ABSTRACT
As the design of Micro-Electro-Mechanical System (MEMS) devices matures and their application extends to critical areas, the issues of reliability and long-term survivability become increasingly important. This paper reviews some general approaches to addressing the reliability and qualification of MEMS devices for space applications. The failure modes associated with different types of MEMS devices that are likely to occur, not only under normal terrestrial operations, but also those that are encountered in the harsh environments of space, will be identified.
Keywords: MEMS devices, reliability, qualification, failure mode, analysis, space environments.
INTRODUCTION
Micro-Electro-Mechanical System (MEMS) devices have successfully been used in terrestrial applications for many years. Light-weight, low-cost, functionally-focused MEMS sensors and actuators promise to revolutionize space exploration in the next millennium. While the potential applications of MEMS are vast, the utilization of MEMS technologies in space missions have been limited thus far due to concerns of reliability and qualification of MEMS devices. Long-term reliability and survivability of MEMS devices for space applications require effective ground demonstration of reliable and robust operation in the hostile environment of space since they cannot be brought back to Earth for service. The establishment of qualification requirements and guidelines has been made difficult in part due to many types of MEMS devices, with different sets of failure modes stemming from different fabrication and construction techniques.
Most of the research on MEMS devices in the past couple of decades have focused on developing advanced fabrication techniques and improving their functional performance. It is only in the past few years when MEMS technologies and device performance have advanced sufficiently and the applications have become so critical that researchers have paid more attention to the issues of reliability and long-term survivability.
FAILURE MODES
Material Incompatibility:
A number of materials are used for the construction of MEMS devices. Even before evaluation of the functional performance of a design, selection of the materials and their compatibility are the first considerations that determine the reliability of the device.
MEMS devices often use silicon and other electronic materials as mechanical structures. In addition to single crystal silicon, polycrystalline silicon (polysilicon) of different impurity concentrations and grain structures is another common structural material used for MEMS fabrication. Silicon is a monocrystal, mechanically-strong material that does not show creep or exhaustion and is well-suited for MEMS elements requiring bending. Polysilicon is readily compatible with micromachining processes. Sputtered thin films and traces of various metals, such as aluminum, tungsten, platinum, and gold, are used as electrical conductors and wires. Both silicon dioxide and silicon nitride have traditionally been used for electrical and thermal isolation, masking, and encapsulation. Silica glass is also increasingly being used for this purpose. Other materials used primarily for electrical isolation are aluminum oxide and polyimide. In addition to these structural materials, there has been an increasing interest of the use of amorphous and diamond-like carbon films and diamond structures in MEMS devices.
Fracture and Fatigue:
Fracture occurs when the load on a device is greater than the strength of the material. Fracture is a serious reliability concern, particularly for brittle materials, since it can immediately or would eventually lead to catastrophic failures. Additionally, debris can be formed from fracturing of microstructure, leading to other failure processes. For less brittle materials, repeated loading over a long period of time causes fatigue that would also lead to the breaking and fracturing of the device. In principle, this failure mode is relatively easy to observe and simple to predict. However, the fatigue properties of thin films is often not known, making fatigue predictions error prone. There are several ways to avoid fracture failure from occurring. One approach is to design the device with the maximum applied stress safely below the stress at which failure occurs or use a material that has a material strength far exceed the maximum stress expected.
Fig.01: Cantilever beams stuck at the free end
Stiction:
One of the biggest problems in MEMS has been designing structures that can withstand surface interactions. This is due to the fact that, when two polished surfaces come into contact, they tend to adhere to one another.While this fact is often unimportant in macroscopic devices due to their rough surface features and the common use of lubricants.MEMS surfaces are smooth and lubricants create, rather than mitigate, friction. As a result, when two metallic surfaces come into contact, they form strong primary bonds, which joins the surfaces together. This is analogous to grain boundaries within polycrystalline materials, which have been found to be often stronger than the crystal material itself. However, adhesive boundaries are usually not as strong as grain boundaries, due to the fact that actual area of contact is limited by localized surface roughness and the presence of contaminants, such as gas molecules.
Adhesion is caused by Van der Waals forces bonding two clean surfaces together. The Van der Waals force is a result of the interaction of instantaneous dipole moments of atoms. In most MEMS devices, surface contact causes failure. When MEMS surfaces come into contact, the Van der Walls force is strong enough to irrevocably bond the two surface.
Vibration:
Fig.02 (a, b): Cracks in single crystal silicon support beams caused by vibrations from a launch test
Vibration is a large reliability concern in MEMS. Due to the sensitivity and fragile nature of many MEMS structures, external vibrations can have disastrous implications. Either through inducing surface adhesion or through fracturing the device support structures, external vibrations can cause catastrophic failure. Long-term vibration will also contribute to fatigue. For space applications, vibration considerations are important, as devices are subjected to large vibrations in the launch process.
Friction and Wear:
Many MEMS devices involve surfaces contacting or rubbing against one and other, leading to friction and wear. The operation of micromachined devices that have contacting joints and bearings is significantly affected by friction and wear of the contact surfaces involved. Friction and wear properties of materials used in the fabrication of contact surfaces must be improved for the long-term reliability and high performance.
Adhesive wear occurs when elements in a device rub together causing small pieces to rip off. These pieces attract and stick to each other, particularly in high-humidity environments, resulting in regions where micromachines get jammed up and fail. Adhesive wear was found to be a major contributor to MEMS failures. Abrasive wear is a cutting or material removal of the surface increasing its roughness. Silicon is a widely used material for MEMS fabrication. However, friction and wear properties of silicon may not be adequate for many sliding applications. The mechanical and tribological properties may have to be improved to meet the functional performance and reliability requirements. One method to improve mechanical properties and possibly tribological properties is by ion implantation.
Thermal Effects:
Temperature changes are a serious concern for MEMS. Internal stresses in devices are extremely temperature dependent. The temperature range in which a device will operate within acceptable parameters is determined by the coefficient of thermal expansion. In devices where the coefficients are poorly matched, there will be a low tolerance for thermal variations. Since future space missions anticipate temperatures in the range of -100 to 150°C, thermal changes are a growing concern in MEMS qualification efforts.
Beyond these issues, there are other difficulties caused by temperature fluctuations. Thermal effects cause problems in metal packaging, as the thermal coefficient of expansion of metals can be greater than ten times that of silicon. For these packages, special isolation techniques have to be developed to prevent the package expansion from fracturing the substrate of the device.
Another area that has yet to be fully examined is the effect of thermal changes upon the mechanical properties of semiconductors. It has long been known that Young’s modulus is a temperature-dependent value. While it is more or less locally constant for a terrestrial operating range, it may vary significantly for the temperature ranges seen in the aerospace environment.
CONCLUSION
The acceptance of MEMS devices for space and critical applications depends largely on their reliability. In this paper, the predominant failure modes of MEMS devices operating in different environments have been identified.
