TIME DEPENDENT DEFORMATION OF THIN FILM PLATINUM DURING NANOINDENTATION TESTING

Open Access
Author:
Tremper, Amber Leigh
Graduate Program:
Materials Science and Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
December 18, 2008
Committee Members:
  • Dr Christopher Muhlstein, Thesis Advisor
  • Christopher Muhlstein, Thesis Advisor
  • Lawrence Friedman, Thesis Advisor
Keywords:
  • nanoindentation
  • platinum
Abstract:
Previous nanoindentation studies of thin film platinum have reported elastic moduli values that are approximately 10 to 25 percent lower than the expected polycrystalline aggregate values. However, no explanation of these low moduli values was given. This study attempts to explore the cause of the low moduli values and further evaluate thin film platinum properties. Instrument artifact, spatial variation, and time-dependency were examined as possible explanation of the low moduli values. Instrument artifacts such as pile-up, machine compliance, residual stresses, surface roughness, and delamination and/or microcracking were eliminated as potential sources of the low moduli values. After correction for both pile-up and machine compliance, the platinum film in this study was found to have a reduced modulus approximately 10 percent lower than that expected from anisotropic elasticity and indentation theories. The film did show spatial variation in mechanical properties, with one region having a modulus that was approximately five percent higher than the theoretical prediction. Additionally, the two different regions showed differing sensitivity to loading and unloading rate during indentation testing. These two combined facts suggest that the material has a spatial variation of mechanical properties, which could be caused by surface chemistry or morphology, localized processing effects, or discohered platinum from the silicon substrate. In the low moduli regions, the film exhibited a time dependent behavior, likely due to anelasticity, or reversible linear viscoelasticity. The loading rate dependency of the load-displacement curves and the independency of the residual indentation depth support this theory. Ultimately, the reported low moduli values previously reported in the literature and in this study are likely due to spatial variation and an anelastic response of thin film platinum, and are not caused by instrument artifact. During the testing of thin film platinum, it was discovered that several indentation tests showed anomalous behavior at very slow loading rates that could not be explained in conjunction with the other, normal results. The abnormal tests had bulging loading and unloading curves, causing the material to appear inconsistently stiff during loading and soft during unloading. When the system displacement drift was examined, it was found to be an important factor in the quality of slow loading rate, long duration test data. Drift experiments showed that the drift rate during a test is non-constant, uncorrectable, and may cause large uncertainties and errors in the values derived from long duration tests. The accumulated drift as a percentage of maximum indentation depth proved to be a good criterion for identifying unreliable data. When the accumulated drift percentage was larger than 100 percent, the elastic moduli values were non-physical due to gross abnormalities in the force-displacement curves. Therefore, tests with accumulated drift percentages larger than 100 percent should be discarded since they are likely to cause error and uncertainty in indentation test data. Additionally, errors in displacement and subsequently calculated values due to non-constant drift should be reported, particularly in long test times.