This paper describes two optical devices based on linear arrays of micro mirrors. The first is a phased array of micro mirrors that can be rotated as well as translated vertically to maintain coherence across the array. We demonstrate experimentally that such micro mirrors are capable of high-diffraction-efficiency, phased-array scanning of laser beams. The second device is a Gires- Tournois interferometer with a micro mirror array that provides tunable phase modulation for the multitude of partially reflected beams within the interferometer. We demonstrate experimentally that the MEMS-GT interferometer can operate as a tunable deinterleaver for dense Wavelength Division Multiplexed fiber optic communication.

Some advantages of predicting reliability include providing advance warning signs of failure, and the reduction of life cycle costs by reducing inspection and unscheduled maintenance. However, predictions can be inaccurate if they do not account for the actual environments that the product is subjected to in its li8fe cycle. This paper describes an in-situ sensor (prognostic monitor) approach, which can be used to estimate the accumulated damage and the remaining life of semiconductor devices.

Electrostatic attraction is a favored principle of actuation in MEMS (e.g. mirrors, relays, membrane devices). In this work we use an electrostatically actuated membrane as demonstrator. Physically based device models require the coupling of the electrostatic and the two domains. One way to reduce this expense consists in reduced order modeling by introducing a local approximation of the electric field using the Differential-Plate-Capacitor-Approximation (DPCA). This semi-analytical approximation can be directly (matrix coupled transducer element) or sequentially (load vector coupling) coupled with the mechanical solver. Both approaches yield results which agree well with those of coupled 3D-field solvers. It turns out that the transducer element converges much faster than the sequentially coupled relaxation scheme, as ong as the voltage is not close to the pull-in voltage. If this is the case then the transducer element has problems to find the equilibrium state at all. To avoid this difficulty we propose the use of a homotopy method to give the transducer element the same accuracy and robustness in the stable and the unstable regions of the operating area. The electrostatic charge and the electrostatic force turn out to be proper homotopy parameters for the given example.

This paper describes a new approach for 3-D capacitance extraction using the boundary element method. The new formulation described in this paper improves condition number of the linear system to be solved and reduces the number of unknowns in the system. Implementation of the described method, in conjunction with an acceleration algorithm, will significantly reduce the computational time and memory required for the simulation of electrostatically actuated MEMS devices.

The paper presents an algorithm for the thermal simulation of MEMS problems for the cases when certain parts of the MEMS structure are given with their detailed structural description while others are given with the RC network compact models. The presented co-siumulation algorithm enables fast simulation in the frequency or in the time domain, and can be a very useful extension of any kind of field solvers. The thermal simulation of a hot-plate problem demonstrates the advantages and the simpli8city of the method.

A closed form expression for the electrical field of a line source over a multi-layered material can only be derived fro special geometries and cases. The use of the spectral domain, but this expression can normally not be transferred back into the space-domain. The solution is an approximation in the spectral domain by use of exponential terms which can easily be transferred back into the space-domain. This is done by use of Prony's method. The number of exponential terms has to be correctly estimated to avoid numerical difficulties. This paper presents a sophisticated method to determine the necessary number of exponential terms and shows how this approximation can be used to calculate inductances and capacitances for conductors with rectangular cross-sections.

Self-alignment in the fluidic phase is an alternative technique to conventional pick-and-place assembly, providing cost-effective, precise assembly of millions of microparts. For accurate alignment, the control of unwanted surface defects lowering alignment precision. Local minima are investigated and the modulation of the energy curve is simulated. Furthermore, hytsteresis effects are studied. The simulation results allow predictions for the modeling of the fluidic surface tension driven self alignment and thus provide conditions for the robustness of the fabrication process.

To improve productivity and design space exploration in MOEMS design, new high levels specification and validation methodologies are required. These methodologies have to deal with systems heterogeneity. In this paper we present SystemC based cosimulation methodology for the global validation of MOEMS which is starting from a heterogenous specification where the different modules may be described at different abstraction levels or using different specification languages.

There are considerable interests in integrating Polymerase chain reaction (PCR) on a microchip can have much fast heating and cooling rate, the delicacy in its structure makes the PCR experiment difficult and cracks often occur particularly for the thin membrane type of PCR chips. Design study and experiment of silicon PCR chips are presented with the aim of identifying the problems encountered in experiment and finding an optimum chip structure. Heating characteristics of four different heater designs have been compared, so have the PCR chambers with fixed frame and with suspended frame. The thermal stress analysis has shown that the structure and heater design can make a significant difference in heating characteristics and in reducing the failure of PCR chips. Different solutions to reduce PCR chip failure have been proposed. One of the solutions was implemented in the experiment, confirming the design study results. Silicon PCR chips have been fabricated. Thermal cycling and initial DNA amplification results are presented.

The development of Wireless Local Loop applications require low-cost technologies in the Ka-band. As transmission lines and passive circuit components on standard low-resistivity silicon substrates have high loss in this frequency range, a new microstrip MEMS technology is proposed for high- frequency applications. It has many advantages as the low cost of the substrates (glass and standard silicon), the simple technological process, the good electrical performances, and the low sensitivity to electromagnetic radiation. This technology consists in inverted microstrip lines, either on low loss silicon or on glass, which can be combined to obtain compact circuits associating active and passive components. Moreover, this technology seems suitable for current microwave and radiofrequency applications because it could be easily adapted to silicon fabrication technologies dedicated to planar circuits.

This paper reports the results of the modeling silicon microsystem flow sensor based on Thermal Time-Of-Flight (TTOF) mode. The basic heat transfer equations and the modeling approach are first presented. The problem domain is decomposed into two subdomains which represent the fluid and the sensor chip structure, respectively. The thermal boundary layer where the interaction between the two subdomains is taking place is modeled using flow-dependent equivalent thermal resistance elements. The two subdomains and the boundary layer are subsequently implemented using the combination of SPICE and analog HDL. An experimental chip of silicon thermal flow sensor is used to validate the present model. The model has been used to predict the behavior of the flow sensor in free-running TTOF mode and also in Thermal-Convection Delay-Line Oscillator (TC-DLO) mode. Both the agreement and discrepancy found between the model and the experiments are shown and discussed.

This paper presents design of novel on-chip mi9cromachined spiral inductors developed in 0.8micrometers commercial CMOS technology. Proposed innovative architecture maximizes the quality factor Q. Suspended inductors can be placed vertically above the substrate, thus, the parasitic substrate effects are considerably reduced. Four designed passive inductors have values from 1nH to 7nH. The Q-factor greater than ten is estimated from calculations, and is experimentally verified by S-parameter measurements.

It is known the micro-EDM is a proper and flexible technology to machine freeform three-dimensional microstructures. Unfortunately, its capabilities are underestimated by the ruling microsystem designers due to the lack of widespread modeling and simulation tools. In this paper the strength of micro-EDM is reflected onto a solid modeling design environment, in which all designs are parametric and feature based. On top of standard features, user defined features can be created, which are automatically assessed on their producibility. The producibility of user defined features is verified by generating a tool electrode, that is able to machine the proposed user defined features. When the tool electrode is producible by WEDG and when it passes the strength check, the user defined feature is added to the feature library of the design environment. To compensate electrode wear the required number of tool electrodes for the multiple electrode strategy is calculated using an analytical expression, which is introduced in this paper. Finally, a number of microstructures are designed and machined to illustrate the implemented micro-EDM design environment.

A methodology for combined modeling of capacitance and force 9in a multi-layer electrostatic comb is demonstrated in this paper. Conformal mapping-based analytical methods are limited to 2D symmetric cross-sections and cannot account for charge concentration effects at corners. Vertex capacitance can be more than 30% of the total capacitance in a single-layer 2 micrometers thick comb with 10 micrometers overlap. Furthermore, analytical equations are strictly valid only for perfectly symmetrical finger positions. Fringing and corner effects are likely to be more significant in a multi- layered CMOS-MEMS comb because of the presence of more edges and vertices. Vertical curling of CMOS-MEMS comb fingers may also lead to reduced capacitance and vertical forces. Gyroscopes are particularly sensitive to such undesirable forces, which therefore, need to be well-quantified. In order to address the above issues, a hybrid approach of superposing linear regression models over a set of core analytical models is implemented. Design of experiments is used to obtain data for capacitance and force using a commercial 3D boundary-element solver. Since accurate force values require significantly higher mesh refinement than accurate capacitance, we use numerical derivatives of capacitance values to compute the forces. The model is formulated such that the capacitance and force models use the same regression coefficients. The comb model thus obtained, fits the numerical capacitance data to within +/- 3% and force to within +/- 10%. The model is experimentally verified by measuring capacitance change in a specially designed test structure. The capacitance model matches measurements to within 10%. The comb model is implemented in an Analog Hardware Description Language (ADHL) for use in behavioral simulation of manufacturing variations in a CMOS-MEMS gyroscope.

High volume ICs production companies show a growing interest in MEMS components. Telecom MEMS are reaching the industrialization stage. Prototypes of integrated inductances and optical switches demonstrate very promising performances. The transition to the high volume production implies the development of Design For Manufacturability (DFM) tools featured to handle MEMS specific processes and related problems such as yield loss due to process dispersion. This paper presents an original statistical optimization method for yield enhancement. The corresponding algorithm is currently developed by MEMSCAP and LIRMM, based on response variability minimization.

A prototype of a 433.92 MHZRF receiver using a micro- mechanical resonator as channel-selection filter in the 2^nd IF stage has been realized (94.5 kHz center frequency, 20 Hz bandwidth). A novel approach is used for managing the temperature drift of the center frequency of the micro-mechanical filter. Instead of stabilizing the filter's center frequency by complex device-level and technological modifications (bias voltage tuning, mechanical or thermal compensation), we modified the architecture of the IF stage in order to continuously adapt the IF frequency to the filter's center frequency deviations. This is achieved by periodical real-time measurements of the filter's center frequency and by then generating the appropriate second LO frequency. The measurement of the center frequency is achieved by putting the filter in an oscillating closed loop. The measured relative matching error between the 2^nd IF frequency and the filter's center frequency is better than 0.005%.

A sensor designer has to keep in mind a couple of things developing a new product. These are performance, reliability and last but not least cost. Cost is mainly a function of yield. To be successful on the market, it is therefore necessary to predict yield for a new design, based on its process steps. This can be done only by modeling a sensor based on process parameters as input to the models. How to realize this, will be shown for the design of a high pressure sensor.

This paper describes how models of slender MEMS components can be generated using symbolic simulations based on Cosserat theory. Due to the generality of the method, models can be generated for a wide range of components under different stress conditions. The structure of the Cosserat- based models is presented, showing their concise, mathematically accurate representation. This potentially results in faster simulation results than finite element analysis, requiring detailed 3d meshed models. The use of Cosserat models in symbolic simulations enables generation of closed-form expressions for the dynamics of components with complex shapes. The influence of various stress conditions, such as package-induced and residual stress, on the behavior of the component can be included in these expressions.

This paper describes an approach to the simulation of optoelectronic integrated circuits based on the development of so-called element stamps for the models of optoelectronic devices. Using stamps, lumped-constant device models can be easily integrated in s a simulator, and built-in models greatly reduce the computational effort required for simulation, as compared to user-provided models written in a Hardware Description Language. As an example, the derivation of the stamp representing a quantum-well laser is described in detail, and numerical simulation results are presented.

The microsystems industry is promising a rapid and widespread growth for the coming years. The automotive, network, telecom and electronics industries take advantage of this technology by including it in their products; thus, getting better integration and high energetic performances. Microsystems related software and data exchange have inherited from the IC technology experience or standards, which appear not to fit the advanced level of conception currently needed by microsystems designers. A typical design flow to validate a microsystem device involves several software from disconnected areas like layout editors, FEM simulators, HDL modeling and simulation tools. However, and fabricated microsystem is obtained through execution of a layered process. Process characteristics will be used at each level of the design and analysis. Basically, the designer will have to customize each of his tools after the process. The project introduced here intends to unify the process description language and speed up the critical and tedious CAD customization task. We gather all the information related to the technology of a microsystem process in a single file. It is based on the XML standard format to receive worldwide attention. This format is called XML-MTD, standing for XML Microsystems Technology Description. Built around XML, it is an ASCII format which gives the ability to handle a comprehensive database for technology data. This format is open, given under general public license, but the aim is to manage the format withing a XML-MTD consortium of leader and well-established EDA companies and Foundries. In this way, it will take profit of their experience. For automated configuration of design and analysis tools regarding process-dependant information, we ship the Technology Manger software. Technology Manager links foundries with a large panel of standard EDA and FEA packages used by design teams relying on the Microsystems Technology Description in XML-MTD format.

This paper describes the design, modeling and simulation of an acoustic microsystem in a pulse-echo ultrasonic application. The microsystem, used as emitter and receiver of ultrasonic signals, consists of a bulk micromachined suspended membrane. During emission, the membrane is placed in an oscillator loop and is thermally actuated at its resonance frequency (approximately equals 40 kHz). This frequency is slightly dependent on the membrane average temperature. The electronic interface circuit monitors this temperature. The membrane oscillations generate an ultrasonic signal (pulse) that propagates in the air and interacts with a solid body. As a result of this interaction, the ultrasonic signal is reflected on the solid surface and is received by the microsystem (echo). During reception, a piezoresistive bridge placed on the membrane is used for monitoring the membrane deflections. The resonance frequency of the membrane is tuned to the emitted frequency by keeping the membrane at the same temperature, achieving then maximum sensitivity. This paper presents in detail the behavioral modeling and simulation of the complete system. Some MEMS parts and the acoustic waves propagation are modeled using an Analogue Hardware Description Language (Verilog-A). The associated electronics are implemented in CMOS and the overall system is simulated with the SpectreHDL simulator in the CADENCE environment.

The necessary information for geometric modeling of MEMS consists of a mask layout file and a process definition file. CIF and GDSII are commonly accepted formats for the mask layout file. However, there are no standard ways to express the process definition file. In this paper, we propose an XML (Extensible Markup Language) data model to describe the process definition file for geometric modeling of surface-micro-machined MEMS. This model has been incorporated in our CAD system. The results of implementation show that the XML approach not only facilitates the development of applications but also helps the exchange of information.

We present a novel approach to produce a micromachined low cost hotplate gas sensor with reduced number of technology steps. The basic idea was to realize a simple device on common silicon substrates using conventional photolithography, sputtering and evaporation techniques. Two main performance parameters were targeted: the power consumption should not exceed 200 mW for an operation at 350 degree(s)C-400 degree(s)C and the thermal response time should be faster than 1 second. Fast thermal time constants allows the operation of device in temperature pulse mode. The first step of the development was the theoretical determination of the power consumption of the micromachined substrates, even temperature distribution on the sensitive area and sufficient mechanical stability. For this we build models describing the thermal behavior of the devices by means of the finite element method (FEM) and corresponding resistance-capacitor-networks (RC-network). Then we developed technological processes to fabricate sensor structures according to the optimal geometry resulting from the model calculations. A first prototype is introduced in this publication.

We propose a novel method for pull-in analysis in electrostatic actuation. In this method, pull-in angle can be found once the capacitance of electrodes is calculated as function of rotation angle. This is based on our theory, which is that the pull-in angle only depends on the capacitance but not on spring constants or applied voltage. The theory is derived analytically and its limit of application is described as well. This theory and method can be applied to the translational motion case in the similar fashion. With the proposed method, the computation time can be reduced considerably since it deals with only one domain rather than executing coupled-domain analysis. This method can be used more effectively where the complexity of electrode structures or spring shape is more severe. By way of example, it is applied to the design of three different types of actuators: parallel-plate torsion mirrors, staggered vertical comb-drives, and scanning micromirror with hidden vertical comb-drives. The theoretical results are compared with experimental data as well.

On-chip electronics performing weak-signal recovery for CMOS BDJ (Buried Double p-n Junction) optical detector is proposed. It includes two identical channels for simultaneous processing of both detector's output signals in continuous-time configuration. Each channel consists of a transimpedance amplifier, a fully differential amplifier, a multiplier and a low-pass filter, thus performing low-noise preamplification and synchronous demodulation. Some key building blocks have been designed with performance optimization. In order to validate the proposed architecture, to verify the system operation and to estimate its performances and characteristics, system-level simulations have been carried out. It has been evaluated that, in a typical case, the integrated system can detect an input optical signal of 21 fW/mm².

In this work, we propose a new technique to improve the transient response of the comb drive actuator using a digital compensation technique. This technique is mainly based on the use of a specific digital pulse train that satisfies the equal area criterion on the force displacement diagram of the actuator. Applying this technique to a specific actuator design, we show that it increases the forward displacement without sticking, reduces the overshoot, and improves the speed of response.

Manufacturing test structures of microsensors and microactuators is very expensive in terms of time and materials. In a conventional design process, this limits the number of design variants to be considered. For this reason, computer supported design techniques are becoming more and more important in microsystems technologies. The modular structure of hybrid systems requires single components to be manufactured in isolation and later combined into one total system. Combining single components into one overall system is bound to be subject to certain tolerances. The concept presented in this paper is the computer-aided design of a modular system rugged enough to be employed in mass fabrication. In mass fabrication, it is not the ideal arrangement of individual components which results in the most effective system. Instead, tolerances in position individual optical elements need to be taken into account already in modeling. Furthermore, environmental influences like e.g. variations of the temperature can have an impact on the performance of the optical function module.

The planar pellistor is a thin film microsensor used for the detection of inflammable gases. In its simplest, it consists of a metallic meander structure that serves both as a heater for the catalyst and as a temperature sensor. For low power operation thermal insulation between the sensor and the substrate is vital and is achieved by a thin silicon nitride membrane. The optimization of such a structure in terms of the sensor's dimensions is discussed.

Computer aided MEMS optimization regarding performance, power consumption, and reliability is an important design task due to high prototyping costs. In the MEMS design flow, a variety of specialized tools is available. FEM tools (e.g. ANSYS, CFD-ACE+) are widely used for simulation on component level. Simulations on system level are carried out with simplified models using simulators like Saber, ELDO, or Spice. A few simulators offer too-specific optimization capabilities but there is a lack of simulator-independent support of MEMS optimization. The paper presents a modular approach for simulation-based optimization, which aims at a flexible combination of simulators and optimization algorithms by partitioning the optimization cycle into separate modules for model generation, simulation, error calculation, and optimization. Available optimization algorithms include direct and indirect methods as well as stochastic approaches. Interfaces to the simulators ANSYS, ELDO, Saber, MATLAB, and SPICE are implemented. Thus the optimization task can be solved on different levels of model abstraction (FEM, ordinary differential equations, generalized networks...). A graphical user interface (GUI) supports control and visualization of the optimization progress. The modules of the optimization system may communicate via the internet (web-based optimization, distributed optimization).

The way of how an efficient and reliable design strategy for Microsystems should be set up is discussed and put into practice. Within this design strategy automatic design optimization plays a decisive role. The model based design optimization system MODOS is presented which has especially been developed for the development of Microsystems. Finally, the proposed design strategy is utilized to optimize a micro mechanical pressure sensor with integrated readout circuitry.

The application of optimization algorithms is especially interesting in areas, where the design process needs several iterations, every single one costing time and money. This is particularly the case with the design of Microsystems. The disadvantage in practice is usually the construction of the quality function, with which the quality of the optimal design is measured. This holds especially true whenever more than one optimization goal is pursued, because every goal needs to be quantified as part of an overall quality function. Especially the user of the optimization system is often at a loss when asked to specify and quantify the single goals' contributions to the overall quality function. This paper demonstrates the problem using a radiometric sensor as an example and examines several computer assisted quality function approaches to transform the multi criteria optimization problem into a single goal problem.

In this work, multi-physics simulation software (CA/MEMS) and design-optimization software (DS/MEMS) tailored for MEMS devices are introduced. The CA/MEMS, which is a simulation engine for DS/MEMS, is a 3-D multi-physics analysis code utilizing various numerical methods such as FEM, BEM and FVM to efficiently model MEMS application problems. The current CA/MEMS includes analysis- modules for structural, thermal, electric, electromagnetic and fluidic fields and is capable of the analyses of various coupled- field problems for MEMS applications. DS/MEMS is design optimization engine for MEMS devices. With integrating CA/MEMS and pre/post processor into CAD environment, DS/MEMS is organized to work in parametric CAD platform. DS/MEMS consists of optimal design module and robust design module. The optimal design module provides users three methods nonlinear programming, Taguchi parameter design and the response surface method. The robust design module, which is specially developed for MEMS application, can be used to minimize the perturbation of performances of MEMS devices under uncertainties of MEMS devices, such as process tolerance and the change of operating environments. To verify the efficiency and accuracy of CA/MEMS and the practical usefulness of DS/MEMS, we have been comparing the simulated results of CA/MEMS with those of other commercial codes and experimental data of manufactured MEMS devices, and investigating the performances of the optimized designs through DS/MEMS.

A prototype contact type micro piezoresistive shear-stress sensor that can be utilized to measure the shear stress between skin of stump and socket of Above-Knee (AK) prosthesis was designed, fabricated and tested. Micro-electro-mechanical system (MEMS) technology has been chosen for the design because of the low cost, small size and adaptability to this application. In this paper, the Finite Element Method (FEM) package ANSYS has been employed for the stress analysis of the micro shear-stress sensors. The sensors contain two X-ducers that will transform the stresses into an output voltage. In the developed sensor, a 3000X3000X3000 micrometers (superscript 3/ square membrane is formed by bulk micromachining of an n-type <100> monolithic silicon. The piezoresistive strain gauges were implanted with boron ions with a dose of 10(superscript 15/ atoms/cm(superscript 2/. Static characteristics of the shear sensor were determined through a series of calibration tests. The fabricated sensor exhibits a sensitivity of 0.13mV/mA-Mpa for a 1.4N full scales shear force range and the overall mean hysteresis error is than 3.5%. In addition, the results simulated by FEM are validated by comparison with experimental investigations.

Currently, hard disk drives (HHD) use rotating disks to store digital data and magnetic recording heads are flying on the disk to read/write data. The recording heads are mounted on a slider- suspension assembly, which makes heads move from one track to another on the disk. The heads movement is controlled by close-loop feedback servo system. It is well known that dynamic behaviors of head-slider-suspension-assembly (HSA) system are of great influence on the track per inch capacity of HDD^1,2. As the problem is structurally complex, it is usually investigated using experimental methods or finite element simulation models ³. Furthermore, the dual-stage servo system, that is, a conventional VCM as the primary stage and a MEMS actuator as the secondary stage for MEMS device embedded HAS, has resulted in more difficulties in predicting HDD dynamic performance. This paper presents studies of the problem using macromodeling simulation approach. It applies efficient FEM based sub-structuring synthesis (SSS)⁴ and fast boundary element method (BEM) approaches incorporated with system dynamics technology to investigate dynamic characteristics of MEMS actuator embedded HAS system for HDD.

The design and simulation of a new optical add-drop multiplexer (OADM) made of two pairs of grating-frustrated directional couplers is presented. Design theory and simulation are introduced to support is feasibility. The directional couplers are proposed to use polymeric waveguides as the cores. The inverted-ridge waveguide fabrication is proposed using lithography and etching process. The proper dimensions and parameters of each waveguides for OADM are simulated. The designed OADM is 1550 nm wavelength with add-drip filters, bandwidth 20 nm, and 80% peak transmissions showing the application for wavelength division multiplexing (WDM).

The Buried Double Junction (BDJ) detector [1], which can be used either as a wavelength-sensitive device and as a photodetector, can be implemented in standard CMOS IC technologies, with no requirement for additional post process step. It has recently been applied to fluorescence detection [2]. The wavelength-sensitive operation of the CMOS BDJ detector is based on the wavelength dependence of the Silicon absorption coefficient a(X) in the visible range. An absorption length defined as 1(X) varies monotonically, from about 0. 1pm to several micrometers when the wavelength of an incident monochromatic light changes from 0.4 jim to 0.8 jim.

This paper reports on JPL's on-going research into MEMS gyroscopes. [1-4] This paper will describe the gyroscope's fabrication- methods, a new 8-electrode layout developed to improve performance and performance statistics of a batch of six gyroscopes (of the 8- electrode design) recently rate tested. Previously in our group, T. Tang and R. Gutierrez presented the results of their extensive use of ethylene diamine pyrocatechol (EDP) to deep-etch the inertial- sensitive r4esonators and post-supporting structures in a 4- electrode gyroscope design. Today, JPL is utilizing an in-house STS DRIE, replacing the old wet-etching steps. This has demonstrated superior precision in machining symmetry of the resonators, thus significantly reducing native rocking mode frequency splits. A performance test of six gyros has shown an average, un-tuned, frequency split of 0.4% (11Hz split for rocking modes at 2.7KHz). The new JPL MEMS gyroscope has a unique 8-electrode layout, whose large electrodes can provide significant electrostatic softening of the resonator's springs. This allows matching of the Coriolis sensitive rocking modal frequencies to be improved from the native 0.4% to an average tuned frequency split of 0.02%. In separate tests, electrostatic tuning in the 8-electrode design has demonstrated the ability to match frequency-splits to within 10mHz, thus ensuring full degeneracy in even a very high Q device. In addition, a newly selected ceramic package-substrate has improved the device's dampening loses such that a mean Q of 28,000 was achieved in the six gyroscope tested. These Q's ere measured via the ring-down time method. The improved fabrication development and other modifications described have led to the JPL's MEMS gyroscope achieving an average bias instability (Allan variance 1/f floor estimate) of 11degree/hr with best in the group being 2degree/hr. In an independent test, Honeywell Inc. reported one of our MEMS gyroscopes as achieving 1degree/hr bias instability flicker floor estimate measured at constant temperature.

Vibration monitoring has become an important mean for wear state recognition at cutting tools, bearings, gears, engines and other highly stressed machine components [1], [2]. The majority of mechanical vibration used to identify the wear state is found in the frequency range from several Hertzto 10kHz [2]. At present, vibration measurement systems are usually based on wideband piezoelectric sensors completed with sophisticated analyzing electronics to observe the spectrum. Because of high costs permanent monitoring is only practicable in safety related applications or at extremely expensive machinery. Future developments in the field of vibration measurement equipment are expected to lead to "smart" sensors with fully digital interface, self-test functionality and on-board storage [3]. In this paper we present a frequency selective capacitive sensor for vibration detection with electrically tunable band selectivity fabricated using near-surface silicon micromechanics (SCREAM). The selectivity is based on the mechanical resonance of the structure, the center frequency of which is variable by direct electrostatic stiffness modulation. This represents a capability of resonance frequency tuning by a control voltage to adjust the measurement range to the desired value. Linearity of the sensor characteristics has been achieved by an optimization of the detection and tuning comb capacitors including FEM-analysis and MATLAB optimization algorithms. By grouping sensor structures with stepped base frequencies into an array the frequency range can be largely extended. To be used as a measurement system the sensor array requires a control unit such as a microcontroller. It handles tasks like cell selection, AD-conversion of the conditioned measurement signal and the generation of the tuning voltage. This measurement system will match the idea of a "smart" sensor since it includes calibration, self test and a digital interface to the outer world. Furthermore it can store data and even draw decissions based on the measured data and in- system algorithms for fault detection.

Modem status of technique of new generation and "critical" technologies in a number of the most developed countries is characterized by Microsystem technology dynamic development from the beginning of the 90-th years ("microsystems technology " - MST) that determined the creation of a new scientific and technical direction - "microsystem engineering " ("mirosystems engineering " MSE).

The Transmission Line Matrix method (TLM) provides an easy to use and fast scheme to calculate acoustic fields. Thus it is possible to couple this method to system simulation tools in order to investigate the behavior of acoustic systems and the interaction between acoustic domain and control electronics at an early design stage. These investigations are very important in MEMS design, especially for ultrasonic devices, transducers and SAW filters. As a proof of concept an active noise control system was designed and the performance was simulated in a 2D acoustic environment. The results clearly show the feasibility of acoustic field calculations along with system level simulations of electronic devices and transducers. Using this scheme, acoustic devices may be simulated and optimized under realistic acoustic conditions.

The potential due to a distribution of the sources and normal dipoles over a flat panel is evaluated. The described method gives a more intuitive and generalized approach to the potential computation. In particular, it provides a formulation for the potential which does not depend on the coordinate system and which is much easier to implement than previously known methods.

The large variety of the technologies and physical, chemical or biochemical effects united in micro systems requires a new generation of quality assurance. An appropriate quality management should check all conceivable influences on the product and perform a complete tolerance synthesis based on its function and with regard to manufacturing and cost optimization. In the following, a new approach to function-oriented tolerance analysis and synthesis is presented. The Institutes for Engineering Design and Micro Technology in Braunschweig, Germany, have developed a computer program, which assists Design Engineers in tolerance analysis and synthesis in micro technology. This contribution presents the theoretical background of this software. For a better understanding two short MEMS examples are shown.

In the development of Microsystems, FEM simulators are used to investigate the behavior of system components with high accuracy. Generally, FEM simulations are time consuming. System-level models of all components are needed to allow a fast but sufficiently exact investigation of the system behavior to simulate entire microsystems. Typically, microsystems consist of nonelectrical components and electronic circuits. Providing models for electronic components and languages t describe the behavior of nonelectrical subsystems, simulators like Eldo, Saber, and VHDL-AMS simulators become more and more popular in the development of Microsystems. For simple structures such as mechanical beams, models of microsystem components can be derived from analytical descriptions. Another possibility to consider more complex structures is to use FEM descriptions to generate models for system simulation. Some FEM simulators like ANSYS allow access to the numerical values of the system matrices. They are established based on the description of geometry and material data. Usually, these system matrices are very large (10,000 up 10 100,000 system variables or more). For system simulation, models with about 10 up to 100 variables are often required. Therefore, methods for order reduction are applied to derive smaller system matrices. An improvement of an order reduction method based on a projection method is introduced in the paper. Using the reduced systems, behavioral models in languages like MAST, HDL-A OR VHDL- AMS can be generated automatically. The described method was applied successfully to simulate mechanical microsystem components on a system level.

Circuit simulation is important for circuit design communities. Relaxation-based algorithms have been proven to be faster and more flexible than the standard direct approach used in SPICE. Signal flow analysis of the simulated circuit is very important in using Relaxation-based algorithms. However, there is no specific research undertaken for it. This paper discusses signal flows in circuit simulation, which gives two definitions for the strength of signal flow (SSF), discusses how to calculate SSF, and proposes techniques to utilize SSF in one of the Relaxation-base algorithms, ITA (Iterated Timing Analysis). Experimental examples on digital as well as analog circuits are given to prove the value of exploiting SSF in circuit simulation.

In cooperation with the industrial partner STEAG microParts GmbH and the Karlsruhe Research Center, the authors at the University of Bremen have created a behavioral model for the microsystem 'medical test strips'. As there is no complete theory for the underlying system yet, with capillary interaction between up to three different materials in the microfluidical/micromechanical world, the flow of liquids in such an application cannot be described analytically. Medical test strips made from plastics are an innovative approach toward cheap and therefore seminal one-way diagnostic systems and can be considered as step towards the long discussed 'lab on a chip' as well. The modeling is based on simulation results from the Finite Volume Method (FVM), yielding partial information for the investigated structures by solving partial differential equations. As this numerical approach cannot provide results for a complete channel, due to calculation efforts, a behavioral model for optimization purpose has been created on a higher abstraction level.

We present models of two types of microsprings namely box- spring and zig-zag spring that can be used to measure the force generated by microactuators. The spring constant for both springs is calculated by FEM using ANSYS software. In these models, the effects of short beams that act as connectors in the spring structures are considered and analyzed by changing their width. Also, from the results, we find that the box spring appears more balanced than the zig- zag spring when the force is applied in the single central direction. A series of SDAs with b ox spring have been fabricated and forces of those SDAs have been calculated.

Reducing design cycle time is a main concern of CAD tools. MEMS designers particularly suffer from a long and difficult design procedure in which different phases are involved: solid modeling, meshing and simulation. In this paper, we present, MEMS Max, the new MEMSCAP environment fully oriented towards RF-MEMS designers. It brings to the designer flexible, complete and easy-to-use RF MEMS-oriented tools. At the same time, it reduces efficiently the design cycle time.

Until recently, MEMS design was only handled by experts. This situation resulted from the lack of efficient and easy-to-use integrated tools covering the complete MEMS system design procedure, from individual MEMS component design to complete system simulation. MemsMaster, provided by MEMSCAP, implements a new design methodology for MEMS prototyping. The underlying mechanism to achieve such an objective is the adoption of a library concept supported by a nodal analysis approach. A library is a means of encapsulating of predefined parameterized elements in a standard way so that users can assemble them to design their own systems. Finite element analysis (FEA) methods offer high efficiency and are widely used to model and simulate behavior of MEMS components subject to multiple coupled physical phenomena. However, finite element models may involve large numbers of degrees of freedom (variables) so that full simulation, especially in the case of transient analysis, can be prohibitively expensive. As a consequence, designers must simplify models or limit available results in order to obtain accurate but fast results. Based on Reduced Order Modeling (ROM) MemsModeler, also proved by MEMSCAP, allows MEMS designers to easily and automatically generate a representative behavioral model of a MEMS component from a multiphysics finite element model. These behavioral models are typically well suited to enrich MemsMaster libraries.

In this paper, we report on feedback interferometric measurements on a micromachined gyroscope and on a micromachined linear accelerometer. Characterization has been performed for different values of pressure and of other parameters using a laser diode. Resonance frequencies and quality factors have been measured. Moreover, hysteresis and other nonlinear phenomena on specific samples have also been detected. The proposed method is based on optical injection and represents an efficient alternative to the standard electrical measurements, which actually shows some limitations for bare prototype testing.

Mechanical oscillators that have a stable resonance with high quality factor have applications as reference oscillators, sensors and in even very sophisticated high- precision experiments for observing quantum effects. In order to obtain a high quality factor the mechanical energy dissipation has to be minimised. At atmospheric pressure the most significant loss mechanism is gas damping, but for an oscillator working in vacuum the major part of the mechanical energy losses is caused by the coupling to the support structure and by internal friction which in turn is a result of a variety of physical mechanisms like thermoelastic effects and phonon scattering. Thus a great attention should to be paid to the material choice and the oscillator design.

This paper presents an innovative device for self-parking in a v-groove and a self-latching vertical mirror on the suspension diaphragm using the out of plane fiber-optical switch array technique. The self-parking offers integrating the optical fiber and mirror within the same optical switch. The self-latching vertical mirror is supported on the suspension diaphragm by four cantilever beams. The theoretical analysis includes a dynamic simulation using the ANSYS software and corner compensation using the IntelliCAD software. The fabrication process consists of wet etching mircromachining, lithography, and excimer laser ablation. This proposed process is simpler than those proposed in other works. An electrostatic driving voltage is used to operate the optical switch. The mirror is made of a photoresist coating with gold film as the switching element. The reflectivity of the gold film mirror is higher than 85% using a wavelength of 1310nm. The micro-optical switch has a maximum of displacement of 48 micrometers and the switching time is below 0.4 ms with a driving voltage of 100 VDC.

We report on the combination of the well established 1.55 micrometers monolithic VCSEL's concept with the Micro-Opto-Electro- Mechanical System (MOEMS) technological breakthrough in order to develop a novel tunable laser device for wavelength division multiplexing optical systems. Technological issures are presented for fabricating surface micromachined InP-based tunable VCSELs.

In this paper, we present a new concept of tunable active Micro-Opto-Electro-Mechanical System (MOEMS) microdevice for specific applications in the 1.3-2.5 micrometers wavelength range such as near infrared spectroscopy or optical telecommunications. The proposed optical structure can be used for the realization of tunable wavelength selective devices (photodetectors or emitters). The device uses a radically new optical design which separates the detector (or emitter) from the filter but place it on top of the filter. As compared to the existing micro-mechanical tunable devices, this concept does present two main advantages such as the improvement of the optical spectral response (forward and backward traveling of the optical waves through the active part) and relaxation of the technological constraints for fabrication (planar monolithic integration of the active component with post-process micromachining). We present here the design, the optical simulations and the fabrication procedure of a first demonstrator consisting in an optically pumped wavelength selective and tunable light emitting diode. The gain active region comprises InAs quantum wires designed for light emission aro7und 1500 nm. The MOEM structure is made of InP/air gap layers. We have obtained an increase of the spontaneous emission by a factor of about 40 and tuning range about 60 nm for actuation voltages up to 15 V.

Design, microfabrication, and integration of a micromachined spatial light modulator ((mu) SLM) device are described. A large array of electrostatically actuated, piston-motion MEMS mirror segments make up the optical surface of the (mu) SLM. Each mirror segment is capable of altering the phase of reflected light by up to one wavelength for infrared illumination ((lambda) equals 1.5 micrometers ), with 4-bit resolution. The device is directly integrated with complementary metal- oxide semiconductor (CMOS) electronics, for control of spatial optical wavefront. Integration with electronics is achieved through direct fabrication of MEMS actuators and mirror structures on planarized foundry-type CMOS electronics. Technical approaches to two significant challenges associated with manufacturing the (mu) SLM is discussed: integration of the MEMS array with the electronic driver array and production of optical-quality mirror elements using a metal-polymer surface micromachining process.

This paper presents a high-speed resolution phase-only microelectromechanical system (MEMS) spatial light modulator (SLM), integrated with driver electronics, using through- wafer vias and solder bump bonding. It employs a polysilicon thin film MEMS technology that is well suited to micromirror array fabrication and offers significant improvement in SLM speed in comparison to alternative modulator technologies. Vertical through-wafer interconnections offer scalability required to achieve 1M-pixel array size. The design, development, fabrication and characterization of a scalable driver integrated SLM is discussed. Pixel opto- electromechanical performance has been quantified experimentally on prototypes, and results are reported.

Variable optical attenuators (VOAs) have wide applications in DWDM optical communication systems, for example, equalizing the power levers of different wavelength channels, flattening the gain of optical amplifiers, etc. A MEMS variable optical attenuator with fibers connectorized has been developed using proprietary drawbridge structure, which has achieved 1.5 dB insertion loss, 45 dB dynamic range and 37 ms response time, and requires only 8 V driving voltage. Finite element model and analytical model have also been studied and compared with the experimental data, showing that the two models predict the mechanical characteristics of MEMS VOA with reasonable accuracy.

Optoelectronic subsystems are becoming increasingly important to reduce the costs of assembly and packaging. The mechanical properties of vapor-deposited thin films can be used to advantage; for example three-micrometer thick silicon nitride microclips to hold single mode optic fibers in place in silicon V-shaped grooves. This paper describes the proposed use of pairs of thin film microcantilevers to precisely locate an optic component such as a filter or a mirror in an optical bench. In this configuration the precision of the lithographic process for the cantilevers determines the exact location of the component in the package, and to first order the etched shape in the substrate is unimportant. Simulation software based on variational principles has been developed to examine the behavior of structures undergoing large-scale elastic deflections. The design software consists of spreadsheet front end to enter parameters, and then Visual Basic (VBA) code and Frontline 'solver' software to run simulations. The fabrication process is described for 5 micrometers thick silicon carbide beams which are then tested by bending them using a surface profiler (such as a Dektak) to deflect the fifty micrometer wide silicon nitride cantilevers through large angles. The possible consequences for more efficient optoelectronic packaging are briefly assessed.

MEMS devices may exhibit delicate structures sensitive to damage during handling or environmental influences. Their functionality may furthermore depend on sealing out the ambient or being in direct contact. Stress, thermal load or contaminations may change their characteristics. Here packaging technology is challenged to extend from microelectronics towards MEMS and MOEMS. Today's approaches typically rely on housing the miniature devices in bulky ceramic or metal casing or putting them into very high volume production-, benefiting from the microelectronics packaging infrastructure. Alternatively, besides focusing on available technology, device manufacturers develop individual packages for their product typically at high cost. While there is nowadays a good infrastructure for MEMS realization from universities to MEMS foundries, packaging still remains as a bottleneck at the end of the design cycle, sometimes stopping a device from being commercialized. Selecting the proper packaging method may tip the scale towards a product success and a product failure. Choosing the right technology therefore is not only a marginal work package but also a crucial part of the product design. Three approaches to be applied for MEMS/MOEMS devices will be presented and highlighted by examples. Single die packaging, die-to-wafer processes as well as wafer level packaging options are detailed with their individual benefits and challenges. Mechanical, fluidic and optical aspects are reflected in the package technologies selected for the individual examples presented. Accelerated testing of the packaged devices under actual conditions may also be a stumbling block; one example on an optical microsystem will showcase this issue, highlighting system response under different environmental loads.

Damage, fatigue and failure of advanced electronic packages of MEMS and related systems are often caused by their growing use under harsh environmental conditions as well as extreme temperatures. Consequently, their thermo mechanical reliability becomes more and more one of the most important preconditions for adopting it in industrial applications. Various kinds of inhomogeneity, residual stresses from several steps of the manufacturing process along with the fact that microelectronic packages are basically compounds of materials with quite different Young's moduli and thermal expansion coefficients contribute to interface delamination, chip cracking and fatigue of solder interconnects. Subsequently, numerical investigations by means of nonlinear FEA, fracture mechanics concepts are frequently used for design optimizations using sensitivity analyses. Parameters used for such sensitivity analyses are typically materials parameters, geometrical and physical boundary conditions but, especially, the influence of geometrical design parameters on the thermo-mechanical reliability is more and more asked for. This paper intends to demonstrate and discuss advantages and needs of using fully parameterized finite element modeling techniques for design optimizations on the basis of damage evaluation and fracture mechanics approaches. For improving that method, the evaluation of mixed mode interface delamination phenomena and thermal fatigue of solder joints were combined with experimental investigations using a gray scale correlation method. Some results of examinations with respect to different meshing techniques and rules should help to clarify their influence on FEA results. So, the combination of numerical and experimental investigations should give a reliable basis for understanding and evaluating failure mechanisms for arising the thermo-mechanical reliability of microcomponents.

One of NASA's challenging projects for advancing the exploration of space is the development and deployment of the Next Generation Space Telescope (NGST) for superseding the existing Hubble Space Telescope. The NGST will be equipped with several camera/spectrometer systems including a 0.6 to 5 micron Multi-Object-Spectrometer. To selectively direct light rays from different regions of space into the spectrometer, an option is to use individually addressable micro-electro-mechanical-mirror arrays serving as the slit mask for the spectrometer. The NASA team at Goddard Space Flight Center has designed an integrated micro-mirror array/CMOS driver chip that can meet the system requirements. The fabrication and testing of prototype chips have yielded promising results. To build the entire MEMS- based slit mask, a design requires accurate placement and alignment of four large (at least 9 cm X 9 cm) pieces of the integrated chips in a 2X2 mosaic pattern. In addition, the mask will have to function at temperatures below 40 K. These requirements pose a serious challenge to the packaging of these integrated MEMS chips. In this paper, we discuss a concept for attaching and aligning the large- area MEMS chips into the 2X2 mask and interconnecting it to the rest of the system. The concept makes use of the flip-chip technology to bump-bond the large chips onto a silicon substrate such that the concern for global thermo- mechanical stresses due to mismatched coefficients of thermal expansion between chip and substrate is eliminated. It also makes use of the restoring force of the solder bumps during reflow to self-align the chips. A critical experiment involving the use of 'mechanical' chips with two-dimensional arrays of bonding pads was carried out to evaluate the feasibility of the packaging concept. Preliminary results indicate that the chips can be attached to form a closely packed mosaic pattern with a relative tilt angle between the chips to less than 0.05 degree, which is within the system specifications. Modeling results of the thermo-mechanical stresses gave small distortion as a result local CTE mismatch between the solder bump and silicon when the package is cooled from the solder reflow temperature down to 40 degrees Kelvin.

Technical challenges and directions for microoptical and optoelectronical packaging include the development of special handling tools for manual and automated assembly stations. These tools not only have to be precise, but also reliable in a mass production environment. The design process for handling devices therefore is a major issue and an approach to it will be presented. Several examples illustrate how the proposed design algorithm was used to solve assembly tasks like fiber handling and microlens- positioning.

In this paper, we report the development of a new 1X4 micro optical switching device which utilizes electrostatic actuation and vertical silicon mirrors. This device is fabricated using a bulk micromachining process, which allows the fabrication of vertical mirrors and U-grooves through deep reactive ion etching (DRIE) of silicon. A limited number of process steps are required in the fabrication. Moreover, the device is patterned in a single lithographic step. A relatively high yield (up to 70%) is achieved during the microfabrication due to this compact process flow. More importantly, a small footprint (<13mm² in die size) is realized. A single mode fiber with a tapered end is placed into a U-groove and positioned passively by a fiber stopper, prior to adhesive bonding with a silicon substrate and a glass cover. Preliminary characterization on the mechanical and optical performance of this device has been carried out, which reveals the promising characteristics of this 1X4 optical switch for use in optical networks.

A gas flow sensor has been developed for home-appliances applications. The main requirements were to obtain a low cost single device able to work in the range 0-1slm with high linearity, low power, reliability and robustness. A thermal flow sensor has been designed with the help of thermal and flow FEM simulation for the design of the sensor chip as well as its packaging. The process flow is based on a simple silicon micromachined technology. Chip-on-board solution has been selected to simplify the packaging. Electronics for driving the sensor and for compensation offsets and temperature dependence and for linearising the output signal has been implemented. Final device shows good sensitivity and linearity in different zones of the range of interest.

This paper presents an overview of the current technology development activities of the MEMS Technology Group at JPL> The group, in collaboration with other research groups at JPL and outside institutions, pursues the development of a wide range of MEMS/NEMS technologies that are primarily applicable to NASA's needs in the area of robotic planetary exploration. The broad classes of technologies being developed include inertial guidance devices, micro- propulsion devices, and adaptive optics for telescope applications, micro-instruments and nano-mechanical resonator devices. End-to-end prototype development of these MEMS/NEMS technologies is conducted at JPL's state-of-the- art Microdevices Laboratory. The group is also pursuing the establishment of a rapid, space-testing program in collaboration with the Aerospace Corporation, in an effort to overcome the traditional barriers to the insertion of new technologies into space missions.

We present in this paper a novel tactile fingerprint sensor composed by a single row of microbeams realized by the way of front side bulk micromachining from a standard CMOS circuit. When the user passes his finger on the sensor, the ridges and the valleys that compose the fingerprint induce deflections in the different microbeams. Using a piezoresistive gauge placed at their base, the deflections can be detected by means of a resistivity change. In addition of the MEMS part, this sensor includes in the same substrate the electronics control that allows to scan the row of microbeams and to amplify the signal from the gauges. A first prototype has been implemented and tested. This sensor dedicated to pixel tests includes three different rows composed by 38 microbeams that allow us to obtain a fingerprint image width of about 2 millimeters (spatial resolution of 50 micrometers i.e. 508 dpi).

Microsystems are exposed to thermo-mechanical loads causing strains and stresses in the used materials because of the mismatch of thermal expansion. To protect microsystems against thermo-mechanical loads and environmental impacts, polymer encapsulation is widely used. Thus the polymer encapsulation substantially influences the reliability of the component. In order to forecast the reliability of such a component, the damage mechanical behavior of the polymer encapsulation material has to be investigated. At a microscale thermally induced stresses lead to initiation of microcracks and microvoids. Damage mechanics is applied to describe the progressive deterioration of the material. Within damage mechanics the state of the damaged material is characterized by a damage measure. Basic concepts of damage mechanics are outlined in this paper. As the used polymer material shows viscoelastic material behavior, we characterize the viscoelastic material properties by a relaxation experiment. Simulation of the established material model is compared with experimental data. As simulation and experiment show good agreement, this material model can be used for further damage analysis. The accumulation of damage in the polymer material is measured by a uniaxial tension test with repeated unloading. The experiment shows monotonically increasing damage of the material, which reflects the irreversibility of damage. Comparison with a simulation of undamaged viscoelastic material behavior shows, that the measured damage accumulation is a significant effect.

In this paper, the response to the first harmonic component (2f) of the electrostatic force in single terminal driven electrostatic comb-drive and parallel-plate drive was used as a signal to extract device parameters, namely, the Q- factor and resonant frequency instead of the fundamental (1f) resonance response. It is shown that the difficulty in motional measurement due to electrical cross-talk (parasitics) using 1f measurement can be overcome with a higher signal-to-noise ratio of the 2f signal. Both atmospheric (low-Q) and reduced pressure environment were investigated using off-chip electronics and lock-in amplifier. The measurements were done on the electrostatic comb-drive and capacitive parallel plate sensing plates that form the two core modules of a yaw rate sensor (dual-axis resonator). The effects of AC and DC bias voltages on the measured response have been investigated. Experimental amplitude and phase response data have been analyzed using the Lorentzian curve-fit, Resonance Curve Area (RCA) method, the half-power bandwidth method (3dB) and the Nyquist plot for data fitting and determination of the Q-factor and resonance frequency.

One of the authors has proposed an electrostatically driven torsional resonator with two degrees of freedom (TDF). The main characteristic of the TDF resonator, in which the electrode gap does not directly affect to inconsistencies between low voltage driving and a large range of motion, is reported. The TDF structure is also beneficial for achieving high Q values. However, size reduction was difficult because of the limitations of the fabrication process. In this study, the TDF resonator is miniaturized by the UV-LIGA (UV exposed lithography and electroplated structure) process. The process and the frequency characteristics of the resonator are reported.

Atomic Force Microscope operating in contact mode is used in this paper for probing the spatial distribution of adhesive forces versus the topography of a silicon nanotip. This nanotip consists in an ultra sha4rp silicon tip with radius less than 15 nm fabricated using a combination of high- resolution electron beam lithography and plasma dry etching. The amplitude of the forces is determined from force versus distance curve measurements. Hence, by determining the contact point and the pull-off force from the force curves, the surface topography and the adhesive forces are simultaneously obtained at various locations on the surface. This paper reports both measurements and the modeling of adhesive forces versus the contact point on the nanotip. As the nanotip is sharper and has got a smaller aperture angle than the employed Atomic Force Microscope tip, the measurements are focused on the nanotip apex.

Waferbonding techniques are frequently used for MEMS/MOEMS fabrication. In this paper, the potential application and methodical limitations of different strength testing approaches including tensile testing and double-cantilever- beam testing for wafer-bonded components are investigated. Special attention is given to the influence of the interfacial atomic bonding strength, the role of interface voids and notches caused by chemical or physical etching steps prior to bonding on the fracture limit. A particular aim of the paper is to discuss the potential of the Micro Chevron-Test for the assessment of the wafer bonding process with particular respect to the quality control during MEMS fabrication. In addition, the methods can also be applied to investigate the lifetime and fatigue properties of wafer- bonded samples exposed to constant or cyclic stresses.

RF switches are believed to replace PIN diodes and MESFETs in numerous future RF applications. But most of the actual applications require high reliability and long lifetimes for their devices. As MEMS is a new technology, aging tests and qualification procedures have yet to be demonstrated. MEMSCAP and CNES are developing an environmental test bench for the study of RF switch failure modes. This paper focuses, in particular, on the influence of the temperature in metallic RF switches/ Actually, architectures, such as the metallic air bridge, the membrane switch and the dielectric switch, display good RF performances. We will show in this paper that most of today's switches are sensitive to buckling. In particular, a few tens of degrees Celsius are enough to create a deformation that drives the air bridge switch (in ON or OFF position) to fatal failure. The influence of tensile pre-stress is also studied since it increases the buckling critical temperature. However, the required pre-stress range will degrade significantly the actuation voltage. Finally, stress relaxation structures are believed to decrease the sensitivity to high temperature while keeping a reasonable actuation voltage.

An original design of a polysilicon loop-shaped microheater on a thin stacked dielectric membrane is presented. This design ensures high thermal uniformity and insulation (20.000 degree(s)C/W) and very low power consumption (20 mW for heating at 400 degree(s)C). Moreover, the use of CMOS-IC compatible TMAH-based bulk-micromachining techniques will allow easy and low cost gas sensor integration.

Silicon has been widely used as the material of choice for the fabrication of MEMS I MST devices. The compatibility of MEMS manufacturing equipment with standard IC equipment presents one ofthe main reasons for this choice. However, over the past years, we have seen new equipment dedicated to MEMS fabrication enter the market place. One such example is the Deep Reactive Ion Etcher, which is capable of vertically etching silicon at a rate of several microns per minute. This type of equipment, now available from several vendors, has revolutionized the MEMS fabrication capabilities and has opened the door to a whole new family of MEMS devices [1].

A new, robust method for double sided processing bulk micromachined components made of an SOI (silicon on insulator) material is presented. Transfer of the SOI film to a second wafer using a bonding step enables double sided processing on 'flat' surfaces of the device film of the SOI wafer, thus avoiding lithography in deep cavities, high temperatures and high voltage processing steps. Room temperature oxygen plasma assisted wafer bonding was used to transfer the SOI film. This bonding step is expected to be CMOS compatible, which, makes it possible to integrate standard process electronics with micromachined devices on the wafer level. Before the bonding step the processing of the SOI and of the second wafer put no requirements on common compatibility, i.e. CMOS compatibility. In addition to double sided lithography and room temperature oxygen plasma assisted wafer bonding, the fabrication process makes use of the CMOS compatible anisotropic TMAH (tetramethyl ammonium hydroxide) etch and etch back technique. Fabrication of bulk micromachined accelerometer structures has shown that the process is successful.

This paper reports on a low-cost low-temperature (2. High-Q inductors and capacitors integrated on low-loss organic package substrates can find numerous RF and microwave SOP applications (such as VCO, IF/RF bandpass filters, LNA, etc., in which IC chips are flip-chip mounted on the package substrate. Integration of passives in organic substrates eliminates the excess cost of assembly, enables miniaturization by allowing for added functionality at the board level, and provides an attractive alternative to higher temperature processes such as ceramic and deposition technologies for high-Q passive implementation.

Hot embossing allows directly integrating conduction paths made of gold in the channel structures required for applications in lab-on-chip systems. In experiments, ditch depths of more than 100 micrometers wide and 2-3 micrometers thick construction paths. It turned out to be of no relevance whether the inclination of the lateral walls was 45 degree(s) or 90 degree(s).

Rapid laser-beam reflowing turned out as an effective method for improving the structure based characteristics of solders and joints all the more with the application of higher melting contemporary Pb-free solders. The basic investigations on laser-beam induced rapid reflowing processes and the generation of grain-refined, restructured solder joints and metallization systems have been carried out on several unleaded alloys. The experimental tests have been concerned with structure, hardness; shear strength, wettability and long-term behavior. Grain refining and higher strengths of solder joints have been detected as the most substantial results after rapid reflowing. Leads that were laser-beam soldered under rapidly reflown solder showed about 10-40% higher shear strengths than in cases of using non-reflown solder. The maximum shear strengths have been attained by using rapidly reflown SnAg3, 8CuO, 7Sb0,2. Higher strength appeared in consequence of grain refining and increasing hardness, but also of structure stabilization due to Cu diffusion. Structure characteristics and strength did not weaken even after extended TCT.

Packaging and electrical interconnection are an important part of the fabrication path of MEMS and MOEMS. This article shows one example for alignment, packaging and interconnection using anodic bonding in combination with vertical v-grooves for precise mechanical alignment. Stacks of twelve alternating layers of borosilicate glass and silicon have been fabricated with an alignment accuracy of better than +/21.5micrometers . Furthermore, a new anodic bonding method is presented that avoids any sodium compound formation at critical areas of a device.

A low-loss RF package has been realized using silicon- micromachining technique. As the frequency of the device increases, the loss caused by packaging increases. Therefore, low loss high frequency packaging is needed. For frequency above 10GHz, a metal package is used but a metal package has spurious resonance caused by interaction between radiation of RF circuits and cavity geometry. To suppress spurious resonance, microwave absorbers are deposited inside the metal package; it is very hard to process with these materials in microscale. In this work, high resistivity silicon is simulated as a high frequency packaging material, and is realized using silicon deep etching process. And metal-filled thorough-hole via is formed to provide surface mount type interconnection using sandblast and copper electroplating process. The silicon package has reflection loss of 15dB, insertion loss of 0.7db from 500MHz up to 40GHz region.

This paper aims at providing a detailed analysis of the consequences of shocking a packaged microstructure to a solid surface. Particularly it concerns the response of homogeneous, packaged clamped-clamped polysilicon microfabricated beams, which can be used in the case of MEMS switches. The theoretical analysis is composed of three parts; first, the micro-machined is considered as a single degree of freedom oscillator the motions of which is governed by an ordinary differential equation. It allows calculation of the deflection induced by the shock. Secondly, we determine the deflections inducing a stress of 1.5Gpa corresponding to the destruction of the polysilicon structures, for small and large deflections considerations. Finally, from the two first parts, pre-shock conditions (velocity and acceleration) are found to avoid stiction, impact and destruction of the packaged microstructures. It is shown that these pre-shock conditions depend on structural properties and damping. Moreover, a shock can generate decelerations ranging of millions of g's.

The purposes of failure analysis are to localize the failure site, to establish the type of failure and to understand the failure mechanism. Some MEMS, such as Analog Device's accelerometer ADXL50¹, are confronted with stiction problems. Either we get around the problem by creating designs, which are not subject to this type of failure, or we try to understand its mechanism. The residual stresses have an important role in the stiction because according to the configurations of the structures, they can stiffen beams and so oppose this failure, or amplify it by bending it towards underlying mechanical elements. That's why we carry an interest in the residual stresses to understand and counteract this type of failure. Besides, the residual stresses have also repercussion on properties such as fatigue, fracture, friction, which are at the origin of other types of failure. In this article, we attempt to set up a method to qualify and quantify the residual stresses in the different layers of the MUMP's process. The role of the residual stresses will be studied through the dimensions of the layers, their types, their coating temperatures to afford a feedback between the designers and the manufacturers. We will use a failure example in MEMS, the variable capacitance, to introduce the impact of the residual stress, on the nominal working of the mechanisms. To understand the failure of the capacitance, we chose to build test structures made of cantilevers to show the influence of the residual stresses of the different layers of multilayers. So the paper will be dedicated to the implementation of test structures and theoretical elements appropriate for revealing and determining the failure due to residual stresses.

Detecting thermal and mechanical defects within multilayered microstructures is an important research area within the microdevice community. The detection of material flaws, mechanical damage, and packaging irregularities is often critical to the overall performanc eof the end product. The technique presented hereafter uses a series of surface temperature measurements, generated by a step function heat flux, to determine the thermal properties of a one- dimensional structure. These properties can either be used directly in a design effort, or they can be used as an indicator of problems that may exist within the structure. This technique is essentially non-invasive and it places no requirements on structure size, thus it is predisposed to semiconductor and MEMS applications. The technique exploits a thermal-electrical analog to match a measured thermal resistance pattern with the pattern of a corresponding thermal structure. Typically, the dimensions of the structure and the disturbance amplitude are required for property value determination.

A rapid and accurate static and quasi-static method for determining the out-of-plane spring constraints of cantilevers and a micromachined vibratory sensor is presented. In the past, much of the effort in nanoindentation application was to investigate the thin-film mechanical properties. In this paper, we have utilized the nanoindentation method to measure directly some micromachined device (e.g. microgyroscope) spring constants. The cantilevers and devices tested were fabricated using the MUMPS process and an SOI process (patent pending). Spring constants are determined using a commercial nanoindentation apparatus UMIS-2000 configured with both Berkovich and spherical indenter tip that can be placed onto the device with high accuracy. Typical load resolution is 20micrometers N to 0.5N and a displacement resolution of 0.05nm. Information was deduced from the penetration depth versus load curves during both loading and unloading.

Thermo-mechanical reliability in advanced electronic, MEMS and MOEMS packaging requires additional material testing approaches. Namely, the necessary understanding of the impact of very local material stressing on component reliability leads to the need of material testing and characterization on microscopic and even on nanoscopic scale. E.g., defect initiation and propagation in multilayer structures applied in electronics, MEMS and MOEMS technology, the influence of material migration to mechanical behavior or defect development in ultra thin silicon dies often are not well understood. A key for micro material testing and characterization is the measurement of strains and displacements inside microscopic regions. Correlation techniques (e.g. micro DAC, nanoDAC) are one of the promising tools for that purpose. There application potentials to micro testing for packaging materials and components are demonstrated in the paper. More in detail approaches to CTE measurement, analysis of moisture-induced strains in polymers and crack testing are discussed. Furthermore, it is shown how the method can be used to study the mechanical response of complex micro components to thermo-mechanical loading.

A new technique to fabricate arbitrarily shaped microstructures by using LCD (liquid crystal display) real- time mask is reported in this paper. Its principle and design method are explained. Based on partial coherent imaging theory, the process to fabricate micro-axicon array and zigzag grating has been simulated. The experiment using a color LCD as real-time mask has been set up. Micro-axicon array and zigzag grating has been fabricated by the LCD real-time mask technique. The 3D surface relief structures were made on pan chromatic silver-halide sensitized gelatin (Kodak-131) with trypsinase etching. The pitch size of zigzag grating is 46.26micrometers . The caliber of axicon is 118.7micrometers , and the etching depth is 1.332micrometers .

In this paper, we present a new method to improve the image quality and resolution of photolithography by filtering in fractional Fourier domain. Introducing a filter into fractional Fourier domain can not only increase the flexibility of the filtering operation, but also enhance the image quality and the depth of focus in photolithography. The corresponding simulation results are illustrated.

The full-field interferometry is very well suited for evaluation of micromechanical and material properties of microsystems. In this paper, we presented a Twyman-Green interferometer for MEMS/MOEMS testing. The measurements of out-of-plan displacements of special silicon membranes with thin film of SiO_xN_y deposited by PECVD enable the analysis of opto=mechanical properties.

In this research, the manufacturing processes of non- spherical refraction microles array by gray-scale mask is investigated. Compared to the conventional multi-lithography fabricating method, the gray-scale mask approach requires only a single lithography action to fabricate a non- spherical refraction microlens array. In the firs part of this research, we emphasize the gray-scale mask based microlens array fabrication processes through the UV-LIGA approach. Furthermore, a two-stage process-modeling scheme is proposed to reduce the time-consuming trial-and-error parameters tuning labor works. At the first stage, a multi- layer feedforward neural-network is employed to model the relationships between the diameter and height of the microlens are obtained, the surface profile of this microlens can be predicted by an empirical equation. The empirical equation is derived through the regressing analysis method with data points sampled from the real microlens surface profile. Experimental results demonstrate that the proposed two-stage scheme can precisely predict the surface profile of the gray-scale mask based microlens.

In this work, we describe a front-side Si micromachining process for the fabrication of suspended membranes for thermal sensors. Membrane release is achieved by means of lateral isotropic etching of the bulk silicon substrate, the etching being optimized for high rates and high selectivity with respect to the membrane material and the photoresist that is used to protect the device. Lateral Si etch rates of the order of 6-7 micrometers /min have been achieved in a high- density F-based plasma, which permits a reasonable etching time for the release of the membrane and the simultaneous formation of the cavity underneath that ensure thermal isolation of the final device.

A PMMA-based reactive ion etching (RIE) process for the fabrication of high aspect ration microstructures is described in this paper. Although the resolution of this process is lower than that of the LIGA process, this process provides a simpler way to get higher height and high aspect ration microstructures. In the process, Ni material is selected as mask and patterned using photochemical etching. The self-bias, which is determined by different etching parameters such as etching power and gas pressure etc., is very important. By optimizing the etching parameters, vertical PMMA profile and 5:1 aspect ratio structures can be obtained. The micro-mask effect and high power etching damage are also demonstrated and discussed in this paper.

This paper describes LISA (Lateral isolated Silicon Accelerometer) technology developed by IME< Singapore and its application on silicon vertical optical switch fabrication. Key processes in LISA technology for optical switch fabrication include deep trench etch and oxide refill to enable insulating anchors in silicon substrate, second deep trench etch to fabricate movable microstructures and metal layer covering for switch surface improvement. In this paper, deep trench (deeper than 35 um) oxide refill process is introduced, the dielectric characteristic of the isolation is evaluated, and more than 100V breakdown voltage is obtained, which is much higher that the requirement in optical switch driving voltage. Some process issues related to high aspect ratio trench etch and release such as notching on silicon beam top and sidewall are shown and discussed, a double spacer process is utilized accordingly to solve the issues. Besides, a mask free metal coating process is presented to improve the mirror surface and light reflectivity. The vertical optical mirrors fabricated by the LISA technology is 35um in height and um in width, the switch displacement is larger than 40um under 35V DC bias, the optical characteristics of the switch is under testing.

This paper presents results on optical cross-connect switches based on novel MEMS vertical mirrors. The switch consists of two mirror arrays to redirect optical beams from an input fiber array to an output fiber array. Each mirror is actuated by two electrostatic comb drive actuators, and can be rotated bi-directionally and perpendicularly to the chip surface. Finite element model (FEM) and Gaussian beam optics have been used to simulate and optimize the optical cross-connect switch architecture. Results have shown that the switch is much less constrained by the scaling distance of light propagation as the port count grows. However, the coupling efficiency is sensitive to angular alignment for large port-counts; thus mechanism for ensuring precise angular control of the micro-mirror is crucial for the proposed MEMS optical cross-connect switches.

Thermally activated bimorph actuators suitable for a range of MOEMS applications, that provide large deflections (up to approximately 200micrometers ) at low voltages and low power, are being developed at the Rutherford Appleton Laboratory (RAL) UK using state of the art microfabrication techniques in a variety of materials. Interdigitated bimorph cantilevers using structural materials (polyamide and gold) with different coefficients of thermal expansion (CTE) and sandwiched microheaters, curl-up from substrate due to residual stress induced during the fabrication process. The initial deflection is approximately 300 micrometers in the vertical plane and 150micrometers in horizontal plane. Operating frequency of up to 50 Hz was achieved with moderate power dissipation (4-5MW) and approximately 50micrometers displacement in vertical direction.

The complex task of measuring forces of interaction between a pliable firm fuel and a rigid wall of the rocket engine is practically important in order to define maximum service time of the rocket engines and to prevent their destruction because of redistribution and loading change. Therefore, it is necessary to repeat measurements receiving the reliable information with an aim of revealing change of the interaction forces.

This Paper presents the design, simulation and fabrication of a single-bridge capacitive switch and a double-bridge reflectively tuned capacitive switch for X-band applications. These switches possess low loss, high isolation and low pull-down voltage. The switching element consists of a thin metallic membrane suspended over the center conductor and fixed at both ends to the ground conductors of a Coplanar Waveguide (CPW) line. The switches have two states, actuated and unactuated, depending on the applied bias voltage between the metallic membrane and the bottom electrode. The serpentine folded hinges are used to get higher inductance so as to make the resonant frequency to be at X-band. The single-bridge switch achieves over 27 dB isolation with an insertion loss of 0.2 0.1 dB at X-band. The largest isolation. 37 dB is obtained at 10.5 GHz. The double-bridge reflectively tuned switch contains two single metal membrane separated by a short length of high-impedance transmission line. The reflectively tuned switch provides an isolation of 51 dB at 10 GHz with an insertion loss of 0.5 dB. At X-band frequency range, its isolation exceeds 42 dB with an insertion loss of 0.5 ±0.1 dB. Besides, the serpentine folded hinges provide a very low spring constant resulting in 3.6-volt pull-rn voltage. The static mechanical model predicts the effective stiffness constant and the pull-in voltage. Deformation of the bridge and its contact behaviour with the dielectric layer are also precisely analysed using Finite Element Method. Finally this paper discusses the fabrication of the RF capacitive switches, which combines bulk- micromachining and surface-micromachining techniques. The thermal oxide layer (buffer layer) is patterned to be discontinuous so as to decrease the attenuation loss.