M. Steven Rodgers

Advanced micromechanisms in a multi-level polysilicon technology

Quad-level polysilicon surface micromachining technology, comprising three mechanical levels plus an electrical interconnect layer, is giving rise to a new generation of micro-electromechanical devices and assemblies. Enhanced components can now be produced through greater flexibility in fabrication and design. New levels of design complexity that include multi-level gears, single-attempt locks, and optical elements have recently been realized. Extensive utilization of the fourth layer of polysilicon differentiates these latter generation devices from their predecessors. This level of poly enables the fabrication of pin joints, linkage arms, hinges on moveable plates, and multi-level gear assemblies. The mechanical design aspects of these latest micromachines will be discussed with particular emphasis on a number of design modifications that improve the power, reliability, and smoothness of operation of the microengine. The microengine is the primary actuation mechanism that is being used to drive mirrors out of plane and rotate 1600-micrometers diameter gears. Also discussed is our most advanced micromechanical system to date, a complex proof-of-concept batch-fabricated assembly that, upon transmitting the proper electrical code to a mechanical lock, permits the operation of a micro-optical shutter.

Routes to failure in rotating MEMs devices experiencing sliding friction

Gear systems rotating on hubs have been operated to failure using Sandias microengine as the actuation device. Conventional failure modes such as fatigue induced fracture did not occur, indicating that the devices are mechanically extremely robust. The generic route to failure observed for all rotating devices involves sticking of structures that are in sliding contact. This sticking evidently results from microscopic changes in the sliding surfaces during operation. The rate at which these changes occur is accelerated by excessive applied forces, which originate from non-optimized designs or inappropriate drive voltages. Precursors to failure are observed, enabling further understanding of the microscopic changes that occur in the sliding surfaces that ultimately led to failure.

A Microelectromechanical High-Density Energy Storage/Rapid Release System

One highly desirable characteristic of electrostatically driven microelectromechanical systems (MEMS) is that they consume very little power. The corresponding drawback is that the force they produce may be inadequate for many applications. It has previously been demonstrated that gear reduction units or microtransmissions can substantially increase the torque generated by microengines. Operating speed, however, is also reduced by the transmission gear ratio. Some applications require both high speed and high force. If this output is only required for a limited period of time, then energy could be stored in a mechanical system and rapidly released upon demand. We have designed, fabricated, and demonstrated a high-density energy storage/rapid release system that accomplishes this task. Built using a 5-level surface micromachining technology, the assembly closely resembles a medieval crossbow. Energy releases on the order of tens of nanojoules have already been demonstrated, and significantly higher energy systems are under development.

Ultrasonic actuation for MEMS dormancy-related stiction reduction

The use of ultrasonic pulses incident on surface micromachines has been shown to reduce dormancy-related failure. We applied ultrasonic pulses from the backside of a silicon substrate carrying SUMMiT processed surface micromachined rotors, used earlier as ultrasonic motors. The amplitude of the pulses was less than what is required to actuate the rotor (sub-threshold actuation). By controlling the ultrasonic pulse exposure time it was found that pulsed samples had smaller actuation voltages as compared to non-pulsed samples after twelve-hour dormancy. This result indicates that the micromachine stiction to surfaces during dormant period can be effectively eliminated, resulting in long-term stability of surface micromachines in critical applications.

SAMPLE (Sandia agile MEMS prototyping, layout tools, and education)

The SAMPLE (Sandia agile MEMS prototyping, layout tools, and education) service makes Sandias state-of-the-art surface micromachining fabrication process, known as SUMMiT, available to U.S. industry for the first time. The service provides a short course and customized computer-aided design (CAD) tools to assist customers in designing micromachine prototypes to be fabricated in SUMMiT. Frequent small-scale manufacturing runs then provide SAMPLE designers with hundreds of sophisticated MEMS (microelectromechanical systems) chips. SUMMiT (Sandia ultra-planar, multi-level MEMS technology) offers unique surface-micromachining capabilities, including four levels of polycrystalline silicon (including the ground layer), flanged hubs, substrate contacts, one-micron design rules, and chemical-mechanical polishing (CMP) planarization. This paper describes the SUMMiT process, design tools, and other information relevant to the SAMPLE service and SUMMiT process.

Development of a compact optical MEMS scanner with integrated VCSEL light source and diffractive optics

In this work the design and initial fabrication results are reported for the components of a compact optical-MEMS laser scanning system. This system integrates a silicon MEMS laser scanner, a Vertical Cavity Surface Emitting Laser (VCSEL) and passive optical components. The MEMS scanner and VCSEL are mounted onto a fused silica substrate which serves as an optical interconnect between the devices. Two Diffractive Optical Elements (DOEs) are etched into the fused silica substrate to focus the VCSEL beam and increase the scan range. The silicon MEMS scanner consists of an actuator that continuously scans the position of a large polysilicon gold- coated shuttle containing a third DOE. Interferometric measurements show that the residual stress in the 50 micrometer X 1000 micrometer shuttle is extremely low, with a maximum deflection of only 0.18 micrometer over an 800 micrometer span for an unmetallized case and a deflection of 0.56 micrometer for the metallized case. A conservative estimate for the scan range is approximately plus or minus 4 degrees, with a spot size of about 0.5 mm, producing 50 resolvable spots. The basic system architecture, optical and MEMS design is reported in this paper, with an emphasis on the design and fabrication of the silicon MEMS scanner portion of the system.

Integration of optoelectronics and MEMS by free-space micro-optics

This report represents the completion of a three-year Laboratory-Directed Research and Development (LDRD) program to investigate combining microelectromechanical systems (MEMS) with optoelectronic components as a means of realizing compact optomechanical subsystems. Some examples of possible applications are laser beam scanning, switching and routing and active focusing, spectral filtering or shattering of optical sources. The two technologies use dissimilar materials with significant compatibility problems for a common process line. This project emphasized a hybrid approach to integrating optoelectronics and MEMS. Significant progress was made in developing processing capabilities for adding optical function to MEMS components, such as metal mirror coatings and through-vias in the substrate. These processes were used to demonstrate two integration examples, a MEMS discriminator driven by laser illuminated photovoltaic cells and a MEMS shutter or chopper. Another major difficulty with direct integration is providing the optical path for the MEMS components to interact with the light. The authors explored using folded optical paths in a transparent substrate to provide the interconnection route between the components of the system. The components can be surface-mounted by flip-chip bonding to the substrate. Micro-optics can be fabricated into the substrate to reflect and refocus the light so that it can propagate from one device to another and them be directed out of the substrate into free space. The MEMS components do not require the development of transparent optics and can be completely compatible with the current 5-level polysilicon process. They report progress on a MEMS-based laser scanner using these concepts.