Image Correlation and Capacitance Measurement Students: Y. Chen, B. Peng, and Z. He Mentor: Y. Zhu March 20, 2003

2. Northwestern University H. D. Espinosa ME 495 Introduction The reliability of MEMS devices is a major issue and it can only be addressed by direct measurements on small specimens with dimensions on the same order of magnitude as the fabricated micro-devices Knowledge on bulk material behavior fails to describe material response in this size regime Various techniques have been developed in the past to address the issues of mechanical integrity and evaluation of the elastic properties of polysilicon. Tensile tests are less vulnerable to geometry-induced errors and the measurements are easier to interpret from an error analysis point of view

3. Northwestern University H. D. Espinosa ME 495 Experimental setup A Actuation Pads S Sensing Pads S S A A V Thermal ActuatorLoad Sensor High-Resolution AFM Imaging of Specimen Film Specimen +V o -V o V sense S Three components: Thermal actuator, specimen, load sensor Scheme of the MEMS tensile testing device

4. Northwestern University H. D. Espinosa ME 495 AFM image shows the surface roughness and grain size of the specimen Optical image of the microtensile specimen and the MEMS testing device. 1 micron

5. Northwestern University H. D. Espinosa ME 495 Experiment Step 1: Characterize the system error by scanning the same area at zero loading to eliminate the results of the shifting and minimize the measuring error Step2: Using AFM records of deformed and undeformed specimen configurations to measure the strains of the surface of the specimen

6. Northwestern University H. D. Espinosa ME 495 Step 1. Scanning the same area without load Two AFM images of the undeformed specimen surface. The images are scanned at the same location with a frequency of 2 Hz. The scanning area is 5x5  m 2. It is very hard to find the difference of the two images by naked eyes.

7. Northwestern University H. D. Espinosa ME 495 Analyze the displacement by DIC method (1)  xx, 10  m x 10  m  xx = 0.000521,  yy = -0.0484

8. Northwestern University H. D. Espinosa ME 495 Analyze the displacement by DIC method (2)  xx, 5  m x 5  m  yy,, 5  m x 5  m  xx = 0.000569,  yy = -0.0509

9. Northwestern University H. D. Espinosa ME 495 Analyze the displacement by DIC method (1)  xx, 2  m x 2  m  yy, 2  m x 2  m  xx = 0.000518,  yy = -0.0533

10. Northwestern University H. D. Espinosa ME 495 (a) S train of undeformed specimen at x direction (b) S train of undeformed specimen at y direction (c) S train of deformed specimen at x direction (d) S train of deformed specimen at y direction X Y 1  m Step 2: DIC results of strain analysis on both undeformed and deformed specimen. The four images are at the same location. X is the direction of the load.

11. Northwestern University H. D. Espinosa ME 495 Calculated Strain:

12. Northwestern University H. D. Espinosa ME 495 Simulation

13. Northwestern University H. D. Espinosa ME 495 Conclusion The absence of surface developed charges makes the use of a probe microscope possible in the regime of small interaction forces between the film and the AFM probe. This technique is less vulnerable to geometry-induced errors and the measurements are easier to interpret from an error analysis point of view. Although there is an initial drifting of the AFM image at zero load, the strain of the specimen can be compensated by subtract the measured strain by the initial shifting Future work will be conducting measurements at various voltages (loads) to find the sensitivity of the strain measurement.