This paper presents a finite element method approach for modeling a thermal-electrical-mechanical coupled-field contact comprised of an elastic hemisphere and an elastic half space. The goal of this model is to develop a fundamental understanding of the behavior of a multi-physics contact with a particular interest in the contact area. The results from the model illustrate a clear difference in contact behavior between force control and displacement control in the presence of an applied electrical potential. It is shown that, while Hertzian contact theory can be used to accurately predict the behavior of the contact under force control, a new relationship must be established to accurately predict the behavior of the contact under displacement control. This relationship can be applied to RF MEMS switch contact bumps as well as macro scale electrical contacts.

Electrical contact resistance is important to the performance of electrical switches and other current-carrying interfaces. This study investigates the behavior of electrical contact resistance for an aluminum sphere-on-flat contact as s function of current through the interface. It is observed that the contact resistance may either increase or decrease with increasing current, depending on the current level as well as the current history. At low current levels the voltage drop across the interface increases initially with increasing current until it saturates, after which the voltage level remains constant. If the current is increased beyond the value corresponding to saturation, a subsequent decrease in current yields a corresponding decrease in voltage, so that the associated current cycle shows substantial hysteresis. However, subsequent cycles of current are reversible so long as the voltage remains below the saturation point. Such behavior suggests that irreversible morphological changes occur at the interface when the current exceeds the level associated with the attainment of voltage saturation.

Microswitches have the potential to be used in numerous applications [1]. Microswitches are slower and currently less reliable than semiconductor switches. However they have the advantage of reduced power consumption, less insertion loss, and better isolation. The operation of a microswitch involves many interesting phenomena which will be reviewed. In an electrostatically actuated switch a voltage difference is applied between the beam and the gate. The resulting electrostatic force pulls the beam toward the gate such that its tip contacts the drain and completes the circuit. However the momentum of the beam causes it to deflect further, even after the tip is in contact with the drain. This stored elastic energy can be sufficient to cause the tip to separate from the drain resulting in what is often referred to as “contact bounce” [2]. Contact bounce causes several problems. First the switch cannot be used until the bouncing has stopped, thus increasing the effective time-to-closure. Second the tip and drain are subjected to more mechanical open-close contacts than are intended. Third the impact force between the tip and drain can be several times greater than the force in the closed position. These last two phenomena can prematurely degrade the contact surfaces. Attempts have been made to tailor the actuation voltage in order to reduce both the impact force and contact bounce [3]. Switches can fail in either of two failure modes. In one mode the contact resistance decreases with cycling. This behavior, while seemingly not a problem, is symptomatic of the contact being cleaned due to repeated contacts (i.e. “contact scrub” as it is known in the industry). As the resistance decreases the force of adhesion increases between these clean metal-to-metal contacts, eventually causing the switch to stick shut when the restoring force in the beam is insufficient to pull the contacts apart [4]. Another failure mechanism is a steady increase in contact resistance until the point at which the switch is no longer useable. These failures are attributable to the growth of contamination in the form of a friction polymer on the contacting surfaces [5]. Another measure of switch performance is intermodulation distortion which quantifies the distortion of a pair of harmonic RF signals with nearly equal frequencies. The beating between two such components causes the power to vary relatively slowly, i.e. at the beat frequency. This beat frequency may be sufficiently close to the thermal response time of the switch leading to distortion and false signals in the frequency range of interest [6].

Topographical changes within the contact area as a function of cycling could be a critical factor causing failure and reliability issues in RF MEMS switch operation. In this paper, gold-to-gold contact, cantilever-type RF MEMS switches were tested (cold-switching mode) for different number of cycles, namely, 10, 10 2 , 10 3 , 10 4 , 10 5 , and 10 6 . After the cycling tests, the contact area of each switch was scanned using optical microscopy, scanning electron microscopy and atomic force microscopy to quantify the exact gold-to-gold contact surface changes, leading to adhesion failures (at about 10 6 cycles). Detailed roughness analysis was carried out to better quantify topographical changes on the contact surface and relate them to failures. It was found that the material transfer from the top beam to the bottom substrate was dominant, and observed after only few cycles. Adhesion failure of gold-to-gold contact switches could be attributed to large protrusions formed on the bottom surface as the switch cycles over 10 5 times.

A numerical simulation is presented for several loading-unloading cycles of an adhesive contact between an elastic-plastic sphere and a rigid flat. The main goal of the simulation is to study the plastic deformation evolution in a contact bump material — the microscopic electrode found in a MEMS micro-switch for providing a good electric contact. This bump is subjected to a cyclic contact interaction with a harder substrate and cyclic plasticity of the bump material can lead to its wear and as result to a failure of the whole MEMS device.

In this paper, we present some current research issues in contact failures in ohmic RF MEMS microswitches. Such experimental and modeling research encompasses contact mechanics (including elastic, plastic, and adhesive contacts) and the related phenomena of brittle and ductile separation. Furthermore the contact interface is not static, but rather is known to evolve over the lifetime of the switch. Also discussed are contamination failures.

This presentation will focus on recent developments in the localized characterization of the mechanical properties of Microsystems and MEMS devices and structures. Conventional indentation techniques provide a highly powerful method for measuring the load and depth response of bulk and coated materials, but can also be used to measure the mechanical properties of very small micro-machined silicon structures. Beam structures, such as are used for accelerometers, need to be characterized in terms of the number of cycles to failure, the spring constant or the energy required to bend the beam by a required amount. Such localized testing needs to be adapted to work at various distances from the origin of the beam with a positioning accuracy of less than a micron. Initial studies have proved to be highly repeatable. A variety of examples are presented which cover a range of application areas, including accelerometer beam structures, micro-switches and printer head components. In addition, the localized testing of friction and wear in MEMS devices will be covered with some examples of the technology available and how it may be applied to such small contact areas in an accurate and reproducible way.

An innovative RF-MEMS varactor has been developed for use in distributed RF-MEMS circuits such as phase shifters and tunable matching networks where the capacitance ratio between on- and off- state must be limited to 2–10. By fabricating standoffs on the bottom side of the beam, this device eliminates the intimate contact between the FR-MEMS and dielectric, that is found in typical RF-MEMS beam capacitive switches. These standoffs limit the range of motion, allowing the capacitance ratio to be set, and also greatly reduces the contact area thus preventing stiction from occurring. The RF-MEMS switch has been fabricated with initial measurements demonstrating a capacitance ratio of 2.5. Preliminary reliability testing results demonstrate that this RF-MEMS design is very robust.

Previous models and measurements of MEMs adhesion have focused on steady-state adhesion behavior. This approach is inadequate for micromachined switches (such as RF MEMS switches) because it gives no information about the time required for contact opening, an important specification. We propose a technique to measure the switch opening time and present substantial experimental data for switches with gold-gold contacts. The data demonstrate that contact opening time increases dramatically as contact dimple area increases or as pull-apart force or contact resistance decrease. A model of opening time is also presented with model parameters that fit the experimental data.

Microelectromechanical systems (MEMS) technology employed to generate switches and relays has the potential to replace traditional solid-state devices due to broader operating temperature range, higher breakdown voltage, and much higher off-state resistance. Interest in MEMS switches and relays has surged recently, principally due to the demonstrated performance in switching radio-frequency signals. The role of surface topography in the performance and lifetime of MEMS switches and relays is the objective of this review paper. Special emphasis is given to an approach that accounts for effect of roughness and to the role of surface topography in the pull-in voltage and contact bouncing. A critical assessment of emerging topics where experimental and theoretical studies are still required in order to obtain better insight into the role of surface topography in MEMS switches and relays is presented.

Many state-of-art microelectronic, photonic and MEMS devices are based upon or created using small-scale contacts. These include, for example, high frequency, microscale electromechanical switches and nanopatterning of organic optoelectronic materials by contact adhesion, cold welding, and lift-off. The initial stages of contact occur between asperities of micro- and/or nano-scopic dimensions. As a consequence, understanding the processes that occur at the atomic level when two rough surfaces are bought into contact is fundamentally important for a wide range of problems including adhesion, contact formation, contact resistance, materials hardness, friction, wear, and fracture. The centrality of single asperities in the fundamental micromechanical response of contact between two rough surfaces has long been recognized. A wide range of experiments has shown that the conductance of small contacts changes abruptly as a function of contact size. In some cases, the conductance through individual asperities increases in a stepwise manner as the two surfaces are pressed into contact. These jumps conductance appear to be correlated with jumps in the force. The observed force-displacement relation appears to be poorly described by JKR theory during loading, while JKR provides a reasonable description of the behavior in unloading. In this presentation (see Acta Materialia 52 , 3983 (2004) for more details), we report the results of molecular dynamics simulations of single asperity contact during multiple cycles of loading and unloading at room temperature. We focus on the mechanisms by which contact deformation occurs and the relationship between contact conductance (and contact area) and the deformation. These simulations account for adhesion, elastic deformation, dislocation generation and migration, the formation of other types of defects and morphology evolution. In order to study the elastic and plastic deformation of the asperities on a rough surface, we set up a model system, as shown in Fig. 1. For simplificity, we consider a single deformable asperity on a deformable substrate that interacts with a flat, rigid plate. We calculate the conductance of the contact during loading and unloading through the modified Sharvin model [12]. To our knowledge, this study represents the first dynamic, atomistic simulation of the elastic and plastic deformation behavior of a single asperity and the corresponding evolution of the contact area and contact conductance. The present simulation results reproduce a large body of existing nano-contact experimental results, including the stepwise variation of contact area and conductance with displacement and the hysteresis in the contact radius and contact resistance versus force curves.

When two surfaces are brought together, contact occurs initially between asperities on the surface. If the mechanical loads are small, complete contact is never achieved and the behavior is dominated by asperity contact. The contact area and asperity morphology may evolve in time as a result of mechanical and capillarity (surface tension) effects, mediated by plastic deformation and/or diffusion. If a current passes through the contact, as in the case of micro-electro-mechanical switches, the evolution may be controlled by electromigration. This effect may be especially important if the voltage drop across the contact is fixed and the fractional contact area is small, such that the current is concentrated in a small number of contacts (see Fig. 1). Electromigration occurs as a result of the voltage driven electrons scattering off and imparting momentum to the atoms in the solid (see Fig. 2). Typically, the electromigration atom flux is opposite the direction of the electrical current (i.e., in the same direction as the electron flux). At small homologous temperatures (i.e., the temperature normalized by the melting temperature) and in small structures (such as asperities), atomic transport will be dominated by surface, rather than bulk, diffusion. In this presentation, we consider the evolution of an idealized asperity under the action of both capillarity and electromigration.

A model for the electrical degradation of MEMS switch metal microcontacts under many-cycle operation is presented. This model utilizes an asperity-based description of contacts and a general model for the degradation evolution. Calculations of switch resistance using this model predict a resistance evolution that shows good qualitative agreement with measurements. Implications of this model are discussed.

Cycling of a micromechanical switch with gold-on-gold contacts demonstrates that the contact resistance decreases and the adherence force increases. An experimental setup using a Scanning Probe Microscope (SPM) is allowing the fundamental physics of this behavior to be better understood. The setup includes two side-by-side cantilevers — one of high stiffness which applies the repeated loading and the other a standard profiling cantilever allowing in-situ measurements of topographical changes.

Microelectromechanical systems (MEMS) radio frequency (RF) switches hold great promise in a myriad of commercial, aerospace, and military applications. MEMS switches offer important advantages over current electromechanical and solid state technologies including high linearity, low insertion loss, low power consumption, good isolation, and low cost [1–21]. However, there is little fundamental understanding of the factors determining the performance and reliability of these devices. Our previous work investigated fundamentals of hot-switched direct current (DC) gold (Au) contacts using a modified microadhesion apparatus as a switch simulator [1]. Those experiments were conducted under precisely controlled operating conditions in air at MEMS-scale forces with an emphasis on the role of surface forces and electric current on switch performance, reliability, and durability [1]. Electric current had a profound effect on deformation mechanisms, adhesion, contact resistance (R), and reliability/durability. At low current (1–10 μA), asperity creep and switching induced adhesion were the most important observations, whereas, at high current (1–10 mA), lack of adhesion and switch shorting by nanowire formation were prominent [1].

The quality of electrical contact resistance is important to the performance and reliability of metal-contact microelectromechanical system (MEMS) relays and switches. The cyclic degradation of contact resistance was examined for a metal multilayer sphere-on-flat contact at low contact force. The relationships between the degradation and underlying mechanisms, particularly surface contamination effects, were investigated through experimentation and analysis. Results indicated that the degradation severity could be significantly decreased with environmental control and contact current reduction. Micro-arcs at very small surface gaps, which can decompose adsorbed contaminants and leave an insulating surface film, were proposed as a cause of the resistance degradation.

Micromechanical switches for RF communications often use gold as a contact material. While significant data exists for such contacts at contact forces larger than 200 mN and smaller 1 mN, little data exists at medium forces between these extremes. Moreover, previous studies have been unclear about the use of contact heating, particularly since voltages less than 100 mV can cause significant heating. This paper presents contact resistance data for both heated and unheated contacts with force from 0.01 to 1000 mN. Without heating, 1–10 mN of force is required to produce low, predictable resistance. However, with heating, forces as small as 0.05 mN produce low-resistance contacts, suggesting that contact heating is a valuable tool in removing insulating films and creating low-resistance, predictable contacts.

This study presents a basic step towards the selection methodology of electric contact materials for microelectromechanical systems (MEMS) metal contact switches. This involves the interrelationship between the two important parameters, resistivity and hardness, since they provide the guidelines and assessment of the contact resistance, wear, deformation, and adhesion characteristics of MEMS switches. For this purpose, thin film alloys of three noble metals; platinum (Pt), rhodium (Rh) and ruthenium (Ru) with gold (Au) were investigated. The interrelationship between resistivity and hardness was established for three amounts of alloying of these metals with gold. Thin films of gold (Au), platinum (Pt), ruthenium (Rh), and rhodium (Ru) were also characterized to obtain their baseline data for comparison. All films were deposited on silicon substrates. When Ru, Rh, and Pt are alloyed with Au, their hardness generally decreases but resistivity increases. This decrease or increase was, in general, dependent upon the amount of alloying.

For many applications, radio frequency microelectromechanical systems (RF MEMS) switches require very fast switching times so that they are capable of switching thousands of times each second. Contact adhesion tends to slow switches by increasing the time required to break the contact and open the switch. Previous work showed that typical switches have a contact opening time exceeding 1 ms, which is far too slow for many applications. This paper presents a dynamic, transient model to predict the time required to break the contact. The model considers both the motion of the switch body and the adhesion at the contact. The model has been validated by comparing it to experimental data. The model shows that vibrations in the switch body can be instrumental in dramatically reducing contact opening time by a factor of nearly 1000, leading to acceptable switch speeds.