ME PhD Oral Examination: Violet Qu (Tuesday, May 31, 10am, CISX-101)

Stanford University Ph.D. Dissertation Defense

Title: "Using a MEMS resonant strain gauge to study thin film stress relaxation"

Violet Qu

Department of Mechanical Engineering

Advisor:  Prof. Thomas W. Kenny

Date:  Tuesday, May 31st, 2011

Time: 10:00 am (Refreshments at 9:45 am)

Location: Paul G. Allen Building 101X auditorium (CISX-101)

Abstract: 

Strain gauges have a wide range of applications. Besides measuring strain, they can be used to make other sensors such as load cells, pressure gauges, accelerometers, and gyroscopes. Of the various strain sensing techniques, a MEMS resonator based approach is particularly attractive because they offer superior sensitivity and resolution. Resonant strain gauges operate on the principle that the resonant frequency of a vibrating structure is a function of the applied strain. After reviewing some state of the art microresonator strain gauges, I will introduce our device -- a double ended tuning fork (DETF) resonator packaged at the wafer level using an epitaxial sealing technology (subsequently called the eSensor). These eSensors have sensitivities of up to 1500 ppm/microstrain, and dynamic strain resolution comparable to state-of-the-art devices (9 nanostrain with 10 kHz bandwidth). However, it is the mid- to long-term measurements where the eSensor really shines, thanks to the unparalleled long-term stability offered by the epi-seal. On the time scale of about a minute, the eSensor has a 0.4 nanostrain resolution. 

With its high sensitivity and resolution, the eSensor is a good candidate for thin film stress relaxation studies. In this part of the talk, I will show how stress changes in a thin film deposited on the outside of the eSensor can be measured. This approach is validated experimentally by using a sputtered platinum (Pt) film to generate a known thermal stress by performing a temperature sweep. The minimum stress change that can be resolved using the eSensor is 0.09 MPa for a 1 micron thick film. 

The last part of the presentation consists of preliminary results from probing the mechanical properties of ALD (atomic layer deposition) alumina films. ALD is a new and fast-growing field. The ALD technique promises to produce films of superior quality than the more conventional thin film deposition methods. Though ALD is finding more and more applications in research as well as manufacturing at lightening pace,  characterization of the films' mechanical properties lags behind. The reason is, in part, because detecting stress signals from films only nanometers thick presents a real challenge to current technologies. Our experimental data show that the eSensor can clearly measure stress changes in ALD alumina films that are 65 to 80 nm thick. I will share interesting discoveries from these first experiments on the time evolution of stress in ALD alumina, and our attempts at understanding them.