Adhesion is important to the yield and performance of MEMS (Micro-Electro-Mechanical Systems) and NEMS (Nano- Electro-Mechanical Systems). Measuring the work of adhesion, with an AFM (Atomic-Force Microscope), allows one to study surface interactions from an energy perspective as opposed to an adhesion (force) perspective. The works of adhesion were measured with different AFM tip radii on multiple types of samples. Two sample variables were examined: four die attach conditions (no attachment, silicone, polyimide silicone, and silver glass), and two surface conditions (native oxide with and without a few angstroms of vapor-deposited diphenyl siloxane). For a normal silicon cantilever, the work of adhesion seems to be slightly less for the treated surfaces than the untreated, except for the silverglass die attach material. There were at least three orders of magnitude difference in the works of adhesion for the different AFM tip radii. Presumably, the trend is due to some combination of material properties, interfacial roughness, and torque on the AFM cantilever.

The debugging-time for polycrystalline MEMS structural layers is the aim of this paper. A description of the fatigue phenomenon for elementary structure in the case of microactuators is presented. It is obtained by combining elementary in situ test benches of varying dimensions and cyclic actuation; in the same way, the debugging-time is determined according to the design and the structural material. Test benches have been developed allowing performance of bending fatigue tests of polycrystalline structural layers, describing the fatigue phenomenon and obtaining a constant debugging-time for a same polycrystalline layer, whatever the excitation frequency and the beams length.

Microelectromechanical systems (MEMS) that sense, think, and act are enabling technologies currently employed in many industrial applications. To operate these devices, a stimulus is required to produce motion. In MEMS, this stimulus may be thermal actuation using current to produce joule heating, or electrostatic actuation using voltages to create electric fields. To qualify MEMS technology, these devices must undergo repeated characterization and testing and at both the die and system level. Electrical overstress (EOS) and electrostatic discharge (ESD) are two important tests used to assess the robustness of a device to steady state and sharp voltage and current transients. Identifying the failure mechanism and understanding the root causes for failure is paramount to the overall improvement and success of any MEMS based system. In this paper we will focus on the effects of EOS and ESD events on surface micromachined polysilicon based electrothermal actuators fabricated using the SUMMiT V™ process.

An ellipsometry based measurement protocol was developed to evaluate changes to MEMS sensor surfaces which may occur during packaging using unpatterned test samples. This package-level technique has been used to measure the 0-20 Angstrom thin films that can form or deposit on die during the packaging process for a variety of packaging processing conditions. Correlations with device performance shows this to be a useful tool for packaged MEMS device and process characterization.

In this paper we discuss reliability and failure analysis issues of RF-MEMS capacitive switches. We describe specific instrumentation and methods that can be used for testing and examination of these switches. These include SEM, AFM, SAM, static and dynamic optical investigation and electrical lifetime testing. Processing as well as testing and packaging issues are discussed.

Surface micromachined micromirror technologies are being employed for various commercial and government applications. One application of micromirror technologies in the commercial sector can be found in Digital Light Projection (DLP™) systems used for theater and home entertainment centers. DLP™ systems developed by Texas Instruments uses DMD™ technology (Digital Mirror Device), an array of micromirrors, to project light onto a screen [1]. This technology is also used by Infocus™ projection systems and widescreen tabletop televisions [2]. Here, the micromirrors act as individual pixels, reflecting light onto the screen with high ¡§digital¡¨ resolution. The most recent application of surface micromachined micromirror technology is optical switching [3], which uses micromirrors to switch optical signals from fiber to fiber for lightwave telecommunications [4]. Companies such as Lucent have fabricated entire optical micromirror switching systems based on their Microstar™ technology [5]. For government applications, surface micromachined micromirror arrays have been developed for potential use in a spectrometer system planned for NASA's Next Generation Space Telescope (NGST) [6]. Various processing technologies are used to fabricate surface micromachined micromirrors. The micromirror arrays developed by TI and Lucent [1,4] uses metal for their structural and reflective components. Micromirrors fabricated at Sandia National Laboratories use the SUMMiT™ (Sandia's Ultra-planar MEMS Multi-level Technology) process with metal deposited on the surface of mechanical polysilicon components to reflect light. Optical micromirror arrays designed and fabricated at Sandia for potential use in the NGST have undergone reliability testing and failure analysis. This paper will discuss the failure modes found in these micromirrors after reliability testing. Suggestions and corrective actions for improvements in device performance will also be discussed.

The Digital Micromirror Device (DMD) is a spatial light modulation Micro-Optical Electro-Mechanical Systems (MOEMS) device used in tabletop projectors, televisions and cinema projection systems. This device creates high resolution, high quality images by deflecting/modulating light with microscopic mirrors. Failure analysis of these devices requires superstructure, package, optics, and substructure approaches. Particles within the active array of a DMD are often killer defects, but those are the subjects of an entire discussion of their own. This paper will show evidence of failures associated with: windows in the package lids, failures of the superstructure area, and failures within the substructure. Methods for removal of the mirrors, as well as other structures, will be covered in greater detail. We will conclude with examples of analysis areas in DMD devices that show how they differ from other types of devices.

A new imaging technique called Wavefront Coding allows real-time imaging of three-dimensional structures over a very large depth. Wavefront Coding systems combine aspheric optics and signal processing to achieve depth of fields ten or more times greater than that possible with traditional imaging systems. Understanding the relationships between traditional and modern imaging system design through Wavefront Coding is very challenging. In high performance imaging systems nearly all aspects of the system that could reduce image quality are carefully controlled. Modifying the optics and using signal processing can increase the amount of image information that can be recorded by microscopes. For a number of applications this increase in information can allow a single image to be used where a number of images taken at different object planes had been used before. Having very large depth of field and real-time imaging capability means that very deep structures such as surface micromachined MEMS can be clearly imaged with one image, greatly simplifying defect and failure analysis.