Sandia improves first-pass success in MEMS by providing expertise in reliability, testing, and characterization thereby enabling our national security customers to effectively utilize state-of-the-art technologies.
The area of MEMS reliability, test, and characterization is unique in that the “product” is generally data collection leading to model validation or predictive model development. Our focus is on a science-based understanding of failure mechanisms, material properties, and surface interactions. We use established reliability methods to develop acceleration factors for MEMS device lifetime prediction. The greatest challenge in MEMS reliability is that failure mechanisms appear to be device dependent. For this reason, we classify devices according to their operational interactions to group mechanisms of interest. We find that stiction/adhesion/resistance between contacting surfaces is a consistent yield-limiting and failure mechanism.
- Class I – no moving parts
- Class II – moving parts with no contact
- Class III – moving with impacting surfaces
- Class IV – moving with rubbing surfaces
- We have developed characterization tools (interferometers, image capture techniques) enabling validation of models.
- We have developed large, parallel test systems (SHiMMeR, environmentally-controlled shaker, Vena chambers) to perform long-term tests in various environments (temperature, humidity, vibration)
- We have developed test structures and techniques for MEMS surface analysis and MEMS material properties measurement.
- We leverage a solid science base to understand failures and develop predictive reliability models.
- We have expertise in electrostatic/thermal actuators, adaptive optics µmirror arrays, RF MEMS, inertial switches, optical switches, and resonators. (Bibliography)
In our lab, we have one of the first WYKO white-light profilometers with Dynamic MEMS capability. This enables both in-plane and out-of-plane resonant frequency measurements. In addition, we have access to a Laser-Doppler Vibrometer to measure out-of-plane motion (with org. 1525). For further characterization or functionality testing, we have developed MEMScript™ which is an integrated vision and actuation automation tool for MEMS.
Figure 1 WYKO NT1100 with DMEMS
Figure 2 Analysis of skipping gear using MEMScript
Parallel Test Systems
Reliability testing requires the development of large parallel test systems where environments can be controlled and long-term stress tests can be realized. Both optical and electrical inspection must be available with capability to halt the actuation signal immediately after a device has failed. We have built two packaged part test systems (named SHiMMeR I or II, for Sandia High Volume Measurement of Micromachine Reliability). SHiMMeR is a flexible, modular system capable of testing a wide variety of device types. The environments can be air or nitrogen, controllable up to 65% RH and at ambient temperature. SHiMMeR II has the added capability of stress testing at higher temperatures through the use of resistive strips under the packages.
Figure 3 SHiMMeR vibration isolation and chamber
Figure 4 SHiMMeR chamber loaded with devices under test.
In addition, we have modified Vena Chambers (eight chambers and one controller) that can hold the same board design as SHiMMeR for temperatures up to 85C and humidity levels up to 85%.
Figure 5 Vena Chambers
We also have a one-board design unit completely enclosed in a glove box named SHiMMeR Lite. This unit can serve as an environmentally-controlled experiment space or a controlled functionality testing. The glove box environment can range up to 85% RH in air or nitrogen (with less that 10 ppmv oxygen).
Figure 6 SHiMMeR Lite
Many of our experiments require long time intervals (up to 2000 hours â€“ 80 days) so we have our own bank of ovens. We have 6 Delta Design 3900 ovens which range in temperature from -65C to +315C. All are equipped with temperature recording equipment. We also have 3 Delta Design ovens with controllers that can cycle from -65C to +315C over time. Any JEDEC/Mil-SPEC standard can be matched.
We recently developed a long-term reliability shaker system to investigate fatigue in MEMS spring/mass systems. We can stress 16 spring/masses at a time to bands of frequencies and displacements. A Laser Doppler Vibrometer is used to measure resonant frequency of the specimen after stress intervals.
Figure 7 Long-Term Reliability Shaker System
Test Structures and Techniques
We have developed a series of structures for mechanical property measurement in MEMS using an interferometry approach. The slide below shows the properties addressed so far.
Figure 8 MEMS Material Properties
A technique to perform surface characterization and calculate adhesion energy between a cantilever beam and the surface below it is described in the slide below. This has proven extremely useful in accelerated aging experiments.
Figure 9 Surface Adhesion Measurements
Solid Science Base
We follow the standard reliability method as shown in the slide; however we link to basic materials properties and surface characterization when needed to understand the physics of failure. This method enables predictive model development with the smaller sample sizes prevalent in national security applications.
Figure 10 Using a solid science base
Many different Sandia organizations are tapped for their knowledge and expertise. Here are some links: Materials Sciences; Microsystems; Failure Analysis
Predictive Reliability Models
We use traditional reliability methods to perform accelerated testing outside the normal operating specifications to produce failures. Understanding the physics of failure enables use of acceptable functional forms for the mechanism of interest.
Figure 11 Accelerated Aging of sliding contact MEMS with one hub
Figure 12 Wear in microengine pin joints