http://ieeexplore.ieee.org TOC Alert for Publication# 84 2012 February 10 21 1 C1 C4 43 21 1 C2 C2 58 21 1 1 3 34 $muhbox{V/K}$, and the power generation output achieves up to 0.56 $muhbox{W/cm}^{2}$ with a temperature gradient of 18$^{circ}hbox{K}$. $hfill$[2011-0233]]]> 21 1 4 6 396 $hfill$[2011-0237]]]> 21 1 7 9 292 $hfill$[2011-0205]]]> 21 1 10 12 565 $hfill$[2011-0149]]]> 21 1 13 22 1621 $hfill$[2010-0349]]]> 21 1 23 33 854 $ hfill$[2011-0118]]]> 21 1 34 43 923 $ muhbox{m}$ have been fabricated from clay-reinforced polyimide. Several mechanical tests were performed on solvent cast solid structures to find the optimum clay percentage in the polyimide that would lead to the highest compressive strength. The fabricated needles were tested for robustness, and it was observed that the needles were capable of withstanding on average compressive loads of up to 0.32 N. The suitability of the microneedles for skin penetration and drug delivery was demonstrated by injection of fluorescent beads into a skin sample.$hfill$[2011-0200]]]> 21 1 44 52 802 $hfill$[2011-0130]]]> 21 1 53 61 949 $^{circ}hbox{C}$. In this study, we investigate the detailed influence of process parameters on the binding condition and hydrophobicity of the perfluoropolymer surface. Even when the postannealing temperature was reduced to 145 $^{circ}hbox{C}$ , by employing 50-nm-thick Cu as an etching mask, it was possible to form highly hydrophobic perfluoropolymer patterns having a contact angle of 110.3$^{circ}$ . The hydrophobicity of the formed surface could be restored to its original value by postannealing at 145 $^{circ}hbox{C}$. This was possible because the depth of the damaged perfluoropolymer surface that contained unsaturated bonds was shallower when a Cu mask was used than when an Al mask was used. The developed technology can be employed in biochip applications such as cell patterning and droplet generation.$hfill$[2011-0076]]]> 21 1 62 67 812 $muhbox{m}$ are reported. Two microsieve architectures are explored: 2.1 and 5.3 $muhbox{m}$ thin dielectric layer membranes supported by high aspect ratio rib features (10 : 1), and 50 $muhbox{m}$ thick silicon membranes from silicon-on-insulator wafers with no supporting ribs and high aspect ratio pore features (16 : 1). The flow throughput of each design is evaluated experimentally and theoretically, and the expected relative robustness is assessed and compared to typical microsieve structures reported in the literature. The experimental and theoretical work suggests that both structures have the potential for higher robustness than the typical micromachined microsieve architectures, with reduced but still reasonable flow throughputs in the range of $10^{5} hbox{to} 10^{6} hbox{L/m}^{2}hbox{-hr-bar}$, depending on the microsieve porosity. Both microsieve architectures are shown to block monodisperse polymer beads 3 $muhbox{m}$ in diameter.$hfill$[2011-0127]]]> 21 1 68 75 990 $11 hbox{mm} times 6 hbox{mm} times 3 hbox{mm}$. Bistable operation is shown for differential pressures up to 300 kPa for gas $(N_{2})$ at high flow rates of 2200 sccm.$hfill$[2011-0195]]]> 21 1 76 84 962 $hfill$ [2011-0143]]]> 21 1 85 99 2078 $muhbox{m}$ to 28 $mu hbox{m}$), the distances between columns (from 20 $muhbox{m}$ to 100 $muhbox{m}$), and the capping layer thickness (from 3 $muhbox{m}$ to 7 $muhbox{m}$). The packages are tested at different pressures up to 12.5 MPa (125 bar). Failure points agree well with the mechanical model. The largest package fabricated is a square package of 300 $muhbox{m}$ side length and with four columns (10 $muhbox{m}$ diameter) in the middle. It withstands a pressure of 10 MPa with a thin SiN capping layer with a thickness of 6 $mu hbox{m}$. Moreover, the packages are carried through grinding, dicing, and transfer molding, demonstrating that the presented thin-film encapsulation approach is robust enough for comm- rcial first-level packaging.$hfill$[2011-0038]]]> 21 1 100 109 1039 $hbox{mm}^{2}$ and different sealed-in pressures ( $sim$100 kPa and $sim$ 2 kPa) are demonstrated. The use of a porous cover on top of the release holes avoids deposition inside the cavity during sealing, but leads to a sealed-in pressure of approximately 100 kPa, i.e. atmospheric pressure. Vacuum ($sim$2 kPa) sealing has been achieved by direct deposition of a sealing material on the SiGe capping layer. Packaged functional accelerometers sealed at around 100 kPa have an equivalent performance in measuring accelerations of about 1 g compared to a piezoelectric commercial reference device. Vacuum-sealed beam resonators survive a 1000 h 85$^{circ} hbox{C}$/85%RH highly accelerated storage test and 1000 thermal cycles between $-40 ^{circ}hbox{C}$ and 150 $^{circ}hbox{C}$.$hfill$[2011-0131]]]> 21 1 110 120 2001 $hfill$[2011-0060]]]> 21 1 121 131 1471 $^{rm TM}$ process producing no apparent detrimental effect due to outgassing. The bumping processes were successfully applied on Al-deposited bond pads of $100 muhbox{m} times 100 muhbox{m}$ with an average bump height of 101.3 $muhbox{m}$ for Ag and 184.8 $muhbox{m}$ for solder (63Sn, 37Pb). DPAT for solder paste produced bumps with the aspect ratio of 1.8 or more. The average shear strengths of Ag and solder bumps were 78 MPa and 689 kPa, respectively. The electrical test on Ag bumps at 794 $hbox{A/cm}^{2}$ demonstrated reliable electrical interconnects with negligible resistance. These scalable FLI technologies are potentially useful for MEMS flip-chip packaging and 3-D stacking.$hfill$[2011-0191]]]> 21 1 132 144 1564 $hfill$[2010-0298]]]> 21 1 145 160 1258 $^{circ}hbox{C}$. All analytical models are verified through finite-element analysis and, for the first time, experimentally validated for deflections over 120 $muhbox{m}$ and temperatures above 200 $^{circ}hbox{C}$. We start by extending a multimorph model originally proposed to analyze piezoelectric actuators, to also consider thermal and residual (postfabrication) strains under large-displacement conditions. The model is experimentally validated through measurements conducted on $hbox{SiO}_{2}$/Ti/Au cantilevers fabricated on silicon wafers. The average error between the measured and simulated displacements is less than 4% and 3% at the beam midpoints and tips, respectively, for the entire range of 20–250 $^{circ}hbox{C}$ . Capacitance measurements conducted to over 200 $^{circ}hbox{C}$ show an average deviation from the macromodel of 6.4% over the range of 20–213 $^{circ}hbox{C}$. The standard deviation for capacitance error is 5.4%, and the maximum error is 15.3%.$- hfill$[2011-0093]]]> 21 1 161 170 1001 $Q$ -factor) and in the same time allowing increased cavity length $L$ . Our approach consists of providing a solution to the main loss mechanism in conventional FP cavities related to the expansion of the Gaussian light beam after multiple reflections inside the cavity. The first design is based on all-silicon cylindrical Bragg mirrors, which provide 1-D confinement of light. In addition to wavelength selectivity, the first design also demonstrates its potential for a new class of applications, including wavelength selective extinction through mode-selective excitation, where the fiber-to-cavity distance is used as the control parameter. The second design is based on cylindrical Bragg mirrors combined with a fiber rod lens to provide a complete solution for 2-D confinement of light. This approach outperforms the first design in terms of $Q$-factor, of nearly 9000 for around 250 $muhbox{m}$-long cavity, which suggests its potential use for biochemical sensing and analysis as well as cavity enhancement applications requiring high $Q.L$ values.$hfill$ [2011-0162]]]> 21 1 171 180 1152 $hbox{cm}^{-1}$ and 334 $hbox{cm}^{-1}$, respectively, corresponding to a measured finesse of 5.$hfill$[2011-0027]]]> 21 1 181 189 886 $hfill$[2011-0183]]]> 21 1 190 197 1061 $muhbox{J}$ of kinetic energy delivered to the robot and a vertical height of approximately 8 cm.$hfill$[2011-0030]]]> 21 1 198 205 544 $-$45.3 ppm/K for the drive mode (bending) and $-$35.5 ppm/K for the sense mode (torsional). The temperature coefficient of sensitivity was determined to be $-$858 ppm/K in the case of operating with constant drive amplitude. This value could be reduced to 17.5 ppm/K over the full observed temperature range by adapting the drive amplitude as a linear function of temperature in an optimal way. The resolution limit of the sensor element was found to be about 0.08$^{circ}/ hbox{s}/sqrt{hbox{Hz}}$ at application-relevant ambient pressure levels ranging from 1 to 10 mbar.$hfill$[2009-0321]]]> 21 1 206 216 2307 $ times$ 8 array. The proposed sensing array shows considerable potential in pressure imagers, inertia switches, and other applications.$hfill$[2011-0129]]]> 21 1 217 223 1179 $muhbox{m}$ wide plate. Varying shapes of the buried fluidic channels (rectangular and ellipsoidal) are studied, resulting in mechanical resonance frequencies between 70 and 78 MHz. Volumes of those nanochannels range between 223 and 833 fL. For fluid-filled rectangular channels, quality factors demonstrate increasing values up to 4300 (compared with 3200 in air), when the resonance frequency $times$ quality factor product exceeds 300 GHz. An extensive study has been carried out with the rectangular channels, showing downward frequency shifts proportional to the liquid mass, as compared with dry devices. Experimental mass responsivities for all tested devices approached 1 $hbox{kHz.pg}^{-1}$. We examined different sources of dissipation, taking place in our structures, through qualitative analysis to explain the quality factor variations between dry and wet devices. We particularly focused on flow conditions, the vibration amplitude, the channel design and position relative to the resonator anchor, as well as the variation of acoustic energy dissipation within the embedded-channel cavities.$hfill$ [2011-0098]]]> 21 1 224 234 1279 $hfill$[2011-0091]]]> 21 1 235 242 1118 $^{circ}$ to 180 $^{circ}$ in 10$^{circ}$ steps with respect to the [110] direction, allowing the angular dependence of Young's modulus to be experimentally mapped out. The Young's modulus was measured to have a value of 170 GPa $pm$ 3 GPa at 0$^{circ}$ and 90$^{ circ}$ to the [110] direction and a value of 131 GPa $pm$ 3 GPa at $pm 40^{circ}$ and $pm 50^{circ}$ to the [110] direction. The measured values of Young's modulus and their angular dependence agree very well with the theoretical values that were recently reported, thereby experimentally verifying the theoretical calculations.$hfill$ [2011-0037]]]> 21 1 243 249 549 21 1 250 250 1585 21 1 251 251 887 21 1 252 252 5173 21 1 C3 C3 2520