Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

The development of Micro-Electro Mechanical Systems (MEMS) and introduction of MEMS-enabled products in the market have made amazing strides in the last two decades; fulfilling a vision of “cheap complex devices of great reliability”.Footnote 1 MEMS are integrated micro-scale systems combining electrical, mechanical or other (magnetic, fluidic/thermal/etc.) elements typically fabricated using conventional semiconductor batch processing techniques that range in size from several nanometers to microns or even millimeters [1]. These systems are designed to interact with the external environment either in a sensing or actuation mode to generate state information or control it at a different scale.

In recent years, MEMS technology has gained wide-spread acceptance in several industrial segments including automotive, industrial, medical and even military applications. The size and growth of the MEMS market is typically represented in volume of a particular kind of sensor device, and in 2009 this market was roughly US$7 Billion and was dominated by pressure sensors, accelerometers, optical devices and microfluidic devices (Fig. 1.1 below)Footnote 2 and represents roughly 8–10 billion units.

Fig. 1.1
figure 1

Estimate of World Wide MEMS Market 2009 (reprinted with permission Copyright – Databeans)

Full size image

MEMS present several daunting technical challenges quite unlike those seen in typical semiconductor microelectronics which have no moving parts. In comparison, MEMS designers create a variety of different 3D structures and highly complex shapes (see Fig. 1.2 below) all at similar micrometer scales using a variety of materials. Another unique challenge in MEMS is that the end-product functionality is often tightly linked to the process used to create it leading to the “one product, one process.” This is in marked contrast to the IC industry where many products share a common process (i.e., there is no equivalent of a “32 nm node” for MEMS).

Fig. 1.2
figure 2

Polysilicon fabricated accelerometer (reprinted with permission Copyright – Analog Devices)

Full size image

MEMS process or fabrication technology has made great strides in recent years to mass fabricate at the micro-scale with a variety of materials (besides conventional semiconductor materials) using standard photolithographic processes for high volume MEMS device fabrication.

The road has not always been smooth, and evolution of MEMS fabrication technology has taken the better part of the last two decades. In this time, MEMS based products have crossed the threshold of prototype volumes into large-scale volume production. Examples such as the Freescale MPX Series pressure sensor, the Analog Devices ADXL series accelerometers, and the Texas Instruments DLP ® mirror are but a few products that have successfully achieved the performance and cost targets necessary to displace competing technologies in specific markets (Fig. 1.3).

Fig. 1.3
figure 3

MEMS development history (reprinted with permission Copyright – MMC)

Full size image

On average, each of these product development efforts lasted several years from initial concept to final volume production and market insertion, although improvements in time-to-market have been observed with successive generations of products. By far the most significant time-consuming factor in each case has been the persistence of a “traditional” manufacturing approach (Fig. 1.4 below), where the engineering of the product for volume manufacturing has gone through many cycles of learning and consequently has taken much longer than anticipated. The novelty of MEMS technology, lack of adequate design tools, a lack of “standard” process flows, the complex interaction of packaging and MEMS device, and MEMS reliability, are challenges that have hindered quicker time to market. Although DFM (Design for Manufacturing) and TQM (Total Quality management) strategies exist in almost all industries today [3], the adoption of a comprehensive design methodology that links all product engineering groups in the MEMS industry was lacking in the early days [4]. In recent years, product development methodologies for MEMS product design are grounded in powerful top-down design tools. Today, concurrent engineering practices (Fig. 1.5) have reaped benefits in terms of faster design cycles and a faster path to volume manufacture. A major challenge continues to be the reliability of the MEMS enabled product in the intended application.

Fig. 1.4
figure 4

Traditional MEMS product development cycle (reprinted with permission Copyright – Sensors Expo)

Full size image
Fig. 1.5
figure 5

Product development flow based on concurrent engineering practices (reprinted with permission Copyright – Sensors Expo)

Full size image

This book was written to aid in the improvement of MEMS product reliability by providing an understanding of the science, and best practices, as well as to document the methodology to drive improvement within a MEMS enabled product. The intent is to provide readers with a valuable reference containing a detailed description of failure mechanisms, understanding of reliability physics, lifetime prediction, test methods, and numerous examples for making improvements to reliability in all types of MEMS products. With the MEMS industry as diverse as it is today, a reference of industry best-practices is helpful in providing new product development efforts with a guide to address specific reliability challenges ahead of time or through design, thereby reducing time-to-market. The following figure (Fig. 1.6) illustrates the linkage between the topics covered in this book, and although obviously simplified, yet should allow the reader to understand the connections between the topics of MEMS reliability.

Fig. 1.6
figure 6

Reliability topics

Full size image

This book begins with this introductory chapter on the need for improved understanding of MEMS reliability issues.

Chapter 2 provides a review of reliability statistics for lifetime prediction, and includes the Weibull, Lognormal and Exponential distributions. The bathtub curve concept is presented as well as acceleration factors of physics of failure, and accelerated testing in MEMS. This chapter also introduces the reader to MEMS reliability through three basic case studies. Acceleration of hinge related creep failure in the DLP® mirror from Texas Instruments and a predictive model of mechanical shock related to stiction in an accelerometer test vehicle from Analog Devices are presented. These are examples of two very different high-volume and high reliability MEMS products. The third case study is a MEMS product with great potential that has not yet entered the marketplace due to reliability challenges; MEMTronics’ RF MEMS product is reviewed.

Chapter 3 examines failure mechanisms and modes introduced in MEMS technology and products during the design and manufacturing stages. This chapter examines the impact of design and manufacturing failure modes that have their origin in the product development phase on the reliability of the final product. These failures include both functionally analyzed and non-analyzed behaviors that are repeatedly seen to be major factors in the life of the product. The second half of the chapter presents manufacturing (or process) failures due to typical MEMS fabrication process failure modes including defects like contamination, release stiction, intrinsic and extrinsic material failures, handling and packaging failure modes.

Chapter 4 covers the physics of failures modes in the field (operational and non-operational). This comprehensive chapter covers mechanical mechanisms, electrical mechanisms and environmental mechanisms by combining theory and data. For each failure mode, mitigation techniques and performance/reliability trade-offs are presented. It should be kept in mind that each failure mechanism will have a different level of predominance depending on the device, fabrication process, and the operational environment and range of the device. Mechanical failure modes specific to MEMS are discussed in detail, including fracture, shock resistance, fatigue, and plastic deformation. Electrical failure modes include dielectric breakdown, dielectric charging, ESD, and electromigration. Environmental failures cover the effects of ionizing and non-ionizing radiation on MEMS, as well as different types of corrosion for metals and silicon.

Chapter 5 defines strategies for identifying root cause and begins with the Failure Modes and Effect Analysis (FMEA), which is an excellent tool to determining root cause and corrective action to assure that failure mechanisms are contained and eliminated. An optical switch Reliability FMEA is presented as an example to teach the reader how to use this methodology. A substantial failure analysis section that describes popular failure analysis techniques with MEMS-based analytical results is also included. A reliability program must contain strategies for identifying potential failure modes, failure mechanisms, risk areas in design and process, and corrective action strategies. Containment of the failure is of crucial importance to minimizing or mitigating the field failure rate while a proper root cause is identified, and corrective action is developed and finally implemented into production. For MEMS technologies, the use of proven methodologies (such as FMEA and failure analysis techniques) to identify potential failure modes and mechanisms are an important part of the reliability approach.

Chapter 6 contains testing and qualification processes and procedures used in MEMS, and are presented with discussions of relevant reliability. This chapter introduces for the first time the unique test equipment and reliability test methods used in the MEMS industry along with quality standards for various target industries that include automotive and military applications. Examples include test data for MEMS specific test equipment as well as qualification and reliability studies.

Chapter 7 offers a summary of the best practices to improve reliability in a MEMS product. The information in this chapter is a synopsis of much more in-depth work and readers should reference other sections of this book or articles listed at the end of this chapter for more details. The chapter discusses the yield-reliability connection specifically to MEMS products where it is common to screen parts at final test for various weaknesses that could result in potential field failures, including ones that impact life of the part. The importance of process and material property characterization [6] is discussed next as well as the use of common test structures and Process Control Monitors. In typical CMOS processes, the importance of process stability and reproducibility is quite well recognized but in MEMS the significance takes on a new dimension because of the additional specialized process steps. Common techniques for yield and quality enhancements are discussed in some depth, as the yield/reliability link has to be kept in mind at all times. Finally, the topic of Design-for-Reliability (DfR) addresses the topic of design methodology that considers probable failure of the device as part of the design process.

In summary, as this unique text was written by authors with extensive industry reliability experience, we have distilled the best practices from a variety of MEMS product development efforts to provide the reader with a clear methodology for developing a solid and comprehensive MEMS reliability program.