MICROELECTROMECHANICAL SYSTEMS ( MEMS ) IN MILITARY SYSTEMS : LESSONS LEARNED AND REMAINING TECHNICAL CHALLENGES

Like commercial system applications, MEMS-based components used in military systems must meet reliability and lifetime requirements. Testing to ensure that MEMS-based components meet these requirements is a critical step in their development for use in military systems. If an efficient approach for developing test and characterization methods is established, the use of MEMS-based components in military systems may be accelerated. Our proposed approach uses a “design for reliability” approach combined with an approach that uses analytical and computational methods. An example is included. Introduction Several military systems currently use microelectromechanical systems (MEMS) technology. Munitions guidance, navigation, and control (GN&C) systems use MEMS-based inertial measurement units; display systems for ships and aircraft use MEMSbased projection systems; and MEMS-based infrared (IR) imaging systems use MEMS-based thermal sensors. MEMS-based fuze/safety and arming (F/S&A) systems, unattended sensor systems, communications systems, and instrumentation systems are under development for use in military systems. Manufacturers have learned that if component reliability for the application is not demonstrated, MEMS-based components will not be used. As a result, practices to develop and produce reliable MEMS products have evolved [1, 2, 3]. These practices entail designing for product reliability early in the development process [1, 2, 4]. Designing for reliability involves testing and characterization throughout development and production. Manufacturers perform a combination of MEMS device and component-specific tests and “standard” tests to develop reliable products and demonstrate that the product meets specifications [1, 2, 5]. Manufacturers have discovered that testing and characterization tool and method development is a critical part of the development and production process for MEMS-based components. Manufacturers establish component specifications to show that the component will meet the requirements for the intended application. Compared with commercial applications for MEMS-based components, military system applications usually involve a greater required component lifetime and exposure to more environments during the component lifetime. Because of these differences, MEMS developers and military system developers will have to develop test and evaluation practices to ensure repeatable, reliable performance of MEMS-based components in military systems [6, 7, 8]. Since successful component developers employ a “design for reliability” approach, this paper begins with a review of MEMS reliability. Next, commercial MEMS design and development practices based on the design for reliability approach are discussed. This is followed by a discussion of testing and characterization performed during design and development. We then propose an approach for identifying and facilitating development of testing and characterization methods and follow this with an example illustrating the proposed approach. MEMS Reliability “Reliability is the probability that a system, component, or device will perform without failure for a specified period of time under specified operating conditions” [9]. Failure may be separated into two distinct categories: • Degradation failure, which consists of device operation departing far enough from normal conditions that the component no longer meets component performance requirements; and • Catastrophic failures, the complete end of device operation [10]. Early in the development of MEMS technologies and components, the emphasis of many research and development (R&D) programs was on device design, process development, and prototype performance demonstration in the laboratory [4]. The processes and devices were not developed to ensure device manufacturability, reliability, or ability to be packaged. Packaging and reliability efforts were relegated to the later stages of product development [5]. Most Armed Service development programs do not have sufficient funding to take an unpackaged prototype MEMS device to an acceptable maturity level. Consequently, many prototype devices targeted for use in military systems are taking longer to develop than initially planned. To facilitate transition of MEMS-based components into military systems, developers should adopt practices that establish the foundation for MEMS-based component manufacturability, packaging, and reliability. The same is true for MEMS-based component transition into commercial system applications: MEMS-based components must have proven manufacturability (i.e., high yield) and proven final component (i.e., packaged device) reliability to be commercially feasible. Manufacturability, packaging, and reliability are inseparable. The experience of MEMS product manufacturers indicates that reliability must be designed into the product using an approach that begins at the product concept phase and continues through the production phase [1, 2, 4]. A design for reliability approach requires a multidisciplinary design team consisting of personnel with experience: chip design, chip-level processing, packaging, system design, manufacturability, and reliability [11]. To determine product reliability, a methodical approach to relate final product reliability to reliability at each of the earlier development and production phases is needed [4]. Another way to look at the relationship between final product reliability and the reliability at each development and production phase is to understand that there is a hierarchy from the chip reliability, to packaged device reliability, to subsystem reliability, to system reliability [11]. MEMS Design and Development Practices Figure 1 shows, in flowchart form, a general component development cycle incorporating a design for reliability approach. To design for reliability, a reliability design goal should be one of the initial design goals established [1]. The requirements for the proposed system application for the product are used to establish product design goals, including a reliability design goal. This is true for military systems. The design goals, including reliability and lifetime goals, should be based on the system requirements. A process flow, a device design, and a component design are generated after the design goals are established. Then, a Failure Mode and Effects Analysis (FMEA) is performed on the process and design [1, 11]. The FMEA approach is a disciplined, systematic procedure designed to identify potential process and product weaknesses that could lead to failure [11]. In the FMEA approach, a group of experts from various technology and manufacturing areas suggest possible failure modes, considering processing methods, design constraints, packaging concerns, test procedures, and other possible factors that could contribute to failure [1]. The group identifies the consequences, or risks, of failures [11]. Based on the possible failures and the related associated consequences, the list of failure modes is ranked according to their effect on the “customer.” This list establishes a priority system for in-process testing, developmental testing, and design improvements [12]. Simulation and analysis is performed to investigate potential failure modes and to introduce possible process and design changes. Once the simulation and analysis is completed, the predicted performance of the device and component is compared to the design goals. If the performance does not meet the design goals, one or more of the following are modified: the process flow, the device design, or the component design. The modifications are based on insight gained from the FMEA and from the simulation and analysis. Compared with approaches that do not use FMEA, the FMEA approach has enabled a faster time to market with lower risk of failures for a new device design [1]. Testing and Characterization Testing and characterization is a critical part of the design for reliability approach [1, 12]. (See Figure 1, which shows that FMEA and test and characterization are part of the development process.) However, test and characterization planning, including the development of test and characterization methods, is not stated explicitly as part of the development process. As noted above, the FMEA identifies and establishes priorities for in-process testing and characterization and for post-process developmental testing and characterization. The experience of MEMS-based component manufacturers indicates that MEMS process-, device-, and component-specific test methods combined with established, accepted test methods are needed to develop reliable components [1–3]. One example of MEMS-based process-, device-, and component-specific testing is the Deformable Mirror Display (DMD) test system developed by Texas Instruments, Inc., for testing of the DMD [1]. In addition, many test and characterization techniques have been developed by researchers in academe, in national laboratories, or in other government laboratories. These researchers typically attempt to gain a general, fundamental understanding of MEMS materials, processes, and device behavior. Testing and characterization research on MEMS materials, processes, and devices has increased over the past 5 years based on the number of presentations at conferences and symposiums sponsored by organizations such as SPIE, the Society for Experimental Mechanics (SEM), the American Society of Mechanical Engineers (ASME), and the Materials Research Society (MRS). Many researchers utilize a combination of analytical and computational techniques to develop testing and characterization tools and methods for advancing the fundamental understanding of MEMS materials, processes, and device behavior [13, 14, 15, 16]. Continued development of MEMS-based process-, device-, and component-specific test methods is needed as existing MEMS-based components are used in new applications, such as military systems