Time division multiplexed (TDM) plasma etch processes have found widespread applications in Micro-Electro-Mechanical Systems (MEMS) device manufacturing. Very often, silicon-on-insulator (SOI) structures are used in MEMS applications with oxide layers used as etch stop/sacrificial layers as well as device function layers. Apart from the conventional requirements for deep silicon etch including high rate, selectivity and sidewall smoothness. SOI structures require finished etches to be free of undercut, commonly referred to as notching, at the silicon/oxide interface. Notching is aggravated due to the aspect ratio dependence (ARDE) effects. The ARDE effects cause structures with different aspect ratio to be etched at different etch rates, and result in the buried oxide layer in bigger features to be exposed while smaller features are still being etched. At Unaxis USA, we have developed a proprietary technique to eliminate the notch formation while maintaining high etch rate. This technique is integrated into time division multiplexed (TDM) Si etch processes, and is implemented in a single etch process. The conventional "bulk" etch to "finish" etch transition is thus made unnecessary, with the benefit of no end point detection and smooth and uniform etch profile. Etch processes are characterized and notch performance is measured as a function over etch percentage and feature aspect ratio. Using the new SOI etching technique, notching is completely eliminated in aspect ratios up to 9:1 and reduced to well below 100 nm for aspect ratios up to 18:1. Moreover, this new technique has been demonstrated to limit the effect of extensive overetch in increasing notch size.

In typical DRIE processes, the etch rate variation across the wafer increases with pattern density, severely limiting the pattern densities that can be used at a specified etch rate tolerance. Here, we present a scheme for including uniformity-improving dummy structures in the etch mask layout that enable the use of high-density patterns in many DRIE process types. The dummy structures take up relatively little space in the layout and reduce the total etch rate variation of a 35% etchable area pattern by 66% while maintaining a high etch rate.

Deep X-ray lithography on PMMA resist is used in the LIGA process. The resist is exposed to synchrotron X-rays through a patterned mask and then is developed in a liquid developer to make high aspect ratio microstructures. This work addresses the thermal analysis and temperature rise of the mask-resist assembly during exposure at the Advanced Light Source (ALS) synchrotron. The concern is that the thermal expansion will lower the accuracy of the lithography. We have developed a three-dimensional finite-element model of the mask and resist assembly. We employed the LIGA exposure-development software LEX-D and the commercial software ABAQUS to calculate heat transfer of the assembly during exposure. The calculations of assembly maximum temperature have been compared with temperature measurements conducted at ALS. The temperature rise in the silicon mask and the mask holder comes directly from the X-ray absorption, but forced convection of nitrogen jets carry away a significant portion of heat energy from the mask surface, while natural convection plays a negligible role. The temperature rise in PMMA resist is mainly from heat conducted from the silicon substrate backward to the resist and from the mask plate through inner cavity air forward to the resist, while the X-ray absorption is only secondary. Therefore, reduction of heat flow conducted from both substrate and cavity air to the resist is essential. An improved water-cooling block is expected to carry away most heat energy along the main heat conductive path, leaving the resist at a favorable working temperature.

We have successfully demonstrated selective trapping, concentration, and release of various biological organisms and inert beads by insulator-based dielectrophoresis within a polymeric microfluidic device. The microfluidic channels and internal features, in this case arrays of insulating posts, were initially created through standard wet-etch techniques in glass. This glass chip was then transformed into a nickel stamp through the process of electroplating. The resultant nickel stamp was then used as the replication tool to produce the polymeric devices through injection molding. The polymeric devices were made of Zeonor 1060R, a polyolefin copolymer resin selected for its superior chemical resistance and optical properties. These devices were then optically aligned with another polymeric substrate that had been machined to form fluidic vias. These two polymeric substrates were then bonded together through thermal diffusion bonding. The sealed devices were utilized to selectively separate and concentrate a variety of biological pathogen simulants and organisms. These organisms include bacteria and spores that were selectively concentrated and released by simply applying D.C. voltages across the plastic replicates via platinum electrodes in inlet and outlet reservoirs. The dielectrophoretic response of the organisms is observed to be a function of the applied electric field and post size, geometry and spacing. Cells were selectively trapped against a background of labeled polystyrene beads and spores to demonstrate that samples of interest can be separated from a diverse background. We have implemented a methodology to determine the concentration factors obtained in these devices.

Spatial light modulators (SLMs) based on micromirrors for use in DUV lithography and adaptive optics require very high mirror planarity as well as mirror stability. The ideal mechanical properties of monocrystalline silicon make this material ideally suited for use in high precision optical MEMS devices. However, the integration of MEMS with CMOS poses certain restrictions on processing temperatures as well as on the compatibility of materials. The key to the successful fabrication of monocrystalline silicon micromirrors on CMOS is the silicon layer transfer process. Here, we discuss two carefully adapted wafer bonding processes that are CMOS compatible and that allow the transfer of a 300nm thick monocrystalline silicon thin film from a SOI donor wafer. One process is based on adhesive bonding using a patterned polymer layer, while the other process is based on direct bonding to a planarization layer of polished glass.

In this paper we present the very promising results for two methods of the so-called Bonding and Deep RIE (BDRIE) technology, characterised by bonding of two wafers with pre-patterned vertical gaps and subsequent RIE trench etching of the active layer. In case of the anodically bonded silicon-glass compound detection electrodes for vertical movement are integrated. The silicon layer contains the movable structure as well as drive and detection electrodes for lateral movement. It is advantageous that finally the mechanical active elements consist of single crystalline silicon without any additional layers. The BDRIE approach allows a great variation of parameters. The active layer thickness can be defined due to application issues. Our examples show active layers thickness ranging from 30 up to 200 μm, patterned by dry etching steps with maximum aspect ratio between 20:1 and 30:1. Structures with trench width variations of more than 50 (widest/smallest trench) have been fabricated successfully. Methods and results of preventing notching and backside etching of the active layer are presented as well. The size of the vertical gap can be as small as 1.5 μm for a very sensitive detection or several tens or hundreds of microns in order to reduce damping and parasitic capacitance. Holes for release in the movable structure are not necessary and will therefore not restrict the design. However, restrictions are given by the minimum size of bond area and the relation between layer thickness, free standing area above the groove and bond pressure, which are discussed within the paper. Applications of BDRIE are inertial sensors like gyroscopes, step-by-step switchgears as well as micro mirrors.

Using tightly focused femtosecond laser pulses, we produced optical waveguide and devices in the transparent materials. This technique has the potential to generate not only channel waveguide but three-dimensional optical devices. In this paper, an optical splitter and U-grooves, which are used for fiber alignment, are simultaneously fabricated in a fused silica glass by using near-IR femtosecond laser pulses. The fiber aligned optical splitter has a low insertion loss, less than 4 dB, including intrinsic splitting loss of 3 dB and excess loss due to the passive alignment of a single-mode fiber. Finally, the utility of the femtosecond laser writing technique is demonstrated by fabricating gratings at the surface and inside the silica glass, respectively.

This paper aims to contribute to the understanding of column formation mechanisms in Al₂O₃-TiC ceramic composites due to processing with excimer laser radiation. The mechanisms proposed in the literature to explain the formation of such columns can be grouped in four categories: shadowing mechanisms, hydrodynamic mechanisms, vapour phase deposition mechanisms, and spatial modulation of absorbed energy mechanisms. In the case of Al₂O₃-TiC ceramics, the hydrodynamic and vapour phase deposition mechanisms can be excluded because experimental results show that the column core is composed of material in a pristine condition. A theoretical simulation of the spatial modulation of absorbed energy due to radiation reflected from preexisting topographic artefacts reveals that this mechanism can explain the growth of columns from those artefacts, but does not explain column growth in Al₂O₃-TiC, because it predicts that the height of the columns will increase indefinitely with increasing number of pulses, whereas it has been experimentally observed that columns only grow during the first 100-200 laser pulses. This model does not explain the observed variation of the columns height with laser fluence either. By contrast, predictions of the shadowing mechanism with TiC globules formed during the first laser pulses shielding the substrate and favouring column growth are in semiquantative agreement with experimental observations. The evolution of surface topography in Al₂O₃-TiC ceramics composite during processing with KrF excimer laser radiation is controlled by the ablation behaviour of individual phases and by the chemical changes of the material surface during laser processing.

Owing to the high bonding energy, most of the glasses are removed by photo-thermal rather than photo-chemical effect when they are ablated by the 193 or 248nm excimer lasers. Typically, the machined surface is covered by re-deposited debris and the sub-surface, sometimes surface as well, is scattered with micro-cracks introduced by thermal stress generated during the process. This study aimed to investigate the nature and extent of the surface morphology and sub-surface damaged (SSD) layer induced by the laser ablation. The effects of laser parameters such as fluence, shot number and repetition rate on the morphology and SSD were discussed. An ArF excimer laser (193 nm) was used in the present study to machine glasses such as soda-lime, Zerodur and BK-7. It is found that the melt ejection and debris deposition tend to pile up higher and become denser in structure under a higher energy density, repetition rate and shot number. There are thermal stress induced lateral cracks when the debris covered top layer is etched away. Higher fluence and repetition rate tend to generate more lateral and median cracks which propagate into the substrate. The changes of mechanical properties of the SSD layer were also investigated.

Multi-photon absorption phenomena induced by ultra fast laser have been considered for many applications of microfabrications such as metal ablation, glass etching and photopolymerization. Among the applications, the photopolymerization by two-photon absorption (TPA) has been regarded as a new microfabricating method. It is possible to be used in photo mask correcting, diffractive optical element and micro machining. The TPA photopolymerization is made possible to fabricate a complicated three dimensional (3D) micro-structure which the conventional photomask technology has not been able to make. In fact, the shape of the voxel (volume pixel: a unit structure of TPA fabrication) is an important factor which could affect the microfabrication process. In this paper, we have reported that 3D micro-structures were fabricated and the generation of voxel shape was analyzed for various optical conditions.

In this paper we report a new practical method of micromachining of transparent materials involving laser-induced plasma, using a conventional Q-switched Nd:KGW laser (1,06 μm). Micromarks and relieve grating with 3 μm pitch have been fabricated in sapphire, silica and glass, including the use of a LC numerical indicator as a mask. Influence of the cross-section size of a shined area on machining results was studied. It was found, that the density and temperature of laser-induced plasma at constant energy density of laser radiation rise with the increase of the cross-section size. Laser-induced plasma influences the ablation rate of transparent materials and metals in the opposite way.

This paper presents recent advances in the HARPSS micromachining technology, which enables implementation of movable Single Crystal Silicon (SCS) structures with high aspect ratio vertical air gaps on low-resistivity silicon substrate. This is suitable for applications of micro-gravity accelerometers, low voltage tunable capacitors, and high-resolution gyroscopes that require aspect ratios as large as 100:1 to achieve high sensitivity and wide tuning range. The single-sided HARPSS process eliminates the need for double-sided processing, and wafer bonding to subsequently package the device. The device thickness and gap spacing can be varied in a wide range, 30-150mm and 0.2-2mm respectively, to select the performance range. The movable MEMS elements are made of bulk silicon substrate, resulting in higher mass, higher quality factor (Q), and better shock resistance, compared to using polysilicon (poly) as the movable structure. This process provides a mechanism for creating corrugation in SCS electrodes to reduce the Brownian noise of sensors. This technique realxes the need to reduce the noise by using the maximum available mass and through-wafer etch. The corrugations are created by a DRIE technique for etching poly surrounded by oxide inside the isolation trenches. Also, uniform capacitive gap spacings are created by growing sacrificial oxide inside the trenches.

New microfabrication technologies in the MEMS domain require novel approaches in computer aided design. Process issues in these technologies affecting the design are becoming increasingly important and Process information held in static design rule sets will be no longer sufficient. This paper describes the methodology and the implementation of a process management system that supports the designer in configuring application specific process flows with predictable properties.

Al₂O₃/ZnO alloy films were grown at 100°C using atomic layer deposition (ALD) techniques. It has been previously established that the resistivity of these films can be tuned over a wide range by varying the amount of Zn in the film. Al₂O₃/ZnO ALD alloy films can therefore be designed with a dielectric constant high enough to provide a large down-state capacitance and a resistivity low enough to promote the dissipation of trapped charges. The material and electrical properties of the Al₂O₃/ZnO ALD films were investigated using Auger electron spectroscopy (AES), nanoindentation, and mercury probe measurements. Chemical analysis using AES confirmed the presence of both Al and Zn in the alloys. The nanoindentation measurements were used to calculate the Young's modulus and hardness of the films. Pure Al₂O₃ ALD was determined to have a modulus between 150 and 155 GPa and a hardness of ~8 GPa, while the results for pure ZnO ALD indicated a modulus between 120 and 140 GPa and a hardness of ~5 GPa. An Al₂O₃/ZnO ALD alloy displayed a modulus of 140-145 GPa, which falls between the two pure films, and a hardness of ~8 GPa, which is similar to the pure Al₂O₃ film. The dielectric constants of the ALD films were calculated from the mercury probe measurements and were determined to be around 6.8. These properties indicate that the Al₂O₃/ZnO ALD films can be engineered as a property specific dielectric layer for RF MEMS devices.

Nondestructive characterization on thin films and their stack in MOEMS device is highly desirable. But, it is often a challenging task because the area is usually small. During processing of thin films, the deposition rates, optical properties, and mechanical properties must be fully understood to fabricate a device with desired performance. With the patterned surface, deposition rate of a typical physical vapor deposition (PVD) technique, such as electron-beam evaporation and sputtering, varies at different location due to shadow effect. In this study, spectroscopic ellipsometry and reflectometry were used to characterize the optical properties of electron-evaporation thin films on a flat substrate. On the other hand, microreflectometer was used to monitor the spectrum of deposited multi-stack of optical thin films inside via-holes. Combination of these two techniques provides a practical way to qualify the processing and ensure the device performance.

The interaction between laser beam and metal, and micromachining technology for small-diameter metal tubes by means of a copper vapour laser are considered. The quality and productivity of cutting have been analysed from a point of view of the laser beam intensity distribution in the treatment zone. The divergence of laser beam and its power have been selected to ensure a minimal cut wall roughness with the maximal production rate being preserved. A stent cutting technology for cardio surgery is presented. The technology equipment has been described for precision cutting of products with a cylindrical surface. Tiny devices being cut out from non-metal material

A characterization of low temperature silicon-glass anodic bonding (AB) parameters is presented here. Silicon-glass couples are bonded at temperature and voltage in the ranges of (200-430)°C and (0.2-2.5)kV, respectively. Two different electrodes are used for applying voltage, single point and planar. Low voltage, low temperature and short bonding time are investigated for different glass thickness and electrode type. The results show that the planar electrode provides a bonding time reduced to less than 5min against the few hours obtained by point electrode, and only slightly dependent on glass thickness. The bond strength of the bonded couples starts to be over the bulk glass strength at 300°C, when using planar electrode, and the high quality bond does not show voids. These results are particularly interesting in case of low temperature, and can be considered better than others presented in literature considering the simpler set-up and the novel electrode type used here. In addition, employment of above mentioned works are demonstrated for the fabrication of sensing microcomponents in lab-on-chip applications. The compatibility of porous silicon (PSi) and the very quick AB process, performed at low temperatures in order to prevent silicon pore filling with thermal oxides, is confirmed here. Satisfactory strength and bond quality was obtained at temperatures as low as 200°C, at voltages of 2500V, with process times lower than 1,5 minutes.

Giant micromirrors with large scanning deflection and good flatness are required for many space and terrestrial applications. A novel approach to manufacturing this category of micromirrors is proposed. The approach combines selective electroplating and flip-chip based technologies. It allows for large air gaps, flat and smooth active micromirror surfaces and permits independent fabrication of the micromirrors and control electronics, avoiding temperature and sacrificial layer incompatibilities between them. In this work, electrostatically actuated piston and torsion micromirrors were designed and simulated. The simulated structures were designed to allow large deflection, i.e. piston displacement larger than 10 um and torsional deflection up to 35°. To achieve large micromirror deflections, up to seventy micron-thick resists were used as a micromold for nickel and solder electroplating. Smooth micromirror surfaces (roughness lower than 5 nm rms) and large radius of curvature (R as large as 23 cm for a typical 1000x1000 um² micromirror fabricated without address circuits) were achieved. A detailed fabrication process is presented. First piston mirror prototypes were fabricated and a preliminary evaluation of static deflection of a piston mirror is presented.

Significant progresses have been made in MOEMS for display, imaging, telecommunication, and bioinstrumentation applications. This talk will first provide an overview of the recent advances in micromirror technologies. Then it will discuss three novel applications using MEMS micromirrors. First, a large port count wavelength-selective switch using a one-dimensional array of two-axis analog micromirrors will be described. Then the fabrication and packaging of two-axis micromirrors for in vivo endoscopic optical coherence tomography (OCT) imaging will be presented. Finally a new "optoelectronic tweezers" for manipulating microparticles and biological cells using direct images of MOEMS spatial light modulators will be described.

Micromechanical RF filters and reference oscillators based on recently demonstrated vibrating on-chip micromechani-cal resonators with Q's >10,000 at 1.5 GHz, are described as an attractive solution to the increasing count of RF components (e.g., filters) expected to be needed by future multi-band wireless devices. With Q's this high in on-chip abun-dance, such devices might also enable a paradigm-shift in transceiver design where the advantages of high-Q are emphasized, rather than suppressed, resulting in enhanced robustness and power savings. An overview of the latest in vi-brating RF MEMS technology is presented with an addendum on remaining issues to be addressed for insertion into tomorrow's handsets.

After billions of years of evolution, nature developed inventions that work, which are appropriate for the intended tasks and that last. The evolution of nature led to the introduction of highly effective and power efficient biological mechanisms that are scalable from micron to many meters in size. Imitating these mechanisms offers enormous potentials for the improvement of our life and the tools we use. Humans have always made efforts to imitate nature and we are increasingly reaching levels of advancement where it becomes significantly easier to imitate, copy, and adapt biological methods, processes and systems. Some of the biomimetic technologies that have emerged include artificial muscles, artificial intelligence, and artificial vision to which significant advances in materials science, mechanics, electronics, and computer science have contributed greatly. One of the newest fields of biomimetics is the electroactive polymers (EAP) that are also known as artificial muscles. To take advantage of these materials, efforts are made worldwide to establish a strong infrastructure addressing the need for comprehensive analytical modeling of their operation mechanism and develop effective processing and characterization techniques. The field is still in its emerging state and robust materials are not readily available however in recent years significant progress has been made and commercial products have already started to appear. This paper covers the state-of-the-art and challenges to making artificial muscles and their potential biomimetic applications.